Industry

A numerical simulation of the role of zooplankton in C, N and P cycling in Lake Kinneret, Israel

Description
A numerical simulation of the role of zooplankton in C, N and P cycling in Lake Kinneret, Israel
Categories
Published
of 25
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  Ecological Modelling 193 (2006) 412–436 A numerical simulation of the role of zooplankton in C, N andP cycling in Lake Kinneret, Israel Louise C. Bruce a , ∗ , David Hamilton b , J¨org Imberger a , Gideon Gal c ,Moshe Gophen d , Tamar Zohary c , K. David Hambright e a Centre for Water Research, University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia b  Department of Biological Sciences, The University of Waikato, Private Bag 3105, Hamilton, New Zealand  c Y. Allon Kinneret Limnological Laboratory, IOLR, P.O. Box 447, Migdal 14950, Israel d  Limnology and Ecology of Wetlands and Freshwater, MIGAL Galilee Technology Centre, Southern Industrial Zone, P.O. Box 831,Kiryat-Shmona 11016, Israel e University of Oklahoma Biological Station and Department of Zoology, HC-71, Box 205, Kingston, OK 73439, USA Received 6 February 2005; received in revised form 30 August 2005; accepted 7 September 2005Available online 10 November 2005 Abstract We quantify the role of zooplankton in nutrient cycles in Lake Kinneret, Israel, using field data and a numerical model.A coupled ecological and hydrodynamic model (Dynamic Reservoir Model (DYRESM)–Computational Aquatic EcosystemDynamicsModel(CAEDYM))wasvalidatedwithanextensivefielddatasettosimulatetheseasonaldynamicsofnutrients,threephytoplanktongroupsandthreezooplanktongroups.Parameterizationofthemodelwasconductedusingfield,experimentalandliterature studies. Sensitivity of simulated output was tested over the full parameter space and established that the most sensitiveparameters were related to zooplankton grazing rates, temperature responses and food limitation. The simulated results predictthat, on average, 51% of the carbon from phytoplankton photosynthesis is consumed by zooplankton. Excretion of dissolvednutrients by zooplankton accounts for 3–46 and 5–58% of phytoplankton uptake of phosphorus and nitrogen, respectively.Comparison of nutrient fluxes attributable to zooplankton with nutrient loads from inflows and release from bottom sedimentsshows that the relative contribution by zooplankton to inorganic nutrients in the photic zone varies seasonally in response to theannual hydrodynamic cycle of stratification and mixing. As a percent of total dissolved organic sources relative contributions byzooplankton excretion are highest (62%) during periods of stratification and when inflow nutrient loads are low, and lowest (2%)during the breakdown of stratification and when inflow loads are high. The results illustrate the potential of a lake ecosystemmodel to extract useful process information to complement field data collection and address questions related to the role of zooplankton in nutrient cycles.© 2005 Elsevier B.V. All rights reserved. Keywords:  Zooplankton; Nutrient cycling; Numerical model; Lake Kinneret ∗ Corresponding author.0304-3800/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.ecolmodel.2005.09.008   L.C. Bruce et al. / Ecological Modelling 193 (2006) 412–436   413 1. Introduction Zooplankton play an important role in lake dynam-ics, as grazers that control algal and bacterial popula-tions, as a food source for higher trophic levels and inthe excretion of dissolved nutrients. Thus, understand-ing their role in the distribution and flux of nutrientsin aquatic systems is critical for effective lake man-agement. Zooplankton grazing on phytoplankton cantransfer more than 50% of carbon fixed by primaryproduction to higher trophic levels (Hart et al., 2000;Laws et al., 1988; Scavia, 1980). Zooplankton excre-tionstronglyinfluencestrophicdynamicsinfreshwaterecosystems by contributing inorganic N and P for pri-maryandbacterialproduction(Gilbert,1998;Lehman,1980; Sterner, 1986; Vanni, 2002; Wen and Peters,1994). Estimates of the fraction of N and P regener-ated by zooplankton and then utilised by phytoplank-ton range from 14 to 50% (Hudson and Taylor, 1996;Hudson et al., 1999; Urabe et al., 1995). The fac-tors controlling this fraction include temperature, zoo-plankton and phytoplankton biomass and species com-position, internal nutrient ratios and mixing regimes.Because these factors interact dynamically, it has beendifficult to quantify the role of zooplankton in nutrientcycling.Models have previously been used to evaluatedifferent aspects of zooplankton dynamics in lakes(Carpenter and Kitchell, 1993; Chen et al., 2002; Coleet al., 2002; H˚akanson and Boulion, 2003; Hongpingand Jianyi, 2002; Ji et al., 2005; Krivtsov et al., 2001;LunteandLeucke,1990;Rukhovetsetal.,2003;Scaviaet al., 1988), reservoirs (Mehner, 2000; Osidele andBeck, 2004; Romero et al., 2004), fjords (Ross etal., 1994), estuaries (Griffin et al., 2000; Robson andHamilton, 2004), lagoons (Lin et al., 1999) and in the marineenvironment(CarlottiandG¨unther,1996;Lawsetal.,1988).Themodelsrangeincomplexityfromsim-ple mass balances to highly parameterized simulationtools that include hydrodynamic processes (Carlottiand G¨unther, 1996; Robson and Hamilton, 2004; Rosset al., 1994). Some models simulate nutrient fluxesbetween trophic levels and provide valuable insightsinto the relative importance of zooplankton in nutri-ent cycles (Urabe et al., 1995). These models may also yield more detailed temporal and spatial informationon nutrient fluxes between different trophic levels ina lake than is often possible with field or laboratorydata. Specifically, models can be used to predict howthe fluxes change in response to environmental factorsor lake management strategies.LakeKinnerethasbeenstudiedintenselybothinsituand experimentally (Hambright et al., 1994; Serruya,1978; Yacobi et al., 1993; Zohary et al., 1994). Thelake supplies approximately 30% of Israel’s drinkingwater,afactthathasmotivatedanextensivewaterqual-ity and ecological monitoring program as well as amajor subsidized fishing effort to rid the lake of plank-tivorous sardines in the hope of fostering the existingzooplanktonpopulation(BlumenshineandHambright,2003). The monitoring program has included routinesamplingofvariouslevelsofthelakefoodweb,includ-ing zooplankton and phytoplankton, water column andtributary chemical and physical parameters, and mete-orological data.In this study, we have applied a coupled ecologi-cal and hydrodynamic model (DYRESM–CAEDYM)to the Lake Kinneret data set to simulate the seasonaldynamics of nutrients, three phytoplankton groups andthree zooplankton groups. The true uniqueness of thisstudy lies in the mechanistic approach to model struc-ture, low vertical resolution of the physical driverrequiring only external forcing, sub-daily time steps,multi-nutrient focus, division of both phytoplanktonand zooplankton into functional groups and the appli-cation against an extensive data set. Although simplemassbalanceboxmodelssuchasHartetal.(2000)canbe useful in determining lake wide patterns, the pro-cessesrepresentedinCAEDYMallowedustofocusonspecific mechanisms responsible for the determinationof important lake phenomenon. The fully mechanis-tic structure of DYRESM means that no calibration isrequired (Yeates and Imberger, 2004); the other advan- tage of using DYRESM is the vertical resolution sothat processes such as sediment release and particu-late settling are fully represented rather than forced asinputs such as in Osidele and Beck (2004). In addi- tion, a sub-daily time step in both the ecological modeland hydrodynamic driver enables resolution of pho-tosynthetic processes so that seasonal trends can bemoreaccuratelyresolvedratherthantheuseofdailyorweekly time steps such as those used by H˚akanson andBoulion (2003). Many ecosystem models are devel-oped based on a single limiting nutrient (Ji et al.,2005). CAEDYM on the other hand explicitly modelstheinorganic,organic,phytoplanktonandzooplankton  414  L.C. Bruce et al. / Ecological Modelling 193 (2006) 412–436  components of carbon, nitrogen and phosphorus. ThisapproachiscrucialinlakesuchasLakeKinneretwherelimitationcanswitchbetweennitrogenandphosphorusdepending on the season. Although some ecologicalmodels divide phytoplankton into functional groups,most treat zooplankton as a single entity (HongpingandJianyi,2002;Krivtsovetal.,2001;Rukhovetsetal.,2003).Byincludingthreephytoplanktonandthreezoo-plankton groups in the model, CAEDYM can be usedto determine how the role of zooplankton changes as afunctionofchangesinseasonaldominancebetweenthemain zooplankton groups. Finally, aquatic ecosystemmodels are often applied to lakes where data are scarceso that rigorous calibration and validation of modelsis difficult (H˚akanson and Boulion, 2003; Osidele andBeck, 2004). The main advantage of the application of the model to such an extensive data set enabled us totest both the choice of parameters and model processesover a 4-year period. Although the application of someaquatic ecosystem models includes one or more of theattributes of DYRESM–CAEDYM described above,the contribution of our study is the inclusion of alltogether with a rigorous calibration and validation.Theobjectiveofthepresentstudywastoinvestigatethe role of zooplankton in the cycling of C, N and Pbetween trophic levels in Lake Kinneret. After validat-ing the coupled DYRESM–CAEDYM model against acomprehensive data set, we used the model to quanti-tatively examine how zooplankton biomass, secondaryproduction and fluxes between trophic levels may beaffected by seasonal changes in hydrodynamic mix-ing regimes. In addition, a sensitivity analysis wasconducted to evaluate the relative importance of keyparameters on model simulation results. The effect of zooplanktongrazingandexcretionontheavailabilityof nutrients was estimated with the calibrated numericalmodel and compared to external sources from inflowsand internal sources from sediment release. 2. Study site Lake Kinneret (32 ◦ 48  N, 35 ◦ 37  E) is a warm-monomictic freshwater lake with maximum depth43m,meandepth25mandsurfaceareaapproximately170km 2 (Fig.1).Thelakeisverticallymixedinwinter (December–February)andthermalstratificationsetsupinspring,persistingfor7–8months.Temperatureinthe Fig.1. MapofLakeKinneretshowinglocationofthemainsamplingstation A and Jordan River inflow. Depth contours are in meters. surface layer typically ranges from 15–17 ◦ C in winterto 26–30 ◦ C in summer (Hambright et al., 1994).The phytoplankton assemblage of Lake Kinneret isgenerally dominated in winter–spring by the dinoflag-ellate  Peridinium gatunense , in summer–autumn by adiverse assemblage of nanoplankton, mostly chloro-phytes, and since the mid-1990s also by filamentouscyanobacteria. A third component to the phytoplank-ton assemblage is the diatom  Aulacoseira granulata that in some years forms a deep-water bloom inJanuary–February(Zohary,2004).Macro-zooplankton biomass in Lake Kinneret is dominated for most of the year by herbivorous cladocerans (Gophen, 1984).The predatory zooplankton assemblage includes adultcopepods, and large rotifers. The micro-zooplanktoncommunity includes copepod nauplii, small herbivo-rousandbactivorousrotifers,ciliatesandheterotrophicflagellates (Hadas and Berman, 1998). A study of    L.C. Bruce et al. / Ecological Modelling 193 (2006) 412–436   415 stable carbon isotopes showed seasonal dietaryvariations of macro-zooplankton, with nanoplanktonas the predominant food source (Zohary et al., 1994).The major food sources for the micro-zooplanktonare bacteria and picophytoplankton for the smallerheterotrophic flagellates, and bacteria, heterotrophicflagellates and nanophytoplankton for the ciliates(Hadas and Berman, 1998; Madoni et al., 1990). 3. Methods 3.1. Model description The model used in this study is a modified versionof the Computational Aquatic Ecosystem DynamicsModel (CAEDYM) coupled to the Dynamic Reser-voir Model (DYRESM). In DYRESM, the lake isrepresented as a series of homogeneous horizontalLagrangian layers of variable thickness; as inflowsand outflows enter or leave the lake, the affectedlayers expand or contract, respectively, and thoseabove move up or down to accommodate any volumechange. Mass, including that of the ecological statevariables, is adjusted conservatively each time layersmerge or are affected by inflows and outflows. Themain processes modeled in DYRESM are surface heat,mass and momentum transfers, mixed layer dynamics,hypolimnetic mixing, benthic boundary layer mixing,inflows and outflows. Local meteorological dataare used to determine penetrative heating due toshort-wave radiation and surface heat fluxes due toevaporation, sensible heat, long-wave radiation andwind stress. The surface wind field introduces bothmomentum and turbulent kinetic energy to the surfacelayercontributingtoverticalmixing.Inadditiontosur-face layer mixing, DYRESM includes algorithms thataccount for internal mixing (encompassing the effectsof shear mixing energized by internal waves) andbenthic boundary layer (BBL) mixing (determined bythe turbulent kinetic energy budget and parameterizedby Lake Number and the Burger number). In this way,mass transfer is enabled from hypolimnetic layers tothe thermocline region via the BBL. A schematic flowchart of the operations performed in DYRESM is pre-sentedinFig.2.Adetaileddescriptionoftheprocesses included in DYRESM is given by Yeates and Imberger(2004). Fig. 2. Schematic representation of: (A) the physical processesincluded in the physical model DYRESM (BBL: benthic boundarylayer; IC: internal cells; BC: benthic boundary layer cells) and (B)the carbon fluxes represented in the ecological model, CAEDYM. The ecological model CAEDYM was set up inthe form of a nutrients–phytoplankton–zooplankton(‘N–P–Z’) model, but with resolution to the levelof individual species or groups of species (Griffinet al., 2000). In the present study, the model wasused to simulate phosphorus and nitrogen in bothparticulate organic and dissolved inorganic forms(POP and PO 4 , PON, NO 3  and NH 4 ), dissolvedoxygen (DO), particulate organic carbon (POC),dissolved organic carbon (DOC), three phytoplanktongroups and three zooplankton groups. The phy-toplankton community was simulated using threegroups in the model: “dinoflagellates”, representing  P.gatunense , “diatoms”, representing  A. granulata , and  416  L.C. Bruce et al. / Ecological Modelling 193 (2006) 412–436  Table 1Equations used to describe the processes included in the ecological model CAEDYM  P  j  /   t  =[P max,j  f  1 ( T  )min(  f  (  I  ),  f  (P),  f  (N)) − (  R  j )  f  2 ( T  ) − Pred  j ]P  j ± S   j  =production − (respiration+excretion+mortality) − predation ± settling  Z i  /   t  =[ G i  A i  f  (Z) i  f  1 ( T  )(1 −  f  ex −  f  eg ) − (  R i  +  M  i )  f  2 ( T  ) − Pred i ]Z i  =(assimilation − excretion − egestion) − (respiration+mortality) − predation  POC/   t  =  [ G i  f  (Z) i  f  1 ( T  ) i ((1 −  A i )+  A i  f  eg )+  M  i  f  2 ( T  ) i ]Z i  +  [  R  j (1 −  f  res )(1 −  f  DOM )  f  2 ( T  )]P  j − Pred POM POC −  R POC  f  (DO)  f  1 ( T  )POC ± S  POM  =(zooplankton messy feeding+zooplankton egestion+zooplankton mortality)+phytoplankton mortality − zooplankton predation − POCdecomposition ± settling  DOC/   t  =  [  R  j (1 −  f  res )  f  DOM  f  2 ( T  )]P  j  +  R POC  f  (DO)  f  2 ( T  )POC −  R DOC  f  (DO)  f  2 ( T  )DOC=phytoplankton excretion+POC decomposition − DOCmineralisation  POP/   t  =  [ G i  f  (Z) i  f  1 ( T  ) i ((1 −  A i )+  A i  f  eg )+  M  i  f  2 ( T  ) i ]IP Zi Z i  +  [  R  j (1 −  f  res )(1 −  f  DOM )  f  2 ( T  )]IP  j − Pred POM POP −  R POP  f  (DO)  f  1 ( T  )POP ± S  POM  =(unassimilated zooplankton food+zooplankton egestion+zooplankton mortality)+phytoplankton mortality − zooplankton predation − POPmineralisation ± settling  PO 4 /t   = R POP f  (DO) f  2 ( T  )POP −  [UP max ,  j f  1 ( T  )  j f  (IP)  j f  (P)  j ]P  j + S  dPO 4 f  (DO) f  2 ( T  )LA / LV = POP mineralisation − phytoplankton uptake + PO 4  sediment flux  PON/   t  =  [ G i  f  (Z) i  f  1 ( T  ) i ((1 −  A i )+  A i  f  eg )+  M  i  f  2 ( T  ) i ]IN zi Z i  +  [  R  j (1 −  f  res )(1 −  f  DOM )  f  2 ( T  )]IN  j − Pred POM PON −  R PON  f  (DO)  f  1 ( T  )PON ± S  POM  =(unassimilated zooplankton food+zooplankton egestion+zooplankton mortality)+phytoplankton mortality − zooplankton predation − PONmineralisation ± settling  NH 4 /t   = R PON f  (DO) f  1 ( T  )PON −  [UN max ,  j P N f  1 ( T  )  j f  (IN)  j f  (N)  j ]P  j − R NO f  (DO) f  2 ( T  )NH 4 + S  dNH 4 f  (DO) f  2 ( T  )LA / LV = PON mineralisation − phytoplankton uptake − nitrification + NH 4  sediment flux  NO 3 /t   = R NO f  (DO) f  2 ( T  )NH 4 − R N 2 f  (DO) f  2 ( T  )NO 3 −  [UN max ,  j (1 − P N ) f  1 ( T  )  j f  (IN)  j f  (N)  j ]P  j  = nitrification − denitrification − phytoplankton uptake  DO /t   = k O 2 (DOatm − DO) +  [P max ,  j f  1 ( T  )  j  min( f  (I) ,f  (P) ,f  (N)) − R  j f  2 ( T  )  j ]P  j Y O 2 :C  [ R i f  2 ( T  ) i ]Z i Y O 2 :C − R DOC f  (DO) f  1 ( T  )DOCY O 2 :C − R NO f  (DO) f  2 ( T  )NH 4 − S  dDO f  (DO) f  2 ( T  )LA / LV = atmospheric flux + (phytoplankton oxygen production − phytoplankton respiratory consumption) − zooplankton respiratory consumption − utilisation of oxygen in mineralisation of DOM − utilisation of oxygen in nitrification − sediment oxygen demandTemperature functions  f  1 ( T  )= θ  T  − 20 − θ  k  ( T  − a ) + b where  k  ,  a  and  b  are constants solved numerically to satisfy the following conditions:  f  1 ( T  )=1, at  T  = T  sta ∂  f  1 ( T  )/  ∂ T  =0, at  T  = T  opt  f  1 ( T  )=0, at  T  = T  max  f  2 ( T  )= θ  T  − 20 Limitation equations  f  (Z) i =(  P  j  +  Z k   +POC)/( K  i  +  P  j  +  Z k   +POC)  f  (  I  )  j  =  I   /   I  s  exp(1 −  I   /   I  s )  f  (IP)  j  =[IP max  /(IP max − IP min )][1 − IP min  /IP]  f  (IN)  j  =[IN max  /(IN max − IN min )][1 − IN min  /IN]  f  (DO)=DO/( K  DO  +DO)  f  (P)=PO 4  /( K  P  +PO 4 )  f  (N)=(NH 4  +NO 3 )/( K  N  +NH 4  +NO 3 )P N  =(NH 4 NO 3 )/[(NH 4  + K  N )(NO 3  + K  N )]+(NH 4 K  N )/[(NH 4  + K  N )(NO 3  + K  N )]Settling S   j  =(ws/    z )P  j S  POM  =( g ( ρ POM − ρ w )(  D POM ) 2  /18 µ )/    z )POMPredationPred i  =  ( G k   f  (Z) k   f  1 ( T  ) k  Z k  PzZOO k,i )Pred  j  =  ( G i  f  (Z) i  f  1 ( T  ) i Z i PzPHY i,j )Pred POM  =  ( G i  f  (Z) i  f  1 ( T  ) i Z i PzPOC)  Abbreviations : Z, zooplankton; P, phytoplankton; POC, particulate organic carbon; DOC, dissolved organic carbon; POP, particulate organicphosphorus; PO 4 , phosphate; PON, particulate organic nitrogen; NH 4 , ammonium; NO 3 , nitrate; POM, particulate organic matter (C, N orP); IP zi , zooplankton internal phosphorus; IN zi , zooplankton internal nitrogen; IP  j , phytoplankton internal phosphorus; IN  j , phytoplanktoninternal phosphorus; DO, dissolved oxygen; DOatm, concentration of oxygen in the atmosphere; LA, layer area; LV, layer volume;    z , layerthickness; ρ w , density of water; µ , viscosity of water; k O 2 , oxygen transfer coefficient.  Subscripts : i, zooplankton group; j, phytoplankton group;k, zooplankton predator group.
Search
Similar documents
View more...
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks