Remote sensing, in situ monitoring and planktonic toxin ...

Remote sensing, in situ monitoring and planktonic toxin

vectors of harmful algal events in the optically complex

waters of the Galician rias (NW Spain)

Memoria presentada por Evangelos Spyrakos:

Licenciado en Ciencias del Mar (University of Aegean, Greece, 2005) MSc en Integrated Coastal Zone Management (University of Aegean, Greece, 2007)

MSc en Física Aplicada (Universidad de Vigo, España, 2009)

para optar al grado de Doctor por la Universidad de Vigo

Dirigida por:

Dr. Jesús M. Torres Palenzuela: Laboratorio de Teledetección y SIG

Física Aplicada Universidad de Vigo

Dra. África González Fernández: Área de Inmunologia Universidad de Vigo

Dr. Cástor Guisande: Departamento de Ecologia y Biologia animal

Universidad de Vigo & FAO Roma

Vigo, October 2011

El Dr. Jesús M. Torres Palenzuela, profesor titular del departamento de Física Aplicada de la Faultad de Ciencias de la Universidad de Vigo, la Dra África González Fernández, Directora del Centro de Investigaciones Biomédicas y Catedrática de Inmunología de la Universidad de Vigo y el Dr. Cástor Guisande González, Catedrático de Ecología de la Universidad de Vigo, INFORMAN: Que la presente memoria, titulada ¨Remote sensing, in situ monitoring and planktonic toxin

vectors of harmful algal events in the optically complex waters of the Galician rias (NW

Spain)¨, presentada por D. Evangelos Spyrakos para optar al grado de Doctor por la

Universidad de Vigo, ha sido realizada bajo su dirección y reúne los requisitos necesarios

para ser defendida ante el tribunal calificador. Por tanto, autorizan su presentación ante el

Consejo de Departamento y la Comisión de Doctorado.

Y para que conste a los efectos oportunos, firman en la presente certificación

En Vigo a 25 de Octubre 2011

Dr. Jesús M. Torres Palenzuela Dra. África González Fernández Dr. Cástor Guisande González

This PhD project was supported by European Commission's Marie Curie Actions

through a grant within the project 20501 ECOsystem approach to Sustainable Management

of the Marine Environment and its living Resources (ECOSUMMER). MERIS data were

obtained through EUROPEAN SPACE AGENCY/ENVISAT project AO-623. A part of the

chlorophyll data base was provided by the Technological Institute for the Control of the

Marine Environment of Galicia.

i

Content

Summary ............................................................................................................................... iii

Summary in spanish .............................................................................................................. vi

Graphical abstract ................................................................................................................ ix

List of tables ........................................................................................................................... xi

List of figures ....................................................................................................................... xiii

List of abbreviations ............................................................................................................ xv

Acknowledgments .............................................................................................................. xvii

Dissemination of and development from work performed for this thesis ...................... xix

Thesis outline .......................................................................................................................... 1

CHAPTER I ▐

5

Overall introduction

9 1.1 Satellite remote sensing of coastal

optically complex waters

12 1.2 ELISA for DA detection 14 1.3 Harmful algae interactions with marine

planktonic grazers

15

17

1.4 Study area: Galician rias

1.5 Motivations and thesis objectives

21 1.6 References

Development of regionally specific

chlorophyll a algorithms of optically complex

waters for MERIS full resolution data

CHAPTER II ▐

29

Abstract 31

2.1 Introduction 32

2.2 Material and methods 34

2.3 Results 44

2.4 Discussion 52

2.5 References 57

CHAPTER III ▐

63

Application of a regionally specific

chlorophyll a algorithm for MERIS full

resolution data during an upwelling cycle

65 Abstract

66 3.1 Introduction

68 3.2 Material and methods

74 3.3 Results & Discussion

88 3.4 References

Remote sensing, in situ monitoring and

environmental perspectives of toxic Pseudo-

nitzschia events in the surface waters of two

Galician rias

CHAPTER IV ▐

95

Abstract 97

4.1 Introduction 98

ii

4.2 Material and methods 100

4.3 Results 106

4.4 Discussion 116

4.5 References 122

CHAPTER V ▐

129

Ingestion rates of the heterotrophic

dinoflagellate Noctiluca scintillans fed on the

toxic dinoflagellate Alexandrium minutum

(Halim)

131 Abstract

132 5.1 Introduction

133 5.2 Material and methods

137 5.3 Results

141 5.4 Discussion

144 5.5 References

Modelling PST transfer and accumulation in

two planktonic grazers

CHAPTER VI ▐

149

Abstract 151

6.1 Introduction 152

6.2 Methods 154

6.3 Results 163

6.4 Discussion 171

6.5 References 174

CHAPTER VII

181 General discussion & further considerations

183

184

186

187

188

189

7.1 Development of regional specific chla

algorithms from MERIS FR data for optically

complex waters

7.2 Remote sensing chla mapping during an

upwelling cycle

7.3 Harmful Pseudo-nitzschia spp. events in

the surface waters of two Galician rias

7.4 Planktonic grazing on a PST-producer

dinoflagellate

7.5 General discussion to all chapters

7.6 References

Conclusions ........................................................................................................................ 191

Annex I ............................................................................................................................... 195

Annex II............................................................................................................................... 211

iii

Summary

Transient increases of phytoplankton abundance, referred to as blooms, are a

common and recurrent phenomenon in many coastal areas around the world including the

upwelling estuarine systems of the Galician rias (embayments in North-West Spain). Harmful

algal events due to toxic phytoplankton species and/or high-biomass blooms pose an

increasing threat to aquaculture and fishing industries, ecosystem health and diversity and

have possible implications to human health and activities. A multidisciplinary approach was

followed in this PhD thesis in order to detect and study harmful algal events and their

driving forces in the waters of the Galician rias and explore the possible toxin planktonic

vectors and the fate of the toxins in the marine system.

A set of neural network-based chlorophyll a (chla) algorithms for Medium

Resolution Imaging Spectrometer (MERIS) full resolution (FR) data using in situ data sets

and fuzzy clustering techniques was developed specifically for the optically complex waters

(defined as Case 2 waters) of the Galician rias. Three clusters were determined, which

represent the different structures found in the data base. Three different neural networks

(NNs) were developed: one including the whole data set, and two others using only data

points belonging to one of the clusters. The fitting results were fairly good and proved the

capability of the tools to predict chla concentrations in the study area. The best prediction

was given by the NN trained with high-quality data using the most abundant cluster data

set. The NNs developed in this study detected accurately the peaks of chla, in both training

and validation sets, outperforming the algorithms that are routinely used for MERIS data in

Case 2 waters. As a continuation, this thesis takes advantage of these regionally specific

algorithms and the characteristics of MERIS in order to deliver more accurate and detailed

chla maps of optically complex coastal waters during an upwelling cycle and harmful algal

events. The main changes in chla concentration and distribution were clearly captured in

the images. There was a significant variation in the timing and the extent of the maximum

chla areas. The maps confirmed that the spatial structure of the phytoplankton distribution

in the rias Baixas is complex and it is affected by the surface currents and winds on the

adjacent continental shelf. Relatively high biomass “patches” were mapped in detail inside

the rias.

During the sampling campaigns (2007-2009) that were carried out in the study area,

toxigenic events due to Pseudo-nitzshia spp. and high abundances of the harmful

dinoflagellate red Noctiluca scintillans were recorded.

iv

A combined analysis of satellite imagery data, measurements of biotic and abiotic

parameters and mixed effects modelling was used to study Pseudo-nitzschia blooms in the

surface waters of two of the rias. Domoic acid (DA) concentrations from natural Pseudo-

nitzshia populations in the Galician rias were for the first time, to our knowledge, measured

in the study area (below detection limits-2.5 μg L-1). Two DA outbreaks were recorded in the

area. P. australis was the dominant Pseudo-nitzschia species during the blooms. The

application of a regional algorithm in combination with the characteristics of MERIS FR

allowed for accurate mapping of chla and the detection of small, high in chla and Pseudo-

nitzschia spp. “patches” in the rias. The optimal model for the Pseudo-nitzschia spp.

abundance and DA concentration suggested the significant effect of some macronutrients

as well as other abiotic and biotic parameters, approximating in that way the potential

environmental causes and effects of the harmful Pseudo-nitzschia spp. blooms in the area.

In the case of N. scintillans, speciments of this species were collected from the rias

and were established in laboratory cultures. This species was fed with the toxic (Paralytic

Shellfish Poisoning or PSP) dinoflagellate Alexandrium minutum in order to evaluate its

ingestion and clearance rates and test for toxin accumulated in the individuals. N. scintillans

actively fed on A. minutum showing no satiated feeding. No detectable amounts of toxins in

the individuals revealed from the High Performance Liquid Chromatography (HPLC) toxin

analysis. In order to calculate detoxification rates of N. scintillans a new experiment this

time performed using a different strain and species of Alexandrium, A. catenatum which

was characterised by higher levels of toxin content. N. scintillans showed relatively high

detoxification rates (-0.17 pg toxin ind-1 h-1). Using the results of these two experiments in

combination to previously published data, dynamic models were developed in order to

study the PST transfer and accumulation in two planktonic organisms, namely the copepod

Acartia clausi and and the heterotrophic dinoflagellate N. scintillans.

This thesis shows the capability of NN models to predict chlorophyll concentrations

on the Galician coast from MERIS images following the widespread understanding of the

need for regionally specific models. According to the recorded in situ data, the model

presented here is an improvement on other previously used techniques, and made it

possible to obtain reliable chlorophyll maps using almost every image. These maps were

used to study the evolution of local oceanographic processes, which in turn were related to

the development of algal blooms in the area. The present study allows more detailed

examination of the chla distribution and detection of high biomass areas in the Galician rias

and the adjacent area and should be an integral part of the monitoring programs.

Moreover, the results of this study deduce that toxic events due to DA should be an

v

important concern and therefore DA in natural phytoplankton populations should be

measured routinely in order to assess the potential of a DA outbreak. This is the first report

about grazing rates of N. scintillans on Alexandrium cells. In addition, Noctiluca may inflict

grazing pressure on the growth of PST species in the field, and could therefore play an

important role as a regulator against PST-producing phytoplankton. The dynamic model

showed that ingestion of toxic dinoflagellates by different types of planktonic organisms

may be important for the dynamics of the toxins in the food web. However, both organisms

illustrated a rapid (50 h) reduction of ingested toxin suggesting inefficiency to transfer

toxins through predation in the food web.

vi

Resumen

Los incrementos en la abundancia de fitoplancton, conocidos como floraciones

algales, son un fenómeno común y periódico en muchas zonas costeras de todo el mundo,

incluyendo las Rías Gallegas, sistemas estuáricos localizados al noroeste de España en una

zona de afloramiento marino. Algunas de estas floraciones algales están consideradas

como nocivas y suponen una amenaza creciente para la acuicultura, la industria pesquera, la

diversidad del ecosistema y a menudo tienen posibles implicaciones en la salud humana y

en sus actividades.

En esta tesis se ha seguido un enfoque multidisciplinar para detectar y estudiar los

eventos de algas nocivas y sus posibles causas en las aguas de las Rías Gallegas y explorar

posibles vectores de biotoxinas en el sistema marino. Se han utilizado para este trabajo los

datos procedentes del sensor Medium Resolution Imaging Spectrometer (MERIS) del

programa Environmental Satellite (ENVISAT) dentro del proyecto AO-623, datos de clorofila

procedentes del programa de monitorización de la calidad de las aguas llevado a cabo por

el Instituto Tecnológico para o control do medio mariño de Galicia (INTECMAR), asi como

datos obtenidos por la Universidad de Vigo dentro del programa de Marie Curie

ECOSUMMER realizado en los años 2007 hasta 2009 en las Rías Baixas Gallegas. Para estas

aguas ópticamente complejas (aguas de Caso 2), fue desarrollado específicamente un

conjunto de algoritmos de índice de clorofila a para datos de MERIS FR basados en la

utilización de redes neuronales artificiales (RNA), técnicas de fuzzy clustering y datos in-situ.

Se determinaron tres categorías de reflectancias en los datos procedentes del sensor

MERIS, las cuales representan diferentes estructuras. Se desarrollaron tres algoritmos: uno

que utiliza todo el conjunto de datos, y otros dos que incluyen puntos pertenecientes a

cada una de las categorías citadas. Los parámetros estadísticos entre la concentración de

clorofila calculada mediante el modelo desarrollado en este estudio y la estimada a partir de

muestra del campo fueron bastante buenos y demostraron la capacidad de estas

herramientas para predecir las concentraciones de clorofila a en el área de estudio. La

mejor predicción se obtuvo con la RNA entrenada con datos de alta calidad mediante el

conjunto de datos del clúster más abundante. El algoritmo desarrollado en este estudio

detecta con alta precisión los valores máximos de clorofila tanto en el conjunto de

entrenamiento como en el de validación, superando los algoritmos para datos MERIS que

rutinariamente utiliza la ESA (Agencia Europea del Espacio) en aguas de caso 2.

vii

A continuación, esta tesis utiliza estos algoritmos desarrollados específicamente

para las aguas de las Rías Gallegas junto a las características radiométricas y espaciales del

sensor MERIS para producir mapas de clorofila más precisos y detallados durante un ciclo

de afloramiento con eventos de floraciones algales. Los principales cambios en la

distribución y concentración de clorofila a fueron claramente detectados en los mapas de

clorofila obtenidos a partir de las imágenes de MERIS. Hubo una variación significativa en la

distribución en el tiempo y en la extensión de las zonas de máximos índices de clorofila

detectados. Los mapas confirman que la estructura espacial de la distribución de

fitoplancton en las Rías Baixas es compleja y que se ve afectada por las corrientes

superficiales y los vientos de la plataforma continental. Dentro de las Rías fueron

detectadas agregaciones de alta biomasa de fitoplancton (“patches”) utilizando los datos

del sensor de color de la ESA.

Durante las campañas en el área de estudio (2007-2009), realizadas por la

Universidad de Vigo, se registraron eventos toxigénicos debidos a la diatomea Pseudo-

nitzschia spp. y también elevadas abundancias del dinoflagelado Noctiluca scintillans en su

forma roja.

Se utilizaron datos derivados de imágenes de satélite, medidas de parámetros

bióticos y abióticos en la columna de agua y modelos de efectos mixtos para estudiar las

floraciones de Pseudo-nitzschia en las aguas superficiales de dos Rías Gallegas (Vigo y

Pontevedra). En estas campañas fueron medidas por primera vez concentraciones de ácido

domoico (AD) de poblaciones naturales de Pseudo-nitzschia en el área de estudio (0-2.5 μg

L-1). Se registraron en la zona dos eventos de elevada concentración de AD, siento la P.

australis la especie dominante durante la floración. La aplicación de un algoritmo regional

en combinación con las características de MERIS FR permitió obtener mapas de clorofila a

de alta precisión y detectar pequeñas zonas de alta concentración superficial de clorofila a

en presencia de altas concentraciones de Pseudo-nitzschia spp. en las Rías. Los modelos

óptimos de efecto mixto (GAMM) diseñados para predicir las abundancias de Pseudo-

nitzschia y de GLMM para la concentración de AD utilizados en este trabajo, incluyen el

efecto significativo de algunos macronutrientes, así como otros parámetros abióticos y

bióticos. De esta manera se podrían explicar las potenciales causas ambientales y conocer

los efectos perjudiciales de las floraciones nocivas de Pseudo-nitzschia spp. en la zona.

En el caso de N. scintillans, fueron recolectados ejemplares de esta especie en la Ría

de Vigo y posteriormente cultivados en laboratorio. Finalmente, este cultivo fue alimentado

con la microalga Alexandrium minutum, productora de la toxina paralizante (Paralytic

Shellfish Poisoning o PSP) para evaluar sus tasas de ingestión y para testar la toxina

viii

acumulada en los individuos. N. scintillans se alimentó activamente de A. minutum durante

el periodo evaluado con tasas de ingestión máximas de 0.3 μg C ind-1 día-1. Los análisis de los

perfiles cromatográficos de N. scintillans expuesto a A. minutum revelaron que esta especie

no acumuló las toxinas paralizantes en su citoplasma. Para calcular las tasas de

detoxificación de PST de N. scintillans se realizó un nuevo experimento utilizando esta vez

una cepa de A. catenatum que se caracteriza por niveles más altos de contenido de toxina.

En este caso se observaron tasas relativamente altas de detoxificación (-0.17 pg toxina ind-1

h-1). Estos resultados junto con datos previamente publicados, se utilizaron para el

desarrollo de modelos dinámicos con el objetivo de estudiar la transferencia y la

acumulación de PST en dos vectores planctónicos.

En conclusión, esta tesis muestra la capacidad de las RNA para predecir las

concentraciones de clorofila en la costa gallega a partir de imágenes de MERIS mostrando

la necesidad de utilizar modelos específicos para cada región. Comparando nuestros

resultados con datos de clorofila a tomados in situ, se observa que el modelo desarrollado

representa una mejora con respecto a otras técnicas utilizadas anteriormente, y ha

permitido obtener mapas precisos de clorofila a partir de las imágenes utilizadas. Estos

mapas fueron utilizados para estudiar la evolución de los procesos oceanográficos en el

área, que a su vez están relacionados con el desarrollo de floraciones de algas. Los

resultados permiten estudiar de forma más detallada la distribución de clorofila a y la

detección de áreas de alta biomasa en las Rías Gallegas y en la plataforma adyacente. Este

tipo de técnicas y modelos desarrollados específicamente para el estudio de la clorofila en

estas aguas, han mostrado ser una herramienta de gran utilidad en la detección y

seguimiento de la dinámica de proliferaciones algales pudiendo ser integrados en los

programas de monitoreo de fitoplancton. Los resultados de los niveles de AD en las aguas

de las Rías indican que la presencia de esta toxina debería determinanse no solo en bivalvos

sino también en poblaciones naturales de fitoplancton. Debería medirse de manera

rutinaria para evaluar el potencial toxico de un evento de floración algal de Pseudo-nitzschia

spp. En esta tesis de doctorado se ha documentado por primera vez el consumo de

dinoflagelados tóxicos en condiciones de laboratorio por parte de N. scintillans. Además, se

deduce de este trabajo, que la especie Noctiluca podría desempeñar un papel importante

controlando el fitoplacton que produce toxinas del tipo PST. El modelo dinámico

desarrollado, mostró que la ingestión de dinoflagelados tóxicos por diferentes tipos de

organismos planctónicos, puede ser importante para el destino de las toxinas en la cadena

trófica.

ix

Graphical abstract

xi

List of tables

CHAPTER I

1.1 ELISA kits commercially available for the detection and qualification of marine toxins

produced by microalgae ........................................................................................................ 13

CHAPTER II

2.1

2.2

2.3

2.4

2.5

Data sets used in this study for the periods 2002–2004 and 2006–2008 from the rias Baixas

................................................................................................................................................ 36

Summary of clustering results applying the FCM algorithm ................................................. 39

Number and characteristics of data points belonging to each cluster ..................................... 41

Percentage of pixels belonging to each cluster over the study area, obtained from

classification images derived from the MERIS images used in this study ............................. 48

Summary of the validation parameters computed using the training and validation data sets

for each one of the three different NNs .................................................................................. 48

CHAPTER III

3.1

3.2

3.3

3.4

MERIS imagery showing the acquisition time (UTC) and mean view zenith angle from west

and sea-truthing mean values of chla, SPM percentage of inorganic contribution to SPM and

Zsd for ria de Vigo (12 stations) during the two samplings ................................................... 71

Percentage of pixels belonging to each cluster over the study area, obtained from

classification images derived from the MERIS images ......................................................... 77

Performance parameters for the chla neural networks tested in this study ............................ 77

Dominant atmospheric and oceanographic conditions off the rias Baixas categorized as three

different states during the upwelling cycle in summer 2008 .................................................. 79

CHAPTER IV

4.1

4.2

4.3

4.4

4.5

4.6

Distribution of MERIS FR imagery used in this study for the years 2007-2009 showing the

acquisition time, sky conditions and mean view zenith angle from west ............................. 104

The three periods considered during the study in relation to the month and year of the

campaigns and the dominant meteorological conditions off the rias ................................... 106

List of the environmental parameters measured in this study combined with their in-situ

mean values during the 3 periods ......................................................................................... 107

Pearson´s correlations between Pseudo-nitzschia spp. abundances, particulate DA and

cellular DA concentrations with several environmental parameters during October 10, 2007

and July 14, 2009 ................................................................................................................. 110

Percentage of pixels belonging to each cluster over the study area ..................................... 112

Results of the generalised additive models, which were used to model Pseudo-nitzschia

abundance ............................................................................................................................. 115

xii

4.7

Results of the generalised linear model that shows the effect of selected abiotic and biotic

parameters on particulate domoic acid ................................................................................. 116

CHAPTER V

5.1

5.2

5.3

Cell parameters for Alexandrium minutum measured during the experiment ...................... 134

Specific toxin composition of gonyautoxins as a percentage, total toxin per cell and total cell

toxicity of Alexandrium minutum and Noctiluca scintillans during the experimental period

.............................................................................................................................................. 137

Mann-Whitney tests were carried out with ingestion rate as the dependent variable and

Alexandrium minutum abundance as a factor ....................................................................... 141

CHAPTER VI

6.1

6.2

6.3

6.4

6.5

6.6

6.7

Model parameters for nutrient uptake, population growth and toxin synthesis dynamics of

Alexandrium ......................................................................................................................... 155

Model parameters for vector population growth, toxin assimilation and detoxification for

Noctiluca scintillans and Acartia clausi ............................................................................... 156

Model parameters tested for model output sensitivity ......................................................... 161

Cell parameters for Alexandrium catenella and Prorocentrum micans measured during the

experiment ............................................................................................................................ 162

Specific toxin composition of gonyautoxins as a percentage, total toxin per cell and total cell

toxicity of Alexandrium catenella during the experiment .................................................... 163

Coefficient of variation and standard deviation from the sensitivity analysis results for

Noctiluca scintillans and Acartia clausi model with modified initial values of the

parameters ............................................................................................................................ 165

First order sensitivity indices showing the contribution and ranking of parameters to the

overall incertainty of the models output ............................................................................... 169

xiii

List of figures

CHAPTER I

1.1

1.2

1.3

1.4

Schematic for the general approach followed in this study and aspects highlighted in the

introduction ..................................................................................................................................... 9

Principal constituents in oceanic and coastal waters .............................................................. 11

Triangular diagram showing the main water optical classes .................................................. 12

Map of the study area ............................................................................................................. 16

CHAPTER II

2.1

2.2

2.3

2.4

2.5

2.6

2.7

2.8

2.9

2.10

2.11

2.12

Map of the study area and location of sampling stations ....................................................... 35

In situ chla distribution in the 4 rias ....................................................................................... 36

MERIS FR imagery set .......................................................................................................... 37

Basic architecture of the MLP network .................................................................................. 42

Mean MERIS reflectance spectra for each cluster obtained using the 3-class FCM algorithm

................................................................................................................................................ 45

RGB composite of a MERIS FR over the study area ............................................................. 46

Classification of a MERIS image and chla concentration map from the study area

(17/11/2003) ........................................................................................................................... 47

Classification of a MERIS image (15/06/2004) from the study area ..................................... 47

Linear relationship between predicted and observed concentrations of chla, considering only

the data points included in the validation set, for the three different neural networks ........... 49

Observed and predicted chla concentrations for each data point in the training and validation

sets and for the three networks ............................................................................................... 50

Scatter plots of in situ chla concentrations for MERIS C2R Processor and NNRB#3 .......... 51

Average in situ chla concentrations for all sites calculated by the NNRB#3 and C2R in the

study area for different days ................................................................................................... 52

CHAPTER III

3.1

3.2

3.3

3.4

3.5

3.6

3.7

3.8

Map of the study area and location of sampling stations ......................................................... 69

Cha fluorescence vertical profiles in ria de vigo on july 09 2008 and july 22 2008 ................ 75

Regression analysis between tsm and chla ............................................................................... 76

Classification of MERIS images from the study area .............................................................. 78

Daily upwelling index, surface currents and daily average of SST off the rias Baixas during

July 2008 .................................................................................................................................. 80

Maps of MODIS SST from the study area during July 2008 ................................................... 82

Maps of MERIS chla from the study area during July 2008 .................................................... 83

RGB MERIS FR composite image over the study area (22/07/2008) .................................... 87

xiv

CHAPTER IV

4.1

4.2

4.3

4.4

4.5

4.6

Map of the study area and location of sampling stations ..................................................... 100

Daily upwelling index off the rias Baixas for the years 2007-2009 .................................... 106

Scanning electron micrographs of Pseudo-nitzschia species ............................................... 108

Distribution of pDA during the two DA outbreaks recorded in the surface waters of ria de

Vigo on October 19, 2007 and July 14, 2009 ....................................................................... 111

Maps of MERIS chla from rias Baixas and adjasted coastal waters .................................... 114

GAMM smoothing curve for Pseudo-nitzschia spp. abundance .......................................... 115

CHAPTER V

5.1

5.2

5.3

Noctiluca scintillans with ingested Alexandrium minutum cells .......................................... 138

Noctiluca scintillans survival rates as a function of Alexandrium minutum abundance for the

4-day incubation period ........................................................................................................ 139

Ingestion and clearance rates of Noctiluca scintillans as a function of Alexandrium minutum

abundance for the 4-day incubation period ingestion and clearance rates ........................... 140

CHAPTER VI

6.1

6.2

6.3

6.4

6.5

6.6

6.7

6.8

Conceptual schematic representation of the dynamic model ............................................... 155

Toxicity retained in Noctiluca scintillans as function of the elapsed starvation time .......... 164

Sensitivity analysis results showing the mean and percentiles of simulated PST accumulated

in Noctiluca scintillans population from: 200 simulations with randomly modified initial

values of phosphorus concentration, non-toxic food concentration, Alexandrium abundance

and Noctiluca abundance and 800 simulations using all the above parameters ................... 166

Sensitivity analysis results showing the mean and percentiles of simulated PST accumulated

in Acartia clausi population from: 200 simulations with randomly modified initial values of

phosphorus concentration, non-toxic food concentration, Alexandrium abundance and

Acartia abundance and 800 simulations using all the above parameters ............................. 167

Contribution of each parameter to the variability of Noctiluca model output as indicated by

Sobol' index (SI) ................................................................................................................... 168

Contribution of each parameter to the variability of Acartia model output as indicated by

Sobol' index (SI) ................................................................................................................... 168

Scatter plot matrix of all pair combinations of “stock” dynamic variables for Noctiluca. Note

that all units in the graph refer to log-transformed values. ................................................... 170

Scatter plot matrix of all pair combinations of “stock”dynamic variables for Acartia. Note

that all units in the graph refer to log-transformed values. ................................................... 170

xv

List of abbreviations and acronyms

The following table describes the significance of various abbreviations and

acronyms used throughout the thesis.

Abs antibodies

ABTS 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid)

AC Acartia clausi

AOPs apparent optical properties

ASP amnesic shellfish poisoning

cDA cellular domoic acid

C2R case-2-regional

cELISA competitive enzyme-linked immunosorbent assay

CASI compact airborne spectographic imager

CDOM coloured dissolved organic material

chla chlorophyll a

CZCS coastal zone color scanner

DA domoic acid

DAB 3,3' Diaminobenzidine tetrahydrochloride

DOC dissolved organic carbon

DSP diarrhetic shellfish poisoning

ELISA enzyme-linked immunosorbent assay

ENACW eastern north Atlantic central water

ESA European space agency

FCM fuzzy c-mean

FR full resolution

GAMM generalized additive mixed models

GLMM generalized linear mixed models

GTX gonyautoxin

HA harmful algae

HAB harmful algal bloom

HB high biomass

HPLC high performance liquid chromatography

IOPs inherent optical properties

ISM inorganic suspended matter

xvi

l2

LM

level 2

light microscopy

MERIS medium resolution imaging spectrometer

MLP multilayer perceptron

MODIS moderate resolution imaging spectroradiometer

NASA-GSFC national aeronautics and space administration Goddard space flight center

NS Noctiluca scintillans

NSP neurotoxic shellfish poisoning

OC4v4 ocean color 4 version 4

OM organic matter

pDA particulate domoic acid

POC particulate organic carbon

PON particulate organic nitrogen

PSP paralytic shellfish poisoning

PST paralytic shellfish toxin

RLw water leaving radiance reflectance

SeaWiFS sea-viewing wide-field-view sensor

SEM scanning electron microscopy

SPM suspended particulate matter

SST sea surface temperature

STXeq saxitoxin equivalents

TMB 3,3',5,5'-Tetramethylbenzidine

TOA top of atmosphere

TSM total suspended material

WIB western Iberian Peninsula

xvii

Acknowledgments

This thesis would not have been possible without the guidance and support of many

generous and knowledgeable people. First I offer my sincerest gratitude to my main

supervisor Jesus M. Torres Palenzuela for all his continuous support, guidance and

encouragement. You are not only my academic supervisor but also a real mentor and a

friend. One simply could not wish for a better and more helpful supervisor. I equally wish to

express my profound thanks to my supervisors África González Fernández and Cástor

Guisande for their generous and endless advice and encouragement. They were always

available for me to consult on my thesis and I needed to frequently.

It would not have been possible to write this doctoral thesis without the help and

support of the colleagues and friends from the three laboratories I worked in the University

of Vigo. Beginning with the Laboratory of Remote Sensing and GIS I am particularly thankful

to Luis González Vilas for his collaboration in the Chapters II and III, where he added his

insightfulness and his advanced computational skills and for the free tickets he was always

supplied me for the Celta matches. Also to my other colleagues from this lab: Angela,

Marta, Nina, Yolanda, Carmen, Nerea, Leo and Miguel thank you for all the endless help (in

the lab or/and in the field) I received over the last four years and for the coffee/lunch break

discussions. I am also grateful to two Professors from the Applied Physics department: José

Luis Legido Soto for being always there to help me when I had crucial administrative

questions or problems and Gabriel Rosón Porto for inviting me to the campaigns in

Pontevedra and for providing valuable references and information about the dynamics of

the rias. I am most grateful to my Patagonian φίλος Maximo Frangópoulos and Aldo

Barreiro of the Zooplankton Laboratory for teaching so much about phytoplankton ecology

and biology, for their guidance (even during the last year(s) they lived abroad) throughout

the this thesis and for their contribution (writing, revising, analysing data and planning

experiments) in Chapters V and VI. I feel really fortunate to work with them and I will never

forget their support and the good times we had in the lab. I also wish to acknowledge the

wonderful assistance and support given to me by Alberto Acuña in the preparation of the

sampling campaigns, on-board and after with the analysis of the samples. Your help was

essential. I wish to express my profound thanks to the other members of this lab Isa and

Alba. Similarly, I appreciate and acknowledge Daniel Pérez for welcoming me to the

University, helping me with the sampling campaigns, spending an unbelievable amount of

hours in the lab making me familiar with the immunological methods and for always being

xviii

there every time I needed scientific advice. Ευχαριστώ Daniel. Special thanks to Elina Garet

with whom I analysed all my seawater samples for domoic acid with ELISA and to my other

colleagues from the Immunology Lab: Andréa, Belén, Silvia, Rosana, Tamara, Mercedes,

Leonardo, Jose Faro, Juan, Bernardo and my spear fishing-squash buddy Cristian.

I am grateful to Eric Desmond Barton of the Instituto Investigaciones Marinas for

spending his valuable time on my drafts of Chapter III and for ceaselessly providing me with

feedback and discussion. I also wish to acknowledge Ronald Doerffer (GKSS Center) and

Malte Elbraechter (Alfred Wegener Institute) for the advice in my first years of the PhD,

which was critical for me to accomplish this dissertation. I thank Graham Pierce from the

University of Aberdeen for organising high quality training courses where I found answers

in many scientific questions and for persistently asking about my progress. Javier Pérez

Marrero (Instituto Canario de Ciencias Marinas) and the rest of the participants in BLOOM

2008 campaign in Argentina, thank you for extending my knowledge on operational

oceanography and ocean colour techniques and I am sorry I have not included any of this

campaign data in the thesis. I will forever be thankful to my former academic and research

supervisors from the University of Aegean Drosos Koutsoubas and Athanasios

Evagelopoulos. They were the reason why I decided to pursue a career in research, their

enthusiasm and love for lecturing is contagious.

A thesis cannot be produced without substantial financial support and I would like

to thank the European Commission's Marie Curie Actions (project 20501 ECOsystem

approach to Sustainable Management of the Marine Environment and its living Resources

[ECOSUMMER]) for providing the grant which gave me the financial capacity to devote

more than two years to this research and to carry out the extensive field and lab work.

MERIS data were obtained through EUROPEAN SPACE AGENCY/ENVISAT project AO-623. I

would like to thank the Technological Institute for the Control of the Marine Environment

of Galicia, for providing me with chlorophyll data.

I deeply appreciate my multinational group of friends Mateus, Marcelo, Pepe,

Consuelo, Fiona, Sabine, Sara, Rebeca, Sonia, Ruth, Carlos, Lisa, Jose who made my life in

Vigo easier, fun and enjoyable.

To Adeline, for all the help in the manuscript correction and for the patience during

the last years: A grá, go raibh míle maith agat.

Πϊνω απ’όλα θα όθελα να ευχαριςτόςω τουσ γονεύσ μου Τούλα και Κώστα και

τον αδερφό μου Γιάννη που με ςτόριξαν όςο κανϋνασ ϊλλοσ ηθικϊ, οικονομικϊ και

ψυχολογικϊ ςε όλα τα ςτϊδια των ςπουδών και τησ παρούςασ διατριβόσ την οπούα και

τουσ αφιερώνω.

xix

Dissemination of and development from work performed for this thesis

Much of the research carried out in this thesis has been published/submitted (or is in

preparation to be submitted) in appropriate peer-reviewed international journals.

1. Spyrakos, E., González Vilas, L., Torres Palenzuela, J., & Barton E., D. (2011). Remote

sensing chlorophyll a of optically waters (rías Baixas, NW Spain): Application of a

regionally specific chlorophyll a algorithm for MERIS full resolution data during an

upwelling cycle. Remote Sensing of Environment 115(10), 2471-2485.

2. González Vilas, L., Spyrakos, E., & Torres Palenzuela, J. (2011). Neural network

estimation of chlorophyll a from MERIS full resolution data for the coastal waters of

Galician rias (NW Spain). Remote Sensing of Environment 115 (2), 524-535.

3. Frangópulos, M., Spyrakos E.1, & Guisande, C. (2011). Ingestion rates of the

heterotrophic dinoflagellate Noctiluca scintillans fed on the toxic dinoflagellate

Alexandrium minutum (Halim). Harmful Algae 10, 304-309.

4. Spyrakos, E., Frangópulos, M., Barreiro, A., & Guisande, C. Modelling PST transfer and

accumulation in two planktonic grazers.

5. Spyrakos, E., Garet E., González-Fernández Á., & Torres Palenzuela, J. Remote sensing,

in situ monitoring and environmental perspectives of toxic Pseudo-nitzschia events in

the surface waters of two Galician rias (NW Spain). Submitted to Harmful Algae (HARAL-

S-00284).

The work has also been published and provisionally accepted to be published as book

chapters in edited reviewed books.

1. Lorenzo-Abalde, S. Calvo, J. Spyrakos, E., & González-Fernández, Á. (2011). Immunology

applied in marine sciences (In Spanish). In Methods and techniques in marine research

(pp. 209-222).

2. Spyrakos, E., González Vilas, L., & Torres Palenzuela, J. Remote sensing of coastal

optically complex waters. In Y. Chemin (Ed), Remote Sensing. ISBN 979-953-307-541-8.

Part of the results of this thesis has been advertised on the European Space Agency´s

web page.

1 Shared authorship

xx

1. Chlorophyll a estimation in coastal waters from MERIS data. http://eopi.esa.int/esa/esa.

The work has also been presented at national and international conferences and

published in their proceeding books:

1. E. Spyrakos, L. González Vilas, J. Torres Palenzuela & E. D. Barton: Application of

regionally specific chla algorithms from MERIS FR data during an upwelling cycle (In

Spanish). XIV Congreso de la Asociación Española de Teledetección, Mieres, España, 21-

23 Septiembre 2011. (ORAL)

2. E. Spyrakos, L. González Vilas, J. Torres Palenzuela & N. Yarovenko: Chlorophyll a

mapping of optically complex coastal waters using regionally specific neural network-

based algorithms for MERIS full resolution data. 5th Workshop on Remote Sensing of

the Coastal Zone, Prague, Czech Republic, 1-3 June 2011. (ORAL)

3. E. Spyrakos, L. González Vilas, J. Torres Palenzuela & E. Barton: Development and

application during an upwelling cycle of neural network-based chlorophyll a algorithms

from MERIS full-resolution data for optically complex waters. 14th International

Conference on Harmful Algae, Creta, Greece, 1-5 November 2010. (POSTER)

4. E. Spyrakos, M. Frangópulos & C. Guisande: Rapid reduction of the PST content in the

heterotrophic dinoflagellate Noctiluca scintillans fed on a toxic dinoflagellate. 14th

International Conference on Harmful Algae, Creta, Greece, 1-5 November 2010.

(POSTER)

5. M. Frangópulos, E. Spyrakos & C. Guisande: Ingestion rates of the heterotrophic

dinoflagellate Noctiluca scintillans fed on the toxic dinoflagellate Alexandrium minutum

(Halim). 14th International Conference on Harmful Algae, Creta, Greece, 1-5 November

2010. (POSTER)

6. J. Torres Palenzuela, L. González Vilas & E. Spyrakos: Artificial Neural Network model

for predicting Pseudo-nitzschia spp. abundance in the Galician Rias (NW Spain). 14th

International Conference on Harmful Algae, Creta, Greece, 1-5 November 2010.

(POSTER)

7. M. Frangópulos, E. Spyrakos, S. Moraga & C. Guisande: Ingestion rates of Noctiluca fed

with the toxic microalgae Alexandrium minutum (In Spanish). XXX Congreso de Ciencias

del mar. Concepción, Chile, 19-22 October 2010. (POSTER)

8. E. Spyrakos, L. González Vilas, J. Torres Palenzuela & E. Barton: Development and

application of Neural network-based chlorophyll a algorithms from MERIS FR data for

optically complex waters. From the Sea-bed to the Cloud-tops, Remote Sensing and

http://eopi.esa.int/esa/esa

http://www.conferences.earsel.org/conferences/28

http://www.conferences.earsel.org/conferences/28

xxi

photogrammetry society annual conference with Irish Earth Observation Symposium,

Cork, Ireland, 1-3 September 2010. (ORAL)

9. E. Spyrakos, L. González Vilas, D. Pérez, M. Frangópulos, C. Guisande, A. González

Fernández & J. Torres Palenzuela: Harmful algal events in the Galician rias (NW Spain).

Marine EcoSystems and Sustainability Conference, University of Aberdeen, Scotland,

U.K. 9-11 December 2009. (ORAL)

10. J. Torres Palenzuela, L. González Vilas, E. Spyrakos, M. Darriba Estévez & N. Yarovenko:

Applications of neural network technology and remote sensing in Oceanography.

Marine EcoSystems and Sustainability Conference, University of Aberdeen, Scotland,

U.K. 9-11 December 2009. (ORAL)

11. J. Torres Palenzuela, L. González Vilas & E. Spyrakos: Determination of chla index for

Case II waters in the Galician coast using MERIS FR (in Spanish). XIII Congresso de la

Asociación Española de Teledetección, Calatayud, 23-26 September 2009. (ORAL)

12. E. Spyrakos, L. González Vilas & J. Torres Palenzuela. Development and validation of

chlorophyll a algorithms for MERIS full-resolution data in Rias Baixas coastal waters

(NW Spain. Second International Symposium of Marine Sciences, Vigo, 27-30 Abril 2009.

(ORAL)

13. E. Spyrakos, J. Torres Palenzuela, D. Pérez, M. Darriba Estévez & C. Guisande: In situ

and remotely sensed measurements of optical properties in coastal waters (ria of Vigo,

Spain. Second International Symposium of Marine Sciences, Vigo, 27-30 Abril 2009.

(POSTER)

14. E. Spyrakos, M. Frangópulos & C. Guisande: Ingestion rates of the heterotrophic

dinoflagellate Noctiluca scintillans fed on the toxic dinoflagellate Alexandrium minutum

(In Spanish). X Reunion Iberica, Fitoplancon Tóxico y Biotoxinas. Lisboa, Portugal, 12-15

May 2009. (ORAL)

15. L. González Vilas, J. Torres Palenzuela, M. Darriba Estévez & E. Spyrakos. Prediction of

Pseudo-nitzschia spp. events in the Galician rias (In Spanish). X Reunion Iberica,

Fitoplancon Tóxico y Biotoxinas. Lisboa, Portugal, 12-15 May 2009. (ORAL)

16. E. Garet, E. Spyrakos, J. Torres Palenzuela & Á. González-Fernández: Determination of

Domoic Acid in the waters of Rias de Vigo (In Spanish). X Reunion Iberica, Fitoplancon

Tóxico y Biotoxinas. Lisboa, Portugal, 12-15 May 2009. (ORAL)

17. E. Spyrakos & J. Torres Palenzuela: Study of harmful algal events in the ria of Vigo (NW

Spain) using geographical information systems and remote sensing techniques. Fourth

International Symposium in GIS/Spatial Analyses in Fishery and Aquatic Sciences, Rio de

Janeiro, Brazil, 25-29 August 2008. (POSTER)

xxii

18. E. Spyrakos, D. Pérez, Á. Mosquera Giménez, A. Acuña, C. Guisande, A. González

Fernández & J. Torres Palenzuela. Combination of different techniques and methods for

the study of water quality and detection of harmful algal events in Ría of Vigo (NW

Spain). Eastern boundary upwelling systems, integrative and comparative approaches,

Las Palmas, Grand Canaria, Spain, 2-6 June 2008. (POSTER)

Thesis outline

1

Thesis outline

3

Thesis outline

Thesis outline

The dissertation has been written as a series of chapters and its structure is

described below. Excluding the overall introduction and general conclusion all chapters

have been adapted from papers that have been published, submitted to a journal or are in

preparation. The publication details of these papers including their current status, the co-

authors and the journals are given above.

Chapter I provides a general introduction to all the chapters regarding the status,

effects and detection of harmful algal blooms, ocean colour techniques in coastal aquatic

systems, immunological methods and their use in toxin detection and the role of planktonic

grazers as toxin vectors and regulators of harmful algal events. In addition, the chapter

gives a description of the study area.

Chapter II describes the development of regionally specific model for chla

estimation from MERIS data, using in-situ data, neural networks and fuzzy c-mean clustering

techniques. It also compares the results with other models that are routinely used.

Chapter III takes advantage of the regionally specific algorithms and the

characteristics of MERIS in order to deliver more accurate, detailed chla maps of optically

complex coastal waters during an upwelling cycle. This application is demonstrated using

the example of a phase of intense upwelling in July 2008. The potential to map the spatial

extent of algae blooms caused by coastal upwelling has also been tested.

Chapter IV deals with the Pseudo-nitzschia spp. blooms in the study area by using a

combined analysis of satellite imagery data, measurements of biotic and abiotic parameters

and mixed effects modelling. The domoic acid concentration in the seawater was

determined for the first time in the Galician rias. The synoptic view of the spatial

distribution on horizontal scales of the blooms were obtained by applying regionally

specific algorithms to MERIS data and the utility of these tools for the monitoring of

Pseudo-nitzschia was examined.

Chapter V comprises the results of a laboratory experiment in order to evaluate the

ingestion and clearance rates of the heterotrophic dinoflagellate Noctiluca scintillans fed on

the toxic microalgae Alexandrium minutum for the first time to our knowledge.

Chapter VI is motivated by the results of Chapter V and intents to study and

compare PST transfer and accumulation of two different potential PST planktonic vectors

that show different grazing and reproductive behaviour and their role as PST vectors. In

order to perform this comparison, a model of toxin accumulation in vector population was

4

Thesis outline

constructed. The planktonic vectors selected were the heterotrophic dinoflagellate N.

scintillans and the copepod Acartia clausi.

Chapter I. Overall introduction

5

Overall Introduction

1.1 Satellite remote sensing of coastal optically complex waters

1.2 ELISA for DA detection 1.3 Harmful algae interactions with marine planktonic grazers



1.6 References

CHAPTER I

7


Transient proliferations of autotrophic algae and some heterotrophic protists,

referred to collectively as blooms are increasingly frequent in coastal areas around the

world (Sellner et al., 2003; Cullen, 2008). Some of these blooms are perceived as harmful

or/and threatening for the human health and activities and are associated with adverse

effects on marine ecosystems. The term harmful algal blooms (HABs) is used (since 1974: 1st

International conference of blooms of toxic dinoflagellates) to describe these events,

which skims over the fact that some phytoplankton species can have harmful effects in

relatively low abundances.

Anthropogenic actions like nutrient overloading of the coastal waters (Ríos et al.,

1995; Glibert et al., 2005), reduction of grazers (Rothschild et al., 1994), human-caused

climate change (Patz et al., 2005) and improved monitoring methods have been addressed

as possible factors for the escalation of this natural phenomenon. In 1993 Hallegraeff

mentioned the apparent global increase in harmful algal events characterising them as a

truly global epidemic with little in common. The last part reflects the deviation observed in

HABs in terms of harmfulness, bloom-forming organisms and dynamics.

Typically the list of harmful species includes two groups of causative organisms,

namely the toxin producers and the high biomass producers, with potentially toxic species

forming dense blooms also existing (Masó & Garcés, 2006). In terms of species number,

conservative estimations mention around 300 species out of 4000-6000 as causative for

harmful algal events under specific circumstances. Even less and in their majority

dinoflagellates are potentially marine toxins producers. Biotoxins produced by algae

species are often grouped according to the various intoxication syndromes they cause or

their chemical properties (hydrophilic and lipophilic) (see Table 1.1 in this Chapter and

Anderson et al., 1993). The best known cases are paralytic shellfish poisoning (PSP),

diarrhetic shellfish poisoning (DSP), amnesic shellfish poisoning (ASP) and ciguatera fish

poisoning (CFP). The harmful effects of high biomass algal events are described in detail in

Zingone and Enevoldsen (2000).

The global increase of harmful algal events in magnitude and frequency

documented the last decades accompanied in some cases with extensive economic and

ecological impacts has motivated advancing detection and monitoring efforts (Smayda,

1990; Hoagland et al., 2002; Sellner et al., 2003). In their review Hallegraeff et al. (2003)

addressed the substantial efforts that have been provided in order to understand and

monitor these events. Historically, harmful algal events detection relied on direct time-

consuming observation by light microscopy of the material (live or preserved) and ship

campaigns which in the most of the cases are characterised by restricted synopticity and

8


high cost (with the exception of ship-of-opportunity programs). Therefore, there is an

increasing scientific and public interest for cost-effective, less time-consuming quantitative

methods for the detection of the harmful algal events and their spatio-temporal variations

(Vrieling & Anderson, 1996; Subramaniam et al., 2002; Carvahlo et al., 2010). Several

approaches based on morphological/optical properties (Gower & Borstad; 1981; Carder &

Steward, 1985; Brown & Yoder, 1994; Sieracki et al., 1998; Subramaniam et al., 1999;

Robbins et al., 2006) or using previously developed biomedical methods (Peperzak et al.,

2000) have been developed recently for the detection of potentially harmful species.

However, it is generally accepted that HABs monitoring programs cannot be fully based on

indirect detections of the bloom-forming organisms and must be completed with direct

observations.

Summarising, increasing attention has been paid for the development of tools and

methods for different aspects of harmful algal events and their toxins sush as the

detection, monitoring, early-warning and dynamics (Fig. 1.1). The need for integration of

these tools and methods for the study of the harmful algal events is very important from a

scientific and management point of view. In this PhD thesis I present an integrated use of

technologies, data sets and models (Fig. 1.1) that could aid and improve the effectiveness of

monitoring and management programs for harmful algal events.

One tool with considerable potential for the detection and monitoring of harmful

algal events is optical remote sensing and especially reflectance-based techniques (ocean

colour). Section 1.1 provides an overview of satellite remote sensing of coastal areas

introducing the Chapters II and III. As continuation, the section 1.2 deals with another

technique used in this study for the rapid quantitative analysis of marine biotoxins.

Detection of harmful algal events in the seawater is critical to understanding this

phenomenon and in turn essential for their prediction and control. Referring to the control

of harmful algal events it is suggested, mainly after the report of Watras et al., 1985, that

planktonic grazers such as protists and zooplankton can put high pressure at harmful

species during blooms and can possible be linked to their prevention. Moreover, the

feeding interactions between harmful algal species and grazers may determine the fate of

the toxin in the food-wed. Section 1.3 gives a general introduction to the Chapters V and VI

which cope with feeding experiments and dynamic models for toxin transfer and

accumulation in planktonic grazers.

9


Fig. 1.1. Schematic for the general approach followed in this study. White cycles represent the aspects highlighted in the introduction.

1.1 Satellite remote sensing of coastal optically complex waters

Despite the fact that coastal waters represent barely the 10% of the world’s ocean

surface, they reflect an important part in terms of social, economic and ecological value. A

number of human activities like mariculture, fisheries, military operations, energy

production and recreation take place in coastal waters. Harmful algal events as it is

mentioned previously can have serious impacts to some of these activities. An important

component in detection of these events is passive ocean colour sensors.

The term ocean colour is loosely used to refer to the wavelength dependence of the

water leaving radiances at the sea surface. Conceptually, ocean colour remote sensing is

simple. The satellite sensors, with a discrete field of view, monitor the radiance (see the

definition by Mobley et al., 1994) reaching the sensor at many wavebands in the visible and

near-IR parts of the spectrum. The passage of the photons (of the sunlight excluded

emitted light by living organisms-bioluminesence) to the sensor involves: scattering by the

atmosphere (where multiple scattering is possible), specular reflection of direct sunlight at

10


the sea surface, in the case of shallow waters reflection by the sea-bottom and finally

upwelling from the sea surface after back-scattering in water. The last one is the useful

signal in ocean colour studies, and it can be attenuated by absorption and scattering of the

atmosphere in its way to and from the sea. This signal can be interpreted and provide

information about marine constituents such as phytoplankton biomass.

Among ocean-colour derived data, chlorophyll a (chla) concentration is the most

used product since it provides a good estimation of phytoplankton biomass and is common

to almost all taxonomic groups (Jeffrey et al., 1997). The phytoplankton community

responds rapidly to environmental changes (EC, 2000), which can cause visible changes in

chlorophyll in the surface waters.

The estimation of chla concentration in the oceans from the first dedicated ocean

colour scanner (CZCS) that launched in 1978 and operated until 1986 provided useful

information on the global distribution of chla but the quality of the data was limited

(Robinson, 2004). The Sea-viewing Wide-Field-view Sensor (SeaWiFS), Moderate Resolution

Imaging Spectroradiometer (MODIS) and the most recent Medium Resolution Imaging

Spectrometer (MERIS) that succeeded CZCS are using more and narrower spectral bands

and finer spatial resolution. MERIS provides data with a 300 m on-ground resolution in nadir

(Full Resolution) and has a spectral resolution of fifteen bands from visible to near infra red,

supporting one of the mission objectives for delicate coastal zone monitoring (Doerffer et

al., 1999).

Traditionally, chla is estimated using empirical algorithms based on the ratio

between the radiance of blue and green light reflected by the sea. For the retrieval of chla

from ocean colour sensors various empirical spectral-ratio algorithms (Evans & Gordon,

1994; Muller-Karger et al., 1990; Aiken et al., 1995; McClain et al., 2004; O´ Reilly et al., 2000;

Brown et al., 2008) and semi-analytical models (Garder & Steward, 1985; Garder et al., 1999)

were developed. These algorithms are accurately applicable in oceanic waters, where

phytoplankton and material of biological origin are the principal agents responsible for the

variation in the optical properties. These are known as Case 1 waters according to the

classification proposed by Morel and Prieur in 1977. In contrast, in coastal waters influenced

by river discharges, coloured dissolved organic material (CDOM) and suspended sediments

are also present and their concentrations vary independently from those of phytoplankton

(Case 2 waters).

11


Fig. 1.2. Left image: Oceanic waters where the phytoplankton and material of biological origin are the principal agents responsible for the variation in the optical properties (Case 1). Right image: Coastal waters influenced by river discharges where coloured dissolved organic material and suspended sediments are also present and their concentrations vary independently from this of phytoplankton (Case 2).

In typical case 2 waters, where high concentrations of water constituents (CDOM,

detritus) absorb strongly in the blue decoupling the phytoplankton absorbance, this ratio

cannot be used for an accurate retrieval of chla (Morel & Prieur 1977; Gons, 1999; Gitelson

et al., 2007). However, chla algorithms that use green to red and near infrared band ratios

have shown good performance in inland and coastal waters (Gilerson et al., 2010). This

classification is not always a simple distinction of coastal and oceanic waters as Morel and

Maritorena (2001) describe in a later study using a new dataset of optical properties,

indicating the need for models more restricted in geographical and seasonal terms. It is

noteworthy mentioning that methods which work for case 2 water may theoretically also

be applied to case 1 water (but not vice versa).

In the effort for more accurate retrieval of water constituents in optically complex

waters, neural network (NN) techniques can play an important role, since they seem ideal

for multivariate, complex and non-linear data modelling (Thiria, 1993). In the last decades

the application of NN techniques for the estimation of selected water quality parameters

from ocean-colour has increased (Atkinson & Tatnall, 1997; Keiner & Yan, 1998; Shahraiyni

et al., 2009). NN based algorithms are currently used as standard products for the

estimation of chla, SPM (suspended particulate material) and yellow substances by the

European Space Agency (ESA) for MERIS data (Doerffer and Schiller, 2007).

12


Fig. 1.3. Triangular diagram showing the main water optical classes (adapted from Prieur & Sathyendranath, 1981). The three corners represented by the letters P, Y and S indicate dominance of phytoplankton, yellow substances and suspended material, respectively.

Although remote sensing tools can be used with a relatively high precision at global

scale for the calculation of chla, they are not always totally accurate in local areas. Dransfeld

et al., (2004) proposed that studies for development of ocean colour algorithms should be

regionally specific and emphasised the role of NNs in the retrieval of water constituents

especially in Case 2 waters. Validation methods and development or expansion of chla

algorithms for specific areas have been widely used to test and regionalize the satellite

products (e.g. Cota et al., 2004; Witter et al., 2009). A recent ESA project (CoastColour)

attempts to provide regionally tuned ocean colour products applying fuzzy logic

classification to guide the MERIS TOA spectrum to the most appropriate neural network-

based algorithm (Brockman, pers. com.).

1.2 ELISA for DA detection

Nowadays, the replacement of the mouse bioassay method that is routinely used for

the detection of marine toxins with new more accurate methods it is considered as

essential. This need for rapid and specific detection and identification biotoxins produced

by the microalgae has led to the implementation of antibody-based assays in the detection

study of harmful algal blooms (Lewis, 2001; Camacho et al., 2007; Yakes et al., 2011). Among

them, ELISA (enzyme-linked immunosorbent assay) is perhaps the most used method due

to its simplicity, sensitivity and adaptability. In most cases, ELISAs can be applied in

13


different samples, including seawater samples or biological extracts (e.g. shellfish and body

fluids) (Naar et al., 2002; Kirkpatrick et al., 2004; Samdal et al., 2005).

There are different types of ELISA (competitive, sandwich, direct and indirect) that are

performed using plastic microplates, antigens, and antibodies (Abs) marked with an

enzyme (such as horseradish peroxidase or alkaline phosphatase) that will react with a

colourless or chromogenic substrate (examples: ABTS, TMB, DAB), producing a coloured

product. The signal reading is completed in a spectrophotometer equipped with a filter

system that allows the simultaneous reading of all the wells of the plate at a given

wavelength. In general, ELISA techniques are very sensitive, detecting

picograms/nanograms of a substance in a sample. Several ELISAs for detection and

qualification of marine phytotoxins are now available commercially (Table 1.1).

Table 1.1. ELISA kits commercially available for the detection and qualification of marine toxins produced by microalgae.

Toxins Effects/Category* available ELISA kits main producer

species

Saxitoxin and

its relatives

Paralytic (PSP)/

saxitoxin group

- Ridaserren, (R-Biopharm,

Darmstadt, Germany)

- Abraxis LLC, (Warminster PA,

USA)

Alexandrium spp. and

Gymnodinium

catenatum

Domoic acid Amnesic (ASP)/

domoic acid group

- Biosense Laboratories (Bergen,

Norway)

Pseudo-nitzschia spp.

Ocadaic acid

and its

derivatives

Diarrhetic

(DSP)/

ocadaic acid group

- UBE Industries, Ltd, (Tokyo,

Japan)

- Rougier Bio-Tech Ltd, (Montreal,

Canada)

Dinophysis and

Prorocentrum

Yessotoxin yessotoxin group - Biosense Laboratories (Bergen,

Norway)

Protoceratium

reticulatum,

Lingulodinium

polyedrum and

Gonyaulax spinifera

Brevetoxin

and analogs

Neutotoxin

(NSP)/

brevetoxin group


Norway)

Karenia brevis

*PSP: Paralytic Shellfish Poisoning, ASP: Amnesic Shellfish Poisoning, DSP: Diarrhetic Shellfish

Poisoning, NSP: Neurotoxic Shellfish Poisoning.2

2 Other 3 known categories of seafood poisoning are: Ciguatera fish poisoning (CFP) caused by ingestion of coral reef fishes that have become toxic through diet, Azaspirascid poisoning (AZP) causing similar symptoms to DSP and finally poisoning from chronic explosure to moderate toxin levels.

14


This thesis deals with an ELISA kit for domoic acid (DA) detection. This technique has

been recently validated as an official method (Association of Official Analytical Chemists-

AOAC, 2006), alternative to mouse bioassay. However, it is clearly mentioned that in the

case of challenged results, high performance liquid chromatography (HPLC) should be used

as the reference method. DA, a naturally occurring but rare non-protein amino acid, is a

neurotoxin produced mainly by several species of the marine diatom Pseudo-nitzschia. In

mild cases, symptoms after the consumption of toxic seafood may include vomiting,

diarrhea, abdominal cramps, headache and dizziness and normally disappear within few

days. In sterner cases, the victim may experience difficulty breathing, confusion, loss of

short-term memory and even coma and death (Wang, 2008). In comparison with HPLC

detection methods, ELISA is more sensitive with detection limits normally in the picogram

range (Garet et al., 2010). In addition, ELISA for DA detection appears to be specific for DA

while cross-reactivity with kainic acid, glutamate or glutamic acid is not significant.

Nevertheless, Burns et al., 2007 noticed significant sensitivity of cELISA to kainic acid

warning for overestimation of DA when kainic acid is present.

1.3 Harmful algae interactions with marine planktonic grazers

Planktonic grazers have a central role in marine pelagic systems as mediators of

energy to higher trophic levels and as regulators of phytoplankton biomass (Lessard, 1991;

Strom & Morello, 1998; Sherr & Sherr, 2007). In the case of harmful algal events the

interactions between the toxic species and the planktonic organisms feeding on them are

considered important, but also complex, unclear and situation-specific. Several studies (e.g.

Irogoien et al., 2005) address grazing as the main reason of harmful algae mortality in

natural bloom episodes, while others have suggested that reduced grazing pressure can be

an essential factor for monospecific blooms formation (Fiedler, 1982 ; Uye & Takamatsu,

1990). The role of planktonic grazing on harmful species may be important not only to

determine the occurrence of a harmful algal event and its dynamics but also for the

distribution of the toxins in the marine food web (Smayda, 1992). A number of laboratory

studies have shown a variety of responses of planktonic grazers feeding on bloom-forming

and/or toxic algae species ranging from no effect to mortality (for a review see Turner &

Tester, 1997). When they are fed with toxic algae species, planktonic grazers may retain,

eliminate or even transform toxins. In this sense, ingestion of toxic species by different

types of planktonic organisms may be significant for the fate of the toxins in the food web,

since a variety of planktonic vectors such as protists and zooplankton may be involved in

15


the transfer of toxins (White 1981; Teegarden et al., 2003; Jiang et al., 2007). In theory, all

the consumers in the marine environment can eventually convert to toxin vectors.

Although many field studies and laboratory experiments deal with grazing

interactions between zooplankton and especially copepods and toxin-producer algae

species, very few have included protists (ciliates and heterotrophic dinoflagellates).

Heterotrophic protists compose a major component of the planktonic communities with

very different feeding strategies and behaviour patterns and can be found at high

concentrations during harmful algal events suggesting considerable effects on the bloom

dynamics. All of the above give prominence to the need for detailed studies of toxin

transfer and retention in systems which include harmful algae, protists and zooplankton.


The Galician rias are V-like embayments along the northern boundary of the NW

Africa upwelling system formed by sunken river valleys flooded by the sea, whose

ecosystems are strongly influenced by oceanic conditions on the adjacent continental shelf.

The rias Baixas constitute the southern part of the Galician rias (Fig. 1). They are formed by

four large coastal embayments, from north to south: Muros y Noya, Arousa, Pontevedra and

Vigo, all oriented in a SW–NE direction, and characterized by strong tides. Surface area

covers approximately 600 km2 and water depths range from 5 to 60 m. The ria de Vigo is the

longest of the rias whereas the ria de Arousa is the widest. Rias vary in width from 1-3 km in

their inner part to 8-12 km in their external part (Vilas et al., 2005). The main freshwater

inputs in the rias are by rivers located in innermost part of the rias.

In these highly primary productive upwelling estuarine systems (Fraga, 1981; Torres

& Barton, 2007) transient increases of phytoplankton abundance, referred to as blooms,

are a frequent phenomenon occurring mainly between early spring and late autumn (Fraga

et al., 1988; Varela, 1992, Figueiras & Ríos, 1993). Phytoplanktonic blooms or ‘‘red tides’’ as

they were formally called, were mentioned for the first time in the Galician rias in 1918 by

Sobrino (Varela, 1992). Sporadically, some phytoplankton blooms in the Galician rias are

perceived as harmful with direct and indirect impacts to the mussel production that

constitute an important economic activity in the area. Harmful algal events in the area are a

well documented phenomenon. Several studies since the 1950s referred to them and in

general to phytoplankton ecology particularizing favourable conditions to the development

of HABs, their origin, dynamic, distribution and toxicity levels (Margalef, 1956; Tilstone et

16


al., 1994; Figueiras et al., 1994; GEOHAB, 2005), seasonal taxonomic and chemical

composition of phytoplankton and picophytoplankton “patchiness” (Figueiras & Niell, 1987;

Nogueira et al., 1997). Harmful algal events in the Galician coast and generally along the

Iberian upwelling system are linked to several (potentially) toxic species like Pseudo-

nitzschia spp. (GEOHAB, 2005), Dinophysis acuminata and D. acuta (Campos et al., 1982;

Reguera et al., 1993), Gymnodynium catenatum (Fraga et al., 1988), Heterosigma akashiwo

(Crespo et al., 2006), Alexandrium minutum (Franco et al., 1994) and Lingulodinium

polyedrum (in the Northern rias: Arévalo et al., 2006). High biomass events due to bloom

forming species have been also recorded (e.g. Noctiluca scintillans and Ceratium furca) but

were never accompanied with anoxia probably because of the short renewal time that

characterise the rias (Álvarez-Salgado, 2008).

Fig. 1.4. Galician coast and bathymetry of the area. From north to south the rias Baixas: Muros y Noya, Arousa, Pontevedra and Vigo.

Chapter I. Overall introduction-Motivations and objectives

17

Motivations and thesis

objectives

19



The present PhD thesis aims to advance or/and test the potential of remote sensing

and immunological methods for the detection of harmful algal events in the Galician rias

and the adjacent area and improve our understanding of some planktonic organisms as

toxin vectors and HA regulators.

Following the widespread understanding that universally applicable water

constituent retrieval algorithms from ocean colour sensors are currently not feasible, the

research was shifted to regionally specific implementations of powerful inversion methods.

Given the considerable interest for accurate chlorophyll a mapping in the optically

complex waters of the study area deriving from the economic and social importance of the

extensive mariculture activities and high frequency of harmful algal events.

Recognising the need for rapid detection of even small quantities of toxins

produced by algae species in the seawater and for understanding of the relationships

between the toxic species abundance, toxin production and the characteristics of the

environment.

Recalling that the role of heterotrophic dinoflagellates can be of major importance

in the fate of algal toxins in the trophic food web but is still insufficiently studied.

Considering that planktonic vectors with different grazing and reproductive

behaviour can play a different role as regulators of the marine toxins in the planktonic

community.

This dissertation will address the following objectives:

Objective 1. To develop and validate neural network-based chla algorithms for

MERIS FR data, which would be specific for the optically complex coastal waters of the

Galician rias.

Objective 2. To apply these algorithms in a short series of MERIS (FR) images

delivered during an upwelling cycle in order to obtain maps of chla.

Objective 3. To test the potential of these algorithms to map the spatial extent of

possible algal blooms caused by coastal upwelling.

Objective 4. To measure for the first time in the study area the domoic acid (DA)

concentrations in natural populations.

20


Objective 5. To determine the environmental factors that might be responsible for

the Peudo-nitzschia events and the production of DA.

Objective 6. To assess the utility of MERIS and regional algorithms for the

monitoring of Pseudo-nitzschia.

Objective 7. To examine whether the frequently found in the area heterotrophic

dinoflagellate N. scintillans could actively feed on a PST-producing dinoflagellate and assess

if it can eventually act as a PST vector.

Objective 8. To study and compare PST transfer and accumulation of two different

potential PST planktonic vectors that show different grazing and reproductive behaviour

and their role as PST vectors in the planktonic community using dynamic models.

21


1.6 References

Aiken, J., Moore, G. F., Trees, C. C., Hooker, S. B., & Clark, D. K. (1995). The SeaWiFS CZCS type

pigment algorithm. In S. B. Hooker, & E. R. Firestone (Eds.). NASA technical memorandum,

104566. SeaWiFS Technical Report Series, vol. 104529. (pp. 104561-104532).

Álvarez-Salgado, X. A., Labarta, U., Fernández-Reiriz, M. J., Figueiras, F. G., Rosón, R.,

Piedracoba, S., Filgueira, R., & Cabanas, J. M. (2008). Renewal time and the impact of harmful

algal blooms on the extensive mussel raft culture of the Iberian coastal upwelling system (SW

Europe). Harmful Algae, 7, 849-855.

Anderson, D. M., Galloway, S. B., & Joseph, J. D. (1993). Marine Biotoxins and Harmful Algae: A

National Plan. Woods Hole Oceanographic Institution-Technical Report.

Arévalo, F., Pazos, Y., Correa, J., Salgado, C., Moroño, A., Pan, B., & Franco, J. (2006). First report

of yessotoxins in mussels of the Galician Rias during a bloom of Lingulodinium polyedrum Stein

(Dodge). In K. Henshilwood, B. Deegan, T. McMahon, C. Cusak, S. Keaveney, J. O. Silke, M.

Cinneide, D. Lyons & P. Hess (Eds.), 5th International Conference on Molluscan Shellfish Safety

(pp. 184–189). Galway, Ireland.

Atkinson, P. M., & Tatnall, A. R. L. (1997). Introduction Neural networks in remote sensing.

International Journal of Remote Sensing, 18, 699-709.

Brown, C. A., Huot, Y., Werdell, P. J., Gentili, B., & Claustre, H. (2008). The origin and global

distribution of second order variability in satellite ocean color and its applications to algorithm

development. Remote Sensing of Environment, 112, 4186-4203.

Brown, C. W., & Yoder, J. A. (1994). Coccolithophorid blooms in the global ocean. Journal of

Geophysical Research, 99, 7467-7482.

Burns, J. M., Schock, T. B., Hsia, M. H., Moeller, P. D., & Ferry, J. L. (2007). Photostability of kainic

acid in seawater. Journal of Agricultural and Food Chemistry, 55, 9951-9955.

Camacho, F. G., Gallardo Rodríguez, J., Sánchez Mirón, A., Cerón García, M. C., Belarbi, E. H.,

Chisti, Y., & Molina Grima, E. (2007). Biotechnological significance of toxic marine

dinoflagellates. Biotechnology Advences, 25, 176-194.

Campos, M. J., Fraga, S., Mariño, J., & Sánchez, J. (1982). Red tide monitoring programme in NW

Spain. Report of 1977-1981. In, International Council for the Exploration of the Sea (p. 8). Denmark

Carder, K., & Steward, R. G. (1985). A remote-sensing reflectance model of a red-tide

dinoflagellate off west Florida. Limnology and Oceanography, 30, 286-298.

Carder, K. L., Chen, F. R., Lee, Z. P., Hawes, S. K., & Kamykowski, D. (1999). Semianalytic

moderate-resolution imaging spectrometer algorithm for chlorophyll a concentration and

absorption with bio-optical domains based on nitrate-depletion temperatures. Journal of

22



Carder, K. L., & Steward, R. G. (1985). A remote-sensing reflectance model of a red-tide


Carvalho, G. A., Minnett, P. J., Fleming, L. E., Banzon, V. F., & Baringer, W. (2010). Satellite

remote sensing of harmful algal blooms: A new multi-algorithm method for detecting the

Florida Red Tide (Karenia brevis). Harmful Algae, 9, 440-448.

Cembella, A. D., Doucette, G. J., & Garthwaite, I. (2003). In vitro assays for phycotoxins. In G. M.

Hallegraeff, D. M. Anderson & A. D. Cembella (Eds.), Manual on harmful marine microalgae. Paris:

IOC-UNESCO.

Cota, G. F., Wang, J., & Comiso, J. C. (2004). Transformation of global satellite chlorophyll

retrievals with a regionally tuned algorithm. Remote Sensing of Environment, 90, 373-377.

Crespo, B. G., Figueiras, F. G., Porras, P., & Teixeira, I. G. (2006). Downwelling and dominance of

autochthonous dinoflagellates in the NW Iberian margin: the example of the Ria de Vigo.

Harmful Algae, 5, 770-781.

Cullen, J. J. (2008). Observations and prediction of harmful algal blooms. In M. Babin, C. S.

Roesler & J. J. Cullen (Eds.), Real-time Coastal Observing Systems for Marine Ecosystem Dynamics

and Harmful Algal Blooms: Theory, Instrumentation and Modelling (pp. 1-41). Paris: UNESCO.

Doerffer, R., & Schiller, H. (2007). The MERIS Case 2 water algorithm. International Journal of

Remote Sensing, 28, 517-535.

Doerffer, R., Sorensen, K., & Aiken, J. (1999). MERIS potential for coastal zone application.


Dransfeld, S., Tatnall, A. R., Robinson, I. S., & Mobley, C. D. (2004). A comparison of Multi-layer

multilinear regression algorithms for the inversion of synthetic ocean colour spectra.


EC (2000). Directive 2000/60/EC of the European Parliament and of the Council of 23 October

2000 establishing a framework for Community action in the field of water policy. Official Journal

L 327, 22/12/2000: 2001-2073.

Evans, R. H., & Gordon, H. R. (1994). Coastal zone color scanner system calibration: a

retrospective examination. Journal of Geophysical Research, 99, 7293-7307.

Fiedler, P. (1982). Zooplankton avoidance and reduced grazing responses to Gymnodinium

splendens (Dinophyceae). Limnology and Oceanography, 27, 961–965.

Figueiras, F. G., Jones, K. J., Mosquera, A. M., Alvarez-Salgado, X. A., Edwards, E., & MacDougall,

N. (1994). Red tide assemblage formation in an estuarine upwelling ecosystem: Ria de Vigo.

Journal of Plankton Research, 16, 857-878.

23


Figueiras, F. G., & Niell, F. X. (1987). Composición del fitoplankton de la ría de Pontevedra (NO de

España). Investigacion Pesquera, 51, 371-409.

Figueiras, F. G., & Ríos, A. F. (1993). Phytoplankton succession, red tides and the hydrographic

regime in the Rías Bajas of Galicia. In T. J. Smayda & Y. Shimizu (Eds.), Toxic Blooms in the Sea

(pp. 239-244). New York: Elsevier.

Fraga, F. (1981). Upwelling off the Galician coast North West Spain. In E. Suess & J. Thiede (Eds.),

Coastal upwelling (pp. 176-182). New York: Plenum Publishing Corp.

Fraga, S., Anderson, D. M., Bravo, I., Reguera, B., Steidinger, K. A., & Yentsch, C. M. (1988).

Influence of upwelling relaxation on dinoflagellates and shellfish toxicity in Ria de Vigo, Spain.

Estuarine Coastal Shelf Science, 27, 349-361.

Franco, J., Fernández, P., & Reguera, B. (1994). Toxin profiles of natural populations and

cultures of Alexandrium minutum Halim from Galician (Spain) coastal waters, Journal of Applied

Phycology, 6, 275-279.

Garet, E., González-Fernández, Á., Lago, J., Vieites, J., Cabado, A., & . . J. Agric and Food

Chemistry (2010). Comparative evaluation of Enzyme-linked Immunoassay and reference

methods for the detection of shellfish hydrophilic toxins in several presentations of seafood.

Journal of Agricultural and Food Chemistry, 58, 1410-1415.

Garthwaite, I., Ross, K. M., Miles, C. O., Hansen, R. P., Foster, D., Wilkins, A. L., & et al (1998).

Polyclonal antibodies to domoic acid, and their use in immunoassays for domoic acid in sea

water and shellfish. Journal of Natural Toxins, 6, 93-104.

GEOHAB (2005). Global Ecology and Oceanography of Harmful Algal Blooms. In G. Pitcher, T.

Moita, V. Trainer, R. Kudela, F. G. Figueiras & T. Probyn (Eds.), GEOHAB Core Research Project:

HABs in Upwelling Systems (p. 82). Paris and Baltimore: IOC and SCOR.

Gilbert, P. M., Anderson, D. M., Gentien, P., Granéli, E., & Sellner, K. G. (2005). The global,

complex phenomena of harmful algal blooms. Oceanography, 18, 131-141.

Gilerson, A. A., Gitelson, A. A., Zhou, J., Gurlin, D., Moses, W. J., Ioannou, I., & al., e. (2010).

Algorithms for remote estimation of chlorophyll-a in coastal and inland waters using red and

near-infrared bands. Optical Express, 18, 24109-24125.

Gitelson, A. A., Schalles, J. F., & Hladik, C. M. (2007). Remote chlorophyll-a retrieval in turbid,

productive estuaries: Chesapeake Bay case study. Remote Sensing of Environmental Science and

Technology, 109, 464-472.

Gons, H. (1999). Optical teledetection of chlorophyll a in turbid inland waters Environmental

Science and Technology, 33, 1127-1132.

Gower, J. F. R., & Borstad, G. A. (1981). Use of the in vivo fluorescence line at 685nm for remote

sensing of surface chlorophyll a. In J. F. R. Gower (Ed.), Oceanography From Space (pp. 329-338).

24


New York: Plenum Press.

Hallegraeff, G. M. (1993). A review of harmful algal blooms and their apparent global increase.

Phycologia, 32, 79-99.

Hallegraeff, G. M., Anderson, D. M., & Cembella, A. D. (Eds.) (2003). Manual on harmful marine

algae. Paris: UNESCO Publishing.

Hoagland, P., Anderson, D. M., Kaoru, Y., & White, A. W. (2002). The economic effects of

harmful algal blooms in the United States: estimates, assessment issues, and information needs.

Estuaries, 25, 819-837.

Irigoien, X., Flynn, K. J., & Harris, R. P. (2005). Phytoplankton blooms: a ‘loophole’ in

microzooplankton grazing impact? Journal of Plankton Research, 27, 313-321.

Jeffrey, S. W., Mantoura, R. F. C., & Wright, S. W. (Eds.) (1997). Phytoplankton Pigments in

Oceanography: Guidelines to Modern Methods. Paris: Unesco.

Jianga, T.-J., Wang, D.-Z., Niua, T., & Xua, Y.-X. (2007). Trophic transfer of paralytic shellfish

toxins from the cladoceran (Moina mongolica) to larvae of the fish (Sciaenops ocellatus).

Toxicon, 50, 639-645.

Keiner, L. E., & Yan, X. H. (1998). A neural network model for estimating sea surface chlorophyll

and sediments from thematic mapper imagery. Remote Sensing of Environmental Science and

Technology, 66, 153-165.

Kirkpatricka, B., Flemingb, L., Squicciarinib, D., Backerc, L., Clarkd, R., Abrahamb, W., Bensone,

J., Chenge, Y., Johnsond, D., Piercea, R., Zaiasb, J., Bossartb, G., & Badenf, D. (2004). Literature

review of Florida red tide: implications for human health effects. Harmful Algae, 3, 99-115.

Lessard, E. J. (1991). The trophic role of heterotrophic dinoflagellates in diverse marine

environments. Marine Microbial Food Webs, 5, 49-58.

Lewis, R. J. (2001). The changing face of ciguatera. Toxicon, 39, 97-106.

Margalef, R. (1956). Estructura y dinámica de la “purga de mar” en Ría de Vigo. Investigacion

Pesquera, 5, 113-134.

Masó, M., & Garcés, E. (2006). Harmful microalgae blooms (HAB): problematic and conditions

that induce them. Marine Pollution Bulletin, 53, 620-630.

McClain, C. R., Feldman, G. C., & Hooker, S. B. (2004). An overview of the SeaWiFS project and

strategies for producing a climate research quality global ocean biooptical time series. Deep-Sea

Research II, 51, 6-42.

Mobley, C. D. (Ed.) (1994). Light and Water: Radiative Transfer in Natural Waters San Diego: CA:

Academic.

25


Morel, A., & Maritorena, S. (2001). Bio-optical properties of oceanic waters: a reappraisal.

Journal of Geophysical Research, 106, 7163-7180.

Morel, A., & Prieur, L. (1977). Analysis of variations in ocean color. Limnology and Oceanography,

22, 709-722.

Muller-Karger, F. E., McClain, C. R., Sambrotto, R. N., & Ray, G. C. (1990). A comparison of ship

and coastal zone color scanner mapped distribution of phytoplankton in the southeastern

Bering Sea. Journal of Geophysical Research, 95, 483-499.

Naar, J., Bourdelais, A., Tomas, C., Kubanek, J., Whitney, P. L., Flewelling, L., Steidinger, K.,

Lancaster, J., & Baden, D. G. (2002). A competitive ELISA to detect brevetoxins from Karenia

brevis (formerly Gymnodinium breve) in seawater, shellfish, and mammalian body fluid.

Environmental Health Perspectives, 110, 179-185.

Nogueira, E., Perez, F. F., & Ríos, A. F. (1997). Seasonal patterns and long-term trends in an

estuarine upwelling ecosystem (Ria de Vigo, NW Spain). Estuarine Coastal and Shelf Science, 44,

185-300.

O'Reily, J. E., Maritorena, S., Siegel, D. A., O'Brien, M. C., Toole, D., Mitcell, B. G., & al., e. (2000).

Ocean color chlorophyll a algorithms foe SeaWiFS, OC2 and OC4. Version 4. NASA technical

memorandum 2000-206892. In S. B. Hooker & E. R. Firestone (Eds.) (p. 49)

Patz, J. A., Campbell-Lendrum, D., Holloway, T., & Foley, J. A. (2005). Impact of regional climate

change on human health. Nature, 438, 310-317.

Peperzak, L., Vrieling, E. G., Sandee, B., & Rutten, T. (2000). Immuno flow cytometry in marine

phytoplankton research. Scientia Marina, 64, 165-181.

Reguera, B., Mariño, J., Campos, M. J., Bravo, I., Fraga, S., Carbonell, A., In:, (Eds.), & , p. (1993).

Trends in the occurrence of Dinophysis spp. in Galician waters. In T. J. Smayda & Y. Shimizu

(Eds.), Toxic Phytoplankton Blooms in the Sea (pp. 559-564). Amsterdam: Elsevier Science

Publishers.

Ríos, A. F., Fraga, F., Figueiras, F. G., & Pérez, F. F. (1995). New and regenerated production in

relation to the proliferations of diatoms and dinoflagellates in natural conditions. In P. Lassus, G.

Arzul, E. Erard, P. Gentien & C. Marcaillou (Eds.), Harmful marine algal blooms (pp. 663-668).

Paris: Lavoisier.

Robbins, I. C., Kirkpatrick, G. J., Blackwell, S. M., Hillier, J., Knight, C. A., & Moline, M. A. (2006).

Improved monitoring of HABs using autonomous underwater vehicles (AUV). Harmful Algae, 5,

749-761.

Robinson, I. S. (Ed.) (2004). Measuring the Oceans from Space: The Principles and Methods of

Satellite Oceanography. Berlin: Springer-Praxis.

26


Rothschild, B. J., Ault, J. S., Goulletquer, P., & Heral, M. (1994). The decline of the Chesapeake

Bay oyster population: A century of habitat destruction and overfishing. Marine ecology progress

series, 111, 29-39.

Samdal, I. A., Aasen, J. A. B., Briggs, L. R., Dahl, E., & Miles, C. O. (2005). Comparison of ELISA

and LC-MS analyses for yessotoxins in blue mussels (Mytilus edulis). Toxicon, 46, 7-15.

Sellner, K. G., Doucette, G. J., & Kirkpatrick, G. (2003). Harmful algal blooms: Causes, impacts

and detection. Journal of Industrial Microbiology and Biotechnology, 30, 383-406.

Shahraiyni, H. T., Shouraki, S. B., Fell, F., Schaale, M., Fischer, J., Tavakoli, A., & al., e. (2009).

Application of an active learning method to the retrieval of pigment from spectral remote

sensing reflectance data. International Journal of Remote Sensing, 30, 1045-1065.

Sherr, E. B., & Sherr, B. F. (2007). Heterotrophic dinoflagellates: a significant component of

microzooplankton biomass and major grazers of diatoms in the sea. Marine ecology progress

series, 352, 187-197.

Sieracki, C. K., Sieracki, M. E., & Yentsch, C. S. (1998). An imaging-in-flow system for automated

analysis of marine microplankton. Marine ecology progress series, 168, 285-296.

Smayda, T. J. (1990). Novel and nuisance phytoplankton blooms in the sea: evidence for a global

epidemic. In E. Graneli, B. Sundstrom, L. Edler & D. M. Anderson (Eds.), Toxic marine

phytoplankton (pp. 29-40). New York: Elsevier.

Smayda, T. J. (1992). Global epidemics of noxious phytoplankton blooms and food chain

consequences in large ecosystems. In K. Sherman, L. M. Alexander & B. D. Gold (Eds.), Food

Chains, Models and Management of Large Marine Ecosystems (pp. 275-307). San Francisco:

Westview Press.

Strom, S. L., & Morello, T. A. (1998). Comparative growth rates and yields of ciliates and

heterotrophic dinoflagellates. Journal of Plankton Research, 20, 571-584.

Subramaniam, A., Brown, C. W., Hood, R. R., Carpenter, E. J., & Capone, D. G. (2002). Detecting

Trichodesmium blooms in SeaWiFS imagery. Deep-Sea Research, 49, 107-121.

Subramaniam, A., Carpenter, E. J., & Falkowski, P. G. (1999). Bio-optical properties of the marine

diazotrophic cyanobacteria Trichodesmium spp.: II. A reflectance model for remote sensing.

Limnology and Oceanography, 44, 618-627.

Teegarden, G. J., Cembella, A. D., & Carissa, L. C. ( 2003). Phycotoxin accumulation in

zooplankton feeding on Alexandrium fundyense -vector or sink? Journal of Plankton Research,

25, 429-443.

Thiria, S., Mejia, C., Badran, F., & Crepon, M. (1993). A neural network approach for

modellingnonlinear transfer functions: Application forwind retrieval fromspaceborne

scatterometer data. Journal of Geophysical Research, 98(12), 22827-22841.

27


Tilstone, G. H., Figueiras, F. G., & Fraga, F. (1994). Upwelling-downwelling sequences in the

generation of red tides in a coastal upwelling system. Marine ecology progress series, 112, 241-253.

Torres, R., & Barton, E. D. (2007). Onset of the Iberian upwelling along the Galician coast.

Continental Shelf Research, 27, 1759-1778.

Turner, J. T., & Tester, P. A. (1997). Toxic marine phytoplankton, zooplankton grazers, and

pelagic food webs. Limnology and Oceanography, 42, 1203-1214.

Uye S., & Takamatsu, K. (1990). Feeding interactions between planktonic copepods and red-

tide flagellates from Japanese coastal waters. Marine ecology progress series, 59, 97–107.

Varela, M. (1992). Upwelling and phytoplankton ecology in Galician (NW Spain) rías and shelf

waters. Boletín del Instituto Español de Oceanografía, 8, 57-74.

Vilas, F., Bernabeu, A., & Méndez, G. (2005). Sediment distribution pattern in the Rías Baixas

(NW Spain): main facies and hydrodynamic dependence. Journal of Marine Systems, 54, 261-276.

Vrieling, E. G., & Anderson, D. M. (1996). Immunofluorescence in phytoplankton research:

applications and potential. Journal of Phycology, 32, 1-16.

Wang, D.-Z. (2008). Neurotoxins from Marine Dinoflagellates: A Brief Review. Marine Drugs, 6,

349-371.

Watras, C. J., Garcon, V. C., Olson, R. J., Chisholm, S. W., & Anderson, D. M. (1985). The effect of

zooplankton grazing on estuarine blooms of the toxic dinoflagellate Gonyaulax tamarensis.


White, A. W. (1981). Marine zooplankton can accumulate and retain dinoflagellate toxins and

cause fish kills. Limnology and Oceanography, 26, 103-109.

Witter, D. L., Ortiz, J. D., Palm, S., Heath, R. T., & Budd, J. W. (2009). Assessing the application of

SeaWiFS ocean color algorithms to Lake Erie. Journal of Great Lakes Research, 35, 361-370.

Yakes, B. J., DeGrasse, S. L., Poli, M., & Deeds, J. R. (2011 ). Antibody characterization and

immunoassays for palytoxin using an SPR biosensor. Analytical and Bioanalytical Chemistry, 400,

2865-2869.

Zingone, A., & Enevoldsen, H. O. (2000). The diversity of harmful algal blooms: a challenge for

science and management. Ocean and Coastal Management, 43, 725-748.

Chapter II. Development of MERIS regionally specific chla algorithms

29

Development of regionally specific chlorophyll a algorithms of optically complex waters for MERIS FR data

Abstract

2.1 Introduction

2.2 Material and methods

2.3 Results

2.4 Discussion

2.5 References

CHAPTER II

31


Abstract

In typical Case 2 waters, accurate remote sensing retrieval of chlorophyll a (chla) is

still a challenging task. In this study, focusing on the Galician rias (ΝW Spain), algorithms

based on neural network (NN) techniques were developed for the retrieval of chla

concentration in optically complex waters, using Medium Resolution Imaging Spectrometer

(MERIS) data. There is considerable interest in the accurate estimation of chla for the

Galician rias, because of the economic and social importance of the extensive culture of

mussels, and the high frequency of harmful algal events. Fifteen MERIS full resolution (FR)

cloud-free images paired with in situ chla data (for 2002–2004 and 2006–2008) were used

for the development and validation of the NN. The scope of NN was established from the

clusters obtained using fuzzy c-mean (FCM) clustering techniques applied to the satellite-

derived data. Three different NNs were developed: one including the whole data set, and

two others using only points belonging to one of the clusters. The input data for these

latter two NNs was chosen depending on the quality level, defined on the basis of quality

flags given to each data set. The fitting results were fairly good and proved the capability of

the tool to predict chla concentrations in the study area. The best prediction was given for

the NN trained with high-quality data using the most abundant cluster data set. The

performance parameters in the validation set of this NN were R2 = 0.86, mean percentage

error (MPE) = − 0.14, root mean square error (RMSE) = 0.75 mg m− 3, and relative

RMSE = 66%. The NN developed in this study detected accurately the peaks of chla, in both

training and validation sets. The performance of the Case-2-Regional (C2R) algorithm,

routinely used for MERIS data, was also tested and compared with our best performing NN

and the sea-truthing data. Results showed that this NN outperformed the C2R, giving much

higher R2 and lower RMSE values. This study showed that the combination of in situ data

and NN technology improved the retrieval of chla in Case 2 waters, and could be used to

obtain more accurate chla maps. A local-based algorithm for the chla retrieval from an

ocean colour sensor with the characteristics of MERIS would be a great support in the

quantitative monitoring and study of harmful algal events in the coastal waters of the rias

Baixas. The limitations and possible improvements of the developed chla algorithms are

also discussed.

32


2.1 Introduction

Chlorophyll a (chla) is an integrative bioindicator of aquatic ecosystems, as it is

common to almost all photosynthetic organisms, and its concentration is widely used for

estimating phytoplankton biomass in water quality and ecological studies. The chla

concentration is one of the earliest and most commonly used satellite ocean colour

products (Robinson, 2004). From the first dedicated ocean colour scanner (CZCS) to the

currently operated sensors such as SeaWiFS, MODIS and MERIS, various algorithms have

been developed and tested for the retrieval of chla concentration from an ocean colour

spectrum.

Empirical algorithms, based on the ratio between the blue and green light reflected

by the sea or combinations of more spectral bands, are widely used to establish statistical

relationships between ocean colour data (normalized water-leaving radiance or remote

sensing reflectance) and in situ chla (e.g. Gordon & Morel, 1983; Kishino et al., 1998; O'Reilly

et al., 1998). Current standard chla ocean colour derived products are based on empirical

algorithms, such as the OC4v4 for SeaWiFS and the OC3M for MODIS (O'Reilly et al., 2000).

These algorithms have provided valuable data about the temporal and spatial distribution

of chla in Case 1 waters, which according to Morel and Prieur (1977) are dominated by

phytoplankton cells. As in typical Case 2 waters, high concentrations of water constituents

(coloured dissolved organic matter or CDOM, detritus) absorb strongly in the blue

wavelength region, decoupling phytoplankton absorbance, this ratio cannot be used for an

accurate retrieval of chla (Morel & Prieur 1977).

Semi-analytical models constitute another approach for estimating chla from ocean

colour sensors. These models relate the apparent optical properties (AOPs) to the inherent

optical properties (IOPs) of the water (Carder et al., 1999; Maritonema et al., 2002). In Case

2 waters, the semi-analytical algorithms for the retrieval of chla are based on the properties

of the reflectance peak near 700 nm (e.g. Gitelson & Kondratyev, 1991: Gons et al., 2002;

Dall'Olmo et al., 2005; Gitelson et al., 2007; Gitelson et al., 2008). Most of these algorithms

are available with SeaWiFS, MODIS and MERIS for the retrieval of chla concentration in Case

2 waters. An already existing algorithm for Case 2 waters (Gons, 1999) was adapted to

MERIS satellite imagery by Gons et al. (2002) using in situ data from inland waters in

Netherlands, which covered a chla concentration range between 3 and 185 mg m− 3.

The algorithm currently used for chla, as a standard product for Case 2 waters, by

the European Space Agency (ESA) is the MERIS Case-2-Regional Processor (C2R) (Doerffer

http://www.sciencedirect.com/science/article/pii/S0034425710002944#bb0210




















33


& Schiller, 2007; Doerffer & Schiller, 2008). This algorithm takes advantage of the combined

neural network (NN) technique to determine IOPs (absorption of pigments, yellow

substance and scattering of all particles). The NN used is mainly based on North Sea

measurements and is trained with simulated bidirectional water-leaving radiance

reflectances (from 9 MERIS bands between 412 and 708 nm). In their review, Atkinson and

Tatnall (1997) showed the variety of NN applications in remote sensing and their

advantages over typical statistical approaches. Neural networks allow the application of

complex bidirectional radiative transfer models (Thiria et al., 1993), and have been

successfully used, not only for retrieving the concentrations of constituents of coastal and

inland waters, but also for improving the satellite-derived products by minimizing the error

of atmospheric correction (e.g. Keiner & Yan, 1998; Keiner & Brown, 1999; Zhang et al.,

2003; Schiller & Doerffer, 1999; Dzwonkowski & Yan, 2005; Doerffer & Schiller, 2007;

Doerffer & Schiller, 2008).

As spectral reflectance data, acquired from MERIS images for Case 2 waters, are

affected by the concentration of different optically active constituents, the development of

a single chla algorithm in this type of waters might cause unfeasible results, if data from a

normal situation are mixed with data from a specific situation, such as for example

sediment-dominated waters after a storm period. Ideally, the best strategy would probably

be the development of a different chla algorithm for each water type. The difficulty stems

from using only the derived MERIS reflectance spectra to characterize and identify different

water types because the concentrations of the different water constituents are not known

a priori. Several authors (Moore et al., 2001; Moore et al., 2009; Cococcioni et al., 2004;

Ressom et al., 2006) have proposed the application of the fuzzy c-mean (FCM) cluster

algorithm (Bezdek, 1981) to remote sensing reflectance data, to be able to determine

clusters that could be associated to different water types.

There is considerable interest in the accurate estimation of chla for the Galician rias,

because of the economic and social importance of the extensive culture of mussels, and the

high frequency of harmful algal events (GEOHAB, 2005). As part of a routine monitoring

programme, the Technological Institute for the control of the marine environment of

Galicia (INTECMAR) has organized sampling on a weekly basis at 41 stations in the Galicia

rias where, among other water quality parameters, the concentration of chla is recorded.

Although an ocean colour sensor with the characteristics of MERIS is considered to be

adequate for coastal remote sensing, the number of studies using MERIS data in the

Galician rias is very limited (Torres-Palenzuela et al., 2005a; Torres-Palenzuela et al., 2005b).

These are, to our knowledge, the only applications of ocean colour remote sensing in the






















34


Galician rias. In the Western Iberian Peninsula (WIB), data delivered by ocean colour

sensors, such as CZCS, SeaWiFS and MODIS, have been used in previous studies (McClain et

al., 1986; Bode et al., 2003; Ribeiro et al., 2005; Oliveira et al., 2009). However, because of

limited spatial resolution and failure of the retrieval algorithms, these studies focused

particularly on the ocean shelf.

The main aim of this study was to develop and validate neural network-based chla

algorithms for MERIS FR data, which would be specific for the optically complex coastal

waters of the Galician rias, using in situ data sets and fuzzy techniques. The FCM algorithm

was first applied to MERIS reflectance spectra to develop cluster-specific NNs with a scope

predefined by FCM. The performance of the neural network-based chla algorithm is

discussed in comparison to that of other algorithms routinely used for MERIS data, which

are still untested in Galician coastal waters.


2.2.1 Study area

The rias Baixas are located in the northwest part of the Iberian Peninsula (Fig. 2.1),

along the northern boundary of the NW Africa upwelling system. The rias Baixas are formed

by four large coastal embayments from north to south: Muros y Noya, Arousa, Pontevedra

and Vigo all with strong tides and oriented in a SW–NE direction. Initial study in the ria de

Vigo during summer showed that Secchi disk depth ranges from 2 to 12 m. Maximum

concentrations of chla occur in the area primarily in spring and summer (Nogueira et al.,

1997), during the upwelling period that normally lasts from April to October (Tilstone et al.,

1994).

2.2.2 In situ data set

Two data sets of chla concentration were used in this study (Table 2.1). One data set

consisted of analytical measurements of chla concentration provided by INTECMAR from

the monitoring programme established in the study area (Fig. 2.1). It contains

spectrofluorometrically determined chla concentrations from the surface to a depth of 5 m,

for the years 2002–2006.








http://www.sciencedirect.com/science/article/pii/S0034425710002944#f0005







http://www.sciencedirect.com/science/article/pii/S0034425710002944#t0005


35


Fig. 2. 1. Upper right: Iberian Peninsula indicating the Galician rias. Upper left: Location of the sampling stations of the monitoring programme established by INTECMAR in the Galician rias Baixas. From north to south, ria de: Muros y Noya, Arousa, Pontevedra and Vigo. Lower map: Map of the ria de Vigo in relation to MERIS FR pixel size. The location of the ECOSUMMER stations is presented.

The second data set included the results of 5 samplings conducted in 2007 and 2008

on cloud-free days in the ria de Vigo (Fig. 2.1) and ria de Arousa. Water samples were

collected from the surface to a depth of 4 m to determine chloroplast pigments. A High

Performance Liquid Chromatography (HPLC) method, with a reversed phase C8, was used

to separate the pigments. The detailed procedures for pigment extraction and separation

follow those of Zapata et al. (2000). MERIS overpasses were within 30 min to 4 h from the

time that samples were collected in situ. In total, the 228 in situ chla data from 15 cloud-free



36


cruises, which matched-up with MERIS (FR) overpasses, covered a wide range of temporal

and spatial variability. Some 57% of the match-up data came from the ria de Arousa and ria

de Vigo. In situ chla ranged from 0.03 to 7.94 mg m− 3. The highest chla concentration was

recorded in the ria de Pontevedra (Fig. 2.2).

Table 2.1. Data sets used in this study for the periods 2002–2004 and 2006–2008 from the rias Baixas. The numbers of sampling stations per ria are in brackets. n is the number of total observations. Valid match-up refers to the number of observations derived from cloud-free scenes and areas not affected by sun glint, and the range to chla concentrations in near surface waters.

Data set Period Ria n Valid chla

match-up

Range

(mg m -3

)

INTECMAR monitoring

programme 2002-2004

Muros y Noya (7),

Arousa (10),

Pontevedra (11), Vigo

(6)

181 107 0.13-7.94

ECOSUMMER Project 2006-2008 Vigo (12), Arousa (1) 46 43 0.03-6.23

Fig. 2.2. In situ chla (mg m− 3) distribution in the 4 rias.


37


2.2.3 MERIS imagery and data extraction

Fifteen MERIS FR level-1b images over the study area (for 2002–2004 and 2006–

2008) were used for the development and validation of the neural networks (Fig. 2.3). The

dates were selected on the basis of the availability of ground data and the cloud coverage

conditions on the Galician coast. The Doerffer and Schiller (2008) algorithm, including the

Beam-4.6 (Brockmann Consult and contributors, Germany) software's smile correction, was

applied for atmospheric correction. This NN algorithm for dedicated atmospheric

correction over turbid Case 2 waters is based on radiative transfer simulations. More than

200,000 simulated spectra were trained in the NN. The performance test of the NN-based

atmospheric correction showed increasing uncertainty with decreasing values of water-

leaving radiance reflectances. The L2 products of the chla concentrations calculated by the

NNs developed in this study, and those of the MERIS Case 2 Regional Processor (C2R)

(Doerffer & Schiller, 2008), were processed using the same atmospheric correction. The

performance of atmospheric correction over the Galician rias is not within the objectives of

the present study.

Fig. 2.3. Distribution of MERIS FR imagery used in this study for the years 2002–2004 and 2006–2008.

The flags for coastline, land, clouds and invalid reflectance were raised using the

Beam software. No sun-glint effect was observed over the coastal water pixels of the

Galician rias. Data points delivered from cloud-free scenes, and areas that were not flagged

for coastline and invalid reflectance, were considered to be valid match-up data (Table 2.1).

Data from the available MERIS images were extracted and linked to the in situ

database. For each sampling point, the corresponding reflectance values were obtained, as

well as geometry values, including the sun zenith, the view zenith and the difference

between the view and sun azimuths. Regarding the geographical link, the median of





38


9 pixels was computed (approximately 0.8 km2 surface area) around the pixel containing

the exact geographical location of the sampling point. It was considered that a median of

9 pixels was able to reduce MERIS instrument noise and was, therefore, used as a data

quality criterion. Some previously masked pixels were not taken into account in the

computation of the median, because they belonged to land or were suspected of being

affected by low clouds or fog. Moreover, using the median instead of the mean reduced the

effect of possible mixed and non-masked pixels, with extremely high or low values. For

each sampling point, the number of pixels included in the median computation was also

extracted as a quality flag, ranging from 9 (highest quality) to 1 (lowest quality). Low quality

values indicate that the sampling station is located in the proximity of the coast or cloudy or

foggy areas, so that the reflectance values could be affected. The neural models were

developed by selecting the input points based on two different quality levels, to evaluate

the effect of the quality flags on the performance of the model. In addition to reflectance

and geometry values, chla concentrations derived from other algorithms were also

extracted using the same methodology (9-pixel median).

2.2.4 Fuzzy c-means (FCM) clustering

FCM algorithms were applied to MERIS data to determine the number of spectral

clusters in the data set. Cluster-specific NNs were then developed for the retrieval of chla.

From the output data of the FCM application, it is possible to represent different optical

water types, as the analysis is based on reflectances. However, the limited sea-truth

information about water constituents (total suspended material (TSM) and CDOM) and

reflectances prevented us from categorizing the FCM results as different optical water

types. Yet, in this study, FCM were principally used for providing the scope of the

developed NNs.

The FCM technique divides a data set into a specified number of clusters, but unlike

other hard or non-fuzzy clustering algorithms, a given data point (reflectance spectrum)

can belong to more than one class with a membership degree (between 0 and 1), specified

by a membership function defined for each cluster. The FCM algorithm is intended to

minimize an objective function (Moore et al., 2009). Hence, clusters are selected so that the

distance between the data points and the cluster centres is minimized, adjusting the cluster

centres iteratively until the previously established optimization criteria are met, which could

be the maximum number of iterations or the minimum change residual.


39


In this study, an FCM cluster algorithm was applied to MERIS reflectance values

linked to the in situ measurements. A membership function was used, based on the

reciprocal of the distance between points. The algorithm requires two input parameters:

the number of clusters c to be returned by the clustering and the weighting exponent m,

which can be any real number greater than 1. In order to validate the clustering, two

functions were computed: the partition coefficient F (Windham, 1981) and the compactness

and separation index S (Xie & Beni, 1991). The F index is a measurement of the overlap

between clusters, and ranges from 0 (overlap between clusters, weak clustering) to 1 (no

overlap between clusters, strong clustering). The S value is defined as the ratio of the

compactness (measurement of the variance within clusters) to the separation (minimum

distance between cluster centres). Thus, a small S value indicates compact and well-

separated clusters. The partition coefficient F and the compactness and separation index S,

which were used in this study as validity functions, have shown good results in previous

studies (Bezdek, 1981; Xie & Beni, 1991) for the identification of the correct structure in data

sets.

The number of expected clusters was unknown a priori, as we did not have

sufficient in situ data available to be able to define different water types. Therefore, we

established the number of clusters (c) as a function of the maximum degree of separability

between the clusters, defined as a function of the MERIS reflectance values. Hence, data

were clustered several times by varying c from 2 to 8 and m from 1.1 to 3 (Table 2.2), and the

best combination (expected to have a high F value and a small S value) of c and m was

selected.

Table 2.2. Summary of clustering results applying the FCM algorithm. Partition coefficient (F) and compactness and separation index (S) were computed for a range of initial conditions (m, weighting exponent; c, number of clusters).

M

c 1.1 1.5 2 2.5 3

F S F S F S F S F S

2 0.99 0.19 0.91 0.16 0.80 0.17 0.71 0.20 0.64 0.25

3 0.98 0.13 0.87 0.18 0.71 0.16 0.56 0.35 0.45 0.96

4 0.98 0.42 0.85 0.31 0.65 0.30 0.47 0.42 0.31 14.06

5 0.98 0.55 0.85 0.29 0.61 0.21 0.37 2.30 0.25 25.69

6 0.98 0.50 0.84 0.27 0.58 0.33 0.25 7.26 0.21 7.64

7 0.99 0.45 0.82 0.31 0.55 0.34 0.21 15.89 0.18 52.26

8 0.98 0.32 0.82 0.29 0.52 0.41 0.18 18.29 0.16 111.18






40


Once the best FCM algorithm was defined, each data point in the data set was

assigned to the cluster with the highest value for the associated membership function. The

mean reflectance spectrum and the statistics associated to the available in situ data were

calculated in an attempt to examine the possible differences of the water constituents in

the identified clusters.

Classification images were also obtained for the MERIS images involved in the

analysis using the same FCM algorithm. In these images, each sea pixel was assigned to the

cluster with the highest value in its corresponding membership function. Using these

images, the percentage of pixels belonging to each cluster over the study area was

computed.

2.2.5 Multilayer perceptron (MLP)

A multilayer perceptron (MLP) artificial neural network (NN) was implemented for

the retrieval of chla concentration from MERIS images over the study area. In addition to

the preliminary NN using the entire data set, the idea was to develop a different NN for

each cluster derived by applying the FCM algorithm, so that results could be merged to

obtain chlorophyll maps. In practice, only one of the clusters provided a representative and

sufficiently large data set to be able to develop an artificial NN.

An MLP is a feedforward NN that has been widely used in environmental sciences,

because of its ability to approximate a set of input data to the corresponding output data.

Unlike other statistical approaches, MLP does not make assumptions about data

distribution and it can model multivariate, complex and nonlinear data (Cheng &

Titterington, 1994). An MLP consists of a set of nonlinear computational elements, named

neurons or nodes, arranged in multiple layers that are interconnected in a feedforward

way, so that each neuron of a layer is only connected to the nodes of the immediately next

layer, but has no connections to neurons in the previous layers. Each connection is defined

by a set of weight values (Leondes, 1998; Raudys, 2001).

A typical MLP structure includes an input layer, one or more hidden layers and an

output layer. The input layer only distributes the input signals into the network, without

processing them. However, the neurons in the hidden layers and in the output layer

transform their input signal by an activation function.

Once the MLP architecture is designed, the relationship between the input and the

desired output is ultimately dependent on the weight values associated to each connection.






41


These values are established by a supervised learning technique, using a priori information

about the actual output corresponding to a set of input data. This adjusts the network, so

that the best possible approximation to the actual output is achieved. In this particular

case, the weights are iteratively adjusted to minimize an error function, computed as the

mean square difference between the model and the actual output. A back-propagation

learning procedure was used: the learning occurs backwards, layer by layer in a looping

pattern, starting with the output layer and ending with the input layer. The adjustment

continues until no more significant variations in the overall error are observed. Weight

adaptations are produced using a nonlinear optimization method called gradient descent,

which requires a differentiable activation function.

The development of an MLP network implies three phases: the design of the

network, i.e. deciding the different characteristics of the algorithm, including inputs and

outputs, number of hidden layers and the activation function for each layer; the training

phase, in which the previously mentioned back-propagation learning procedure is used

according to a pre-established strategy; and finally the validation phase, in which the

performance of the trained MLP is evaluated using a set of parameters.

Reflectance and geometry data were used as input variables. Reflectance data

included MERIS bands 2 to 13 (except band 11), ranging from 442.5 nm to 865 nm. Band 1

introduced a lot of noise and, therefore, was removed from the analysis. The values were

log-transformed prior to being introduced into the network. Geometry data included sun

zenith, view zenith and difference between the view and the sun azimuth. Although these

values hardly vary in a given image, because the study area was very small compared to the

complete MERIS scene, they showed important variations among different images,

depending on the relative position of the Galician rias. At this point, it should be noted that

MERIS data beyond a solar zenith angle limit of 60° were not considered because of the

sun-glint effect. The range of the viewing zenith angle was 1.44–36.75° from west

(Table 2.3).

Table 2.3. Number of data points (n) belonging to each cluster. View Zenith and chla (mean ± SD and range) per cluster identified.

Cluster n View Zenith (º) chla (mg m-3

)

1 119 12.5±7.3

1.4-36.7

1.9±1.8

0.0-7.9

2 23 17.3±11.0

8.2-36.5

1.5±0.8

0.2-3.0

3 8 12.8±4.4

7.8-22.9

1.0±0.4

0.3-1.5


42


The basic architecture of the MLP network used in this work is shown in Fig. 2.4. It

consists of the following: an input layer with 14 input nodes (11 reflectance values and 3

geometry values); two hidden layers, each one with 4 neurons; and an output layer with a

unique node associated to the desired output, i.e. the chla concentration. The activation

function for both hidden layers is a nonlinear hyperbolic tangent function, while in the

output node it is only applied a bias transfer function.

Fig. 2.4. Basic architecture of the MLP networks used in this work. g is the geometry of the images. The inputs are the MERIS remote sensing reflectance values for 11 bands centred on different wavelengths (nm) and the geometry of the images (g).

Three different NNs were developed using different input data sets: a preliminary

NN including the entire data set (NNRB#1), and two NNs using data points belonging to the

main cluster (Cluster#1): one using the entire cluster data set (NNRB#2), while the other one

only used data points with a quality flag equal to 9 (NNRB#3), thereby, avoiding the

presence of points near the coastline. It should be stressed that the number of data points

included in the development of an NN does not affect its architecture.

The complete data set for each one of the three networks (with a different number

of data points) is divided into two parts: a training set, which includes approximately 80% of

the records to be used in the training phase, and a validating set, with the remaining data


43


points necessary for validating the algorithm. Both subsets must be random and significant

for the whole data set. Therefore, they were created including data points from all the

images and covering the entire range of variation of the chla concentration.

In the training phase, a leave-one-out cross-validation was applied to each training

set to improve the performance of the neural models. Cross-validation is a standard

method for assessing the predictive accuracy of a model in a test sample. Leave-one-out

cross-validation is the best method, when the training set is too small to be able to split it

into significantly sized training and test samples (Fukunaga, 1990). The network is trained N

times (N being the number of elements of the training set), and each time one element is

retained for testing and, therefore, omitted from the training procedure involving the

remaining N−1 elements. Each data point is thereby used exactly once as a test sample.

2.2.6 Performance measurements

For the validation phase, the model fitting was evaluated according to a set of

parameters, which compared the observed chla concentration (ChlO) and the obtained one

using the algorithm (ChlM). These performance measurements were obtained using the

training or/and the validation data sets. The computed parameters were the following:

- Coefficient of determination (R2) between ChlO and ChlM;

- Mean prediction error (MPE) between ChlO and ChlM. MPE is defined as:

where PEi = ChlOi − ChlMi is the prediction error and N is the number of data points;

-Variance of the prediction errors (VAR):

- Root mean square error (RMSE) between ChlO and ChlM:


44


- Relative RMSE:

R2 is a measurement of the correlation between the observed and the predicted

data sets. MPE made it possible to find out if a model tends to underestimate (high positive

values) or overestimate (high negative values) the observed chla concentrations. RMSE and

relative RMSE were used in this work as measurements of absolute error (in mg m− 3) and

relative error respectively, while VAR was used to quantify error variability.

Our results correspond to neural networks selected following a two-step

procedure. In the first stage, the optimal configuration of the network was selected. The

NN model was first trained and validated using the Cluster#1 high-quality (quality flag equal

to 9) data and by varying the input parameters, such as number of hidden layers or

activation functions, to choose the ideal design according to the validation criteria. The

ones selected were those with high R2, MPE near zero and small relative RMSE, RMSE and

VAR, when computed using both the training and the validation set. Overfitting could have

been produced, if only the training set had been used to validate the model. The same

configuration was applied to the other two input data sets (complete data set and

complete Cluster#1 data set). The second step tackles how the fitting result of a neuronal

model can vary between different runs, even though it is trained using the same data, as it

is dependent on the initial value of the weights. Therefore, each one of the three models

was trained N times using leave-one-out cross-validation and, each time, a different data

point was used as the first test sample. Once each one of the NNs was trained, fitting

parameters were computed using both the training and the validation sets, and the best

neural model was selected as in the previous step. Hence, the results shown in this study

are those obtained using the best NN for each case.

2.3 Results

2.3.1 FCM

Table 2.2 shows the results of applying the FCM algorithm to the MERIS reflectance

values associated to in situ data, varying the number of clusters c and the weighting


45


exponent m to establish the optimal number of clusters, as a function of the degree of

separability between them. It can be seen that the best result is achieved with three

clusters and a m value of 1.1. This 3-cluster FCM algorithm was used to assign one cluster to

each one of the 150 valid data points, according to their highest membership grade. Except

for one data point, the selection of the corresponding cluster was unequivocal, with values

of 0.99 for the membership function associated to that cluster. A summary of the cluster

assignment results is shown in Table 2.3. At first sight, the largest cluster (Cluster#1) is seen,

with approximately 80% of the values, and two smaller groups (Cluster#2 and Cluster#3).

Table 2.3 shows basic statistical information about chla and the geometry

associated to the entire data set and to each cluster. Unfortunately, TSM and CDOM data

were unavailable and prevented a more complete characterization of the possible water

types related to the clusters. Fig. 2.5 shows the mean reflectance spectra for each group. It

can be seen that Cluster#2 shows higher reflectance values than Cluster#1 at lower

wavelengths, while Cluster#3 shows higher values throughout the entire spectrum.

Fig. 2.5. Mean MERIS reflectance spectra for each cluster obtained using the 3-class FCM algorithm (c = 3; m = 1.1). Reflectance values are multiplied by 104.




46


Classification images, showing the cluster value for each sea pixel, were also

obtained for the MERIS images (e.g. MERIS image shown in Fig. 2.6) involved in the analysis

(Fig. 2.7a and Fig. 2.8). In theory, the membership grades for each cluster would allow us to

blend the chla concentration, obtained from the different neural networks developed for

each cluster, into a given pixel, so that chla maps with soft transitions would be created

(Moore et al., 2009). In practice, an NN model was only developed for Cluster#1, so that

these classification images were only useful for masking pixels belonging to Cluster#2 or

Cluster#3 (Fig. 2.7b).

Fig. 2.6. RGB composite of a MERIS FR image acquired on 17/11/2003 over the study area.






47


Fig. 2.7. a. Classification of a MERIS image derived from the study area on 17/11/2003. The 3 classes identified using the FCM are shown b. Chla map for the same date, after the application of the NN for high-quality data. Note that gray colour represents pixels that belong to Cluster#2 and Cluster#3.

Fig. 2.8. Classification of a MERIS image derived from the study area on 15/06/2004. The 3 classes identified using the FCM are shown.

Table 2.4 shows the percentage of pixels belonging to each cluster for each image

over the rias Baixas. Cluster#1 includes the majority of the pixels in 12 of 15 images, with

more than 80% of pixels in eight images. On average, 70% of the pixels over the study area

belong to this cluster. Cluster#2 is the predominant one in two images and contains more

than 40% in two other images. Cluster#3 is the least abundant in most of the images, with


48


less than 2% of pixels in ten of them. However, this cluster was also predominant in one

image.

Table 2.4. Percentage of pixels belonging to each cluster over the study area (rias Baixas), obtained from classification images derived from the MERIS images used in this study.

rias Baixas

Cluster#1 Cluster#2 Cluster#3

28/10/2002 30.48 63.22 6.31

08/04/2003 83.69 8.00 8.31

13/05/2003 93.53 4.49 1.98

02/09/2003 52.28 46.63 1.09

04/11/2003 37.39 4.28 58.33

17/11/2003 86.81 3.90 9.30

05/04/2004 72.91 13.44 13.65

15/06/2004 12.00 87.35 0.65

26/07/2004 56.11 43.18 0.71

26/09/2006 98.68 0.15 1.16

10/07/2007 56.22 43.47 0.31

25/07/2007 97.74 1.49 0.77

19/10/2007 97.83 1.48 0.69

09/07/2008 89.55 9.16 1.28

22/07/2008 86.43 12.88 0.69

Average 70.11 22.87 7.02

2.3.2 Neural networks

Table 2.5 summarizes the number of data points (or records) included within each

one of the three data sets used to develop the different NNs.

Table 2.5. Summary of the validation parameters computed using the training and validation data sets for each one of the three different NNs developed in this study.

NN Dataset N R2 MPE VAR RMSE RMSE %

NNRB#1

Complete dataset

(N = 150)

Training

Validation

120

30

0.77

0.63

-0.03

0.08

0.66

1.04

0.81

1.00

67

74

NNRB#2

Cluster#1 dataset

(N = 119)

Training

Validation

92

24

0.86

0.78

0.02

0.02

0.68

0.99

0.68

0.99

60

72

NNRB#3

Cluster#1, Quality = 9

(N = 83)

Training

Validation

66

17

0.97

0.86

0.01

-0.14

0.10

0.57

0.32

0.75

41

66


49


The data sets derived from the main cluster (Cluster#1) are additive, as the most

extended one (NNRB#2) includes all the records belonging to the NNRB#3 data set. A

summary of the validation criteria computed for validation and training data sets used in

the development of each one of the three neural networks is shown in Table 2.5.

Parameters from validation sets were expected to be worse than those attained using the

training sets (Table 2.5), because validation data were not included in the learning

procedure of the algorithm.

Results from the NNs developed using the Cluster#1 data set (NNRB#2 and NNRB#3)

were better than those with the network using the entire data set (NNRB#1). Moreover, the

algorithm (NNRB#3), using only high-quality data points (quality flag equal to 9),

outperforms the neural model (NNRB#2) derived from the entire cluster data set

(Table 2.5).

Fig. 2.9 shows the relationship between the chla concentrations predicted by the

three neural models and the actual values obtained using only the validation set.

Fig. 2.9. Linear relationship between predicted and observed concentrations of chla, considering only the data points included in the validation set, for the three different neural networks: a. NNRB#1 (complete data set); b. NNRB#2 (complete Cluster#1 data set); c. NNRB#3 (Cluster#1 data set, including only data points with quality flag equal 9).





50


A clear linear trend can be seen in all the cases with positive correlations, with

determination coefficients ranging from 0.63 to 0.86. As expected, the best fitting was

obtained when the high-quality Cluster#1 data set was used (Fig. 2.9c). Deviations of the

forecasted chla concentrations from the expected ones (line y = x) are also shown in the

graphics. These deviations were more evident in the network using the complete data set.

Fig. 2.10 shows predicted chla concentrations for each data point, in both the

training set (left part of the graph) and the validation set (right part of the graph), and for

the three networks developed. At first glance, the neural networks fit well the

concentrations shown, particularly in the training set. However, for some records the

model overestimates or underestimates the actual values, preventing a higher correlation.

Fig. 2.10. Observed and predicted chla concentrations for each data point in the training and validation sets (right and left of the vertical division line, respectively), and for the three networks: a. NNRB#1 (complete data set); b. NNRB#2 (complete Cluster#1 data set); c. NNRB#3 (Cluster#1 data set, including only data points with quality flag equal 9). Each black or white strip in the X axis represents a different image, sorted from older to more recent dates. Data points in each image were spatially sorted from north to south.



51


2.3.3 Comparison with other models

To test the performance of the neural network-based chla algorithm developed in

this study, NNRB#3 was applied to all valid Cluster#1 data, and compared with the C2R that

has routinely been used for MERIS data. Fig. 2.11 shows two scatter plots: one of the in situ

chla concentration for the rias Baixas compared to the chla calculated with C2R (Fig. 2.11a),

and the other of the in situ chla concentration compared to the chla calculated with our NN

model (Fig. 2.11b).

Fig. 2.11. Scatter plots of in situ chla concentrations for: a. MERIS C2R Processor and b. NNRB#3 developed in this study. The performance parameters are given for each algorithm.

For the study area, the NNRB#3 developed by us performed better than the C2R,

consequently giving much higher R2 and lower RMSE values. In general, C2R tended to

overestimate the chla concentration (MPE negative). The overestimation was more evident

for the chla concentrations calculated from images derived during the winter months

(lower concentrations of chla) (Fig. 2.12).





52


Fig. 2.12. Average in situ chla concentrations for all sites calculated by the NNRB#3 and C2R in the study area for different days.

2.4 Discussion

2.4.1 FCM

FCM analysis came up with three clusters, representing the different structures that

were found in the data set of reflectances. The 15 images that were used in this study cover

7 different months over 6 years. Although this data set should be a quite good

representation of the chla concentration (Fig. 2.12) and probably of MERIS FR reflectance

found in the area, Cluster#3 was represented by a small number of individual samples

(n = 8) that seem to correspond to a non-typical situation in the Galician rias. There was no

clear pattern to the spatial distribution of the pixels per cluster. Pixels belonging to

Cluster#2 and Cluster#3 appeared close to the coastline, at the innermost and outermost

parts of the rias.

Ideally, in situ data of all water constituents would make possible an optical

classification of the water types in the study area, similar to the classifications mentioned in

(Moore et al., 2001) and (Feng et al., 2005). The limited TSM and CDOM data in an area

where the optical properties of the water were practically unknown, due to the limited

number of ocean colour studies, prevented us from making a qualified separation of water




53


types. However, some simple observations on the different clusters could be made,

assuming accurate atmospheric correction. Cluster#1 pixels contained a wider range of chla

concentrations (0.0–7.9 mg m− 3). For the bands located between 620 and 865 nm, the

reflectance spectra coincided for Clusters#1 and #2. For Cluster#2 pixels, water-leaving

reflectance was high in blue (0.007 sr− 1) and decreased with increasing wavelength,

probably indicating blue waters (Froidefond et al., 2002), while these same pixels were

characterized by chla concentrations ranging from 0.2 to 3.0 mg m− 3. Cluster#3 pixels

showed a broad peak in the reflectance spectrum (Fig. 2.5) at 510 and 560 nm. Chlorophyll

concentration was low in Cluster#3 (range of chla: 0.3–1.5 mg m− 3). Spectral reflectance of

Cluster#3 appeared to be higher in the longer wavelengths, when compared to that for

Cluster#1 and Cluster#2. The higher reflectance values at bands 7–9, observed for the

Cluster#3 data, may be related to atmospheric aerosol features. Five samples of very low

chla concentration, classified in Cluster#3, came from a single image, but from two different

rias on November 17 2003, after a strong storm period.

The advantage of this classification is that the output includes three clusters of

overlapped chla ranges, which can provide the images that reveal the areas where the NN

developed for chla retrieval can be best applied to obtain the most reliable results.

Moreover, these results can be used as a basis for future studies of ocean colour in the

study area, for the collection of in situ water constituents and for the study of particle size

and composition in specific clusters of reflectances.

There were no discernable differences in geometry between the three clusters,

while the degree of mixing was very high. Indeed, the presence of pixels belonging to the

three classes in a single image is quite remarkable, with only small variations in the

geometry values (Fig. 2.7).

The best FCM (c = 3; m = 1.1) algorithm returned three clusters and most of the

points in the data set belonged to Cluster#1. It was the only cluster suitable for the

development of the neural model because, in addition to its size, it covered the entire range

of chla (from 0.03 to 7.94 mg m− 3) seen in the whole data set, and it was the most frequent

cluster in the classification images, appearing in all the MERIS images used in this study.

Chlorophyll mapping with NN models developed using only Cluster#1 required the

masking of pixels belonging to other clusters. Indeed, mapping might have been impractical

if Cluster#2 and Cluster#3 pixels had been predominant over the study area. However, their

presence did not prevent continuous chlorophyll mapping over large areas, because the

predominance of Cluster#1, observed in the data set, was also reflected in the images.

Hence, the classification images showed that Cluster#1 pixels were the most abundant over




54


the rias Baixas area, in twelve of the fifteen images (Table 2.4). Chla mapping of the four rias

Baixas was achievable in eight images, while in another six images complete mapping of at

least one of the rias was feasible. The remaining image (June 15 2004) was more

problematic (Fig. 2.8), although the mapping of a small part in the ria de Arousa was also

possible.

Fig. 2.7b shows a chla map of the study area (November 17 2003) calculated by the

NNs. Chla concentration appears to be higher in adjacent coastal areas than in the rias

Baixas. The observed pattern of chla spatial distribution may be the result of a winter

upwelling event that is unusual, but has been recorded and described in the area in

previous studies (Álvarez et al., 2003; Prego et al., 2007).

2.4.2 Neural networks

The NN trained with the complete data set (NNRB#1) produced well-correlated

results (R2 = 0.77 in the training set; R2 = 0.63 in the validation set). However, although this

model performs well with low-medium chlorophyll values, it does not work properly with

values greater than 4 mg m− 3, which show a mean absolute prediction error greater than

1 mg m− 3. In addition, it does not detect some peaks of high or low chlorophyll (Fig. 2.10a).

If the validation set is considered overall, the model tends to slightly underestimate

chlorophyll values (MPE = 0.08), with a higher percentage of data points appearing below

the identity line (Fig. 2.9a). The main drawback of this NN is that data points with similar in

situ chlorophyll values were underestimated or overestimated depending on the cluster

associated to them. This effect is clearly seen in Fig. 2.9a, for two records with in situ values

of around 5.5 mg m− 3. Therefore, this chlorophyll-mapping model might produce unreliable

results, despite the good fitting results.

The fitting results, using the neural networks trained with the Cluster#1 data set,

were quite good and demonstrated the capability of the tool to predict chla concentrations

in the study area. The best results were achieved with NNRB#3, which was developed using

only high-quality data points. This model detected the peaks of high chla concentration

(Fig. 2.10c) in both the training and validation sets. The negative MPE value (MPE = − 0.14)

and a higher percentage of data points over the identity line (Fig. 2.9c) are indicative of a

slight overestimation of the chla concentrations, when only the validation set is considered,

preventing a better fitting in comparison with the training set. The deviation seen for the

expected values (identity line) is greater at high chla concentrations, but this could be due

to the small number of data points with chla peaks in the validation set, which prevents











55


making a better adjustment. If the training data set is analysed, in addition to the high

correlation (R2 = 0.97) the error parameters are small (Rel. RMSE = 41; RMSE = 0.32;

VAR = 0.10), when compared to other models. This is because around 85% of the data points

show absolute prediction errors (PE) that are lower than 0.5 mg m− 3, while the

overestimation effect observed in the validation set is hardly noticeable (MPE = 0.01).

The results of the network trained using the complete Cluster#1 data set (NNRB#2)

were worse than those obtained only using the high-quality data points (NNRB#3). The

inclusion of pixels with a quality level below nine implies that each one of such pixels is next

to at least one invalid pixel (classified as land, cloud or fog) and, therefore, its reflectance

value could be affected in some way. In the most common case of pixels near the coastline,

there could be a bottom effect on reflectance or the presence of macroalgae and adjacency

effects could also affect reflectance. However, this is difficult to evaluate, because it does

not merely depend upon depth, but also on tidal height, illumination conditions and the

type of bottom. To sum up, there is no reliable way to find out whether or not a reflectance

value is affected and, therefore, it is not known if an error is being introduced. Data of in

situ reflectance may solve this problem.

Despite the uncertainty, NNRB#2 showed correlations greater than 0.75 and an

RMSE value lower than 1 mg m− 3 for both the training and the validation sets. However, this

model presented some problems when compared to NNRB#3: it did not detect some peaks

of chlorophyll values greater than 6 mg m− 3 (Fig. 2.10b), and it also overestimated a record

in the validation set, by giving it a value of 9.4 mg m− 3 (Fig. 2.9 and Fig. 2.10) that was larger

than the maximum concentration found in the in situ data set. In general, excluding this

record, the NNRB#2 model tended to slightly underestimate chlorophyll values (MPE = 0.2

for the training and validation sets), as in seen in Fig. 2.9b, where most of the points appear

below the identity line.

2.4.3 Comparison with other models

Even if there are ongoing efforts for the validation of the C2R algorithm, previously

conducted validation studies have shown good results, especially for the North Sea (Peters,

2006), which is the relevant area of the data set for the development of this algorithm.

Furthermore, in inland waters results showed reasonably good correlation between the

MERIS algal and in situ chla concentrations (Alikas & Reinart, 2008; Odermatt et al., 2010). In

our study, the C2R showed a nonlinear relationship with the in situ data. This may be due to

the relatively low concentrations (mainly < 3 mg m− 3) of chla in the study area. The scope of










56


the C2R algorithm covers a much wider range of chla. Regional characteristics of the study

area, such as upwelling and the numerous small rivers that introduce water into the ria, can

lead to a temporal and spatial change of the phytoplankton abundance and composition, as

well as changes of the optical properties of the water. However, C2R performed much

better for the period 2006–2008.

57


2.5 References

Alikas, K., & Reinart, A. (2008). Validation of the MERIS products on the large European lakes:

Peipsi, Vänern and Vättern. Hydrobiologia, 599, 161−168.

Álvarez, I., deCastro, M., Prego, R., & Gómez-Gesteira, M. (2003). hydrographic characterization

of a winter-upwelling event in the Ria of Pontevedra (NW Spain). Estuarine Coastal and Shelf

Science, 56, 869−876.

Atkinson, P. M., & Tatnall, A. R. (1997). Introduction neural networks in remote sensing.

International Journal of Remote Sensing, 18, 699−709.

Bezdek, J. C. (1981). Pattern recognition with fuzzy objective function algorithms. New York:

Plenum.

Bode, A., Carrera, P., & Lens, S. (2003). The pelagic foodweb in the upwelling ecosystem of

Galicia (NW Spain) during spring: Natural abundance of stable carbon and nitrogen isotopes.

ICES Journal of Marine Sciences, 60, 11−22.




Geophysical Research, 104, 5403−5421.

Cheng, B., & Titterington, D. M. (1994). Neural networks: A review from a statistical perspective.

Statistical Science, 9, 2−54.

Cococcioni, M., Corsini, G., Lazzerini, B., & Marcelloni, F. (2004). Approaching the ocean color

problem using fuzzy rules. IEEE Transactions on Systems, Man and Cybernetics—Part B:

Cybernetics, 34(3), 1360−1373.

Dall'Olmo, G., Gitelson, A. A., Rundquist, D. C., Leavitt, B., Barrow, T., & Holz, J. C. (2005).

Assessing the potential of SeaWiFS and MODIS for estimating chlorophyll concentration in

turbid productive waters using red and near-infrared bands. Remote Sensing of Environment, 96,

176−187.

58



Remote Sensing, 28, 517−535.

Doerffer, R., & Schiller, H. (2008). MERIS regional coastal and lake case 2 water project—

Atmospheric correction ATBD. GKSS Research Center 21502 Geestacht, Germany. Version 1.0, 18

May 2008.

Dzwonkowski, B., & Yan, X. H. (2005). Development and application of a neural network based

ocean colour algorithm in coastal waters. International Journal of Remote Sensing, 26, 1175−1200.

Feng,H., Campbell, J.W., Dowell,M.D. &Moore, T. S. (2005).Modeling spectral reflectance of

optically complex waters using bio-optical measurements from Tokyo Bay. Remote Sensing of

Environment, 99, 232−243.

Froidefond, J. M., Gardel, L., Guiral, D., Parra, M., & Ternon, J. F. (2002). Spectral remote sensing

reflectances of coastal waters in French Guiana under the Amazon influence. Remote sensing of

Environment, 80, 225−232.

Fukunaga, K. (1990). Introduction to statistical pattern recognition, 2nd ed. San Diego, CA:

Academic Press.

GEOHAB (2005). Global ecology and oceanography of harmful algal blooms. Paris and Baltimore:

IOC and SCOR.


productive estuaries: Chesapeake Bay case study. Remote Sensing of Environment, 109, 464−472.

Gitelson, A. A., Dall'Olmo, G., Moses, W., Rundquist, D. C., Barrow, T., Fisher, T. R., Gurlin, D.,

&Holz, J. (2008). A simple semi-analytical model for remote estimation of chlorophyll-a in turbid

waters: Validation. Remote Sensing of Environment, 112, 947−951.

Gitelson, A. A., & Kondratyev, K. Y. (1991). Optical models of mesotrophic and eutrophic water

bodies. International Journal of Remote Sensing, 12, 373−385.

Gons,H. (1999). Optical teledetection of chlorophyll a inturbid inland waters. Environmental

Science & Technology, 33, 1127−1132.

59


Gons, H. J., Rijkeboer, M., & Ruddick, K. G. (2002). A chlorophyll-retrieval algorithm for satellite

imagery (medium resolution imaging spectrometer) of inland and coastal waters. Journal of

Plankton Research, 24, 947−951.

Gordon, H. R., & Morel, A. Y. (1983). Remote assessment of ocean colour for interpretation of

satellite visible imagery: A review. New York: Springer Verlag.

Keiner, L. E., & Brown, C. W. (1999). Estimating oceanic chlorophyll concentrations with neural

networks. International Journal of Remote Sensing, 20, 189−194.

Keiner, L., & Yan, X. H. (1998). The use of a neural network in estimating surface chlorophyll and

sediments from Thematic Mapper imagery. Remote Sensing of the Environment, 66, 153−165.

Kishino, M., Ishimaru, T., Furuya, K., Oishi, T., & Kawasaki, K. (1998). In-water algorithm for

ADEOS/OCTS. Journal of Oceanography, 54, 431−436.

Leondes, C. T. (1998). Neural network systems techniques and applications. New York: Academic

Press.

Maritonema, S., Siegel, D. A., & Peterson, A. R. (2002). Optimization of a semianalytical ocean

color model for global-scale applications. Applied Optics, 41, 2705−2714.

McClain, C. R., Chao, S. -Y., Atkinson, L. P., Blanton, J. O., & Castillejo, F. (1986). Winddriven

upwelling in the vicinity of Cape Finisterre, Spain. Journal of Geophysical Research, 91,

8470−8486.

Moore, T. S., Campbell, J. W., & Dowell, M. D. (2009). A class-based approach to characterizing

and mapping the uncertainty of the MODIS ocean chlorophyll product. Remote Sensing of

Environment, 113, 2424−2430.

Moore, T. S., Campbell, J. W., & Feng, H. (2001). A fuzzy logic classification scheme for selecting

blending ocean color algorithms. IEEE Transactions on Geoscience and Remote Sensing, 39,

1764−1776.


22, 709−722.

60


Nogueira, E., Pérez, F. F., & Ríos, A. F. (1997). Seasonal patterns and long-term trends in an

estuarine upwelling ecosystem (Ría de Vigo, NW Spain). Estuarine, Coastal and Shelf Science, 44,

285−300.

Odermatt, D., Giardino, C., & Heege, T. (2010). Chlorophyll retrieval with MERIS Case-2-Regional

in perialpine lakes. Remote Sensing of Environment, 114, 607−617.

Oliveira, P., Nolasco, R., Dubert, J., Moita, T., & Peliz, A. (2009). Surface temperature,

chlorophyll and advection patterns during a summer upwelling event off central Portugal.

Continental Shelf Research, 29, 759−774.

O'Reilly, J. E., Maritorena, S., Mitchell, G., Siegel, D. A., Carder, K. L., Garver, S. A., Kahru, M., &

McClain, C. (1998). Ocean color chlorophyll algorithms for SeaWiFS. Journal of Geophysical

Research, 103, 24937−24953.

O'Reilly, J. E., Maritorena, S., O'Brien, M. C., Siegel, D. A., Toole, D., Menzies, D., et al. (2000).

Ocean color chlorophyll a algorithms for SeaWiFS, OC2 and OC4: Version 4. In S. B. Hooker &

E.R. Firestone (Eds.), NASA technical memorandum 2000-206892, Vol. 11 49pp.

Peters, S. Water Product Validation Case II waters. Validation at the North Sea. In, Second

Working Meeting on MERIS and AATSR Calibration and Geophysical Validation CD SP-615.

Frascati, Italy.

Prego, R., Guzmán-Zuñiga, D., Varela, M., deCastro, M., & Gómez-Gesteira, M. (2007).

Consequences of winter upwelling events on biogeochemical and phytoplankton patterns in a

western Galician ria (NW Iberian Peninsula). Estuarine Coastal and Shelf Science, 73, 409−422.

Raudys, S. (2001). Statistics and neural classifiers. An integrated approach to design. London:

Springer.

Ressom, H. W., Turner, K., & Musavi, M. (2006). Estimation of ocean water chlorophyll-a

concentration using fuzzy c-means clustering artificial neural networks. International Joint

Conference on Neural Networks (pp. 4118−4125). Vancouver, Canada: IEEE.

61


Ribeiro, A., Peliz, A., & Santos, A. M. P. (2005). A study of the response of chl-a biomass to a

winter upwelling event off western Iberia using SeaWiFS and in situ data. Journal of Marine

Systems, 53, 87−107.

Robinson, I. (2004). Measuring the oceans from space: The principles and methods of satellite

oceanography. Chichester, UK: Spinger.

Schiller, H., & Doerffer, R. (1999). Neural network for emulation of an inverse model operational

derivation of Case II water properties from MERIS data. International Journal of Remote Sensing,

30, 1735−1746.

Thiria, S., Mejia, C., Badran, F., & Crepon, M. (1993). A neural network approach for modelling

nonlinear transfer functions: Application for wind retrieval from spaceborne scatterometer

data. Journal of Geophysical Research, 98(12), 22827−22841.

Tilstone, G. H., Figueiras, F. G., & Fraga, F. (1994). Upwelling–downwelling sequences in the

generation of red tides in a coastal upwelling system. Marine Ecology Progress Series, 112,

241−253.

Torres-Palenzuela, J. M., Vilas-González, L., & Mosqura-Giménez, Á. (2005a). Correlation between

MERIS and in-situ data for study of Pseudo-nitzschia spp. toxic blooms in Galician coastal area (pp.

497−507). Rotterdam: Millpress.

Torres-Palenzuela, J. M., Vilas-González, L., & Mosqura-Giménez, Á. (2005b). Detection of

Pseudo-nitzschia spp. toxic blooms using MERIS images on the Galician coast (pp. 1915−1919). :

European Space Agency (Special Publication).

Windham, M. (1981). Cluster validity for fuzzy clustering algorithms. Fuzzy Sets and Systems, 5,

177−185.

Xie, X., & Beni, G. (1991). A validity measure for fuzzy clustering. IEEE Transactions on Pattern

Analysis and Machine Intelligence, 13, 841−847.

Zapata, M., Rodríguez, F., & Garrido, J. L. (2000). Separation of chlorophylls and carotenoids

from marine phytoplankton: A new HPLC method using a reversed phase C8 column and

pyridine-containing mobile phases. Marine Ecology Progress Series, 195, 29−45.

62


Zhang, T., Fell, F., Liu, Z. -S., Preusker, R., Fisher, J., & He, M. -X. (2003). Evaluating the

performance of artificial neural network techniques for pigment retrieval from ocean color in

Case I. Journal of Geophysical Research, 108, 3286−3298.

Chapter III. Chla mapping during an upwelling event

63

Application of a regionally specific

chlorophyll a algorithm for MERIS

FR data during an upwelling cycle

Abstract

3.1 Introduction


3.3 Results & Discussion

3.4 References

CHAPTER III

65


Abstract

This study takes advantage of a regionally specific algorithm and the characteristics

of Medium Resolution Imaging Spectrometer (MERIS) in order to deliver more accurate,

detailed chlorophyll a (chla) maps of optically complex coastal waters during an upwelling

cycle. MERIS full resolution chla concentrations and in situ data were obtained on the

Galician (NW Spain) shelf and in three adjacent rias (embayments), sites of extensive

mussel culture that experience frequent harmful algal events. Regionally focused

algorithms (Regional neural network for rias Baixas or NNRB) for the retrieval of chla in the

Galician rias optically complex waters were tested in comparison to sea-truth data. The one

that showed the best performance was applied to a series of six MERIS (FR) images during

a summer upwelling cycle to test its performance. The best performance parameters were

given for the NN trained with high-quality data using the most abundant cluster found in

the rias after the application of fuzzy c-mean clustering techniques (FCM). July 2008 was

characterized by three periods of different meteorological and oceanographic states. The

main changes in chla concentration and distribution were clearly captured in the images.

After a period of strong upwelling favourable winds a high biomass algal event was

recorded in the study area. However, MERIS missed the high chlorophyll upwelled water

that was detected below surface in the ria de Vigo by the chla profiles, proving the necessity

of in situ observations. Relatively high biomass “patches” were mapped in detail inside the

rias. There was a significant variation in the timing and the extent of the maximum chla

areas. The maps confirmed that the complex spatial structure of the phytoplankton

distribution in the rias Baixas is affected by the surface currents and winds on the adjacent

continental shelf. This study showed that a regionally specific algorithm for an ocean colour

sensor with the characteristics of MERIS in combination with in situ data can be of great

help in chla monitoring, detection and study of high biomass algal events in an area

affected by coastal upwelling such as the rias Baixas.

66


3.1 Introduction

Although remote sensing tools can be used with a relatively high precision at global

scale for the calculation of chlorophyll a (chla), they are not always totally accurate in local

areas (Cota et al., 2004; Ruddick et al., 2008) and highly dynamic systems such upwelling

regimes. Eastern boundary upwelling systems cover a small percentage of the ocean

surface, but account more than 20% of the global fish catch.

In these high productive systems harmful algal events due to toxic phytoplankton

species and/or high-biomass blooms pose an increasing threat for aquaculture and fishing

industries, ecosystem health and diversity and have possible implications for human health

and activities (Trainer et al., 2010). Harmful algal events in eastern boundary upwelling

systems have been closely associated with the wind properties (Bates et al., 1998; Pitcher et

al., 1998) and have become a focal point of numerous studies (e.g. Kudela et al., 2005;

Pitcher & Nelson, 2006; Fewcett et al., 2007). For example, in their review on harmful algal

events in upwelling systems, Pitcher et al. (2010) noticed that variations in wind-stress

fluctuations and buoyancy inputs in upwelling systems are controlling factors of the bloom

timing and pointed out the role of inner-shelf dynamics on the spatial distribution of the

bloom.

Upwelling waters are characterized by considerable variability in the vertical

distribution of phytoplankton (Brown & Hutchings, 1987) and in their optical properties

(Morel & Prieur, 1977). Optically active water constituents such as SPM which are brought

into the surface because of the strong mixing and vertical advection that takes place during

upwelling events may vary independently of the surface chla, as they do in typically shallow

estuarine case II waters.

In typical case II waters the traditional satellite-derived chla models (empirical:

Muller-Karger et al., 1990; Evans & Gordon, 1994; Aiken et al., 1995; O´ Reilly et al., 2000;

McClain et al., 2004; Brown et al., 2008 and semi-analytical: Carder & Steward, 1985; Carder

et al., 1999) based on the ratio between the radiance of blue and green light reflected by

the surface waters cannot be used for an accurate retrieval of chla (Morel & Prieur 1977;

Gons, 1999; Gitelson et al., 2007). However, chla algorithms that use green to red and near

infrared band ratios have shown good performance in inland and coastal waters (Gilerson

et al., 2010). In the effort for more accurate retrieval of water constituents in optically

complex waters, neural network (NN) techniques can play an important role, since they

seem ideal for multivariate, complex and non-linear data modeling (Thiria et al., 1993).

67


Dransfeld et al., (2004) emphasised the role of NNs in the retrieval of water constituents

especially in Case 2 waters. In recent decades the application of neural network (NN)

techniques for the estimation of selected water quality parameters from ocean-colour has

increased (Atkinson & Tatnall, 1997; Keiner & Yan, 1998; Dzwonkowski & Yan, 2005; Zhang

et al., 2003; Shahraiyni et al., 2009). NN based algorithms are currently used as standard

products for the estimation of chla, suspended particulate matter (SPM) and yellow

substances by the European Space Agency (ESA) for Medium Resolution Imaging

Spectrometer (MERIS) data (Doerffer & Schiller, 2007, 2008).

The Galician rias are V-like embayments along the northwest part of the Iberian

Peninsula formed by sunken river valleys flooded by the sea, whose ecosystems are

strongly influenced by oceanic conditions on the adjacent continental shelf. Interest in

developing an accurate estimation of chla in these rias is considerable, mainly because of

the economic and social importance of the extensive culture of mussels (Rodríguez

Rodríguez et al., 2011), and the frequent occurrence of harmful algal events (GEOHAB,

2005).

Although MERIS is an ocean colour sensor with characteristics considered suitable

for chla monitoring and detection of HABs in coastal areas (Doerffer et al., 1999), to our

knowledge the studies using MERIS data in the Galician rias are limited to those of Torres

Palenzuela et al. (2005a; b), Spyrakos et al., (2010) and González Vilas et al. (2011). The latter

authors developed a chla algorithm based on NNs and classification techniques from MERIS

full resolution data for rias Baixas coastal waters. Previous ocean colour studies by satellite

sensors (CZCS, SeaWiFS, MODIS) during active upwelling in the Iberian system (McClain et

al., 1986; Peliz & Fiuza, 1999; Joint et al., 2002; Bode et al., 2003; Ribeiro et al., 2005; Oliveira

et al., 2009a, Oliveira et al., 2009b) played an important role in the identification of chla

patterns and study of harmful algal blooms and primary production but were restricted to

the ocean shelf because of insufficient spatial resolution. Another problem that affected

many of these previous satellite remote sensing studies in the area was the failure of the

algorithms used to provide reliable chla data during upwelling favourable conditions

especially in the areas closest to the coast.

In the present chapter a set of neural network-based chla algorithms previously

developed for the Galician rias waters (within the rias and for coastal waters on the

continental shelf) are applied for the first time in a short series of MERIS (FR) images

delivered during an upwelling cycle in order to obtain maps of chla. This study tests the

potential of the algorithms to map the spatial extent of possible algal blooms caused by

coastal upwelling. Also, the temporal and spatial distribution of the chla patterns, captured

68


in the MERIS images using the local adapted algorithm, are discussed in relation to the

meteorological and oceanographic conditions in the area. Finally, the performance of the

neural network-based chla algorithm is compared to in situ measurements.

3.2 Methods and data

3.2.1 Study area

The rias Baixas constitute the southern part of the Galician rias (Fig. 3.1). This study

focuses on three rias (Arousa, Pontevedra and Vigo), each connected to the open sea

through two entrances, to the north and south of the islands located at the external part of

each ria.

In these highly primary productive upwelling estuarine systems (Fraga, 1981, Torres

& Barton, 2007; Spyrakos et al., in press) transient increases of phytoplankton abundance,

referred to as blooms, are a frequent phenomenon occurring mainly between early spring

and late fall (Fraga et al., 1988; Varela, 1992, Figueiras & Ríos, 1993). Sporadically, some

phytoplankton blooms in the Galician rias are perceived as harmful with direct and indirect

impacts to the mussel production that constitute an important economic activity in the

area. Harmful algal events in the Galician rias are a well documented phenomenon

(Margalef, 1956; Figueiras & Niell, 1987; Figueiras & Ríos, 1993; Tilstone et al., 1994; Figueiras

et al., 1994; Nogueira et al., 1997; Tilstone et al., 2003; GEOHAB, 2005). Harmful events due

to the Paralytic shellfish toxin (PST) producer Gymnodinium catenatum have been

occasionally (1976: Estrada et al., 1984; 1985: Laiño, 1991; 2005: Bravo et al., 2010a; b) or/and

annually (1985-1995: Pazos et al., 2006) recorded in the Galician rias. It is generally

considered (Fraga et al., 1990; Figueiras et al., 1996) that advection of warmer waters from

the shelf into the rias at the end of the upwelling season coincides with the highest

abundances of G. catenatum. Pazos et al. (2006) observed a northward progression of G.

catenatum along the Iberian Peninsula starting on the Portuguese coast, suggesting this

could provide an early notice of G. catenatum events in the Galician rias.

69


Fig. 3.1. A) Galician coast and bathymetry of the area. From north to south the rias Baixas: Muros y Noya, Arousa, Pontevedra and Vigo. The location of the Seawatch buoy station off Cabo Silleiro is shown by a black rectangle. B) Map of ria de Vigo showing the locations of the sampling stations. The MERIS FR pixel size is presented in relation to the size of the ria.

70


3.2.2 Sampling regime

Two samplings were conducted in 2008 in the ria de Vigo. Twelve fixed stations

were visited on cloud-free days (July 9 and 22). The sampling transect was extended from

the open sea towards the inner part of the ria. Satellite data from MERIS (FR) were

available for the same days. The depth of the sites, where samples were collected, ranged

from 5 m inside the ria to 100 m outside. Triplicate water samples from surface to 3 m were

collected at each station (Fig. 3.1B) from a sampler (3524 cm3) for the determination of chla

and SPM.

3.2.3 In situ measurements

In situ chla fluorescence profile was monitored by a Turner designs CYCLOPS-7

submersible fluorometer. Profiles of water temperature were provided by a portable meter

(HI 9829, Hanna instruments). The depth of the euphotic zone was established with a

Secchi disk. For the High Performance Liquid Chromatography (HPLC) chla determination,

water samples (100-200mL) were filtered through a 9mm diameter Whatman GF/F filters

and stored at -80oC for 2 weeks, and 95% methanol was used as extraction solvent for the

pigments. In this study only chla concentration data are presented, calculated as the sum of

chlorophyllide a, chlorophyll a epimer, chlorophyll a allomer and divinyl chlorophyll a. An

HPLC method using a reversed phase C8 was applied for the separation of the pigments.

Details of pigment extraction and separation are provided in Zapata et al. (2000).

SPM was evaluated in terms of SPM concentration and percent weight of organic

matter (%OM). Pre-combusted (450 oC for 24 h), pre-washed in 500 mL of MilliQ, 47mm

Whatman GF/F filters were used. These filters were then dried at 65 oC to a constant weight.

Particles were collected by filtering a standard volume (1000mL) of seawater samples and

then rinsed with 50mL MilliQ in order to remove salts and dissolved organic matter. For the

determination of SPM the filters were dried at 65oC till no weight changes were observed.

The filters were then re-combusted at 450 ºC for 5 h in order to obtain the inorganic

suspended matter (ISM). The percent weight of organic matter (%OM) was determined by

subtracting the ISM from the SPM. All the filters were weighted on a Precisa 262 SMA-FR

microbalance (10-5 g precision).

71


Table 3.1. MERIS imagery showing the acquisition time (UTC) and mean view zenith angle from west. Sea-truthing mean values (± S.D.) of chlorophyll a (chla), suspended particulate matter (SPM), percentage of inorganic contribution to SPM and Secchi disk depth (Zsd) for ria de Vigo (12 stations) during the two samplings.

Chla (mg m-3

) 0.74±0.80 0.97±0.82

SPM (mg L-1

) 1.85±0.30 2.01±0.65

inorganic matter (%) 48.17±6.21 48.20±7.70

Zsd (m) 7.65±3.30 4.45±1.66

MERIS FR July 03 2008 July 09 2008 July 16 2008 July 19 2008 July 22 2008 July 29 2008

Acquisition time (UTC) 10:59 11:10 10:50 10:56 11:02 10:42

View zenith angle (º) 13.5 13.0 20.7 15.3 11.7 20.7

3.2.4 Oceanographic and meteorological data

Oceanographic and meteorological data off the rias Baixas were provided by the

Spanish Port System (www.puertos.es). More specifically, wind speed (W) and direction,

current data and water temperature were observed at a Seawatch buoy station located off

Cape Silleiro (42° 7.8′N, 9° 23.4′W) (Fig. 3.1). This meteorological station was selected as

fairly representative of the study area (Herrera et al., 2005). Daily upwelling index (IW) was

estimated from wind by Bakun's (1973) method:

IW = - y/( W·f)= - 1000· a·CD·W·Wy/( W·f) m3/(s·km) (1)

where y is the alongshore component of wind stress N·m-2, W is the density of seawater

(1025 kg·m-3), f is the Coriolis parameter (9.9·10-5 s-1 at 42º latitude, a is the density of air (1.2

kg·m-3 at 15ºC), CD is an empirical dimensionless drag coefficient (1.4·10-3 according to Hidy,

1972) and W and Wy are the average daily modulus and northward component of the wind.

Moderate Resolution Imaging Spectroradiometer (MODIS-Aqua) sea surface

temperature (SST) daily level 2 data for July 2008 were downloaded from the website of

the National Aeronautics and Space Administration Goddard Space Flight Center (NASA-

GSFC) (http://oceancolor.gsfc.nasa.gov/). The 1 × 1 km resolution MODIS data were

processed using MATLAB software to derive projected SST maps of the study area. The

study area for the SST maps was expanded to 42-43o N and 9.3-8.3o W. The MODIS imagery

contains 6 images for the dates that MERIS data were available.

http://www.puertos.es/

http://www.sciencedirect.com/science?_ob=RedirectURL&_method=externObjLink&_locator=url&_cdi=5824&_issn=00344257&_originPage=article&_zone=art_page&_plusSign=%2B&_targetURL=http%253A%252F%252Foceancolor.gsfc.nasa.gov%252F

72


3.2.5 MERIS data and MERIS chla algorithms for rias Baixas

3.2.5.1 Regional neural network for rias Baixas (NNRB)

This set of algorithms represents feed-forward NNs trained by supervised learning

using iterative back-propagation of error for the retrieval of chla from MERIS FR data

(Chapter II). These algorithms approximate sets of different classes of water-leaving

radiance reflectances data, determined after the application of Fuzzy c-means clustering

techniques (FCM), to a set of appropriate chla concentrations. Input variables are 11 MERIS

water leaving radiance reflectance and 3 geometry values. It was found that MERIS data

can be classified in 3 clusters (#1, #2 and #3) but only one could be used for the 3 different

NNs that were developed for the retrieval.

The method performs well in the estimation of chla from MERIS (FR) data in the

optically complex waters of the rias Baixas and detects accurately the peaks of chla. NNRB

is based on in situ chla data collected from the rias Baixas during a long period survey (2002-

2008), covering the temporal variability of chla in all the part of the rias. In contrast with

the Doerffer and Schiller algorithm, this algorithm does not use simulated data. The result is

a narrower range (0.03-7.73 mg m-3), but this is considered as sufficient for the study area.

3.2.5.2 Application of chla algorithms to MERIS imagery

The MERIS satellite imagery used in this study contains 6 full-resolution level-1b

images derived from the area in July 2008. MERIS overpasses were within 2 h of the time

that samples and data were collected in situ. Beam 4.6 (Brockmann Consult and

contributors, Germany) software was used for the analysis of the imagery.

The BEAM-4.6´s smile correction was applied to the original level-1b data. For the

atmospheric correction the ocean colour data were processed with a NN-based algorithm

implemented alongside the MERIS Case-2-Regional Processor (C2R) which was developed

by Doerffer and Schiller (2008). This NN algorithm for dedicated atmospheric correction

over turbid case 2 waters is based on radiative transfer simulations. The performance test

of the atmospheric correction showed increasing uncertainty with decreasing values of

water leaving radiance reflectances (Doerffer & Schiller, 2008).

The flags for coastline, land, clouds and invalid reflectance were raised using the

Beam software. Ocean colour data derived from areas significantly affected by sun glint

73


(beyond a solar zenith angle limit of 60o) were characterized invalid and removed from the

analysis.

The FCM algorithm that is presented in Chapter II was applied to the level-2

reflectance data in order to identify the different clusters. Classification images were then

obtained for the available MERIS images using the same FCM algorithm. The pixels in these

images were assigned to the cluster with the highest value in its corresponding

membership function and the percentage of pixels belonging to each cluster was

computed. The performance of the available algorithms was then tested and the NN with

the best performance measures was applied to the MERIS leaving radiance reflectance

values in order to deliver the chla maps for the study area.

Chla data points delivered from cloud-free scenes and areas that were not flagged

for coastline and invalid reflectance were considered to be valid match-up data and were

used for the performance testing of the chla algorithms. Water-leaving radiance

reflectances and chla concentrations were computed as mean values of the pixel

corresponding to the sampling station location and the 8 surrounding pixels. These 9 pixels

cover approximately 0.8 km2 of surface area and it was considered that this averaging was

able to reduce MERIS instrument noise. Although MERIS spatial resolution of chla is

considered suitable to study coastal areas there are cases (e.g. Kutser, 2004) where

variability occurs even within one MERIS pixel. Chla spatial variability in the Galician rias has

not been studied in sufficient detail to determine sub-pixel variability, but available

information from previous campaigns in the ria de Vigo using MiniBAT undulating vehicle

and NERC-CASI instrument (http://www.iim.csic.es/~barton/cria) and previous studies on

the spatial distribution of algae (dinoflagellates and diatoms) blooms from station samples

in the area (Figueiras & Ríos, 1993; Bravo et al., 2010b) show variability well represented by

MERIS.

For each sampling point, the number of pixels included in the median computation

was also extracted as a quality flag, ranging from 9 (highest quality) to 1 (lowest quality).

Low quality values indicate that the sampling station is located in the proximity of the coast

or cloudy or foggy areas, so that the reflectance values could be affected.

The imagery was then remapped using the standard Mercator projection with a

fixed grid of 890 by 890 pixels. Each chla image ranges from 42o 04′ N to 42 o 40′ N latitude

and from 8o 32′W to 9o 32′W longitude, which covers approximately 3.1 x 103 km2.

http://www.iim.csic.es/~barton/cria

74


3.3 Results and discussion

3.3.1 In situ data

Sea-truthing mean values of HPLC chla, SPM, percentage of inorganic matter and

Secchi disk depth for the two samplings are given in Table 3.1, which also summarizes

information about the available MERIS imagery. Water temperatures near surface ranged

from 16.90 to 19.54 oC and from 16.58 to 18.91 oC, respectively, for the two samplings.

Temperature at 10 m depth during the first campaign was between 15.33 and 17.70 oC,

whereas temperatures dropped to 13.50-14.47 oC at the sampling stations on July 22.

Water temperature decreased from the outer part towards the inner part of the ria.

Values of Secchi disk depth between 2 to 12 m were measured in the ria de Vigo, generally

less than half the water column depth.

Chla concentration in the surface water samples did not show a wide variation. Chla

levels were relatively low in comparison with the temporal pattern proposed for the rias

Baixas by Nogueira et al. (1997) where chla concentrations close to 5 mg m-3 are described

as typical during the summer period. Mean chla determined by HPLC varied from 0.03 at

station 12 to 2.65 mg m-3 in the inner part during the first sampling. On July 22 the highest

chla concentration (2.72 mg m-3) was recorded in the innermost ria station. Although the

range in the surface chla concentration was similar in both samplings, differences were

observed in the chla profiles. In the sampling conducted on July 9, small differences were

observed in the chla concentration profiles in the first 10 m of the water column in all

stations except the three at the inner part of the ria where a chla maximum was recorded

at 4 m depth (Fig. 3.2A). On the other hand, a vertical gradient of chla was detected in

almost all sampling stations during the second sampling (Fig. 3.2B), with the highest values

of chla (up to 16 mg L-1) found close to 10 m. The same vertical distribution pattern during

the month of July in ria de Pontevedra is described in Varela et al. (2008) and is imputed to

the presence of upwelled waters. This depth-varying chla distribution may affect the

remote sensing reflectance spectra especially at stations where light penetrates deeper

into the water column. Kutser et al., 2008 showed that different vertical distributions of the

same total cyanobacterial biomass may strongly impact the remote sensing signal.

75


Fig. 3.2. Plots of chorophyll a fluorescence (mg m− 3) vertical profiles for the upper 10 m of the water column in ria de Vigo on A) July 09 2008 and B) July 22 2008.

SPM concentrations varied from 1.17 to 3.15 mg L-1 in the ria de Vigo and showed

decreasing values with distance from st. 1, which is located in the inner, narrow part of the

ria and closer to the main freshwater inputs. In this part of the ria sediment resuspension

and continental runoff are probably higher having as a result high concentrations of SPM.

The results of the chla and SPM analyses of this study combined with available

unpublished data from the same sampling stations using the same methodology showed

that these two variables vary independently (Fig. 3.3, determination coefficient of linear

relationship R2=0.1). This confirms the initial assumption that rias Baixas waters can be

categorized as Case 2 (Morel & Prieur, 1977). This classification is not always a simple

distinction of coastal and oceanic waters as Morel and Maritorena (2001) describe in a later

76


study using a new dataset of optical properties, indicating the need for models more

restricted in geographical and seasonal terms.

Fig. 3.3. Regression analysis between Total Suspended Material (TSM, also termed Suspended

Particulate Matter) and chla concentrations. (y = 1.76 + 0.13x, R2 = 0.1 and sample size N = 41).

3.3.2 Classification results and comparison of MERIS chla algorithms with in situ

data

Classification images, showing the cluster value for each sea pixel, were obtained

for the MERIS images involved in the analysis (Fig. 3.4). In theory, the membership grades

for each cluster would allow us to blend the chla concentration, obtained from the different

neural networks developed for each cluster, into a given pixel, so that chla maps with soft

transitions would be created (Moore et al., 2009). In practice, the NNRB model was only

developed for Cluster#1, so that these classification images presented here were only useful

for detecting the zones where Cluster#1 is the dominant cluster and therefore the areas

where NNRB can be best applied to obtain more reliable results. Figure 3.4 shows that

Cluster#1 is dominant in almost all the images in the rias Baixas and the adjacent area. Table

3.2 shows the percentage of pixels belonging to each cluster for each image over the rias

77


Baixas. Cluster#1 includes the majority of the pixels in ria de Vigo, with more than 72% of

pixels in the six images. On average, 71% and 65% of the pixels over ria de Pontevedra and ria

de Arousa belong to this cluster. Cluster#2 is the predominant one in the ria de Pontevedra

and ria de Arousa in the image delivered on July 29. Cluster#3 is the least abundant in most

of the images, with less than 3.25% of pixels in all of them. However, the presence of

Cluster#2 and Cluster#3 does not prevent the continuous chlorophyll mapping over large

areas in the rias, because of the predominance of Cluster#1. The image acquired on the July

29 is more problematic (referring to the high percentage of pixels belonging to Cluster#2),

although the mapping of a large part of ria de Vigo and small parts of ria de Pontevedra and

ria de Arousa was possible.

Table 3.2. Percentage of pixels belonging to each cluster over the study area (rias Baixas), obtained from classification images derived from the MERIS images used in this study.

Date ria Cluster#1 Cluster#2 Cluster#3

Vigo 74.22 23.95 1.83

July 03 2008 Pontevedra 62.50 37.35 0.15

Arousa 76.33 22.55 1.13

Vigo 95.40 4.04 0.55

July 09 2008 Pontevedra 96.14 3.57 0.30

Arousa 94.24 3.01 2.75

Vigo 84.52 15.48 0.00

July 16 2008 Pontevedra 70.48 28.78 0.73

Arousa 49.56 47.19 3.25

Vigo 82.17 17.66 0.17

July 19 2008 Pontevedra 83.68 16.18 0.15

Arousa 75.68 22.48 1.84

Vigo 87.41 11.88 0.71

July 22 2008 Pontevedra 90.06 9.79 0.15

Arousa 82.92 15.86 1.22

Vigo 72.43 27.57 0.00

July 29 2008 Pontevedra 24.17 75.40 0.43

Arousa 11.67 86.60 1.73

Table 3.3. Performance parameters for the Chla neural networks tested in this study.

Chla model Data base R2

RMS error

(mg m-3

)

Relative RMS

error %

NNRB#1 Whole 0.17 0.74 93

NNRB#2 Cluster#1 0.24 0.69 85

NNRB#3 Cluster#1

High quality 0.70 0.46 65

78


Fig. 3.4. Classification of MERIS images derived from the study area. The 3 classes identified using the FCM are shown.

79


Insofar as the two cloud-free field campaigns were designed specifically to collect

samples within a time period of 2 h from the MERIS overpasses time, 24 valid data were

available to test performance of the MERIS chla algorithms. The performance parameters

for the match-up data of MERIS chla retrieved by the three NNRB algorithms (NNRB#1,

NNRB#2 and NNRB#3) are shown in Table 3.3. The available dataset did not show a wide

range of environmental variation, with chla ranging from 0.03 to 2.72 mg m-3. NNRB for the

Cluster#1, high quality data produced well correlated results (R2=0.70). Bottom effects on

the reflectance or the presence of macroalgae and adjacency effects might be the factors

responsible for the difference in the performance parameters that were observed between

NNRB#1/NNRB#2 and NNRB#3. The good performance of NNRB#3 is not surprising

considering that NNRB#3 can clearly follow the cycle of chlorophyll recorded in the rias

Baixas: concentrations lower than 1 mg m-3 during the winter months, up to 8 mg m-3 during

the spring and autumn maxima and close to 5 mg m-3 during the summer. Moreover, the

NNRB algorithm is trained with MERIS and in situ chla data during upwelling events.

NNRB#3 seems to be robust and ideal for the rias Baixas coastal waters where it can be

used for a more accurate mapping of chla in order to improve the understanding of the

spatial and temporal distributions.

3.3.3 Upwelling cycle

Different meteorological and oceanographic periods were identified and

categorized as three different states in the area during July of 2008 (Table 3.4, Fig. 3.5). The

states lasted from 9 to 11 days which is typical in an upwelling cycle in the area (Nogueira et

al., 1997).

Table 3.4. Dominant atmospheric and oceanographic conditions off the rias Baixas categorized as three different states during the upwelling cycle in summer 2008.

Period Date Dominant atmospheric and oceanographic conditions off rias Baixas

1 July 1-10 winds blowing mainly in south direction (Iw=-108) after a period of

favourable upwelling winds, mostly northward surface flow

2 July 11-21 strong north winds (Iw=900), surface flow towards southwest

3 July 22-31 mainly south blowing winds (Iw=-230), southward surface flow

80


Fig. 3.5. A) Daily upwelling index off the rias Baixas. The Iw in m− 3 s− 1 100 m− 1 represents the offshore Ekman flux in the surface layer. Arrows indicate the days where MERIS FR images were available,

81


numbers define the different states identified during the studied period, and red dots denote the days of sampling. B) Surface currents (cm s− 1) recorded by the Seawatch buoy station located off Cape Silleiro (42° 7.8′N, 9° 23.4′W). Data were daily averaged from 0 a.m. on July 1 2008 to 12 p.m. on July 31 2008. Symbols shown in Fig. 3.2. C) Daily average of sea surface temperature for the month of July 2008 off the rias Baixas. All data are means ± 1 S.D.

State 1 (July 1-11)

This 10 d period state comes after a strong upwelling that occurred in the area at

the end of June (Fig. 3.5A) and it is characterized mainly by weak winds of variable direction

which are typical of upwelling relaxation in the area (deCastro et al., 2004). An exception of

strong downwelling-favourable wind from the south was recorded on July 4.

Surface flow off the rias had a northward direction with a speed ranging between

0.5 and 7.5 cm s-1 (Fig. 3.5B). On July 3, SST ranged between 16 and 17 oC inside the rias, but

an area with temperature higher than 17 oC was observed outside the rias. In the next SST

image (July 9) an increase in temperature was recorded in the rias Baixas and the adjacent

area (Fig. 3.6). The temperature increase was confirmed by the Seawatch data (Fig. 3.5C).

The daily mean water temperature off the rias Baixas increased from 16.4 oC, in the first

days of July up to 18 oC in a period of 10 days after the upwelling.

Two MERIS (FR) images (Fig. 3.7) from the study area were available during State 1,

one on July 3 and the other on July 9. In both, several high chla “patches” were mapped

inside and in the outer parts of the rias. In the area off the external coast of the rias the chla

concentration in the images remained at levels close to 0 mg m-3. This pattern of the

phytoplankton biomass principally confined in the rias while in neighbouring shelf area the

chla levels remained very low, seems to be generated by the northward flow of surface

waters outside the ria. The development of northward currents in the relaxation following

intense north winds, responsible for the upwelling recorded at the end of June, may

introduce water of high chla to the three rias from the ocean area outside them in the first

days of July. The continuing mainly north-westward directed transport over several days

may have been responsible for the chla distribution observed on July 9, where chla

concentration was significantly higher in the ria de Arousa than in the other two rias in the

south.

http://www.sciencedirect.com/science/article/pii/S0034425711001805#ref_f0010

82


Fig. 3.6. MODIS-derived Sea Surface Temperature maps for rias Baixas and adjasted coastal waters during the upwelling cycle of July 2008. White patches represent clouds and land.

83


Fig. 3.7. Chla maps for MERIS FR data derived during the upwelling cycle of July 2008 in the study area. White colour indicates masking of land and clouds.

84


Different patterns of the higher chla areas in the rias were mapped during State 1.

On July 3, areas of high chla concentrations were mapped in the ria de Arousa (3-4.5 mg m-3)

and close to the mouths of the three rias (higher than 2 mg m-3). On July 9, chla

concentrations greater than 2 mg m-3 were mapped mainly in the middle and inner parts of

the ria de Vigo and in the outer part of the ria de Arousa.

In the image obtained on July 3, areas of high chla concentrations were observed in

the outer part of the three rias, whereas chla decreased towards the inner part of the rias.

Varela et al. (2008) reported that this gradient is common in the ria de Pontevedra during

the upwelling period when the meteorological forcing is the main factor responsible for the

circulation of the ria.

Six days after the first available image, the gradient of chla in the rias described

above was observed only in the ria de Arousa. On the contrary, Vigo and Pontevedra were

characterized by a chla gradient where concentration increased toward inshore. In the ria

de Pontevedra areas of higher chla were recorded at the innermost part and close to the

northern mouth of the ria. In the rest of the ria de Pontevedra chla concentration was close

to 0 mg m-3. In the inner part of ria de Vigo, MERIS chla varied between 2 and 3 mg m-3.

MERIS data delivered from areas like the most interior, shallow part of the rias normally

considered as suspicious because the high abundance of macroalgae increases the chla

signal (Gons, 1999) were here characterized as reliable, since they were confirmed by in situ

data. Moreover, water transparency in the 20 m station (St. 1) as determined by the Sechhi

disk measurements during the first campaign was 2 meters, decreasing the effect of the

bathymetry. This part of the rias can be firmly considered as estuary and during nutrient

enrichment from river flows, high concentrations of chla have been recorded (Evans &

Prego, 2003). In this case the observed relatively high concentrations of chla at the inner

part of the two southern rias may be the result of the mixture of estuarine water with

Eastern North Atlantic Central Water (ENACW) combined with high residence times. The

different offshore-inshore gradient of ria de Arousa seems to be formed by material

transferred to the north from the rias de Vigo and Arousa due to the northward surface

currents. Differences in topography and local winds should also be considered as possible

factors for the observed differences. Ria de Arousa is considered to be the most productive

of the rias Baixas (Bode & Varela, 1998). In the classification of Vidal-Romaní (1984) ria de

Arousa is categorized as open bay, while ria de Pontevedra and ria de Vigo as fjord-like.

Though fjords have deep quiescent interiors, only intermittently renewed, and a shallow sill

at the entrance, while the rias are shallow and have a 2 layer circulation that reverses

between up and downwelling.

85


State 2 (July 12-21)

State 2 was characterized by 9 d of sustained upwelling favourable winds (Fig. 3.5A)

and southwest currents of up to 6 cm s-1 (Fig. 3.5B). SST maps showed that temperature

ranged between 16 and 17 oC in the coastal area outside the rias (Fig. 3.6). Temperature

recorded by the Seawatch buoy decreased more than 1 oC during the upwelling (Fig. 3.5C).

The two chla maps (Fig. 3.7) for this state trace the primary results of the upwelling

favourable winds. On July 16 map areas with the highest chla concentrations were recorded

in the middle part of ria de Pontevedra, at the mouth of ria de Arousa and through the entire

ria de Vigo. The distributions were similar in form in the ria de Vigo and ria de Pontevedra but

higher chla (>2.5 mg m-3) was found in the former. Unpublished data showed the outflow of

ria water towards offshore in speeds that reached 4 cm s-1. This situation of the surface

water being advected offshore in the rias, when upwelling favourable wind started to blow

off the rias Baixas is typical of the positive estuarine circulation that has been described in

the area (Fraga & Margalef, 1979; Figueiras & Pazos, 1991). In this two-layer circulation, the

offshore surface Ekman-transport advects the low salinity water out of the ria, while the

denser upwelled water flows into the ria along the sea bed. The zone of enhanced surface

chla concentration that in the MERIS images extend throughout the ria de Pontevedra and

ria de Vigo is probably surface water that is flowing out of the rias due to the positive

estuarine circulation generated during the upwelling favourable conditions.

The July 19 chla image shows a noticeable increase in chla with concentrations

higher than 1 mg m-3 over the entire continental shelf zone, although chla decreased slightly

within the rias. In the study of Ospina-Álvarez et al. (2010) it was found out that during the

upwelling favourable conditions that characterized the Northern Galician rias during the

period of July 13-22 2008 the ENACW did not enter in the rias. While in that period chla in

the Northern Galican rias did not exceed the value of 1 mg m-3 (Ospina-Álvarez et al., 2010),

in the rias Baixas it was generally higher than 1 mg m-3.

State 3 (July 22-31)

As a result of the strong upwelling event a peak of chla with concentrations up to 5

mg m-3 was mapped on July 22 in the coastal area off Galicia. The high chla concentration

was extended from the northern offshore area to the interior of the rias (Fig. 3.7). A

coincident area of relatively low temperature was mapped in the north part of the study

area, whereas an area of warmer water was detected at the south. Differences up to 2 oC

86


were obtained between the rias (Fig. 3.6). This alongshore difference probably reflects the

persistence of stronger coastal upwelling in the north after the event of 10-21 July and an

earlier onset of relaxation in the south. It is often the case that upwelling is more persistent

in the north of the area (Torres & Barton, 2007). Figure 3.8 shows the development of the

upwelling on the Galician coast. With the abrupt decrease of upwelling-favourable to zero

wind on July 22, currents at the Seawatch buoy became briefly northward as expected, but

subsequently returned to southward despite the onset of intermittent northward winds.

The last MERIS image (July 29) is consistent with strong relaxation: the offshore region has

near-zero chlorophyll and a region of moderately high chla is bound to the coast. Within

the rias values tend to be low, reflecting downwelling conditions. It seems probable that

more flow inshore of the Seawatch buoy was northward and convergent to shore. At the

end of July chla in the ria de Vigo showed the lowest concentration of all the images of

previous days.

The high chla concentrations along the Galician shelf coupled with low SST. MERIS

and MODIS images at the start of this state on July 22 show clearly the presence of a cold,

chlorophyll-rich area resulting from the previous 10 days of upwelling. Although high chla

water was recorded below the surface in the in situ profiles (Fig. 3.2B) during the second

sampling in the ria de Vigo, MERIS data recorded the low surface values present. Figueiras

and Pazos (1991) noted the presence of nutrient-rich water during a summer upwelling

event in the rias Baixas that did not reach the surface. As soon as upwelling ceases, the 2-

layer circulation reverses and surface waters flow inwards and sink to the lower layer

carrying with them the higher surface concentrations of chla. The possible non uniformity

of the Inherent Optical Properties (IOP) in the water profiles (Stramska & Stramski, 2005)

and the development and validation of the water constituent algorithms based on water

samples from certain depths (e.g. O'Reily et al., 2000; González Vilas et al., 2011) affirms the

necessity of the in situ data.

Although this high biomass area was not sampled directly, in situ data from the ria

de Vigo revealed relatively high concentrations of diatoms (mainly Chaetoceros spp.) and

small flagellates (personal observation). The potentially domoic acid producing Pseudo-

nitzschia spp. was also present in the ria de Vigo but in relatively low concentrations. This

phytoplankton composition seems to be typical in the rias Baixas during the summer

according to the annual cycle of phytoplankton abundance proposed in 1987 by Figueiras

and Niell. Moreover, Frangópulos et al., (2011) mentioned the presence of the red-tide

dinoflagellate Noctiluca scintillans in high abundances in ria de Vigo during summer 2008.

87


Fig. 3.8. RGB MERIS (l2) FR composite image acquired on July 22 2008 over the study area. Land was masked in black.

88


3.4 References

Aiken, J., Moore, G. F., Trees, C. C., Hooker, S. B., & Clark, D. K. (1995). The SeaWiFS CZCS-type

pigment algorithm. In S. B. Hooker, & E. R. Firestone (Eds.), NASA technical memorandum,

104566. SeaWiFS Technical Report Series, vol. 29 (pp. 1-32).

Atkinson, P. M., & Tatnall, A. R. L. (1997). Introduction Neural networks in remote sensing.


Bakun, A. (1973). Coastal upwelling indices, west coast of North America 1946-71. In NOAA

Technical Report NMFS-SSRF 671 (pp. 1-103). Seattle: US Dept. Commerce.

Bates, S. S., Garrison, D. L., & Horner, R. A. (1998). Bloom dynamics and physiology of domoic-

acid producing Pseudo-nitzschia species. In D. M. Anderson, A. D. Cembella & G. M. Hallegraeff

(Eds.), Physiological ecology of harmful algal blooms (pp.267-292). Heidelberg, Germany:

Springer-Vergal.

Bode, A., Carrera, P., & Lens, S. (2003). The pelagic foodweb in the upwelling ecosystem of

Galicia (NW Spain) during spring: natural abundance of stable carbon and nitrogen isotopes.

ICES Journal of Marine Sciences, 60, 11-22.

Bode, A., & Varela, M. (1998). Primary production and phytoplankton in three Galician Rias Altas

(NW Spain): seasonal and spatial variability. Scientia Marina, 62(4), 319-330.

Bravo, I., Figueroa, R. I., Garcés, E., Fraga, S., & Massanet, A. (2010a). The intricacies of

dinoflagellate pellicle cysts: The example of Alexandrium minutum cysts from a bloom-recurrent

area (Bay of Baiona, NW Spain). Deep Sea Research Part II: Topical Studies in Oceanography, 57,

166-174.

Bravo, I., Fraga S., Figueroa R. I., Pazos Y., Massanet A., & Ramilo I. (2010b). Bloom dynamics and

life cycle strategies of two toxic dinoflagellates in a coastal upwelling system (NW Iberian

Peninsula). Deep-Sea Research Part II: Topical Studies in Oceanography, 57, 222–234.

Brown, C. A., Huot, Y., Werdell, P. J., Gentili, B., & Claustre, H. (2008). The origin and global

distribution of second order variability in satellite ocean color and its applications to algorithm

development. Remote Sensing of Environment, 112, 4186-4203.

Brown, C., & Hutchings, L. (1987). The development and decline of phytoplankton blooms in the

Benguela upwelling system. 1. Drogue movements, hydrography and bloom development.

South African Journal of Marine Science, 5, 171-193.





89


Carder, K. L., & Steward, R. G. (1985). A remote-sensing reflectance model of a red-tide


Cota, G. F., Wang, J., & Comiso, J. C. (2004). Transformation of global satellite chlorophyll

retrievals with a regionally tuned algorithm. Remote Sensing of Environment, 90, 373-377.

deCastro, M., Gómez-Gesteira, M., Alvarez, I., & Prego, R. (2004). Negative estuarine circulation

in the Ria of Pontevedra (NW Spain). Estuarine, Coastal and Shelf Science, 60, 301–312.


Remote Sensing, 28, 517-535.

Doerffer, R., & Schiller, H. (2008). MERIS regional coastal and lake case 2 water project -

Atmoshperic correction ATBD. In: GKSS Research Center 21502 Geestacht. Germany. Version 1.0,

18 May 2008.

Doerffer, R. Sorensen, K., & Aiken, J. (1999). MERIS potential for coastal zone application,

International Journal of Remote Sensing, 20(9), 1809–1818.

Dransfeld, S., Tatnall, A. R., Robinson, I. S., & Mobley, C. D. (2004). A comparison of Multi-layer

multilinear regression algorithms for the inversion of synthetic ocean colour spectra.

International Journal of Remote Sensing, 25(21), 4829-4834.

Dzwonkowski, B., & Yan, X. -H. (2005). Development and application of a neural network based

ocean colour algorithm in coastal waters. International Journal of Remote Sensing, 26, 1175-1200.

Evans, R. H., & Gordon, H. R. (1994). Coastal zone color scanner system calibration: A

retrospective examination. Journal of Geophysical Research, 99, 7293-7307.

Evans, G., & Prego, R. (2003). Rias, estuaries and incised valleys: is a ria an estuary? Marine

Geology, 196, 171-175.

Fawcett, A., Pitcher, G. C., Bernard, S., Cembella, A. D., & Kudela, R. M. (2007). Contrasting wind

patterns and toxigenic phytoplankton in the southern Benguela upwelling system. Marine

Ecology Progress Series, 348, 19–31.

Figueiras, F. G., Gómez, E., Nogueira, E., & Villarino, M. L. (1996). Selection of Gymnodinium

catenatum under downwelling conditions in the Ría de Vigo. In T. Yasumoto, Y. Oshima & Y.

Fukuyo (Eds.), Harmful and toxic algal bloom (pp. 215–218). Paris, France: IOC/ Unesco.

Figueiras, F. G., Jones, K. J., Mosquera, A. M., Alvarez-Salgado, X. A., Edwards, E., & MacDougall,

N. (1994). Red tide assemblage formation in an estuarine upwelling ecosystem: Ria de Vigo.


Figueiras, F. G., & Niell, F. X. (1987). Composición del fitoplankton de la ría de Pontevedra (NO de

España). Investigación Pesquera, 51, 371-409.

90


Figueiras, F., & Pazos, Y. (1991). Microplankton assemblanges in three Rías Baixas (Vigo, Arosa

and Muros, Spain) with a subsurface chlorophyll maximum: their reletionships to hydrography.

Marine Ecology Progress Series, 76, 219-233.

Figueiras, F. G., & Ríos, A. F. (1993). Phytoplankton succession, red tides and the hydrographic

regime in the Rías Bajas of Galicia. In T. J. Smayda & Y. Shimizu (Eds.), Toxic Blooms in the Sea

(pp. 239-244). New York: Elsevier.

Fraga, F. (1981). Upwelling off the Galician coast North West Spain. In E. Suess & J. Thiede (Eds.),

Coastal upwelling (pp. 176-182). New York: Plenum Publishing Corp.

Fraga, S., Anderson, D. M., Bravo, I., Reguera, B., Steidinger, K. A., & Yentsch, C. M. (1988).

Influence of upwelling relaxation on dinoflagellates and shelfish toxicity in Ria de Vigo, Spain.

Estuarine Coastal Shelf Science, 27, 349-361.

Fraga, F., & Margalef, R. (1979). Las Rias Gallegas. In Estudio y explotacion del mar en Galicia

(pp. 101-122), University of Santiago (Ed.).

Fraga, S., Reguera B., & Bravo, I. (1990). Gymnodinium catenatum bloom formation in the

Spanish rías. In Graneli E., B. Sundstrom, L. Edler & D. M. Anderson (Eds.), Toxic marine

phytoplankton (pp. 149-154). New York: Elsevier.

Frangópulos, M., Spyrakos, E., & Guisande, C. (2011). Ingestion and clearance rates of the red

Noctiluca scintillans fed on the toxic dinoflagellate Alexandrium minutum (Halim). Harmful Algae,

10, 304-309.

GEOHAB (2005). Global Ecology and oceanography of Harmful Algal Blooms (GEOHAB). In G.

Pitcher, T. Moita, V. Trainer, R. Kudela, F. G. Figueiras & T. Probyn (Eds.), GEOHAB Core Research

Project: HABs in Upwelling Systems (p. 82). Paris and Baltimore: IOC and SCOR.

Gilerson, A. A., Gitelson, A. A., Zhou, J., Gurlin, D., Moses, W. J., Ioannou, I., & Ahmed S. A.

(2010). Algorithms for remote estimation of chlorophyll-a in coastal and inland waters using red

and near-infrared bands. Optical Express, 18, 24109–24125.


productive estuaries: Chesapeake Bay case study. Remote Sensing of Environment, 109, 464-472.

González Vilas, L., Spyrakos, E., & Torres Palenzuela, J. M. (2011). Neural network estimation of

chlorophyll a from MERIS full resolution data for the coastal waters of Galician rias (NW Spain).

Remote Sensing of Environment, 115, 524-535.

Gons, H. (1999). Optical Teledetection of Chlorophyll a in turbid inland waters. Environmental

Science & Technology, 33, 1127-1132.

Herrera, J. L., Piedracoba, S., Varela, R. A., & Roson, G. (2005). Statial analysis of the wind field

on the western coast of Galicia (NW Spain) from in situ measurements. Continental Shelf

Research, 25, 1728-1748.

91


Joint, I., Groom, S., Wollast, R., Chou, L., Tilstone, G., Figueiras, F., Loijens, M., & Smyth, T.

(2002). The response of phytoplankton production to periodic upwelling and relaxation events

at the Iberian shelf break: estimates by 14C method and by satellite remote sensing. Journal of

Marine Systems, 32, 219-238.

Keiner, L. E., & Yan, X. H. (1998). A neural network model for estimating sea surface chlorophyll

and sediments from thematic mapper imagery. Remote Sensing of the Environment, 66, 153-165.

Kudela, R., Pitcher, G., Probyn, T., Figueiras, F., Moita, T., & Trainer V. (2005). Harmful Algal

Blooms in coastal upwelling systems. Oceanography, 18, 184-197.

Kutser, T. (2004). Quantitative detection of chlorophyll in cyanobacterial blooms by satellite

remote sensing. Limnology and Oceanography, 49, 2179–2189.

Kutser T., Metsamaa L., & Dekker A. G. (2008). Influence of the vertical distribution of

cyanobacteria in the water column on the remote sensing signal. Estuarine Coastal and Shelf

Science, 78, 649-654.

Laiño, C. C. (1991). Gymnodinium catenatum toxins from mussels (mytilus galloprovincialis).

Environmental Technology, 12, 33-40.

Margalef, R. (1956). Estructura y dinámica de la ´´purga de mar´´ en Ría de Vigo. Investigacion

Pesquera, 5, 113-134.

McClain, C. R., Feldman, G. C., & Hooker, S. B. (2004). An overview of the SeaWiFS project and

strategies for producing a climate research quality global ocean bio-optical time series. Deep-Sea

Research II, 51, 6-42.

McClain, C. R., Chao, S. -Y., Atkinson, L. P., Blanton, J. O., & Castillejo, F. (1986). Wind-driven

upwelling in the vicinity of Cape Finisterre, Spain. Journal of Geophysical Research, 91, 8470-8486.

Moore, T. S., Campbell, J. W., & Dowell, M. D. (2009). A class-based approach to characterizing

and mapping the uncertainty of the MODIS ocean chlorophyll product. Remote Sensing of

Environment, 113, 2424−2430.

Morel, A., & Maritorena, S. (2001). Bio-optical properties of oceanic waters: A reappraisal.

Journal of Geophysical Research, 106, 7163-7180.


22, 709-722.

Muller-Karger, F. E., McClain, C. R., Sambrotto, R. N., & Ray, G. C. (1990). A comparison of ship

and coastal zone color scanner mapped distribution of phytoplankton in the southeastern

Bering Sea. Journal of Geophysical Research, 95, 483-499.



92


185-300.

Oliveira, P., Moita, T., Silva, A., Monteiro, I., & Palma, A. (2009a). Summer diatoms and

dinoflagellate blooms in Lisbon Bay from 2002 to 2005: Pre-conditions inferred from wind and

satellite data. Progress in Oceanography, 83, 270-277.

Oliveira, P., Nolasco, R., Dubert, J., Moita, T., & Peliz, A. (2009b). Surface temperature,

chlorophyll and advection patterns during a summer upwelling event off central Portugal.


O'Reily, J. E., Maritorena, S., Siegel, D. A., O'Brien, M. C., Toole, D., & Mitcell, B. G. et al. (2000).

Ocean color chlorophyll a algorithms foe SeaWiFS, OC2 and OC4. Version 4. In: S. B. Hooker &

E.R. Firestone (Eds.), NASA technical memorandum 2000-206892, Vol. 11, 49 pp.

Ospina-Álvarez, N., Prego, R., Álvarez, I., deCastro, M., Álvarez-Ossorio, M. T., Pazos, Y., Campos,

M. J., Bernárdez, P., Garcia-Soto, C., Gómez-Gesteira, M., & Varela, M. (2010). Oceanographic

patterns during a summer upwelling-downwelling event in the Northern Galician Rias:

Comparison with the whole Ria system (NW of Iberian Peninsula). Continental Shelf Research, 30,

1362-1372.

Pazos, Y., Moroño, A., Triñanes, J., Doval, M., Montero, P., Vilarinho, M. G., & Moita, M. T.

(2006). Early detection and intensive monitoring during an unusual toxic bloom of Gymnodinium

catenatum advected into the Galician Rías (NW Spain). Proceedings of 2006 International

Conference on Harmful Algae, September 4–8. Copenhagen, Denmark.

Peliz, A., & Fiuza, A. (1999). Temporal and spatial variability of CZCS-delivered phytoplankton

concentrations off the western Iberian Peninsula. International Journal of Remote Sensing, 20,

1363-1407.

Ribeiro, A., Peliz, A., & Santos, A. M. P. (2005). A study of the response of chla-biomass to a

winter upwelling event off western Iberia using SeaWiFS and in situ data. Journal of Marine

Systems, 53, 87-107.

Pitcher, G. C., Boyd, A. J., Horstman, D. A., & Mitchell-Innes, B. A. (1998). Subsurface

dinoflagellate populations, frontal blooms and the formation of red tide in the southern

Benguela upwelling system. Marine Ecology Progress Series, 172, 243-264.

Pitcher, G. C., Figueiras, F. G., Hickey, B. M., & Moita, M. T. (2010). The physical oceanography of

upwelling systems and the development of harmful algal blooms. Progress in Oceanography, 85,

5–32.

Pitcher G. C., & Nelson G. (2006) Characteristics of the surface boundary layer important to the

development of red tide on the southern Namaqua shelf of the Benguela upwelling system.

Limnology and Oceanography, 51, 2660–2674.

Rodríguez Rodríguez, G., Villasante, S., & García-Negro, M. (2011). Are red tides affecting

economically the commercialization of the Galician (NW Spain) mussel farming? Marine Policy,

93


35, 252-257.

Ruddick, K., Lacroix, G., Lancelot, C., Nechad, B., Park., Y., Peters, S., & Van Mol., B. (2008).

Optical remote sensing of the North Sea. In V. Barale & M. Gade (Eds.), Remote Sensing of the

European Seas (pp. 79-90). New York: Springer-Verlag.

Shahraiyni, H. T., Shouraki, S. B., Fell, F., Schaale, M., Fischer, J., Tavakoli, A., Preusker, R.,

Tajrishy, M., Vatandoust, M., & Khodaparast, H. (2009). Application of an active learning method

to the retrieval of pigment from spectral remote sensing reflectance data. International Journal

of Remote Sensing, 30, 1045-1065.

Spyrakos, E., González Vilas, L., & Torres Palenzuela, J. (2010). Development and application of

neural network-based chlorophyll-a algorithms from MERIS full-resolution data for optically

complex waters. Proceedings of Remote sensing and photogrammetry society annual conference

with Irish earth observation symposium, Septermber 1-3. Cork, Ireland.

Spyrakos, E., Santos-Diniz, T. C., Martinez-Iglesias G., Torres-Palenzuela, J. M, & Pierce, G. J. (in

press). Spatiotemporal patterns of marine mammal distribution in coastal waters of Galicia, NW

Spain. Hydrobiologia.

Stramska, M., & Stramski, D. (2005). Effects of a nonuniform vertical profile of chlorophyll

concentration on remote-sensing reflectance of the ocean. Applied Optics, 44, 1735-1747.

Thiria, S., Mejia, C., Badran, F., & Crepon, M. (1993). A neural network approach for modelling

nonlinear transfer functions: Application for wind retrieval from spaceborne scatterometer

data. Journal of Geophysical Research, 98(12), 22827−22841.

Tilstone, G. H., Figueiras, F. G., & Fraga, F. (1994). Upwelling-downwelling sequences in the

generation of red tides in a coastal upwelling system. Marine Ecology Progress Series, 112, 241-253.

Tilstone, G. H., Figueiras, F. G., Lorenzo, L. M., & Arbones, B. (2003). Phytoplankton composition,

phytosynthesis and primery production during different hydrographic conditions at the

Northwest Iberian upwelling system. Marine Ecology Progress Series, 252, 89-104.

Torres, R., & Barton, E. D. (2007). Onset of the Iberian upwelling along the Galician coast.


Torres-Palenzuela, J. M., Vilas-González, L., & Mosqura-Giménez, Á. (2005a). Correlation

between MERIS and in-situ data for study of Pseudo-nitzschia spp. toxic blooms in Galician

coastal area. (pp. 497-507). Rotterdam: Millpress.

Torres-Palenzuela, J. M., Vilas-González, L., & Mosqura-Giménez, Á. (2005b). Detection of

Pseudo-nitzschia spp. toxic blooms using MERIS images on the Galician coast (pp. 1915-1919).:

European Space Agency (Special Publication).

Trainer, V. L., Pitcher, G. C., Reguera, B., & Smayda, T. J. (2010). The distribution and impacts of

harmful algal bloom species in eastern boundary upwelling systems. Progress in Oceanography,

94


85, 33-52.

Varela, M. (1992). Upwelling and phytoplankton ecology in Galician (NW Spain) rías and shelf

waters. Boletín del Instituto Español de Oceanografía, 8, 57-74.

Varela, M., Prego, R., & Pazos, Y. (2008). Spatial and temporal variability of phytoplankton

biomass, primary production and community structure in the Pontevedra Ria (NW Iberian

Peninsula): oceanographic periods and possible response to environmental changes. Marine

Biology, 154, 483-499.

Vidal-Romaní, J. R. (1984). A orixe das rias galegas: Estado da censtión (1886-1983). Cuardenos

da Área de Ciencias Mariñas, Seminario de Estudos Galegos. Vol. 1. Ediciós do Castro. A Coruña,

Spain.


(NW Spain): main facies and hydrodynamic dependence. Journal of Marine Systems, 54, 261–276.



pyridine-containing mobile phases. Marine Ecology Progress Series, 195, 29-45.

Zhang, T., Fell, F., Liu, Z. -S., Preusker, R., Fischer, J., & He, M. -X. (2003). Evaluating the

performance of artificial neural network techniques for pigment retrieval from ocean color in

Case I waters. Journal of Geophysical Research, 108, 3286-3298.

Chapter IV. Toxic Pseudo-nitzschia spp. events

95

Remote sensing, in situ monitoring and environmental perspectives of toxic Pseudo-nitzschia events in the surface waters of two Galician rias

Abstract

4.1 Introduction


4.3 Results

4.4 Discussion

4.5 References

CHAPTER IV

97


Abstract

In this chapter, a combined analysis of satellite imagery data, measurements of

biotic and abiotic parameters and mixed effects modelling was used to study Pseudo-

nitzschia blooms in the optically complex surface waters of two Galician (NW Spain) rias

(embayments), sites of extensive mussel culture. This included the measurement of domoic

acid (DA) concentrations in natural microalgae populations for the first time, as far as we

know, in the area and the determination of Pseudo-nitzschia abundances along with several

environmental parameters during the years 2007-2009. A regionally/cluster-specific

chlorophyll a (chla) algorithm for Medium Resolution Imaging Spectrometer (MERIS) was

applied to a series of images where in-situ data were available. Three periods were

considered which were linked with the dominant meteorological conditions off the rias. The

first toxic event was recorded on autumn 2007 and was associated with moderate Pseudo-

nitzschia (abundances) and high particulate DA (pDA) (2.5 μg L-1) concentrations. Similar

pDA levels to ours have been measured in the surface waters from other coastal upwelling

areas around the world. A second, but lower in toxin concentration DA outbreak, was

recorded during summer 2009. P. australis seems to be the main source of DA in the study

area. The application of a regional algorithm in combination with the characteristics of

MERIS FR allowed for accurate mapping of chla and the detection of high in chla and

Pseudo-nitzschia spp. “patches” in the rias. The optimal model for the Pseudo-nitzschia spp.

abundance and DA concentration suggested the significant effect of some macronutrients

as well as other abiotic and biotic parameters, approximating in that way the potential

environmental causes and effects of the harmful Pseudo-nitzschia spp. blooms in the area.

The results of this study deduced that toxic events due to DA should be an important

concern and therefore DA in the seawater should be measured routinely in order to assess

the potential of a DA outbreak. Moreover, MERIS FR data and regionally specific algorithms

showed that they can provide valuable information about the blooms and should be an

integral part of the monitoring programs.

98


4.1 Introduction

Pseudo-nitzschia Peragallo is a chain-forming diatom genus widespread in all oceans

of the world. A number of these diatom species are known as potential producers of a

neurotoxic amino acid, domoic acid (DA), which when accumulated via trophic transfer in

the food-wed can have deleterious (domoic amnesic syndrome, DAP) and even fatal effects

to several marine organisms and less frequently to humans. Pseudo-nitzschia is the only

diatom among 27 dinoflagellate and 1 raphidophyte species that is listed as responsible for

harmful algal events in upwelling regions. Besides this, it has been associated with

upwelling events in the California Current System (Bates et al., 1998; Adams et al., 2000;

Trainer et al., 2002; Marchetti et al., 2004; Kudela et al., 2004, 2005; Anderson et al., 2006)

and in the Iberian System (Figueiras & Rios, 1993; Moroño et al., 2000; Mouriño et al. 2010).

Understanding the relationships between the abundance of this diatom, DA

outbreaks and the characteristics of the environment will provide valuable information on

the variable(s) that might determine their growth, distribution and DA production, which

can then be used to predict harmful incidents. Several authors (Smith et al., 1990; Parsons

et al., 2002; Kudela 2004; GEOHAB 2005; Klein et al., 2010) have related these blooms with

changes in their nutrient physiology and therefore with upwelling events and

eutrophication. On the west coast of USA, an area where Pseudo-nitzschia is a common

harmful bloom former, AD outbreaks caused by P. australis have been associated with

upwelling events and elevated but in declining stage nutrients (Horner et al., 1997).

Several laboratory experiments studies (Bates et al., 1991; Pan et al., 1996;

Maldonado et al., 2002; Fehling et al., 2004; Wells et al., 2005) have addressed the silicic

acid and/or phosphorus, iron, copper limitations as significant factors for the toxin

production of Pseudo-nitzschia species. Although the relationship between DA production

and the nutritional status appears to be complex in natural populations since they are often

composed of several species (Trainer et al., 2000, 2009), it seems that the silicic acid supply

plays an important role (Kudela et al., 2004; Marchetti et al., 2004; Anderson et al., 2006).

More specifically, for P. australis, Anderson et al. (2006) suggested that ratio values of

Si(OH)4:NO3- below 2 are found to be linked to DA production from this species. Recently

Trainer et al. (2009) showed that iron limitation and not macronutrient stress is more

essential for the DA production in natural populations.

Harmful algal events due to the diatoms of the genus Pseudo-nitzschia are

considered as a reoccurring phenomenon along the northern boundary of the Iberian–

99


Canary current upwelling system. These blooms have become a focal point of numerous

studies since the first time they were recorded in the autumn of 1994 (Miguez et al., 1996)

and are occasionally related with short-time DA outbreaks which mainly occure at a late

stage of the bloom (Trainer et al., 2010). Although several species of Pseudo-nitzschia have

been recorded in the Galician coastal embayments referred to as rias (NW Spain), only P.

australis is known to produce DA (Fraga et al., 1998). Toxigenic events due to blooms of P.

australis in this area have been recorded to follow PSP (Paralytic Shellfish Poisoning)

outbreaks (Reguera et al., 2003) in thermohaline stratified water masses at the early

upwelling period (March-June).

There is considerable interest in the study of Pseudo-nitzschia and the detection of

DA in the Galician rias due to the economic and social importance of the extensive culture

of mussels in this region and human health concerns (GEOHAB, 2005; Rodríguez Rodríguez

et al., 2011). As part of a routine monitoring programme, the Technological Institute for the

control of the marine environment of Galicia (INTECMAR) has organized sampling on a

weekly basis in the Galician rias where, among other water parameters, the Pseudo-nitzschia

spp. abundance and the biotoxins in mussels and other molluscs are recorded. DA

detection in this area has been studied mainly on shellfish extracts (Arévalo et al., 1998;

Garet et al., 2010) and on Pseudo-nitzschia cultures (Fraga et al., 1998; Maneiro et al., 2005).

However, no information is available, to our knowledge, on the particulate DA

concentrations (pDA) in seawater.

In these coastal embayements transient increases of Pseudo-nitzschia spp. often

exceed abundances of 106 cells L-1 dominating the phytoplankton community. High

concentrations of these potentially toxic species have been observed within high biomass

phytoplankton “patches” (Marić et al., 2011) which can raise problems to field monitoring

programs. An important component for detecting these high biomass blooms is the passive

ocean colour sensors. Medium Resolution Imaging Spectrometer (MERIS) that succeeded

the first satellite ocean colour sensor CZCS is using more and narrower spectral bands and

finer spatial resolution. MERIS provides data with a 300 m on-ground resolution in nadir

(Full Resolution) and has a spectral resolution of fifteen bands from visible to near infrared,

supporting one of the mission objectives for a delicate coastal zone monitoring (Doerffer et

al., 1999). MERIS data has been recently used on the Galician coast for the development of

regionally specific chlorophyll a algorithms (González Vilas et al., 2011) and to map the

spatial extent of high biomass algae events during an upwelling cycle (Spyrakos et al., 2011).

Our approach in this study is an integration of multidisciplinary observations in the

surface waters of two Galician rias in order to: 1) measure for the first time the pDA

100


concentrations in the seawater, 2) determine the environmental factors that might be

responsible for the Peudo-nitzschia events and the production of DA and 3) obtain a

synoptic view of the spatial distribution on horizontal scales of the blooms by applying

regionally specific algorithms to MERIS data and assess the utility of these tools for the

monitoring of Pseudo-nitzschia blooms.

4.2 Materials and methods

4.2.1 Study area

The Galician rias are V-like coastal formations along the northwest part of the

Iberian Peninsula (Fig. 4.1). The rias Baixas constitute the southern part of the Galician rias.

They are formed by four large coastal embayments, from north to south: Muros y Noia,

Arousa, Pontevedra and Vigo, all oriented in a SW-NE direction. This study focuses on two

rias (Pontevedra and Vigo), each connected to the open sea through two entrances, to the

north and south of the islands located at the external part of each ria.

101


Fig. 4.1. A) Map of the Galician coast. From north to south the rias Baixas: Muros y Noya, Arousa, Pontevedra and Vigo. The location of the Seawatch buoy station off Cabo Silleiro is shown by a black rectangle. B) Map of ria de Vigo and ria de Pontevedra showing the locations of the sampling stations.

4.2.2 Sampling regime and in situ measurements

Samples were collected in two Galician rias over a three year period (2007-2009)

(Table 4.1). Ten fixed stations were visited five times in the ria de Vigo. Two additional

sampling stations, shown in Fig. 4.1B as closed cycles, were sampled on July 14, 2009. The

depth of the stations ranged from 5 m inside the ria to 100 m outside. Chla fluorescence

profile was monitored by a Turner designs CYCLOPS-7 submersible fluorometer. Profiles of

water temperature, pH and dissolved oxygen (DO) were provided by a portable meter (HI

9829, Hanna instruments). A five station transect was sampled in ria de Pontevedra during

the summer of 2009. A Seabird Model 25 CTD was used to collect vertical profiles of

temperature, salinity, fluorescence and depth of water column at each site. The sampling

transect extended from the open sea towards to the inner part of the ria de Vigo, while in

the ria de Pontevedra the transects extended from the middle part of the ria to the

innermost part. Triplicate water samples from surface to 3 meters were collected at each

102


station from a sampler (3524 cm3) for further analysis. The depth of the euphotic zone was

established with a Secchi disk.

High performance liquid chromatography (HPLC): For the HPLC chla determination,

water samples (100-200mL) were filtered through 9 mm diameter Whatman GF/F filters and

stored at -80oC for two weeks. 95% methanol was used as extraction solvent for the

pigments. In this study only chla concentration data are presented, calculated as the sum of

chlorophyllide a, chlorophyll a epimer, chlorophyll a allomer and divinyl chlorophyll a. An

HPLC method using a reversed phase C8 was applied for the separation of the pigments.

Details of pigment extraction and separation are provided in Zapata et al. (2000).

Suspended particulate material (SPM): SPM was evaluated in terms of SPM

concentration and percent weight of organic material (%OM). Pre-combusted (450 oC for 24

h), pre-washed in 500 mL of MilliQ, 47mm Whatman GF/F filters were used. These filters

were then dried at 65 oC to a constant weight. Particles were collected by filtering a

standard volume (1000mL) of seawater samples and then rinsed with 50mL MilliQ in order

to remove salts and dissolved organic material. For the determination of SPM the filters

were dried at 65oC till no weight changes were observed. The filters were then re-

combusted at 450 ºC for 5 h in order to obtain the inorganic suspended material (ISM). The

percent weight of organic material (%OM) was determined by subtracting the ISM from the

SPM. All the filters were weighted on a Precisa 262 SMA-FR microbalance (10-5 g precision).

Nutrients and dissolved organic carbon (DOC): Water samples were filtered through

47mm Whatman GF/F filters and the filtered water stored in -40oC for further analysis of

DOM and dissolved inorganic micronutrients (nitrate, nitrite, phosphorus, silicate and

ammonia). All nutrients and DOC were analysed with a Technicon AAII auto-analyser.

Particulate organic carbon and nitrogen (POC, PON): POC and PON content were

determined from seawater samples (100-200mL) filtered through pre-combusted 9mm

diameter Whatman GF/F filters and then stored at -80oC. The filters were then dried at 70°C

and combusted in a Fisons EA-1108 CHN analyzer. Sulfanilamide was used as the standard.

Pseudo-nitzschia spp. light microscopy (LM) counting: 250mL of seawater from the

first series of the samples collected in the stations were fixed with buffered formaldehyde

at a final concentration of 1% and stored under dark and cool conditions. An appropriate

aliquot (10, 50mL) depending on the phytoplankton cells, inorganic suspended material and

detritus density, was settled in a counting chamber. Total abundances of Pseudo-nitzschia

spp. were enumerated using an inverted light microscope at 250x and 400x magnification

(Utermöhl, 1958). Pseudo-nitzschia species were separated into P. delicatissima and P.

seriata complexes depending on their valve width (Tomas, 1997).

103


Pseudo-nitzschia species identification: Phytoplankton samples for Pseudo-nitzschia

species identification were collected in the upper 10 meters of each station during the

campaigns by vertical net tows with 20 μm mesh-size plankton net. Samples were placed in

dark and fixed with paraformaldeyde (to a final concentration of 1%). Scanning electron

microscopy (SEM) was performed on select samples (depending on the presence or not of

Pseudo-nitzschia in the counting chambers) to identify individual Pseudo-nitzschia cells.

Removal of the organic matter from the samples involved treatment with hydrogen

hyperoxide and incubation at 90oC.

Particulate DA (pDA) analysis: Particulate DA analysis was performed to triplicate

sub-samples derived from the first series of the seawater samples collected in the stations

shown in Fig. 1. The DA remained on Whatman GF/F 9 mm filters after the filtration of 200

mL seawater was quantified by ASP cELISA (Biosense Laboratories, Bergen Norway). A

protocol described in Fehling et al., (2004) was followed for the preparation of the

seawater samples and the assay procedure. A Multiskan EX (LabSystems) microplate

spectrophotometer with a 450 nm filter was used to read the absorbance of the samples.

The detection limit for the ELISA assay was 1pg DA mL-1 and for for the analysis of pDA in

the seawater was 0.005 ng DA L-1. Cellular DA (cDA) concentration was calculated by

normalizing the pDA concentration to the abundance of Pseudo-nitzschia spp. cells found in

the preserved samples.

Meteorological buoys: Meteorological data off the rias Baixas were provided by the

Spanish Port System (www.puertos.es). Wind speed (W) and direction were observed at a

Seawatch buoy station located off Cape Silleiro (42° 7.8′N, 9° 23.4′W). This meteorological

station was selected as fairly representative of the study area (Herrera et al., 2005). Daily

upwelling index (IW) was estimated from wind by Bakun's (1973) method:

IW = - y/( W·f)= - 1000· a·CD·W·Wy/( W·f) m3/(s·km)

where a is the density of air (1.2 kg·m-3 at 15ºC), CD is an empirical dimensionless drag

coefficient (1.4·10-3 according to Hidy, 1972), f is the Coriolis parameter (9.9·10-5 s-1 at 42º

latitude), W is the density of seawater (1025 kg·m-3), and W and Wy are the average daily

module and northerly component of the wind.

http://www.puertos.es/

104


4.2.3 MERIS FR imagery

The MERIS satellite imagery used in this study contains 7 full-resolution level-1b

images derived from the area during the years 2007-2009. MERIS overpasses were within 2

hours of the time that samples and data were collected in situ (Table 4.1).

Table 4.1. Distribution of MERIS FR imagery used in this study for the years 2007-2009 showing the acquisition time (UTC), sky conditions and mean view zenith angle from west. The places in brackets represent the rias where the campaigns were carried out on those dates.

Date Sky conditions Acquisition time (UTC) View zenith angle (º)

25/07/2007 (Vigo) clear 11:11 15.3

19/10/2007 (Vigo) clear 11:08 12.6

09/07/2008 (Vigo) clear 11:10 13.0

22/07/2008 (Vigo) clear 11:02 11.7

27/05/2009 (Pontevedra) clear 10:51 20.7

07/07/2009 (Pontevedra) cloudy 10:58 15.7

14/07/2009 (Vigo) partly cloudy 10:42 20.7

Beam 4.6 (Brockmann Consult and contributors, Germany) software was used for

the analysis of the imagery. The BEAM-4.6´s smile correction was applied to the original

level-1b data. For the atmospheric correction, the ocean colour data were processed with a

NN-based algorithm which was developed by Doerffer and Schiller (2008). The flags for

coastline, land, clouds and invalid reflectance were raised using the Beam software. Ocean

colour data derived from areas significantly affected by sun glint (beyond a solar zenith

angle limit of 60o) were considered invalid and were masked. The FCM algorithm described

on Chapter II was applied to the water-leaving radiance reflectance data in order to identify

the different clusters.

Classification images were then obtained for the available MERIS images using the

same FCM algorithm. The pixels in these images were assigned to the cluster with the

highest value in its corresponding membership function and the percentage of pixels

belonging to each cluster was computed. Water-leaving radiance reflectances and chla

concentrations were computed as mean values of the pixel corresponding to the sampling

station location and the 8 surrounding pixels. These 9 pixels cover approximately 0.8 km2 of

surface area and it was considered that this average was able to reduce MERIS instrument

noise. The imagery was then remapped using the standard Mercator projection with a fixed

105


grid of 890 by 890 pixels. Each image ranges from 42o 04' N to 42 o 40' N latitude and from

8o 32'W to 9o 32'W longitude, which covers approximately 3.1 x 103 km2.

4.2.4 Statistical modelling approach

The poor relationships between Pseudo-nitzschia abundance and pDA concentration

in our data set showed that a model based only on the abundance is not always sufficient

for the prediction of toxigenic events in the area. Therefore, two separate models were

developed in this study, which associate Pseudo-nitzschia abundance and pDA

concentration with selected environmental parameters. Data exploration showed that

there are considerable differences of Pseudo-nitzschia spp. density and pDA values between

the stations and the different months, indicating that these two factors have to be included

in the models. The abundance data was analysed using generalized additive mixed models

(GAMMs) as implemented by the mgcv (Wood, 2000) library in R (version 2.9.1, R

development CoreTeam). GAMMs use additive non-parametric functions by smoothing

splines to model covariate effects, while allow correlations between the variables explicitly

by adding random effects to the additive predictor and are applicable to nested data

structures (Lin & Zhang, 1999). Poisson and negative binomial distributions were used for

the model fitting of the Pseudo-nitzschia count data. However, the negative binomial model

did not converge. A linear function was considered and generalised linear mixed model

(GLMM: Zuur et al., 2009) was applied to the DA concentration data. Gaussian distribution

was used for DA data. A square root transformation was applied to the Pseudo-nitzschia

abundances and DA concentrations data. From the list of available variables (Table 4.3) Zd,

N, NO22-, TSIM, Temp, Si(OH)4/PO4

3- were dropped due to collinearity observed when the

variation inflation factors through the corvif-function in the AED library were calculated

(Zuur, 2009). DOC and pH were also removed from the statistical analysis due to inefficient

number of measurements. All the other variables were considered as predictors. The

optimization of the model was based on finding first the optimal random structure

followed by selection of the optimal fixed structure. Both models included each station as a

random effect. The final models were selected stepwise on the basis of the AIC (Akaike

Information Criterion) and individual significance of explanatory variables. The F test was

used to compare the nested models (see Zuur et al., 2007). We used R (R Development

Core Team, 2007) for the analyses.

106


4.3 Results

4.3.1 Environmental and meteorological conditions

Three periods were selected reflecting the different months of the campaigns and

the meteorological conditions off the study area (Table 4.2).

Table 4.2. The three periods considered during the study in relation to the month and year of the campaigns and the dominant meteorological conditions off the rias.

Period Month Year Meteorological conditions off the rias

1 May 2009 downwelling favourable winds

2 July 2007-2009 upwelling cycle

3 October 2007 strong downwelling favourable winds

Fig. 4.2 shows that strong downwelling-favourable wind was recorded on period 1.

The samplings carried out on period 2 were characterized mainly by upwelling conditions or

relaxation after strong upwelling favourable winds. Downwelling favourable conditions

were dominant off the rias Baixas during the sampling conducted on period 3.

Fig. 4.2. Daily upwelling index off the rias Baixas for the years 2007-2009. The Iw in m−3 s−1 100 m−1 represents the offshore Ekman flux in the surface layer. Arrows indicate the days where the in-situ surveys were carried out.

Table 4.3 provides a list of the environmental parameters measured in this study

and their mean values during the 3 periods. The surface concentration of nitrates in the

study area ranged from 0 to 3.775 mg L-1 and nitrites ranged from 0.005-0.4153 mg L-1. The

highest values for these two macronutrients were observed in the ria de Pontevedra on

period 1, and specifically at the three inner stations close to the main freshwater inputs in

the ria. The range of the orthophosphate at the sampling stations was 0-0.850 mg L-1. High

107


orthophosphate concentrations were detected at the ria de Pontevedra on period 1. Silicic

acid concentrations were generally very low and below the detection limit. The highest

surface silicic acid concentrations were recorded at the innermost station during the two

samplings in the ria de Pontevedra. Concentration levels of ammonium on the surface

waters in the two studied rias were low during the samplings. Elevated ammonium

concentrations were observed in the ria de Vigo during July 2008. The POC concentration in

the surface waters ranged between 0.13 (Station 1, period 3) and 0.75 mg L-1 (Station 3,

period 2). Higher levels of POC were observed in the ria de Vigo on July 22 2008. During

period 3 we found high concentrations of DOC in agreement with relatively high surface

chla concentrations.

SPM concentrations varied from 1.17 to 3.74 mg L-1 in the ria de Vigo and ria de

Pontevedra and showed in general decreasing values with distance from Station 1, which is

located in the inner, narrow part of the rias and closer to the main freshwater inputs. In this

part of the rias sediment resuspension and continental runoff is probably higher having as a

result high concentrations of SPM. ISM followed the same spatial and temporal distribution

with SPM. Values of Secchi disk depth, between 2 to 12 m, were measured in the ria de Vigo

and from 6 to 11 m in the ria de Pontevedra, generally less than half the water column depth.

In-situ surface chla values ranged from 0.03 to 6.25 μg L-1 in ria de Vigo and from

0.48-3.63 in the ria de Pontevedra. The maximum concentration in the study area was

revealed on period 3 at station 3V.

Table 4.3. A list of the environmental parameters measured in this study combined with their in-situ mean values during the 3 periods. nd= not determined, bd=below detection limit.

Parameter Abbreviation Units 1 2 3

Chlorophyll a Chla μg L-1

0.95 1.24 3.73

Suspended particulate material SPM mg L-1

1.87 2.09 2.47

inorganic suspended material ISM % 37.8 42.1 48.5

Secchi disk depth Zsd m 8.8 6.7 7.3

pH pH n/a nd 8.2 8.1

Salinity Sal psu 35.5 33.2 29.7

Temperature Temp ºC 13.1 18.6 15.9

Silicic acid Si(OH)4 mg-Si L-1

1.380 0.340 0.630

Orthophosphate PO43-

mg-P L-1

0.606 0.052 0.018

Nitrite NO22-

mg-N L-1

0.354 0.027 0.016

Nitrate NO3- mg-N L

-1 2.689 0.057 0.096

Ammonium NH4 mg-N L-1

bd 0.009 0.005

Dissolved organic carbón DOC mg-C L-1

nd 1.43 2.98

Particulate organic carbon POC mg L-1

0.28 0.37 0.33

Particulate organic nitrogen PON mg L-1

0.07 0.07 0.04

108


4.3.2 Pseudo-nitzschia composition, abundance and particulate domoic acid

Scanning electron microscopy observations of the phytoplankton net samples

revealed the dominance of P. australis Frenguelli (Hasle 1978) (Fig. 4.3). P. pungens and P.

pseudodelicatissima were also found in the samples. Pseudo-nitzschia was recorded during

all the surveys carried out in the study area and was present in all the samples. Pseudo-

nitzschia cells belonging to P. delicatissima complex were detected at less on one-third of

the samples, mainly in lower abundances than P. seriata complex and only when Pseudo-

nitzschia species belonging to seriata complex were present.

Fig. 4.3. Scanning electron micrographs of Pseudo-nitzschia australis A1) whole valve; A2) central part; A3) tips of the valves B) central part of P. pungens collected from the study area.

109


Overall, pDA concentrations were above the detection limit for the ELISA assay in

the 60% of the samples (n=45). Two DA outbreaks were recorded in ria de Vigo, one in

period 3 and the second on July, 14 2009 (period 2), when also the highest abundances of

Pseudo-nitzschia spp. occurred. In general pDA was detected in the surface waters of the

rias when Pseudo-nitzschia spp. were found in moderate/high abundances. Although some

of the samples contained Pseudo-nitzschia up to abundances of 1.7 104 cells L-1 did not show

detectable DA by ELISA. pDA concentration did not follow the changes in Pseudo-nitzschia

spp. abundance. The highest Pseudo-nitzschia spp. abundance recorded in the area did not

coincide with the highest pDA concentrations. Cellular DA (cDA) concentrations showed a

wide range of values (0-70.3 pg DA cell-1) during the survey period averaging 20.7 pg DA cell-

1. DA cell quota showed a poor linear regression relationship with both Pseudo-nitzschia spp.

abundance (r=0.27) and pDA (r=0.18). During the study period the abundance of Pseudo-

nitzschia spp. and pDA in the surface waters of ria de Vigo and ria de Pontevedra showed

considerable differences in their temporal and spatial variations.

The first toxic event was recorded on period 3 in the ria de Vigo during a strong

downwelling period and it was characterized by high abundances of Pseudo-nitzschia spp.

belong to seriata complex and DA concentrations up to 2.5 μg L-1. Surface pDA was

detectable at all the sites where samples were available. Concentrations of pDA were

higher than 0.5 μg L-1 throughout the middle part of the ria de Vigo and were still detectable

but in lower concentrations (close to 0.05 μg L-1) outside and at the inner part of the ria

(Fig. 4.4). pDA in the ria de Vigo showed a strong positive significant relationship with

Pseudo-nitzschia spp. in this dataset (Pearson´s r=0.919, p=0.001). Pseudo-nitzschia spp.

abundances and pDA concentrations were highly correlated with Secchi disk depth, TSM,

POC, chla and the ratios of Si(OH)4/PO43- and Si(OH)4/N (Table 4.4). For the same survey the

mean cellular DA was 31.7 ranging between 9.5 and 70.3 pg DA cell-1. The highest cellular DA

concentration was recorded at the site (V1) where the lowest Pseudo-nitzschia spp.

abundance (0.1 104 cells L-1) was found.

On July 14, 2009 of period 2 a toxic event mostly due to Pseudo-nitzschia seriata

complex was monitored in the ria de Vigo. During this event Pseudo-nitzschia spp. reached

the highest abundance for the sampling period in the area (12 104 cells L-1). Pseudo-nitzschia

abundance was high at all sites averaging 5.6 104 cells L-1. In contrast to the Pseudo-nitzschia

event recorded on period 3, surface chla was relatively low in the ria de Vigo (0.6-2.3 mg m-

3). pDA was detectable in all the samples and ranged between 0.04 to 1.18 μg L-1. The

highest pDA concentrations were found in samples from sites close to the southern mouth

of ria de Vigo while the lowest concentration was recorded close to the northern mouth of

110


ria (Fig. 4.4). The maximum pDA concentration for this sampling was measured in a sample

which contained 6.9 104 cells L-1 Pseudo-nitzschia spp., while the minimum pDA

concentration was found in a sample which contained the same magnitude of Pseudo-

nitzschia spp. (5.6 104 cells L-1). Cellular DA concentrations averaged 14 pg DA cell-1, showing

a wide range of values 0.7-58.6 pg DA cell-1. There was no significant correlation (Pearson´s

test) between Pseudo-nitzschia spp., DA concentration and cellular DA concentration for

this sampling. Pseudo-nitzschia spp. abundance was positively correlated to the Si(OH)4/N

ratio (r=0.753, p=0.05), while a strong relationship was revealed between cellular DA

concentration and TSIM, NO3-, NO2

2-, Si(OH)4/N (Table 4.4)

Table 4.4. Pearson´s correlations between Pseudo-nitzschia spp. abundances (P-N in cells L-1), particulate DA (pDA in μg L-1) and cellular DA (cDA in pg DA cell-1) concentrations with several environmental parameters during October 10, 2007 and July 14, 2009. *significant at p = 0.05 and ** significant at p = 0.001.

Moreover, it is noteworthy to mention that during the 2008 samplings Pseudo-

nitzschia spp. was recorded in only a few stations and in general in low abundances while

pDA was barely detectable in the most of the samples. Only Pseudo-nitzschia species

belonging to seriata complex were detected during these campaigns. During the summer

campaign of 2007 the maximum abundance of Pseudo-nitzschia spp. (>0.44 104 cells L-1) in

the ria de Vigo was detected at the station 1, located at the innermost part of the ria.

Pseudo-nitzschia spp. abundance decreased almost linearly from the warm surface waters

(around 19 oC) at the inner part of the ria de Vigo to the colder (17 oC) more saline waters

outside the ria. Finally, Pseudo-nitzschia spp. abundances in the ria de Pontevedra averaged

80 cells L-1 on May 27, 2009 and 0.4 104 cells L-1 on July 7, 2009. pDA values in all samples

from ria de Pontevedra were above the detection limit but relatively low (range:0.01-0.07 μg

L-1). Cellular DA levels varied from 0 to 54 pg DA cell-1.

pDA Chla SPM Zsd pH Sal Si(OH)4 Temp PO43- N Si(OH)4/PO4

3- Si(OH)4/N POC

10/10/07

P-N 0.92** 0.77* 0.83* -0.75* 0.20 -0.09 -0.49 0.16 -0.02 0.00 -0.72* -0.71* 0.87*

pDA 0.67 0.78* -0.69 0.07 0.04 -0.44 0.32 -0.32 0.02 -0.71* -0.69 0.84**

cDA 0.36 0.28 -0.22 -0.23 -0.62 0.26 0.83* 0.51 0.09 -0.24 0.00 -0.24

14/07/09

P-N -0.08 0.11 0.66 - - -0.46 0.01 0.28 -0.33 -0.59 -0.23 0.75* 0.03

pDA -0.24 -0.15 - - 0.67 0.00 -0.52 0.02 0.21 -0.03 -0.28 0.48

cDA -0.06 -0.70 - - 0.50 0.00 -0.43 0.71 0.96** -0.16 -0.92** 0.22

111


Fig. 4.4. Distribution of particulate domoic acid (pDA) during the two DA outbreaks recorded in the surface waters of ria de Vigo on A) October 19, 2007 and B) July 14, 2009.

4.3.3 MERIS chla

The application of the fuzzy c-means clustering method in the MERIS FR data

revealed zones where Cluster#1 is dominant and therefore the zones where NNRB can be

best applied in order to obtain more reliable results of chla concentration. Table 4.5 shows

the percentage of pixels belonging to each cluster for each image over the ria de Vigo and

ria de Pontevedra. Cluster#1 was dominant in the two rias, averaging almost 92% of the

pixels in the six images. Cluster#2 and Cluster#3 were represented in the study area with 7.9

and 0.3% of the pixels respectively. In most of the images, the presence of Cluster#2 and

Cluster#3 does not prevent the mapping over large areas in the two rias, since Cluster#1 is

the predominant Cluster. However, in the image acquired on July 14 2009, the relatively

high percentages of pixels belong to Cluster#2 in combination to the cloud cover over the

study area averted a reliable continuous chla mapping of the rias.

The spatial distribution of chla in the study area was analysed from the available

MERIS FR images. Phytoplankton biomass as it was recorded by MERIS FR using the NNRB

chla algorithm was distributed heterogeneously in the study area. Elevated chla

concentrations in the shallow innermost parts of the rias which can be observed in some of

the MERIS images, are most likely due to high abundances of macroalgae on the remote

sensing signal since in-situ measurements showed generally lower values.

112


Table 4.5. Percentage of pixels belonging to each cluster over the study area, obtained from classification images derived from the MERIS images used in this study.

Date ria Cluster#1 Cluster#2 Cluster#3

July 25 2007 Vigo 98.53 0.92 0.55

Pontevedra 99.70 0.30 0.00

October 19 2007 Vigo 97.70 2.30 0.00

Pontevedra 97.93 1.48 0.59

July 09 2008 Vigo 95.40 4.04 0.55

Pontevedra 96.14 3.57 0.30

July 22 2008 Vigo 87.41 11.88 0.71

Pontevedra 90.06 9.79 0.15

May 27 2009 Vigo 98.07 1.58 0.35

Pontevedra 98.98 1.02 0.00

July 14 2009 Vigo 70.59 29.01 0.41

Pontevedra 71.20 28.51 0.29

On July 25 2007 MERIS NNRB surface chla concentrations were relatively low (<1 μg

L-1) in the most parts of ria de Vigo and ria de Pontevedra. An area of chla concentrations

close to 2.5 μg L-1, categorised by the FCM method applied in this study as Cluster#2, was

mapped in the area off the external coast of the rias. Generally, surface chla was higher in

the offshore area, where the lowest concentrations of Pseudo-nitzschia spp. were found.

Maximum MERIS NNRB chla concentrations were observed in the study area on October 19

2007 and coincided with the first toxic event during the sampling period due to Pseudo-

nitzschia spp. in ria de Vigo. MERIS chla mapping on October 19 2007 showed zones of

elevated chla (up to concentrations of 10 μg L-1) at the middle part of the ria de Vigo, where

the highest Pseudo-nitzschia spp. concentrations were measured, and close to the northern

and southern mouths of the ria. These high biomass “patches” are clearly defined in ria de

Vigo.

In ria de Pontevedra, chla concentration was significantly lower in comparison to the

concentrations observed in ria de Vigo and ria de Arousa. Areas of relatively high chla

concentrations were noted in the adjoining area of the rias. In the next MERIS image,

acquired on July 09 2008, a different chla distribution pattern is observed. In this image, the

phytoplankton biomass is principally confined in the rias, while in the area off the rias the

chla concentration remained at levels close to 0 μg L-1. On July 22 2008 a large high chla

biomass event was mapped in the neighbouring shelf area. High concentrations were also

extended to the outside part of ria de Arousa and ria de Pontevedra. The chla retrieved by

MERIS NNRB in the ria de Vigo showed wider variation than the concentrations measured

in-situ. During July 2008 Pseudo-nitzschia abundances remained at low levels. On May 27

113


2009 concentrations of chla lower than 1 μg L-1 were mainly noted in the three rias. NNRB

estimates were well above this value off the rias. Due to cloud cover and relatively large

areas of Cluster#2 pixels in the rias only data from selected areas like the outer part of the

ria de Pontevedra and ria de Arousa can be reliable. There surface chla levels were close to

or lower than 1 μg L-1.

114


Fig. 4.5. Chla maps during the surveys retrieved from MERIS FR data using the NNRB#3. White colour indicates masking of land while grey masking of clouds.

115


4.3.4 Mixed modelling

The results of the generalized additive mixed models (GAMMs), which were used to

model the Pseudo-nitzschia abundance in response to the biotic and abiotic parameters are

summarised in the Table 4.6.

Table 4.6. Results of the generalised additive mixed models used to model the Pseudo-nitzschia spp. with selected biotic and abiotic parameters. For linear covariates the parameter estimates with SE and t-values are given. For the additive covariate the estimated degrees of freedom (edf) and F-value are provided. Periods 1, 2 and 3 corresponds to May, July and October.

Period Estimates SE t

1 -6.85 ±2.83 -2.42

2 3.01 ±1.72 1.74

3 3.73 ±1.82 2.05

Linear effects Estimates SE t

Sal 0.21 ±0.06 3.39

chla 0.89 ±0.31 2.90

Additive effects edf F

Si(OH)4/N 3.08 3.61

The results of the GAMMs suggest the significant (P<0.005) linear effect of salinity

and chla. The optimal model fitted the ratio Si(OH)4/N in an additive fashion (P=0.02). The

cross-validation estimated the degrees of freedom for this function to be 3.08. The

smoothing curve shown in Fig. 4.6 suggests a peak of Pseudo-nitzschia spp. abundance in

moderate values (close to 1.7) of the Si(OH)4/N ratio and an almost linear decrease for

higher values.

Fig. 4.6. GAMM smoothing curve describing the effect of Si(OH)4-:N ratio on Pseudo-nitzschia spp.

116


From all the explanatory variables which were tested to model the DA

concentration, the final GLMM model included the statistically significant effect of POC

(P=0.02), NO3- (P=0.03), PO4

3- (P=0.004) and Si(OH)4/N (P=0.02). Parameters estimates with

SE and t-values for the linear covariates are given in the Table 4.7. Examination of the

residuals did not indicate patterns.

Table 4.7. Results of the best generalised linear mixed model that shows the effect of selected abiotic and biotic parameters on particulate domoic acid (pDA) in our database. Station was included as random effect. Parameter estimates with SE and t-values are given. Periods are the same as in table 4.

Period Estimates SE t

1 -0.12 0.29 0.42

2 0.09 0.23 0.39

3 0.79 0.26 3.07

Parameters Estimates SE t

POC 0.70 0.29 2.42

NO3- -0.38 0.17 -2.23

PO43-

0.69 0.22 3.15

Si(OH)4/N -0.18 0.07 -2.51

4.4 Discussion

4.4.1 Pseudo-nitzschia abundance and species

Pseudo-nitzschia spp. were revealed to be a common component of the

phytoplankton assemblages during the summer period in the study area, because they

were recorded in samples from all summer surveys. Concomitantly, in their review on HABs

in upwelling areas Trainer et al. (2010) reached to a similar conclusion that Pseudo-nitzschia

spp. are present in the Galician rias mainly during late spring-summer. Moreover, we

recorded an autumn-time high abundance event in the ria de Vigo. The Pseudo-nitzschia

assemblages during our study contained mostly Pseudo-nitzschia species belong to the

seriata complex. Pseudo-nitzschia species belonging to the delicatissima complex were also

present in some samples and always along with Pseudo-nitzschia seriata complex species. P.

australis was identified as the dominant species in the net samples examined by electron

microscopy. However, field samples from the two rias contained P. delicatissima and P.

pungens cells. The latter species has been found to consist an important part of the

phytoplankton in the ria de Pontevedra during late winter and summer

upwelling/downwelling conditions (Prego et al., 2007). Previous studies show that Pseudo-

117


nitzschia assemblages in the Galician rias may also include other species like P.

pseudodelicatissima, P. fraudulenta, P. multiseries, P. subpacifica and P. cuspidata (Miguez et

al., 1996; Fraga et al., 1997).

4.4.2 Domoic acid

As far as we know, the present study is the first to provide DA concentrations from

natural phytoplankton populations in the Galician rias. DA has been previously reported

only in shellfish extracts and Pseudo-nitzschia cultures (Arévalo et al., 1998) using HPLC-UV.

From our results, P. australis seems to be the main source of DA in the study area. This

appears to be the most likely situation since P. australis is concurrently the only known DA-

producer (Fraga et al., 1997). Moreover, intoxications by DA in Galician mussels have been

linked to dominant abundances of P. australis (Miguez et al., 1996). The results of this

survey show that relatively high Pseudo-nitzschia spp. abundances and DA concentrations

can be distributed throughout the surface waters of ria de Vigo.

Overall, pDA was detected in more than half of the samples obtained during the

surveys (2007-2009) and relatively high values of pDA and cDA were found. In our data,

particulate domoic acid (pDA) ranged from below detection limits to 2.5 μg L-1. Similar pDA

levels have been measured in the surface waters in other coastal upwelling areas around

the world which have experienced DA outbreaks and where P. australis was the dominant

Pseudo-nitzschia species during the blooms (Busse et al., 2006; Schnetzer et al., 2007;

Garcia-Mentoza et al., 2009). For example, the same magnitude of Pseudo-nitzschia (7 104

cells L-1) abundance that gave us the maximum pDA concentration (2.5 μg L-1) were

detected in San Diego (USA) and the pDA levels reached at 2.3 μg L-1 (Busse et al., 2006).

Cellular DA in field samples varied between 0 and 70.3 pg DA cell-1. We are not aware of

other measurements of cellular DA content in field populations from the Iberian upwelling

system. Several studies along the West coast of USA have recorded analogous ranges of

cellular DA during harmful events mainly due to P. australis (7-75 pg DA cell-1: Scholin et al.,

2000; 0.1-78: Trainer et al., 2000; 5-43 pg DA cell-1: Busse et al., 2006; 0-117 pg DA cell-1:

Schnetzer et al., 2007).

118


4.4.3 Pseudo-nitzschia spp. harmful events in the study area

In our study, the first toxic event in the ria de Vigo was recorded on October, 19

2007. It was characterized by high pDA concentrations due to moderate abundances of

Pseudo-nitzschia. It is worth mentioning that high Pseudo-nitzschia spp. abundances have

been previously observed in the Galician rias during autumn causing the closure of the

mussel harvesting (Moroño, 2000). A second DA episode due to Pseudo-nitzschia spp. was

recorded in the ria de Vigo on July 14 2009 but showed lower DA concentrations. However,

for the years 2007 and 2009 the Technological Institute for the control of the marine

environment of Galicia (INTECMAR) announced that there was not any Pseudo-nitzschia

spp. bloom recorded in the Galician rias associated with mussel intoxication

(www.intecmar.org; Moroño & Pazos., 2009). In parallel, in their detailed annual report,

relatively high Pseudo-nitzschia spp. abundances were shown during mid-October (October

15 and October 22 2007), reaching in some stations values of 2 105 cells L-1. Combining our

results with the INTECMAR database, it can be considered that the high abundances of

Pseudo-nitzschia spp. remained in the ria de Vigo for a maximum of 2 weeks during October.

On July 2009 the closures on the mussel production in the ria de Vigo were associated to

the presence of DSP (Diarrhetic Shellfish Poisoning). On the same year, Pazos and Moroño

(2010) referred very high abundances of P. australis on May in the ria de Vigo and

Pontevedra during a week. Mussel and oyster harvesting closures due to ASP were not

reported during the same periods even if the DA levels in the surface waters were relatively

high. This suggests that these bivalves were not contaminated probably due to the short-

term duration of the harmful events, to the coastal dynamics and to the affinity and

depuration rates of mussels. Spyrakos et al. (2011) describe great changes on

phytoplankton biomass in the Rias Baixas within 3 days. According to Blanco et al. (2002b)

DA events measured as accumulated toxin in the local mussels have been shown to be

characterized by short duration. More specifically, Pazos et al. (2003) mention DA

intoxication in mussels at the end of Pseudo-nitzchia blooms, which normally last for one or

two weeks. Nonetheless, these short-term DA outbreaks, if they are recurrent, can pose a

significant threat to the harvesting of species that show slow DA depuration rates like

scallops (Blanco et al., 2002a; Mauriz & Blanco, 2010). The results of this study deduce that

toxic events due to DA in the Galician rias can be a recurrent phenomenon, be associated

with moderate Pseudo-nitzchia abundances and should be an important concern. Even if in

the Galician rias it has never been reported serious illness or fatalities caused by ASP, DA

levels such the ones detected here have the potential for significant impacts on the

119


ecosystem and human health security and therefore DA in the seawater and not only in

bivalves should be measured routinely in order to assess the potential of a DA outbreak.

4.4.4 Environmental conditions associated with Pseudo-nitzschia spp. abundance

and DA concentration

Coastal upwelling and nutrients availability at the end of the blooms have been

suggested as possible factors implicated in the appearance of toxic Pseudo-nitzschia spp.

events in the Galician rias (Pazos et al., 2003; Trainer et al., 2010). However, the

environmental conditions related to the DA occurrence and distribution in the area is poorly

defined. This study attempts to link environmental properties with the Pseudo-nitzschia spp.

abundances and DA concentrations in order to proximate potential environmental causes

and effects of the regional Pseudo-nitzschia spp. blooms. Therefore, several parameters

(Table 4.3) were determined during the surveys.

The results of the GAMMs for Pseudo-nitzschia spp. revealed the significant linear

effect of salinity. The preference of Pseudo-nitzschia spp. for higher values of salinity found

in our data set confirms previous observations on Pseudo-nitzschia spp. in coastal waters of

Galicia. For example Torres Palenzuela et al. 2010 analysing time series of Pseudo-nitzschia

spp. abundance between the years 1999-2000 concluded that Pseudo-nitzschia spp. blooms

were recorded mainly in high salinity values (>35). Salinity has been suggested as an

important regulator of specific Pseudo-nitzschia species or Pseudo-nitzschia spp. abundance

in other observational studies around the world (e.g. Baltic Sea: Caroppo et al., 2005 and

West Coast of USA: Trainer et al, 2000). Besides salinity, Trainer et al. 2000 and 2002 have

associated high Pseudo-nitzschia abundances with low temperature values and therefore

upwelling events. Palma et al. (2010) in the study on the Pseudo-nitzschia spp. modelling on

the central Iberian coast found a four days delay between the upwelling and the presence

of Pseudo-nitzschia but no direct relation of SST (Sea Surface Temperature) with the

intensity of the upwelling. Our surveys were conducted mainly after or during the

downwelling events off the rias Baixas. Annual data would probably model better the

effect of temperature on the Pseudo-nitzschia abundance. Nevertheless, the two surveys

during an upwelling cycle on July 2008 were accompanied with a small decrease on the

seawater surface temperature in the ria de Vigo (see Figure 3.6 in Chapter III) but not with

high Pseudo-nitzschia abundances. In parallel, after the strong upwelling on July 2008 (July

19), very high values of chla were found a 10m depth, which were missed by the surface

120


sampling. The toxic Pseudo-nitzschia spp. event described here during the autumn survey

on 2007, took place during a downwelling event and after a transition from strong

downwelling at the end of September to strong upwelling the first week of October. In the

ria de Vigo it has been described that diatoms are removed from the surface and bottoms

layers during strong downwelling in the first half of October and after being accumulated

there at the beginning of the downwelling at the end of September (Crespo et al., 2006).

Unpublished data (available only for this survey) of the total phytoplankton revealed that

autrotrophic flagellates dominate (in terms of total abundance) at the station 1V,

cryptophycea at the station 2V, diatoms at the stations 3V-8V and dinoflagellates at stations

9V-10V. Pseudo-nitzschia spp. accounted from 4% (Station 4V and Station8V) to 32% (1V) of

the total diatom abundance. Taken together the results of our study and data published by

INTECMAR it is suggested that diatoms and therefore Pseudo-nitzschia spp. can remain and

bloom in the ria de Vigo for several weeks after the strong downwelling at the end of

September.

The final GAMM fitted to the Pseudo-nitzschia spp. abundance includes the linear

effect of chla. This probably reflects a dominance or co-dominance of Pseudo-nitzschia spp.

in the phytoplankton community. Trainer et al. (2000) and Kudela et al. (2005) and more

recently Trainer et al. (2009) pointed out that there is higher probability for Pseudo-nitzschia

spp. blooms when chla concentration is high. The trend recognised during the sampling

period for higher Pseudo-nitzschia spp. abundances along chla concentrations can be

valuable in the remote sensing detection of high biomass areas in the Galician rias where

these potentially harmful species could be found in high abundances. In this study chla is

retrieved from MERIS FR data using a regional algorithm previously developed. A part of

the chla data from the surveys described here has been used for the development of this

regional neural network-based algorithm, which has been shown to permit accurate

mapping of chla of the rias (Spyrakos et al., 2011). In our MERIS imagery set, Cluster#1 was

dominant in the three rias and the adjacent area allowing reliable chla mapping over the

area. The autumn-time DA event was exemplified by high chla concentrations. The highest

Pseudo-nitzschia spp. abundances (3 to 6 104 cells L-1) were found in a relatively small area in

the ria de Vigo where elevated surface chla biomass (> 6 mg m-3) was mapped from the

MERIS data. These small areas of high chla accompanied with moderate abundances of

toxic Pseudo-nitzschia spp. can easily missed by conventional monitoring techniques. The

MERIS FR and MODIS (not shown here) images from the extended area of the Iberian

Peninsula show very high chla concentrations coupled with colder waters along the

Portuguese and Galician coast indicating a general high biomass event.

121


According to GAMM results Pseudo-nitzschia spp. abundance increases till a

maximum value at moderate values of the ratio of Si(OH)4 to N (mainly NO3-), followed by a

sharp decline at higher values. This may suggest that lower, but not the lowest, ratio values

favour higher Pseudo-nitzschia spp. abundances and partly supports the consideration that

low Si(OH)4/N errands the growth of several Pseudo-nitzschia species (Pan et al., 1996b;

Anderson et al., 2006; Marchetti et al., 2004). Otherwise, these results may reflect the

decline of Si(OH)4 due to high Pseudo-nitzschia spp. abundances. Although the final GAMM

model for Pseudo-nitzschia abundance showed the statistically significant effect of the

explanatory variables: salinity, chla and Si(OH)4/N, these results are limited by the relatively

small number of observations, the no species-specific nature of Pseudo-nitzschia abundance

data and the lack of micronutrients data, which have shown to play an important role on

the Pseudo-nitzschia growth and DA production.

The GLMM identified POC and macronutrients as significant factors influencing pDA

concentration. The enhancement of higher pDA concentrations by low Si(OH)4/N observed

in our data are consistent with the results reported in Anderson et al. (2006) in the Santa

Barbara Channel, California and in other field studies in which DA seems to be associated

with Si-limitation (e.g. Marchetti et al., 2004). In addition, our results associate PO43- and

NO3- concentrations with pDA in the surface seawater. DA production by several Pseudo-

nitzschia species in relation to the availability of these two macronutrients has been studied

in several laboratory experiments since it is considered that they can play an important role

in the DA production (Bates et al., 1991; Pan et al., 1996a, 1996b; Fehling et al., 2004). In

some of these studies PO43--limitation and N-sufficiency (since DA is an amino-acid) have

been pointed out as driving factors for DA production. However, this is in contradiction

with our results where moderate-high pDA values were found in relatively high PO43- and

low NO3- concentrations. Notably, the fact that our results associate high pDA values with

high POC concentrations, implicates rich in pDA POC in the surface waters.

122


4.5 References

Adams, N. G., Lesoing, M., & Trainer, V. L. (2000). Environmental conditions associated with

domoic acid in razor clams on the Washington coast. Journal of Shellfish Research, 19, 1007-1015.

Anderson, C. R., Brzezinski, M. A., Washburn, L., & Kudela, R. (2006). Circulation and

environmental conditions during a toxigenic Pseudo-nitzschia australis bloom in the Santa

Barbara Channel, California. Marine Ecology Progress Series 327, 119-133.

Arévalo, F. F., Bermúdez, M., & Salgado, C. (1998). ASP toxicity in scallops: individual variability

and tissue distribution. In B. Reguera, B. Blanco, M. L. Fernández & T. Wyatt (Eds.), Harmful

Microalgae (pp. 499-502). Santiago de Compostela: Xunta de Galicia and the IOC of UNESCO.

Bakun, A. (1973). Coastal upwelling indices, west coast of North America 1946–71. In, NOAA

Technical Report NMFS-SSRF 671 (pp. 1-103). Seattle.

Bates, S. S., Defreitas, A. S. W., Milley, J. E., Pocklington, R., Quilliam, M. A., Smith, J. C., &

Worms, J. (1991). Controls on domoic acid production by the diatom Nitzschia-pungens f.

multiseries in culture—nutrients and irradiance. Canadian Journal of Fisheries and Aquatic

Sciences, 48, 1136-1144.

Bates, S. S., Garrison, D. L., & Horner, R. A. (1998). Bloom dynamics and physiology of domoic-

acid-producing Pseudo-nitzschia species. In D. M. Anderson, A. D. Cembella & G. M. Hallegraeff

(Eds.), Physiological ecology of harmful algal blooms. Heidelberg: Springer-Verlag.

Blanco, J., Acosta, C. P., de la Puente, M. B., & Salgado C. (2002a). Depuration and anatomical

distribution of the amnesic shellfish poisoning (ASP) toxin domoic acid in the king scallop Pecten

maximus. Aquatic Toxicology, 60, 111-121.

Blanco, J., De La Puente, M. B., Arevalo, F., Salgado, C., & Morono, A. (2002b). Depuration of

mussels (Mytilus galloprovincialis) contaminated with domoic acid. Aquatic Living Resources,

15(1), 53-60.

Busse, L., Venrick, E., Antrobus, R., Miller, P., Vigilant, V., Silver, M., Mengelt, C., Mydlarz, L., &.

Prezelin, B. (2006). Domoic acid in phytoplankton and fish in San Diego, CA, USA. Harmful Algae,

5, 91–101.

Caroppo, C., Congestri, R., Bracchini, L., & Albertano, P. (2005). On the presence of Pseudo-nitzschia calliantha Lundholm, Moestrup et Hasle and Pseudo-nitzschia delicatissima (Cleve) Heiden in the Southern Adriatic Sea (Mediterranean Sea, Italy). Journal of Plankton Research, 277, 63–774.

Crespo, B.G., Figueiras, F. G., Porras, P., & Teixeira, I. G. (2006). Downwelling and dominance of

autochthonous dinoflagellates in the NW Iberian margin: the example of the Ría de Vigo.

Harmful Algae, 5, 770-781.

Cziela, C. (1999). The transport and distribution of the toxic diatom Pseudo-nitzschia spp. in the

123


Coos Bay estuary and the adjacent shelf.

Doerffer, R., & Schiller, H. (2008). MERIS regional coastal and lake case 2 water project—

Atmoshperic correction ATBD, Germany: GKSS Research Center 21502 Geestacht, Version 1.0, 18

May 2008.

Doerffer, R., Sorensen, K., & Aiken, J., 20(9), 1809–1818. (1999). MERIS potential for coastal zone

application. International Journal of Remote Sensing, 20, 1809-1818.

Fehling, J., Davidson, K., Bolch, C. J., & Bates, N. R. (2004). Growth and domoic acid production

by Pseudo-nitzschia seriata (Bacillariophyceae) under phosphate and silicate limitation. Journal

of Phycology, 674-683.

Figueiras, F. G., & Rios, A. F. (1993). Phytoplankton succession, red tides and the hydrographic

regime in the Rías Bajas of Galicia. In T. J. Smayda & Y. Shimizu (Eds.), Toxic Phytoplankton

Blooms in the Sea. Amsterdam: Elsevier.

Fraga, S., Ivarez, M. J., Miguez, A., Fernbndez, M. L., Costas, E., & Lopez-Rodas, V. (1997).

Pseudo-nitzschia species isolated from Galician waters. Proccedings of VIII International

Conference on Harmful Algae. June 25-28, Vigo, Spain.

Fraga, S., Alvarez, M. J., Míguez, A., Fernández, M. L., Costas, E., & Lopez-Rodas, V. (1998).

Pseudo-nitzschia species isolated from galician waters: toxicity, DNA content and lectin binding

assay. In B. Reguera, B. Blanco, M. L. Fernández & T. Wyatt (Eds.), Harmful Algae: Xunta de

Galicia and Inter¬governmental Commission of UNESCO.

Garcia-Mendoza, E., Rivas, D., Olivos-Ortiz, A., Almazan-Becerril, A., Castaneda-Vega, C., & Pena-

Manjarrez, J. L. (2009). A toxic Pseudo-nitzschia bloom in Todos Santos Bay, northwestern Baja

California, Mexico. Harmful Algae, 8, 493–503.

Garet, E., González-Fernández, Á., Lago, J., Vieites, J., & Cabado, A. (2010). Comparative

evaluation of Enzyme-linked Immunoassay and reference methods for the detection of shellfish

hydrophilic toxins in several presentations of seafood. Journal of Agricultural and Food

Chemistry, 20, 1410-1415.

GEOHAB (2005). Global Ecology and Oceanography of Harmful Algal Blooms, GEOHAB Core

Research Project: HABs in Upwelling Systems. In P. Pitcher, Moita, T., Trainer, V.L., Kudela, R.,

Figueiras, P., Probyn, T. (Ed.) (p. 82). Paris and Baltimore.

González Vilas, L., Spyrakos, E., & Torres Palenzuela, J. M. (2011). Neural network estimation of


Remote Sensing of Environment, 115, 524-535.

Herrera, J. L., Piedracoba, S., Varela, R. A., & Roson, G. (2005). Statial analysis of the wind field

on the western coast of Galicia (NW Spain) from in situ measurements. Continental Shelf

Research, 25, 1728-1748.

124


Hidy, G. M. (1972). A view of recent air–sea interaction research. Bulletin of American

Meteorological Society, 53, 1083-1102.

Horner, R. A., Garrison, D. L., & Plumley, F. G. (1997). Harmful algal blooms and red tide

problems on the US west coast. Limnology & Oceanography, 42, 1076-1088.

Klein, C., Claquin, P., Bouchart, V., Le Roy, B., & Véron, B. (2010). Dynamics of Pseudo-nitzschia

spp. and domoic acid production in a macrotidal ecosystem of the Eastern English Channel

(Normandy, France). Harmful Algae, 9, 218-226.

Kudela, R., Cochlan, W., & Roberts, A. (2004). Spatial and temporal patterns of Pseudo-nitzschia species in central California related to regional oceanography. In K. A. Steidenger, J. H. Landsberg, J.H., Tomas, C.R., Vargo, G.A. (Eds.), Harmful Algae 2002. Florida Fish and Wildlife Conservation Commission, Florida Institute of Oceanography, and Intergovernmental Oceanographic Commission of UNESCO, St. Petersburg, Florida, pp. 347–349.

Kudela, R., Pitcher, G., Probyn, T., Figueiras, F., Moita, T., & Trainer, V. L. (2005). Harmful algal

blooms in coastal upwelling systems. Oceanography, 18, 185-197.

Landsberg, G. H, Tomas, C. R., & Vargo, G.A. (Eds.), Harmful Algae 2002 (pp. 347–349). St.

Petersburg, Florida: Florida Fish and Wildlife Conservation Commission, Florida Institute of

Oceanography, and Intergovernmental Oceanographic Commission of UNESCO.

Lin, X., & Zhang, D. (1999). Inference in generalized additive mixed models using smoothing

splines. Journal of the Royal Statistical Society, 61, 381-400.

Maldonado, M. T., Hughes, M. P., Rue, E. L., & Wells, M. C. (2002). The effect of Fe and Cu on

growth and domoic acid production by Pseudo-nitzschia multiseries and Pseudo-nitzschia

australis. Limnology & Oceanography, 47, 515-526.

Maneiro, I., Iglesias, P., Guisande, C., Riveiro, I., Barreiro, A., Zervoudaki, S., & Granéli, E. (2005).

Fate of domoic acid ingested by the copepod Acartia clausi. Marine Biology, 148, 123-130.

Marchetti, A., Trainer, V. L., & Harrison, P. J. (2004). Environmental conditions and

phytoplankton dynamics associated with Pseudo-nitzschia abundance and domoic acid in the

Juan de Fuca eddy. Marine Ecology Progress Series, 281, 1-12.

Marić, M., Ljubešić, Z., Godrijan, J., Viličić, D., Ujević, I., & Precali, R. (2011). Blooms of the

potentially toxic diatom Pseudo-nitzschia calliantha Lundholm, Moestrup & Hasle in coastal

waters of the northern Adriatic Sea (Croatia) Estuarine, Coastal and Shelf Science, 1-9.

Mauriz, A., & Blanco, J. (2010). Distribution and linkage of domoic acid (amnesic shellfish

poisoning toxins) in subcellular fractions of the digestive gland of the scallop Pecten maximus.

Toxicon, 55, 606-611.

Miguez Á., L., F. M., & S., F. (1996). First detection of domoic acid in Galicia (NW Spain). In T.

Yasumoto, Y. Oshima & Y. Fukuyo (Eds.), Harmful and Toxic Algal Blooms (pp. 143-145). Paris,

France: IOC of UNESCO.

125


Moroño, A., & Pazos, Y. (2009). Floraciónes algales nocivas y condiciones oceanográficas en las

rias gallegas durante los años 2007 y 2008. In Proceedings of X Reunion Iberica, Fitoplancon

Tóxico y Biotoxinas. Lisboa, Portugal, 12-15 May 2009.

Moroño, A., Pazos, Y., & Maneiro, J. (2000). Evolución del fitoplancton tóxico y condiciones

oceanográficas asociadas en los años 97–98 en las Rías Gallegas. In, VI Reunión Ibérica sobre

Fitoplancton Tóxico y Biotoxinas (pp. 59-66). Sevilla: Junta de Andalucía.

Mouriño, H., Silva, A., Barão, M., I., & Moita, M., T. (2010). Can Pseudo-nitzschia blooms be

modeled by coastal upwelling in Lisbon Bay? Harmful Algae, 9, 294-303.

Pan, Y, Subba Rao, D. V., &Mann, K. H. (1996a). Changes in domoic acid production and cellular

chemical composition of the toxigenic diatom Pseudo-nitzschia multiseries under phosphate

limitation. Journal of Phycology, 32, 371–381.

Pan, Y. L., Rao, D. V. S., Mann, K. H., Brown, R. G., & Pocklington, R. (1996b). Effects of silicate

limitation on production of domoic acid, a neurotoxin, by the diatom Pseudo-nitzschia

multiseries. 1. Batch culture studies. Marine Ecology Progress Series, 131, 225-233.

Parsons, M. L., Dortch, Q., & Turner, R. E. (2002). Sedimentological evidence of an increase in

Pseudo-nitzschia (Bacillariophyceae) abundance in response to coastal eutrophication.

Limnology & Oceanography, 47, 551-558.

Pazos, Y., & Moroño, A. (2011). Monitoración de condiciones oceanograficas y fitoplankton en

las zonas de producción de moluscos de Galicia (2007-2010). In Proceedings of XI Reunion Iberica,

Fitoplancon Tóxico y Biotoxinas. Bilbao, Junio 2009.

Pitcher, G. C., G., F. F., Hickey, B. M., & Moita, M. T. (2010). The physical oceanography of

upwelling systems and the development of harmful algal blooms. Progress in Oceanography 85,

5-32.

Prego, R., Guzmán-Zuñiga, D., Varela, M., & deCastro M (2007) Consequences of winter upwelling events on biogeochemical and phytoplankton patterns in a western Galician ria (NW Iberian peninsula). Estuarine, Coastal and Shelf Science, 73, 409–422. R Development Core Team, 2009. R: A Language and Environment for Statistical Computing, Vienna. <www.R-project.org>.

Reguera, B., Garcés, E., Bravo, I., Pazos, Y., Ramilo, I., & González, G. (2003). Cell cycle patterns

and estimates of in situ division rates of dinoflagellates of the genus Dinophysis by a postmitotic

index. Marine Ecology Progress Series 249, 117-131.

Rodríguez Rodriguez, G., Villasante, S., & García-Negro, M. (2011). Are red tides affecting

economically the commercialization of the Galician (NW Spain) mussel farming? Marine Policy,

35, 252-257.

Smith, J. C., Cormier, R., Worms, J., Bird, C. J., Quilliam, M. A., Pocklington, R., Angus, R., &

Hanic, L. (1990). Toxic blooms of the domoic acid containing diatom Nitzschia pungens in the

126


Cardigan River, Prince Edward Island. In E. Graneli, B. Sundstrom, L. Edler & D. M. Anderson

(Eds.), Toxic Marine Phytoplankton (pp. 227–232). New York.

Schnetzer, A., Miller, O., Schaffner, R., Stauffer, B., Jones, B., Weisberg, S., DiGiacomo, P.,

Berelson, W., & Caron, D. (2007). Blooms of Pseudo-nitzschia and domoic acid in the San Pedro

Channel and Los Angeles harbor areas of the Southern California Bight, 2003–2004. Harmful

Algae, 6, 372–387.

Scholin, C., Gulland, F., Doucette, G., Benson, S., Busman, M., Chavez, F., Cordaro, J., Delong, E.,

Vogelaere, A., Harvey, J., Haulena, M., Lefebvre, K., Lipscomb, T., et al., (2000). Mortality of sea

lions along the central California coast linked to a toxic diatom bloom. Nature, 403, 80–84.

Spyrakos, E., González Vilas, L., Torres Palenzuela, J. M., & Barton, E. D. (2011). Remote sensing

chlorophyll a of optically complex waters (rias Baixas, NW Spain): Application of a regionally

specific chlorophyll a algorithm for MERIS full resolution data during an upwelling cycle. Remote

Sensing of Environment, 115, 2471-2485.

Tomas, C. R. (1997). Identifying Marine phytoplankton. California, USA: Academic press.

Torres Palenzuela, J., González Vilas, L., & Spyrakos, E. (2010). Artificial Neural Network model

for predicting Pseudo-nitzschia spp. abundance in the Galician Rias (NW Spain). In proceeding of

14th International Conference on Harmful Algae, May 2010, Creta, Greece.

Trainer, V. L., Bill, B. D., Stehr, C. M., Wekell, J. C., Moeller, P., Busman, M., & Woodruff, D.

(2000). Domoic acid production near California coastal upwelling zones, June 1998. Limnology &

Oceanography, 45, 1818-1833.

Trainer, V. L., Hickey, B. M., & Horner, R. A. (2002). Biological and physical dynamics of domoic

acid production off the Washington coast. Limnology & Oceanography 47, 1438-1446.

Trainer, V. L., MacFadyen, A., & Moore, S. K. (2009). Variability of Pseudo-nitzschia and domoic

acid in the Juan de Fuca eddy region and its adjacent shelves. Limnology & Oceanography, 54,

289-308.

Trainer, V. L., Pitcher, G. C., Reguera, B., & Smayda, T. J. (2010). The distribution and impacts of

harmful algal bloom species in eastern boundary upwelling systems. Progress in Oceanography,

85, 33-52.

Utermöhl, H. (1958). Zur vervollkommnung der quantitativen phytoplankton-methodik.

Mitteilung Internationale Vereinigung Theoretische unde Amgewandte Limnologie, 9, 1-38.


(NW Spain): main facies and hydrodynamic dependence. Journal of Marine Systems, 54, 261-276.

Wells, M. L., Trick, C. G., Cochlan, W. P., Hughes, M. P., & Trainer, V. L. (2005). Domoic acid: The

synergy of iron, copper, and the toxicity of diatoms. Limnology & Oceanography 50, 1908-1917.

127


Wood, S. N. (2000). Modelling and smoothing parameter estimation with multiple quadratic

penalties. Journal of the Royal Statistical Society, 62, 413-428.



pyridine-containing mobile phases. Marine Ecology Progress Series, 195, 29-45.

Zuur, A., Ieno, E., & Smith, G. (2007). Analysing ecological data. New York, USA: Springer.

Zuur, A., Ieno, E., Walker, N., Saveliev A., & Smith, G. (2009). Mixed effects models and extensions

in ecology with R. New York, USA: Springer.

Chapter V. N.scintillans PST vector?

129

Ingestion and clearance rates of the red Noctiluca scintillans fed on the toxic dinoflagellate Alexandrium minutum (Halim)

Abstract

5.1 Introduction


5.3 Results

5.4 Discussion

5.5 References

CHAPTER V

131


Abstract

In this study, we evaluated the ingestion and clearance rates of the heterotrophic

dinoflagellate Noctiluca scintillans (Macartney) fed on the toxic microalgae Alexandrium

minutum Halim in laboratory cultures. Specimens of the red form of N. scintillans were

collected in the ria de Vigo and kept in culture. The results showed that N. scintillans actively

fed on A. minutum, with maximum ingestion rates of 0.37 μg C ind−1 day−1, showing that

there was non-satiated feeding by N. scintillans on A. minutum. The mean cellular toxin

content of A. minutum in the stock culture ranged from 2.57 to 3.44 fmol cell−1. The HPLC

analyses of the extracts from N. scintillans fed with 4 different abundances of A. minutum

revealed no detectable amounts of toxin. This suggested that although N. scintillans was

able to ingest a toxic dinoflagellate such as A. minutum into its vacuoles, the ingested toxin

was bioconverted or excreted by Noctiluca. Our findings support the theory that Noctiluca

may inflict grazing pressure on the growth of Paralytic Shellfish Toxin (PST) species in the

field, and could therefore play an important role as a regulator against PST-producing

phytoplankton. This is the first report about grazing rates of N. scintillans on Alexandrium

cells.

132


5.1 Introduction

Heterotrophic dinoflagellates are common members of the plankton and play an

important role in the marine pelagic systems as regulators of phytoplankton biomass

(Lessard, 1991; Strom & Morello, 1998; Jeong, 1999; Sherr & Sherr, 2007). One of the most

common organisms in this group is Noctiluca scintillans (Ehrenberg) Macartney. N. scintillans

is a large non-photosynthetic dinoflagellate with worldwide distribution which dominates

some communities of plankton grazers and occasionally causes spectacular and striking red

tides in several coastal regions of the world (Fung & Trott, 1973; Schaumann et al., 1988;

Huang & Qi, 1997; Elbrächter & Qi, 1998; Dela-Cruz et al., 2002; Fonda-Umani et al., 2004;

Fukuda & Endoh, 2006).

N. scintillans appears in the water column in two distinctive morphotypes: the

tropical or green form, which contains endosymbiotic algae (Sweeney, 1976; Hansen et al.,

2004; Furuya et al., 2006) and the strictly heterotrophic or red form that inhabits more

temperate waters (Kiørboe & Titelman, 1998). One of the most noteworthy characteristics

of N. scintillans is its highly omnivorous feeding behavior. It consumes a wide range of prey

such as phytoplankton cells (Nakamura, 1998a; Dela-Cruz et al., 2002; Hansen et al., 2004),

copepods eggs (Kimor, 1979; Daan, 1987; Quevedo et al., 1999), copepod nauplii (Dela-Cruz

et al., 2002), fish eggs (Enomoto, 1956; Hattori, 1962), fecal pellets (Kiørboe, 2003), marine

snow (Shanks & Walters, 1996) and bacteria (Kirchner et al., 1996). The main factors that

control the population growth of N. scintillans are water temperature, salinity,

phytoplankton biomass and natural nutrient enrichment processes (Uhling & Sahling, 1990;

Lee & Hirayama, 1992a; Buskey, 1995; Huang & Qi, 1997; Dela-Cruz et al., 2002; Dela-Cruz et

al., 2003).

Although N. scintillans does not produce toxins, it is categorized as harmful algae

due to the high levels of ammonium accumulated inside the cells and oxygen depletion

which occurs during the decay of high abundance events (Smayda, 1997). Several cases of

mass mortalities in caged fish and shrimp farms due to ammonium release by N. scintillans

have been reported (Okaishi & Nishio, 1976; Suvapepun, 1989; Naqvi et al., 1998; Montani et

al., 1998). More recently, Kirchner et al. (2001) and Seibold et al. (2001) pointed out that

Noctiluca cells may contain a large number of endocytic bacteria, which could be involved in

the production of harmful substances.

It is assumed that diatoms are generally the phytoplankton group most preferred

by N. scintillans as food items, both in field populations and in cultures (Buskey, 1995;

http://www.sciencedirect.com/science/article/pii/S1568988310001228#ref_bib0195




















































133


Kirchner et al., 1996; Jeong & Shim, 1996; Kiørboe & Titelman, 1998; Nakamura, 1998b;

Tiselius & Kiørboe, 1998). Hanslik (1987) reported that the diatom Skeletonema costatum is

a high value prey item, when compared to a dinoflagellate like Scrippsiella trochoidea.

Phytoplankton species such as Heterocapsa inlandica, Tetraselmis tetrathelle, and

Gymnodinium nagasakiense, have been tested as food items of N. scintillans. Takayama

(1977) showed that N. scintillans fed on H. inlandica increased in density up to 200× in 3

weeks, while Lee and Hirayama (1992b) showed no differences in clearance rates between

T. tetrathelle and G. nagasakiense. Although there is no clear trend in terms of diet, it seems

that the relationship between N. scintillans and its potential prey may be species-specific.

Few laboratory studies have addressed the interaction of Noctiluca with toxic

phytoplankton. Hinz et al. (2001) used toxic and non-toxic strains of Alexandrium fundyense

and Alexandrium tamarense to feed Noctiluca, but their study only assessed possible

ingestion by crab larvae, and they did not determine grazing rates or possible toxic effects.

Hansen et al. (2004) and Azanza et al. (2010) fed green N. scintillans with the dinoflagellate

Pyrodinium bahamense var. compressum, a tropical PST producer. Results showed that N.

scintillans ingested P. bahamense var. compressum, but at relatively low rates. In general, it

seems that heterotrophic dinoflagellates exhibit low ingestion rates compared to other

groups of predators (e.g. ciliates, see Strom & Morello, 1998).

In field studies, predation by N. scintillans on toxic dinoflagellate species, such as

Dinophysis caudata and D. acuta (Escalera et al., 2007) and Gymnodinum catenatun (Alonso-

Rodríguez et al., 2005) has been recorded, with the prey being ingested into the vacuoles of

the cell. Although there are no available quantitative data on the removal rates of toxin-

producing phytoplankton species by N. scintillans, it is generally considered that N.

scintillans could act as a regulator of the dynamics of harmful algal events.

The aim of this study was to examine whether N. scintillans could actively feed on a

PST-producing dinoflagellate. Also, to assess if N. scintillans can eventually act as a PST

vector, individuals were tested for the presence of toxin accumulated in their cells.

5.2 Materials and methods

5.2.1 Noctiluca scintillans collection

Specimens of the red N. scintillans were collected by vertically integrated tows, at

39 m depth, close to a field station located in the ria de Vigo, Spain (42°13.3′N, 8°47.7′W), an



















134


area that has experienced blooms of toxic and non-toxic Alexandrium spp. (Centro de

Calidad del Medio Marino de la Xunta de Galicia, Vilaxoan, monitoring reports). Samples were

transported to the laboratory within 2 h of collection. In the laboratory ∼500 N. scintillans

cells were transferred by a pipette into 250 ml beakers of aged natural seawater previously

filtered through 0.2 μm membrane filter. Samples were incubated at 18 °C with an

illumination of 45 μmol photon m−2 s−1 provided by cool-white fluorescent tubes with a 12:12

light/dark cycle for two weeks, using the dinoflagellate Heterocapsa triquetra as a food

source.

5.2.2 Algal species

The non-axenic dinoflagellate strain of Alexandrium minutum (AL1V) used in this

study was isolated from the Galician rias, and came from long-established populations

cultured in the Instituto Español de Oceanografía (Vigo). N. scintillans were fed algae at the

exponential phase, because toxin composition is constant in exponential growth-phase

cultures (Franco et al., 1994; Parkhill & Cembella, 1999). A. minutum is a toxic strain that

only contains gonyautoxins 1–4 (GTX1–4) (Franco et al., 1994).

Cell carbon and nitrogen content were determined on the first and last day of the

experiment, from triplicate subsamples filtered on pre-combusted GF/F filters at low

pressure, dried at 70 °C and combusted in a Fisons EA-1108 CHN analyzer. Sulfanilamide was

used as the standard. Cell diameters and carbon and nitrogen content of A. minutum are

shown in Table 5.1. Biovolume was calculated by the most likely geometrical shape method

(Olenina et al., 2006), with A. minutum cells considered to be a rotational ellipsoid.

Table 5.1. Cell parameters for Alexandrium minutum measured during the experiment. Carbon and nitrogen content, C:N ratio, cell diameter measured with inverted microscope and biovolume (means ± S.D., n = number of determinations).

Cell parameter n Alexandrium minutum

Carbon (pg cell-1

) 7 798.9±10.9

Nitrogen (pg cell-1

) 7 176.2±3.1

C:N 7 4.5±0.0

Cell diameter (μm) 60 21.7±0.4

Volume (μm-3

10-3

) 60 4.0±0.3




http://www.sciencedirect.com/science/article/pii/S1568988310001228#ref_tbl0005


135


5.2.3 Experimental design

Noctiluca cells were starved for 72 h in 250 ml beakers containing filtered seawater.

Fifteen individuals in the trophon stage were then pipetted out into each of twenty-four 40

ml beakers which contained experimental food concentrations. Four concentrations of A.

minutum, ranging from 50 to 400 cell ml−1 (low: 50; moderate: 100, 200; high: 400) were

prepared with six replicate samples per concentration. Only individuals with active tentacle

movement were used.

The culture medium was prepared with aged natural seawater (salinity 33.6 ‰),

filtered through GF/F Whatman filters and autoclaved. N. scintillans cells were kept

throughout the experiment at 18 °C under two different conditions: twelve replicates (3 for

each experimental food concentration) in 12:12 h light/dark cycle (condition 1) and twelve

more in completely dark conditions (condition 2), with the same experimental food

concentrations as condition 1. Each day, N. scintillans cells were isolated and pipetted to a

new beaker with A. minutum suspensions at the experimental concentration. Mortality in N.

scintillans was lower than 6% day−1 in all experimental food concentrations. Cell abundance

of A. minutum, toxin content per cell, clearance rates (the volume of water swept clear of

A. minutum cells) and ingestion rates were all estimated on a daily basis for the different

experimental food concentrations.

5.2.4 Grazing estimation

All the replicates (15 individuals per 40 ml beaker) were used to estimate the

ingestion rates of the red N. scintillans fed on A. minutum at each experimental food

concentration. Initial beakers without N. scintillans were prepared simultaneously. Initial

and subsequent, subsamples were immediately preserved using 4% formaldehyde for cell

counting. Grazing experiments were run for 24 h at the temperature and light conditions

described above. The mortality of N. scintillans was checked everyday after 24 h incubation

using a steromicroscope. A. minutum subsamples were preserved with 4% formaldehyde for

cell counting. Abundances of A. minutum were determined by quadruplicate counting of 1

ml of sample with a Sedgewick-Rafter chamber using an inverted microscope. Frost's (1972)

equations were used to calculate clearance and ingestion rates.


136


5.2.5 Toxin analysis

To estimate the cell toxin content of A. minutum, algal cells were collected on pre-

combusted 13 mm GF/F Whatman filters and stored at −30 °C in ultracentrifuge plastic tubes

and lyophilized; 300 μl of 0.05 M acetic acid was added to the lyophilized material, and the

sample was homogenized using a pipette tip adapted to fit the shape of the vial. The

sample was shaken then frozen twice. Finally, the extract was centrifuged twice at 9000

rpm for 5 min, after which 200 μl of the supernatant was carefully collected with a Hamilton

syringe, and stored at −30 °C. For analysis of PST content in N. scintillans, on the last day of

the experiment a minimum of 15 and a maximum of 25 individuals from each experimental

food concentration were transferred to filtered seawater then to distilled water and

collected with a known volume (≤1 ml). Samples of N. scintillans were stored at −30 °C in

ultracentrifuge plastic tubes and lyophilized; 100 μl of acetic acid (0.05 M) was added to the

lyophilized material which was then processed as for A. minutum samples.

Analysis of the Paralytic Shellfish Poisoning (PSP) toxins (by high-performance liquid

chromatography (HPLC) with fluorescence detection) was carried out following a

modification of the method of Oshima et al. (1989), described by Franco and Fernández

(1993). Chromatographic profiles of A. minutum cells were determined by quadruplicate

injections of 35 μl of extracts (diluted with 0.05 M acetic acid, as necessary).

Chromatographic profiles of N. scintillans were determined by injection of 35 μl of the

extracts. Toxins from the National Research Council of Canada (Halifax) were used as toxin

standards. The toxicity of A. minutum in saxitoxin equivalents (STXeq) was calculated from

the HPLC chromatograms. The toxin concentrations were multiplied by a toxin-specific

conversion factor to yield toxicity. The specific toxicity conversion factors of the individual

toxins were adopted from Oshima (1995), based upon empirical mouse bioassay data

determined using purified standards, and assuming the conversion factor of 1 mouse unit

(MU) = 0.23 μg STXeq for the ddy mouse strains: 567.6 (GTX1), 205.2 (GTX2), 364.3 (GTX3)

and 414.7 (GTX4).






137


5.3 Results

Table 5.2 shows the specific toxic composition, total toxin per cell and total cell

toxicity of Alexandrium minutum during the experiment, and of the red Noctiluca scintillans

cells surviving on day 5, when the experiment was completed. The mean cellular toxin

content of A. minutum in the stock culture ranged from 2.57 up to 3.44 fmol cell−1 on days 3

and 4, respectively. There were no significant differences in total cellular toxin content of A.

minutum throughout the experiment (ANOVA, F2,6 = 0.838, p = 0.478).

Table 5.2. Alexandrium minutum and Noctiluca scintillans. Specific toxin composition of gonyautoxins, as a percentage (GTX1-4, mean ± SE, fmol cell-1), total toxin per cell (combined GTX1, GTX2, GTX3, GTX4, mean ± SE, fmol cell-1), and total cell toxicity (mean ± S.E., fg STXeq cell-1) of each one during the experimental period. (STXeq = saxitoxin equivalents, n.d. = not detected). Data presented in order of decreasing toxin content.

HPLC analyses of the extracts from N. scintillans cells fed with 4 different

abundances of A. minutum for the two different treatments revealed no detectable

amounts of toxin (Table 5.2). Notwithstanding, N. scintillans was able to ingest the toxic

dinoflagellate A. minutum into its vacuoles (Fig. 5.1). Survival rates (Fig. 5.2) of N. scintillans

were higher than 75% in all food concentrations and treatments, for the 4 days of the

experiment.

Species Day GTX4 GTX1 GTX3 GTX2 Total toxin Cell toxicity

A. minutum 2 56.53 ± 0.88 38.18 ± 0.55 3.00 ± 0.26 0.63 ± 0.02 2.69 ± 0.83 1260 ± 389

3 46.80 ± 2.71 44.02 ± 1.89 3.93 ± 0.20 1.78 ± 0.24 2.57 ± 0.22 1211 ± 101

4 67.87 ± 3.49 28.57 ± 3.22 1.96 ± 0.11 0.50 ± 0.06 3.44 ± 0.21 1563 ± 85

N. scintillans 5 n.d. n.d. n.d. n.d. n.d. n.d.



http://www.sciencedirect.com/science/article/pii/S1568988310001228#ref_fig0005


138


Fig. 5.1. Noctiluca scintillans with ingested Alexandrium minutum cells. Scale bar = 20 μm.

139


Fig. 5.2. Noctiluca scintillans survival rates as a function of Alexandrium minutum abundance for the 4-day incubation period. Open circles represent condition 1, when N. scintillans was fed with A. minutum under light/dark conditions of 12:12 h, whereas closed circles represent condition 2, when N. scintillans was fed with A. minutum under conditions of complete darkness. All data are means ± 1 S.E.

The ingestion rate of N. scintillans increased continuously under conditions of

complete darkness, with increasing A. minutum concentration up to a maximum value of

0.37 μg C ind−1 day−1 (i.e. 466 cells ind−1 day−1) at the high concentration during the 4 days of

the experiment (Fig. 5.3A). When N. scintillans cells were fed with A. minutum under

light/dark conditions of 12:12 h, the ingestion rate became saturated at a concentration of

0.30 μg C ml−1 (i.e. 370 cells ml−1). A Mann–Whitney test between the ingestion rate and A.

minutum abundance showed that there were only significant differences in ingestion rates

between the 2 different experimental treatments at concentration 4 (Table 5.3). Clearance

rates of N. scintillans fed on A. minutum were in the ranges 0.44–1.55 and 0.55–1.91 ml ind−1

day−1 for conditions 1 and 2, respectively (Fig. 5.3B). N. scintillans had a higher clearance at

moderate concentrations of A. minutum, followed by a steep, sharp decline at the high

concentration. At a low concentration, clearance rates were extremely low. When




140


comparing clearance rates between treatments, N. scintillans only displayed a higher rate

for condition 1 than for condition 2 at low concentrations.

Fig. 5.3. Noctiluca scintillans. Ingestion (A) and clearance (B) rates of Alexandrium minutum as a function of A. minutum abundance for the 4-day incubation period. All data are means ± 1 S.E. Symbols as in Fig. 5.2.


141


Table 5.3. Mann-Whitney tests were carried out with ingestion rate as the dependent variable and Alexandrium minutum abundance as a factor. The 4 cell concentrations of Alexandrium minutum versus the 2 different experimental conditions were compared: (1) 46, (2) 100, (3) 201, (4) 406 mean values in cells ml-1. *Significant p-values (α = 0.05).

Mann-Whitney

All concentrations 1 2 3 4

Condition 1 U=15167

p=0.045*

U=903

p=0.475

U=935

p=0.652

U=926

p=0.391

U=651

p=0.006*

Condition 2

5.4 Discussion

The red Noctiluca scintillans feeds on the toxic strain of Alexandrium minutum used

in this study (Fig. 5.1 and Fig. 5.3A). Irrespective of the treatment, the relationship between

ingestion rate and A. minutum abundance in the feeding experiment was similar to those

reported for Noctiluca feeding on other algae, such as Tetraselmis tetrathelle and

Gymnodinium nagasakiense (Lee & Hirayama, 1992b) and Chatonella antiqua and

Heterosigma akashiwo (Nakamura, 1998a). Maximum rates were recorded at high

abundances and minimum rates at low abundances of food.

There was an interesting result in light/dark conditions, where the ingestion rate

became saturated at 0.30 μg C ml−1 (i.e. 370 cells ml−1), indicating limitation in the food

consumption ability of Noctiluca under these conditions. In contrast, under conditions of

complete darkness, the highest mean ingestion rate of the experiment (466 cells ml−1, 0.37

μg C ml−1) was recorded at the highest A. minutum concentration, showing that there is

non-satiated feeding of N. scintillans on A. minutum. Similar behavior was observed by Dutz

(1998) and Frangópulos et al. (2000) with the copepod Acartia clausi fed on the low toxic

strain of Alexandrium lusitanicum and A. minutum, respectively. In a study of the growth

and grazing responses of green Noctiluca feeding on the PST-producing dinoflagellate

Pyrodinium bahamense var. compressum, Hansen et al. (2004) also found that the green

Noctiluca preyed actively on Pyrodinium, without exhibiting signs of reduced ingestion or

satiation, at values of 540 P. bahamense cells ml−1, close to the 466 A. minutum cells ml−1

observed in our study for condition 2 (complete darkness). Nevertheless, the ingestion

rates of N. scintillans fed on A. minutum are considerably lower than those of the copepod

Acartia clausi fed on the same species and the same strain (Frangópulos et al., 2000;

Guisande et al., 2002; Barreiro et al., 2006). As pointed out by Kiørboe and Titelman (1998),

this is mainly because Noctiluca preys more efficiently on immobile (diatoms) than on

mobile food (for example dinoflagellates). The use of a plankton wheel to rotate bottles

http://www.sciencedirect.com/science/article/pii/S1568988310001228#fig0005















142


instead of a static culture, as suggested by Buskey (1995), would keep the phytoplankton

more disperse and could possibly have improved the contact rate of Noctiluca with its food.

Even so, our findings support the likelihood that the red Noctiluca is capable of limiting the

growth of PST species in the field, and may play an important role as a regulator against

PST-producing phytoplankton.

The clearance rates of the red N. scintillans fed on A. minutum are directly related to

food concentration, and showed a trend of decreased clearance as the prey concentration

increased, which is similar to that seen by Lee and Hirayama (1992b) in feeding experiments

with T. tetrathelle, and by Hansen et al. (2004) with P. bahamense. In our case, the clearance

rate curve began to decline when the food concentration increased over 0.1 μg C ml−1 (or

130 cells ml−1), which is very similar to the densities reported by Hansen et al. (2004) for the

green Noctiluca. Depending upon the cell concentration of A. minutum, N. scintillans was

able to remove A. minutum cells with a clearance rate of 0.44–1.55 and of 0.55–1.91 ml ind−1

day−1 for treatments 1 and 2, respectively. These are higher values than those previously

reported: 0.48 ml ind−l day−1 (Lee & Hirayama, 1992b), 0.10–0.35 ml ind−1 day−1 (Nakamura,

1998b) and 0.06 ml ind−1 day−1 (Hansen et al., 2004). As indicated in Fig. 5.2, survival rates of

N. scintillans were always higher than 75% in all food concentrations and treatments for the

4 days of the experiment. The cell loss that did occur was attributed to manipulation, as

each day N. scintillans cells were isolated and pipetted to a new beaker with new food

suspensions. The mortality observed was therefore not related to consumption of toxic

algae or other variables such as cannibalistic feeding behavior, which has previously been

detected in N. scintillans (Lirdwitayaprasit, 2002) and in other heterotrophic dinoflagellates

(Jeong & Latz, 1994; Latz & Jeong, 1996). Consequently, in our study there was no evidence

that the toxicity of A. minutum has an inhibitory effect on the growth and feeding activities

of N. scintillans.

The growth rates of Noctiluca were not evaluated in our study, but during the

experimental phase it was noted that they were very low. This was probably because the

experimental food concentrations used were not sufficiently high, and the growth rate of

this form of Noctiluca depends very much on the concentration, size and quality of the food

item (Lee & Hirayama, 1992a; Buskey, 1995; Kiørboe & Titelman, 1998; Nakamura, 1998a;

Nakamura, 1998b). Moreover, with light intensity values identical to ours, but using green

Noctiluca which metabolism is different to the red Noctiluca used in this experiment,

Hansen et al. (2004) recorded growth rates of 0.058 day−1, indicating that intermediate or

low growth rate values are typically obtained using dinoflagellates or poor food quality

species as prey. It is probable that as a mobile prey, A. minutum is not the most suitable






















143


prey for Noctiluca. However, as Noctiluca is a voracious non-selective feeder that shows no

evidence of food preference (Uhling & Sahling, 1990), dinoflagellates constitute an

alternative food item that should be considered as a food source for Noctiluca.

The strain of A. minutum used in this study is at the low end of the toxicity range

observed for species of the genus Alexandrium (Chang et al., 1997). In both conditions,

HPLC analyses of the extracts from N. scintillans cells fed with A. minutum revealed no

detectable amounts of toxin (see Table 5.2). Noctiluca was still, however, able to ingest A.

minutum as shown in Fig. 5.1 and Fig. 5.3, indicating that Noctiluca bioconverted or excreted

the ingested toxins rather than bioaccumulating them. Recently Azanza et al. (2010) found

similar results studying the effects of the green Noctiluca fed on the PST-producing

dinoflagellate P. bahamense. These authors found that when Noctiluca is exposed to similar

concentrations to those used in our study (the authors’ setups 1:500 and 5:1000 are

equivalent to our 200 and 400 A. minutum ml−1), the predators did not contain any toxins, or

they were not present in detectable amounts, indicating the capacity of Noctiluca to

bioconvert and/or excrete the saxitoxins. The differences between the results of the two

studies could be because the toxin analyses of Azanza et al. (2010) were undertaken with

the culture, whereas ours were directly measured on a Noctiluca extract. Although Hansen

et al. (2004) also found grazing pressure of the green Noctiluca on Pyrodinium, toxic

concentrations were not measured. From our photographic evidence (Fig. 5.1), we believe

that Noctiluca is capable of accumulating PST in its vacuoles, perhaps at lower rates, but

detoxification/excretion processes could occur rapidly after feeding, as has been seen for

the ciliate Favella taraikaensis feeding on Alexandrium tamarense (Kamiyama & Suzuki,

2006).

Future work and further assays with a higher toxicity strain are necessary to clarify

this process and to achieve a better understanding of the ecological role of N. scintillans in

the planktonic community and also further research should include the use of plankton

wheels since cells may accumulate on the side of the container and/or one part of the

predator organism thus affecting the quality of the experiment.















144


5.4 References

Alonso-Rodríguez, R., Ochoa, J. L., & Uribe-Alcocer, M. (2005). Grazing of heterotrophic

dinoflagellate Noctiluca scintillans (Mcartey) Kofoid on Gymnodinium catenatum Graham. Revista

Latina de Microbiologia, 47(1-2), 6-10.

Azanza, R. V., Cruz, L. J., Cariño, F. A., Blanco, A. G., & Butardo, V. M. (2010). Paralytic shellfish

toxin concentration and cell density changes in Pyrodinium bahamense-Noctiluca scinitllans

feedings experiments. Toxicon, 55, 1017-1023.

Barreiro, A., Guisande, C., Frangópulos, M., González-Fernández, A., Muñoz, S., Pérez, D.,

Magadán, S., Manerio, I., Riveiro, I., & Iglesias, P. (2006). Feeding strategies on the copepod

Acartia clausi, on single and mixed diets of toxic and non-toxic strains of the dinoflagellate

Alexandrium minutum. Marine Ecology Progress Series, 316, 115-125.

Buskey, E. J. (1995). Growth and bioluminescence of Noctiluca scintillans on varying algal diets.


Chang, F. H., Anderson, D. M., Kulis, D. M., & Till, D. G. (1997). Toxin production of Alexandrium

minutum (Dinophyceae) from the Bay of Plenty, New Zealand. Toxicon 35, 393–409.

Daan, R. (1987). Impact of egg predation by Noctiluca miliaris on the summer development of

copepod populations in the southern North Sea. Marine Ecology Progress Series, 37, 9-17.

Dela-Cruz, J., Ajani, P., Lee, R., Pritchard, T., & Suthers, I. (2002). Temporal abundance patterns

of the red tide dinoflagellate Noctiluca scintillans along the southeast coast of Australia. Marine

Ecology Progress Series, 236, 75-88.

Dela-Cruz, J., Middleton, J. H., & Suthers, I. M. (2003). Population growth and transport of the

red tide dinoflagellate, Noctiluca scintillans, in the coastal waters off Sydney Australia, using cell

diameter as a tracer. Limnology & Oceanography, 48(2), 656-674.

Dutz, J. (1998). Repression of fecundity in the neritic copepod Acartia clausi exposed to the toxic

dinoflagellate Alexandrium lusitanicum: relationship between feeding and egg production.

Marine Ecology Progress Series, 175, 97–107.

Elbrächeter, M., & Qi, Y. Z., (1998). Aspects of Noctiluca (Dinophyceae) populations dynamics. In

D. M Anderson, A. D. Cembella & G. M. Hallegraeff (Eds.), Physiological ecology of harmful algal

blooms. NATO ASI Series, Vol G41: Ecol. Sciences, Springer-Verlag, Berlin, pp. 315-335.

Enomoto, Y. (1956). On the occurrence on Noctiluca scintillans (Macartney) in the waters

adjacent to the west coast of Kyushu, with special reference to the possible damage caused to

the fish eggs by that plankton. Bulletin of the Japanese Society for the Science of Fish, 22, 82-88.

Escalera, L., Pazos, Y., Moroño, A., & Reguera, B. (2007). Noctiluca scintillans may act as a vector

of toxigenic microalgae. Harmful algae, 6, 317-320.

145


Fonda-Umani, S., Beran, A., Parlato, S., Virgilio, D., Zollet, T., De Olazabal, A., Lazzarini, B., &

Cabrini, M. (2004). Noctiluca scintillans Macartney in the Northern Adriatic Sea: long-teerm

dynamics, relationships with temperature and eutrophication, and role in the food web. Journal

of Plankton Research, 26(5), 545-561.

Franco, J. M., & Fernández, P. (1993). Separation of PSP toxins by reversed phase high

performance liquid chromatography, with postcolumn reaction and fluorimetric detection.

Chromatographia, 35, 613–620.

Franco, J. M., Fernández, P., & Reguera, B. (1994). Toxin profiles of natural populations and

cultures of Alexandrium minutum Halim from Galician (Spain) coastal waters. Journal of Applied


Frangópulos, M., Guisande, C., Maneiro, I., Riveiro, I., & Franco, J. (2000). Short-term and long-

term effects of the toxic dinoflagellate Alexandrium minutum on the copepod Acartia clausi.

Marine Ecology Progress Series, 203, 161-169.

Frost, B. W. (1972). Effects of size and concentration of food particles on the feeding behaviour

of the marine planktonic copepod Calanus pacificus. Limnology & Oceanography, 40, 476-482.

Fukuda, Y., & Endoh, H. (2006). New details from the complete life cycle of the red-tide

dinoflagellate Noctiluca scintillans (Ehrenberg) McCartney. European Journal of Protistology, 42,

209-219.

Fung, Y. C., & Trott, L. B. (1973). The occurrence of Noctiluca scintillans (Macartey) induced red

tide in Hong Kong. Limnology & Oceanography, 18, 472-476.

Furuya, K., Saito, H., Sriwoon, R., Omura, T., Furio, E.E., Borja, V.M., & Lirdwitayaprasit, T. (2006).

Vegetative growth of Noctiluca scintillans containing the endosymbiont Pedinomonas noctilucae.

African Journal of Marine Science, 28(2), 305-308.

Guisande, C., Frangópulos, M., Carontenuto, Y., Maneiro, I., Riveiro, I., & Vergara, A. R. (2002).

Fate of paralytic shellfish poisoning toxins ingested by the copepod Acartia clausi. Marine

Ecology Progress Series, 240, 105-115.

Hansen, P. J., Miranda, L., & Azanza, R. V. (2004). Green Noctiluca scintillans: a dinoflagellate

with its own green house. Marine Ecology Progress Series, 275, 79-87.

Hanslik, M. (1987). Nahrungsaufnahme und nahrungsverwertung beim meeresleuchttierchen

Noctiluca miliaris. Diss. Univ. Bonn, 1-92.

Hattori, S. (1962). Predatory activity of Noctiluca on anchovy eggs. Bulletin of the Japan Sea

Regional Fisheries Research Laboratory, 9, 211-220.

Hinz, S., Sulkin, S., Strom, S., & Testermann, J. (2001). Discrimination in ingestion of protistan

prey by larval crabs. Marine Ecology Progress Series, 222, 155-162.

146


Huang, C., & Qi, Y. (1997). The abundance cycle and influence factors on red tide phenomena of

Noctiluca scintillans (Dinophyceae) in Dapeng Bay, the South China Sea. Journal of Plankton

Research, 19, 303-318.

Jeong, H. J. (1999). The ecological role of heterotrophic dinoflagellates in marine planktonic

community. Journal of Eukaryotic Microbiology, 46(4), 390-396.

Jeong, H. J., & Latz, M. I. (1994). Growth and grazing rates of the heterotrophic dinoflagellate

Protoperidinium spp. on red tide dinoflagellates. Marine Ecology Progress Series, 106, 173-185.

Jeong, H. J., & Shim, J. H. (1996). The interactions between red-tide dinoflagellate Gymnodinium

sanguineum and microzooplanktonic grazers. In: T., Yasumoto, Y., Oshima & Fukuyo, Y. (Eds.),

Harmful and Toxic Algal Blooms, 7th toxic phytoplankton conference. Sendai, Japan,

Intergovernmental Oceanographic Commission of UNESCO, pp. 377-380.

Kamiyama, T., & Suzuki, T., 2006. Lack of accumulation of paralytic shellfish poisoning (PSP)

toxins in the tintinnid ciliate Favella taraikaensis feeding on the toxic dinoflagellate Alexandrium

tamarense. Limnology & Oceanography, 317, 57-65.

Kimor,, B. (1979). Predation by Noctiluca miliaris Souriray on Acartia tonsa Dana eggs in the

inshore waters of southern California. Limnology & Oceanography, 24, 568-572.

Kiørboe, T. (2003). High turnover rates of copepods fecal pellets due to Noctiluca scintillans

grazing. Marine Ecology Progress Series, 258, 181-188.

Kiørboe, T., & Titelman, J. (1998). Feeding, preys election and prey encounter mechanisms in the

heterotrophic dinoflagellate Noctiluca scintillans. Journal of Plankton Research, 20, 1615-1636.

Kirchner, M., Sahling, G., Uhlig, G., Gunkel, W., & Klings, K. W. (1996). Does the red tide forming

dinoflagellate Noctiluca scintillans feed on bacteria? Sarsia, 81, 45-55.

Kirchner, M., Wichels, A., Seibold, A., Sahling, G., Schütt, C. (2001). New and potentially toxic

isolates from Noctiluca scintillans (Dinoflagellata). In: G. Hallegraeff, (Ed.), Proceedings on

Harmful Algae, 9th International Conference on Harmful Algae Blooms, Tasmania 2000,

Intergovernmental Oceanographic Commission of UNESCO, pp. 379-382.

Latz, M. I., & Jeong, H. J. (1996). Effect of red tide dinoflagellate diet and cannibalism on the

bioluminescence of the heterotrophic dinoflagellates Protoperidinium spp. Marine Ecology

Progress Series, 132, 275-285.

Lee, J. K., & Hirayama, K. (1992a). Effects of salinity, food level and temperatura on the

population growth of Noctiluca scintillans (Macartney). Bull. Fac. Fish. Nagasaki Univ., 71, 163-168.

Lee, J. K., & Hirayama, K., (1992b). The food removal rate by Noctiluca scintillans feeding on

Tetraselmis tetrathelle and Gymnodinium nagasakiense. Bull. Fac. Fish. Nagasaki Univ., 71, 169-175.

147



environments. Marine Microbial Food Webs, 5, 49-58.

Lirdwitayaprasit, T. (2002). Culture of green Noctiluca under laboratory conditions: I. Feeding

behavior and sexual reproduction. Proceedings of the Fifth IOC/WESTPAC International Science

Symposium.

Montani, S., Pithakpol, S., & Tada, K. (1998). Nutrient regeneration in coastal seas by Noctiluca

scintillans, a red tide-causing dinoflagellate. Journal of Marine Biotechnology, 6, 224-228.

Nakamura, Y. (1998a). Growth and grazing of a large heterotrophic dinoflagellate, Noctiluca

scintillans, in laboratory studies. Journal of Plankton Research, 20, 1711-1720.

Nakamura, Y. (1998b). Biomass, feeding and production of Noctiluca scintillans in the Seto Island

Sea, Japan. Journal of Plankton Research, 20, 2213-2222.

Naqvi, S. W., George, M. D., Narvekar, P. V., Jayakumar, D. A., & 9 more (1998). Severe fish

mortality associated with “red tide” observed in the sea off Cochin. Current Science, 75(6), 543-

544.

Olenina, I., Hajdu, S., Andersson, A., Edler, L., Wasmund, N., Busch, S., Göbel, J., Gromisz, S.,

Huseby, S., Huttunen, M., Jaanus, A., Kokkonen, P., Ledaine, I., & Niemkiewicz, E. (2006).

Biovolumes and size-classes of phytoplankton in the Baltic Sea. Baltic Sea Environment

Proceedings No.106, 144 pp.

Okaishi, T., & Nishio, Y. (1976). Identification of ammonia as the toxic principle of red tide of

Noctiluca miliaris. Bulletin of the Plankton Society of Japan, 23, 25-30.

Oshima, Y., Sugino, K., Yasumoto, T. (1989). Latest advances in HPLC analysis of paralytic

shellfish toxins. In: S. Natori, K. Hashimoto & Y. Ueno (Eds.), Mycotoxins and phycotoxins.

Elsevier, Amsterdam, pp. 319–326.

Oshima, Y. (1995). Post-column derivatization HPLC methods for paralytic shell fish poisons. In:

G. M. Hallegraeff, D. M. Anderson, A. D. Cembella, (Eds.), Manual on harmful marine microalgae.

IOC (Intergovtl. Oceanogr. Comm.), Manuals and guides, No. 33, UNESCO, Rome, pp. 81–94.

Parkhill, J. P. & Cembella, A. D. (1999). Effects of salinity, light and inorganic nitrogen on growth

and toxigenicity of the marine dinoflagellate Alexandrium tamarense from northeastern Canada.


Quevedo, M., González-Quiros, R., & Anadon, R. (1999). Evidence of heavy predation by

Noctiluca scintillans on Acartia clausi (Copepoda) eggs off the central Cantabrian coast (NW

Spain). Oceanologica Acta, 22, 127-131.

Schaumann, K., Gerdes, D., & Hesse, K. J. (1988). Hydrographic and biological characteristics of a

Noctiluca scintillans red tide in the German Bight, 1984. Meeresforschung, 32, 71-91.

148


Seibold, A., Wichels, A., & Schütt, C. (2001). Diversity of endocytic bacteria in the dinoflagellate Noctiluca scintillans. Aquatic Microbial Ecology, 25, 229-235.

Shanks, A., & Walters, K. (1996). Feeding by the heterotrophic dinoflagellates (Noctiluca

scintillans) in marine snow. Limnology & Oceanography, 41, 177-181.

Sherr, E. B., & Sherr, B. F. (2007). Heterotrophic dinoflagellates: a significant component of

microzooplankton biomass and major grazers of diatoms in the sea. Marine Ecology Progress

Series, 352, 187-197.

Smayda,T. J. (1997). What is a bloom? A commentary. Limnology & Oceanography, 42(5), 1132-

1136.

Strom, S.L., & Morello, T. A. (1998). Comparative growth rates and yields of ciliates and

heterotrophic dinoflagellates. Journal of Plankton Research, 20(3), 571-584.

Suvapepun, S., (1989). Occurrences of red tides in the Gulf of Thailand. In: T. Okaichi & T.

Nemoto (Eds.), Red tide: Biology, Environmental Science and Toxicology, Elsevier, New York, pp.

41-44.

Sweeney, B. M., (1976). Perinomonas noctilucae (Prasinophyceae), the flagellate symbiotic in

Noctiluca (Dinophyceae) in Southeast Asia. Journal of Phycology, 7, 53-58.

Takayama, H., (1977). Culture of Noctiluca scintillans (Macartney). Bulletin of the Plankton Society

of Japan, 24(2), 83-86.

Tiselius, P., & Kiørboe, T. (1998). Colonization of diatom aggregates by the dinoflagellate

Noctiluca scintillans. Limnology & Oceanography, 43(1), 154-159.

Uhling. G. & Sahling, G. (1990). Long-term studies on Noctiluca scintillans in the German Bight

population dynamics and red tide phenomena 1968-1988. Netherlands Journal of Sea Research,

25(1/2), 101-112.

http://www.google.es/url?sa=t&source=web&cd=8&ved=0CHUQFjAH&url=http%3A%2F%2Fwww.oeaw.ac.at%2Flimno%2Ffiles%2Fpdf%2FMueller_et_al06.pdf&ei=h_CMTp3NBcSu0QW328nUBQ&usg=AFQjCNGTqaD3RkVMB9vx-xjqT0aufSEpKA&sig2=asFa96L0Vx_tQitqJg_XDw

Chapter VI. PST transfer and accumulation in two planktonic grazers

149

Modelling PST transfer and

accumulation in two planktonic

grazers

Abstract

6.1 Introduction

6.2 Methods

6.3 Results

6.4 Discussion

6.5 References

CHAPTER VI

151


Abstract

A model was developed in this chapter in order to study and compare the PST

transfer and accumulation of two different potential PST planktonic vectors, namely the

heterotrophic dinoflagellate in its red form Noctiluca scintillans and the copepod Acartia

clausi, and their role as PST vectors in the planktonic community. Different factors that

influence the toxin transfer such as toxin synthesis, grazing on toxic and non-toxic food and

population size of PST producers and vectors were considered in the model. Moreover, a

laboratory experiment was conducted in order to calculate the detoxification rates of

Noctiluca fed on Alexandrium catenella. According to the model results, the two grazers

showed a significant difference mainly in the timing of the PST accumulation. Noctiluca

exhibited a rapid response to the grazing of Alexandrium with high initial toxin

accumulation, followed by a reduction to zero concentration of toxins in a period of almost

two days. In contrast Acartia showed a considerable delay in comparison to Noctiluca to

accumulate the same amount of toxin in the population. This delay is linked to the slower

reproduction rates that characterise the copepod. The range of the initial values used for

the sensitivity analysis of the model is representative of the coastal environment of a

Galician ria (embayment located at the NW of Iberian Peninsula), where the two grazers

and Alexandrium may co-exist. The model for Acartia showed less sensitivity to these key

parameters probably due to the time delay in accumulation of important amount of toxins.

Both grazers showed a rapid (50 h) reduction of ingested toxin suggesting inefficiency to

transfer toxins through predation in the food web.

152


6.1 Introduction

Paralytic shellfish toxins (PSTs), a class of potent neurotoxin produced among other

microalgae by a number of Alexandrium species and strains when are accumulated via

trophic transfer in the food-wed can cause a neurological affliction to humans known as

Paralytic shellfish poisoning (PSP). The transfer of PSP toxins to higher trophic levels may

involve a variety of planktonic vectors (e.g. mesozooplankton: White 1981; Turner & Tester

1997; Turner et al. 2000a; Hamasaki et al., 2003; Teegarden et al., 2003; Jiang et al., 2007,

tintinids: Turner et al., 2005; heterotrophic dinoflagellates: Sampayo, 1998; Matsuyama et

al., 1999). The role of mesozooplankton and especially of copepods in the toxin dynamics

has received more attention than any other type of planktonic grazer and has become a

focal point of numerous studies (e.g. Robineau et al. 1991; Turriff et al., 1995; Teegarden &

Cembella 1996; Frangópulos et al., 2000). More specific, certain copepod species have

shown in several laboratory experiments and field studies the ability to accumulate toxin by

grazing on PST-producing dinoflagellates (White 1981; Boyer et al., 1985; Turriff et al., 1995;

Teegarden & Cembella 1996; Turner et al. 2000b), serving as potential vectors for PST

transfer to their predators. Toxin accumulation is not the only process observed in

copepods preying on toxic dinoflagellates.

Several authors have demonstrated that copepods may retain, eliminate or even

transform PSP toxins when are fed on Alexandrium (Shimizu, 1978; Guisande et al., 2002;

Hamasaki et al., 2003; Dam & Hasey, 2011). However, PST may also have toxic effects on

copepods, causing mortality (Bagoien et al., 1996), behavioural changes (Sykes & Huntley

1987) or reduction of ingestion/filtering rates and egg hatching (Dutz 1998; Frangópulos et

al., 2000; Barreiro et al., 2006).

In contrast, few studies deal with the grazing interactions between protists (ciliates

and heterotrophic dinoflagellates) and toxic dinoflagellates. Protists as well as

mesozooplankton are common members of the plankton and play an important role in the

marine pelagic systems, as regulators of phytoplankton biomass (Lessard, 1991; Strom &

Morello, 1998; Sherr & Sherr, 2007). Based on observations of natural plankton populations,

several authors showed that ciliates and heterotrophic dinoflagellates (Watras et al., 1985;

Calbet et al., 2003; Duccete et al., 2005; Turner, 2010) can be found at high concentrations

during Alexandrium bloom events suggesting considerable effects on the bloom dynamics.

Certain marine protists have been reported to actively feed on toxic Alexandrium species

(Kamiyama & Suzuki, 2006; Kamiyama et al., 2006; Frangópulos et al., 2011) and grow in

153


relatively good rates at low Alexandrium abundances independently of their cellular PST

quota (Hansen et al., 1992; Kamiyama et al., 2005). However, there are several cases of no

evident ingestion of Alexandrium by protozoan grazers. At high Alexandrium densities,

Hansen (1989, 1992) noted toxic effects on the swimming bahaviour of a ciliate tintinid

followed by swelling and lysis of the cell, which was considered to occur due to PST

exudates. In a later study, Tillman & John (2002) suggested a currently unknown toxic

substance as responsible for the toxic effects of Alexandrium on different protists. In one of

the few studies on the fate of PST in protists previously fed with toxic dinoflagellates,

Kamiyama and Suzuki (2006) showed rapid detoxification in the ciliate Favella taraikaensis.

More recently, Frangópulos et al. (2011) reported that although Noctiluca scintillans was

able to ingest low toxicity A. minutum into its vacuoles, no accumulated toxin was detected

suggesting rapid detoxification/excretion after feeding. These findings show that certain

protist grazers like N. scintillans may be important as a regulator of PST in the planktonic

community by preventing its transfer or accumulation in the food web.

The heterotrophic dinoflagellate N. scintillans and the calanoid copepod Acartia

clausi can dominate (occasionally and for prolonged periods, respectively) some

communities of plankton grazers in several coastal regions of the world (Cataletto et al.,

1995; Elbrächter & Qi, 1998; Fukuda & Endoh, 2006). Nevertheless, Quevedo et al., (1999)

showed that these two potential planktonic PST vectors may co-exist in relatively high

densities and noted out predation by N. scintillans on A. clausi eggs. N. scintillans is the most

studied organism among Noctilucales, an order which contains several aberrant species of

dinoflagellates. Occasionally, this large omnivorous, non-photosynthetic dinoflagellate

forms, mainly by accumulation of buoyant cells, pronounced and striking red tides. In field

studies, predation by N. scintillans has been recorded against toxic dinoflagellate species,

such as Dinophysis caudata and D. acuta (Escalera et al., 2007) and Gymnodinum catenatun

(Alonso-Rodríguez et al., 2005). Even if there are evidences that N. scintillans could play a

significant role in the dynamics of harmful algal events by toxin removal from the web,

there are no available data of detoxification rates. Detoxification process has been

observed also in Acartia clausi. Guisande et al., 2002 demonstrated toxin accumulation and

relatively high detoxification rates of the copepod A. clausi after ingesting A. minutum,

suggesting different fate of toxins which may be significant for higher trophic level

organisms. Using different Alexandrium species as food items, Dutz (1998) and Frangópulos

et al. (2000) showed negative effects on the gross growth-efficiency of A. clausi. The later

authors mentioned reduced egg hatching and naupliar production with increasing toxin

accumulation in the A. clausi.

154


It is generally considered that planktonic grazers can have a central ecological role

in harmful blooms dynamics acting as potential toxin vectors to their predators. Ingestion

of toxic dinoflagellates by different types of planktonic organisms may be important for the

fate of the toxins in the food web. In the present work, our aim is to study and compare

PST transfer and accumulation of two different potential PST planktonic vectors that show

different grazing and reproductive behaviour and their role as PST vectors in the planktonic

community. In order to perform this comparison, a model of toxin accumulation in vector

population was constructed. This model accounts for the influence of different factors in

toxin transfer: toxin synthesis, grazing on toxic and non-toxic food, and population size of

PST producers and vectors. The planktonic vectors selected were the heterotrophic

dinoflagellate N. scintillans and the copepod A. clausi.

6.2 Methods

6.2.1 The model

General description of the model. Dynamic models for Noctiluca scintillans and

Acartia clausi were developed in order to study the link between population dynamics and

PST transfer and accumulation in vectors. Our models describe the changes in phosphorus

quota, Alexandrium and vector population, toxin content in Alexandrium and toxin in vector

population. A simplified schematic diagram of the model is presented in Fig. 6.1. Model

parameterisation was based on published information (Frangópulos et al., 2002; Guisande

et al., 2002; John & Flynn 2002; Barreiro et al 2006; Kiørboe & Titelman 1998) and new data

from laboratory experiments, which are described in detail below. All published and

unpublished data from the authors of this paper were based on experiments carried out in

similar laboratory conditions (aged natural seawater of 33.6 salinity, 12:12 cycle of light:dark

and temperature range of 15-18 ºC). A summary of all the parameters used in this model is

also given in tables 6.1 and 6.2.

155


Fig. 6.1. Conceptual schematic representation of the dynamic models developed for the two different planktonic vectors. Non-continuous lines refer to processes related only to copepod population.

Table 6.1 Model parameters for nutrient uptake, population growth and toxin synthesis dynamics of Alexandrium.

Parameter Description Value Unit Source

µmax Maximum growth rate 0.009 - Barreiro (unpublished)

vmaxp Maximum uptake rate of P 1.55 pg cell-1

h-1

Barreiro (unpublished)

Hsp Half saturation constant for P uptake 0.86 µmol Barreiro (unpublished)

Qmax Maximum intracellular quota of P 3.4 pg cell-1


Qmin Minimum intracellular quota of P 0.028 pg cell-1


Tsr Rate of toxin synthesis 0.0036 gN gC-1

John & Flynn 2002

sP1 Scalar for toxin synthesis 12.3 - John & Flynn 2002

QP Power constant for toxin synthesis 4.2 - John & Flynn 2002

Kts Constant for toxin synthesis 0.054 - John & Flynn 2002

hl Half life of toxins 0.2 % tuned

156


Table 6.2. Model parameters for vector population growth, toxin assimlation and detoxification. NS: Noctiluca scintillans; AC: Acartia clausi.

Model. This model considers that the growth rate and the losses of Alexandrium

due to ingestion by the planktonic vectors are the factors controlling the population of

Alexandrium. Alexandrium growth is based on intracellular phosphorus quota (Q). Thus,

population can be modelled as a function of phosphorus:

(Eq. 1)

The relationship between the growth rate (μ) or division rate and the intracellular

level of phosphorus quota (Q) in the model is been described by the following equation

introduced by Droop (1983):

(Eq. 2)

Parameter Description Vector Value Unit Source

k Detoxification rate NS

AC

-0.17

-0.024 pg toxin ind-1

h-1

Parameterised

Guisande et al.,

2002

tε Toxin assimilation

efficiency

NS

AC

0.15

0.038 pg toxin ind

-1

Experimental

data

Guisande et al.,

2002

Igmax Maximum ingestion

rate

NS

AC

15000

602818 pg C ind-1

h-1

Parameterised

Kig

Half saturation

constant for

ingestion rate

NS

AC

240000

2026018 pg C ml-1

Parameterised

m Natural mortality

NS

AC

0.001

0.06

Ind h-1

%

Experimental

data tuned

Frangópulos

2002

Epmax Maximum egg

production rate AC

1.543 Egg ind-1

h-1

Parameterised

KEp

Half saturation

constant for egg

production

AC 12279 pg C ml-1

Parameterised

ε Carbon assimilation

efficiency NS

AC

65498 ind pg-1

Kiørboe &

Titelman 1998

ω

Effect of toxins in

hatching success AC -0.14 - Parameterised

157


where, μmax is the maximum μ and Qmin (pg cell-1) the subsistence intracellular quota of

phosphorus below which there is no further cell division.

Internal nutrient concentration expressed here as cell quota is given by the balance

of uptake (vp) minus the phosphorus consumed through growth of Alexandrium population

and toxin synthesis:

(Eq. 3)

A modified Michaelis-Menten equation (Eq. 3) was used to simulate Alexandrium

uptake of P. It assumes an increase with external phosphorus inputs (P) to a level that

approaches a maximum uptake rate (vmax) and a decrease with intracellular phosphorus

quota (Q) according to Morel (1987).

(Eq. 4)

The constants appeared in Eq. 4 were estimated using experimental unpublished

data (Table 6.1).

Toxin dynamics of Alexandrium was modeled as the function of toxin synthesis and

degradation:

(Eq. 5)

The degradation rate was estimated at 0.2 %. This model is assuming no limitation of

toxin synthesis by nitrogen. Nitrogen is needed for toxin synthesis, but it is usually coming

from recycling organic nitrogen inside the cell, rather than direct uptake (Anderson et al.,

1990, John & Flynn 2000). Phosphorus stress favours toxin production in terms that low

levels of P associate with high toxin synthesis rates. The equation for toxin synthesis by

Alexandrium was taken from John & Flynn (2000):

(Eq. 6)

158


The parameters introduced in this equation are the rate of toxin synthesis (Tsr), a

scalar for toxin synthesis (sP1) and a constant for toxin synthesis (Kts).

Our model describes the dynamics of Noctiluca population in terms of growth rate

related to the ingestion of toxic and no toxic phytoplankton and mortality (m1):

(Eq. 7)

Mortality (m1) represents natural mortality caused by other factors than toxins. It

was tuned to 0.001 ind h-1 based on experimental data on A. catenella described in the

section 2.3 in this Chapter. Simulated ingestion by Noctiluca was modelled according to Eq.

8 based on assumption that the basic factor that determines its population is the biomass

of phytoplankton. The model considers two food sources, Alexandrium population (CA) and

non-toxic food (phytoplankton and detritus) (CP). There is no evidence of food preference

by Noctliluca, suggesting a voracious non-selective feed (Uhlig & Sahling, 1990). Thus, we

assumed in the model that feeding is unselective. We also consider that Noctiluca is not

affected at all by toxins. The available information on the red Noctiluca (Frangopulos et al.,

2011; Chapter V, this study) propose no inhibitory effects of toxic Alexandrium on the

growth and feeding activities of Noctiluca. The ingestion rate is derived from the calculation

of the maximum ingestion rate (Igmax= 15000 pg C ind-1h-1) and half saturation constant (KIg=

240000 pg C ml-1). The results of the experiment described in Frangopulos et al. (2011) were

fitted to the function:

(Eq. 8)

The growth rate (μ1) of Noctiluca follows the responses of the ingestion rate

(IgNOC):

(Eq. 9)

The carbon assimilation efficiency (ε) of Noctiluca was fixed at 65498 ind pg-1

(Kiørboe & Titelman, 1998). The latter authors mention that the size and quality of the food

item may also affect the growth rates of the Noctiluca.

159


The dynamic of PST in the Noctiluca population (Pn) was calculated as follows:

(Eq. 10)

Toxin kinetic in Noctiluca involves a balance between toxin assimilation and

detoxification rate:

(Eq. 11)

Toxin assimilation by Noctiluca is related to the toxin assimilation efficiency (tε), the

ingestion rate and the toxin content of Alexandrium:

(Eq. 12)

where, tε was calculated from experimental data at 0.15 pg toxin ind-1. The second term

represents the ingestion rate of Alexandrium by Noctiluca (Eq. 8), while the last one the

toxin kinetics of Alexandrium (Eq. 5).

The detoxification kinetic rate by Noctiluca was expressed as an exponential decay

rate defined by the equation:

(Eq. 13)

where Ct is toxin concentration in Noctiluca (pg ind.-1) and dt is the difference between t and

t-1. The hourly decay rate of PST in the red N. scintillans, which was assumed to be constant

and time-dependent, was calculated at 0.17 pg toxin ind-1 by experimental data.

The model for the study of copepods as PST vector is assuming no affect by the

toxins on the copepods’ ingestion, egg production rates and mortality, because of dilution

effects by non-toxic food (Barreiro et al., 2006). Two alternative food sources (Alexandrium

population and non-toxic food) of equal selectivity by Acartia were assumed. The ingestion

rate of Acartia can be described by the Eq. 8, where maximum ingestion rate (Igmax) and the

half saturation constant (Kig) can be replaced by the parameterised values 602818 pg C ind-1

and 2026018 pg C ml-1.

160


In this model egg production rates (eggs female-1 ind-1) were formulated as a factor

of ingestion following the type II function:

(Eq. 14)

where, Epmax (1.543 Egg female-1 ind-1) is the maximum egg production rate and Kep the half

saturation constant (Table 6.2).

On the other hand, we assume that copepod egg hatching success is affected by

toxins because they act in a dose dependent manner (Barreiro et al., 2006). A linear

regression was selected to formulate the experimental data in order to model the egg

hatching of Acartia. The effect of toxins in hatching success (ω) was calculated at -0.14

(Table 6.2):

(Eq. 15)

The growth rate of Acartia (μ2) was derived from the knowledge of the Ep and Eh:

(Eq. 16)

A time interval of 48 hours was assumed for egg hatching and was introduced to

the calculation of the growth rate. In our model the dynamics of the copepod population

( ) involve a balance between growth rate (μ2) and mortality (m2). Acartia mortality

represents natural mortality and has been calculated in laboratory experiments

(Frangopulos et al., 2002). Here, it is expressed as a percentage of the copepod population

(Table 6.2). Thus, it was calculated by:

(Eq. 17)

Daily PST assimilation efficiency (tε) and detoxification rate (k) by copepods have

been estimated by Guisande et al. (2002). The latter authors mention that detoxification in

copepods includes toxin output in eggs and faecal pellets and toxins eliminated as

161


dissolved form. Toxin kinetic in copepod population was modelled based on the Eq. 11, 12

and 13.

Model implementation and sensitivity analysis. The models were implemented by

the system dynamic software Ventana VENSIM (Vensim PLE Plus Version 5.9). The model

simulation time lasted 200 h, which corresponds to an approximate time for Alexandrium

natural population bloom with a time step of 1 hour. For all the simulations we used the

Euler integration method. A Monte Carlo sensitivity analysis of the models was conducted

to reveal the effect of changes of the stock variables (the external phosphorus sources, the

initial abundance of Alexandrium, the abundance of non-toxic food and the initial

population density of vectors). For each of these variables, a random normal distribution

was parameterised based on in-situ data from the Galician rias (NW Spain) and the adjacent

area (Spyrakos et al., 2011; González Vilas et al., 2011; Valdes et al., 1990; Pazos et al., 2007,

2008; Spyrakos, unpublished data). Table 6.3 shows details on the values used for each

variable in the sensitivity analysis runs. A single sensitivity analysis was run independently

for each of the variables presented. In addition a multivariate sensitivity analysis was run

testing all five variables together. The simulations were run 3 times. The method of Sobol'

(Sobol', 1990) was applied as implemented by the sensitivity (Pujol, 2007) library in R

(version 2.9.1, R development CoreTeam). This method can handle nonlinear and non-

monotonic functions and models and it is based on variance decomposition. The result of

this method are expressed as a sensitivity index (SI), which is used to quantify the amount

of variance that selected parameters contribute to the overall incertainty of the model

output.

Table 6.3. Model parameters tested for model output sensitivity.

Parameter Range Mean±SD

Phosphorus (μmole) 0-9 0.92±2.02

Alexandrium (cells ml-1

) 150-29840 5012±66

non-toxic food (particles ml-1

) 750-156246 39258±349

Noctiluca (ind ml-1

) 0.1-1.8 0.8±0.6

Acartia (ind ml-1

) 0.0003-0.001 0.0007±0.0009

6.2.3. Laboratory experiments (PST effect on Noctiluca scintillans)

Algal species. The non-axenic dinoflagellate strains of Alexandrium catenella

(ACC06) and Prorocentrum micans (PM1V) were used in this study. A. catenella was isolated

from fjords in Southern Chile while P. micans was isolated from the Galician rias, and both

162


came from long-established populations cultured in the Instituto Español de Oceanografía

(Vigo). N. scintillans was fed algae at the exponential phase, because toxin composition is

constant in exponential growth-phase cultures (Franco et al., 1994; Parkhill & Cembella

1999). Details for the collection and culture establishment of N. scintillans specimens are

given in Chapter V. Cell carbon and nitrogen content were determined on the first day of

the experiment, from five replicate subsamples filtered on pre-combusted GF/F filters at

low pressure, dried at 70°C and combusted in a Fisons EA-1108 CHN analyzer. Sulfanilamide

was used as the standard. Table 6.4 shows the carbon and nitrogen content of A. catenella

and P. micans.

Table 6.4. Cell parameters for Alexandrium catenella and Prorocentrum micans measured during the experiment. Carbon and nitrogen content, C: N ratio (means ± S.D., n = number of determinations).

Cell parameter n Alexandrium catenella Prorocentrum micans

Carbon (pg cell-1

) 5 1492.5±118.4 3833.9±509.0

Nitrogen (pg cell-1

) 5 276.2±26.4 611.3±87.7

C:N 5 5.4±0.1 6.3±0.1

Experimental design. Noctiluca cells were starved for 72 hours in 250 ml beakers

containing filtered seawater in 12:12 h light:dark cycle. Fifty individuals in the trophon stage

were then pipetted out into each of fifteen/four 40 ml beakers which contained 400 cells

ml-1 A. catenella/P. micans, respectively. A previous study (Chapter V) showed that the

ingestion rate of the red N. scintillans fed with A. minutum in 12:12 h light:dark cycle

conditions saturates at a concentration of 370 cells ml-1. Only Noctiluca individuals with

active tentacle movement were used. Each day, N. scintillans cells were isolated,

enumerated and pipetted to a new beaker with A. catenella/P. micans suspensions at the

experimental concentration. On the last day of the experiment an average of 43 individuals

were transferred in replicate samples to filtered seawater and then starved for 0, 3, 6, 9

and12 hours. The toxin content retained in N. scintillans preyed upon A. catenella was

qualified through HPLC. In parallel, the toxicity of N. scintillans fed on the dinoflagellate P.

micans was measured as a control.

Grazing estimation. The replicates were used to estimate the ingestion rates of the

red N. scintillans fed on A. catenella and P. micans at the last day of the experiment. Initial

beakers without N. scintillans were prepared simultaneously. Initial and subsequent,

subsamples were immediately preserved using 4% formaldehyde for cell counting. Grazing

experiments were run for 24 h at the temperature and light conditions described above.

The mortality of N. scintillans was checked every day after 24 h incubation using a

163


stereomicroscope. A. catenella and P. micans subsamples were preserved with 4%

formaldehyde for cell counting. Abundances of A. catenella and P. micans were determined

by quadruplicate counting of 1 ml of sample with a Sedgewick-Rafter chamber using an

inverted microscope. Frost’s (1972) equations were used to calculate clearance and

ingestion rates.

Toxin analysis is described in the section 2.5 in Chapter V.

6.3 Results

6.3.1 Laboratory experiments

Table 6.5 shows the specific toxic composition, total toxin per cell and total cell

toxicity of A. catenella during the experiment. In the stock culture of A. catenella the mean

cellular toxin content varied between 17.19 and 23.31 fmol cell-1. No C toxins were detected.

The mean survival rate of red N.scintillans fed on A. catenella was 65% while in the N.

scintillans population fed on P. micans was 85%. Ingestion rates were calculated for the last

day of the experiment and showed a mean value of 0.07 g C ind-1 day-1 for the red N.

scintillans fed on A. catenella. In parallel, ingestion rates of the red N. scintillans using as

food item P. micans were negative (-0.04 g C ind-1 day-1).

Table 6.5. Alexandrium catenella. Specific toxin composition of gonyautoxins, as a percentage (GTX1-4, mean ± S.D., fmol cell-1), total toxin per cell (combined GTX1, GTX2, GTX3, GTX4, mean ± S.D., fmol cell-1), and total cell toxicity (mean ± S.D., fg STXeq cell-1) of each one during the experimental period. (STXeq = saxitoxin equivalents).

The toxin concentrations found here are higher than the ones measured in the

previous experiment (Chapter V) using A. minutum and permitted us to detect retained

toxin Noctiluca individuals. The retained toxin in N. scintillans was gradually decreased with

increasing the starvation time till a concentration of 12 fmol ind-1 at Hour: 12 (Fig. 6.2).

Day GTX1 GTX2 GTX3 GTX4 Total toxin Cell toxicity

1 5.89 ± 1.04 4.24 ± 1.08 36.79 ± 0.75 53.07 ± 2.23 17.19 ± 1.83 6813 ± 742

2 4.89 ± 0.14 3.51 ± 0.24 37.12 ± 0.92 54.48 ± 1.07 23.31 ± 1.59 9235 ± 644

3 4.26 ± 0.22 3.83 ± 0.34 40.75 ± 0.21 51.17 ± 0.26 20.18 ± 5.94 7926 ± 2350

4 4.28 ± 0.39 4.24 ± 0.28 43.00 ± 0.53 48.48 ± 0.50 22.05 ± 2.67 8614 ± 1044

164


Fig. 6.2. Toxicity retained in Noctiluca scintillans fed with Alexandrium catenella at the end of the experiment as function of the elapsed starvation time. For each cycle the number of observations was 50.

6.3.1 Model

The evolution of the simulated PST accumulated in Noctiluca scintillans and Acartia

clausi populations with respect to the randomly varied parameters presented in the Table

6.3 is depicted in the Figures 6.3 and 6.4. Note that the graphs do have 2 different scales

depending on the data.

In general, the mean concentration of PST accumulated in N. scintillans continuously

increases with the start of the simulation till a maximum value shortly after, indicating an

immediate PST accumulation in response to the toxic Alexandium grazing. In all the graphs

dynamics of toxin accumulation in the population are reaching fast the maximum value in

around 10 hours from the start of the simulation. This maximum in the PST in the Noctiluca

is followed by a rapid drop to zero values where it remained throughout the rest of the

simulation period. In average, no remained toxin was modelled in the population after 50

hours. Nevertheless, with modified non-toxic food concentration and initial Noctiluca

abundance the simulation results above the 75th percentile are displaced to the right

indicating a showing longer detoxification times (Fig. 6.3 c, d, e). Lowest concentrations of

non-toxic food and lowest initial abundances of Noctiluca are likely the reasons for these

longer detoxification times observed. The model output is tangibly sensitive to initial

Alexandrium abundance (Table 6.6). The highest mean simulated value (0.25-0.3) of toxin in

165


Noctiluca among the random normal distributed parameters tested was observed with

initial Alexandrium abundance. In all other cases the 100th percentile line is always below

0.08. A second peak a shortly time after the first one was appeared between the

percentiles 75 and 100 in the sensitivity analysis results from the 200 simulations with

randomly modified initial Noctiluca concentration (Fig. 6.3 d). In the same graph the 25th

percentile is displaced to the right indicating a time delay to the accumulation of toxin in

Noctiluca population.

In contrast the model for Acartia shows a considerable delay in PST accumulation in

the population (Fig. 6.4). Accumulation of similar amounts of toxins in the population takes

longer in copepods than in Noctiluca because of a delay caused by copepod reproduction

(around 150 hours). The graphs of the sensitivity analysis illustrate a very slight increase of

the simulated toxin in Acartia population after 4 days from the simulation start followed by

a sharp increase towards the end. The appearance of these 2 peaks is due to dynamics of

copepod reproduction. The influence of initial Noctiluca abundance to the model reveals a

more clear form of the 2 peaks. The model seems to be substantially sensitive to non-toxic

food concentration and phosphorus availability (Table 6.4). At the end of this simulation the

toxin in Acartia has been significantly decayed. Dynamics of toxin accumulation in Noctiluca

are more deeply influenced by the variation in the selected variables. They show a lower

degree of variation than Acartia (Table 6.6).

Table 6.6. Coefficient of variation and standard deviation (CV±S.D) from the sensitivity analysis results for NS and AC model with randomly modified initial values of the parameters shown in Table 6.3.

Parameter NS AC

Coefficient of variation

Phosphorus 0.63±0.69 0.63±0.69

Alexandrium 2.93±2.02 0.33±0.21

non-toxic food 0.41±0.13 0.68±0.09

planktonic grazer 1.76±0.89 0.26±0.07

all the above 7.58±4.41 1.26±0.60

166


Fig. 6.3. Sensitivity analysis results showing the mean and percentiles of simulated PST accumulated in Noctiluca scintillans population from: 200 simulations with randomly modified initial values of a) phosphorus concentration, b) non-toxic food concentration, c) Alexandrium abundance and d) Noctiluca abundance and 800 simulations using all the above parameters. The ranges of the parameters are provided in Table 6.3.

167


Fig. 6.4. Sensitivity analysis results showing the mean and percentiles of simulated PST accumulated in Acartia clausi population from: 200 simulations with randomly modified initial values of a) phosphorus concentration, b) non-toxic food concentration, c) Alexandrium abundance and d) Acartia abundance and 800 simulations using all the above parameters. The ranges of the parameters are provided in Table 6.3.

168


For the Noctiluca model Sobol' analysis showed that 4 parameters out of 8 were

responsible for the 80% of the output variability (Fig. 6.5). Table 6.7 shows the rank and the

sensitivity index when the Sobol' analysis was repeated only for the parameters listed in

Table 6.6. The model output was mainly influenced by initial Alexandrium concentration.

The Sobol' analysis for the Acartia model indicated that initial non-toxic food concentration,

initial Alexandrium abundance and phosphorus concentration acounted 66% of all the SI

values (Fig. 6.6). In this case, the initial concentration of non-toxic food was the parameter

that contributed the most to the overall incertainty of the model output.

Fig. 6.5. Contribution of each parameter to the variability of Noctiluca model output as indicated by Sobol' index (SI).

Fig. 6.6. Contribution of each parameter to the variability of Acartia model output as indicated by Sobol' index (SI).

169


Table 6.7. First order sensitivity indices showing the contribution and ranking of parameters to the overall incertainty of the models output.

Planktonic grazer Parameter Sobol´ rank SI

P 3 0.19

non-toxic 2 0.22

NS Alex 1 0.25

Noc 4 0.11

P 3 0.19

AC non-toxic 1 0.23

Alex 2 0.21

Ac 4 0.02

In order to analyse the paired relationships among avariables within our model, we

plotted all combinations of pairs of “stock” variables (except non-toxic food, which in not

dynamic) for Noctiluca (Fig. 6.7) and Acartia (Fig. 6.8). Averaged values of the 800

simulations from the multivariate sensitivity analysis were used for these scatter plots.

Fig. 6.7 graph indicates that high phosphorus concentrations coincide with high

Alexandrium abundances, since both parameters decrease with the time. Alexandrium

population is monotically decreasing due to high grazing and Noctiluca population growth

and phosphorus because of uptake from Alexandrium. The toxin in Alexandrium population

is closely related to intracellular phosphorus quota, while in Noctiluca population the

relationship is more complicated showing a positive response to Alexandrium population,

phosporus and intracellular phosphorus. An exception in this responce appears in the first

steps of the simulation, where grazing of Noctiluca on the toxic dinoflagellates and toxin

accumulation are fast. The relationship between Noctiluca population and toxin in

Alexandrium can be characterised as mostly negative with an increase at low-intermediate

abundances of Noctiluca. The same response is observed between toxin in Noctiluca and

toxin in Alexandrium. However, an increase of accumulated toxin in Noctiluca is shown

following the increasing toxicity in Alexandrium at intermediate values of Alexandrium

toxicity. This occurs at the beginning of the simulation when intracellular phosphorus

decreases, and therefore toxin in Alexandrium increases, while Noctiluca population still

grows and accumulates toxin (because Alexandrium is still at high densities).

On the other side, the delay observed in the growth of Acartia is reflected in the

plots of Alexandrium abundance against phosphorus, intracellular phosphorus

concentration and Acartia abundance. Alexandrium abundance increases rapibly in relation

to these parameters during the first steps of the model and then decreases due to high

grazing and Aartia population growth. Toxin in Alexandrium shows a negative relationship

with phosphorus and intracellular phosphorus and an almost vertical increase with Acartia

170


when the population of this planktonic grazer grows. The latter pattern is detected

between toxin in Acartia and toxin in Alexandrium.

Fig. 6.7. Scatter plot matrix of all pair combinations of “stock”dynamic variables for Noctiluca. Note that all units in the graph refer to log-transformed values.

Fig. 6.8. Scatter plot matrix of all pair combinations of “stock”dynamic variables for Acartia. Note that all units in the graph refer to log-transformed values.

171


6.4 Discussion

In the laboratory experiment mortality of Noctiluca with both dinoflagellates used

as food items was considered as the effect of manipulation since the individuals were

isolated and pipetted to new beakers every day during the experiment. The relatively low

ingestion rates of N. scintillans on A. catenella and P. micans measured at the last day of the

experiment probably reflects food-repletion in Noctiluca. This was confirmed by

microscopic observations.

Referring to our model now, it is noteworthing mentioning that for all the key

variables used in the model the input values represent distributions from sea-truth

measurements in the same coastal environment (embayment of ria de Vigo). Apparently,

our model conceptually simplifies certain aspects of the complex nature. Among the

simplifications, Noctiluca and Acartia are considered to grow without predation limiting

their population. Even if this assumption is unrealistic at the same time is necessary in order

to approximate the amount of secondary production that can accumulate toxins, which is

the aim of this study. In order to develop the model we also assumed that there is no

competition for phosphorus, proposing Alexandrium as the only consumer of this nutrient.

Nevertheless, this assumption does not affect the toxin dynamics and is necessary to

establish an increasing Alexandrium population in our model. Moreover, we considered

phosphorus and not nitrogen limitation in the growth of Alexandrium, since phosphorus is

involved in the toxin production dynamics. In a different case the results of the model

would be similar but without the effect of phosporus stress, which is more interesting in

our case. The non-toxic food was added to the model to provide an alternative to

Alexandrium source of food to the vectors and it was assumed not to vary in order to make

it always available to them. We assume non-toxic food is always available above limiting

levels, without taking care of its own dynamics that would unnecessarily complicate our

modeling. The low abundance of Alexandrium and its relatively low toxicity along with the

use of non-toxic food in the model permits us to take for granted that the toxic effects of

Alexandrium to the copepod are limited to the egg hatching success (Barreiro et al., 2006).

Even so, taking into account the short-term dynamics which characterises the toxin

accumulation and the fast detoxification (50 h in the majority of the cases) in Noctiluca,

shown after the model simulation, we could conclude that toxin transfer through predation

on this vector might not be important. As a PST vector N. scintillans seems to be of less

significance. This result might be pertinent to other protozoan grazers which like Noctiluca

have to be non-selective feeders and not affected by the toxins. For example, the ciliate

172


Favella taraikaensis has shown actively grazing on A. tamarense in laboratory conditions

without toxic effects (Kamiyama et al., 2005).

On the other hand, the copepod dynamic model results show that population

requires more time to accumulate comparable to Noctiluca amounts of toxin. Because of

this delay, high predation on copepods during an Alexandrium toxic event, would probaply

result less or even no toxin accumulation in these organisms. Simulated accumulation and

detoxification in Acartia appear to be comparable to those of Noctiluca. Total detoxification

from the beginning of accumulation peak lasts 50 h, resticting in that way the availability of

this vector for toxin transfer to higher trophic levels. This could be of great importance in

the area of Galician coast, where Acartia is the most abundant mesozooplankton in all

seasons (Valdes et al., 1990). The results of the model developed here could be possibly

generalized for other calanoid copepods found in the area and other coastal environments,

which are also characterised by omnivorous feeding, similar size and high egg production.

In any case, variations in grazing rates and preferences and toxic effects must be

considered.

The slower reproduction rates in copepods have as effect the delayed (100-150 h)

noticeable increase in toxin accumulation in comparison to Noctiluca and thus the different

behaviour of these two planktonic grazers as PST vectors. This finding is considerable

because it shows that in the case of co-existence of these two planktonic grazers

(confirmed by unpublished data for the summer period in the Galician rias and recorded in

other regions: e. g. Quevedo et al., 1999). The presence of two abundant organisms of the

plankton community with high accumulated amount of toxin in marine environment at

different times from the start of the toxic event might increases the risk for toxin transfer

through predation.

The range of the input data for the model parameters controls whether or not a

factor will have a strong influence in the dynamics of the toxin accumulation. Using a widely

changed range of input values, as for example lower phosphorus concentrations or higher

amount of toxin in Alexandrium, which can be found in another coastal ecosystem the

obtained model results would be different. An important feature of the toxin dynamics for

the heterotrophic dinoflagellate is that the key variable explaining the total amount of toxin

that could be accumulated is the initial abundance of Alexandrium (cv=2.93±2.02, CI=0.25),

as it was expected. The time delay observed in the simulated PST toxin accumulated in a

copepod population probably conceals the variation of the initial concentration of this toxic

dinoflagellate. Generally, this seems to be the case for the lower sensibility that the model

for Acartia showed for the tested parameters. Nevertheless, the non-important effect of

173


initial Acartia abundance on toxin dynamics is possibly linked to the relatively low variation

of values included in the sensitivity analysis that correspond to the abundances found in the

Galician coast. Noctiluca, on the other side, shows a wider distribution in the initial values

used in the model, ranging from total absent to bloom-forming abundances. In natural

assemblages, Noctiluca blooms appear to create pink “patches” and slicks with significant

deviation of the abundance (Huang & Qi, 1997). Moreover, in their review on heterotrophic

dinoflagellates Jeong et al. (2010) mention that protozoan grazers are typically 100 to

10000 timess more abundant than copepods. For the copepod studied in this paper, the

initial concentration of the non-toxic food seems to be an important parameter involved in

the toxin dynamics (cv=0.68±0.09, SI=0.23). The presence of alternative to the toxic has

been demonstrated to effect by dilution (Barreiro et al., 2006).

Conclusively, the model predicts very fast detoxification in both the planktonic

organisms suggesting that these vectors of PST are inefficient. However, simulated toxin

accumulation in the vetors reached at certain moments comparatively high amounts.

Several explanations have been proposed for this reduction of the ingested toxin in

copepods such as excresion (Guisande et al., 2002), metabolic degradation or regurgitation

(Teegarden et al., 2003). Detoxification or excretion processes have been also suspected in

Noctiluca (Frangópulos et al., 2011) and ciliates (Kamiyama & Suzuki, 2006). The latter

authors mention potential negative effects in several organisms due to concentrated

dissolved toxin.


174


6.5 References

Alonso-Rodríguez, R. Ochoa, J. L., Uribe-Alcocer, M. Blanco A. G., & Butardo, V. M. (2005).

Grazing of heterotrophic dinoflagellate Noctiluca scintillans (Macartney) Kofoid on Gymnodinium

catenatum Graham. Revista Latinoamericana de Microbiología, 47(1–2), 6–10.

Anderson, D. M., Kulis, D. M., Sullivan, J. J., Hall, S., & Lee, C. (1990) Dynamics and physiology of

saxitoxin production by the dinoflagellates Alexandrium spp. Marine Biology, 104, 511–524.

Barreiro, A., Guisande, C., Frangópulos, M., Gónzalez- Fernández. A., Muñoz, S., Pérez, D.,

Magadán, S., Maneiro, I., Riveiro, I., & Iglesias, P. (2006) Feeding strategies of the copepod

Acartia clausi on single and mixed diet of toxic and non-toxic strains of the dinoflagellate

Alexandrium minutum. Marine Ecology Progress Series, 316, 115-125.

Bagoien, E., Miranda, A., Reguera, B., & Franco J. M. (1996). Effects of two paralytic shellfish

toxin producing dinoflagellates on the pelagic hapacticoid copepod Euterpina acutifrons. Marine

Biology, 126, 361–369.

Boyer, G. L., Sullivan, J. J., LeBlanc, M., & Andersen, R. J. (1985). Assimilation of PSP toxins by

the copepod Tigriopus californicus from dietary Protogonyaulax catanella. In: D. M. Anderson, A.

W. White & D. G. Baden (Eds). Toxic Dinoflagellates. (pp 407-412). New York: Elsevier.

Calbet, A. , Vaqué, D., Felipe, J., Vila, M., Sala, M. M., Alcaraz, M., & Estrada, M. (2003). Relative

grazing impact of microzooplankton and mesozooplankton on a bloom of the toxic

dinoflagellate Alexandrium minutum. 259, 303-309.

Carreto, J., Carignan, M., & Montoya, N. (2001). Comparative studies on mycosporine-like amino

acids, paralytic shellfish toxins and pigment profiles of the toxic dinoflagellates Alexandrium

tamarense, A. catenella and A. minutum. Marine Ecology Progress Series, 223, 49-60.

Cataletto, B., Feoli, E., Fonda Umani, S., & Cheng-Yong S. (1995). Eleven years of time-series

analysis on the net-zooplankton community in the Gulf of Trieste. Journal of Marine Science, 52,

669–678.

Dam, H., G. & Haley S., T. (2011). Comparative dynamics of paralytic shellfish toxins (PST) in a

tolerant and susceptible population of the copepod Acartia hudsonica. Harmful Algae, 10, 245-

253.

Droop, M. R. (1983). 25 years of algal growth kinetics. A personal view. Botanica Marina, 26, 99-

112.

Doucette, G. J., Turner, J. T., Powell, C. L., Keafer, B. A., & Anderson, D. M. (2005). ECOHAB-Gulf

of Maine. Trophic accumulation of PSP toxins in zooplankton during Alexandrium blooms in

Casco Bay, Gulf of Maine, April–June, 1998. I. Toxin levels in Alexandrium and zooplankton size

fractions. Deep-Sea Research Part II, 52 (19-21), 2764–2783.

http://www.sciencedirect.com/science/journal/15689883

175


Dutz, J. (1998). Repression of fecundity in the neritic copepod Acartia clausi exposed to the

toxic dinoflagellate Alexandrium lusitanicum: relationship between feeding and egg production.


Elbrächter, M. & Qi, Y. Z. (1998). Aspects of Noctiluca (Dinophyceae) populations dynamics, In:

D. M. Anderson, A. D. Cembella & G. M. Hallegraeff (Eds). Physiological Ecology of Harmful Algal

Blooms. NATO ASI Series. (pp 315-335). Berlin: Springer-Verlag.

Escalera, L., Pazos, Y., Moroño A., & Reguera, B. (2007). Noctiluca scintillans may act as a vector

of toxigenic microalgae. Harmful Algae, 6, 317–320.

Franco, J. M., Fernández, P., & Reguera, B. (1994). Toxin profiles of natural populations and

cultures of Alexandrium minutum Halim from Galician (Spain) coastal waters. Journal of Applied


Frangópulos, M., Guisande, C., Maneiro, I., Riveiro, I., & Franco, J. (2000). Short-term and long-

term effects of the toxic dinoflagellate Alexandrium minutum on the copepod Acartia clausi.


Frangópulos, M., Spyrakos, E., & Guisande, C. (2011). Ingestion and clearance rates of the red

Noctiluca scintillans fed on the toxic dinoflagellate Alexandrium minutum (Halim), Harmful Algae,

10, 304–309.

Frost, B. W. (1972). Effects of size and concentration of food particles on the feeding behaviour

of the marine planktonic copepod Calanus pacificus. Limnology & Oceanography, 40, 476-482.

Fukuda, Y. &Endoh, H. (2006). New details from the complete life cycle of the red-tide

dinoflagellate Noctiluca scintillans (Ehrenberg) Macartney. European Journal of Protistology, 42,

209–219.

González Vilas, L., Spyrakos, E., & Torres Palenzuela, J. (2011). Neural network estimation of


Remote Sensing of Environment 115 (2), 524-535.

Guisande, C., Frangópulos, M., Carotenuto, Y., Maneiro, I. Riveiro I., & Vergara, A. R. (2002). Fate

of paralytic shellfish poisoning toxins ingested by the copepod Acartia clausi. Marine Ecology

Progress Series, 240, 105–115.

Hamasaki, K., Takahashi, T., & Uye, S. (2003). Accumulation of paralytic shellfish poisoning

toxins in planktonic copepods during a bloom of the toxic dinoflagellate Alexandrium tamarense

in Hiroshima Bay, western Japan. Marine Biology, 143, 981–988.

Hansen, P. J. (1989). The red tide dinoflagellate Alexandrium tamarense: effects on behaviour

and growth of a tintinnid ciliate. Marine Ecology Progress Series, 53, 105-116.

176


Hansen, P. J., Cembella, A. D., & Moestrup, Ø. (1992). The marine dinoflagellate Alexandrium

ostenfeldii: paralytic shellfish toxin concentration, composition, and toxicity to a tintinnid ciliate.

Journal of Phycology, 28, 597–603.

Huang, C., & Qi, Y. (1997). The abundance cycle and influence factors on red tide phenomena of

Noctiluca scintillans (Dinophyceae) in Dapeng Bay, the South China Sea. Journal of Plankton

Research, 19, 303-318.

Jiang, T. H., Wang, D. Z., Niu, T., & Xu, Y. X. (2007). Trophic transfer of paralytic shellfish toxins

from the cladoceran (Moina mongolica) to larvae of the fish (Sciaenops ocellatus). Toxicon, 50(5),

639-645.

Kamiyama, T., & Suzuki, T. (2006). Lack of accumulation of paralytic shellfish poisoning (PSP)

toxins in the tintinnid ciliate Favella taraikaensis feeding on the toxic dinoflagellate Alexandrium

tamarense. Marine Ecology Progress Series, 317, 57-65.

Kamiyama, T., Suzuki, T., & Okumura, Y. (2006). Feeding of the tintinnid ciliate Favella

taraikaensis on the toxic dinoflagellate Alexandrium tamarense and the fate of prey toxins.

African Journal of Marine Science, 28(2), 343-346.

Kamiyama, T., Tsujino, M., Matsuyama, Y., & Uchida, T. (2005). Growth and grazing rates of the

tintinnid ciliate Favella taraikaensis on the toxic dinoflagellate Alexandrium tamarense. Marine

Biology, 147, 989–997.

Kim, C.-J. Kim, C.-H., & Sako, Y. (2005). Paralytic shellfish poisoning toxin analysis of the genus

Alexandrium (Dinophyceae) occurring in Korean coastal waters. Fisheries Science, 71, 1-11.

Kiørboe, T., & Titelman, J. (1998). Feeding, preys election and prey encounter mechanisms in the

heterotrophic dinoflagellate Noctiluca scintillans. Journal of Plankton Research, 20, 1615-1636.


environments. Marine Microbial Food Webs, 5, 49–58.

Matsuyama, Y., Miyamoto, M., & Kotani, Y. (1999). Grazing impact of the heterotrophic

dinoflagellate Polykrikos kofoidii on a bloom of Gymnodinium catenatum. Aquatic Microbial

Ecology, 17, 91–98.

Morel, F. M. (1987). Kinetics of nutrient uptake and growth in phytoplankton. Journal of

Phycology, 23, 137-150.

Parkhill, J. P., & Cembella, A. D. (1999). Effects of salinity, light and inorganic nitrogen on growth

and toxigenicity of the marine dinoflagellate Alexandrium tamarense from northeastern Canada.


Pazos, Y., Moroño, A., & Doval, M. (2007). Anuario oceoanográfico de Galicia 2007. X. Romarís &

L. González (Eds). Galicia: Trisquel Ediciones.

177


Pazos, Y., Moroño, A., & Doval, M. (2007). Anuario oceoanográfico de Galicia 2007. X. Romarís &

L. González (Eds). Galicia: Trisquel Ediciones.

Pujol, G. (2007). Sensitivity: sensitivity analysis, R package version 1.3-0.

Quevedo, M., González-Quiros, R., & Anadon, R. (1999). Evidence of heavy predation by

Noctiluca scintillans on Acartia clausi (Copepoda) eggs off the central Cantabrian coast (NW

Spain). Oceanologica Acta, 22, 127–131.

R Development Core Team, 2009. R: A Language and Environment for Statistical Computing, Vienna. <www.R-project.org>.

Robineau, B., Gagne, J. A., Fortier, L., & Cembella, A. D. (1991) Potential impact of a toxic

dinoflagellate (Alexandrium excavatum) bloom on survival of fish and crustacean larvae. Marine

Biology, 108, 293–301.

Sampayo, M. A. (1998). Polykrikos kofoidii Chatton predation on Gymnodinium catenatum

Graham and its effects. In: B. Reguera, J. Blanco, M. L. Fernandez & T. Wyatt T (Eds). Harmfull

algae. (pp 182-183). Santiago de Compostella: Xunta de Galicia and Intergovernmental

Oceanographic Commision of UNESCO.

Sherr, E. B., & Sherr, B. F., 2007. Heterotrophic dinoflagellates: a significant component of

microzooplankton biomass and major grazers of diatoms in the sea. Marine Ecology Progress

Series, 352, 187–197.

Shimizu Y. (1978). Dinoflagellate toxins. In: P.J. Scheuer, (Ed). Marine Natural Products, Chemical

and Biological Perspectives. (pp. 1–42). New York: Academic Press.

Sobol’, I. M. (1990). On sensitivity estimation for nonlinear mathematical models.

Matematicheskoe Modelirovanie, 112-118.

Spyrakos, E., González Vilas, L., Torres Palenzuela, J., & Barton E., D. (2011). Remote sensing

chlorophyll a of optically waters (rías Baixas, NW Spain): Application of a regionally specific

chlorophyll a algorithm for MERIS full resolution data during an upwelling cycle. Remote Sensing

of Environment 115(10), 2471-2485.

Strom, S. L., Morello T. A. (1998). Comparative growth rates and yields of ciliates and

heterotrophic dinoflagellates. Journal of Plankton Research, 20(3), 571-584.

Sykes, P. F., & Huntley, M. E. (1987). Acute physiological reactions of Calanus pacificus to

selected dinoflagellates: direct observations. Marine Biology, 94, 19-24.

Teegarden, G. J., & Cembella, A. (1996). Grazing of toxic dinoflagellates, Alexandrium spp., by

adult copepods of coastal Maine: implications for the fate of paralytic shellfish toxins in marine

food webs. Journal of Experimental Marine Biology and Ecology, 196, 145–176.

178


Teegarden, G. J., Cembella, A. D., Capuano, C. L., Barron, S. H., Durbin, E. G. (2003). Phycotoxin

accumulation in zooplankton feeding on Alexandrium fundyense–vector or sink? Journal of

Plankton Research, 25, 429-443.

Tillmann, U. & John, U. (2002). Toxic effects of Alexandrium spp. on heterotrophic

dinoflagellates: an allelochemical defence mechanism independent of PSP toxins, Marine

Ecology Progress Series, 230, 47–58.

Turner, J. T. (2006). Harmful algae interactions with marine planktonic grazers. In: E. Granéli & J.

T. Turner (Eds.), Ecology of harmful algae. (pp. 259-270). Berlin: Springer-Verlag.

Turner, J. (2010). Zooplankton community grazing impact on a bloom of Alexandrium fundyense

in the Gulf of Maine. Harmful Algae, 9(6), 578-589.

Turner, J. T., Doucette, G. J., Keafer, B. A., & Anderson, D. M. (2005). ECOHAB–Gulf of Maine.

Trophic accumulation of PSP toxins in zooplankton during Alexandrium blooms in Casco Bay,

Gulf of Maine, April–June 1998. II. Zooplankton abundance and size-fractionated community

composition. Deep-Sea Research II, 52 (19-21), 2784-2800.

Turner, J. T., Doucette G. T., Powell C. L., Kulis D. M., Keafer, B. A., & Anderson, D. M. (2000).

Accumulation of red tide toxins in larger size fractions of zooplankton assemblages from

Massachusetts Bay, USA. Marine Ecology Progress Series, 203, 95–107

Turner, J. T., Powell, C. L., Kulis, D. M., Keafer, B. A., Anderson, D. M., & Doucette, G. J. (2000)

Accumulation of PSP toxins in zooplankton assemblages in the Gulf of Maine. In: D. Anderson

(Ed.), Symposium on Harmful Marine Algae in the US. Marine Biological Laboratory, Woods Hole

MA.

Turner, J. T., & Tester, P. A. (1997). Toxic marine phytoplankton, zooplankton grazers, and

pelagic food webs, Limnology and Oceanography, 42, 1203–1214.

Turriff, N., Runge, J. A., & Cembella, A.D. (1995) Toxin accumulation and feeding behaviour of

the planktonic copepod Calanus finmarchicus exposed to the red-tide dinoflagellate Alexandrium

excavatum. Marine Biology, 123, 55–64.

Uhling, G., & Sahling, G. (1990). Long-term studies on Noctiluca scintillans in the German Bight

population dynamics and red tide phenomena 1968–1988. Netherlands Journal of Sea

Research, 25(1/2), 101-112.

Valdés, L., Roman, M. R., Alvarez-Ossorio, M. T., Gauzens, A. L., & Miranda, A. (1990).

Zooplankton composition and distribution off the coast of Galicia, Spain. Journal of Plankton

Research, 12, 629-643.

Watras, C. J., Carcon, V. C., Olson. R. J., Chisholm, S. W, & Anderson, D. M. (1985). The effect of

zooplankton grazing on estuarine blooms of the toxic dinoflagellate Gonyaulax tamarensis.

Journal of Plankton Research, 7, 891–908.

179


White, A.W. (1981). Marine zooplankton can accumulate and retain dinoflagellate toxins and

cause fish kills, Limnology and Oceanography, 26, 103–109.

Chapter VII. General discussion & further considerations

181

General discussion & further considerations

7.1 Development of regional specific chla algorithms from MERIS FR data for optically

complex waters

7.2 Remote sensing chla mapping during an upwelling cycle

7.3Harmful Pseudo-nitzschia spp. events in the surface waters of two Galician rias

7.4 Planktonic grazing on a PST-producer dinoflagellate


7.6 References

CHAPTER VII

183


7.1 Development of regional specific chlorophyll a algorithms from MERIS

full-resolution data for optically complex waters

The results of this study show the capability of neural network models to predict

chlorophyll a (chla) concentrations on the Galician coast from MERIS (Medium Resolution

Imaging Spectrometer) images following the widespread understanding of the need for

regionally specific models. According to the recorded in situ data, our model is an

improvement on other previously used techniques, and made it possible to obtain reliable

chlorophyll maps using almost every image. These maps can be used to study the evolution

of local oceanographic processes, which in turn could be related to the development of

algal blooms in the area. The direct relationship between these blooms and chlorophyll

concentration can also be analysed. In the 150 match-up points obtained over 6 years, we

hope that we have been able to incorporate a sufficiently wide range of the other factors

responsible for the variation in the MERIS reflectance spectra in the area. The NNRB#3

developed in this study covers the range of the temporal variations of chla concentration

that have been recorded in the Rias Baixas. Nogueira et al. (1997) suggested a temporal

pattern of chla for the Rias Baixas. According to this pattern, chla concentration is lower

than 1 mg m− 3 during the winter months and rises (up to 8 mg m− 3) during the spring and

autumn maxima. Although higher concentrations have been recorded during algal blooms,

in general, chla concentration is close to 5 mg m− 3 during the summer. As an analytical tool,

the main limitations of NNRB#3 are that it requires good MERIS images (free of cloud and

without sun glint) and, also, the masking of Clusters#2 and #3. In addition, total suspended

material (TSM) and coloured dissolved organic matter (CDOM) data would certainly

improve the neural networks (NNs). Further research should include a more thorough

analysis of the water types in the area, including CDOM and TSM data, to provide better

discrimination. Despite these potential problems, NNRB#3 was able to successfully identify

a significant relationship between reflectances and chla in the Galician rias. All NN-based

algorithms for the retrieval of water constituents have an operational range defined by the

optical characteristics of the waters for which they are trained. In our study, the scope of

NNRB#3 (the algorithm proposed here as adequate for chla mapping) is defined by the

fuzzy c-mean (FCM) results and the chla concentrations (0.03–7.94 mg m− 3). Moreover, if

sufficient points were available, it may be possible to develop a neural model for Clusters#2

and #3, or to improve the Cluster#1 model by adding new data. Chlorophyll maps would,

thereby, become more complete by blending different algorithms for different water types.


184


7.2 Remote sensing chlorophyll a mapping during an upwelling cycle

Three different states of meteorological and oceanographic periods were identified

in the area during July 2008. Surface currents and winds off the rias Baixas affected the

distribution of chla in the rias Baixas. At the beginning of July (State 1) the variable and

weak wind together with the resulting northward surface currents limited the high chla

concentrations to the rias so that only low chla values were found in the offshore area.

Differences in the topography of the rias, effects of local winds and transport by currents

between the rias seem to be the main factors for the observed differences in the gradients

of chla along the rias between ria de Arousa and the two southern rias (Vigo and

Pontevedra). MERIS images obtained during State 2 showed the first response of chla

distribution due to the strong favourable winds that were blowing in the area. With the

development of strong upwelling, the circulation in the rias is reinforced in the estuarine

sense so that chla increases rapidly there. After a period of six days of continued upwelling,

chla concentrations higher than 1 mg m-3 were observed in all the area mapped according to

MERIS data. State 3 commences with the appearance of a high biomass algal event

coincident with the area of low SST (Sea Surface Temperature) as the culmination of the

preceding, extended upwelling. The weak northward winds that characterized this state

permitted downwelling that transferred chla rich water toward the rias. The upwelled

water was recorded in the chla profiles below the surface in the ria de Vigo but was missed

by the MERIS. The continuing downwelling circulation resulted in a decay of the bloom and

a subduction of surface waters in the rias compatible with the decrease of chla observed in

the last MERIS image. Although MERIS has a repeated interval of 3 days, cloud cover

prevented acquisition of all possible images. Nevertheless, the 6 images obtained in July

2008 captured the main changes in chla concentration and distribution during the three

periods of different meteorological and oceanographic states.

The application of FCM revealed the areas where the NNRB for chla retrieval can be

best applied to obtain the most reliable results. Even if NNRB#1 and #2 did not give reliable

results, NNRB#3 showed good performance indices and seems suitable for chla

determination in the area. The chla concentrations observed in this study fall into the scope

of the NNRB algorithms proposed in Chapter 2. According to the data recorded in situ, our

models perform better than the C2R (Case-2 Regional) algorithm proving the necessity for

regionally specific models.

185


The present study allows more detailed examination of the chla distribution and

detection of high biomass “patches” in the Galician rias and the adjacent area during a

summer upwelling cycle due to the finer spatial resolution and precise atmospheric

correction offered by MERIS. The application of an algorithm specially developed for the

study area provides more accurate mapping of chla, which has resulted, for the first time to

our knowledge, surface chla mapping of the interior of the rias Baixas.

There was a significant variation in the timing and the extent of the chla peak areas.

The maps confirmed that the complex spatial structure of the phytoplankton distribution in

the rias Baixas is affected by the surface currents and winds on the adjacent continental

shelf. Field studies have limited spatial coverage and temporal frequency. Some of these

areas of high chla levels and the dynamic changes in chla distribution that are apparent in

satellite images can be missed by in situ monitoring. High chla levels in the rias due to the

increase in the concentration of harmful phytoplankton species have been recorded in the

past especially in summer (GEOHAB, 2005). Moreover, some potentially toxic species such

as Pseudo-nitzschia spp., which form blooms in the study area and in other upwelling

systems, can be found in high concentrations within these high biomass phytoplankton

“patches”.

The approach followed in this study can be particularly useful in the case of blooms

of Gymnodynium catenatum. The progression of G. catenatum blooms from the Portuguese

coast to the Galician rias can be tracked by MERIS and current data, providing useful

advanced information, of importance to the local mussel industry. It is worth mentioning

that a harmful algal event due to G. catenatum caused the closure of the shellfishery in the

Galician rias from October 2005 till February 2006 (Caballero Miguez et al., 2009).

An example of a localized feature is that constantly high surface chla was observed

in the Bay of Baiona, located in the southern mouth of ria de Vigo. This bay is characterized

as a zone where harmful algal events due to species like Alexandrium minutum are a

frequent and recurrent phenomenon (Bravo et al., 2010). It is worth noting that for the area

seaward of the rias all the algorithms used in this study came up with very similar values and

patterns for chla. Overall, this study showed that the synergy of two space borne sensors

(MERIS, MODIS) in combination with in situ data can be of great help in the monitoring,

detection and study of high biomass algal events in an coastal upwelling areas.

186


7.3 Harmful Pseudo-nitzshia events in the surface waters of two Galician rias

Domoic acid (DA) was measured for the first time in seawater samples from the

Galician rias and revealed the presence of high levels in several cases. P. australis seems to

be the main source of DA in the area which agrees with previous findings (Miguez et al.,

1996; Fraga et al., 1998). The actual method for public health protection from DA

intoxication involves the monitoring of molluscs from the rias using mouse bioassay. The

European Union´s official limit for DA is 20 mg/kg of edible part of the shellfish, while

European Food Safety authority (EFSA) has recently proposed the value 4.5 as regulatory

limit (EFSA 2009:1181). On the other hand there are no official limits, at the time of writing

this thesis, for DA concentration in the seawater and calculation or even estimation of DA

concentration in the seawater from shellfish tissue data is extremely difficult mainly due to

different depuration times. Pseudo-nitzschia spp. blooms have recurrently appeared in the

study area and as this study shows even in moderate abundances it can be accompanied by

high pDA concentrations. Despite the fact that no serious illnesses caused by ASP (Amnesic

Shellfish Poisoning) have been reported in the area, DA levels such as the ones detected

here could have the potential for significant impacts on the ecosystem and human health

(e.g by chronic explosure to moderate toxin levels as it is mentioned by Thessen & Stoecker

in 2008) and therefore should be regularly monitored. Furthermore, the DA content in

some potential vectors that can be consumed by humans and in other marine animals, has

not been studied in the Galician rias.

Ocean colour techniques can be helpful in tracking potentially harmful events due

to this diatom species and guide monitoring programs. Satellite data can be even more

useful following the approach of this PhD study with near real-time fine resolution imagery

and regionally/cluster-specific algorithms for the retrieval of chla and in-situ data.

The optimal models for the Pseudo-nitzschia spp. abundance and DA concentration

suggested the significant effect of some macronutrients as well as other abiotic and biotic

parameters, approximating in that way the potential environmental causes and effects of

the harmful Pseudo-nitzschia spp. blooms in the area. However, these models might be

improved with the addition of some parameters previously identified in other studies as

important (urea: Kudela et al. 2008; iron and copper: Rue & Bruland 2001, Wells et al. 2005).

The preliminary results (Torres et al., 2010) of models developed framework of this project

(but not presented here) applying mixed modelling and neural networks in time-series of

187


Pseudo-nitzschia abundance have shown a great potential for an operational forecast of this

diatom blooms in the area but still need some improvement.

7.4 Planktonic grazing on a PST-producer dinoflagellate

First of all, this study showed that the red form of the heterotrophic dinoflagellate

Noctiluca scintillans can actively ingest toxic Alexandrium minutum in laboratory conditions.

These two species can co-occur in the Galician rias as well as in other coastal areas around

the world. N. scintillans fed on A. minutum with no obvious toxic effects and with noticeably

lower ingestion rates than those observed in some copepods. However, this omnivorous

planktonic organinism can be found in very high abundances in coastal waters and seems

that it may have a significant grazing impact on prey population and could contribute to the

cessation of toxic blooms due to Alexandrium. The results of the first experiment were not

conclusively about the toxin accumulated in Noctiluca, probably because of the low toxin

content of Alexandrium cells. In a second experiment using this time a more toxic strain of

Alexandrium, Noctiluca revealed relatively high removal rates. This study examined the

dynamics of the toxin accumulated in two different planktonic organisms: N. scintillans and

the copepod Acartia. We found that both copepod and heterotrophic dinoflagellate seem

to be inefficient but not ineffective vectors of PSP (Paralytic Shellfish Poisoning). These

planktonic organisms show fast detoxification rates but relatively high amounts of

accumulated toxin and thus can be hardly characterised as PSP vectors. These results are

quite different from those published for bivalve shellfish (e.g. Brikelj et al., 1991). Mussels,

clams and scallops have shown high efficiency in PSP toxins accumulation while the

decontamination process can take several weeks.

An important feature of the model is that the range of the initial concentrations of

the key parameters responds to values found in the Galician rias. In this area Noctiluca has

been found in very high abundances during the end of the summer, while A. clausi is the

dominant mesozooplankton organism throuout the year. A. minutum is also present in

moderate abundances. Thus, these results could be important for the study of toxin

dynamics in this coastal system. In a next step we will consider the joined influence of the

two planktonic organisms in the toxin dynamics.

188



Galician coast, one of the most intensive areas of mussel production in the world,

suffers from a frequent occurrence of a variety of harmful algal events. Coastal

embayments like rias are physically complex systems and pose a challenge for the

detection, monitoring and forecast of harmful algal events. The high temporal and spatial

variability in biogeochemical processes that characterises these systems leads often to

patchy distribution of harmful algal events in terms of time and space.

MERIS FR data are an impotant source of synoptic observations which are focused

among other to the study of harmful algal blooms in coastal areas since it offers the

necessary spatial resolution and sensor sensibility. For remote sensing applications of

harmful algal events studies the accurate estimation of the concentration of chla is

essential. Here a new algorithm concept that makes use of neural network tehnology and

takes account different spectral properties to derive chla concentrations is presented.

Given specific assumptions this study made successfull the use of ocean colour imagery for

the estimation of a key property of harmful algal blooms such as chla in this optically

complex and dynamic system.

Moreover, management of the ecosystem requires a rapid and sensible way for

toxin detection and a greater understanding of the planktonic grazers’ role as regulators of

toxic species and their importance as vectors in the marine food-web. Our results were the

first reports, as far as we know, of DA concentrations in natural Pseudo-nitzschia spp.

populations in the study area and of accumulation of PST in the dinoflagellate red N.

scintillans. Given the fast reduction of accumulated toxin, it seems that N. scintillans and the

copepod Acartia are not an important link in the transfer of PST in the Galician rias.

Conclusively, this PhD thesis shows the benefits of an integrated use of in-situ data,

satellite data and models in the monitoring, detection, prediction and management of

harmful algal events in the Galician rias. The next step will integrate in-situ data, satellite

products and model results into really operational, user-relevant information on harmful

algal events in the area.

189


7.6 References

Bravo, I., Fraga S., Figueroa R. I., Pazos Y., Massanet A., & Ramilo I. (2010b). Bloom dynamics and

life cycle strategies of two toxic dinoflagellates in a coastal upwelling system (NW Iberian

Peninsula). Deep-Sea Research Part II: Topical Studies in Oceanography, 57, 222–234.

Bricelj, V. M., Lee, J. H. and Cembella, A. D. (1991) Influence of dinoflagellate cell toxicity on uptake

and loss of paralytic shellfish toxins in the northern quahog Mercenaria mercenaria. Marine Ecology

Progress Series, 74, 33–46.

Caballero Miguez, G., Garza Gil, M. D., & Varela Lafuente, M. M. (2009). The institutional

foundations of economic performance of mussel production: the Spanish case of the Galician

floating raft culture. Marine Policy, 33, 288–296.

Fraga, S., Alvarez, M. J., Miguez, A., Fernandez, M. L., Costas, E., & Lopez-Rodas, V. (1998).

Pseudonitzschia species isolated from Galician waters: toxicity, DNA content and lectin binding

assay. In B. Reguera, J. Blanco, M. L. Fernadez & T. Wyatt (Eds.). Harmful algae. Xunta de Galicia

and IOC of UNESCO.

GEOHAB (2005). Global Ecology and Oceanography of Harmful Algal Blooms. In G. Pitcher, T.

Moita, V. Trainer, R. Kudela, F. G. Figueiras & T. Probyn (Eds.), GEOHAB Core Research Project:

HABs in Upwelling Systems (p. 82). Paris and Baltimore: IOC and SCOR.

Kudela, R., Banas, N., Barth, J., Frame, E., Jay, D., Largier, J., Lessard, E., Peterson, T., & Van der

Woude, A. (2008). New insights into the controls and mechanisms of plankton productivity in

coastal upwelling waters of the northern California Current System. Oceanography, 21, 40–54.

Miguez Á., L., F. M., & S., F. (1996). First detection of domoic acid in Galicia (NW Spain). In T.

Yasumoto, Y. Oshima & Y. Fukuyo (Eds.), Harmful and Toxic Algal Blooms (pp. 143-145). Paris,

France: IOC of UNESCO.



185-300.

Rue, E., & Bruland. (2001).Domoic acid binds iron and copper: A possible role for the toxin

produced by the marine diatom Pseudonitzschia. Marine Chemistry, 76, 127–134.

Thessen, A. E., & Stoecker D. K. (2008). D istribution, abundance and domoic acid analysis of the

toxic diatom genus Pseudo-nitzschia from the Chesapeake Bay. Estuaries and Coasts, 31, 664-672.

Wells, M. L., Trick, C. G., Cochlan, W. P., Hughes M. P., & Trainer, V. L. (2005). Domoic acid: The

synergy of iron, copper, and the toxicity of diatoms. Limnology and Oceanography, 50,1908-1917.

Torres Palenzuela, J., González Vilas, L., & Spyrakos, E. (2010). Artificial Neural Network model

for predicting Pseudo-nitzschia spp. abundance in the Galician Rias (NW Spain). In proceeding of

14th International Conference on Harmful Algae, May 2010, Creta, Greece.

Conclusions

191

Conclusions

193

Conclusions

Conclusions

The main results of this thesis can be summarised as follow:

1. A set of chla algorithms was developed from MERIS FR data specifically for the

optically complex waters of the Galician rias and the adjacent area. These

algorithms performed better than C2R, which is routinely used for MERIS images in

Case 2 waters. The algorithm proposed here as adequate for chla mapping is the

NNRB#3. Its scope is defined by the FCM results and the chla concentrations (0.03–

7.94 mg m−3). The application of FCM reveals the areas where the NNRB for chla

retrieval can be best applied to obtain the most reliable results.

2. The application of an algorithm specially developed for the study area provided for

the first time, to our knowledge, surface chla mapping of the rias Baixas. As an

analytical tool, the main limitations of this approach are that it requires good MERIS

images (free of cloud and without sun glint) and, also, the masking of Clusters#2

and #3.

3. Three different states of meteorological and oceanographic periods were identified

in the area during the July of 2008. There was a significant variation in the timing

and the extent of the chla peak areas. The maps confirmed that the complex spatial

structure of the phytoplankton distribution in the rias Baixas is affected by the

surface currents and winds on the adjacent continental shelf.

4. The present study allows more detailed examination of the chla distribution and

detection of high biomass “patches” in the area during a summer upwelling cycle

due to the finer spatial resolution and precise atmospheric correction offered by

MERIS and the application of the local specific algorithms. This can be of great help

in the chla monitoring in any coastal upwelling area providing high-quality near real-

time cha maps and showing possible and actual harmful algal events.

5. pDA was measured for the first time, as far as we know, in seawater samples in the

Galician rias and revealed the presence of high levels in several cases. P. australis

seems to be the main source of DA.

6. The results of this study deduce that toxic events due to DA should be an important

concern and therefore DA in the seawater should be measured routinely in order to

assess the potential of a DA outbreak. Moreover, MERIS FR data and regionally

specific algorithms showed that they can provide valuable information about these

blooms and should be an integral part of the monitoring programs.

194

Conclusions

7. The optimal model for the Pseudo-nitzschia spp. abundance and DA concentration

suggested the significant effect of some macronutrients as well as other abiotic and

biotic parameters, approximating in that way the potential environmental causes

and effects of the harmful Pseudo-nitzschia spp. blooms in the area.

8. This is the first study providing evidence for ingestion of the PST microalgae

Alexandrium by the red form of the heterotrophic dinoflagellate Noctiluca

scintillans. Differences in light/dark conditions revealed different responses in food

consumption. These findings support the assumption that the red Noctiluca is

capable of limiting the growth of PST species in the field, and may play an important

role as a regulator against PST-producing phytoplankton.

9. The dynamic model predicts very fast detoxification and relatively high toxin

accumulation in both Noctiluca and Acartia suggesting that these planktonic

organisms are inefficient but perhaps not ineffective PST vectors.

10. The present PhD improves our understanding of harmful algal events in the study

area, moves forward remote sensing techniques for their detection and contributes

to the limited knowledge on aspects of toxin pathways in the food web. The

principal results of this multidisciplinary approach can be used to aid and improve

the effectiveness of monitoring and management programs.

ANNEX I

195

ANNEX I

197

ANNEX I

Resumen del Capítulo II

Desarrollo de algoritmos de índice de clorofila específicos para aguas de Caso 2 en

la costa gallega a partir de datos de MERIS FR

Las técnicas de teledetección permiten estimar la productividad del océano

mediante la obtención de la concentración de clorofila, la cual a su vez está asociada con la

abundancia de fitoplancton. Con este objetivo se utilizan sensores de color que operan en

el rango visible del espectro, consiguiéndose muy buenos resultados a escala global y en

océano abierto. Sin embargo, la exactitud es menor en áreas costeras con aguas de caso 2

más complejas en su composición y afectadas por fenómenos locales. Hay un interés

considerable en la estimación precisa de la clorofila a de las Rías Gallegas, debido a la

importancia económica y social del cultivo extensivo de mejillón, y la alta frecuencia de

eventos de algas nocivas. En este trabajo se han desarrollado algoritmos para la obtención

de mapas de clorofila para el área costera de Galicia a partir de imágenes del sensor

ENVISAT MERIS (Medium Resolution Imaging Spectrometer) de alta resolución (FR). Los

algoritmos se basan en la utilización de redes neuronales (RN) del tipo perceptrón

multicapa (MLP) que han sido entrenadas y validadas utilizando medidas de campo

obtenidas entre los años 2002 y 2008. Estos modelos se aplican a un conjunto de datos

(Clusters) definidos mediante técnicas de agrupamiento (FCM) aplicadas anteriormente a

los datos obtenidos por satélite. Se desarrollaron tres diferentes RN: una que incluye el

conjunto de datos, y otras dos con sólo los puntos que pertenecen a una de las categorías

(Clusters). Los buenos resultados obtenidos demuestran el gran potencial del sensor MERIS

para obtener mapas de clorofila en el interior de las Rías Gallegas. La mejor predicción fue

dada por la RN entrenada con datos de alta calidad mediante el conjunto de datos de la

categoría (Cluster) más abundante. Los parámetros estadísticos en el conjunto de

validación de esta RN fueron R2 = 0,86, MPE= − 0.14, RMSE= 0,75 mg m−3 y RMSE relativa =

66%. La RN desarrollada en este estudio detecta con precisión los altos valores de clorofila,

en los conjuntos de entrenamiento y validación. Los resultados mostraron que este

algoritmo superó el C2R (Case-2 Regional Processor), dando mayor valor de R2 y valores

RMSE inferiores. Este estudio demostró que la combinación de datos in situ y tecnología de

RN permite obtener mapas de clorofila más precisos. Un algoritmo de índice de clorofila

específico para la zona, obtenido a partir de los datos de un sensor de color del océano con

las características de MERIS, sería un gran apoyo en el seguimiento y estudio de eventos de

198

ANNEX I

algas nocivas. También se describen las limitaciones y posibles mejoras de los algoritmos

desarrollados.

199

ANNEX I

Resumen del Capítulo III

Aplicación de algoritmos, a partir de imágenes MERIS FR, de índice de clorofila

específicos para la costa gallega durante un ciclo de afloramiento

Este estudio utiliza algoritmos regionalmente específicos y las características de

Medium Resolution Imaging Spectrometer (MERIS) para obtener mapas de clorofila en el

área costera de Galicia (aguas caso 2), con una precisión mayor que la obtenida por otros

algoritmos ya existentes. Los algoritmos empleados se basan en la utilización de redes

neuronales y técnicas de “fuzzy clustering” utilizando datos in-situ. El algoritmo que mostró

la mejor eficacia fue aplicado en una serie de seis imágenes MERIS (FR) durante un ciclo de

afloramiento que se observó en el área en Julio de 2008. Los cambios principales en la

concentración y distribución de clorofila a fueron captados claramente en las imágenes. Los

mapas confirmaron que la estructura espacial de la distribución de fitoplancton en el área

de estudio puede ser compleja. Las corrientes oceánicas y los vientos de superficie

afectaron a la distribución de clorofila en las Rias Baixas. Este estudio mostró que un

algoritmo regionalmente específico para un sensor de color de océano con las

características de MERIS, en combinación con datos in situ, puede ser de gran ayuda en la

comprensión de la dinámica marina en el interior de las Rías y en concreto en la distribución

de biomasa algal durante un ciclo de afloramiento.

Introducción

Entre los productos derivados de los sensores del color del océano, la

concentración de clorofila a es el más utilizado ya que permite estimar de forma precisa la

biomasa de fitoplancton, ya que este pigmento es común a casi todos los grupos

taxonómicos. Además, las poblaciones de fitoplancton responden rápidamente a los

cambios ambientales provocando variaciones en su concentración.

Se han desarrollado varios algoritmos empíricos, como el tradicional cociente entre

las bandas azul y verde, así como diferentes modelos semi-analíticos para la estimación de

la clorofila a partir de sensores de color. Sin embargo, con este tipo de modelos o

algoritmos no se obtiene una estimación precisa para aguas costeras de Caso 2, con altas

concentraciones de otros componentes como detritus o sustancias amarillas, los cuales

absorben la radiación en la banda azul.

200

ANNEX I

Se ha demostrado que las técnicas basadas en redes neuronales (RN) son más

apropiadas para trabajar con datos complejos y no lineales, por lo que pueden ser muy

útiles para estimar de una forma más precisa las concentraciones de los diferentes

componentes del agua en aguas ópticamente complejas.

Las Rías Gallegas (noroeste de España, Figura 8.1) se caracterizan por su alta

productividad y son muy importantes en la economía de la región, ya que en ellas se

desarrolla un cultivo intensivo de mejillón en bateas y sustentan significativas pesquerías

locales. La productividad se debe a los eventos periódicos de afloramiento de aguas

profundas más frías y ricas en nutrientes que se producen principalmente entre Mayo y

Septiembre, bajo condiciones de viento del noreste y gracias a la particular morfología y

orientación de las Rías. En el presente trabajo se ha aplicado un algoritmo de clorofila

basado en una RN y desarrollado previamente para las aguas de las Rías Gallegas a un

conjunto de 6 imágenes MERIS FR durante un ciclo de afloramiento. Los resultados se

compararon con las medidas in situ y con otros algoritmos que se utilizan habitualmente

para imágenes MERIS (CR2). Finalmente, se analizó la distribución temporal y espacial de

los patrones de clorofila obtenidos a partir de las imágenes en relación con las condiciones

meteorológicas y oceanográficas observadas en la zona.

Fuentes de datos

MERIS es un sensor óptico a bordo del satélite ENVISAT que permite obtener imágenes

con 15 bandas espectrales en un rango entre 412 nm y 900 nm y con una resolución

temporal de 3 días. En este estudio se utilizaron 6 imágenes MERIS de alta resolución (300

m) adquiridas en el mes de Julio de 2008 (Tabla 8.1).

Tabla 8.1: Imágenes MERIS utilizadas, con indicación de la fecha, hora de la adquisición y ángulo cenital medio desde el oeste.

Fecha de la Imagen Hora de la adquisición

(UTC)

Angulo cenital (º)

03-07-08 10:59 13.5

09-07-08 11:10 13.0

16-07-08 10:50 20.7

19-07-08 10:56 15.3

22-07-08 11:02 11.7

29-07-08 10:42 20.7

201

ANNEX I

Además, se llevaron a cabo dos campañas de muestreo en la ría de Vigo los días 9 y

22 de Julio del 2008, coincidentes con sendas imágenes. En estas campañas se tomaron

muestras de agua por triplicado en 12 estaciones, desde la superficie hasta tres metros de

profundidad, para la determinación de pigmentos mediante cromatografía líquida de alta

resolución (CLAE) y de las partículas en suspensión utilizando gravimetría. También se

estableció la profundidad de la zona eufótica con un disco de Secchi.

Finalmente, se utilizaron los datos de velocidad y dirección del viento y de la

corriente obtenidos con la boya oceanográfica Seawatch localizada frente a Cabo Silleiro

(42 ° 7.8 'N, 9 ° 23.4' W; Fig. 8.1). A partir de estos datos se estimó el índice diario de

afloramiento.

Fig. 8.1. Costa gallega y batimetría de la zona. La boya Seawatch se muestra con un rectángulo negro.

Metodología

Las imágenes utilizadas son del nivel 2 de procesamiento, es decir, se distribuyen

con los valores de reflectancia ya calculados para cada pixel. En un primer paso, se les aplicó

la corrección radiométrica para corregir el efecto “smile” implementada en BEAM-4.6 y se

llevó a cabo la corrección atmosférica mediante la aplicación del algoritmo basado en una

RN desarrollado por Doerffer and Schiller (2008).

El segundo paso fue la aplicación a los datos de reflectancia del clasificador no

supervisado basado en la teoría de agrupación por lógica difusa, en concreto en la técnica

FCM (Fuzzy c-Means Clustering). Según este trabajo, las reflectancias de MERIS en la zona

202

ANNEX I

de Galicia se clasifican en tres cluster (#1, #2, #3), que son los que se observan en las

imágenes de clasificación obtenidas (Fig. 8.2).

Fig. 8.2. Clasificación de una imagen de MERIS (22-07-08). Se muestran las tres clases identificadas por el clasificador FCM.

El siguiente paso fue la aplicación a los datos de reflectancia de las tres RN (NNRB#1,

NNRB#2, NNRB#3) desarrollas en el trabajo anterior (Capitulo II) para la estimación de la

clorofila. Las redes solo se aplicaron a los píxeles previamente clasificados dentro del

Cluster#1, ya que la estimación utilizando los otros grupos no sería fiable, y nos permitió

obtener mapas de clorofila. Posteriormente, se compararon las concentraciones de

clorofila obtenidas para las estaciones de muestreo con los valores correspondientes en los

mapas de clorofila derivados de las RN, y se calculó el coeficiente de correlación (R2) y el

error cuadrático medio (RMSE) para evaluar las redes. La comparación con los datos in situ

se extendió también al algoritmo desarrollado por Doerffer para aguas de Caso 2.

Finalmente, los mapas derivados de la red con la que se obtuvieron los mejores resultados

en la comparación con los datos in situ (mayor R2 y menor RMSE) se utilizaron para analizar

la distribución espacial y temporal de la clorofila en relación con los datos de vientos,

203

ANNEX I

corrientes e índices de afloramiento.

Resultados

Datos in-situ

La concentración de la clorofila en las muestras de agua no mostró una amplia

variación. La concentración de la clorofila determinada por CLAE varió de 0.03 mg m-3 a 2.72

mg m-3 durante los dos muestreos.

Clasificador FCM

El Cluster#1, en el cual se pueden aplicar los algoritmos de clorofila, es claramente

predominante en la serie de las imágenes para la zona de estudio. Así, considerando las 6

imágenes, al menos el 72%, 71% y 65% de los pixeles corresponden a esta categoría en las Rías

de Vigo, Pontevedra y Arousa, respectivamente. El Cluster#2 es predominante en la Ría de

Pontevedra y la Ría de Arousa en la imagen del 29 de julio. El Cluster#3 es el menos

abundante en la mayoría de las imágenes, ya que supone menos de 3,25% de los píxeles en

todos los casos. La Figura 8.2 muestra un ejemplo de imagen de clasificación, en la cual se

observa como el Cluster#1 es predominante en todas las Rías Baixas y el área adyacente.

Algoritmos para la estimación de clorofila

En general, las tres redes neuronales mejoraron significativamente los resultados

obtenidos con el algoritmo C2R (R2 = 0.04 y RMSE = 0.9 mg m-3), mostrando mayores

valores de R2 y valores más bajos de RMSE. Los mejores resultados se obtuvieron para la

red NNRB#3 (R2 = 0.70 y RMSE = 0.46 mg m-3)

Ciclo de afloramiento

Se identificaron tres periodos meteorológicos y oceanográficos diferentes con una

duración de entre nueve y once días (Fig. 8.3).

204

ANNEX I

Fig. 8.3. Índice diario de afloramiento, que representa el flujo de Ekman en la capa superficial. Las

flechas indican los días que había imágenes de MERIS FR disponibles, los números los diferentes

periodos identificados durante el mes estudiado y los círculos los días de muestreo.

En esta parte se muestra tan solo uno de los mapas de clorofila obtenidos de la

aplicación del algoritmo en la zona (Fig. 8.4). El resto de las imágenes se describen en

relación con las condiciones meteorológicas y oceanográficas (Capitulo III).

Fig. 8.4. Mapa de clorofila durante el ciclo de afloramiento de julio de 2008 en el área de estudio. La tierra y las nubes estaban enmascarados y aparecen en color blanco.

205

ANNEX I

A principios de Julio (periodo 1), los vientos y las corrientes superficiales causan una

acumulación elevada de clorofila en el interior de las Rías. Durante el segundo período, las

imágenes MERIS mostraron los primeros efectos, en la distribución de clorofila, de los

fuertes vientos favorables para el afloramiento que soplaban en la zona. Con el desarrollo

de fuertes corrientes, la clorofila aumentó rápidamente, y después de un período de seis

días de afloramiento continuo, se observaron concentraciones de clorofila superiores a 1

mg m-3 en toda el área a través de los datos de MERIS. El periodo 3 comienza con la

aparición de un evento de elevada biomasa de algas (Fig. 8.4) que coincide con el final de

un fuerte afloramiento. Los vientos del norte débiles al inicio de este período, permiten que

el agua con alta concentración de clorofila se transfiera hacia las Rías. El hundimiento

continuo que caracterizó a este período tuvo como resultado la dispersión de las altas

cantidades de clorofila y el hundimiento de las aguas superficiales en las Rías, lo cual

coincide con la disminución de la clorofila superficial observada en la última imagen de

MERIS.

Conclusiones

El presente estudio demostró que las imágenes MERIS de alta resolución son una

herramienta muy potente para el análisis detallado de la distribución de la clorofila en las

Rías de Galicia y en la plataforma adyacente durante un ciclo de afloramiento. La aplicación

de un algoritmo especialmente desarrollado para el área de estudio nos permite obtener

mapas precisos de clorofila con una resolución nunca antes lograda para esta zona. Los

mapas muestran que la estructura espacial de la distribución del fitoplancton en el área de

estudio puede ser compleja. Algunas de estas áreas de alta concentración de clorofila que

se pueden identificar en las imágenes de satélite podrían incluso no ser detectadas por los

programas de monitorización in situ.

206

ANNEX I

Resumen del Capítulo IV

Teledetección, monitoreo in situ y perspectivas ambientales de eventos tóxicos de

Pseudo-nitzschia en las aguas superficiales de dos Rías Gallegas (noroeste de

España)

En este capítulo se aborde un análisis que combina datos de imágenes de satélite,

medidas in situ de parámetros bióticos y abióticos y modelos estadísticos para estudiar las

floraciones de Pseudo-nitzschia en las aguas superficiales de dos Rías gallegas (Vigo y

Pontevedra). Este análisis incluye también la determinación de las concentraciones de ácido

domoico (AD) en las poblaciones naturales realizadas por primera vez en la zona y la

determinación de las abundancias de Pseudo-nitzschia junto con varios parámetros

ambientales durante los años 2007-2009. Un algoritmo de índice de clorofila a a partir de

datos de MERIS (Medium Resolution Imaging Spectrometer) FR fue aplicado a una serie de

imágenes. Se consideraron tres períodos relacionados con las condiciones meteorológicas

dominantes. El primer evento tóxico fue obtenido en el otoño de 2007 y se caracterizó por

moderadas abundancias de Pseudo-nitzschia spp. y concentraciones altas de AD (2,5 mg L-1).

Un segundo episodio de menor toxicidad de AD fue registrado durante el verano de 2009.

P. australis parece ser la principal fuente de AD en el área de estudio. La aplicación de un

algoritmo regional en combinación con las características de MERIS FR permitió obtener de

forma precisa concentraciones de clorofila a y detectar pequeñas areas de alta

concentración de clorofila y elevada abundancia de Pseudo-nitzschia spp. Los modelos

óptimos de efecto mixto (GAMM) para las abundancias de Pseudo-nitzschia y de GLMM para

la concentración de AD incluyen el efecto significativo de algunos macronutrientes, así

como otros parámetros abióticos y bióticos. El uso de estos modelos podrían ayudar a

conocer las causas medioambientales y los efectos perjudiciales en el desarrollo de las

floraciones nocivas de Pseudo-nitzschia spp. en la zona.

207

ANNEX I

Resumen del Capítulo V

Tasas de ingestión del dinoflagelado heterótrofo Noctiluca scintillans sobre la

microalga tóxica Alexandrium minutum

Se evaluaron las tasas de ingestión del dinoflagelado heterótrofo Noctiluca scintillans

(McCartney) sobre la microalga tóxica Alexandrium minutum Halim en cultivos de

laboratorio. Ejemplares de N. scintillans fueron recolectados en la Ría de Vigo y

posteriormente mantenidos en cultivo. Se prepararon 4 concentraciones experimentales de

A. minutum (entre 50 y 400 células ml-1) y se seleccionaron 360 individuos de N. scintillans

que fueron transferidos a botes de 40 ml que contenían alimento. Las cepas fueron

cultivadas durante 5 días y separadas en dos grupos: con ciclo luz-oscuridad 12:12h y sin luz.

Al término del experimento se evaluaron las tasas de ingestión diarias, la toxicidad en A.

minutum y en N. scintillans (mediante cromatografía líquida). Los resultados muestran que

en ambos grupos N. scintillans se alimentó activamente de A. minutum, durante el periodo

evaluado, a tasas de ingestión máximas de 0.3 μg C individuo-1 día-1, valor en el cual se

alcanza la saturación, manteniéndose la presión de consumo a través del tiempo. La

toxicidad en A. minutum estuvo dentro del rango normal detectado para esta especie, con

un valor medio de 2.90 fmol célula-1. Los análisis de los perfiles cromatográficos de los

individuos de N. scintillans expuestos a A. minutum revelaron que la especie no acumuló las

toxinas paralizantes en su citoplasma, lo que sugiere que la toxina sería metabolizada o

bien liberada. Estos datos son los primeros realizados sobre el consumo de dinoflagelados

tóxicos en condiciones de laboratorio por parte de N. scintillans.

208

ANNEX I

Resumen del Capítulo VI

Modelización de transferencia y acumulación de PST en dos consumidores planctónicos

En este capitulo se desarrolló un modelo con el objectivo de estudiar y comparar la

transferencia y la acumulación de PST en dos diferentes potenciales vectores planctónicos

de PST. El dinoflagelado heterótrofo en su forma roja Noctiluca scintillans y el copépodo

Acartia clausi fueron los organismos planctónicos que se seleccionaron para este estudio.

También se investigó la importancia de estos dos organismos como vectores de PST en la

comunidad planctónica. Diferentes factores que influyen en la tranferencia de toxinas

fueron considerados en el modelo, como la síntesis de las toxinas, la ingestión de alimento

tóxico y no tóxico y el tamaño de la población tanto en los vectores como en los

productores de toxina. Por otra parte, se llevó a cabo un experimento de laboratorio para

calcular las tasas de detoxificación en individuos de Noctiluca alimentados con A. catenella.

Los resultados del modelo mostraron diferencias en los tiempos que se aparecieron los

picos de acumulación de PST en los dos organismos. Noctiluca presenta una respuesta

rápida a la ingestión de Alexandrium que se caracteriza con una alta acumulación inicial de

toxina, seguida por una reducción a cero en un período de casi dos días. Por el contrario,

Acartia mostró un considerable retraso para acumular los mismos niveles de PST en la

población observados en Noctiluca. Este retraso está ralacionado con las tasas de

reproducción más lentas que caracterizan al copépodo. El rango de valores iniciales

utilizados para el análisis de sensibilidad del modelo es representantivo del medio costero

de una Ría Gallega, donde los dos consumidores y el dinoflagelado Alexandrium pueden

coexistir. Los resultados del modelo revelaron menor sensibilidad de Acartia para los

parámetros más importantes, probablemente debido al retraso en la acumulación de

toxinas. Ambos organismos mostraron una reducción rápida (50 h) de la toxina ingerida,

sugiriendo que no son muy eficaces en la tranferencia de toxinas a través la cadena trofica

marina.

209

ANNEX I

Conclusiones

1. Un conjunto de algoritmos para la estimación de la clorofila a partir de datos del

sensor MERIS FR se desarrolló específicamente para las aguas ópticamente

complejas de las Rías de Galicia y el área adyacente. Estos algoritmos superaron el

C2R, que se utiliza habitualmente para imágenes MERIS en aguas de Caso 2. El

algoritmo adecuado propuesto en esta tesis para la estimación de clorofila a es el

NNRB # 3. Su ámbito de aplicación viene definido por los resultados de la FCM y las

concentraciones de clorofila a (0.03–7.94 mg m−3). La aplicación de la FCM muestra

las áreas donde puede aplicarse mejor el NNRB para obtener los resultados más

fiables.

2. La aplicación de un algoritmo especialmente desarrollado para el área de estudio

nos permite obtener mapas precisos de clorofila con una resolución nunca antes

lograda para esta zona. Las principales limitaciones de esta metodología son en

primer lugar, que se requieren buenas imágenes MERIS (libres de nubes y sin

reflección solar en la zona de estudio) y por otro lado también se necesita

enmascarar las categorías #2 y # 3.

3. Se identificaron tres periodos de diferentes condiciones meteorológicas y

oceanográficas en la zona durante Julio de 2008. Hubo una variación significativa en

el tiempo y la extensión de las áreas de alta concentración de clorofila. El presente

estudio demostró que las imágenes MERIS de alta resolución son una herramienta

muy potente para el análisis detallado de la distribución de la clorofila en las Rías de

Galicia y en la plataforma adyacente durante un ciclo de afloramiento.

4. Los mapas muestran que la estructura espacial de la distribución del fitoplancton en

el área de estudio puede ser compleja. Algunas de estas áreas de alta concentración

de clorofila que se pueden identificar en las imágenes de satélite podrían incluso no

ser detectadas por los programas de monitorización in situ.

5. La concentración de ácido domoico (AD) se midió por primera vez en muestras de

agua de mar de las Rías Gallegas y además se ha detectado la presencia de esta

toxina en altos niveles en varios casos. P. australis parece ser la principal fuente de

DA.

6. Los resultados de este estudio sugieren que los eventos tóxicos debidos a AD

deberían ser un motivo de preocupaión importante y por lo tanto, el AD debería

medirse. Por otra parte, los datos de MERIS FR y los algoritmos específicos

210

ANNEX I

mostraron que pueden suministrar información valiosa de casos de floraciones

algales y deberían ser parte integrante de los programas de monitoreo.

7. El modelo óptimo para la abundancia de Pseudo-nitzschia spp. y para la

concentración DA sugirió el efecto significativo de algunos macronutrientes, así

como otros parámetros bióticos y abióticos, aproximando de ese modo las posibles

causas ambientales y los efectos de las floraciones perjudiciales de Pseudo-nitzschia

spp. en el área.

8. Se presentan los primeros datos sobre el consumo de dinoflagelados tóxicos en

condiciones de laboratorio por parte de N. scintillans. Las tasas de ingestión

máximas fueron obtenidas con altas abundancias de alimentos. Las diferencias en

las condiciones de luz y oscuridad muestran diferentes respuestas en el consumo de

alimentos. Estos resultados soportan la hipótesis que Noctiluca es capaz de limitar

el crecimiento de especies productoras de PST en el campo y pueden tener un papel

importante como regulador de fitoplancton toxico.

9. El modelo dinámico predice una detoxificación muy rápida y una acumulación de

toxinas relativamente alta en Noctiluca y Acartia. Estos resultatos muestran que los

dos organismos planctónicos pueden ser ineficientes, pero quizás no ineficaces

como vectores de PST.

10. La presente tesis doctoral ha mejorado nuestra comprensión sobre los eventos de

algas nocivas en las aguas interiores de las Rias Gallegas y ha mejorado la

metodología preexistente, basada en técnicas de teledetección, para el estudio de

índices de clorofila superficial y su relación con la dinámica oceanográfica de la zona

de estudio. Los principales resultados de este enfoque multidisciplinar pueden ser

de gran interés y utilidad en los programas de monitorización relacionados con

episodios de floraciones algales nocivas/tóxicas.

ANNEX II

211

ANNEX II

213

ANNEX II

Técnicas de inmunodetección y en concreto ELISAs aplicadas al campo

marino

El uso de anticuerpos (Acs) en técnicas de detección aplicadas al campo marino

permite identificar de forma específica organismos o moléculas de interés. Entre las

técnicas más empleadas se encuentran:

1) Enzyme Linked InmunoSorbent Assay (ELISA)

2) Inmunofluorescencia (IF)

3) Western Blot (WB).

La descripción de estos métodos, con ejemplos de aplicaciones concretas, es el

objeto del presente capítulo. La técnica más conocida y extendida, debido principalmente a

su simplicidad, es el ELISA. Permite detectar y/o determinar la concentración de una

molécula (o antígeno) con alta sensibilidad. Se inmoviliza el antígeno a detectar o los Acs

específicos sobre el fondo de un pocillo de plástico, y, posteriormente se lleva a cabo la

detección con Acs unidos a enzimas, mediante una reacción colorimétrica. Esta técnica

puede aplicarse en la identificación de especies, para evitar fraudes, para detectar

sustancias tóxicas o prohibidas, etc.

La immunofluorescencia usa Acs unidos a fluorocromos y suele aplicarse para la

detección de organismos enteros. Muestras de agua de mar conteniendo larvas de

moluscos y dinoflagelados pueden ser visualizadas bajo microscopio de fluorescencia o

confocal, o de forma automática, con un citómetro de flujo.

El Western blot consiste en la identificación de una proteína concreta en una

muestra compleja, tras una separación electroforética en gel de poliacrilamida y posterior

transferencia a membrana. El revelado con anticuerpos específicos conjugados a enzimas

permitirá identificar la proteína de interés en la membrana. Esta técnica permite

diagnosticar enfermedades y adulteraciones en productos marinos, localizar una proteína

en un tejido, identificar y/o diferenciar especies marinas, etc.

Las técnicas de inmunodetección pueden ser utilizadas con éxito para identificar

especies sin necesidad de conocimientos taxonómicos previos, incluso en productos

procesados, cocidos o subproductos que carecen de caracteres morfológicos.

214

ANNEX II

ELISA

La necesidad de identificar o detectar de forma rápida y específica diversos

compuestos ha llevado a que las técnicas de inmunodetección utilizando Acs se estén

aplicando en muchos campos científicos. De todas ellas, el ELISA es quizás el método más

empleado debido a su sencillez, sensibilidad y versatilidad. Mediante esta técnica se están

llevando a cabo controles de procedencia de muchos productos, con el fin de evitar el

fraude o la comercialización de especies protegidas que, una vez procesadas, no pueden ser

identificadas morfológicamente.

Descripción del método

Hay distintos tipos de ELISAs (competitivo, sándwich, directo e indirecto) que se

realizan utilizando microplacas de plástico, antígenos y Acs marcados con una enzima

(como la peroxidasa de rábano picante o fosfatasa alcalina) que reaccionará con un

sustrato incoloro o cromógeno (ejemplos: ABTS, TMB, DAB) produciendo un producto

coloreado (Figura 1). La lectura de la señal se realiza en un espectrofotómetro equipado con

un sistema de filtros que permite la lectura simultánea de todos los pocillos de la placa a

una longitud de onda determinada. En los últimos años, se han desarrollado técnicas

semejantes al ELISA, de mayor sensibilidad, utilizando agentes quimioluminiscentes o

fluorescentes, así como el marcaje del ligando o antígeno en vez del anticuerpo

(EnzimoInmunoAnálisis o EIA). Las técnicas de ELISA son muy sensibles, detectando entre

picogramos-nanogramos de sustancia en una muestra.

1) ELISA directo: los pocillos se recubren con la solución donde se encuentra el antígeno a

detectar, y a continuación, la placa se incuba con Acs específicos frente a dicho antígeno,

conjugados a una enzima. Tras lavar para eliminar aquellos anticuerpos no unidos, se lleva a

cabo la reacción colorimétrica. La coloración que aparece es proporcional a la cantidad de

antígeno presente en la muestra.

2) ELISA indirecto: Al igual que en el caso anterior, los pocillos se recubren con el antígeno,

pero primero se incuban con Acs primarios no marcados específicos frente al antígeno, y, a

continuación, después de lavados, la placa se incuba con Acs secundarios frente a los

primarios, conjugados a una enzima. Presenta mayor sensibilidad que el ensayo directo, ya

215

ANNEX II

que a cada Ac primario se le pueden unir varios Acs secundarios, amplificando la señal. Este

ensayo es útil para la cuantificación de Acs en una muestra.

3) ELISA tipo "sándwich": esta variante utiliza dos tipos de Acs, ambos específicos frente al

antígeno. Los pocillos se recubren con Acs primarios específicos. Tras bloquear y lavar la

placa, se incuba con la muestra donde se encuentra el antígeno a detectar, y, tras sucesivos

lavados, se incuba con Acs conjugados a enzimas dirigidos frente a un epítope diferente del

antígeno. Esta variante presenta gran especificidad y es muy útil para la detección y

cuantificación de moléculas grandes.

4) ELISA competitivo. Es el más sensible y versátil. En este caso, antígeno o anticuerpo se

adhieren al plástico y va a existir una competencia por aquellos que estén en solución,

marcados o no con enzima. Como ejemplo, los pocillos de la placa se recubren con

antígenos conocidos conjugados a proteínas inertes. Posteriormente se añade una mezcla

de muestra problema + una cantidad conocida y limitante de anticuerpo específico marcado

con enzima. En el proceso de lavado se retiran aquellos anticuerpos que reconocen al

antígeno en forma soluble y por tanto la señal detectada en el pocillo será inversamente

proporcional a la cantidad de antígeno presente en la muestra. Es una técnica muy útil para

la detección y cuantificación de moléculas pequeñas.

En todos los tipos de ELISA es necesario incluir controles, tanto negativos (muestra

donde la sustancia a detectar está ausente), como positivos (muestra con cantidades

conocidas de la sustancia a detectar). Para que el ensayo sea cuantitativo deben incluirse

diluciones seriadas de una muestra con concentración conocida, para realizar una curva

estándar donde interpolar los datos de las muestras. Para evitar uniones inespecíficas se

requiere el bloqueo de las placas con una proteína inerte como albúmina sérica bovina o

caseína.

Entre cada paso, la placa se lava con una solución tamponada salina que contenga una baja

concentración de un detergente suave (ej. PBS 0.001% Tween), para eliminar todas las

moléculas no unidas de modo específico.

Utilización

El uso de la técnica de ELISA en el campo marino es muy amplio. Se ha empleado en

programas de monitorización de calidad del agua para la detección de diferentes

216

ANNEX II

patógenos como bacterias, virus, hongos y parásitos protozoarios (Schloter et al., 1995).

Gracias a la elevada sensibilidad de la técnica, y capacidad de procesar un gran número de

muestras, puede ser útil en el estudio de enfermedades víricas y bacteriológicas de peces

moluscos marinos y cetáceos, así como para estudios de reproducción de peces y moluscos

(Kang et al., 2003). Desde hace años se han utilizado en estudios de procesos bioquímicos

oceánicos y costeros y en la detección y cuantificación de especies de fitoplancton. Algunos

ELISAs permiten detectar la presencia de determinadas microalgas nocivas y toxinas

marinas (Garet et al., 2010) lo cual tiene gran utilidad para la industria de la acuicultura y

protección de la salud pública.

En el campo de las toxinas marinas es fundamental encontrar métodos alternativos

al bioensayo en ratón; en las últimas décadas se están desarrollando inmunoensayos

enzimáticos. En la mayoría de los casos los ELISAs pueden ser aplicados en varias matrices,

incluyendo muestras de agua, extractos de algas o muestras biológicas (mariscos, líquidos

corporales...). Un buen ejemplo del uso de esta técnica es la reciente validación de un ELISA

como método oficial para la detección de ácido domoico (toxina marina de la categoría de

las amnésicas) como alternativa al método de bioensayo en ratón. Actualmente se

encuentran ya comercializados diversos ELISAs para toxinas marinas (Tabla 9.1).

Tabla 9.1. Kits de ELISA comerciales para toxinas marinas.

Toxinas Efecto/

Categoría*

kits ELISA disponibles Principales microalgas

productoras

Saxitoxina y

relacionadas

Paralizantes (PSP)/

grupo saxitoxina

- Ridaserren, (R-Biopharm,

Darmstadt, Germany)

- Abraxis LLC, (Warminster PA,

USA)

Alexandrium spp. y

Gymnodinium catenatum

Ácido domoico Amnésicas (ASP)/

grupo ácido

domoico


Norway)

Pseudo-nitzschia spp.

Ácido ocadaico y

derivados

Diarreicas

(DSP)/

grupo ácido

ocadaico

- UBE Industries, Ltd, (Tokyo,

Japan)

- Rougier Bio-Tech, (Ltd of

Montreal, Canada)

Dinophysis y

Prorocentrum

Yesotoxina grupo yesotoxina - Biosense Laboratories (Bergen,

Norway)

Protoceratium

reticulatum,

Lingulodinium polyedrum

y Gonyaulax spinifera.

Brevetoxina y

análogos

Neurotóxicas

(NSP)/

grupo brevetoxina


Norway)

Karenia brevis

217

ANNEX II

ELISA: Ejemplo de aplicación

El ácido domoico es una ficotoxina amnésica, producida principalmente por varias

especies de diatomeas del género Pseudo-nitzschia. Los síntomas de las toxinas amnésicas

incluyen vómitos, diarrea, y en casos más graves pérdida de memoria, desorientación, e

incluso el coma o la muerte.

El ensayo autorizado como método oficial para la detección y cuantificación de

ácido domoico y estandarizado por la Association of Official Analytical Chemists (AOAC,

2006), se basa en un ensayo de ELISA tipo competitivo (Figura 9.1). Toxina conjugada a

proteina es depositada en un pocillo y la toxina presente en las muestras problema compite

por la unión con el anticuerpo específico frente a la toxina, que está conjugado a una

enzima (peroxidasa).

Fig. 9.1. Procedimiento para calcular la concentración intracelular de ácido domoico (AD) en células de Pseudo-nitzschia spp. mediante un Elisa competitivo. HRP: peroxidasa de rábano picante.

Remote sensing, in situ monitoring and planktonic toxin ...

Documents

Transcript of Remote sensing, in situ monitoring and planktonic toxin ...