Recent habitat degradation in karstic Lake Uluabat, western Turkey: A coupled...

B I O L O G I C A L C O N S E R V A T I O N 1 4 1 ( 2 0 0 8 ) 2 7 6 5 – 2 7 8 3

. sc iencedi rec t . com
ava i lab le a t www
journal homepage: www.elsevier .com/ locate /b iocon

Recent habitat degradation in karstic Lake Uluabat, westernTurkey: A coupled limnological–palaeolimnological approach

Jane M. Reeda,*, Melanie J. Lengb,c, Sandra Ryana,1, Stuart Blackd,Selcuk Altinsaclie,2, Huw I. Griffithsa,3

aDepartment of Geography, University of Hull, Cottingham Road, Hull HU6 7RX, UKbNERC Isotope Geosciences Laboratory, British Geological Survey, Keyworth, Nottingham NG12 5GG, UKcSchool of Geography, University of Nottingham, Nottingham NG7 2RD, UKdSchool of Human and Environmental Science, University of Reading, Whiteknights, P.O. Box 227, Reading RG6 6AB, UKeDepartment of Biology, University of Istanbul, Istanbul, Turkey

A R T I C L E I N F O

Article history:

Received 13 November 2007

Received in revised form

1 August 2008

Accepted 11 August 2008

Available online 21 September 2008

Keywords:

Biomonitoring

Palaeolimnology

Diatoms

Ostracods

Isotopes

Eutrophication

0006-3207/$ - see front matter � 2008 Elsevidoi:10.1016/j.biocon.2008.08.012

* Corresponding author: Tel.: +44 1 482 466 0E-mail addresses: [email protected] (

reading.ac.uk (S. Black), [email protected] Present address: Entec, 155 Aztec West, P2 Present address: Merdivenkoy Mahallesi3 Deceased.

A B S T R A C T

The Ramsar site of Lake Uluabat, western Turkey, suffers from eutrophication, urban and

industrial pollution and water abstraction, and its water levels are managed artificially.

Here we combine monitoring and palaeolimnological techniques to investigate spatial

and temporal limnological variability and ecosystem impact, using an ostracod and mol-

lusc survey to strengthen interpretation of the fossil record. A combination of low inverte-

brate Biological Monitoring Working Party scores (<10) and the dominance of eutrophic

diatoms in the modern lake confirms its poor ecological status. Palaeolimnological analysis

of recent (last >200 yr) changes in organic and carbonate content, diatoms, stable isotopes,

ostracods and molluscs in a lake sediment core (UL20A) indicates a 20th century trend

towards increased sediment accumulation rates and eutrophication which was probably

initiated by deforestation and agriculture. The most marked ecological shift occurs in the

early 1960s, however. A subtle rise in diatom-inferred total phosphorus and an inferred

reduction in submerged aquatic macrophyte cover accompanies a major increase in sedi-

ment accumulation rate. An associated marked shift in ostracod stable isotope data indic-

ative of reduced seasonality and a change in hydrological input suggests major impact

from artificial water management practices, all of which appears to have culminated in

the sustained loss of submerged macrophytes since 2000. The study indicates it is vital

to take both land-use and water management practices into account in devising restoration

strategies. In a wider context, the results have important implications for the conservation

of shallow karstic lakes, the functioning of which is still poorly understood.

� 2008 Elsevier Ltd. All rights reserved.

er Ltd. All rights reserved.

61; fax: +44 1 482 466 340.J.M. Reed), [email protected] (M.J. Leng), [email protected] (S. Ryan), s.black@

(S. Altinsacli).

ark Avenue, Almondsbury, Bristol BS32 4UB, UK.

Ortabahar Sokak, Ozdin Apartımanı No. 20, Daire 4, Istanbul, Turkey.

mailto:[email protected]



mailto:s.black@ reading.ac.uk

mailto:s.black@ reading.ac.uk


2766 B I O L O G I C A L C O N S E R V A T I O N 1 4 1 ( 2 0 0 8 ) 2 7 6 5 – 2 7 8 3

1. Introduction

Concerns over the degradation of inland water quality and

aquatic habitats have become key environmental issues. Over

much of Europe, for example, the dictates of the EU Water

Framework Directive (WFD; EU Directive 2000/60/EC) have

led to the development of internationally coherent water

chemistry and biomonitoring programmes. These are sup-

ported by a growing body of research into the pattern and pro-

cess of biological response, aimed in part at predicting the

‘undisturbed’ reference state as a target for restoration of sus-

tainable good ecological (e.g. Freshwater Biological Associa-

tion, 2005).

Turkish water resources have high economic and biodiver-

sity value, but many lakes are suffering pollution impact (e.g.

Green et al., 1996; Roberts and Reed, in press) and long-term

monitoring programmes are not well developed. Given Tur-

key’s desire to become an EU member state, there is addi-

tional pressure to achieve good ecological and chemical

status. Recent short-term studies of lakes, streams and

groundwater indicate widespread problems of eutrophication

(e.g. Akkoyunlu, 2003; Beklioglu et al., 2003; Karakoc et al.,

2003; Karafistan and Arık-Colakoglu, 2005; Dalkıran et al.,

2006; Dere et al., 2006; Duran, 2006; Aksoy and Scheytt, 2007;

Duran and Suicmez, 2007), pesticide and heavy metal con-

tamination (e.g. Arcak et al., 2000; Barlas et al., 2005; Kazancı

et al., 2006; Demirel, 2007; Yılmaz, 2007), and hydrological im-

pact (Magnin and Yarar, 1997; Lammens and van den Berg,

2001). A variety of potential causes can be identified, promi-

nent amongst which are those associated with the expansion

of agriculture, industry and urban populations. Restoration

efforts include measures such as sewage diversion (e.g. Kara-

koc et al., 2003), biomanipulation (e.g. Beklioglu et al., 2003)

and socio-economic studies (e.g. Ozesmi and Ozesmi, 2003).

However, the definition of successful sustainable manage-

ment or restoration plans is hindered by a lack of detailed

understanding of the nature of environmental impact and

ecosystem response (Karakoc et al., 2003; Ozesmi and Oze-

smi, 2003).

The potential of palaeolimnological research to recon-

struct water pollution impact in lakes is well demonstrated

(e.g. Battarbee, 1999; Smol, 2002). For countries such as Tur-

key, which lack large-scale monitoring programmes, palaeo-

limnology offers the only reliable means of establishing the

timing and magnitude of ecosystem impact. Palaeolimnolog-

ical data can be of particular value in contributing informa-

tion on the reference state of a lake (e.g. Bennion et al.,

2004; Leira et al., 2006; Bennion and Battarbee, 2007). Interna-

tionally, karstic tectonic systems have received less attention

than the glacial lakes of northwestern Europe, and such stud-

ies should benefit environmental managers throughout the

circum-Mediterranean.

To date, palaeolimnological studies in Turkey have had fo-

cused on long-term palaeoclimate and vegetation history

(Van Zeist and Bottema, 1982; Bottema and Woldring, 1984;

Landmann et al., 1996; Eastwood et al., 1999, 2007; Leng

et al., 1999; Reed et al., 1999; Roberts et al., 1999; Kashima,

2002; Jones et al., 2002, 2006, 2007). More recently, this has ex-

tended to recognising signatures of volcanic eruptions (East-

wood et al., 2002) and seismic activity (Leroy et al., 2002).

Apart from a study of mineralogy, geochemistry and basic

diatom and ostracod analysis in Lake _Iznik, western Turkey

(Franz et al., 2006), and of sediment accumulation rates in

Lake Uluabat itself (Kazancı et al., 2004), conservation-based

palaeolimnology is lacking.

Here we combine a range of monitoring and palaeolimno-

logical techniques to investigate spatial and temporal limno-

logical variability and ecosystem change in an internationally

important wetland in northwestern Turkey. In the modern

environment, standard invertebrate biomonitoring tech-

niques are used to assess current ecosystem health, and a

survey of ostracod and mollusc distribution is used to

strengthen palaeolimnological interpretation of fossil assem-

blages. In line with the approach taken in large-scale monitor-

ing programmes (Bennion and Battarbee, 2007), our

palaeolimnological study aims to establish the degree of re-

cent (last ca. 200 yr) human impact on the physico-chemical

and ecological character of the lake and to assess how far

the ecosystem may have shifted from a former reference

state. The focus on recent change assumes a degree of long-

term stability, such that the 19th century character of the lake

is taken as its ‘natural’ state prior to expansion of human

activities, and marked change thereafter is likely to represent

a deviation from the norm. We combine a well-tested range of

proxy indicators (sediment organic and carbonate content,

diatoms, ostracods and molluscs, authigenic and ostracod

oxygen and carbon isotopes) to reconstruct the timing and

magnitude of recent limnological change by analysis of a

short sediment core, dated by 210Pb and 137Cs analysis of a

parallel core at the same site.

2. The study site

The Ramsar site, Lake Uluabat (or Lake Ulubat, Apolyont

Golu), is shallow (currently <2 m deep) and eutrophic. It is lo-

cated south of the Sea of Marmara at 8 m a.s.l. in Bursa,

northwestern Turkey (40�10 0N, 28�35 0E) (Fig. 1). The climate

is Mediterranean-continental, with long hot summers and

cold winters. Temperatures range from �16 to +40 �C and

mean annual precipitation is 668 mm. As with Lake Manyas

(or Kus�) to the west, it is a karstic tectonic basin within Meso-

zoic limestones which was probably formed by the damming

of a Holocene river channel (Kazancı et al., 2004). It is ca.

24 km long and 12 km wide, with an estimated catchment

area of 10,555 km2 (Gurluk and Rehber, 2006). The lake is

groundwater-fed, with additional inflow through the River

Mustafakemalpas�a and minor drainage channels. The only

outflow is the Kocacay River in the north-west; during periods

of drought its flow has been known to reverse, forming a tem-

porarily closed hydrological system.

The Uluabat catchment is in one of the most productive

agricultural regions of Turkey, with approximately 16 urban

settlements on the lake shores (Dalkıran et al., 2006), the

largest of which are shown in Fig. 1. Most of the catchment

is dedicated to arable farming and willow plantations, with

some fruit plantations and stock breeding. During 1937–

1993, 148 km2 of natural floodplain was drained for agricul-

ture, with the progressive construction along the western

littoral and the lower reaches of the River Mastafakemalpas�a

of artificial silt embankments (Magnin and Yarar, 1997).

Fig. 1 – Map of Lake Uluabat, showing its location in western Turkey (inset), and modern land-use in the catchment.

B I O L O G I C A L C O N S E R V A T I O N 1 4 1 ( 2 0 0 8 ) 2 7 6 5 – 2 7 8 3 2767

Some remnant natural sand banks along the western littoral

are populated by Tamarix scrub. Reed beds dominated by

Phragmites australis and Typha spp. are extensive along the

western and southern shores, and occur in patches to the

north and east.

Lake Uluabat has high economic and biodiversity value

(Ozesmi and Ozesmi, 2003). The fisheries industry is of na-

tional importance, dominated by pike, Israeli carp (Tilapia)

and smaller fish such as rudd and roach (Lammens and van

den Berg, 2001). The lake once provided 30% of national Turk-

ish crayfish production (Astacus leptodactylus) until they were

almost extirpated by a disease in 1986. Until recently Lake

Uluabat boasted the largest beds of the water lily (Nymphaea

alba) in Turkey, with a diverse aquatic flora including Cerato-

phyllum demersum and Potamogeton crispus (Secmen and Leb-

lebici, 1996), but plant macrophyte cover has declined

seriously since 2001. In contrast, migratory bird populations

have increased significantly, apparently linked to declining

water quality in Lake Manyas (Magnin and Yarar, 1997). Lake

Uluabat supports over 400,000 water birds, including the

endangered Pelecanus crispus.

As outlined below, Lake Uluabat is under pressure from

eutrophication, industrial and urban pollution and water

abstraction, and its naturally fluctuating water levels are

now managed artificially. The lake was unprotected until its

listing as a Ramsar site in 1998, and its biodiversity value is

also reflected in its inclusion, in 2000, in the Living Lakes Net-

work (http://www.livinglakes.org/partnership.htm). An eco-

system management plan has been set up (Ozesmi and

Ozesmi, 2003; Salihoglu and Karaer, 2004; Gurluk and Rehber,

2006), but management attempts are hampered by the need

for additional limnological monitoring and deeper under-

standing of ecosystem functioning (Degirminci et al., 2005;

Gurluk and Rehber, 2006).

3. Current knowledge and uncertainties

Monitoring programmes have been of limited value in assess-

ing temporal trends and specific impacts since the majority

have generated a maximum of 1–2 yr data. Degirminci et al.

(2005) noted that there was no system for monitoring external

impacts. The lake is eutrophic; total phosphorus data are

lacking, but maximum mean chlorophyll a (52.0 mg m�3), ni-

trate (5.5 mg l�1 N-NO3) and soluble reactive phosphorus

(0.1 mg l�1) values for 1998–1999 are high (Dalkıran et al.,

2006). External impacts are poorly understood; diffuse agri-

cultural pollution may be an important source of nutrient in-

put (Magnin and Yarar, 1997) and, from a Holocene sediment

chronology, Kazancı et al. (2004) suggest increased 20th cen-

tury sediment accumulation rates are linked to deforestation

and the expansion of agriculture. In contrast, other authors

suggest that inflow from the R. Mustafakemalpas�a is the ma-

jor source of both nutrients and accelerated siltation (Lam-

mens and van den Berg, 2001; Ozesmi and Ozesmi, 2003;

Barlas et al., 2005). Lammens and van den Berg (2001) report

fluctuating nutrient levels linked to changing external river

input and reduced internal plant macrophyte cover compared

to a decade earlier. Other ecological studies are limited to

phytoplankton surveys (Dalkıran, 2000; Karacaoglu et al.,

2004), a survey of Ephemeroptera (Tanatmis�, 2002) and an

analysis of littoral ostracod communities linked to the pres-

ent study (Altınsaclı and Griffiths, 2001).

Industrial activity includes extraction and mining of sand,

lignite and heavy metals, which has been extensive

http://www.livinglakes.org/partnership.htm


historically along the R. Mustafakemalpas�a (Barlas et al.,

2005). Mining activities have reportedly reduced since the

1960s and since 1999 are reported to mainly comprise sand

extraction (Lammens and van den Berg, 2001). This is again

at odds with a statement by Kazancı et al. (2006), that open

cast mining and associated pollution are still intense. Urban

population data are lacking.

Water levels fluctuate naturally on an intra- and inter-an-

nual basis. Historical data are fragmentary and contradictory.

Most authors agree that a progressive reduction in lake vol-

ume has occurred over recent decades, possibly coupled with

reduced seasonal fluctuation in water depth. Water abstrac-

tion is via four pumping stations on the shores (Fig. 1). Lam-

mens and van den Berg (2001) suggest there has been a 1 m

rise in minimum water-level since 1989 due to construction

works on the outlet. In 1997, abstraction directly from the lake

was for small-scale irrigation of 64 km2 of land, but the inflow

was being reduced by river abstraction for irrigation of

203 km2 of land (Magnin and Yarar, 1997). Reported maximum

depth has reduced from an estimated historic high of 10 m, to

6 m in 1997 (Magnin and Yarar, 1997) and recent estimates of

3.0–3.5 m in winter/spring and 1.0–1.5 m in summer (Lam-

mens and van den Berg, 2001; Karacaoglu et al., 2004; Barlas

et al., 2005, this study). From field observations in 2000, re-

duced water levels are evident in Golyazı, where a bridge con-

structed in the 1970s straddles dry ground. Kazancı et al.

(2004) estimate a ca. 7% reduction in surface-area from May

1987 to July 2000 based on Landsat images, although from

our observations this contrast will have been due in part to

seasonally low summer lake levels and to a drought in 2000.

Siltation is thought to be an additional contributor to a rather

larger estimate of a 12% reduction in lake area from 1984 to

1993 (Aksoy and Ozsoy, 2002 cited in Dalkıran et al., 2006).

The latter contradicts trends derived from basic State moni-

toring during 1982–2000 (Ryan, 2001; Beklioglu et al., 2006;

Dalkıran et al., 2006), which indicate a drop in mean water-le-

vel of ca. 3 m by 2000 from a peak in 1988, but also indicate

water-level was only ca. 0.5 m above the 2000 level in 1982.

4. Materials and methods

4.1. Modern limnology

Due to their specific ecological preferences, ostracods and

molluscs can provide valuable palaeolimnological proxy data

on parameters such as habitat availability, dissolved oxygen

content (DOC) and nutrient status, to complement inferences

based on other proxies such as diatoms (Griffiths and Holmes,

2000; Griffiths et al., 2002). To strengthen interpretation of fos-

sil assemblages based on their observed modern habitat pref-

erences, a transect survey was carried out of species

distribution in profundal habitats of the lake and in the open

waters of the littoral zone. This complemented ostracod data

from littoral habitats generated by Altınsaclı and Griffiths

(2001) during 1995–1996, when the lake had extensive sub-

merged macrophyte cover. Samples were collected along

two 10-sample transects (Fig. 2) on 14 April, 2001, using a Gar-

min GPS to record the location. Transect A sampled the open

waters at the centre of the lake basin, running towards the

putative point-source of pollution, the R. Mustafakemalpas�a,

and Transect B ran southwards from reed beds in the north-

east. Conductivity, pH and DOC were measured with What-

man probes at 19 of the 20 sites. Water depth was measured

with a hand-held echo sounder, and transparency with a Sec-

chi disk.

Samples were collected using an Ekman Grab, sieved in the

field through a 63 lm sieve and preserved in 95% ethanol. In

the laboratory samples were washed through a 180 lm sieve

and hand picked while viewing at 10· on a Prior dissecting

microscope. Standard counting techniques were used (Mou-

thon, 1986; Altinsaclı and Griffiths, 2001). Identification was

made by reference to standard texts and local studies

(Bronshtein, 1947; Meisch, 2000 for ostracods; Fitter and Man-

uel, 1986; Kubilay and Timur, 1995 for molluscs). Candonid

ostracods were not identified to a lower level owing to difficul-

ties in identifying congeneric juveniles; from Altınsaclı and

Griffiths (2001), they may include benthic Candona angulata

and Candona neglecta, or limnic Fabaeformiscandona caudata.

The data were expressed as total number of individuals per

Ekman grab sample (a volume of ca. 25 · 25 · 25 cm3).

The same samples provided material for a simple inverte-

brate biomonitoring assessment of current water quality sta-

tus, by calculation of Biological Monitoring Working Party

(BMWP) scores to assess bulk organic pollution and/or eutro-

phication (Armitage et al., 1983). Samples were prepared as

above and identified using standard texts (Croft, 1986; Wil-

liams and Feltmate, 1992). This biotic index was developed

for running waters but is effective for lakes (e.g. Krno et al.,

2006). Modified versions such as the Spanish BMWP (Alba

Tercedor and Sanchez Ortega, 1988), may be more robust for

diverse faunas of carbonate-rich waters of the circum-Medi-

terranean, but diversity was sufficiently low, and representa-

tivity of groups high, that the original system was deemed

applicable. Rare occurrences of damsel flies (Odonata) were

not identified to family level. Of the three relevant families

in the scoring system, two (Platycnemidae and Coenagriidae)

command a score of six and one (Lestidae) scores eight. An

approximate score of six was adopted.

Spatial variation in invertebrate distribution was explored

using principal components analysis (PCA), which assumes a

linear species response and is appropriate for small ecological

data-sets (Jongman et al., 1995). In the absence of a larger

environmental data-set, direct gradient analysis (redundancy

analysis) was not appropriate. Since each sample represented

organisms preserved in a ca. 25 cm-depth sediment sample,

and some organisms (ostracods, molluscs, Chironomidae)

preserve in sediment, whereas others (e.g. Odonata, Hemip-

tera) do not, the two groups are not equivalent numerically.

Rather than transform species data to proportions, the analy-

sis was performed on abundance data, transformed to Cornell

format using C2 v. 1.4.2 (Juggins, unpublished), and PCA was

performed using Canoco v. 3.14 (ter Braak, 1990). Results were

displayed graphically using C2.

4.2. Palaeolimnology

4.2.1. Collection of sediment coresA coring location was selected away from the immediate influ-

ence of the R. Mustafakemalpas�a, with the aims of (1) obtain-

ing a sufficiently long chronological record with a short core

Fig. 2 – Map showing the location of transect sites A and B. Inset A: location of sampling sites within Transect A. Inset B:

location of sampling sites within Transect B, and location of the coring site for UL20A and UL20B.


(with the reported local siltation impact offshore from the del-

ta), and (2) assessing lake-wide water quality status in an off-

shore location which was not reportedly subject to maximum

impact (again, from the river). Co-ordinates of the sampling

site were recorded using a Garmin GPS. Two cores, UL20A

(32 cm) and UL20B (21 cm) were collected from within 1 m

apart on 9 December 2000 (40�11 00700N, 28�40 02000E) in 2 m of

water to the northeast of the lake basin, using a Glew gravity

corer (Glew, 1991) (Fig. 2). Both cores retrieved an undisturbed

sediment–water interface. Sediment cores were described and

extruded in the field into sterile plastic bags, subsampling at

0.5 cm resolution for the top 1 cm, and 1 cm thereafter. In

the laboratory sediment samples were stored at 4 �C. The long-

er UL20A was selected as the master core and UL20B was re-

tained for 210Pb and 137Cs dating.

4.2.2. Basic sediment propertiesSubsamples of ca. 1–2 g were taken at 1 cm intervals for mea-

surement of percentage organic and carbonate content, by

oven drying overnight at 50 �C, and loss on ignition at 550 �C

and 925 �C, respectively, using standard techniques (Dean,

1974). Following calculation of organic and carbonate content,

the additional parameter of carbonate-free organic carbon

was calculated as the proportion of weight loss at 550 �C, tak-

ing 100% as the sum of organic + final residue weight after

ignition at 925 �C.

4.2.3. Diatom analysisSubsamples of ca. 0.1 g were prepared using standard tech-

niques (Battarbee, 1986) with hot H2O2 and HCl to oxidise or-

ganic matter and remove carbonates, respectively, and using

Naphrax� as a slide mountant. A resolution of 2 cm was ap-

plied initially, counting the upper sequence at 1 cm thereafter

to improve interpretation of recent change. Where diatoms

were well preserved, ca. 500 valves per slide were counted un-

der oil immersion with phase contrast at a magnification of

1000·, on a Leica DMLB light microscope. Counts were re-

duced in some samples due to poor preservation. Taxonomy

and nomenclature follows Krammer and Lange-Bertalot

(1986, 1988, 1991a,b).


Diatom-inferred total phosphorus (DI-TP) reconstruction

was performed by classic weighted averaging regression and

calibration, using C2, and based on the Combined European

TP training set (Battarbee et al., 2000). The training set was

collected largely from temperate Western Europe: shallow,

hypereutrophic British and Scandinavian lakes (Anderson

et al., 1993; Bennion, 1994; Bennion et al., 1996; Lotter et al.,

1998), French crater lakes (Rioual, 2000) and central Alpine oli-

gotrophic–mesotrophic lakes (Wunsam and Schmidt, 1995).

There is not a more local TP training set. Results were dis-

played using Tilia, Tiliagraph and TGView (Grimm, 1991).

Stratigraphic zone boundaries were defined using constrained

incremental sum of squares, on square-root transformed

data, using the program CONISS (Grimm, 1987).

4.2.4. Ostracod and mollusc analysisSediment subsamples were taken at 1 cm intervals and pre-

pared using standard techniques of drying and subsequent

wet sieving at 180 and 63 lm (Griffiths and Holmes, 2000).

Samples were hand picked while viewed at 10· on a Prior bin-

ocular microscope and mounted on micropalaeontological

slides using gum tragacanth; larger gastropods were stored

in glass tubes. Adult ostracods were identified and counted

separately from juveniles. Abundance was expressed as indi-

viduals per 100 g dry sediment.

4.2.5. Stable isotope analysisOxygen and carbon stable isotope ratios were measured on

the authigenic (<80 lm fraction) carbonate fraction and on

biogenic carbonate (ostracod valves). For authigenic carbon-

ate, 1–2 g bulk sediment was subsampled at each level of

UL20A and disaggregated in 5% sodium hypochlorite solu-

tion (10% chlorox) for 24 h to oxidise reactive organic mate-

rial. Samples were then sieved at 80 lm to remove biogenic

carbonates. The <80 lm fraction was washed with deionised

water, dried at 40 �C and ground in agate. The authigenic

carbonate was reacted with anhydrous phosphoric acid in

vacuo overnight at a constant 25 �C. The CO2 liberated was

separated from water vapour under vacuum and collected

for isotope analysis. Measurements were made on a VG Op-

tima dual inlet mass spectrometer. Overall analytical repro-

ducibility was >0.1& for d13C and d18O (2r). Isotope values

(d13C, d18O) are reported as permil (&) deviations of the iso-

topic ratios (13C/12C, 18O/16O) calculated to the VPDB scale

using a within-run laboratory standard calibrated against

NBS standards.

To avoid within- and between-species variability, analysis

of ostracod carbonate was based on up to 10 valves per sam-

ple of the dominant species, Physocypria kraepelini, cleaned

with deionised water using a 0000 paintbrush. The species

was present throughout the sequence, is relatively short

lived, and is nektonic, that is, living in the water column; nek-

tonic taxa should respond more directly to hydrologically

mediated isotopic changes than bottom dwellers. Individuals

comprising ca. 60 lg of ostracod carbonate were analysed

using an automated common acid bath VG Isocarb + Optima

dual inlet mass spectrometer. Where sufficient numbers of

adults were preserved in a subsample, analysis was per-

formed on several individuals to estimate the mean and stan-

dard deviation. Isotope values (d13C, d18O) are reported as

permil (&) deviations of the isotopic ratios (13C/12C, 18O/16O)

calculated to the VPDB scale using a within-run laboratory

standard calibrated against NBS standards. Analytical repro-

ducibility is typically <0.1& for both d13C and d18O (based on

similarly sized laboratory standards).

4.2.6. 210Pb and 137Cs radiometric countingSubsamples from UL20B were weighed and counted on a Har-

well Instruments, Broad Energy (BE5030) high purity germa-

nium coaxial photon detector. This detector has an ultra

low background set up (detector and cryostat) with a

0.5 mm thick carbon-epoxy window and remote detector

chamber. Detector specifications were full width at half max-

imum (‘FWHM’) @ 5.9 keV = 0.45 keV, FWHM @ 1.3 MeV =

<1.2 keV. To keep self-absorption differences negligible, stan-

dard samples were used to calibrate the detectors using mate-

rial of similar density and dimensions to the sediment

analysed. The NIST standard reference material, SRM4357

was spiked with NIST SRM’s 3164 and 3159, which was config-

ured into the same geometry. A secondary standard was also

made in the form of a disc (80 mm diameter) from the same

material to which the detectors had been calibrated

previously.

The 210Pb/226Ra activity ratio was determined from the

activities at the 46.5 and 186.3 keV gamma-ray peaks, the lat-

ter following correction methods in De Corte et al. (2005). The

activity of the 210Pb/214Pb, using the 46.5 keV/295 and 352 keV

gamma-rays for 214Pb, and the 210Pb/214Bi ratios using the

46.5 keV/609 and 1764 keV gamma-rays for 214Bi, were also

determined. All the ratios were then averaged. The 137Cs

activity was determined using the 661 keV peak. Samples

were counted for approximately 7–10 days each in order to re-

duce the uncertainties by accumulating a large number of

counts in each analyte peak. Most analyte peaks were

>10,000 net counts (i.e. <1% uncertainty). External reproduc-

ibility was checked using international standards (NIST-SRM

4357) and was found to be within 0.76%. Ages were obtained

using the constant rate of supply (‘CRS’) model (Ivanovich

and Harmon, 1992).

5. Results

5.1. Limnological survey

From the basic water chemistry and physical limnology (Table

1), pH values were high (>8), with only minor variation be-

tween samples. Transparency was relatively low (50–

115 cm), with minimum values in samples A1–2 and B1. The

shallower-water samples A1–A5, closest to the R. Musta-

fakemalpas�a inflow (Fig. 2), had slightly lower dissolved oxy-

gen content (DOC; range 5.6–8.2 mg l�1; water depth 60–

110 cm), compared to DOC values of 8.0–10.8 mg l�1 and water

depth of 125–170 cm in A6–A10 and B1–10. The deepest waters

were located close to the coring site at Golyazı, but on water

chemistry alone the two sampling zones showed marked

similarities.

Invertebrate results are given in Table 2 and results of PCA

in Fig. 3. PCA Axes 1 and 2 together explain a relatively high

percentage of total variance in the species data (49.5%).

Fig. 3 shows clear separation of samples from the two tran-

Table 1 – Table showing the site co-ordinates and basic limnological parameters of transect samples for the moderninvertebrate, ostracod and mollusc survey

Sample Site location Waterdepth (cm)

Watertemperature (�C)

pH Conductivity(lS cm�1)

Dissolved oxygencontent (mg l�1)

Oxygensaturation (%)

Secchi discdepth (cm)

A1 40�07 04700N; 28�35 01800E 60 21 8.2 533 5.6 61.0 50

A2 40�07 05200N; 28�35 02200E

A3 40�07 05700N; 28�35 02600E 100 23 8.3 561 7.5 91.1 60

A4 40�07 06000N; 28�35 02900E 110 26 8.3 544 6.6 80.2 90

A5 40�08 00300N; 28�35 02600E 110 24 8.4 546 8.2 95.6 100

A6 40�08 00600N; 28�35 03500E 125 23 8.4 540 7.8 91.3 100

A7 40�08 00900N; 28�35 03300E 130 25 8.4 527 9.8 111.7 120

A8 40�08 01100N; 28�35 03300E 150 25 8.4 553 8.7 105.0 140

A9 40�08 01600N; 28�35 03900E 140 24 8.5 548 10.8 119.3 100

A10 40�08 01900N; 28�35 04400E 150 24 8.5 533 9.5 110.8 115

B1 40�10 04900N; 28�39 05800E 140 22 8.6 550 9.6 104.0 60

B2 40�10 04900N; 28�40 00100E 140 22 8.7 553 9.2 101.0 80

B3 40�10 04200N; 28�40 00600E 150 22 8.7 535 10.4 121.0 80

B4 40�10 03700N; 28�40 00700E 160 22 8.7 530 9.3 109.2 90

B5 40�10 03100N; 28�40 01000E 170 22 8.6 528 8.7 96.6 80

B6 40�10 02300N; 28�40 01300E 170 22 8.6 526 8.5 91.4 85

B7 40�10 01800N; 28�40 01500E 160 22 8.6 528 8.3 96.6 80

B8 40�10 01100N; 28�40 01700E 170 22 8.6 529 7.9 90.8 100

B9 40�10 00500N; 28�40 01800E 170 22 8.6 523 9.0 105.6 100

B10 40�10 00000N; 28�40 02100E 170 22 8.7 522 9.0 104.4 110

Table 2 – Table showing invertebrate abundance (per Ekman grab sample; ca. 253 cm3) and BMWP scores for Transect Aand B samples

Sample

A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 B1 B2 B3 B4 B5 B6 B7 B8 B9 B10

Crustacea

Ostracoda

Physocypria kraepelini 0 15 12 25 10 41 22 35 40 125 519 279 554 222 168 102 113 36 38 36

Darwinula stevensoni 0 2 1 14 3 16 40 23 91 1 0 0 1 1 5 46 0 0 11

Ilyocypris gibba 0 7 3 5 5 9 3 5 8 21 1 0 0 0 3 2 26 10 7 19

Candona spp. (adult) 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0

Candona spp. (juvenile) 0 0 2 0 0 3 1 2 13 77 75 105 41 38 61 53 50 74 102

Total Ostracoda 0 24 19 44 15 56 41 81 73 250 598 354 659 264 210 170 238 97 119 168

Copepoda 0 32 83 54 7 243 46 92 21 152 3 11 14 3 1 4 3 9 2 21

Cladocera 0 40 16 18 15 61 4 1 0 8 6 1 25 0 0 3 0 0 0 0

Insecta

Chironomidae 0 21 10 15 21 17 14 17 17 0 48 20 28 17 31 12 54 9 19 24

Odonata (damsel fly larva) 1 0 0 0 0 0 0 0 0 0 118 0 0 0 0 0 0 0 0 0

Hemiptera 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Notonecta 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

ACARI 0 0 0 0 8 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0

Gastropoda

Bithynia spp. 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0

Valvata piscinalis 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0

Lymnaea peregra 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0

Bivalvia

Unio spp. 0 1 3 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Dreissena polymorpha 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0

Oligochaeta 27 49 22 36 9 302 30 35 31 4 15 14 101 1 2 26 520 0 12 173

Nematoda 12 101 36 21 16 32 66 114 122 268 156 204 130 404 141 309 312 123 233 128

BMWP score 7 9 9 3 14 3 3 3 3 1 15 3 3 3 3 3 3 3 3 3


-1.0 0.0 1.0 2.0 3.0 4.0 5.0-1.0

0.0

1.0

2.0

3.0

4.0

5.0

Axis

2

PhysocypriakraepeliniCandona spp.

Copepoda

Chironomidae

Odonata

Nematoda

Oligochaeta

-1.0 0.0 1.0 2.0-1.0

0.0

1.0

2.0

Axis 1 Axis 1

Axis

2

A1

A2

A3

A4A5

A6

A7 A8A9 A10

B1B2

B3B4

B5

B6

B7

B8

B9

B10

Fig. 3 – Scatter plots of Axis 2 against Axis 1 showing the distribution of (a) transect samples and (b) invertebrate taxa in

principal components analysis (k1 = 0.28; k2 = 0.22; cumulative percentage variance of the species data for Axis 1 plus Axis

2 = 27.2%).


sects in the direction of Axis 1. This is driven by high abun-

dance of ostracods (P. kraepelini and candonids) and nematode

worms in Transect B, vs. higher abundance of zooplanktonic

copepods in Transect A. Zooplanktonic cladocera were not

abundant, but were also more common in Transect A. Rarer

occurrences of other groups such as Odonata (damsel flies)

and gastropods did not exert a significant influence on PCA

results. It is noteworthy, however, that sample B1, which

along with B3 has maximum scores on Axis 1, was the most

0

2

4

6

8

10

12

14

16

A1 A2 A3 A4 A5 A6 A7 A8 A9 A10

BMW

P sc

ore

Transect A

0

2

4

6

8

10

12

14

16

B1 B2 B3 B4 B5 B6 B7 B8 B9 B10

BMW

P sc

ore

Transect B

Fig. 4 – Graphs of invertebrate BMWP scores for Transect A

and B samples.

diverse assemblage of the data-set (Table 2); this was the only

sample from reed beds.

BMWP scores (Fig. 4) were low. Only two samples scored

>10 (A5 and B1). The only common groups were the low scor-

ing, pollution-tolerant Chironomidae and Oligochaeta. BMWP

has the weakness that it does not take abundance into ac-

count. Apart from sample B1, the higher scores were due to

the presence of sensitive taxa (e.g. Odonata) at very low

abundance.

5.2. Palaeolimnology

5.2.1. Radiometric datingResults of 210Pb and 137Cs dating (Table 3 and Fig. 5) indicated

that the sediment record extends back prior to at least the

start of the 19th century. The estimated 210Pb age of 1964 ± 6

corresponds well with the peak in 137Cs values at 12.5 cm

depth, indicative of the peak in weapons testing in 1963.

Fig. 5a shows a clear discontinuity in 210Pb values between

samples 15–16 cm and 14–15 cm. This reflects a major shift

in sediment accumulation rate, to more rapid accumulation

(0.26 kg m�2 yr�1) after the 1950s (Fig. 5b). Both sets of results

indicated some reworking of sediments from the 1960s in the

top 2 cm, with anomalously low 210Pb and high 137Cs values,

respectively. Sediment disturbance was not observed during

coring, and (see Diatom Analysis) the surface sediment of

the master core was intact. Recent sediments may have been

affected by the location of Lake Uluabat close to the epicentre

of the 1999 _Izmit earthquake (Homan and Eastwood, 2001).

5.2.2. Basic sediment propertiesThe two cores, UL20A and UL20B, were of homogeneous

brown silty clay with no major lithostratigraphic boundaries.

Both showed a trend towards less compaction in the upper

sequence, with compact sediment from the core base to ca.

22 cm depth and soft, unconsolidated sediment above ca.

5 cm (ca. 1987).

Loss-on-ignition results for UL20A (Fig. 6) show a clear

trend towards increasing organic content from low values of

Table 3 – Table showing the results of 210Pb and 137Cs radiometric analysis of UL20B

Depth (cm) Unsupported 210Pb(Bq kg)

Unsupported 210Pb(Bq kg)a

137Cs(Bq kg)

137Cs(Bq kg)a

Estimated yearof deposition

Year ofdepositiona

lnUnsupported 210Pb(Bq kg)

0.25 34.91 2.09 39.70 1.82 1999.66 5.91 3.55

0.75 16.51 1.42 37.21 2.03 1999.42 4.06 2.80

1.50 99.20 5.14 35.87 0.85 1997.76 9.96 4.60

2.50 85.98 6.76 33.67 1.20 1995.57 9.27 4.45

3.50 101.30 7.84 31.20 1.50 1992.66 10.06 4.62

4.50 69.90 4.63 31.82 0.91 1989.33 8.36 4.25

5.50 69.07 6.67 41.86 1.73 1986.74 8.31 4.24

6.50 54.27 3.67 45.80 1.41 1983.55 7.37 3.99

7.50 44.96 3.83 47.17 1.61 1980.79 6.71 3.81

8.50 38.61 2.90 55.74 1.78 1978.27 6.21 3.65

9.50 34.25 2.15 55.75 1.43 1974.36 5.85 3.53

10.50 34.07 3.02 59.60 2.98 1973.30 5.84 3.53

11.50 34.36 2.56 60.27 1.58 1968.03 5.86 3.54

12.50 31.68 3.44 72.97 1.88 1964.09 5.63 3.46

13.50 21.51 7.52 62.62 3.53 1959.84 7.64 3.07

14.50 20.59 6.50 49.47 1.37 1955.60 8.54 3.02

15.50 48.17 15.47 52.82 1.42 1943.25 9.94 3.87

16.50 45.02 14.60 49.02 1.27 1927.18 14.71 3.81

17.50 22.65 7.83 45.31 1.22 1911.38 13.76 3.12

18.50 12.96 7.26 43.63 1.19 1892.07 23.60 2.56

19.50 12.40 7.03 48.70 1.24 1867.56 33.52 2.52

20.50 5.89 3.81 37.15 0.88 1692.02 92.43 1.77

a Standard error.

1820

1840

1860

1880

1900

1920

1940

1960

1980

2000

2020

0 5 10 15 20 25Depth (cm)

Estim

ated

yea

r of D

epos

ition

0.0

1.0

2.0

3.0

4.0

5.0

0 5 10 15 20 25Depth (cm)

Ln21

0 Pb

(Bq

kg)

Middle = 0.26 kg m-2 a-1

Base = 0.07 kg m-2 a-1

Fig. 5 – Graph of unsupported ln210Pb against depth for core

UL20B (a) and of estimated age against depth for (b),

showing the location of the inferred change in sediment

accumulation rate (dashed line).


8–10% from the core base to 13 cm depth (ca. 1962), and rising

thereafter to a maximum of 17% at the sediment surface. The

CaCO3-free organic curve gives a better representation of

changes in the delivery of organic carbon to the sediment,

exhibiting a minor rising trend from ca. 24 cm depth (pre-

19th century) to 17 cm depth (1920s), and a consistent rise

thereafter from 12% to 29% at the sediment surface, which

also exhibits an acceleration above ca. 13 cm depth. The car-

bonate curve almost paralleled these trends, with relatively

stable values of ca. 15% from 32 to 21 cm, where compaction

was greatest, rising to a peak of 29% at 4 cm depth.

5.2.3. Stable isotopesOxygen and carbon stable isotope values for bulk authigenic

carbonate and ostracod calcite (Fig. 6) are highly dissimilar

although this in part is a function of the different sampling

regimes. Authigenic carbonate values are relatively stable

and negative. d13Cauth varies in the range �4.0& to �6.6&

and d18Oauth in the range �6.8& to �8.1&.

The ostracod, P. kraepelini, was sufficiently well preserved

for stable isotope analysis of 11 sample levels above 21 cm

depth, but with a gap between 13 and 7 cm depth. At each le-

vel, 5–10 individual valves were analysed and the data are

shown as the mean and range. In contrast to the relatively

stable authigenic carbonate values, the ostracod data show

a marked shift at ca. 13 cm depth (ca. 1962). Values from the

base of the profile to ca. 13 cm depth are higher (d13Cost range

of mean values �0.9& to �3.6&; d18Oost range of mean values

�2.8& to �4.1&). There is a marked shift to more negative

values above ca. 13 cm (d13Cost range of mean values �0.9&

to �3.6&; d18Oost range of mean values �4.4& to �6.0&).

The error bars are also much wider for both isotopes in the

Dep

th (c

m)

10 200

5

10

15

20

25

30

35

30 10 20 30 -6.0 -8.0 -6.0 -4.0 -2.0 0.0

Loss on ignition (%)

organiccontent

non-carborganiccontent

carbonatecontent

Stable isotopes (‰ vs. VPDB)

18Oauth13Cauth

18Oost ( ) 13Cost ( )

ostracod carbonatebulk authigeniccarbonate

1956

20001998199619931989198719841981197819741973196819641960

194319271911189218681692

Estim

ated

dat

e

δ δ δ δ

Fig. 6 – Diagram comparing organic and carbonate content, bulk authigenic and ostracod stable isotope data in the sediment

core, UL20A. Organic content is expressed both as % loss on ignition (LOI) at 550 �C, and as a percentage recalculated by

exclusion of the % carbonate. % carbonate was calculated from LOI at 925 �C. Subscript ‘auth’ denotes bulk authigenic

carbonate; ‘ost’ denotes ostracod carbonate.


lower sequence, reflecting the greater variability of results for

analysis of individual valves than in the upper sequence.

Above ca. 13 cm depth, ostracod stable isotope values show

weak covariance, but there is no relationship below 13 cm

depth.

5.2.4. Diatom analysisOne hundred and thirty-nine diatom taxa were identified, of

which 95 were present in the training set for diatom-inferred

total phosphorus (DI-TP) reconstruction, allowing reconstruc-

tion based on >80% of taxa in each sample. The summary dia-

tom diagram (Fig. 7) shows lower total counts indicative of

poor preservation in subsamples 11–12 cm, 14–15 cm, 20–

21 cm and 24–25 cm, below which diatoms were either absent

or present as uncountable fragments.

Three biostratigraphic zones could be recognised using

CONISS. Diatom zone D1 (24.5–17.5 cm; pre-19th century to

ca. 1911) was dominated by the common epiphytes (Germain,

1981), Epithemia sorex, Epithemia adnata, Rhoicosphenia abbrevi-

ata and Rhopalodia gibba, along with the cosmopolitan taxa,

Cocconeis placentula and Amphora pediculus. The planktonic

taxon, Aulacoseira granulata, was common in two of the four

subsamples, a species characteristic of turbid, eutrophic

waters. The transition to zone D2 (17.5–11.0 cm; ca. 1911–

1970) was marked by increased relative abundance of eutro-

phic planktonic species, A. granulata, Stephanodiscus hantzschii,

Stephanodiscus medius, Cyclostephanos dubius and Cyclotella

meneghiniana. These increases were at the expense of the epi-

phytic species. Zone D3 (11.0–0.0 cm; ca. 1970–2000) was

marked by a further decline in relative abundance of Epith-

emia spp. and other epiphytes, with the consistent high rela-

tive abundance of the previous range of eutrophic planktonic

species, and a peak in DI-TP at 2.5 cm depth (ca. 1996). The rel-

ative abundance of A. pediculus also increased compared to

zone D1; this species is often found in open-water, plank-

ton-dominated samples (e.g. Roberts et al., 2001; Wilson

et al., 2008), presumably exploiting floating algae or plants

as a substrate, but is also a common coloniser (with C. placen-

tula) on P. australis in neighbouring Lake Manyas under condi-

tions of high turbidity and sediment load (Albay and

Akcaalan, 2003). A range of more fragile taxa including small

Navicula and Nitzschia spp. were also present consistently,

some of which (Nitzschia palea, Nitzschia hungarica, Navicula

cryptotenella) are known for their pollution tolerance. The

more fragile of these taxa may have been under-represented

in the lower sequence due to dissolution.

The surface sediment sample was separated out in CON-

ISS. This was probably because it contained a higher relative

proportion of live frustules than sediment samples below,

which in a well-mixed lake should integrate the diatom flora

of the lake as a whole. The absence of E. adnata in particular

indicates there was no reworking of sediment during coring,

and the master core was undisturbed.

The DI-TP reconstruction indicates that the lake has been

eutrophic (>100 lg l�1 TP) since prior to the 19th century. With

the increase in relative abundance of eutrophic taxa, the sub-

tle trend towards rising values towards the sediment surface,

in Zone D1, provides evidence for anthropogenic eutrophica-

tion since the start of the 20th century.

5.2.5. OstracodsOstracod results (Fig. 8) lacked evidence for major shifts in

species assemblage composition. The most marked biostrati-

graphic transition occurred at 6 cm depth (ca. 1985). Spanning

0

2

4

6

8

10

12

14

16

18

20

22

24

Dep

th (c

m)

20%

Aulaco

seira

gran

ulata

%

Stepha

nodis

cus h

antzs

chii

%

Stepha

nodis

cus m

edius

20%

Cyclos

tepha

nos d

ubius

%

Cyclot

ella m

eneg

hinian

a

20%

Epithe

mia so

rex

20%

Epithe

mia ad

nata

20%

Cocco

neis

place

ntula

%

Rhoico

sphe

nia ab

brevia

ta

%

Rhopa

lodia

gibba

%

Fragilar

ia uln

a

%

Fragila

ria ca

pucin

a var.

mes

olepta

%

Navicu

la rad

iosa

%

Navicu

la ca

pitora

diata

%

Navicu

la pla

centu

la

%

Gyrosig

ma acu

minatum

%

Gomph

onem

a trun

catum

%

Gomph

onem

a oliva

ceum

%

Cymato

pleura

solea

%

Nitzsc

hia am

phibi

a

20%

Amphora

pedic

ulus

%

Cymbe

lla lep

tocero

s

%

Amphora

libyc

a

%

Navicu

la cry

ptoten

ella

%

Amphora

vene

ta

%

Nitzsc

hia fo

ntico

la

%

Nitzsc

hia so

lita

%

Nitzsc

hia pa

lea

%

Nitzsc

hia in

cogn

ita

%

Nitzsc

hia ca

pitell

ata

20 40 60 80 100

Plankto

nic

Benthic

573599611535569815580569600

575382550

311

562

440

272

574

208

Count

50 100 150 200

DI-TP

D3

D2

D1

Planktonic Benthic

ConissCONISS

Total sum ofsquares

% μg l-11.0 2.0 3.0

Fig. 7 – Summary diatom diagram for UL20A showing all taxa present at P2% relative abundance, the proportion of

planktonic and benthic taxa, the diatom count per slide and diatom-inferred total phosphorus (DI-TP).


most of the sequence, Zone O1 (31.5–6.0 cm; pre-19th century

to ca. 1985) was dominated by Candona spp., with lower abun-

dance of P. kraepelini, Cypridopsis vidua and Ilyocypris gibba. The

only identifiable adult Candona specimens were of C. angulata.

Ostracods were present at very low abundance (<1 g�1), rising

slightly towards the upper zone boundary. The transition to

Zone O2 (6.0–0.0 cm; ca. 1985–2000) was marked by an in-

crease in the relative abundance of the limnic species, P. kra-

epelini, to >60% at the expense of benthic candonids, and the

disappearance of I. gibba. Ostracod abundance rose to maxi-

mum values of ca. 100 g�1 at 1–2 cm depth. Changes in abun-

dance may be partly a function of higher preservation

potential towards the sediment surface.

5.2.6. MolluscaMollusc assemblages (Fig. 9) were dominated by gastropods;

apart from at 1–2 cm depth, bivalves were absent or rare. All

molluscs were rare in the compact sediments of Zone M1

(31.5–22.0 cm depth; pre-19th century), with occasional shells

of the gastropods, Planorbis spp., Physa acuta and an indeter-

minate larval protoconch, and the bivalve genus, Pisidium.

An increase in abundance and diversity occurred in Zone

M2 (22.0–0.0 cm; pre-19th century to 2000), partly reflecting

better preservation, and peaked at 237–238 inds. 100 g�1 be-

tween 4 and 6 cm. Zone M2 exhibited co-dominance by the

aquatic prosobranch (gill-breathing) Valvata piscinalis and pul-

monate (air-breathing) Planorbis spp. Larval shells were abun-

dant in the first half of the zone prior to the overall increase in

mollusc abundance above 10 cm depth. Mainly above this

depth, other gastropods were present sporadically, including

the pulmonates, Lymnaea peregra, Lymnaea truncatula, P. acuta

and Physa fontanalis, and the prosobranchs Bithynia sp. and

Bythinella sp.

6. Discussion

6.1. Current chemical and ecological status

The eutrophic status of Lake Uluabat is recognised from water

chemistry monitoring (e.g. Dalkıran et al., 2006) and macro-

phyte surveys (Lammens and van den Berg, 2001), but ecolog-

ical status at lower trophic levels had not been previously

explored. The high pH values are typical of alkaline, karstic

lakes, and the low water clarity is typical of shallow, turbid,

eutrophic lakes. The general lack of between-sample variabil-

ity in measured water chemistry parameters is consistent

with full mixing in a shallow lake and suggests that, in spite

of reported differences in residence time across the basin

(Lammens and van den Berg, 2001), there is no strong evidence

for local variability. The slightly lower DOC values in the shal-

lower waters close to the Mustafakemalpas�a delta may just

reflect water depth/temperature effects, since these samples

do not also exhibit the lowest BMWP scores. Our invertebrate

biomonitoring study had the limitation of being restricted to a

single season (spring), but the combination of consistently low

scores and the presence in the surface sediment of diatoms

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

Dep

th (c

m)

20 40 60 80

Physo

cypri

a krae

pelin

i(A)

20 40 60

Physo

cypri

a krae

pelin

i(J)

20

Cyprid

opsis

vidu

a(A)

20

Cyprid

opsis

vidu

a(J)

20

lIyoc

ypris

gibb

a

Cando

na an

gulat

a

20 40 60 80%

Cando

nasp

p.(J)

500 1000 1500Inds./100g dry sed.

Ostrac

od ab

unda

nce

Zone O2

Zone O1

% % % % % %

Fig. 8 – Diagram showing the relative and absolute abundance of ostracod taxa in UL20A. Counts are separated into adult (A)

and juvenile (J) valves.


typical of eutrophic to hypereutrophic lakes, indicates poor

ecological status at all levels of the ecosystem. The low BMWP

scores (<10 in most samples) tends to indicate significant bulk

organic or nutrient pollution (Abel, 1996). Although designed

primarily to improve palaeolimnological interpretation, the

homogeneity of ostracod transect samples also indicates low

habitat diversity across the open waters.

It is noteworthy that the highest BMWP scores (with abun-

dant damsel flies and the only presence of gastropods) de-

rived from sample B1, the only sample located in reed beds

rather than the open water. This suggests low scores are

partly a function of habitat availability, which must have been

reduced with the loss of submerged aquatic macrophyte cov-

er. Our study in 2001 was performed during a year which ap-

peared to be unusual in the lack of development of extensive

submerged macrophyte beds; less pollution-tolerant inverte-

brate groups tend to be more common within the plant-dom-

inated habitat and it is likely that habitat degradation in

addition to low water quality is a cause of low diversity. Since

2001 submerged aquatic plant cover has remained low and

our study appears to have coincided with the start of a major

decline in Lake Uluabat’s biodiversity. It is well demonstrated

that a loss of plant cover could have important ramifications

at all levels of the ecosystem, since plant habitats provide

important refugia for invertebrates such as cladocera which

consume algae and can function to maintain a well-oxygen-

ated, clear water state (Moss et al., 1996).

Gastropods (V. piscinalis, L. peregra, Bithynia sp.) were also

only present in the reed-bed sample, B1, and we would pre-

dict major changes to molluscan community structure if

plant loss is prolonged. The results of a study of English ponds

by Lodge and Kelly (1985) demonstrated that the loss of sub-

merged macrophytes caused 99% of Lymnaea and ca. 35% of

Valvata populations to die, with more rapid recolonisation

with regrowth by Lymnaea, but Planorbis and Bithynia were

not reduced. While some taxa such as Planorbis spp. can occu-

py both submerged and emergent macrophyte habitats, V.

piscinalis is typically associated only with submerged vegeta-

tion, and was restricted to this habitat in eutrophic Radley

Pond, UK (Lodge and Kelly, 1985). The Lake Uluabat gastropod

community was not found in the open-water mud habitat; at

best the gastropod fauna would be reduced to a lower diver-

sity community restricted to shoreline emergent vegetation

habitats should plant loss be sustained.

6.2. Ostracods and molluscs as palaeolimnological proxies

The results of the ostracod and mollusc survey provide infor-

mation on habitat distribution. The widespread occurrence

and abundance of the ostracod, P. kraepelini, is consistent with

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

Dep

th (c

m)

20 40 60 80 100

Valvata

piscin

alis

20 40 60 80 100

Planorb

is spp.

20

Lymna

ea (ra

dix) p

eregra

Lymna

ea (g

alba)

trunc

atula

20 40 60 80 100

Physa

acuta

20

Physa

fontan

alis

20

Bithyn

iasp

p.

Bythine

llasp

p.

20 40 60 80 100

Protoc

onch

ia ind

et

20 40 60 80 100

Psidium

spp.

(j)

20 40%

Bivalve

s indet

100 200 300

Inds./100g dry sed.

Mollus

c abu

ndan

ce

Zone M2

Zone M1

% %% % %%%%%%

Fig. 9 – Diagram showing the relative and absolute abundance of mollusc taxa in UL20A.


its nektonic (swimming) life habit in open waters, although,

from its common occurrence close to reed beds in the littoral

zone (Altınsaclı and Griffiths, 2001), it is not a good proxy indi-

cator for water depth. The higher abundance of the shallow,

limnic ostracod, Darwinula stevensoni, in Transect A is consis-

tent with the conjecture of Altınsaclı and Griffiths (2001) that,

based on its distribution in littoral reed beds, the taxon is

common throughout the lake in all areas apart from the

northeast. The distribution of I. gibba is more widespread

than thought previously, being common in the waters of ca.

1 m depth in this study, rather than showing a preference

for depths <0.5 m as suggested by Altınsaclı and Griffiths

(2001). The marked preference of candonids for the habitats

of Transect B may indicate the presence of the benthic species

C. angulata, which Altınsaclı and Griffiths (2001) found to be

particularly common in this part of the lake. In general, the

data-set shows strong similarities to the faunal composition

of the littoral zone described by Altınsaclı and Griffiths

(2001), suggesting that in this lake it is not possible to distin-

guish definitively between open water and littoral zone habi-

tats based on ostracod species composition alone. The

exception is C. vidua, which is absent here but was common

in the previous study. Since our samples are mainly from

open-water habitats, it is possible its presence could be indic-

ative of the proximity of reed beds, since it has a marked pref-

erence for well-vegetated zones (Bronshtein, 1947).

The apparent habitat preferences of molluscs are dis-

cussed above; although rare, their clear, restricted distribu-

tion suggests these taxa are useful palaeolimnological

indicators for the proximity of plant macrophytes.

6.3. Palaeolimnological evidence for changes in ecosystemstatus

The palaeolimnogical data provide evidence for changes in

the physical, chemical and biological environment of Lake

Uluabat over time. The modern values for organic content

are similar to other Mediterranean karstic eutrophic lakes

such as Mikri Prespa, NW Greece (20%; Stevenson and Flower,

1991), and the pattern of increased organic content over time

is a common function of eutrophication, with increased inter-

nal productivity. The high carbonate content is typical of kar-

stic lakes, and the parallel rise over time may reflect a

combination of increased productivity and/or input of miner-

ogenic matter from the catchment. The negative bulk authi-

genic stable isotope values, and their lack of strong

covariance, suggests an open lake with little evaporative con-

centration. Since there is little surface outflow, this indicates

major groundwater aquifer throughput in the karstic system.

The results are comparable to modern isotopic values from

carbonate precipitating in the lake during 2001, of d13C

�4.8&; d18O �7.6& (Leng, unpublished data). The large range

in ostracod stable isotope values, which do not follow the

same pattern of variation as the bulk authigenic values, re-

flects the fact that the ostracods capture a moment in time

and the lake water is highly seasonal.

The diatom data indicate that the lake has been eutrophic

since prior to the 19th century and possibly as early as the

17th century, but a subtle diatom-inferred trend of increasing

eutrophication correlates with the inferred increases in sedi-

ment accumulation rate and organic and carbonate content,


and in combination provides clear evidence for 20th century

nutrient enrichment which accelerated after ca. 1962 (ca.

13 cm depth). The inferred acceleration in nutrient status

after ca. 1962 starts soon after the acceleration of sediment

accumulation rates indicated by the 210Pb data at ca. 14–

15 cm depth; the apparent 1 cm offset may simply be a func-

tion of having used a second core for radiometric dating. In a

previous study of sediment accumulation rates in Lake Ulu-

abat, Kazancı et al. (2004) estimated an increasing trend from

0.2 cm yr�1 prior to 2000 years BP, to >0.4 cm yr�1 in the 20th

century.

The changes in diatom species assemblage composition

are not as dramatic as in some studies of temperate lakes of

northwestern Europe (e.g. Soppensee, Switzerland; Lotter,

2001), where recent eutrophication is marked by the first

appearance and expansion of eutrophic taxa. In this lake,

however, which has been eutrophic since at least the start

of the 19th century and is known to lack a well-developed

planktonic flora due to high turbidity (Karacaoglu et al.,

2004), the gradual rise in eutrophic planktonic taxa does indi-

cate accelerated eutrophication.

Perhaps the most serious implication of the diatom data is

that the rise of planktonic taxa occurs largely at the expense

of epiphytic taxa. The decline in Epithemia spp. to <10% abun-

dance above ca. 13 cm depth (ca. 1962) is likely to reflect a sus-

tained reduction in submerged aquatic macrophyte cover. As

noted, macrophyte cover is often the mainstay of ecosystem

health in a eutrophic lake. From pre-2000 data, Beklioglu

et al. (2006) observed that plant growth in Lake Uluabat was

most extensive in its shallow phases. The lake is currently

in a shallow state, so the loss of plant cover, which has accel-

erated since 2000, indicates major ecological impact.

Other palaeoecological data are provided by ostracods and

molluscs, which do not exhibit a major transition in ca. 1962.

The ostracod fauna is similar to that of the northeastern tran-

sect survey, being dominated by candonids, with increased

abundance of the nektonic taxon, P. kraepelini, after ca. 1985.

The increase in the latter would be consistent with the effects

of reduced plant cover since 1985, but co-dominants are com-

mon in the reed beds of the modern littoral zone (Altınsaclı

and Griffiths 2001). C. vidua, for example, is common above

26 cm depth. This may indicate on the one hand that until

very recently the lake has been well vegetated; the weakness

of P. kraepelini as a proxy for open water habitats was dis-

cussed above. On the other hand, C. vidua has a reported

intolerance of very low oxygen availability (Meisch, 2000),

and its presence is consistent with the directly measured

water column parameters. From the foregoing, the co-domi-

nance of V. piscinalis and Planorbis spp. in the molluscan fau-

na, with rare associated taxa which are common the

modern lake, supports the inference of a well-vegetated lake

throughout the recent past. As with the ostracod record, the

similarity to the modern fauna (which is common in eutro-

phic ponds; Lodge and Kelly, 1985), and the abundance of pul-

monate taxa which can tolerate relatively low oxygen

conditions, again suggests that the lake has been eutrophic

throughout.

The transition in ca. 1962 also correlates, as far as ostracod

preservation permits the definition of the transition, with the

major shift in d18Oost values, from more positive, fluctuating

values in the lower record to more negative, stable values in

the upper record. This we suggest is due a greater seasonality

in the lower record. The weak covariance between d13Cost and

d18Oost above ca. 13 cm depth, and lack of covariance below,

may indicate the lake became more hydrologically ‘closed’,

or endorheic, after ca. 1962. In a hydrologically stable lake, a

shift to more negative oxygen stable isotope values would

indicate reduced rather than enhanced evaporative concen-

tration. Here, it is more likely to indicate a seasonal shift in

the provenance (spring vs. river water) and isotopic composi-

tion of inflowing water, since it is not reflected in an associ-

ated shift in bulk stable isotope values. Equally, since both

oxygen and carbon isotope values change (rather than carbon

alone), it is unlikely to have been driven by ecosystem shifts

such as changes in the trophic structure.

The diatom planktonic:benthic ratio may also be influ-

enced by changing lake level, with increases in the proportion

of plankton at the coring site with rising lake level. In this

study, however, it is impossible to infer lake-level change with

confidence. Although the results of water-level and surface-

area monitoring data are so contradictory, the consensus is

that, if anything, the lake has shallowed over the last decades.

The increase in diatom plankton in the upper sequence is

consistent both with higher productivity and higher water

levels, but can only be interpreted logically in terms of the

former.

6.4. External impacts

In the light of the fragmentary and contradictory record of

changing land-use practices, industrial and urban activity,

and water management practices outlined above, all of which

lack clear chronological detail, it is impossible to determine

precisely the relative impacts of these multiple stressors,

but some useful conclusions may be drawn. Our results con-

firm (as in the Holocene study of Kazancı et al., 2004), that a

pattern of increased sediment accumulation rates was initi-

ated in the early 20th century, and suggest that eutrophica-

tion has been driven at least in part by nutrient input from

sediment supply. The most likely cause, as suggested by Kaz-

ancı et al. (2004), is deforestation and the expansion of agri-

culture in the catchment. High estimated sediment

accumulation rates in the early mid-20th century in karstic

Lake _Iznik, western Turkey (Franz et al., 2006) also correlated

well with inferred mineralogical evidence for deforestation,

although accumulation rates appeared to decrease subse-

quently with recent nutrient enrichment. An alternate expla-

nation is changing river discharge and siltation from the R.

Mustafakemalpas�a. Lammens and van den Berg (2001) sug-

gest discharge has reduced rather than increased since the

1960s, however. They link this with reduced mining activity,

although water abstraction (Magnin and Yarar, 1997) must

also have had an influence. In either case, this general pattern

does not match the inferred trend of accelerated sediment

accumulation rate and eutrophication since ca. 1960 in our re-

cord. Urban sources may also be significant; Green et al. (1996)

noted that phosphorus stripping in sewage treatment plants

was deemed too expensive by the State to reduce urban pol-

lution in Lake Burdur, central Turkey. The 1960s are com-

monly a time of intensification of both agricultural and


urban pollution across Europe but in the absence of relevant

information we cannot ascertain this.

The trend from ca. 1960 towards plant loss with acceler-

ated sediment accumulation rate does correlate with ostracod

stable isotope evidence for reduced seasonal fluctuation and a

shift in the hydrological regime. This suggests that acceler-

ated ecological impact is at least in part a function of water

management practices. The wide variation in the lower ostra-

cod stable isotope record is consistent with the summer

reproductive biology of P. kraepelini (Meisch, 2000) and proba-

bly reflects a greater degree of summer seasonal fluctuation

prior to embankment of the lake, which started in 1937, and

the subsequent management of the outflow during the

1980–1990s (Lammens and van den Berg, 2001). The reduction

in variability in the upper record is consistent with more sta-

ble lake levels from the 1980s; we lack ostracod data for ca.

1960–1980 to define these trends in more detail. With artificial

management of the River Mustafakemalpas�a, it is possible

that with reduced inflow the hydrological input from springs

is more significant than previously, although the lack of a re-

sponse in the bulk authigenic data suggests a more seasonal

and/or local change in hydrological regime.

6.5. Implications for lake management

From this study, the most serious impact on the physico-

chemical and ecological status of Lake Uluabat appears to

have been during the 1960s, against a trend of rising nutrient

levels since the start of the 20th century. The most important

outcome of the palaeolimnological study is that there is clear

evidence for the impact of both catchment land-use activities

and water management practices, and that the most rapid

and marked change in ecological status appears to be associ-

ated more with the latter. Since the 1960s, an additional

source of pressure from nutrient enrichment must also derive

from the reported recent switch of bird migration routes from

Lake Manyas to Lake Uluabat as water quality deteriorated in

the former (cf. Noordhuis et al., 2002). The loss of aquatic

macrophyte communities since 2000 has also been observed,

and the poor ecological status of the modern lake at lower lev-

els of the trophic web. Without intervention, it seems likely

that the endpoint for Lake Uluabat will be to become as de-

graded as Lake Manyas.

In the context of the WFD approach to conservation and

restoration of lakes, it has been ascertained that there is no

ecologically distinct ‘natural baseline’ to define as a restora-

tion target. A clear target would be to restore the lake to its

pre-1960s state, however. Our study confirms the suggestion

of Lammens and van den Berg (2001) that a reduction in nutri-

ent loading alone would be insufficient, and that natural

water-level fluctuations must also be restored. An obvious

restoration strategy in addition to reducing external nutrient

input, would be to encourage regrowth of littoral reed beds,

which were formerly more extensive, to act as a buffer zone

for nutrient input (Moss et al., 1996). The restoration of natu-

ral fluctuations again becomes significant, since Lammens

and van den Berg (2001) predicted the loss of reed beds over

time under the current regime.

A key to improving ecological status is to encourage sub-

merged plant regrowth. In Lake Uluabat, a reduction in exter-

nal nutrient input alone would be unlikely to be sufficient. In

lake restoration, plant regrowth is often hampered by a vari-

ety of factors including grazing and/or bioturbation by birds

(e.g. Noordhuis et al., 2002; Irfanullah and Moss, 2004), or fish

(e.g. Korner, 2001; Matsuzaki et al., 2007). While fish stocks

including the turbidity-inducing carp have declined (Lam-

mens and van den Berg, 2001), reported pressure from bird

grazing is extreme. Our results also indicate that the reduc-

tion in aquatic macrophytes is linked to changes in water

management practices. A similar phenomenon was observed

by van Geest et al. (2005), who demonstrated the negative im-

pact of reduced water-level fluctuations on aquatic vegetation

succession in floodplain lakes of the Lower Rhine. In Lake

Uluabat, which appears to have undergone a long-term trend

towards reduction in submerged macrophyte cover driven by

multiple ecosystem impacts, it is unlikely that ecosystem

‘stability’ could be restored without recourse to biomanipula-

tion (cf. Bootsma et al., 1999). Again, an obvious strategy

which has proved successful in other lakes would be to

encourage recolonisation by exclusion of bird and fish popu-

lations in artificial embayments. In the longer term, however,

short of culling, the only viable solution to the increased pres-

sure from migratory birds would be to carry out a major res-

toration project on neighbouring Lake Manyas.

6.6. Wider implications: the conservation of karstic lakes

Even in research focused specifically on the EU WFD, there

has been a marked bias in conservation-based palaeolimnol-

ogy towards temperate glacial lakes of northwestern Europe

(e.g. Jeppesen et al., 2002; Battarbee et al., 2005; McGowan

et al., 2005; Bradshaw et al., 2006; Leira et al., 2006), and

there is a dearth of palaeolimnological impact studies across

the circum-Mediterranean. In shallow, productive lakes,

much of the research, as here, has exploited the value of

diatoms as sensitive environmental indicators of trophic sta-

tus, while strengthening interpretation of ecosystem status

by employing a multi-proxy approach (Bennion and Battar-

bee, 2007).

The problem is not simply one of a lack of relevant diatom-

based training sets (cf. Reed, 2007) and of sufficient knowl-

edge of pre-impact reference states. Many lakes across the

circum-Mediterranean are karstic, groundwater-fed systems

in tectonically active areas and the assumption that the ‘tem-

perate-lake’ approach to evaluation of ecological status is va-

lid has not been tested. The results of this study are

important in demonstrating the response of a karstic lake

ecosystem to external and internal stresses. Here, the results

of diatom analysis showed a relatively subtle eutrophication

response. In a palaeolimnological study of shallow Lake Mikri

Prespa (Greece), Stevenson and Flower (1991) also demon-

strated little diatom response in spite of accelerated sediment

and nutrient input and water abstraction (Stevenson and

Flower, 1991; Hollis and Stevenson, 1997). Similarly, in shallow

Lake Dojran (Greece-Former Yugoslav Republic of Macedonia),

Griffiths et al. (2002) interpreted a subtle diatom-inferred

trend towards eutrophication, and a stable ostracod fauna,

as indicative that in spite of high nutrient loading and exces-

sive water abstraction, the ecological status of the lake had

not been seriously affected.


Several authors (e.g. Hollis and Stevenson, 1997 for Mikri

Prespa; Bertahas et al., 2006 for Lake Trichonis, Greece; de

Vicente et al., 2006 for Laguna Nueva, Spain) have argued that

high groundwater discharge in karstic lakes buffers them

from the effects of accelerated nutrient input. In a temperate

lake, a subtle diatom response may indeed be indicative of

minimal impact, with little cause for concern. We have sug-

gested previously (Griffiths et al., 2002), however, that karstic

lakes may exhibit a stable phase of apparent buffering by

groundwaters, but that this may be followed by an ecological

threshold of major ecosystem impact. In Lake Uluabat, this

pattern is suggested by the rapid loss of aquatic macrophytes

since 2000. In Mikri Prespa, the eutrophic diatom A. granulata

is now abundant (G. Wilson, personal communication), while

Lake Dojran has all but disappeared and is highly eutrophic

(S. Krstic, personal communication). Although rather anec-

dotal, all these case studies support the contention that signs

of apparently minor ecological change should be taken more

seriously in karstic, groundwater-fed lakes than in glacial

lakes.

The study also demonstrated the significant impact of

changed water management practices on the ecological sta-

tus of Lake Uluabat. Water abstraction is not a major issue

in humid regions, and glacial lakes do not exhibit the same

degree of natural lake-level fluctuation as karstic lakes such

as Lake Uluabat and many other fresh lakes across the cir-

cum-Mediterranean. For obvious reasons, the larger monitor-

ing projects in northwestern Europe do not have a focus on

hydrological impact. As cited here, individual studies have

been carried out, but the results of this study demonstrate

clearly the need for larger-scale monitoring and palaeolimno-

logical research into the additional impacts of water manage-

ment practices on ecology.

Acknowledgements

We would like to offer our sincere thanks to the various

organisations who provided support for this project, compris-

ing the Royal Society (RSRG 22163 to J.M.R.), the British Coun-

cil Britain–Turkey Partnerships Programme (H.I.G., S.A.,

J.M.R.), and the Rotary Club of Turkey (S.R.). The project was

carried out during the tenure of a Leverhulme Special Re-

search Fellowship to J.M.R. (SRF/66). Thanks are also due to

John Garner and Keith Scurr for help with the map, to Songul

Altınsaclı for help in the field, and to two anonymous referees

for their constructive comments on the manuscript. We ded-

icate this paper to the memory of Huw and his undying

enthusiasm.

R E F E R E N C E S

Abel, P.D., 1996. Water Pollution Biology. Taylor and Francis,London.

Akkoyunlu, A., 2003. Evaluation of eutrophication process in Lake_Iznik. Fresenius Environmental Bulletin 12, 801–807.

Aksoy, A.O., Scheytt, T., 2007. Assessment of groundwaterpollution around Torbali, Izmir, Turkey. EnvironmentalGeology 53, 19–25.

Alba Tercedor, J., Sanchez Ortega, S., 1988. Un metodo rapdio ysimple para evaluar la calidad biologica de las aguas corrientesbasado en el de Hellawell (1978). Limnetica 4, 51–56.

Albay, M., Akcaalan, R., 2003. Comparative study of periphytoncolonisation on common reed (Phragmites australis) andartificial substrate in a shallow lake, Manyas, Turkey.Hydrobiologia, 531–540.

Altınsaclı, S., Griffiths, H.I., 2001. Ostracoda (Crustacea) of LakeUluabat (Apolyont Golu) (Bursa Province, Turkey). Limnologica31, 109–117.

Anderson, N.J., Rippey, B., Gibson, C.E., 1993. A comparison ofsedimentary and diatom-inferred phosphorus profiles:implications for defining pre-disturbance nutrient conditions.Hydrobiologia 253, 357–366.

Arcak, S., Haktanır, K., Kibar, M., Dengiz, O., 2000. Ecologicalchanges in Manyas lake related to boron pollution and waterregime. In: Proceedings of the International Symposium onDesertification, Konya, Turkey, June 13–17, 2000.

Armitage, P.D., Moss, D., Wright, J.F., Furse, M.T., 1983. Theperformance of a new biological water quality score systembased on macroinvertebrates over a wide range of unpollutedrunning-water sites. Water Research 17, 333–347.

Barlas, N., Akbulut, N., Aydogan, M., 2005. Assessment of heavymetal residues in the sediment and water samples of UluabatLake, Turkey. Bulletin of Environmental Contamination andToxicology 74, 286–293.

Battarbee, R.W., 1986. Diatom analysis. In: Berglund, B.E. (Ed.),Handbook of Holocene Palaeoecology and Palaeohydrology.John Wiley, Chichester, pp. 527–570.

Battarbee, R.W., 1999. The importance of palaeolimnology in lakerestoration. Hydrobiologia (395/396), 149–159.

Battarbee, R.W., Juggins, S., Gasse, F., Anderson, N.J., Bennion, H.,Cameron, N.G., 2000. European Diatom Database (EDDI). Aninformation system for palaeoenvironmental reconstruction.In: European Climate Science Conference, Vienna City Hall,Vienna, Austria, October 19–23, 1998, pp. 1–10.

Battarbee, R.W., Anderson, N.J., Jeppesen, E., Leavitt, P.R., 2005.Combining palaeolimnological and limnological approachesin assessing lake ecosystem response to nutrient reduction.Freshwater Biology 50, 1772–1780.

Beklioglu, M., _Ince, O., Tuzun, I., 2003. Restoration of the eutrophicLake Eymir, Turkey, by biomanipulation after a major externalnutrient control I. Hydrobiologia 489, 93–105.

Beklioglu, M., Altınayar, G., Tan, C., 2006. Water level control oversubmerged macrophyte development in five shallow lakes ofMediterranean Turkey. Archiv fur Hydrobiologie 166, 535–555.

Bennion, H., 1994. A diatom-phosphorus transfer function forshallow, eutrophic ponds in southeast England. Hydrobiologia(275/276), 391–410.

Bennion, H., Battarbee, R., 2007. The European Water FrameworkDirective: opportunities for palaeolimnology. Journal ofPaleolimnology 38, 285–295.

Bennion, H., Juggins, S., Anderson, N.J., 1996. Predictingepilimnetic phosphorus concentrations using an improveddiatom-based transfer function and its application to lakeeutrophication management. Environmental Science andTechnology 30, 2004–2007.

Bennion, H., Fluin, J., Simpson, G.L., 2004. Assessingeutrophication and reference conditions for Scottishfreshwater lochs using subfossil diatoms. Journal of AppliedEcology 41, 124–138.

Bertahas, I., Dimitriou, E., Karaouzas, I., Laschou, S., Zacharias, I.,2006. Climate change and agricultural pollution effects on thetrophic status of a Mediterranean lake. Acta Hydrochimica etHydrobiologica 34, 349–359.

Bootsma, M.C., Barendregt, A., van Alphen, J.C.A., 1999.Effectiveness of reducing external nutrient load entering aeutrophicated shallow lake ecosystem in the Naardermeer


nature reserve, The Netherlands. Biological Conservation 90,193–210.

Bottema, S., Woldring, H., 1984. Late Quaternary vegetation andclimate of southwestern Turkey, Part II. Palaeohistoria 26, 123–149.

Bradshaw, E.G., Nielsen, A.B., Anderson, N.J., 2006. Using diatomsto assess the impacts of prehistoric, pre-industrial andmodern land-use on Danish lakes. Regional EnvironmentalChange 6, 17–24.

Bronshtein, Z.S., 1947. Freshwater Ostracoda. In: Fauna of theUSSR, Crustaceans, vol. 2(1). Amerind Publications, New Delhi(English translation 1988).

Croft, P.S., 1986. A key to the major groups of British freshwaterinvertebrates. Field Studies 6, 531–579.

Dalkıran, N., 2000. The seasonal variation of the epipelic,epiphytic and epilithic algae in Lake Uluabat (Bursa).Unpublished M.Sc. Thesis, University of Bursa, Turkey (inTurkish with English abstract).

Dalkıran, N., Karacaoglu, D., Dere, S., S�enturk, E., Torunoglu, T.,2006. Factors affecting the current status of a eutrophicshallow lake (Lake Uluabat, Turkey): relationships betweenwater physical and chemical variables. Chemistry and Ecology22, 279–298.

Dean, W.E., 1974. Detection of carbonate and organic matter incalcareous sediments and sedimentary rocks by loss onignition. Chemical Geology 44, 151–185.

De Corte, F., Umans, H., Vandenberghe, H., De Wispelaere, A., vanden Haute, P., 2005. Direct gamma-spectrometricmeasurement of the 226Ra 186.2 keV line for detecting238U/226Ra disequilibrium in determining the environmentaldose rate for the luminescence dating of sediments. AppliedRadiation and Isotopes 63, 589–598.

de Vicente, I., Moreno-Ostos, E., Amores, V., Rueda, F., Cruz-Pizarro, L., 2006. Low predictability in the dynamics of shallowlakes: implications for their management and restoration.Wetlands 26, 928–938.

Degirminci, H., Alp, A., Buyukcangaz, H., 2005. Diagnostic analysisof Lake Uluabat in Turkey. Journal of Environmental Biology 27,431–436.

Demirel, Z., 2007. Monitoring of heavy metal pollution ofgroundwater in a phreatic aquifer in Mersin-Turkey.Environmental Monitoring and Assessment 132, 15–23.

Dere, S�., Dalkıran, N., Karacaoglu, D., Elmacı, B., S�enturk, E., 2006.Relationships among epipelic diatom taxa, bacterialabundances and water quality in a highly polluted streamcatchment, Bursa, Turkey. Environmental Monitoring andAssessment 112, 1–22.

Duran, M., 2006. Monitoring water quality using benthicmacroinvertebrates and physicochemical parameters ofBehzat Stream in Turkey. Polish Journal of EnvironmentalStudies 15, 709–717.

Duran, M., Suicmez, M., 2007. Utilization of both benthicmacroinvertebrates and physicochemical parameters forevaluating water quality of the stream Cekerek (Tokat, Turkey).Journal of Environmental Biology 28, 231–236.

Eastwood, W.J., Roberts, N., Lamb, H.F., Tibby, J.C., 1999. Holoceneenvironmental change in southwest Turkey: apalaeoecological record of lake and catchment-relatedchanges. Quaternary Science Reviews 18, 671–696.

Eastwood, W.J., Tibby, J.C., Roberts, N., Birks, H.J.B., Lamb, H.F.,2002. The environmental impact of the Minoan eruption ofSantorini (Thera): numerical analysis of palaeoecological datafrom southwest Turkey. The Holocene 12, 431–444.

Eastwood, W.J., Leng, M.J., Roberts, N., Davis, B., 2007. Holoceneclimate change in the eastern Mediterranean region: acomparison of stable isotope and pollen data from LakeGolhisar, southwest Turkey. Journal of Quaternary Science 22,327–341.

Fitter, R., Manuel, R., 1986. Guide to the Freshwater Life of Britainand Northwest Europe. Collins, London.

Franz, S.O., Schwark, L., Bruchmann, C., Scharf, B., Klingel, R., vanAlstine, J.D., Cagatay, N., Ulgen, U.B., 2006. Results from amulti-disciplinary sedimentary pilot study of tectonic LakeIznik (NW Turkey) – geochemistry and paleolimnology of therecent past. Journal of Paleolimnology 35, 715–736.

Freshwater Biological Association, 2005. New tools for themonitoring, modelling and management of fresh waters.Special Topic. Freshwater Forum 23, 170 pp.

Germain, H., 1981. Flore des diatomees. Diatomophycees EauxDouces et Saumatres du Massif Armoricain et des ContreesVoisines d’Europe Occidentale. Boubee, Paris.

Glew, J.R., 1991. Miniature gravity corer for recovering shortsediment cores. Journal of Paleolimnology 5, 285–287.

Green, A.J., Fox, A.D., Hilton, G., Hughes, B., Yarar, M., Salathe, T.,1996. Threats to Burdur Lake ecosystem, turkey and itswaterbirds, particularly the white-headed duck Oxyuraleucocephala. Biological Conservation 76, 241–252.

Griffiths, H.I., Holmes, J., 2000. Non-marine ostracods andquaternary palaeoenvironments. QRA Technical Guide No. 8.Quaternary Research Association, London.

Griffiths, H.I., Reed, J.M., Leng, M.J., Ryan, S., Petkovski, S., 2002.The conservation status and recent palaeoecology of BalkanLake Dojran. Biological Conservation 104, 35–49.

Grimm, E., 1991. TILIA and TILIA-GRAPH. Illinois State Museum,Springfield, USA.

Grimm, E.C., 1987. CONISS: a FORTRAN 77 program for thestratigraphically constrained cluster analysis by the method ofincremental sum of squares. Computers and Geosciences 13,13–35.

Gurluk, S., Rehber, E., 2006. Evaluation of an integrated wetlandmanagement plan: case of Uluabat (Apollonia) Lake, Turkey.Wetlands 26, 258–264.

Hollis, G.E., Stevenson, A.C., 1997. The physical basis of the LakeMikri Prespa systems: geology, climate, hydrology and waterquality. Hydrobiologia 351, 1–19.

Homan, J., Eastwood, W.J., 2001. The 17 August 1999 Kacaeli(_Izmit) Earthquake: historical records and seismic culture.Earthquake Spectra 17, 617–634.

Irfanullah, H., Moss, B., 2004. Factors influencing the return ofsubmerged plants to a clear-water, shallow temperate lake.Aquatic Botany 80, 177–191.

Ivanovich, M., Harmon, R.S., 1992. Uranium Series DisequilibriumApplications to Environmental Problems, second ed.. OxfordUniversity Press, Oxford.

Jeppesen, E., Jensen, J.P., Amsinck, S., Landkildehus, F., Lauridsen,T., Mitchell, S.F., 2002. Reconstructing the historicaldevelopment in planktivorous fish abundance in lakes fromsize of Daphnia ephippia in the sediment. Journal ofPaleolimnology 27, 133–143.

Jones, M.D., Leng, M.J., Keen, D.H., Turney, C.S.M., 2002.Interpreting stable-isotope records from freshwater snail-shellcarbonate: a Holocene case study from Lake Golhisar, Turkey.The Holocene 12, 629–634.

Jones, M.D., Roberts, C.N., Leng, M.J., Turkes, M., 2006. A high-resolution late Holocene lake isotope record from Turkey andlinks to North Atlantic and monsoon climate. Geology 34, 361–364.

Jones, M.D., Roberts, C.N., Leng, M.J., 2007. Quantifying climaticchange through the last glacial–interglacial transition basedon lake isotope palaeohydrology from Turkey. QuaternaryResearch 67, 463–473.

Jongman, R.H.G., ter Braak, C.J.F., van Tongeren, O.F.R., 1995. DataAnalysis in Community and Landscape Ecology. CambridgeUniversity Press, Cambridge.


Karacaoglu, D., Dere, S., Dalkıran, N., 2004. A Taxonomic Study onthe Phytoplankton of Lake Uluabat (Bursa). Turkish Journal ofBotany 28, 473–485.

Karafistan, A., Arik-Colakoglu, 2005. Physical, chemical andmicrobiological water quality of the Manyas lake, Turkey.Mitigation and Adaptation Strategies for Global Change 10,127–143.

Karakoc, G., Erkoc, F.U., Katırcıoglu, H.K., 2003. Water quality andimpacts of pollution sources for Eymir and Mogan Lakes(Turkey). Environment International 29, 21–27.

Kashima, K., 2002. Environmental and climatic changes duringthe last 20,000 years at Lake Tuz, central Turkey. Catena 48, 3–20.

Kazancı, N., Leroy, S., _Ileri, O., Emre, O., Kibar, M., Oncel, S., 2004.Late Holocene erosion in NW Anatolia from sediments of LakeManyas, Lake Ulubat and the southern shelf of the MarmaraSea, Turkey. Catena 57, 277–308.

Kazancı, N., Toprak, O., Leroy, S.A.G., Oncel, S., Ileri, O., Emre, O.,Costa, P., Erturac, K., McGee, E., 2006. Boron content of LakeUlubat sediment: a key to interpret the morphological historyof NW Anatolia, Turkey. Applied Geochemistry 21, 134–151.

Korner, S., 2001. Development of submerged macrophytes inshallow Lake Muggelsee (Berlin, Germany) before and after itsswitch to the phytoplankton-dominated state. Archiv furHydrobiologie 152, 395–409.

Krammer, K., Lange-Bertalot, H., 1986. Bacillariophyceae 1. Teil:Naviculaceae. In: Ettl, H., Gartner, G., Gerloff, J., Heynig, H.,Mollenhauer, D. (Eds.), . In: Susswasserflora von Mitteleuropa,Bacillariophyceae, vol. 2/1. Gustav Fischer Verlag, Stuttgart.

Krammer, K., Lange-Bertalot, H., 1988. Bacillariophyceae 2. Teil:Bacillariaceae, epithemiaceae, surirellaceae. In:Susswasserflora von Mitteleuropa. Bacillariophyceae, vol. 2/2.Gustav Fischer Verlag, Berlin.

Krammer, K., Lange-Bertalot, H., 1991a. Bacillariophyceae 3. Teil:Centrales, fragilariaceae, eunotiaceae. In: Susswasserflora vonMitteleuropa. Bacillariophyceae, vol. 2/3. Gustav FischerVerlag, Stuttgart.

Krammer, K., Lange-Bertalot, H., 1991b. Bacillariophyceae 4. Teil:Achnanthaceae. In: Susswasserflora von Mitteleuropa.Bacillariophyceae, vol. 2/4. Gustav Fischer Verlag, Stuttgart.

Krno, I., Sporka, F., Galas, J., Hamerlik, L., Zatovicova, Z., Bitusik,P., 2006. Littoral benthic macroinvertebrates of mountain lakesin the Tatra Mountains (Slovakia, Poland). Biologia 61, 147–166.

Kubilay, A., Timur, G., 1995. Identification, zonation and seasonaldistribution of mollusks in the benthic fauna at Kopru fishingground, Lake Egirdir. Turkish Journal of Zoology 19, 241–248.

Lammens, E., van den Berg, M., 2001. Evaluation of the ecologicalcondition of Lake Uluabat. RIZA Work Document 2001.197X,The Netherlands.

Landmann, G., Reimer, A., Lemcke, G., Kempe, S., 1996. DatingLate Glacial abrupt climate changes in the 14,570 yr longcontinuous varve record of Lake Van, Turkey. Palaeogeography,Palaoeclimatology, Palaeoecology 122, 107–118.

Leira, M., Jordan, P., Taylor, D., Dalton, C., Bennion, H., Rose, N.,Irvine, K., 2006. Assessing the ecological status of candidatereference lakes in Ireland using palaeolimnology. Journal ofApplied Ecology 43, 816–827.

Leng, M.J., Roberts, N., Reed, J.M., Sloane, H.J., 1999. LateQuaternary palaeohydrology of the Konya Basin, Turkey, basedon isotope studies of modern hydrology and lacustrinecarbonates. Journal of Paleolimnology 22, 187–204.

Leroy, S., Kazancı, N., _Ileri, O., Kibar, M., Emre, O., McGee, E.,Griffiths, H.I., 2002. Abrupt environmental changes within alate Holocene lacustrine sequence south of the Marmara Sea(Lake Manyas, N-W Turkey): possible links with seismicevents. Marine Geology 190, 531–552.

Lodge, D.M., Kelly, P., 1985. Habitat disturbance and the stability offreshwater gastropod populations. Oecologia 68, 111–117.

Lotter, A.F., 2001. The palaeolimnology of Soppensee (CentralSwitzerland), as evidenced by diatom, pollen, andfossil-pigment analysis. Journal of Paleolimnology 25, 65–79.

Lotter, A.F., Birks, H.J.B., Hofmann, W., Marchetto, A., 1998.Modern diatom, cladocera, chironomid, and chrysophyte cystassemblages as quantitative indicators for the reconstructionof past environmental conditions in the Alps II Nutrients.Journal of Paleolimnology 19, 443–463.

Magnin, G., Yarar, M., 1997. Important Bird Areas in Turkey. DogalHayatı Koruma Dernegi, Istanbul, Turkey.

Matsuzaki, S.S., Usio, N., Takamura, N., Washitani, I., 2007. Effectsof common carp on nutrient dynamics and littoral communitycomposition: roles of excretion and bioturbation. Archiv furHydrobiologie 168, 27–38.

McGowan, S., Leavitt, P.R., Anderson, N.J., Hall, R.I., Jeppesen, E.,Odgaard, B.V., 2005. Changes in algal production andcommunity composition during a shift from a clearwater to aturbid state in two shallow lakes: a palaeolimnologicalapproach. Ecology 86, 2200–2211.

Meisch, C., 2000. Freshwater Ostracoda of Western and CentralEurope. Spektrum Akademischer Verlag, Heidelberg.

Moss, B., Madgwick, J., Phillips, G., 1996. A Guide to theRestoration of Nutrient-enriched Shallow Lakes. BroadsAuthority, Norwich.

Mouthon, J., 1986. General principles for a method of assessingthe overall quality of lacustrine sediments using a simplifiedanalysis of the mollusc community. Annals of Limnology 22,209–217.

Noordhuis, R., van der Molen, D.T., van den Berg, M.S., 2002.Response of herbivorous water-birds to the return of Chara inLake Veluwemeer, The Netherlands. Aquatic Botany 72, 349–367.

Ozesmi, U., Ozesmi, S., 2003. A participatory approach toecosystem conservation: fuzzy cognitive maps andstakeholder group analysis in Uluabat Lake, Turkey.Environmental Management 31, 518–531.

Reed, J.M., 2007. The role of quantitative diatom-basedpalaeolimnology in environmental archaeology. In: Lillie, M.,Ellis, S. (Eds.), Wetland Archaeology and Environments.Regional Issues, Global Perspectives. Oxbow, Oxford, pp. 198–210.

Reed, J.M., Roberts, N., Leng, M.J., 1999. An evaluation of thediatom response to Late Quaternary environmental change intwo lakes in the Konya Basin, Turkey, by comparison withstable isotope data. Quaternary Science Reviews 18, 631–647.

Rioual, P., 2000. Diatom assemblages and water chemistry of lakesin the French Massif Central: a methodology forreconstruction of past limnological and climate fluctuationsduring the Eemian period. Unpublished Ph.D. Thesis,University College London.

Roberts, C.N., Reed, J.M., in press. Lakes, wetlands and holoceneenvironmental change. In: Woodward, J.W. (Ed.), The PhysicalGeography of the Mediterranean Basin. Oxford UniversityPress, Oxford.

Roberts, N., Black, S., Eastwood, W.J., Griffiths, H.I., Lamb, H.F.,Leng, M.J., Parish, R., Reed, J.M., Twigg, D., Yigitbasioglu, H.,1999. Chronology and stratigraphy of Late Quaternarysediments in the Konya Basin, Turkey: results from the KOPALProject. Quaternary Science Reviews 18, 611–631.

Roberts, N., Reed, J., Leng, M.J., Kuzucuoglu, C., Fontugne, M.,Bertaux, J., Woldring, H., Bottema, S., Black, S., Hunt, E.,Karabıyıkoglu, M., 2001. The tempo of Holocene climaticchange in the eastern Mediterranean region: new high-resolution crater-lake sediment data from central Turkey. TheHolocene 11, 721–736.

Ryan, S., 2001. An eco-hydrological assessment of the presentcondition and recent palaeoecology of a Turkish Ramsar site,


Lake Uluabat (NW Turkey). Unpublished M.Res. Thesis,University of Hull.

Salihoglu, G., Karaer, F., 2004. Ecological risk assessment andproblem formulation for Lake Uluabat, a Ramsar State inTurkey. Environmental Management 33, 899–910.

Secmen, O., Leblebici, E., 1996. Wetland Plants of Turkey. EgeUniversity, _Izmir, 870 pp. (in Turkish).

Smol, J.P., 2002. Pollution of lakes and rivers. APaleoenvironmental Perspective. Arnold, London.

Stevenson, A.C., Flower, R.J., 1991. A palaeoecological evaluationof environmental degradation in Lake Mikri Prespa, NWGreece. Biological Conservation 57, 89–109.

Tanatmıs�, M., 2002. The ephemeroptera (insecta) fauna of LakeUlubat Basin. Turkish Journal of Zoology 26, 53–61.

ter Braak, C.J.F., 1990. Update Notes: CANOCO version 3.10.Agricultural Mathematics Group, Wageningen, TheNetherlands, 35 pp.

van Geest, G.J., Coops, H., Roijackers, R.M.M., Buijse, A.D.,Scheffer, M., 2005. Succession of aquatic vegetation driven byreduced water-level fluctuations in floodplain lakes. Journal ofApplied Ecology 42, 251–260.

Van Zeist, W., Bottema, S., 1982. Vegetational history of theEastern Mediterranean and the Near East during the last20,000 years. In: Bintliff, J.L., Van Zeist, W. (Eds.),Palaeoclimates, Palaeoenvironments and HumanCommunities in the Eastern Mediterranean Region in LaterPrehistory. BAR International Series, vol. 133, pp. 277–321.

Williams, D.D., Feltmate, B.W., 1992. Aquatic Insects. C.A.B.International, Wallingford.

Wilson, G.P., Reed, J.M., Lawson, I.T., Frogley, M.R., Preece, R.C.,Tzedakis, P.C., 2008. Diatom response to the last glacial–interglacial transition in the Ioannina basin, northwestGreece: implications for Mediterranean palaeoclimatereconstruction. Quaternary Science Reviews 27, 428–440.

Wunsam, S., Schmidt, R., 1995. A diatom-phosphorus transferfunction for Alpine and pre-alpine lakes. Memorie dell’IstitutoItaliano di Idrobiologica 53, 85–99.

Yılmaz, D.D., 2007. Bioaccumulation and distribution of heavymetals between different compartments of the KızılırmakRiver (Turkey). Fresenius Environmental Bulletin 16, 14–18.

Recent habitat degradation in karstic Lake Uluabat, western Turkey: A coupled...

Documents

Transcript of Recent habitat degradation in karstic Lake Uluabat, western Turkey: A coupled...