Estuarine, Coastal and Shelf Science (2002) 54, 473–499doi:10.1006/ecss.2000.0659, available online at http://www.idealibrary.com on

Modelling the Danube-influenced North-westernContinental Shelf of the Black Sea. II: EcosystemResponse to Changes in Nutrient Delivery by theDanube River after its Damming in 1972

C. Lancelota,e, J. Stanevab, D. Van Eeckhoutc, J.-M. Beckersd and E. Stanevb

aEcologie des Systemes Aquatiques, Universite Libre de Bruxelles, CP 221, Campus Plaine, Bd. du Triomphe,B-1050, Bruxelles, BelgiumbUniversity of Sofia, Department of Meteorology and Oceanography, 5 James Bourchier Str., BG-1126, Sofia,BulgariacBureau Federal du Plan, Avenue des Arts, 47–49, B-1000, Bruxelles, BelgiumdUniversite de Liege, GHER, Sart-Tilman B5, B-4000, Liege, Belgium

Received 23 October 1998 and accepted in revised form 17 December 1999

The ecological model BIOGEN, describing the carbon, nitrogen, phosphorus and silicon cycling throughout aggregatedchemical and biological compartments of the planktonic and benthic marine systems, has been implemented in thenorth-western Black Sea to assess the response of this coastal ecosystem to eutrophication by the Danube River. Thetrophic resolution of BIOGEN was chosen to simulate the major ecological changes reported in this coastal area since the1960s. Particular attention was paid to establishing the link between quantitative and qualitative changes in nutrients,phytoplankton composition and food-web structures. The BIOGEN numerical code structure includes 34 state variablesassembled in five interactive modules describing the dynamics of (1) phytoplankton composed of three distinct groups,each with a different trophic fate (diatoms, nanophytoflagellates, non-silicified opportunistic species); (2) meso- andmicrozooplankton; (3) trophic dead-end gelatinous organisms composed of three distinct groups (the omnivorousNoctiluca and the carnivores Aurelia and the alien Mnemiopsis), and organic matter degradation and associated nutrientregeneration processes by (4) planktonic and (5) benthic bacteria. The capability of the BIOGEN model to simulate therecent ecosystem changes reported for the Black Sea was demonstrated by running the model for the period 1985–1995.The BIOGEN code was implemented in an aggregated and simplified representation of the north-western Black Seahydrodynamics. The numerical frame consisted of coupling a 0-D BIOGEN box model subjected to the Danube with a1-D BIOGEN representing the open-sea boundary conditions. Model results clearly showed that the eutrophication-related problems of the north-western Black Sea were not only driven by the quantity of nutrients discharged by theDanube, but that the balance between them was also important. BIOGEN simulations clearly demonstrated thatphosphate, rather than silicate, was the limiting nutrient driving the structure of the phytoplankton community and theplanktonic food-web. In particular, it showed that a well-balanced N:P:Si nutrient enrichment, such as that observed in1991, had a positive effect on the linear, diatom–copepod food-chain, while the regenerated-based microbial food-chainremained at its background level. When present, the gelatinous carnivores also benefited from this enrichment throughouttheir feeding on copepods. A synergetic effect of fishing pressure and cultural eutrophication was further indirectlysuggested by modifying the mortality coefficient of copepods. However, BIOGEN scenarios with unbalanced nutrientinputs, such as nitrogen or phosphate deficiency recorded in 1985 and 1995, predicted the dominance of an activemicrobial food-web in which bacteria and microzooplankton played a key role; the former as nutrient regenerator, thelatter as a trophic path to the copepods and hence to the carnivorous. In such conditions, however, a significant biomassreduction of all gelatinous organisms was simulated, in perfect agreement with recent observations. From these modelscenarios it is suggested that the observed positive signs of Black Sea ecosystem recovery might well be related to thereduction of nutrient loads in particular phosphate, by the Danube. � 2002 Elsevier Science Ltd. All rights reserved.

Keywords: eutrophication; ecological modelling; Black Sea; Danube

eCorresponding author. E-mail: lancelot@ulb.ac.be

Introduction

The Black Sea, having unique features, such as beingthe largest enclosed catchment basin and receivingfreshwater and sediment inputs from rivers draining

0272–7714/02/030473+27 $35.00/0

half of Europe and parts of Asia, is very sensitive to theprocess of eutrophication. In fact, the Black Sea hasundergone several changes over the last few decades,driven by human perturbations in the coastal eco-system itself and in the drainage basins of the rivers.Among these, the Danube River, as the recipient of

� 2002 Elsevier Science Ltd. All rights reserved.

474 C. Lancelot et al.

the effluents from eight European countries, affectsthe north-western Black Sea ecosystem and representsthe most significant source of river-borne nutrientsflowing into the Black Sea (Tolmazin, 1985). Sincethe early 1960s, noticeable alterations have beenobserved at various trophic levels of the Black Seaecosystem and are well documented. In less than30 years, the Black Sea has evolved from a highlybiologically diverse ecosystem to that of a low bio-diversity, dominated by jellyfish (Gomoiu, 1990; Mee,1992; Bologa et al., 1995). During the late 1980sand early 1990s, there was an almost total collapseof the fisheries industry, which coincided with anunprecedented increase of the jellyfish Aurelia and thecombjelly Mnemiopsis, unintentionally introduced intothe Black Sea in the mid-1980s. At that time, it wasthought that these unpalatable carnivores—feeding onzooplankton, fish eggs and larvae—were responsiblefor dramatically reducing the recruitment of fish to theadult carnivore stocks. A comprehensive analysis ofexisting data (Bologa et al., 1995), in combinationwith idealized ecological modelling (Van Eeckhout &Lancelot, 1997), however, suggests that the explosivedevelopment of jellyfish was the consequence ofdiverse human activities occurring almost synchro-nously in the drainage basin of the Danube River andin the marine system. There are: (1) the manipulationof hydrologic regimes of outflowing rivers, in particu-lar the damming in 1972 of the Danube River by the‘ Iron Gates ’, approximately 1000 km upstream(Bondar, 1977); (2) urban and industrial expansionand the intensive use of agricultural fertilizers (Bologaet al., 1984; Gomoiu, 1990; (3) the introduction ofexotic species, such as Mnemiopsis (Vinogradov &Tumantseva, 1993; Mutlu et al., 1994); and (4)selective and excessive fishing (Ivanov & Beverton,1985; Stepnowski et al., 1993; Bingel et al., 1993;Gucu, 1997). In particular, increased nutrientloads to the Black Sea modified the quantitative andqualitative nutrient environment of the coastal phyto-plankton, increasing nitrogen and phosphorus andlowering silicate (Popa, 1993; Cociasu et al., 1997;Humborg et al., 1997). After 1970, this nutrientchange stimulated the development of numerousphytoplankton blooms in summer and led, in general,to the phytoplankton community being dominated bynon-siliceous species, some of them of little palat-ability for mesozooplankton (Bodeanu, 1992; Bologaet al., 1995). As an immediate response to increasedprimary production, large developments of secondary(copepods) and higher trophic levels (fishes) were firstobserved (Porumb, 1989). The higher production offish led to an increase of fish catches and, as a matterof consequence, of fishing pressure (Zaitsev, 1993;

Gucu, 1997). By decreasing fish stocks, humanactivity indirectly stimulated the growth of the gelati-nous top predators Aurelia and the alien Mnemiopsis,which competed for the same food as the fish, but athigher concentrations. This situation accelerated in anexplosive way due to the lack of known predators ofthese gelatinous carnivores and to their voraciousfeeding on fish eggs and larvae.

Since 1994, however, some positive signs for therecovery of the coastal Black Sea ecosystem have beenobserved. Phosphorus and nitrogen loads to the BlackSea have considerably decreased, but the Danubeconcentrations are comparable to other pollutedEuropean rivers such as the Rhine, Rhone and Po(Cociasu et al., 1997). Some planktonic and benthicspecies considered to be extinct or very rare nowadaysin the Black Sea have again become common(Lancelot et al., 1998). The abundance of jellyfish haslevelled out and the number of anchovy eggs andlarvae has increased (Shiganova, 1997). Incidentally,these improvements in the marine ecosystem corre-spond to the economic decline recorded in Easternand Central European countries since 1991 (Masaryk& Varley, 1997) (not studied here). The link betweenhuman pressures and changes in the marine ecosys-tem is indeed complex and cannot be understood bysimple correlation between historical and ecologicalevents. Mechanistic models, which are based onphysical, chemical and biological principles anddescribe carbon and nutrient cycles as a function ofnatural and anthropogenic pressures, are ideal toolsfor handling this complex system. When validated,these mathematical models are of great value in deter-mining the measures to be taken for recoveringsustainable water quality and biodiversity in thenorth-western Black Sea.

As a first step in the direction, the conceptualecological model BIOGEN (Van Eeckhout &Lancelot, 1997), developed for the analysis of theresistance (or lack of resistance) of the Black Seaecosystem to destabilization, has been implemented inthe Danube-influenced north-western Black Sea. Thishigh trophic-resolution ecological model explicitlydescribes the bottom-up and top-down controls ofthe Black Sea pelagic food-chain, and considers theexchanges of nutrients at the water–sediment inter-face. These properties make it generic and distinctfrom existing, more simple ecological models devel-oped for the surface layer of the Black Sea (Oguz et al.,1996; Cokasar & O} zsoy, 1998; Gregoire et al., 1998).The complexity of the physical features in theDanube-influenced Black Sea continental shelf(Tolmazin, 1985) would ideally require the imple-mentation of BIOGEN in a 3-D frame of high

Modelling the north-western Black Sea: the ecosystem 475

spatio-temporal resolution (Beckers et al., 2000). Thecomplexity of the BIOGEN model makes its directcoupling with the required 3-D physical model scien-tifically and practically unworkable without a detailedanalysis of the ecological features, such as in a moresimple frame. As necessary steps before the imple-mentation of the 3-D BIOGEN in the north-westernBlack Sea, we performed two numerical implementa-tions of the BIOGEN model. Firstly, the BIOGENmodel was coupled with a 1-D vertically resolvedphysical model in order to test the response of theecological model to changing nutrient conditions in aclosed water column. Secondly, the BIOGEN modelwas implemented as a two-box model resulting fromthe coupling between the 1-D vertically resolvedopen-area model and a volume-variable 0-D boxmodel of the coastal area submitted to Danube inputsand bordered at a salinity of 17. This salinity levelcorresponds to the transition between the Danube-influenced waters and the open Black Sea at a salinityof 18·2 (Ragueneau et al., 2000). The capability of themathematical tool to simulate the eutrophication-related changes of the Black Sea ecosystem wasappraised by running scenarios corresponding to three‘ post-Iron Gates ’ periods with contrasting nutrientloads by the Danube River (Table 1). These are theperiods 1980–1985 (nitrogen enriched), 1989–1993(phosphorus increase and silicic acid impoverish-ment), and 1995–1997 (decrease of all nutrient con-centrations, except for ‘ nitrate-excess ’ signature ofthe out-flowing Danube waters).

BIOGEN model description

T 1. Average concentrations of inorganic nutrients at the Danube outflow, calculated for threemajor periods from Cociasu et al.’s (1997) data. Molar ratios are compared with nutrientrequirements of coastal diatoms (after Brzezinsky, 1985)

PeriodDischarge

(km3 yr�1)

Inorganic nutrients Molar ratios

PO4

(�M)SiO

(�M)NO3

(�M)NH4

(�M) N/P Si/P N/Si

1980–1985 210 2·9 69 — 14·8 — 241989–1993 170 4·5 47 232 6·2 54 16 51994–1995 191 2·2 36 178 5·8 84 16 5·2

Coastal diatoms 16 16 1

General structure

The structure of the BIOGEN model—state variablesand processes linking them—is schematically illus-trated in Figure 1 and documented in the Appendix

(Tables A1 and A2). The model describes the cyclingof carbon, nitrogen, phosphorus and silicon throughaggregated chemical and biological compartments ofthe planktonic and benthic systems. Each biologicalcomponent represents a set of different organismsgrouped together according to their trophic level andfunctional ecological behaviour. BIOGEN thusincludes 34 state variables (Appendix, Table A1)assembled in five models. These describe: (1) thegrowth physiology (photosynthesis, growth, exu-dation, respiration) of phototrophic flagellates (NF),diatoms (DA) and opportunistic non-siliceous micro-phytoplankton (OP); (2) the dynamics (grazing,growth, nutrient regeneration, egestion) of thedominant micro- (MCZ) and meso- (COP) zoo-plankton populations; (3) the feeding and growthactivity of the omnivorous giant dinoflagellateNoctiluca (NOC) and the carnivorous Aurelia (AUR)and Mnemiopsis (MNE) gelatinous organisms; (4) thedynamics of organic matter (particulate, POM anddissolved, DOM; each with two classes of biodegrad-ability) degradation by bacteria (BAC) and itscoupling with nutrient regeneration; and (5) thebenthic diagenesis and nutrient release by local sedi-ments. The model is closed by gelatinous organismmortality and by fish pressure. The latter is indirectlyincluded in the model through the mortality of meso-zooplankton (COP) and is described as first-orderkinetics.

All forms of major inorganic nutrients are con-sidered. All phytoplankton groups assimilate NO3,NH4 and PO4. Silicic acid (SiO) is taken up only byDA and released into the surrounding medium afterdiatom lysis and zooplankton feces dissolution. PO4

and NH4, the latter only when the chemical com-position of the bacterial substrate (BS) is N-depleted,are directly used by bacteria (BAC). Both NH4

and PO4 are regenerated through BAC, MCZ,COP, NOC, AUR and MNE catabolic activity. All

476 C. Lancelot et al.

organisms undergo autolytic processes, which releasedissolved (DOM) and particulate (POM) polymericorganic matter, each with two classes of biodegrad-ability, into the water column. Large phytoplankton(DA and OP), detrital particulate organic matter(POM) and zooplankton (COP, NOC, AUR, MNE)fecal pellets undergo sedimentation. Sedimented bio-genic material is pooled as benthic particulate organicmatter (BPOM) with two classes of biodegradability.Benthic nutrient exchanges are calculated fromorganic matter degradation, oxygen consumption andnutrient release and transformation (nitrification/denitrification), taking into account PO4 and NH4

adsorption on particles and mixing processes in theinterstitial and solid phases of the sediment.

NH4 OP DOMi BAC

POMi

NOC

MNE AUR

COP

NF

MCZ

DA

SiO

NO3

PO4

Sediment

F 1. Diagrammatic representation of the structure of the BIOGEN model. Inorganic nutrients include ammonium(NH4), nitrate (NO3), phosphate (PO4) and silicic acid (SiO). Organic matter is composed of dissolved (DOM1,2) andparticulate (POM1,2) matter each with two different biodegradability classes. Phytoplankton is composed of three groups:diatoms (DA), autotrophic nanoflagellates (NF) and opportunists (OP). Bacterioplankton is represented by BAC.Zooplankton includes microzooplankton (MCZ) and copepods (COP). The gelatinous food-chain is composed of Noctiluca(NOC), Aurelia (AUR) and Mnemiopsis (MNE).

Equations and parameterization

The differential equations describing the conservationof state variables are listed in the Appendix (Table

A3). The kinetics of processes relative to changes inmost biological state variables have been describedin extenso elsewhere (Billen & Servais, 1989; Billenet al., 1989; Lancelot et al., 1991, 1997a, b, 2000).Mathematical expressions are listed in the Appendix(Tables A4–A7).

BIOGEN innovation is the representation of thefeeding mode of the gelatinous organisms NOC, AURand MNE. The omnivorous Noctiluca (NOC) feed onall auto- and heterotrophic micro-organisms anddetrital POM. The gelatinous carnivores AUR andMNE eat mesozooplankton (COP). All gelatinousorganisms are ‘ top-predators ’ and no trophic linkexists between them. Thus, lysis is the only mortalityprocess affecting them. The feeding mechanisms ofthe gelatinous organisms have been shown to greatlydiffer from that of copepods and microzooplankton.The latter are described by Monod-type kineticswith saturation of the specific ingestion rate athigh food concentrations (Appendix, Table A5).

Modelling the north-western Black Sea: the ecosystem 477

Conversely, observations (Hoffmeyer, 1990) showthat, above a minimum food threshold, the specificfood ingestion rate of gelatinous organisms (omnivor-ous and carnivorous) increases linearly with respect tofood without saturation at high concentration. Thecorresponding mathematical expression is given in theAppendix (Table A5).

The parameterization procedure constitutes a keystep for the successful implementation of a mech-anistic model and the assessment of its predictioncapability. Most biological parameters were derivedfrom process-level measurements conducted duringthe 1995 and 1997 EROS cruises of the RV ProfessorVodyanitsky, as well as from other relevant literaturedata. Additionally, a large number of numericalprocess-oriented sensitivity studies have been carriedout to understand the model’s response in differentconditions. The sensitivity experiments were directedtowards (1) tuning the set of unknown or poorlydefined biological parameters and (2) investigatingthe response of the ecosystem to the initial conditions.All sensitivity analyses were carried out with the0-D version of the BIOGEN model (Staneva,unpubl. data) and will be described elsewhere.BIOGEN parameters are summarized in theAppendix (Table A4–A7).

Model implementations

1-D BIOGEN

The BIOGEN model was coupled with the verticallyresolved 1-D physical model of Staneva et al. (1998).This model is a 1-D-turbulent hydrodynamical modelthat calculates the thickness of the wind-mixed layerfrom the balance between the kinetic turbulent energyinduced by the wind and the buoyancy input byheating at the surface and the entrainment of heavywaters from below. The model is an adaptation ofthe formulation of Gill and Turner (1976) and isdescribed in extenso in Staneva et al. (1998).

The 1-D BIOGEN model has been run as a closedsystem for several years until reaching a steady statefor all state variables. This was generally achievedafter less than 5-year-runs. The atmospheric forcing(heat flux and wind stress) was obtained from theclimatological data set of Sorkina (1974), corrected byadding high frequency (twice daily) signals obtainedfrom the US National Meteorological Centre for theperiod 1980–1986. Vertical discretization was 5 mand the sediment module was connected to thebottom layer.

Two sets of nutrient initial conditions were used inorder to test the model’s response in two ‘ water

columns ’ representative of the surface open Black Sea(200 m deep), on the one hand, and of the shelf waters(40 m) under the influence of the Danube, on theother. Initial nutrient concentrations (Table 2) werechosen from data collected during the EROS 21cruise of the RV Professor Vodyanitsky during thewinter–spring transition period of 1997.

T 2. 1-D BIOGEN: initial conditions for nutrientconcentrations in the surface layer (source: cruise report ofEROS 21, leg 1 of the RV Professor Vodyanitsky, 1997)

Nutrient (�M) Open sea Shelf area

NO3 2·2 30·7NH4 0·2 1·2PO4 0·1 0·4Si(OH)4 5·5 21·5

0-D 1-D BIOGEN

The numerical frame consisted of the ‘ on-line ’coupling of the ‘ open-sea ’ 1-D BIOGEN with a‘ coastal ’ BIOGEN box model of variable volume.The former model represents the open-sea boundaryconditions as described above. The box model sym-bolizes the coastal area influenced by freshwater andnutrient loads by the Danube River. Two fixedlatitudes, the coast and a moving interface defined bya salinity of 17 determine its geographical limits. Thevolume of the coastal box and its exchange fluxes withthe ‘ open-sea ’ water column were obtained at eachtime-step from the data diagnosed by the 3-D GHERgeneral circulation model constrained by climatologi-cal atmospheric and freshwater (Altman & Kumish,1986) forcing. The procedure is described in extensoin Beckers et al. (2002). In short, the average con-centration of any biogeochemical variable C in thebox model is calculated according to the followingequation:

dC/dt=Pc+Qr/V(Cr�C)+Qin1 /V(C1�C)+Qin

3 /V(C3�C)+A/V(C2�C) (1)

in which Pc results from the internal biogeochemicaltransformations of C calculated by the BIOGENequations, Qr is the Danube River input and Qin

l

and Qin3 represent the north and south water inflow

into box volume V, respectively. The coefficient Aparameterizes the diffusion between the coastal andthe open-sea boxes. C2 is the open-sea surface layerconcentration of C predicted by the ‘ open-sea ’ 1-D

478 C. Lancelot et al.

BIOGEN model. C1 and C3 are the concentrations ofthe northern and southern boundaries of the box(equal to C or C2 depending on the sign of Qin

1 andQin

3 , respectively).Qin

3 , Qin3 , A, V and the average salinity S of the

coastal box were calculated in the coupled modelinteractively at each time-step (1 h) from the 15-daydata diagnosed by the circulation model of Beckerset al. (2000). A zero gradient condition was imposedfor lateral ‘ open-sea ’ boundary conditions of biogeo-chemical variables. River boundary conditions wererestricted to inorganic nutrients. Hourly nutrientinputs by the Danube River were provided from theinterpolation of monthly nutrient concentrationsmeasured at the Danube mouth for the period 1980–1995 (Cociasu et al., 1997) and the climatologicalfreshwater discharges estimated by Altman andKumish (1986). In such conditions, the mathematicaltool is able to test the ecosystem response to changingnutrient concentrations without interfering with theinterannual variations of atmospheric forcing.

Model results

��

��

�

������� ���

�����

� � � � �� �� �� �� �� �� ��

������������

��

����

�

��� ���

� � � � �� �� �� �� �� �� ��

��

��

��

��

��

��

��� �� �� �� �� �� �� ��������� ������

F 2. Two-year simulations of seasonal variations of(a) the vertical profiles of temperature and (b) the uppermixed layer depth.

Modelling seasonal physicochemical changes and biologicalsuccessions in the stratified Black Sea water column:1-D BIOGEN

The seasonal variation of the surface layer physicalstructure of the open Black Sea due to the climato-logical atmospheric forcing is illustrated in Figure 2 bythe vertically resolved temperature [Figure 2(a)] anddepth of the upper mixed layer [Figure 2(b)]. Thesimulated thickness of the upper mixed layer fluctu-ates between 5 and 60 m, which corresponds well withprevious observations (Oguz et al., 1996; Stanevaet al., 1998). Convective mixing is maximal at the endof February, deepening the mixed layer depth up toabout 60 m [Figure 2(b)], with temperature charac-teristic of 6·5–7·0 �C [Figure 2(a)]. Deeper winterconvection is prevented by the shallow, permanentpycnocline that characterizes the Black Sea. Thisfeature results from the strong freshwater dilution ofBlack Sea surface waters and the inflow of denserMediterranean Sea waters through the BosphorusStrait. In spring, the water column warms up gradu-ally and a strong thermocline is predicted at the baseof a shallow, upper mixed layer in mid-summer[5–10 m; Figure 2(b)]. Accordingly, the maximumsea-surface temperature is observed during mid-summer and is about 24 �C [Figure 2(a)], in perfectagreement with observations recorded during theEROS 2000 cruises of July–August 1995 [e.g. Figure5(b)] and by other (e.g. Vedernikov & Demidov,1997).

Figures 3 and 4 show the steady-state time evol-ution of the vertical profiles of nutrients and biologicalstate variables predicted for nutrient conditionsmimicking the ‘ open area ’ and ‘ shelf area ’ watercolumns, respectively. The steep gradients simulatedat 60 m and below (Figure 3) result from the BlackSea’s strong stratification, which prevents the enrich-ment of surface layers with high concentrations ofnutrients from deeper layers.

Reasonable agreement is observed between modelpredictions and data available for the central basin,both seasonally and in magnitude. As an example,Figure 5 compares selected vertical profiles ofBIOGEN simulations and nutrients and chlorophyll aobservations for spring [Figure 5(a)] and summer[Figure 5(b)] periods.

In accordance with the 1997 data (Ragueneau et al.,2000), the predicted steady-state winter nutrient sig-nature of the open Black Sea is ‘ silicate-excess ’, but‘ nitrogen- and phosphate-deficient ’ compared to theSi/N/P molar ratio of 16:16:1, summarizing thestoichiometry of phytoplankton and diatoms (Redfieldet al., 1963; Brzezinski, 1985). Steady-state open-seawinter concentrations of nitrates [Figure 3(a)] are

Modelling the north-western Black Sea: the ecosystem 479

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

���� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

!����"������ #����"�

�����

$!����"����%� &�!���"����'�

��� ����(��")������� �����* ��"��+ ���(��")�����,�

������'��(�����(� ��%��-���"��+ ���(�������

.�% �����(������� ����������(�����/�

������(�����+� � ���������(�����"�

��� ����������������� ��� �����������������

��� ���������� ��� ����������

� �������������� ��� �����������������

��� ���������� ��� ����������

� ������������� � �������������

��� ���������� ��� ����������

��� ����������������������� ��� �����������������������

��� ���������� ��� ����������

��� ����������������������� � ������������������

��� ���������� ��� ����������

��� ����������������� ��� �����������������

��� ���������� ��� ����������

F 3. 1-D BIOGEN predictions in the open-sea water column of the Black Sea obtained after the fifth year-run:(a) nitrates; (b) ammonium; (c) phosphate; (d) silicate; (e) diatom–Chl �; (f) nanoflagellate–Chl a; (g) copepods;(h) microzooplankton; (i) bacteria; (j) Noctiluca; (k) Aurelia; (l) Mnemiopsis.

480 C. Lancelot et al.

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

���� ���

����������

��

�

��� ���

����������

��

�

��� ���

����������

��

���� ���

����������

��

�

��� ���

����������

� ��������������0� � ��������������0�

��� ����0������� ��� ����0�������

� ��������������0� � ��������������0�

��� ����0������� ��� ����0�������

� ��������������� � ���������������

��� ����0������� ��� ����0�������

�1� �1��1��1��1��1��1��1��10�1� �1� ��10�1��1��1��1��1��1��1�

��� ����0������� ��� ����0�������

�1� ��10�1��1��1��1��1��1��1� � ������������������

��� ����0������� ��� ����0�������

� ������������������ �1� ��1���1���1���1��

��� ����0������� ��� ����0�������

!����"�2���� #����"�

2����

$!����"�2��%� &�!���"�2��'�

��� ����(��")��2���� �����* ��"��+ ���(��")��2��,�

������'��(��2��(� ��%��-���"��+ ���(��2����

.�% �����(��2���� ����������(��2��/�

������(��2��+� � ���������(��2��"�

F 4. 1-D BIOGEN predictions in the nutrient-enriched water column obtained after the fifth year-run: (a) nitrates;(b) ammonium; (c) phosphate; (d) silicate; (e) diatom–Chl a; (f) nanoflagellate–Chl a; (g) copepods; (h) microzooplankton;(i) bacteria; (j) Noctiluca; (k) Aurelia; (l) Mnemiopsis.

Modelling the north-western Black Sea: the ecosystem 481

0Silicic acid (µM)

50

–150

Dep

th (

m)

–10–30–50–70–90

–110–130

40302010 0.0Chlorophyll a (mg m–3)

1.0

–150

Dep

th (

m)

–10–30–50–70–90

–110–130

0.80.60.40.2

0Nitrate (µM)

6

–150

Dep

th (

m)

–10–30–50–70–90

–110–130

42 0.0Phosphate (µM)

2.0

–150

Dep

th (

m)

–10–30–50–70–90

–110–130

1.51.00.5

0Temperature (°C)

10

–150

Dep

th (

m)

–10–30–50–70–90

–110–130

8642 17Salinity

21

–150D

epth

(m

)

–10–30–50–70–90

–110–130

201918(a)

17 21201918

0Silicic acid (µM)

50

–150

Dep

th (

m)

–10–30–50–70–90

–110–130

40302010Chlorophyll a (mg m–3)

–150

Dep

th (

m)

–10–30–50–70–90

–110–130

Nitrate (µM)

–150D

epth

(m

)

–10–30–50–70–90

–110–130

Phosphate (µM)

–150

Dep

th (

m)

–10–30–50–70–90

–110–130

0Temperature (°C)

30

–150

Dep

th (

m)

–10–30–50–70–90

–110–130

2010Salinity

–150

Dep

th (

m)

–10–30–50–70–90

–110–130

(b)

0.0 1.00.80.60.40.2

0.0 2.01.51.00.50 642

F 5. 1-D BIOGEN simulations in the open Black Sea water column. Observations recorded in (a) April 1997 and(b) July 1995. ( ), data; (——), model.

typically those characteristic of the central basin(Konovalov, 1999). In contrast, simulated winterconcentrations of phosphate, ammonium and silicicacid are slightly higher than generally observed. Asexpected in stratified marine systems in general and inthe Black Sea in particular (Vedernikov & Demidov,1997; Vinegradov et al., 1999), the model simulatesthe development in mid-March of a phytoplanktonspring bloom composed of diatoms [Figure 3(e)]. Theonset of the spring diatom bloom corresponds tothe installation of the thermocline at 35 m, when thesurface layer temperature is about 8 �C (Figure 2).The bloom occupies the upper mixed layer andreaches a maximum of 1·2 mg Chl-a m�3 [Figure3(e)], in fair agreement with observations [Figure5(a); Vinegradov et al., 1999]. Its magnitude is con-trolled by nutrient resources—nitrate [Figure 3(a)],but mostly phosphate [Figure 3(c)]—rather than bythe grazing pressure of copepods, whose biomassremains at its lowest level until late April [Figure3(g)]. As a result, significant concentrations of silicate[5–10 �M; Figures 3(d) and 5(a)] are leftover at the

decline of early spring diatoms. The disappearance ofearly April diatoms, either because of their autolysis orsedimentation, releases dissolved and particulateorganic matter (not shown) into the water column,which in turn stimulate the growth of bacteria. Thelatter reach a biomass of 10 mg C m�3 approxi-mately, 10 days after the spring diatom bloom [Figure3k,e]. This delay is explained by the macromolecularstructure of diatom-derived substrates that have tobe hydrolysed in monomers prior to being taken upby bacteria (see equations in Appendix). Based onregenerated nitrogen and phosphorus associated withthe bacterial degradation of ungrazed spring diatoms,the model predicts a cascade of auto- and hetero-trophic successions of low amplitude in April–May(Figure 3). Firstly, a rather modest surface layerphytoplankton bloom, almost equally composed ofdiatoms [�0·6 mg Chl-a m�3; Figure 3(g)] andnanophytoflagellates [�0·4 mg Chl-a m�3; Figure3(f)], is simulated in early May. The amplitude ofthese blooms is regulated by phosphate availabilityand by the pressure of their respective grazers; the

482 C. Lancelot et al.

copepods, whose grazing activity on diatoms is stimu-lated by the higher temperature in May [Figure 3(g)],and the ubiquitous microzooplankton [Figure 3(h)].Interestingly, model simulations show a response de-lay between copepods and diatoms [Figure 3(e,g)],and a strong coupling between microzooplanktonand nanophytoplankton, which peak almost synchro-nously [Figure 3(f,h)]. In mid-May, the microzoo-plankton bloom is repressed by the grazing pressure ofcopepods that shift from diatoms to microzooplanktonprey after the decline of the former[Figure 3(e,g,h)].The predicted biomass of copepods [Figure 3(g)]reaches a maximum of about 14 mg C m�3 in Juneand exceeds that of the microzooplankton [Figure3(h)]. The maximum biomass of copepods correlateswith secondary diatom blooms sustained by regener-ated phosphate [Figure 3(c,e,g)], as suggested by theshort-term oscillations of the latter and its co-occurrence with high bacterial biomass [Figure 3(i)].In such conditions of low phosphate availability, thesimulated phosphate pattern of June–July could wellreflect a competition between diatoms and bacteria forhis essential nutrient.

Furthermore, several subsurface, summer diatomblooms are predicted [Figure 3(e)], supported by thesupply of nutrient-rich waters from below under con-ditions of sufficient light. As observed by severalauthors (Sorokin, 1982; Vedernikov & Demidov,1997; Vinogradov et al., 1999), an autumn diatombloom of �0·7 mg Chl-a m�3 is predicted andcorresponds to the increased vertical mixing withnutrient-rich deeper waters [Figure 2(b)]. In contrastwith early spring, the autumn diatom bloom sustainssome copepod growth [Figure 3(e,g)]. The simulatedbiomass of all gelatinous organisms—Noctiluca, Aure-lia, Mnemiopsis—is not very significant [less than5 mg C m�3; Figure 3(j,k,l)] due to the low concen-tration of their food resources in comparison withtheir needs. Noctiluca peak in May and July–augustwhen microzooplankton and/or nanophytoplanktonare abundant [Figure 3(f,h,j)]. Aurelia and Mnemiopsisdevelop in late summer–early autumn when copepodsreach their maximum biomass [Figure 3(g,k,l)]. Thepredicted biomass of both gelatinous carnivores is afactor of two lower compared to observations in 1992(Vinogradov et al., 1999).

In summary, 1-D BIOGEN simulations in the openBlack Sea indicate that the surface layer planktonicsystem is driven by winter phosphate availability,which determines the magnitude and extent of theearly spring diatom bloom. The mismatch predictedbetween early spring diatoms and copepod grazinglimited by low temperature is, via the microbial deg-radation of organic matter derived from ungrazed

diatoms, at the basis of a complex nutrient-regenerating food-web with strong interactions. Asexpected in summer in oceanic waters, this food-webincludes an active microbial network (bacteria, nano-phytoplankton and microzooplankton) in addition todiatoms and copepods. In this silicon-excess system,the summer blooms of diatoms are regulated byphosphate availability. The success of copepods insummer–autumn is explained by their ability to switchfrom diatoms to microzooplankton prey and viceversa, as well as by the lack of control by food-limitedgelatinous carnivores. Finally, it must be stressed thatthe predicted silicate never reached a limiting leveland was always higher than predicted nitrates, whichis consistent with observations (Konovalov, 1999;Figure 5).

BIOGEN predictions in the ‘ Danube-influencedshelf ’ water column, mimicking the 1997 nutrientconditions (Figure 4), show a quite different steady-state winter nutrient regime due to the differentsignatures of river and marine waters. Steady-statewinter concentrations are strongly enriched in nitro-gen and phosphate and show an excess of nitrate withrespect to silicon and phosphorus when compared tothe phytoplankton stoichiometry [Figure 4(a,b,c,d)].As in the open Black Sea, a diatom bloom is simulatedin early March, but is three times higher in biomass[Figure 4(e)]. Again, this early spring diatom bloom isseverely limited by ambient phosphate [Figure 4(c,e)]and is apparently poorly grazed by copepods, whosebiomass is at its lowest level, although still more thantwice that of the open Black Sea [Figures 4(g) and3(g)]. Furthermore, the bacterial regeneration pro-cesses associated to the degradation of ungrazeddiatoms stimulate an active microbial network,with bacteria and microzooplankton reaching a sig-nificantly higher biomass compared to copepods[Figure 4(g,h)] and to similar predictions in the openBlack Sea [Figure 3(g,h)]. Although not directlycomparable, these simulations are qualitatively andquantitatively supported by the 1995 microbiologicalobservations in the north-western Black Sea (Bouvier,1998). Model predictions show two peaks of micro-zooplankton separated by a huge bloom of Noctilucalasting from mid-May to the end of June [Figure 4(l)].The maximum biomass simulated [�30 mg C m�3;Figure 4(j)] is higher by far than that of copepods andmicrozooplankton, and equivalent to bacteria, all inperfect agreement with 1995 observations (Bouvier,1998; Weisse et al., 2002). At this time, the watercolumn is literally cleaned of other micro-organisms(Figure 4), which is explained by the voracious feed-ing of this omnivorous gelatinous dinoflagellate. Inthe absence of significant phytoplankton biomass, a

Modelling the north-western Black Sea: the ecosystem 483

transient accumulation of the ammonium released byNoctiluca catabolic losses is predicted at Noctilucamaximal biomass [Figure 4(b,e,f,j)]. Such an accumu-lation is not simulated for phosphate due to the stronglimitation of this nutrient. Compared to the openBlack Sea simulations, where carnivorous gelatinousorganisms were quasi-absent [Figure 3(m,n)], signifi-cant biomasses of both Aurelia and Mnemiopsis aresimulated in this nutrient-enriched water columnover the whole season, but mostly during thesummer–autumn period [Figure 4(m,n)].

In summary, model predictions clearly suggest thatthe unbalanced nutrient enrichment of the Black Sea,mimicking the 1997 situation, stimulates the diatomcomponent of the phytoplankton community as longas phosphate is not limiting. The diatom–copepodlinear food-chain is also enhanced, but its extent iscontrolled by the strong feeding pressure of the gel-atinous carnivores, which maintains copepods at abiomass of between 5 and 10 mg C m�3. Conse-quently, much of the diatom production is notgrazed and a very active microbial network develops,sustained by organic matter derived from ungrazeddiatoms. High biomass is predicted for all micro-organisms that allow for the development ofNoctiluca in an explosive way. Finally, it is interestingto note that the maximal biomass predicted for thedifferent gelatinous organisms is quite similar—�30 mg C m�3—although delayed in time [Figure4(j,k,l)].

Modelling the response of the north-western Black Seaecosystem to changes in nutrient delivery by the DanubeRiver after its damming in 1972: coupling a‘ Lagrangian-like ’ BIOGEN box model to 1-DBIOGEN

The previous comparative analysis of 1-D BIOGENresults, constrained by climatological atmosphericforcing and nutrient conditions mimicking thosetypical of the open Black Sea and the Danube–BlackSea mixing zone, indicates that this model is suitablefor simulating the response of the coastal ecosystemto changing nutrient loads. The capability of theBIOGEN model to simulate the ecological changestaking place in the north-western Black Sea over thelast decade in response to changing nutrient deliveryby the Danube was further investigated. For thispurpose, the coupled ‘ box-1-D-water-column ’ BIO-GEN has been run under three contrasting nutrientloads by the River Danube, corresponding to thepost-Danube damming by the Iron Gates (Table 1).This period was chosen because of the availability ofthe complete sets of nutrient concentrations at the

Danube out-flow (Cociasu et al., 1977). Before 1985,some nutrient forms were not measured.

Model results verification was first conducted in theDanube-Black Sea mixing zone by constraining themodel with the Danube nutrient concentrations of1995 (Figure 6). This year was chosen because fieldobservations are available for most state variables,although they are restricted to the summer periodwhen heterotrophs are reaching their maximum bio-mass [Figure 6(b)] and primary production is sus-tained by regenerated nutrients. Table 3 comparesnutrient and biological data collected in the surfacewaters of the 12–17 salinity transition zone during leg1 of the EROS 2000 cruise of the RV ProfessorVodyanitsky in July 1995 and BIOGEN model resultsaveraged over the 1-month cruise duration. Such acomparison was chosen due to the inherent differencebetween the cruise (grid along eutroph–oligotrophgradients) and model (homogeneous box) samplingmode. Examination of Table 3 shows a reasonablecorrelation between predictions and observations. Asa general trend, however, predictions are closer to thecorresponding minimal than average field values(Table 3). This is explained by the higher averagesalinity of the field grid compared to that of the boxmodel (Table 3). From this it can be concluded thatmodel runs with the coupled ‘ box-1-D-water-column ’ BIOGEN are simulating chemical and eco-logical changes close to the frontal area between theDanube and the open water.

Figure 6 compares the seasonal predictions ofnutrients and biological variables in the Danube–Black Sea mixing zone obtained by running thecoupled ‘ box-1-D-water-column ’ BIOGEN underthe forcing of the Danube nutrient loads of 1985,1991 and 1995 [Figure 6(a); Cociasu et al., 1997]. Asignificant interannual and seasonal variability ofDanube nutrient concentrations distinguishes thesethree years [Figure 6(a)]. The fluctuations reflecthuman-induced changes in the Danube watershedand/or different biogeochemical transformations inthe river system (Garnier et al., 2002). This haddramatic results for nutrient concentrations at theDanube out-flow in the Black Sea, which weremodified both quantitatively and qualitatively[Figure 6(a)]. Phosphate was the most changeablenutrient, peaking at tremendously high concentrationsin 1991 [Figure 6(a)]. As a general trend, nitrate andphosphate Danube concentrations in 1995 were attheir lowest level for this ‘ post-Iron Gates ’ period.No real trend can be concluded from silic acid fluc-tuations shown by Figure 6(a), although river loadsof silicon have been reported to decrease after theDanube damming (Humborg et al., 1997). BIOGEN

50

012

120

Time (months)1

100

80

60

40

20

111098765432

Danube SiO (mmol m–3)

0

Time (months)1

45403530252015105

111098765432

SiO (mmol m–3)

012

10

1

98765432

111098765432

Danube PO4 (mmol m–3)

0

2.5

1

2

1.5

1

0.5

111098765432

PO4 (mmol m–3)

012

450

1

40035030025020015010050

111098765432

Danube NO3 (mmol m–3)

0

45

1

403530252015105

111098765432

NO3 (mmol m–3)(a)

1

35

012

40

Time (months)1

3530252015105

111098765432

Aurelia (mg C m–3)

012

30

Time (months)1

25

20

15

10

5

111098765432

Mnemiopsis (mg C m–3)

012

50

Time (months)1

45403530252015105

111098765432

Noctiluca (mg C m–3)

0121

30

25

20

15

10

5

111098765432

Copepods (mg C m–3)

012

25

1

20

15

10

5

111098765432

Microzooplankton (mg C m–3)

012

50

1

15105

111098765432

Bacteria (mg C m–3)

012

10

1

987654321

111098765432

Diatoms (mg Chl-a m–3)

012

10

1

987654321

111098765432

Nanophytoplankton (mg Chl-a m–3)

012

10

1

987654321

111098765432

Opportunists (mg Chl-a m–3)(b)

2520

3530

4540

F 6. Time evolution of (a) Danube nutrient forcing and corresponding 0–1-D BIOGEN nutrient simulations in theDanube–Black Sea mixing zone and (b) 0–1-D BIOGEN predictions of nutrients and biological state variables inthe Danube–Black Sea mixing zone. (——), 1985; (– – –), 1991; (· · ·), 1995.

Modelling the north-western Black Sea: the ecosystem 485

T 3. Comparison between model predictions in the Danube–Black Sea mixing zone and datacollected during the EROS 2000, leg 1 cruise of the RV Professor Vodyanitsky, in July 1995

State variable

Field data

BIOGENMean Min. Max.

Salinity 14·52 12·55 16·59 15·7NO3

a (�M) 2·54 0·0 15·32 5·9NH4

b (�M) 0·87 0·24 2·42 1·53PO4

c (�M) 0·14 0·10 0·4 0·15Si(OH)4

d (�M) 5·33 1·96 12·41 8·89Phytoplanktone (mg Chl a m�3) 3·42 0·47 6·28 1·44Microzooplanktonc (mg C m�3) 115·2 9·11 224·6 12·35Bacteriaf (mg C m�3) 34·68 11·1 53·9 18·3Copepodsg (mg C m�3) 8·62 6·94Noctilucag (mg C m�3) 8·37 6·31

aA. Krastev (in EROS cruise report, 1995); bJ. Vervlimmeren (in EROS cruise report, 1995); cR. Pencheva (inEROS cruise report, 1995); dL. Popa (in EROS cruise report, 1995); eS. Moncheva (in EROS cruise report,1995); fT. Bouvier (1998); gB. Alexandrov, T. Weisse and U. Scheffel (unpubl. data).

simulations in the Danube–Black Sea mixing zonereproduce the observed interannual variability ofwinter concentrations of nitrates and phosphates inthe Danube [Figure 6(a)]. However, due to mixingwith silicate-rich marine waters, the predicted winternutrient balance is significantly different from that ofthe river [Figure 6(a)]. Consequently, a higher winterstock of nutrients is predicted in the Danube–BlackSea mixing zone in 1991 compared to 1985 and 1995.Moreover, the simulated winter nutrients of 1991are relatively well balanced with respect to coastaldiatom and phytoplankton stoichiometry [Figure6(a)]. Predicted winter concentrations of nitratesand silicate are very comparable in 1985 and 1995.Interestingly, the model predicts severe phosphatedepletion in 1995 [Figure 6(a)]. Nutrient simulationsin the Danube–Black Sea mixing zone show similarseasonal fluctuations between years with the lowestvalues occurring in the spring–summer period due tobiological uptake [Figure 6(a,b)]. However, the simu-lated delay of the nutrient spring decreases suggests,in conditions of climatological atmospheric forcing, avery sensitive response from the phytoplankton com-munity and related heterotrophs to changing nutri-ents. Accordingly, contrasting phytoplankton andheterotrophc successions are simulated for the threedifferent years [Figure 6(b)]. As a general trend, awell-balanced nutrient enrichment, such as that of1991, is shown to stimulate the growth of all com-ponents of the phytoplankton community [Figure6(b)]. The most visible effect on the planktonic food-chain is the enhancement in early spring to its linearbranch composed of diatoms and copepods, the latter

blooming after the diatoms following a 1 month delay[Figure 6(b)]. Moreover, these nutrient-rich con-ditions also stimulate the gelatinous organisms—Noctiluca, Aurelia, Mnemiopsis—which reach, insummer, a predicted biomass level similar to that ofthe copepods [Figure 6(b)]. In contrast, the modelpredicts little development of the microbial networkcomposed of bacteria and microzooplankton, allowingnanophytoplankton to reach a non-negligible biomassthat persists during the whole vegetative period[Figure 6(b)]. On the other hand, the phosphate-deficient nutrient conditions in 1985 and 1995 arepredicted to enhance the microbial network relative tothe linear food-chain and as much as phosphate isstrongly depleted [Figure 6(b)]. Bacteria, in particu-lar, reach a tremendously high biomass and appearto play a key role in phosphate regeneration pro-cesses. Incidentally, such high bacterial biomass wasmeasured during the 1995 EROS 2000 cruise(Bouvier, 1998). A lower biomass of gelatinouscarnivores is simulated in 1985 and 1995 compared tolevels in 1991, which corresponds with a lower simu-lated copepod biomass [Figure 6(b)]. Also, in 1985and 1995, the predicted biomass of Noctiluca is sig-nificantly lower than in 1991. Although the simulatedNoctiluca biomass is similar for both years, the modelpredicts a 1-month shift in their maxima, with Noc-tiluca blooming in mid-June and mid-July in 1995 and1985, respectively. As a general trend, the modelpredicts a strong interannual variability in thetiming of autotrophic and heterotrophic successions(Figure 6), largely driven by changing nutrients, sincelight and temperature conditions remain the same.

486 C. Lancelot et al.

Discussion

012

50

Time (months)1

40

30

20

10

111098765432

Bacteria (mg C m–3)

012

60

Time (months)1

50

40

30

20

10

111098765432

Bacteria (mg C m–3)

012

20

1

15

10

5

111098765432

Copepods (mg C m–3)

012

20

1

15

10

5

111098765432

Copepods (mg C m–3)

012

8

1

7654321

111098765432

Diatoms (mg Ch-a m–3)

012

8

1

7654321

111098765432

Diatoms (mg Chl-a m–3)

F 7. Sensitivity of the 0–1-D BIOGEN predictions for 1991 in the Danube–Black Sea mixing zone to the coefficientsdescribing the exchanges between the coastal area and the adjacent zones. Left panel: sensitivity to the exchange coefficientbetween the coastal and open-sea areas; right panel: sensitivity to the southern and northern inflow fluxes.

Sensitivity of the BIOGEN predictions to hydrodynamics

Prior to its further ‘ on-line ’ coupling with a 3-DGeneral Circulation Model of the Black Sea, thehigh-trophic resolution ecological model BIOGENhas been numerically implemented in the north-western Black Sea by coupling a 0-D ‘ Lagrangian-like ’ BIOGEN box model subjected to the Danubewith a 1-D BIOGEN representing the open-seaboundary conditions. The extent to which this aggre-gated and simplified representation of the complexhydrodynamics prevailing in this shelf area is realisticwith respect to the eutrophication-related questionaddressed by the mathematical model has been inves-tigated by conducting sensitivity studies of BIOGENpredictions to physical fluxes between the Danube–Black Sea mixing zone and the adjacent open-seaareas. Figure 7 shows BIOGEN simulations of threekey biological state variables obtained when changing(1) the exchange coefficient between the shelf box andthe open-sea water column (Figure 7, left panel) and

(2) the water inflow from the north and south, takingthe mixing coefficients that specify the errors in esti-mating the fluxes into consideration (Figure 7, rightpanel). The exchange coefficients and fluxes, and theirsensitivity to the physical state variables, are describedin extenso in Beckers et al. (2002). Comparing thesesimulations with those obtained by changing Danubenutrient loads (Figure 6) suggests little sensitivity ofBIOGEN predictions to the parameterization of theexchange coefficient between the shelf and open areas.This indicates that the current numerical implemen-tation can be used as a first approach to explore theresponse of the north-western Black Sea ecosystem tochanges in Danube nutrient delivery. Higher sensi-tivity of BIOGEN predictions is, however, obtainedby changing the mixing of inflowing waters (Figure 7).This suggests that the inflow conditions from the Rimcurrent have a marked influence on the ecosystemdynamics of the Danube–Black Sea mixing zone,bordered at a salinity of 17. Interestingly, however, thechanging inflow conditions mainly affect the magni-tude of the state variables, with little effect on their

Modelling the north-western Black Sea: the ecosystem 487

time evolution (Figure 7, right panel). In contrast,changing the Danube nutrient inputs to the Black Seaover the season has a marked incidence on the mag-nitude, bloom extent and time appearance of allbiological state variables (Figure 7). This differenceis consistent with the chosen proportional changeof inflow conditions. Further implementation ofBIOGEN in a high-resolution 3-D model of the shelfdynamics, nested into the GHER general circulationmodel of the Black Sea (Beckers et al., 2002), wouldprovide an accurate estimation of the combined influ-ence of the Danube and the Rim current on thestructure and functioning of the north-western BlackSea shelf ecosystem.

012

60

Time (months)1

555045403530252015105

111098765432

Aurelia (mg C m–3)

012

60

Time (months)1

555045403530252015105

111098765432

Mnemiopsis (mg C m–3)

012

50

Time (months)1

45403530252015105

111098765432

Noctiluca (mg C m–3)

012

45

1

403530252015105

111098765432

Copepods (mg C m–3)

012

6

1

5

4

3

2

1

111098765432

Microzooplankton (mg C m–3)

012

20

1

15

10

5

111098765432

Bacteria (mg C m–3)

012

10

1

987654321

111098765432

Diatoms (mg Chl-a m–3)

012

10

1

987654321

111098765432

Nanophytoplankton (mg Chl-a m–3)

012

10

1

987654321

111098765432

Opportunists (mg Chl-a m–3)

F 8. Sensitivity of the 1991 BIOGEN predictions to fishing pressure obtained indirectly by changing copepod mortalityto fish predation. (– – –), current 1991 prediction; (——), twofold decrease of copepod mortality by fish pressure.

Nutrient changes and trophic structures

Model scenarios of changing Danube nutrientinputs to the north-western Black Sea observed overthe 1985–1995 period show that the mechanisticBIOGEN model, based on food-chain structure andphysiological concepts, has the required trophic reso-lution to address the ecological changes evident in the

Black Sea since the 1960s. Model result analysesindicate that coastal eutrophicated-related problemsare not only driven by the quantity of nutrientsdischarged into the coastal system, but that the bal-ance between them is just as important. BIOGENpredictions clearly demonstrate that limiting nutrientsdetermines the structure of the phytoplankton com-munity, which in turn constrains the structure andfunctioning of the planktonic food-web. In particular,it shows that a well-balanced N:P:Si nutrient enrich-ment, as observed, for example, in 1991, has a positiveeffect on the diatom–copepod linear food-chain, whilethe regenerated-based microbial food-chain remainsat its basic level. When present in the system, gelati-nous carnivores also benefit from this enrichmentthrough their feeding on the increased copepod bio-mass. Nitrogen or phosphate limitation, on the otherhand, directs the structure of the planktonic food-webtowards the dominance of an active microbial food-web in which bacteria and microzooplankton play akey role; the former as nutrient regenerator, the latteras a trophic path to the copepods and hence to thelinear food-chain. In the typically ‘ silicate-excess ’

488 C. Lancelot et al.

Black Sea ecosystem, phosphate rather than nitrogenconstrains the structure and functioning of the BlackSea ecosystem. Hence, the successful development ofgelatinous carnivores is inversely correlated to phos-phate limitation. Therefore, the observed positivesigns of recovery of the Black Sea ecosystem (Lancelotet al., 1998) might well be related to the reduction ofnutrient loads in particular phosphate, by the Danube.

More generally, model results indicate that thecurrent concept of diatoms as better competitors forinorganic nitrogen and phosphate when there is suf-ficient silicon—the general concept behind the currentunderstanding of anthropogenic eutrophication—doesnot hold. BIOGEN predictions, in agreement withrecent field observations (Ragueneau et al., 2002),clearly suggest that diatom growth in the coastal andopen Black Sea has been regulated over the lastdecade by ambient phosphate in spring and nitro-gen and phosphate in summer; with silicon neverdepleted. This contrasts with the current understand-ing of the eutrophication phenomenon of the BlackSea attributed to a combination of a concomitantincrease in nitrogen but mostly phosphorus riverineloads and a decrease in silicate delivery (Cociasu et al.,1997), the latter of which is attributed to the Danubedamming in 1972 (Humborg et al., 1997).

Eutrophication, fishing pressure and trophic structures

The BIOGEN model has further been used to test therecent hypothesis of Gucu (1997, 2002) on the crucialrole of overfishing rather than man-made eutrophi-cation as being responsible for the successful develop-ment of gelatinous carnivores in the north-westernBlack Sea in the late 1980s–early 1990s. The influ-ence of the fisheries industry on the blooming ofgelatinous carnivores was tested by running BIOGENwith the Danube nutrient loads of 1991 and changingthe fishing coefficient. The latter was indirectly con-sidered by modifying the first-order mortality coef-ficient of copepods, where a lower value correspondsto a higher fish pressure. Model simulations (Figure8) suggests, under conditions of well-balancednutrient enrichment, a positive link between fishingpressure and gelatinous carnivores. A greater thantwo-fold increase of the biomass of both carnivorousgelatinous organisms is predicted for a doubling offishing pressure (Figure 8). Interestingly, no change ispredicted among the whole auto- and heterotrophiccommunity (Figure 8). Together, this first assaysuggests that overfishing, in addition to eutrophi-cation, could have played a role in the destabilizationof the Black Sea ecosystem reported for the years1989–1991. This demonstrates the capability of

mechanistic models to handle the complex and non-linear nature of the link between human activities andthe functioning of the Black Sea ecosystem.

Acknowledgements

This work is a contribution to the EROS 21project funded by the programmes Environment andClimate (contract no. ENV4-CT96-0286) andINCO-Copernicus (contract IC20-CT96-0065) ofthe European Commission. It is publication no. 185of the ELOISE initiative. The authors would like tothank Dr J.-M. Martin, Prof. N. Panin and Prof. V.Egorov, whose dedication has made the collaborationbetween Western and Eastern European institutions asuccessful one. We are grateful to T. Weisse and B.Alexandrov for helpful discussions about the par-ameterization of zooplankton feeding. Our thanks toT. Bouvier, A. Cociasu, A. Krastev, S. Moncheva,R. Pencheva, L. Popa and U. Scheffel for providingfield data. Finally, we thank two anonymous reviewersfor their critical and constructive comments.

References

Altman, E. N. & Kumish, N. I. 1986 Interannual andseasonal variability of the Black Sea freshwater balance. TrydiGosudarstvenae Oceanographicheskogo Institute 145, 3–15 (inRussian).

Beckers, J.-M., Gregoire, M., Nihoul, J. C. J., Stanev, E., Staneva,J. & Lancelot, C. 2002 Modelling the Danube-influenced north-western continental shelf of the Black Sea. I: Hydrodynamicalprocesses simulated by 3-D and box models. Estuarine, Coastaland Shelf Science 54, 453–472.

Billen, G. & Servais, P. 1989 Modelisation des processus dedegradation bacterienne de la matiere organique en milieuacquatique. In Micro-organismes dans les Eucosystemes Oceaniques(Bianchi, M., Marty, D., Bertrand, J.-C., Caumette, P. &Gauthier, M.). Paris, Masson, pp. 219–245.

Billen, G., Dessery, S., Lancelot, C. & Meybeck, M. 1989 Seasonaland inter-annual variations of nitrogen diagenesis in the sedi-ments of a recently impounded basin. Biogeochemistry 8, 73–100.

Bingel, F., Kideys, A., Ozsoy, E., Turgul, S., Basturk, O. & Oguz,T. 1993 Stock assessment studies for the Turkish Black Seacoast. NATO-TU Fisheries Final Report. IMS, METU, 108 pp.

Bodeanu, N. 1992 Algal blooms and development of the mainphytoplanktonic species at the Romanian Black Sea littoral inconditions of intensification of the eutrophication process. InMarine Coastal Eutrophication: The Response of Marine TransitionalSystems to Human Impact. Problems and Perspectives for Restoration(Vollenweider, R. A., Marchetti, R. & Viviani, R., eds). Science ofthe Total Environment Suppl. 1992, 891–906.

Bologa, A. S., Skolka, H. V. & Frangopol, P. T. 1984 Annual cycleof planktonic primary productivity off the Romanian Black Sea.Marine Ecology Progress Series 19, 25–32.

Bologa, A. S., Bodeanu, N., Petran, A., Tiganus, V. & Zaitsev, Y.1995 Major modifications of the Black Sea benthic andplanktonic biota in the last three decades. Bulletin de l’InstitutOceanographique de Monaco 15, 85–110.

Bondar, C. 1977 Changes in hydrological pattern induced byengineering works on the lower Danube. Hidrotecnica 22, 87–89.

Modelling the north-western Black Sea: the ecosystem 489

Bouvier, T. 1998 Structure et dynamique du reseau trophiquemicrobien dans le bassin nord occidental de la Mer Noire sousinfluence du Danube. Ph.D. Dissertation, Universite Libre deBruxelles, Belgium.

Brzezinsky, A. M. 1985 The Si:C:N ratio of marine diatoms:interspecific variability and the effect of some environmentalvariables. Journal of Phycology 2, 347–357.

Cokasar, T. & O} zsoy, E. 1998 Comparative analysis and modellingfor regional ecosystems of the Black Sea. In Proceedings of theSymposium on the scientific results of the NATO-TU Black SeaProjects (Oguz, T. & Ivanov, L., eds). NATO-ASI, KluwerAcademic Publisher, pp. 323–359.

Cociasu, A., Diaconu, V., Popa, L., Nae, I., Dorogan, L. & Malciu,V. 1997 The nutrient stock of the Romanian Shelf of the BlackSea during the last three decades. In Sensitivity to Change: BlackSea, Baltic Sea and North Sea (Ozsoy, E. & Mikaelyan, A., eds).NATO ASI Series, 2—Environment 27, 49–65.

Garnier, J., Billen, G., Hannon, E., Fonbonne, S., Videdina, Y. &Soulie, M. 2002 Modelling the transfer and retention of nutrientsin the drainage network of the Danube River. Estuarine, Coastaland Shelf Science 54, 285–308.

Gill, A. & Turner, J. S. 1976 A comparison of seasonal thermo-cline models with observations. Deep Sea Research 21, 391–401.

Gomoiu, M. T. 1990 Marine eutrophication syndrome in thenorth-western part of the Black Sea. In Marine Coastal Eutrophi-cation: The Response of Marine Transitional Systems to HumanImpact. Problems and Perspectives for Restoration (Vollenweider,R. A., Marchetti, R. & Viviani, R., eds) Science of the TotalEnvironment Suppl. 1992, 683–703.

Gregoire, M., Beckers, J.-M., Nihoul, C. J. C. & Stanev, E. 1998Reconnaissance of the main Black Sea’s ecohydrodynamics bymeans of a 3-D interdisciplinary model. Journal of Marine Systems16, 85–105.

Gucu, A. C. 1997 Role of fishing in the Black Sea ecosystem. InSensitivity to Change: Black Sea and North Sea (Ozsoy, E.& Mikaelyan, A., eds) NATO ASI Series, 2—Environment 27,149–162.

Gucu, A. C. 2002 Can overfishing be responsible for the successfulestablishment of Mnemiapsis leidyi in the Black Sea? Estuarine,Coastal and Shelf Science 54, 439–451.

Hoffmeyer, M. S. 1990 Alganus observaciones sobre la alimen-tation de Mnemiopsis mccradyi Mayer, 1990 (Ctenophora,Lobata). Set. Zool. Porto Alegre 70, 55–65.

Humborg, C. 1997 Primary productivity regime and nutrientremoval in the Danube Estuary. Estuarine, Coastal and ShelfScience 45, 579–589.

Humborg, C., Ittekkot, V., Cociasu, A. & Bodungen, B. 1997Effect of Danube River dam on Black Sea biogeochemistry andecosystem structure. Nature 386, 385–388.

Ivanov, L. & Beverton, R. J. H. 1985 The fisheries resources of theMediterranean. Part 2: Black Sea. G.F.C.M. Studies and Reviews60, FAO, Rome, 135 pp.

Konovalov, S. K., Ivanov, L. I., Murray, J. W. & Eremeeva, L. V.1999 Eutrophication: a plausible cause for changes in hydro-chemical structure of the Black Sea anoxic layer. In EnvironmentalDegradation of the Black Sea: Challenges and Remedies (Besiktepe,S. T., Unluata, U. & Bologa, A. S., eds) NATO Science Series,2—Environmental Security 56, 61–74.

Lancelot, C., Veth, C. & Mathot, S. 1991 Modelling ice-edgephytoplankton bloom in the Scotia–Weddell Sea sector of theSouthern Ocean during spring 1988. Journal of Marine Systems 2,333–346.

Lancelot, C., Panin, N. & Martin, J.-M. 1998 The north-westernBlack Sea: a pilot site to understand the complex interactionbetween human activities and the coastal environment. In(Barthel, K.-G., Barth, H., Bohle-Carbonell, M., Fragakis, C.,Lipiatou, E., Martin, P., Ollier, G. & Weydert, M., eds). ThirdEuropean Marine Science and Technology Conference—Project Syn-opsis, Vol. II: Strategic Marine Research, European Commission(Luxembourg) EUR 18220 EN, pp. 706–713.

Lancelot, C., Rousseau, V., Billen, G. & Van Eeckhout, D. 1997aCoastal eutrophication of the Southern Bight of the North Sea:assessment and modelling. In Sensitivity to Change: Black Sea,Baltic Sea and North Sea (Ozsoy, E. & Mikaelyan, A., eds).NATO ASI Series, 2—Environment 27, 439–454.

Lancelot, C., Becquevort, S., Menon, P., Mathot, S. & Dandois,J.-M. 1997b Ecological modelling of the planktonic microbialfood-web. In Belgian Research Programme on the Antarctic,Scientific Results of Phase III (1992–1996) (Caschetto, S., ed.) 1,1–78.

Lancelot, C., Hannon, E., Becquevort, S., Veth, C. & de Baar,H. J. W. 2000 Modeling phytoplankton blooms and carbonexport production in the Southern Ocean: Dominant controls bylight and iron of the Atlantic sector in Autral spring 1992. DeepSea Research I 47, 1621–1662.

Masaryk, T. G. & Varley, I. 1997 Nutrient balances for Danubecountries. Danube Applied Research Program (95-0614.00,PHARE: ZZ911/0102), Final Report. 98 pp.

Mee, L. D. 1992 The Black Sea in crisis: a need for concertedinternational action. Ambio 21, 278–286.

Mutlu, E., Bingel, F., Gucu, A. C., Melnikov, V. V., Nierman, U.,Ostr, N. A. & Zaika, V. E. 1994 Distribution of the new invaderMnemiopsis sp. and the resident Aurelia aurita and Pleurobrachiapileus populations in the Black Sea in the years 1991–1993. ICESJournal of Marine Sciences 51, 407–421.

Oguz, T., Ducklow, H., Malanote-Rizzoli, P., Turgul, S., Nezlin,N. & Unluata, U. 1996 Simulation of annual plankton cycle inthe Black Sea by a one-dimensional physical biological model.Journal of Geophysical Research 101, 16551–16569.

Popa, A. 1993 Liquid and sediment inputs of the Danube River intothe north-western Black Sea. In Transport of Carbon and Nutrientsin Lakes and Estuaries, Part 6. SCOPE/UNEP Sonderband,pp. 137–149.

Porumb, F. 1989 On the development of Noctiluca Scintillans undereutrophication of Romanian Black Sea coastal waters. In MarineCoastal Eutrophication: the Response of Marine TransitionalSystems to Human Impact. Problems and Perspectives for Restoration(Vollenweider, R. A., Marchetti, R. & Viviani, R., eds). Science ofthe Total Environment Suppl. 1992, 907–920.

Ragueneau, O., Egorov, V., Vervlimmeren, J., Cociasu, A., Deliat,G., Krastev, A., Daoud, N., Rousseau, V., Popovitchev, W.,Brion, N., Popa, L., Gauwet, G. & Lancelot, C. 2002 Biogeo-chemical transformations of inorganic nutrients in the mixingzone between the Danube River and the north-western Black Sea.Estuarine, Coastal and Shelf Science 54, 321–336.

Redfield, A. C., Ketchum, B. H. & Richards, F. A. 1963 Theinfluence of organisms on the composition of sea water. InThe Sea Vol. II (Hill, M. N., ed.). John Wiley, New York,pp. 26–77.

Shiganova, T. A. 1997 Mnemiopsis Leidyi abundance in the BlackSea and its impact on the pelagic community. In Sensitivity toChange: Black Sea, Baltic Sea and North Sea (Ozsoy, E. &Mikaelyan, A., eds). NATO ASI Series, 2—Environment 27,117–131.

Sorkina, A. I. 1974 Reference Book on the Black Sea Climate.Gydrometeoizdat, Moscow, 406 pp. (in Russian).

Sorokin, Yu. I. 1982 The Black Sea. Moscow, 216–pp.Staneva, J. V., Stanev, E. V. & Oguz, T. 1998 The impact of

atmospheric forcing and water column stratification on the yearlyplankton cycle. In Proceedings of the Symposium on the ScientificResults of the NATO-TU-Black Sea Project (Oguz, T. & Ivanov, L.,eds). NATO-ASI Series, 301–323.

Stepnowski, A., Gucu, A. C. & Bingel, F. 1993 Assessment of thepelagic fish resources in the southern Black Sea using echointegration and dual-beam processing. Archives of Acoustics, PolishAcademy of Science Journal 18, 83–104.

Tolmazin, D. 1985 Changing coastal oceanography of the BlackSea. Progress in Oceanography 15, 217–276.

Van Eeckhourt, D. & Lancelot, C. 1997 Modelling the functioningof the north-western Black Sea ecosystem from 1960 to present.

490 C. Lancelot et al.

In Sensitivity to Change: Black Sea, Baltic Sea and North Sea(Ozsoy, E. & Mikaelyan, A., eds). NATO-ASI Series 2—Environment 27, 455–468.

Vedernikov, V. I. & Demidov, A. B. 1997 The vertical distributionof the primary production and chlorophyll in the different seasonof the deep parts of the Black Sea. Oceanologya 37, 414–423.

Vinogradov, M. E. & Tumantseva, N. 1993 Some results ofinvestigations of the Black Sea biologic communities. In BlackSea Research Country Profiles (Level II). UNESCO, pp. 80–94.

Vinogradov, M. E., Shuskina, E. A., Mikaelyan, A. S. & Nezlin,N. P. 1999 Temporal (seasonal and interannual) changes of

ecosystem of the open waters of the Black Sea. In EnvironmentalDegradation of the Black Sea: Challenges and Remedies (Besiktepe,S. T., Unluata, U. & Bologa, A. S., eds). NATO Science Series,2—Environmental Security 56, 109–129.

Weisse, T., Gomoiu, M.-T., Scheffel, U. & Brodrecht, F. 2002Biomass and size composition of the comb jelly Mnemiopsis sp. inthe north-western Black Sea during spring 1997 and summer1995. Estuarine, Coastal and Shelf Science 54, 423–437.

Zaitsev, Yu. P. 1993 Impacts of eutrophication on the Black Seafauna: Studies and Reviews. General Fisheries Council for theMediterranean 64, 59–86.

Appendix

T A1. The BIOGEN model: state variables

Variable Symbol

Biological state variables:Diatoms: DA=DAF+DAS+DAR

Functional and structural metabolites DAFMonomers DASReserves DAR

Phototrophic nanoflagellates: NF=NFF+NFS+NFRFunctional and structural metabolites NFFMonomers NFSReserve products NFR

Opportunistic phytoplankton: OP=OPF+OPS+OPRFunctional and structural OPFMonomers OPSReserve products OPR

Bacteria BAC

Microzooplankton MCZ

Copepods COP

Noctiluca NOC

Aurelia aurita AUR

Mnemiopsis sp. MNE

Organic matterMonomeric: carbon, nitrogen BSC, BSN

Dissolved polymers (high biodegradability):carbon, nitrogen, phosphorus DC1, DN1, DP1

Dissolved polymers (low biodegradability):carbon, nitrogen, phosphorus DC2, DN2, DP2

Particulate organic matter (high biodegradability):carbon, nitrogen, phosphorus PC1, PN1, PP1

Particulate organic matter (low biodegradability):carbon, nitrogen, phosphorus PC2, PN2, PP2

Detrital biogenic silica BSi

Inorganic nutrientsNitrate NO3

Ammonium NH4

Phosphate PO4

Silicic acid SiO

Modelling the north-western Black Sea: the ecosystem 491

T A2. The BIOGEN model: processes

Description Symbol

Phytoplankton dynamics:Photosynthesis �i i=DA, NF, OPGrowth �i i=DA, NF, OPReserve synthesis si i=DAR, NFR, OPRReserve catabolism ci i=DAR, NFR, OPRExudation ei i=DAS, NFS, OPSRespiration respi i=DA, NF, OPAutolysis lysif i=DA, NF, OP; j=F, S, RSedimentation sed if i=DA, OP; j=F, S, RNutrient uptake uptk

PHY k=NO3, NH4, PO4, SiO; PHY=DA+NF+OP

Zooplankton dynamics:Grazing gl/q l=MCZ, COP, NOC, AUR, MNE for l=MCZ q=BAC, NF

l=COP q=DA, MCZl=NOC q=DA, NF, OPP, BAC, MCZ, PC1,2

l=AUR, MNE q=COPGrowth �l l=MCZ, COP, NOC, AUR, MNEMortality lysl l=MCZ, COP, NOC, AUR, MNEEgestion egl l=MCZ, COP, NOC, AUR, MNENutrient regeneration regk

l l=MCZ, COP, NOC, AUR, MNE k=NH4, PO4

Microbial loop dynamics:C and nutrient uptake uptkBAC k=BSC, BSN, NH4, PO4

Growth �BACMortality lysBAC

Ammonification regNH4BAC

Nitrification niDenitrification dniEctoenzymatic hydrolyse of DOM elysDi Di=DC1,2=DN1,2=DP1,2

Ectoenzymatic hydrolyse of POM elysPi Pi=PC1,2=PN1,2=PP1,2

Sedimentation of POM sedPi Pi=PC1,2=PN1,2=PP1,2

Benthic dynamics:Nutrient exchanges at the sediment–water interface Jk k=NO3, NH4, PO4, SiO

492 C. Lancelot et al.

T A3. The BIOGEN model: conservation equations

Modelling the north-western Black Sea: the ecosystem 493

494 C. Lancelot et al.

Modelling the north-western Black Sea: the ecosystem 495

T

A4.

The

BIO

GE

Nm

odel

:eq

uati

ons

and

para

met

ers.

Phy

topl

ankt

ondy

nam

ics

Pro

cess

Equ

atio

nP

aram

eter

s

Sym

bol

Des

crip

tion

DA

NF

OP

Uni

t

Pho

tosy

nthe

sis

�k m

ax(1

�ex

p�

(�I/

k max))

Fk* m

ax

Max

imal

phot

osyn

thes

isra

te0·

355

0·35

50·

355

h�

1

aP

hoto

synt

heti

ceffi

cien

cy0·

0012

0·00

120·

0012

h�

1(�

Em

�2

s�1)�

1

Res

erve

synt

hesi

ss R

�* max

Max

imal

rate

ofre

serv

esy

nthe

sis

0·3

0·3

0·4

h�

1

k sH

alf-

satu

rati

onco

nsta

nt0·

060·

060·

06di

men

sion

less

Res

erve

cata

bolis

mc R

k cR

k* cR

ate

ofre

serv

eca

tabo

lism

0·2

0·4

0·4

h�

1

Gro

wth

��* m

ax

Max

imal

grow

thra

te0·

155

0·16

50·

155

h�

1

wit

hln

ut=

k NH

alf-

satu

rati

onco

nsta

ntfo

rin

orga

nic

nitr

ogen

upta

ke1

0·2

2·2

�M

ork P

Hal

f-sa

tura

tion

cons

tant

for

PO

4up

take

0·1

0·02

0·2

�M

ork S

iH

alf-

satu

rati

onco

nsta

ntfo

rS

iup

take

1—

—�M

Res

pira

tion

resp

k mF

+��

k mM

aint

enan

ceco

nsta

nt0·

0003

0·00

050·

0005

h�

1

�***

Bio

synt

hesi

sco

st0·

4–0·

80·

4–0·

80·

4–0·

8di

men

sion

less

Exu

dati

one

k e�

k eE

xcre

tion

cons

tant

0·01

0·1

0·01

h�

1

Aut

olys

isly

s PH

Yk l

ys(

F+

S+

R)

k lys*

*A

utol

ysis

cons

tant

0·00

10·

0025

0·00

2h

�1

Sed

imen

tati

on(D

A,

OP

)se

d PH

Yk s

ed(F

+S

+R

)k s

ed**

Sed

imen

tati

onra

teco

nsta

nt0·

008

00·

008

mh

�1

Inor

gani

cN

upta

keup

tN�C

NC

NP

hyto

plan

kton

stoi

chio

met

ryC

/N4

44

�mol

CF

/�m

olN

Am

mon

ium

pref

eren

ce(r

pi)

� rp

iR

elat

ive

pref

eren

cein

dex

0·3

0·4

0·75

dim

ensi

onle

ss

PO

4up

take

uptP

�/C

PC

PP

hyto

plan

kton

stoi

chio

met

ryC

/P80

8080

�mol

CF

/�m

olP

Si(

OH

) 4up

take

uptS

i�/

CS

iC

Si

Phy

topl

ankt

onst

oich

iom

etry

C/ S

i2·

8�m

olC

F/�

mol

Si

For

cin

gte

mp

erat

ure

dep

end

ence

:p

aram

eter

sw

ith

*p(

T)

p*ex

p(�

(T�

Top

t)2/d

ti2)

Top

tO

ptim

alte

mpe

ratu

re22

2224

�Cdt

iT

empe

ratu

rein

terv

al12

1212

�C

Lim

itin

gn

utr

ien

td

epen

den

ce:

par

amet

ers

wit

h**

p(ln

ut)

p**(

1+

5)(1

�ln

ut)

Inor

gan

icn

itro

gen

sou

rce—

dep

end

ence

:�*

**�(

N)

0·4r

pi+

0·8(

1�

rpi)

PH

Y=

F+

S+

R:

PH

Y=

DA

,N

F,

OP

.

496 C. Lancelot et al.

T

A5.

The

BIO

GE

Nm

odel

:eq

uati

ons

and

para

met

ers.

Zoo

plan

kton

dyna

mic

s

Pro

cess

Equ

atio

n

Par

amet

ers

Sym

bol

Des

crip

tion

Val

ueU

nit

Mic

rozo

opla

nkt

on:

2p

reys

:1

=B

AC

;2

=N

FG

razi

ngg M

CZ

gMC

Zm

ax�[

Pre

y 1,2

/(kg

ap

p1,2

+P

rey 1

,2)]

gMC

Zm

ax*

Max

imal

graz

ing

rate

0·10

h�

1

wit

h:kg

ap

p1=

kgB

AC

(1+

NF

)/k g

NF

kgB

AC

Hal

f-sa

tura

tion

cst

onB

AC

15m

gC

m�

3

kgap

p2=

kgN

F(1

+B

AC

)/k g

BA

Ckg

NF

Hal

f-sa

tura

tion

cst

onN

F10

mg

Cm

�3

Gro

wth

� MC

ZY

MC

Zg M

CZ

YM

CZ

Gro

wth

effici

ency

0·3

dim

ensi

onle

ssA

utol

ysis

lys M

CZ

kdM

CZ

MC

Zkd

MC

ZM

orta

lity

cons

tant

0·00

3h

�1

Cop

epod

sC

OP

:2

pre

ys:

1=

DA

;2

=M

CZ

Gra

zing

g CO

PgC

OP

max�[

Pre

y 1,2

/(kg

ap

p1,2

+P

rey 1

,2)]

gCO

Pm

ax*

Max

imal

graz

ing

rate

0·12

h�

1

wit

h:kg

ap

p1=

kgD

A(1

+M

CZ

)/k g

MC

Zkg

DA

Hal

f-sa

tura

tion

cst

onD

A50

mg

Cm

�3

kgap

p2=

kgM

CZ(1

+D

A)/

k gD

Akg

MC

ZH

alf-

satu

rati

oncs

ton

MC

Z25

mg

Cm

�3

Gro

wth

� CO

PY

CO

Pg C

OP

YM

ZO

Gro

wth

effici

ency

0·33

dim

ensi

onle

ssE

gest

ion

egC

OP

EZ

g CO

PE

ZE

gest

ion

cst

0·25

dim

ensi

onle

ssM

orta

lity

lys C

OP

kdC

OP

CO

Pkd

CO

P*

Cst

for

tota

lm

orta

lity

0·00

1h

�1

Noc

tilu

caG

razi

ngg N

OC

kgN

OC

*In

gest

ion

coeffi

cien

t5�

10�

5h

�1

mg

C�

1m

�3

TH

NO

CF

eedi

ngth

resh

old

10m

gC

�1

m�

3

and

i=D

A,

NF

,O

PP

,B

AC

,M

CZ

,P

C1,2

Gro

wth

� NO

CY

NO

Cgr

azN

OC

YN

OC

Gro

wth

effici

ency

0·15

dim

ensi

onle

ssE

gest

ion

egN

OC

Ez

g NO

CE

zE

gest

ion

cst

0·25

dim

ensi

onle

ssM

orta

lity

lys N

OC

kdN

OC

NO

Ckd

NO

C*

Mor

talit

ycs

t0·

007

h�

1

Au

reli

aG

razi

ngg A

UR

kgA

UR

(CO

P�

TH

AU

R)

kgA

UR

*In

gest

ion

coeffi

cien

t4�

10�

4h

�1

mg

C�

1m

�3

TH

AU

RF

eedi

ngth

resh

old

5m

gC

�1

m�

3

Gro

wth

� AU

RY

AU

Rg A

UR

YA

UR

Gro

wth

effici

ency

0·15

dim

ensi

onle

ssE

gest

ion

Eg A

UR

Ez

g AU

RE

zE

gest

ion

cst

0·25

dim

ensi

onle

ssM

orta

lity

lys A

UR

kdA

UR

AU

Rkd

AU

R*

Mor

talit

ycs

t4�

10�

4h

�1

Mn

emio

psi

sG

razi

ngg M

NE

kgM

NE(C

OP

�T

HM

NE)

kgM

NE*

Inge

stio

nco

effici

ent

6�

10�

4h

�1

mg

C�

1m

�3

TH

MN

EF

eedi

ngth

resh

old

10m

gC

�1

m�

3

Gro

wth

� MN

EY

MN

Eg M

NE

YM

NE

Gro

wth

effici

ency

0·2

dim

ensi

onle

ssE

gest

ion

egM

NE

Ez

g MN

EE

zE

gest

ion

cst

0·25

dim

ensi

onle

ssM

orta

lity

lys M

NE

kdM

NE

MN

Ekd

MN

E*

Mor

talit

ycs

t6�

10�

4h

�1

Modelling the north-western Black Sea: the ecosystem 497

T

A5.

Con

tinue

d

Par

amet

erE

quat

ion

Sym

bol

Sig

nific

atio

nM

CZ

CO

PN

OC

AU

RM

NE

Tem

per

atu

reco

ntr

ol:

par

amet

ers

wit

h*

p(T

)p*

exp�

(T�

Top

t)2/

dti2

)T

opt

Opt

imal

tem

pera

ture

23�C

23�C

23�C

23�C

23�C

dti

Tem

pera

ture

inte

rval

12�C

8�C

8�C

12�C

12�C

Des

crip

tion

Val

ueU

nits

Zoo

pla

nkt

onst

oich

iom

etry

C/N

rati

o63

�g/�

MN

/Pra

tio

16�M

/�M

498 C. Lancelot et al.

T

A6.

The

BIO

GE

Nm

odel

:eq

uati

ons

and

para

met

ers.

Mic

robi

allo

opdy

nam

ics

Pro

cess

Equ

atio

nP

aram

eter

s

Sym

bol

Sig

nific

atio

nV

alue

Uni

t

PC

1,2

prod

ucti

onby

lysi

san

deg

esti

on� p

1,2

(lys

PH

Y+

lys B

AC

+ly

s MC

Z+

lys C

OP

+ly

s NO

C+

lys A

UR

+ly

s MN

E)

+� p

1,2

(eg C

OP

+eg

NO

C+

egA

UR

+eg

MN

E)

� p1

� p1

PC

1fr

acti

onin

lysi

spr

oduc

tsP

C1

frac

tion

ineg

esti

on0·

10·

3di

men

sion

less

� p2

PC

2fr

acti

onin

lysi

spr

oduc

ts0·

4di

men

sion

less

� p2

PC

2fr

acti

onin

eges

tion

0·4

PC

1,2

ecto

enzy

mat

ichy

drol

ysis

elys

P1,2

k 1,2

bPC

1,2

k 1b*

k 2b

PC

1hy

drol

ysis

cst

PC

2hy

drol

ysis

cst

0·00

50·

0002

5h

�1

h�

1

PC

1,2

sedi

men

tati

onse

d P1,2

(vsm

/dep

th)

vsm

PC

1,2

sink

ing

rate

0·01

mh

�1

DC

1,2

prod

ucti

onby

lysi

san

deg

esti

on� d

1,2

(lys

PH

Y+

lys B

AC

+ly

s MC

Z+

lys C

OP

+ly

s NO

C+

lys A

UR

+ly

s MN

E)

+� d

1,2

(eg C

OP

+eg

NO

C+

egA

UR

+eg

MN

E)

� d1

DC

1fr

acti

onin

lysi

spr

oduc

ts0·

3di

men

sion

less

� d1

DC

1fr

acti

onin

eges

tion

0·1

dim

ensi

onle

ss� d

2D

C2

frac

tion

inly

sis

prod

ucts

0·2

dim

ensi

onle

ss� d

2D

C2

frac

tion

ineg

esti

on0·

2di

men

sion

less

Ect

oenz

ymat

ichy

drol

ysis

ofD

C1,2

elys

d1,2

e 1m

ax*

Max

imum

rate

ofD

C1

hydr

olys

is0·

75h

�1

e 2m

ax*

Max

imum

rate

ofD

C2

hydr

olys

is0·

25h

�1

kh1

Hal

fsa

tura

tion

cst

for

DC

1hy

drol

ysis

250

mg

Cm

�3

kh2

Hal

fsa

tura

tion

cst

for

DC

2hy

drol

ysis

2500

mg

Cm

�3

Mon

omer

sup

take

upt B

AC

b max*

Max

imum

rate

ofB

Sup

take

0·4

h�

1

BS

=B

SC

,B

SN

k BS

Hal

fsa

tura

tion

cst

for

BS

upta

ke25

mg

Cm

�3

Bac

teri

algr

owth

� BA

CY

BA

Cup

t BA

CY

BA

CG

row

theffi

cien

cy0·

2di

men

sion

less

PO

4up

take

uptP

O4

BA

CC

PB

AC

Bac

teri

alC

/Pra

tio

106

�M�M

�1

Bac

teri

ally

sis

lys B

AC

kdB

AC

BA

Ckd

BA

C*

Bac

teri

ally

sis

rate

cst

0·01

h�

1

Am

mon

ifica

tion

regN

H4

BA

C[(

1�

YB

AC

)/Y

BA

C]�

BA

C/C

NB

AC

CN

BA

CB

acte

rial

C/N

rati

o4

�M�M

�1

Nit

rific

atio

nni

nim

ax

Max

imum

nitr

ifica

tion

rate

0·03

mM

m�

3h

k ni

Hal

f-sa

tura

tion

cst

5�M

Tem

per

atu

refo

rcin

g:p

aram

eter

sw

ith

p(T

)p*

exp(

�(T

�T

opt)

2/d

ti2)

Top

tO

ptim

alte

mpe

ratu

re30

�Cdt

iT

empe

ratu

rein

terv

al18

�C

Modelling the north-western Black Sea: the ecosystem 499

T A7. The BIOGEN model: equations and parameters. Benthic diagenesis

Process EquationParameters

Symbol Signification Value Unit

Diffusion(interstitial phase)

Fick law Di Apparent diffusioncoefficient

1·8�10�5 cm2 s�1

Mixing (solid phase) Fick law Ds Mixing coefficient 1·8�10�6 cm2 s�1

Organic N mineralization k1,2b PN1,2 k1b* PN1 hydrolysis cst 0·005 h�1

k2b PN2 hydrolysis cst 0·00025 h�1

Organic P mineralization k1,2p PP1,2 k1p* PP1 hydrolysis cst 0·05 h�1

k2p PP2 hydrolysis cst 0·0025 h�1

Benthic nitrification(oxic layer)

knib1,2 NH4 knib1,2 First order nitrification cst 1 h�1

NH4 adsorption/desorption 1st order equilibrium Kam NH4 adsorption cst 6 dimensionless

PO4 adsorption/desorption(in benthos)

1st order equilibrium KpaKpe

PO4 adsorption cst (oxic)PO4 adsorption cst (anoxic)

351·7

dimensionlessdimensionless

BSi dissolution kBSiBSi kBSi Silica dissolution cst 0·000075 h�1

Temperature control: parameters with*p(T) p*exp(�(T�Topt)2/dti2) Topt Optimal temperature 30 �C

dti Temperature interval 18 �C