1

Paper accepted in Marine Pollution Bulletin

Title: Setting the maximum ecological potential of benthic communities, to assessecological status, in heavily morphologically-modified estuarine water bodies

Authors: Ángel Borja*, Guillem Chust, Andrea del Campo, Manuel González, CarlosHernández

Affiliation: AZTI-Tecnalia; Marine Research Division; Herrera Kaia, Portualdea s/n;20110 Pasaia (Spain).

* Corresponding author: A. Borja E-mail: aborja@azti.es; Phone number: +34-667174430, Fax number: +34-943004801

Abstract

Investigations on setting benthic macroinvertebrates reference conditions innatural waters have increased recently. Under the European Water FrameworkDirective, importance is given to research in morphological heavily-modified waterbodies (HMWB), which are very common in countries with high human pressure.However, research has not been undertaken on setting the maximum ecologicalpotential (MEP), as a reference in HMWB. The objective of the present investigation isto set the MEP of two metrics (diversity and richness), used in assessing the ecologicalstatus in different benthic indices. The Oiartzun estuary (Basque Country) is used as acase study, which changed morphologically in the 19th Century, following harbourconstruction. Data obtained from 1874 and the present were used to model changes incurrents, water residence time, salinity, volume, and intertidal area. Benthicmacroinvertebrate data, from 1995 to 2011, were used to predict 19th Century andpresent MEP. Changes in the estuary were described: loss of all of the intertidal areas;doubling of the volume; residence time, changing from 2 to 95 days; current velocityreduced by 50%; salinity increase. All these factors have led to changes in the benthiccommunities and the structural variables. Predicted richness and diversity, for 1874,were lower (48-76%) than those at present. Taking into account the differences betweennatural and modified waters, it is proposed to utilize 75% of the natural referenceconditions, as the MEP values for Basque HMWB.

Keywords: maximum ecological potential; Water Framework Directive; benthic

metrics; reference conditions; modelling; Basque Country

2

1.- Introduction

The European Water Framework Directive (WFD; 2000/60/EC) has several

objectives, for example: to prevent water ecosystems deterioration and to protect and

enhance the status of water resources. However, the most important is to achieve a

‘good ecological status’ (GES) for all waters, by 2015.

Ecological quality is based upon the status of the biological (phytoplankton,

macroalgae, macrophytes, benthos and fishes), hydromorphological and physico-

chemical quality elements. Consequently, many methodologies have been proposed in

recent years to assess the ecological status of water bodies, within the WFD (Birk et al.,

2012).

Most of these methodologies were developed for its use within natural water

bodies, in which the ecological status is a perceived, or measured, deviation from a

reference condition (Hering et al., 2010). However, the WFD also defines and considers

‘Artificial Water Bodies’, i.e. those created by human activity, such as an artificial lake,

and ‘Heavily Modified Water Bodies’ (HMWB), a water body resulting from physical

alterations by human activity, which substantially change its hydrogeomorphological

character, e.g. a harbour. In both cases of definition, Member States may designate a

body of surface water as being ‘artificial’ or ‘heavily modified’ when there are

significant adverse effects to the hydromorphological characteristics of that body, which

otherwise would be necessary for achieving GES. Such effects include: (i) the wider

environment; (ii) navigation, including port facilities, or recreation; (iii) activities for

the purposes for which water is stored, such as drinking-water supply, power generation

or irrigation; (iv) water regulation, flood protection, land drainage; or (v) other equally-

important sustainable human development activities.

In terms of implementing the WFD, environmental managers are required to

3

assess the status of HMWB, in relation to achieving at least ‘Good Ecological Potential’

(GEP). A water body shows a GEP when there are slight changes in the values of the

relevant abovementioned biological quality elements, compared to the values found at

the maximum ecological potential (MEP). The MEP is considered as the reference

condition for HMWB; it is intended to describe the best approximation to a natural

aquatic ecosystem, which could be achieved given the hydromorphological

characteristics that cannot be changed without significant adverse effects on the

specified use, or the wider environment (CIS, 2003a, 2003b; Borja and Elliott, 2007).

The meaning of potential is something that can only be achieved if something else

happens e.g. the hydromorphological causes of change are removed. Accordingly, the

MEP biological conditions should reflect, as far as possible, those associated with the

closest comparable natural water body type, at reference conditions; in accordance with

the established hydromorphological and associated physico-chemical conditions.

Although the WFD implementation process for natural waters has experienced

important advances since approval of the Directive (Hering et al., 2010), at present only

minimal advancement has taken place, in terms of understanding the meaning of GEP,

especially within an ecological context; likewise, how to define MEP (CIS, 2006;

Wetzel et al., 2012). However, Borja and Elliott (2007) have proposed an approach to

establish the MEP, through several steps, as outlined below.

(i) Select the appropriate quality elements for MEP, identifying the closest

comparable natural surface water category.

(ii) Establish the hydromorphological conditions required for the MEP, which

determine the values for the biological and general physico-chemical quality

elements.

(iii) Establish the MEP physico-chemical conditions, identifying the closest

4

comparable surface water body type, taking into account the MEP

hydromorphological conditions.

(iv) Finally, establish the MEP biological conditions that reflect, as far as possible,

those associated with the closest comparable water body type.

In the case of natural coastal and transitional waters, much research has been

undertaken in setting reference conditions and boundaries for quality classes, especially

in terms of benthic macroinvertebrates (see Borja et al. (2009a), for a review of

European methods, and Borja et al. (2012), for the problems associated with the setting

of reference conditions). After the WFD, any macroinvertebrates assessment must be

based upon richness, diversity and the ratio between opportunistic and sensitive species.

One of the methods being used by many European countries, in determining this ratio,

is the AZTI’s Marine Biotic Index (AMBI), as described in Borja et al. (2000). More

recently, amongst the methods used in assessing benthic status (see Borja et al., 2009a),

the multivariate-AMBI (M-AMBI; Muxika et al., 2007) is being used in several

countries; this includes the three single metrics mentioned above (diversity, richness

and the AMBI). From these metrics, AMBI (as the ratio of opportunistic/sensitive

species) values for MEP should not differ considerably from those required in natural

waters, because most of the sensitive taxa of the type-specific communities should be

present in the HMWB, under good conditions. However, richness and diversity will

change, compared to natural conditions.

Hence, the objective of the present investigation is to set the reference conditions

(i.e. MEP) of these two single metrics (diversity and richness), to assess the ecological

potential of benthic macroinvertebrates in an HMWB, following the approach proposed

previously by Borja and Elliott (2007).

5

2.- Materials and Methods

2.1. Study area and main human pressures

In order to investigate this particular topic, an HMWB in the Basque Country,

southeastern Bay of Biscay, has been selected for investigation. This area is the

Oiartzun estuary, which was modified considerably to create the Pasaia Harbour (Figure

1). The morphology of this estuary has changed dramatically since the mid-19th

Century, following the development of the harbour (Reizabal et al., 1987; Rivas and

Cendrero, 1992). Dredging activities, to maintain the adequacy of the navigation

channels, together with the building of the port infrastructures, have resulted in

morphological changes of the natural environment (Uriarte et al., 2004; Tueros et al.,

2009; Montero et al., 2013). Furthermore, the estuary has been highly contaminated by:

industrial development on the area, e.g. with a thermal power station and a paper mill;

and the settlement of several villages, with hydrometallurgical- and mining-related

discharges (Cantón and Legorburu, 1991).

The Oiartzun estuary is 5.5 km in length, with a total surface of about 1 km2 and

an average water depth of around 10 m (Valencia et al., 2004). The Oiartzun river flows

into this estuary, and drains an 87 km2 area and has an annual mean flow of 4.8 m3 s−1

(Borja et al., 2006).

The estuary is sheltered and the main natural driving force is the tide, which is

semidiurnal in the Bay of Biscay. The maximum spring tidal range exceeds 4.5 m. The

region is defined as ‘low meso-tidal’ during neap tides, but ‘high meso-tidal’ during

springs (González et al., 2004).

2.2. Approach followed in the study and the data used

The harbour is located within the North-East Atlantic (NEA)-1/26 typology of

6

the WFD. Following the approach discussed in the Introduction (based upon Borja and

Elliott, 2007), a further series of steps were completed (see below).

2.2.1. Benthic macroinvertebrates data

As macroinvertebrates are the biological quality element to be analysed for the

MEP, some old benthic data extracted from Navaz (1948) and for the present (1995-

2011), were obtained for the Oiartzun estuary and other 11 estuaries from the Basque

Country (see sampling details in Borja et al., 2009b). The use of data from other Basque

estuaries is related to the fact that the present MEP conditions, from the HMWB, should

be associated with the closest comparable water body type.

2.2.2. Modification of the hydromorphological conditions

The study of this modification, required for the MEP, was based upon changes in

the docks and seabed within the estuary; these produce distinct alterations in water

residence times, tidal prisms, etc. In order to estimate such changes, existing

bathymetric and morphological charts of the previous and present configuration of the

estuary were used. The oldest nautical chart available for the Oiartzun estuary,

published by the Spanish Directorate of Hydrography in 1874, was digitized. As it was

neither georeferenced, nor scaled, a more recent and referenced nautical chart published

by the French Service Hydrographique et Océanographique de la Marine, in 1992, was

used for rectification purposes (6375 Chart, Port de Pasajes, at 1:7500). Hence, 54

common points were identified on both of the charts; these permitted a graphical

rectification of the 1874 chart. Although 118 years separate the charts, some common

locations were identified, i.e. a lighthouse, several castles, churches, the train rail and

the non-modified coastline of Pasajes San Juan village. Additionally, some bathymetric

information gaps, as detected for the inner part of the estuary, were completed using a

bathymetric chart published in 1909. Figure 2 shows the overlapping of the rectified

7

1874 nautical chart, with the coastline published in 1992. With respect to the present

bathymetric and morphological characterization of the estuary, the bathymetric data

from a multibeam survey undertaken on 28th July 2009 for the Port Authority of Pasajes,

together with the coastline published in 1992, were used.

Intertidal and subtidal volumes and areas were calculated for the past and

present configurations of the estuary. To achieve this objective, two 5 m resolution

Digital Terrain Models (DTM) were constructed: (i) Pasajes 1874: 5 m grid DTM

referred to the hydrographical zero at that time; and (ii) Pasajes 2009: 5 m grid DTM

referred to the local hydrographical zero. It is of note that intertidal volumes and areas

were computed by considering the maximum tidal amplitude indicated on the chart of

1874 (i.e. 3.3 m) and, for the chart of 2009; the latter is the value considered in Valencia

et al. (2004) i.e. 4.5 m.

Tidal currents and water residence times within the estuary were estimated for

both morphological configurations, in 1874 and 2009. Firstly, a depth-integrated 2-D

hydrodynamical model, based upon the Finite Element Method and named MAREAS

(González et al., 1995), was applied to both morphological configurations to simulate

the tidal currents. In this model, shallow-water equations are solved with an harmonic

decomposition of the free surface elevation and the velocity fields (Walters, 1986). Sea

level measurements (a 5 year data series), as obtained from the oceano-meteorological

station located at the mouth of the estuary (http://estacion.itsasnet.com), were used as

external boundary conditions for astronomic tidal current simulations. The 24

amplitudes and phases of the tidal components were obtained; these were assumed to be

identical, in both of the morphological configurations.

High-resolution Finite Element computational meshes were constructed for the

model implementation, for both of the morphological configurations. The values from

8

the DTMs elaborated previously were interpolated to the Finite Element mesh, using the

Delaunay Triangulation Method (Press et al., 2007). The mean area of the quadrilaterals

of the 1874 and 2009 configuration meshes were 50 m2 and 37 m2, respectively.

In order to evaluate water renewal, the estuary volume, both in its present and

past configurations, was filled in the model by a uniformly-distributed passive tracer

(with an initial concentration equal to unity (1)); its Eulerian dispersion was simulated,

without considering diffusion, until the tracer mean concentration fell below e−1, i.e.

37%, of its initial concentration. The evolution in the tracer concentration was fitted to

an exponential curve, as proposed by Tartinville et al. (1997); from this, the renewal

time parameter was obtained. The adjustment of this equation to the results of the

dispersion model enabled the calculation of the water residence time. A more detailed

description of the method applied here can be found in Montero et al. (2013).

2.2.3. Establishing physico-chemical conditions

The physico-chemical conditions of the harbour should be based upon those in a

reference water body (see Bald et al., 2005)). However, the changes in

hydromorphological conditions could lead to a change in typology and, subsequently, to

new reference conditions associated with the new type. The physico-chemical

conditions will have an important influence on the values for the biological quality

elements within the MEP. Taking into account the morphological changes, probably the

most important elements affected by the harbour presence could be salinity, sediment

grain-size, dissolved oxygen, water transparency, etc.; these are through changes in

residence time and current velocities.

Hence, changes in salinity distribution, in the past and present morphological

configurations of the estuary, were estimated by applying numerical modelling. The

Eulerian dispersion model applied was the same as that used for the water residence

9

time calculations, as described previously. Boundary conditions in the model were: 35

psu at the mouth of the estuary; and 0 psu at the headwaters of the river. The model

simulated the salinity dispersion, due to tidal currents, over a year. The salinity

distribution was estimated by averaging the salinity values of the past 9 months,

simulated at each point of the numerical mesh (both in the 1874 and 2009

configurations).

2.2.4. Establishing the MEP biological conditions

These conditions reflect, as far as possible, those associated with the closest

comparable water body type (see above). The biological conditions at the MEP will be

influenced by the hydromorphological and physico-chemical conditions. In the case of

the harbour, the changes in residence times can lead to changes in physico-chemical

conditions, resulting again in changes in richness and diversity in the benthic fauna.

Hence, modelling of the single metrics of the benthic communities is needed, to

determine the benthic MEP of the Oiartzun Estuary, before and after the

hydromorphological changes. It is necessary to determine the values of two of the single

metrics used in M-AMBI, such as species richness and Shannon’s diversity. Ysebaert et

al. (2002) proposed to use logistic regression to predict benthic community composition

in estuaries. The approach used in our investigation consisted of the statistical

modelling of the metrics (not the species composition), as a function of estuarine

features and local conditions, such as intertidal area and salinity, using Generalized

Additive Models (GAM, Hastie and Tibshirani, 1990). Subsequently, they were used to

predict the diversity and species richness of Oiartzun, in 1874; then, to extract metric

values for each station and the overall estuary, both in intertidal and subtidal areas.

However, all intertidal data used in this investigation were obtained from the lower

intertidal (0-1 m tidal height). Hence, it should be noted that, as the value of intertidal

10

areas in supporting the biota differs with tidal height, the results from this investigation

are only valid for subtidal and lower intertidal zones.

As MEP conditions reflect those associated with the closest comparable water

body type, a dataset, with 715 observations of benthic data from the 12 Basque

estuaries, since 1995, was used (Borja et al., 2009b) to derive the above information.

Some of these estuaries represent similar types to that in the past for the Oiartzun

estuary, e.g. Oka, Oria, Butrón (see Figure 1), whilst others represent similar types in

present times (e.g. Nervión or Bidasoa). Since several estuaries have experienced a

recent increase in biodiversity, as a result of water treatment (e.g. Nervion estuary,

Pascual et al., 2012), a time-series analysis has been undertaken prior to the model

building. In this way, the effect of habitat and geomorphological features, on

biodiversity indices, can be isolated. As explanatory variables, those experiencing

variation with anthropogenic-driven hydromorphological changes were used: salinity

(classified into four gradient classes: euhaline, polyhaline, mesohaline, and oligohaline);

tide zonation (intertidal, subtidal); estuarine volume; subtidal volume; estuarine area;

and the percentage of the intertidal area.

The steps undertaken in the statistical modelisation were: (i) time-series analysis

of the biodiversity metrics, for each of the estuaries; (ii) the removal of the sampling

data (over half of the period, i.e. after 2003) for those estuaries that biodiversity metrics

increase significantly, due to recent water treatment and quality improvement; (iii) the

building of the GAM (Poisson distribution for species richness, and Gaussian for

Shannon’s diversity index) using all of the estuaries and selected data at the present

time; (iv) model selection – firstly, by removing highly correlated variables (r > 0.97),

and subsequently using AIC (Akaike Information Criterion)-, which permits the

removal of variables that do not improve the model; and (v) the prediction of

11

biodiversity metrics for the Oiartzun Estuary, under the hydromorphological scenario of

1874.

3.- Results

3.1. Changes in hydromorphological conditions

Important hydromorphological changes were estimated for the Oiartzun estuary.

Figure 3 shows the bathymetry (DTM) of both present and past morphologies of the

estuary. The areas, volumes and water residence times, computed for both years, are

listed in Table 1. The overall surfaces of the intertidal areas, which represented 71% of

the total area of the estuary, were lost; at present, a subtidal habitat dominates the

estuary. Consequently, the estuary volume has increased. In the present configuration,

the estuary has more than the double the volume it had in the past. All these

morphological changes are reflected in a very different water residence time, which, in

the past configuration, was found to be of the order of 2 days; presently it is of the order

of 95 days.

Figure 4 shows the pattern of the tidal currents, computed for both

configurations. Maximum tidal currents were around two times higher during the past

configuration, than at present; changing from 25 cm s-1 to 13 cm s-1 in the channel, and

from 7 cm s-1 to 3 cm s-1 in the middle part of the estuary.

3.2. Changes in physico-chemical conditions

The numerical computations have identified that the Oiartzun estuary was less

saline during the past configuration, than at present. Such differences in salinity

distribution are represented in Figure 5. Whilst, in the past morphological configuration,

the middle part of the estuary appeared to have an important mesohaline area (i.e. from

5 to 18 psu), in the present configuration it was mainly polyhaline (i.e. from 18 to 30

psu).

12

3.3. Modelling single metrics of benthic communities

The time-series analysis undertaken has revealed that four estuaries (the

Nervion, Oiartzun, Urumea, and Orio) presented significant increases in species

richness (Table 2, Figure 6). Hence, all sampling stations monitored prior to 2003 have

been removed (i.e. 216 samplings) for these estuaries, from the dataset (i.e. the overall

dataset retained 499 selected samplings).

Amongst the explanatory variables, estuary volume and subtidal volume were

correlated highly with estuarine area (r>0.97); hence, the former 2 variables were

removed. For modelling species richness (on a logarithmic scale), two variables were

included as parametric factors (salinity and tidal zone) and two as smooth terms

(estuarine area and percentage of intertidal area). Hence, although hydrological and

morphological changes were important in the estuary, the most important variables, in

terms of biological response, were those related to changes in salinity and tidal zone

characteristics. Overfitting was prevented, by restricting to 3 the degree of smoothness

of model terms. All of the variables entered were retained according to AIC and the

significant p-value of the model coefficient (Table 3). The generalised additive model

for richness explained 56.7% of the variance (Figure 7), compared with 51.3% using all

715 samples, i.e. prior to data selection. Species richness was greater in saltwater than

in freshwater; likewise, in subtidal rather than intertidal zones (Figure 8). Large

estuaries, with a low percentage of intertidal areas, incorporate higher species richness

(Figure 8). The mean richness prediction for the Oiartzun Estuary, for 1874 (14.6), was

less than that estimated (28.7) for the present geomorphological conditions (Table 4).

Slight underestimation was detected in the model (estimated at 28.7 taxa, compared

with observed of 34.4). As has been shown, geomorphological conditions in the estuary,

in the past, were less saline and with a higher percentage of intertidal area; this, in turn,

13

predicts less species richness.

For the Shannon’s diversity index (H), the statistical model was similar to that

used for richness, since the two metrics were highly correlated (r = 0.76), although less

accurate (44.3% of the explained variance). H depended only upon salinity and

estuarine area (Table 3). The prediction of mean H for 1874 (2.12 bits ind-1) was also

less than that estimated (2.53 bits ind-1) for present geomorphological conditions (Table

4). As in the case of species richness, hydrological conditions in the past were less

saline, combined with an higher percentage of intertidal area, which predicts less

diversity.

Discussion

Setting reference conditions is one of the most critical issues in assessing the

ecological status of marine waters (Borja et al., 2012). The WFD identifies four options

for setting such conditions: (i) utilising the conditions at pristine locations, or sites with

only very minor disturbance; (ii) utilising historical data and information, on formerly

undisturbed sites; (iii) applying modelling techniques, to predict reference conditions; or

(iv) utilising expert judgement (Vincent et al., 2002). Advantages and disadvantages of

the four options have been reviewed by Borja et al. (2012).

The setting of reference conditions, for macroinvertebrate communities in

natural water bodies, has been investigated for different European areas: Denmark

(Nielsen et al., 2003); the Basque Country (Muxika et al., 2007); Portugal (Teixeira et

al., 2008); Italy (Paganelli et al., 2011); or Mediterranean and Black Sea lagoons

(Basset et al., 2013). However, there is an absence of investigations in Europe,

regarding the setting of macroinvertebrates MEP; this remains an important challenge in

the WFD implementation (Hering et al., 2010).

In this investigation undertaken within the Basque Country MEP setting in

14

HMWB, a statistical modelling approach has been used, together with reference to some

historical information. Hence, following prediction of the MEP values for 1874,

compared with those of present natural areas in the Basque Country estimated by

Muxika et al. (2007) (Table 5): richness and diversity in the oligo-mesohaline area

represented 71% and 75% of the present natural reference conditions; MEP values in

the polyhaline area were 47-55% of the natural ones; and MEP values in the euhaline

area were 61-76% of the reference conditions. In the oligo-, meso- and polyhaline

sections of the estaury, richness and diversity values in 1874 were very similar to those

associated with natural reference conditions of the Cerastoderma edule-Scrobicularia

plana community (as described in Borja et al., 2004). Most probably, this is because the

area laid close to natural conditions in 1874, i.e. low human pressure, as can be deduced

from historical studies (Reizabal et al., 1987). At that time, the estuary was essentially

intertidal, with low residence times and meso-polyhaline conditions (as shown in the

results of this investigation). Hence, the estuary was dominated probably by the

Cerastoderma edule-Scrobicularia plana community, as described by Navaz (1948);

this is as it occurs in similar small Basque natural estuaries, presently (Borja et al.,

2004, 2009b). Subsequently, the estuary is now heavily modified, mainly subtidal, with

long residence times, is more euhaline, and the expected benthic communities are those

of Venus fasciata and Abra alba (described in Borja et al., 2004). These communities

are richer and more diverse than those within the oligo-mesohaline areas. HMWB,

specifically harbours, are characterized by changes in residence and flushing times

(Sámano et al., 2012); as such, by alterations of salinity stretches (when the HMWB is a

transitional water), resulting in changes in structuring benthic communities, as

highlighted in the present contribution.

Furthermore, comparing the estimated present MEP and natural reference

15

conditions, for single benthic metrics, could be of value (Table 5). Richness and

diversity estimated MEP, in polyhaline areas, represented 66% and 48%, of the natural

ones, respectively; the estimated MEP in euhaline areas represented around 76%. Thus,

in present HMWB, the MEP conditions should be lower than those in equivalent natural

areas.

It is well known that species richness and diversity of benthic fauna is greater in

euhaline, rather than in oligohaline conditions (following the Remane’s diagram);

likewise, in subtidal, rather than intertidal zones (Whitfield et al., 2012). In addition,

large estuaries, incorporating a mixture of subtidal and intertidal areas, hold higher

structural benthic community values, than small estuaries dominated by intertidal areas;

this is because of the diverse number of habitats (Turpie et al., 2004). Salinity, tidal

zonation and the percentage contribution of the intertidal zone appear to represent the

habitat factors affecting species richness. The relationship of the size of estuaries, with

species richness, is associated probably with habitat size effects (Shaffer, 1981). Thus,

populations inhabiting larger estuaries would be affected less by demographic

stochasticity; therefore, a larger number of species could be found per unit of area.

As commented upon above, studies have not been undertaken in setting MEP in

HMWB, in transitional or coastal waters; nonetheless, some research has been

undertaken in assessing the status of these HMWB, e.g. Borja et al. (2009) and Wetzel

et al. (2012). In The Netherlands, with the closure of the Afsluitdijk in 1932, the former

Southern Sea estuary was transformed into the freshwater Ijsselmeer Lake (Lammens et

al., 2008). These latter authors stated that, in this particular case, a natural lake rather

than an estuary will be the starting point for determining ecological potential; including

in the MEP the effect of the hydromorphological changes, after all mitigation measures

have been considered.

16

In the Oiartzun Estuary, the effect of hydromorphological changes, before the

main harbour construction and after more than a century of its functioning, has been

investigated. From this investigation, it is interesting to note that the mean values of

structural variables, such as richness and diversity, increased when a small estuary, with

extensive intertidal flats and short residence times, was transformed into a HMWB. This

followed a harbour construction, with extensive subtidal areas and long residence times.

However, the new values for that HMWB, associated with MEP conditions, were

always lower than those of the reference conditions in unaltered areas.

Hence, taking into account the above differences, between natural and modified

waters (the range of structural values in modified waters was 48-76% of the natural

ones), the proposal of this investigation is that present MEP values should lie at around

75% of the natural reference conditions (Table 5) i.e. areas have a potential to be 33%

higher than they are now if the hydromorphological stressors were removed; this takes

into account a precautionary approach. The WFD indicates that this potential reduction

can be accepted (and the area classed as a HMWB) as long as there is cost-benefit

reason for not removing the hydromorphological stressors. In this particular case, our

proposal is only for richness and diversity, not for AMBI; this is used also in the

multivariate analysis, when calculating M-AMBI to assess the ecological status, within

the WFD (Muxika et al., 2007). For AMBI, taking into account the comments in the

Introduction, our recommendation is to utilise the same values as in the reference

conditions of natural waters. Finally, it is necessary to highlight that this approach is

based upon a data-rich area and the methods described here may not be as accurate in

data-poor areas.

Acknowledgements

The Basque Water Agency (URA) funded this study, through a Convention with

17

AZTI-Tecnalia (project: Ecological Monitoring Network of the Basque Country). The

Port Authority of Pasajes provided the bathymetric data of 2009. Professor Michael

Collins (School of Ocean and Earth Science, University of Southampton (UK) and

AZTI-Tecnalia (Spain)) advised us on some details of the manuscript. A reviewer has

provided interesting opinions, which have improved the manuscript. This is contribution

number 621 from the Marine Research Division (AZTI-Tecnalia).

References

Bald, J., A. Borja, I. Muxika, J. Franco, V. Valencia, 2005. Assessing referenceconditions and physico-chemical status according to the European WaterFramework Directive: A case-study from the Basque Country (Northern Spain).Marine Pollution Bulletin, 50: 1508-1522.

Basset, A., E. Barbone, A. Borja, M. Elliott, G. Jona-Lasinio, J.C. Marques, K. Mazik,I. Muxika, J.M. Neto, S. Reizopoulou, I. Rosati, H. Teixeira, 2013. Naturalvariability and reference conditions: setting type-specific classificationboundaries for lagoon macroinvertebrates in the Mediterranean and Black Seas.Hydrobiologia, 704: 325-345.

Birk, S., W. Bonne, A. Borja, S. Brucet, A. Courrat, S. Poikane, A. Solimini, W. van deBund, N. Zampoukas, D. Hering, 2012. Three hundred ways to assess Europe'ssurface waters: An almost complete overview of biological methods toimplement the Water Framework Directive. Ecological Indicators, 18: 31-41.

Borja, A., M. Elliott, 2007. What does `good ecological potential' mean, within theEuropean Water Framework Directive? Marine Pollution Bulletin, 54: 1559-1564.

Borja, A., J. Franco, V. Pérez, 2000. A marine biotic index to establish the ecologicalquality of soft-bottom benthos within European estuarine and coastalenvironments. Marine Pollution Bulletin, 40: 1100-1114.

Borja, A., F. Aguirrezabalaga, J. Martínez, J.C. Sola, L. García-Arberas, J.M.Gorostiaga, 2004. Benthic communities, biogeography and resourcesmanagement. Borja, A. and Collins, M. (Eds.). Oceanography and MarineEnvironment of the Basque Country, Elsevier Oceanography Series, 70: 455-492.

Borja, A., I. Galparsoro, O. Solaun, I. Muxika, E.M. Tello, A. Uriarte, V. Valencia,2006. The European Water Framework Directive and the DPSIR, amethodological approach to assess the risk of failing to achieve good ecologicalstatus. Estuarine, Coastal and Shelf Science, 66: 84-96.

Borja, A., A. Miles, A. Occhipinti-Ambrogi, T. Berg, 2009a. Current status ofmacroinvertebrate methods used for assessing the quality of European marinewaters: implementing the Water Framework Directive. Hydrobiologia, 633: 181-196.

Borja, A., I. Muxika, J.G. Rodríguez, 2009b. Paradigmatic responses of marine benthiccommunities to different anthropogenic pressures, using M-AMBI, within theEuropean Water Framework Directive. Marine Ecology, 30: 214-227.

Borja, Á., D.M. Dauer, A. Grémare, 2012. The importance of setting targets and

18

reference conditions in assessing marine ecosystem quality. EcologicalIndicators, 12: 1-7.

Cantón, L., I. Legorburu, 1991. Heavy metal concentration in sediments from Pasajesharbour, Spain. Marine Pollution Bulletin, 22: 207-209.

CIS, 2003a. Overall Approach to the Classification of Ecological Status and EcologicalPotential. Water Framework Directive Common Implementation StrategyWorking Group 2 A Ecological Status (ECOSTAT), Rome, 53 pp.

CIS, 2003b. Identification and Designation of Heavily Modified and Artificial WaterBodies. Water Framework Directive Common Implementation StrategyWorking Group, Guidance Document No 4, Produced by Working Group 2.2 –HMWB, European Communities, 118 pp.

CIS, 2006. WFD and hydromorphological pressures. Good practice in managing theecological impacts of hydropower schemes; flood protection works; and worksdesigned to facilitate navigation under the Water Framework Directive. WaterFramework Directive Technical Report, European Communities, 68 pp.

González, M., M.A. García, M. Espino, A. Sánchez-Arcilla, 1995. Un modelo numéricoen elementos finitos para la corriente inducida por la marea astronómica.Aplicaciones al Estrecho de Gibraltar. RIMNI, 11: 383–400.

González, M., A. Uriarte, A. Fontán, J. Mader, P. Gyssels, 2004. Marine dynamics.Borja, A. and Collins, M. (Eds.) Oceanography and Marine Environment of theBasque Country, Elsevier Oceanography Series, 70: 133-157.

Hastie, T.J., R.J. Tibshirani, 1990. Generalized Additive Models. Chapman & Hall,London, XV + 335 pp.

Hering, D., A. Borja, J. Carstensen, L. Carvalho, M. Elliott, C.K. Feld, A.-S. Heiskanen,R.K. Johnson, J. Moe, D. Pont, A.L. Solheim, W. v. de Bund, 2010. TheEuropean Water Framework Directive at the age of 10: A critical review of theachievements with recommendations for the future. Science of The TotalEnvironment, 408: 4007-4019.

Lammens, E., F. Van Luijn, Y. Wessels, H. Bouwhuis, R. Noordhuis, R. Portielje, D.Van Der Molen, D. 2008. Towards ecological goals for the heavily modifiedlakes in the IJsselmeer area, the Netherlands. Hydrobiologia, 599(1): 239-247.

Montero, N., M.J. Belzunce-Segarra, A. Del Campo, J.M. Garmendia, L. Ferrer, J.Larreta, M. González, M.A. Maidana, M. Espino, 2013. Integrativeenvironmental assessment of the impact of Pasaia harbour activities on theOiartzun estuary (southeastern Bay of Biscay). Journal of Marine Systems, 109-110: S252–S260.

Muxika, I., A. Borja, J. Bald, 2007. Using historical data, expert judgement andmultivariate analysis in assessing reference conditions and benthic ecologicalstatus, according to the European Water Framework Directive. Marine PollutionBulletin, 55: 16-29.

Navaz, J.M., 1948. Estudio de la ría de Pasajes en relación con su producción demoluscos comestibles (con una carta bionómica). Boletín del Instituto Españolde Oceanografía, 11: 1-31.

Nielsen, K., B. Somod, C. Ellegaard, D. Krause-Jensen, 2003. Assessing referenceconditions according to the European Water Framework Directive usingmodelling and analysis of historical data: an example from Randers Fjord,Denmark. Ambio, 32: 287-294.

Paganelli, D., G. Forni, A. Marchini, C. Mazziotti, A. Occhipinti-Ambrogi, 2011.Critical appraisal on the identification of Reference Conditions for theevaluation of ecological quality status along the Emilia-Romagna coast (Italy)

19

using M-AMBI. Marine Pollution Bulletin, 62: 1725-1735.Pascual, M., A. Borja, J. Franco, D. Burdon, J.P. Atkins, M. Elliott, 2012. What are the

costs and benefits of biodiversity recovery in a highly polluted estuary? WaterResearch, 46: 205-217.

Press, W.H., S.A. Teukolsky, W.T. Vetterling, B.P. Flannery, 2007. Numerical Recipes:The Art of Scientific Computing (3rd ed.). New York: Cambridge UniversityPress. 1235 pp.

Reizabal, G., P. Gónzalez-Amezúa, A. González, 1987. Pasajes, un puerto, una historia.Junta del Puerto de Pasajes, ISBN 8450552036, 9788450552034, 209 pp.

Rivas, V., A. Cendrero, 1992. Análisis histórico de la evolución superficial de losestuarios del País Vasco. Lurralde, 15: 199-227.

Sámano, M. L., J. F. Bárcena, A. García, A. G. Gómez, C. Álvarez, J. A. Revilla, 2012.Flushing time as a descriptor for heavily modified water bodies classificationand management: Application to the Huelva Harbour. Journal of EnvironmentalManagement, 107: 37-44.

Shaffer, M.L., 1981. Minimum population sizes for species conservation. BioScience,31: 131-134.

Tartinville, B., E. Deleersnijder, J. Rancher, 1997. The water residence time in theMururoa atoll lagoon: sensitivity analysis of a three-dimensional model. CoralReefs, 16: 193-203.

Teixeira, H., F. Salas, J. M. Neto, J. Patrício, R. Pinto, H. Veríssimo, J. A. García-Charton, C. Marcos, A. Pérez-Ruzafa, J. C. Marques, 2008. Ecological indicestracking distinct impacts along disturbance-recovery gradients in a temperate NEAtlantic Estuary - Guidance on reference values. Estuarine, Coastal and ShelfScience, 80: 130-140.

Tueros, I., Á. Borja, J. Larreta, J.G. Rodríguez, V. Valencia, E. Millán, 2009.Integrating long-term water and sediment pollution data, in assessing chemicalstatus within the European Water Framework Directive. Marine PollutionBulletin, 58: 1389-1400.

Turpie, J., B. Clark, D. Knox, P. Martin, C. Pemberton, C. Savy, 2004. Contributions toinformation requirements for the implementation of resource directed measuredfor estuaries. Volume 1. Improving the biodiversity importance rating of SouthAfrican estuaries. Water Research Commission Report nº 1247/1/04, 145 pp.

Uriarte, A., M. Collins, A. Cearreta, J. Bald, G. Evans, 2004. Sediment supply, transportand deposition: contemporary and Late Quaternary evolution. Borja, A. andCollins, M. (Eds.) Oceanography and Marine Environment of the BasqueCountry, Elsevier Oceanography Series, 70: 97-131.

Valencia, V., J. Franco, A. Borja, A. Fontán, 2004. Hydrography of the southeasternBay of Biscay. Borja, A. and Collins, M. (Eds.) Oceanography and MarineEnvironment of the Basque Country, Elsevier Oceanography Series, 70: 159-194.

Vincent, C., H. Heinrich, A. Edwards, K. Nygaard, J. Haythornthwaite, 2002. Guidanceon Typology, Reference Conditions and Classification Systems for Transitionaland Coastal Waters Produced by: CIS Working Group 2.4 (Coast), CommonImplementation Strategy of the Water Framework Directive, EuropeanCommission.

Walters, R.A., 1986. Finite element model for tidal and residual circulation.Communications in Numerical Methods in Engineering, 2(4): 393–398.

Wetzel, M.A., P.C. von der Ohe, W. Manz, J.H.E. Koop, D.-S. Wahrendorf, 2012. Theecological quality status of the Elbe estuary. A comparative approach on

20

different benthic biotic indices applied to a highly modified estuary. EcologicalIndicators, 19: 118-129.

Whitfield, A.K., M. Elliott, A. Basset, S.J.M. Blaber, R.J. West, 2012. Paradigms inestuarine ecology - A review of the Remane diagram with a suggested revisedmodel for estuaries. Estuarine, Coastal and Shelf Science, 97: 78-90.

Ysebaert, T., P. Meire, P. M. J. Herman, H. Verbeek, 2002. Macrobenthic speciesresponse surfaces along estuarine gradients: prediction by logistic regression.Marine Ecology Progress Series, 225: 79-95.

21

Figure captions

Figure 1: The Oiartzun Estuary study area and sampling locations, within the Basque

Country, showing also the position of the remaining estuaries included in the

investigation of benthic metrics.

Figure 2: The present coastline of the Oiartzun Estuary superimposed on the nautical

chart of the area, in 1874.

Figure 3: Bathymetric changes observed within the Oiartzun Estuary, between 1874 (a)

and 2009 (b).

Figure 4: Changes in the current velocity, modelled in the Oiartzun Estuary, in 1874 (a)

and 2009 (b)

Figure 5: Changes in salinity, modelled in the Oiartzun Estuary, between 1874 (a) and

2009 (b)

Figure 6: Species richness trends for those estuaries where statistically-significant

increase was detected: a) Nervion; b) Oiartzun; c) Orio; and d) Urumea

estuaries.

Figure 7. Species richness observed versus that predicted by the selected Generalised

Additive Model (GAM).

Figure 8: Explanatory factors for benthic species richness, according to the selected

Generalised Additive Model (GAM). The grey shadow is the standard error = 2.

Key: Oligoh- oligohaline; Mesoh- mesohaline; Polyh- polyhaline; Euh- euhaline.

22

Fig. 1 Borja et al.

23

Fig 2

24

Fig 3

25

Fig 4

26

Fig 5

(PSU)

27

Fig 6.

15

20

25

30

35

40

1994 1999 2004 2009

Spec

ies r

ichn

ess

10

15

20

25

30

35

1994 1999 2004 2009

02468

101214

1994 1999 2004 2009

Spec

ies r

ichn

ess

Year

0

2

4

6

8

10

12

1994 1999 2004 2009

Year

(a) (b)

(c) (d)

28

Fig 7.

1

10

100

1 10 100

Spec

ies r

ichn

ess

(obs

erve

d)

Speciess richness (predicted)

29

Fig 8.

30

Table 1. Changes experienced in the Oiartzun Estuary, between 1874 and 2009, in

terms of intertidal and subtidal surface (area), total and subtidal volume; these

are presented together with the percentages (between brackets), and water

residence times (Rt).

Year Area (ha) Volume (Mm3) Rt (days)

Intertidal Subtidal Total Subtidal

1874 78.27 (71%) 32.41 (29%) 5.13 2.36 (46%) 2

2009 0.00 (0%) 94.99 (100%) 12.91 8.63 (67%) 95

31

Table 2. Time-series analysis between 1995 and 2011, of benthic species richness for

each of the Basque estuaries. Significant p-values are shown in bold.

Estuary Linear Regression coefficient p-valueArtibai -0.241 0.212Barbadun 0.001 0.990Bidasoa 0.107 0.501Butron 0.248 0.193Deba 0.031 0.871Lea 0.283 0.083Nervión 0.599 0.0004Oiartzun 0.824 0.0008Oka 0.027 0.852Orio 0.386 0.0009Urola 0.281 0.130Urumea 0.292 0.019

32

Table 3. Generalised Additive Models (GAM) for species richness and the Shannon’s

diversity index, using hydrographical and geomorphological variables of the estuaries.

Key: EDF- estimated degrees of freedom of the smooth term (s).

Explanatoryvariables

Parametriccoefficients

EDF p-value

Species richness

Salinity (mesohaline) -0.562 <0.0001Salinity (oligohaline) -1.099 <0.0001Salinity (polyhaline) -0.358 <0.0001Tide (subtidal) 0.523 <0.0001s(estuarine area) 1.67 <0.0001s(% intertidal Area) 0.96 <0.0001

Shannon’sdiversity (H)

Salinity (mesohaline) -0.732 <0.0001Salinity (oligohaline) -1.632 <0.0001Salinity (polyhaline) -0.853 <0.0001Tide (subtidal) Not sig. Not sig.s(estuarine area) 0.99s(% intertidal Area) Not sig. Not sig.

33

Table 4. Single metrics (richness and diversity) observed at present conditions and

predicted by the statistical model at present and under hydrographic and

geomorphological conditions of 1874, in the Oiartzun Heavily Modified Water Body.

Sampling station Observedpresent

Estimatedpresent

Predictedfor 1874

Species richness

8 41.1 30.2 24.521 32.3 30.2 8.318 38.0 30.2 10.116 25.3 30.2 10.124 21.7 21.1 10.114 47.9 30.2 24.5Mean 34.4 28.7 14.6

Shannon’s diversity(H)

8 4.18 2.67 2.6721 2.55 2.67 1.9418 3.31 2.67 1.8216 2.04 2.67 1.8224 1.61 1.82 1.8214 3.35 2.67 2.67Mean 2.84 2.53 2.12

34

Table 5. Maximum ecological potential (MEP) for single benthic metrics (richness and

diversity), predicted for 1874 and estimated for 2009, in the Oiartzun Heavily Modified

Water Body, compared with present natural reference conditions used in the Basque

Country (Muxika et al., 2007); this is applied in different salinity stretches and

asssociated benthic communities. Key: SD- Standard Deviation.

Salinity stretch Associated community MEP predictedfor 1874

MEP estimatedat present

Present naturalreferenceconditions

Presentproposed

MEPRichness

Oligo/mesohaline Cerastoderma edule/Scrobicularia plana 9.2 (SD: 1.27) - 13 10Polyhaline Venus fasciata 14.9 (SD: 8.31) 21.1 32 24Euhaline Abra alba 24.5 30.2 (SD: 0) 40 30

Shannon’s diversityOligo/mesohaline Cerastoderma edule/Scrobicularia plana 1.88 (SD: 0.08) - 2.5 1.9Polyhaline Venus fasciata 2.10 (SD: 0.49) 1.82 3.8 2.8Euhaline Abra alba 2.67 2.67 (SD: 0) 3.5 2.7