JPR Advance Access originally published online on November 23, 2007
Journal of Plankton Research 2008 30(1):1-5; doi:10.1093/plankt/fbm094
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
HORIZONS |
How best to include the effects of climate-driven forcing on prey fields in larval fish individual-based models
1 Institute of Hydrobiology and Fisheries Science, Center for Marine and Climate Research, University of Hamburg, Olbersweg 24, 22767 Hamburg, Germany 2 Geophysical Institute, Bjerknes Center for Climate Research, University of Bergen, AllÉgaten 70, 5007 Bergen, Norway
* CORRESPONDING AUTHOR: ute.daewel{at}uni-hamburg.de
Received on ; accepted on November 19, 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT IntroductionPrey fields from observations:...Prey fields from stage-based...A new approach for...ConclusionsFundingReferences |
|---|
If we intend to examine the indirect effects of climate variability on the vital rates of key marine species, climate-induced changes in the spatial-temporal dynamics of prey must be resolved. Recently, structured population simulations have been coupled to ecosystem (nutrient-phytoplankton-zooplankton-detritus, NPZD) models to derive prey fields. Model-derived prey fields offer advantages (e.g. increased spatial-temporal coverage, direct links to climate forcing). In contrast, employing structured population simulations (e.g. stage-based copepod models) has several disadvantages, including the lack of realistic utilization of phytoplankton production, the absence of boundary condition data and a vastly increased coupled model complexity. To avoid the pitfalls limiting the utility of structured population models, we argue for a more simple approach for obtaining a size-structured prey field using NPZD model estimates of bulk zooplankton carbon and in situ zooplankton abundance-at-size data. The approach was developed to obtain prey fields for a larval fish individual-based model (IBM), but the method may offer wide applicability. Moreover, our approach greatly simplifies the coupling of NPZD models and larval fish IBMs and is an example of the reduction in model complexity that will be critical to the development of end-to-end ecosystem models that use, for example, a rhomboid approach to examine trophodynamic climate impacts at basin scales.
|
Introduction |
|---|
|
TOPABSTRACTIntroductionPrey fields from observations:...Prey fields from stage-based...A new approach for...ConclusionsFundingReferences |
|---|
Understanding and predicting the effects of changes in physical forcing due to climate change on the population dynamics of key species are critical for the development of an ecosystem approach to the management of marine resources. One of the tools for understanding how variability in climatic forcing influences the recruitment of marine fish population is the application of individual-based models (IBMs) (e.g. Werner et al., 2001
). In their simplest form, IBMs allow the exploration of the effects of climate-driven changes in transport dynamics addressing the so-called advection recruitment hypothesis proposed by Parrish (Parrish et al., 1983). IBMs have also been employed to examine the potential role that climate-driven changes in ecosystem production may have on the survival and growth of early life stages of key fish species (e.g. Werner et al., 2001; Hinrichsen et al., 2002; Peck and Daewel, 2007) addressing Cushing’s (Cushing, 1975) match-mismatch hypothesis. |
Prey fields from observations: limitations and aliasing errors |
|---|
|
TOPABSTRACTIntroductionPrey fields from observations:...Prey fields from stage-based...A new approach for...ConclusionsFundingReferences |
|---|
To study indirect effects of climate on the vital rates of key species, the spatial-temporal dynamics of prey resources resulting from changes in physical forcing need to be resolved. For larval fish IBMs, the use of observed prey distributions has been suggested as an optimal approach but this is untenable for a variety of reasons. First, due to resource limitations, in situ prey fields employed in larval fish IBMs are generally derived from measurements of the species- and stage-specific abundance of copepods (e.g. Lough et al., 2005) derived from only a limited number of cruises, providing information that are restricted in time and space. Extrapolation of these sparse measurements to provide prey fields for high resolution 3D models is likely to create aliasing errors of unknown magnitude and direction. Second, from a practical standpoint, measurements made on zooplankton captured in nets often lack replication (and variance estimates) and, when replicate measurements are made at the same station, they often agree poorly. Moreover, in situ measurements of prey fields cannot be used to project future situations since there is no recognition of the impact of climate-driven bottom-up processes affecting prey dynamics. Hence, in situ measurements of prey are limited in their ability to address the potential effects of climate change necessary for the development of ecosystem-based management strategies and model strategies have to be developed including prey field simulations.
Estimates of zooplankton abundance and production gained by using nutrient-phytoplankton-zooplankton-detritus (NPZD) models provide highly spatially- and temporally-resolved larval fish prey fields. More importantly, in contrast to in situ measurements, model parameterizations include interactions between environmental forcing variables (e.g. temperature, light and nutrients) and the responses of phyto- and zooplankton which enables one to investigate the impacts of climate change on ecosystem components. Nonetheless, this NPZD-based approach contains some caveats. Larval fish do not eat bulk zooplankton carbon—the standard output from an NPZD model (e.g. mg C m–2 d–1) and NPZD-derived estimates of zooplankton must be transformed and partitioned into relevant, size-specific prey categories.
|
Prey fields from stage-based copepod models: some pitfalls |
|---|
|
TOPABSTRACTIntroductionPrey fields from observations:...Prey fields from stage-based...A new approach for...ConclusionsFundingReferences |
|---|
Since they provide more complex, species-specific information compared to NPZD models, stage-based copepod models (Carlotti and Radach, 1996
; Carlotti and Wolff, 1998; Fennel and Neumann, 2003; Neumann and Fennel, 2006; Moll and Stegert, 2007) have been suggested to adequately derive model-based prey fields. However, coupling stage-resolved copepod models to ecosystem models has a number of potential weak points. First, in many ecosystems, a large number of zooplankton species (other than copepods) exploit phytoplankton production (from microzooplankton to euphasiids and salps). For example, >26 zooplankton taxa have been reported to occur in the North Sea and northeast Atlantic Ocean (Colebrook, 1984). In light of such diversity, it is unlikely that combining a stage-based model for only one copepod species with bulk estimates of production of the remaining species (e.g. see Moll and Stegert, 2007) offers more detailed information about the size- or stage-structure of the zooplankton community than the exclusive use of a bulk formulation. Therefore, modeling a plethora of secondary consumers may be required to accurately simulate the magnitude, timing and especially the size and stage composition of zooplankton production and to provide prey fields that reflect the observed diversity of zooplankton in the diet of larval fish. Unfortunately, in most cases, it is almost impossible to parameterize models for all key species and, even if this was possible, the complexity of such a model would make its utility questionable (Fulton et al., 2003). A stage-resolved copepod IBM published by Fennel (Fennel, 2001) comprised 10 state variables for one species. In a five-species system, 50 variables would need to be solved instead of one or two bulk formulations. This increases not only the complexity of the ecosystem model but also its uncertainties, especially with regard to species interactions. Moreover, computational limits imposed by such a high degree of model complexity will make simulation and analyses of large spatial and/or long temporal scales problematic if not impossible. Even in less complex systems where only a few copepod species constitute the majority of zooplankton biomass (e.g. Baltic Sea: Möllmann and Köster, 2002), computational limitations associated with the spatially-explicit, stage-resolved approach would lead to other problems, specifically, underestimates of the variability and spurious trends in the prey field due to the relevance of neglected species that would definitely limit the applicability of the model to long time series. For complex shelf-sea systems such as the North Sea, the stage-resolved approach may be particularly inappropriate since this region has a relatively high number of relevant larval fish prey species (Munk and Nielsen, 1994) that exhibit high inter-annual variability (e.g. Beaugrand and Ibanez, 2004). A second pitfall related to employing stage-based copepod models is that the dynamics of specific zooplankton species heavily depend on initial conditions that are difficult to adequately capture within models. For example, the timing and magnitude of the invasion of species such as Calanus finmarchicus into shelf sea regions and/or the hatching of resting eggs from the sediments of Acartia spp. are critical constraints on the seasonal dynamics of these copepods.
In short, we can conclude that larvae fish IBMs based on stage-resolved zooplankton models or observed zooplankton prey fields are in-appropriate modeling strategies for addressing climate induced variability and future climate impacts in complex regional systems. For developing a sustainable modeling strategy, it is necessary to focus on the processes we are able to observe and parameterize, instead of increasing the number of uncertain parameters.
For these reasons, we developed an alternative approach to generate size-structured prey fields based on estimates of zooplankton bulk biomass from a 3D ecosystem model "ECOSMO" (Schrum et al., 2006). This NPZD model reflected well the spatial and temporal trends in the zooplankton in the North Sea (for details, see Schrum et al., 2006), and we raise the potential here for such a structure to be used in other ecosystems.
|
A new approach for size-structured prey fields |
|---|
|
TOPABSTRACTIntroductionPrey fields from observations:...Prey fields from stage-based...A new approach for...ConclusionsFundingReferences |
|---|
In our method, the modelled bulk biomass is partitioned into different prey size categories (pli) using a relative size frequency distribution of zooplankton and the individual prey mass in a given category. To accomplish this, a relative size frequency distribution (SDpli) of zooplankton has to be assumed. In our approach for the North Sea, we have used an exponentially decreasing function of abundance with increasing prey length (Fig. 1) based on mean, size-specific zooplankton concentrations gained from two North Sea field sampling programs: LIFECO (www.lifeco.dk/default.htm) and the ongoing German GLOBEC program (e.g. Fig. 2A)
|
|
|
|
Since the prey size categories were composed of different zooplankton species, particularly by different copepod species, the individual prey mass in a certain category mi (µg carbon) was defined as:
|
|
; Berggreen et al., 1988; Köster, 2003).
The total biomass within a particular prey size category was calculated by multiplying the corresponding relative size frequency, representing the numbers of individuals, with the individual prey mass. By normalizing this value to the total biomass (tm), the percentage of the total biomass represented by a specific prey size category (SPpli) (Fig. 1) can be represented as:
|
|
|
|
The exponential size distribution (SDpl) that we derived from in situ zooplankton abundance estimates in the North Sea agrees with the findings of a number of primary publications dealing with the topic (e.g. Sheldon and Parsons, 1967; Gordon, 1970; Sheldon et al., 1972). For example, observations from Sheldon and Parson (Sheldon and Parsons, 1967) and Gordon (Gordon, 1970) indicated an exponential decrease in the number of particles within a size distribution encompassing particles 
128 µm. Owing to trophic considerations, it is reasonable that these observations can be extrapolated to larger sizes. This appears to be the case since the relative biomass distribution that we generated corresponded well to calculations made by Munk and Nielsen (Munk and Nielsen, 1994) using in situ data on size-specific zooplankton abundance at Dogger Bank, North Sea (Fig. 2B), thus providing confidence in our conversion method and prey field assumptions. Although the assumptions are considered to be a weakness, they are part of the price that must be paid to reduce model complexity.
It should be noted that field data such as those presented by Munk and Nielsen (Munk and Nielsen, 1994) and Kuehn et al. (Kuehn et al., unpublished data) provide a snapshot of a zooplankton size distribution that will inevitably be undergoing change. We have addressed this problem by performing sensitivity analyses to quantify the impact of different zooplankton size distributions on IBM model estimates of potential larval fish survival and growth (Daewel et al., unpublished data). In that study, large differences existed in estimates of potential growth and survival of larval fish as a result of changes in the prey size spectrum (particularly changes that limited the abundance of the sizes of prey consumed by first-feeding larvae). It is therefore apparent that inclusion of a deterministic description of size spectra variability on different temporal and spatial scales as affected by changes in total biomass is critical for a proper description of the functional response of climatic effects on larval growth and survival. The approaches from size spectra modelling (e.g. Pope et al., 1994; Andersen and Beyer, 2006) appear to be a productive avenue for future research in this direction.
|
Conclusions |
|---|
|
TOPABSTRACTIntroductionPrey fields from observations:...Prey fields from stage-based...A new approach for...ConclusionsFundingReferences |
|---|
To summarize, the concept presented here represents a new approach to developing a size-structured prey field using NPZD model estimates of bulk zooplankton carbon. Although we applied this conversion method to derive prey fields for a larval fish IBM in the North Sea (Daewel et al., submitted), it has potential for ready application to other regions where ecosystem models currently provide bulk carbon estimates of zooplankton biomass and where in situ measurements of zooplankton abundance-at-size have been made. The technique avoids problems that limit the utility of copepod structured population models such as the lack of realistic utilization of phytoplankton production and the lack of zooplankton boundary condition data. Furthermore, the approach identifies potential nodes for the coupling of NPZD models and larval fish IBMs that will be critical for the development of end-to-end ecosystem models that use, for example, a rhomboid approach (de Young et al., 2004
). The method we proposed here would allow a decrease in the complexity of the zooplankton rhomboid thus decreasing the overall model complexity. |
Funding |
|---|
|
TOPABSTRACTIntroductionPrey fields from observations:...Prey fields from stage-based...A new approach for...ConclusionsFundingReferences |
|---|
Global Ocean Ecosystem Dynamics (GLOBEC Germany) program funded through the German Federal Ministry for Education and Research (BMBF 03F0320E) and the German Science Foundation (DFG) AQUASHIFT program cluster Resolving the Trophodynamic Consequences of Climate Change ("RECONN," DFG # JO556/1-1).
|
Acknowledgements |
|---|
We would like to thank four anonymous reviewers for their helpful comments on an earlier draft of this manuscript.
|
Notes |
|---|
Written responses to this article should be submitted to R. P. Harris at r.p.h@pml.ac.uk within two months of publication. For further information, please see the Editorial ‘Horizons’ in Journal of Plankton Research, Volume 26, Number 3, Page 257.
Communicating editor: K.J. Flynn
|
References |
|---|
|
TOPABSTRACTIntroductionPrey fields from observations:...Prey fields from stage-based...A new approach for...ConclusionsFundingReferences |
|---|
Andersen K. H., Beyer J. E. Asymptotic size determines species abundance in the marine size spectrum. Am. Naturalist (2006) 168:54–61.[CrossRef][ISI]
Beaugrand G., Ibanez F. Monitoring marine plankton ecosystems. II: long-term changes in North Sea calanoid copepods in relation to hydro-climatic variability. Mar. Ecol. Prog. Ser. (2004) 284:35–47.[CrossRef]
Berggreen U., Hansen B., Kiørboe T. Food size spectra, ingestion rate and growth of the copepod Acartia tonsa during development: implications for determination of copepod production. Mar. Biol. (1988) 99:341–352.[CrossRef]
Carlotti F., Radach G. Seasonal dynamics of phytoplankton and Calanus finmarchicus in the North Sea as revealed by a coupled one-dimensional model. Limnol. Oceanogr. (1996) 41:522–539.
Carlotti F., Wolff U. A Lagrangian ensemble model of Calanus finmarchicus coupled with a 1-D ecosystem model. Fish. Oceanogr. (1998) 7:225–242.
Colebrook J. M. Continuous plankton records: relationships between species of phytoplankton and zooplankton in the seasonal cycle. Mar. Biol. (1984) 83:313–323.[CrossRef]
Cushing D. H. Marine Ecology and Fisheries (1975) Cambridge: Cambridge University Press. 278. p.
Davis C. S. Predatory control of copepod seasonal cycles on Georges Bank. Mar. Biol. (1984) 82:31–40.[CrossRef]
de Young B., Heath M., Werner F., et al. Challenges of modeling ocean basin ecosystems. Science (2004) 304:1463–1466.
Fennel W. Modeling of copepods with link to circulation models. J. Plankton Res. (2001) 23:1217–1232.
Fennel W., Neumann T. Variability of copepods as seen in a coupled physical-biological model of the Baltic Sea. ICES Mar. Sci. Symp. (2003) 219:208–219.
Fulton E. A., Smith A. D. M., Johnson C. R. Effect of complexity on marine ecosystem models. Mar. Ecol. Prog. Ser. (2003) 253:1–16.[CrossRef]
Gordon D. C. Jr. A microscopic study of organic particles in the North Atlantic Ocean. Deep-Sea Res. (1970) 17:175–185.[ISI]
Hinrichsen H.-H., Möllmann C., Voss R., et al. Bio-physical modelling of larval Baltic cod (Gadus morhua) growth and survival. Can. J. Fish. Aquat. Sci. (2002) 59:1858–1873.[CrossRef]
Köster P. J. Körpermaße und Kohlenstoffgehalte dominanter Zooplanktonarten der Ostsee in Abhängigkeit von Salzgehalt, Ernährungszustand und Temperatur. (2003) Diploma Thesis, University of Rostock.
Lough R. G., Buckley L. J., Werner F. E., Quinlan J. A., Edwards K. P. A genernal biophysical model of larval cod (Gadus morhua) growth applied to populations on Georges Bank. Fish. Oceanogr (2005) 14:241–262.[CrossRef]
Moll A., Stegert C. Modelling Pseudocalanus elongatus stage-structured population dynamics embedded in a water column ecosystem model for the northern North Sea. J. Mar. Syst. (2007) 64:35–46.[CrossRef]
Möllmann C., Köster F. W. Population dynamics of calanoid copepods and the implications of their predation by clupeid fish in the Central Baltic Sea. J. Plankton Res. (2002) 24:959–977.
Munk P., Nielsen T. G. Trophodynamics of the plankton community at Dogger Bank: predatory impact by larval fish. J. Plankton Res. (1994) 16:1225–1245.
Neumann T., Fennel W. A method to represent seasonal vertical migration of zooplankton in 3D-Eularian models. Ocean Model. (2006) 12:188–204.[CrossRef]
Parrish R. H., Bakun A., Husby D. M., et al. Comparative climatology of selected environmental processes in relation to eastern boundary current pelagic fish reproduction. In: Proceedings of the Expert Consultation to Examine Changes in Abundance and Species Composition of Neritic Fish Resources San José’ Costa Rica April 1983. FAO. Fish. Rep—Sharp G. D., Csirke J., eds. (1983) 731–777.
Peck M. A., Daewel U. Physiologically based limits to food consumption, and individual-based modeling of foraging and growth of larval fishes. Mar. Ecol. Prog. Ser. (2007) 347:171–183.[CrossRef]
Pope J., Shepherd J., Webb J. Successful surf-riding on size spectra: the secret of survival in the sea. Proc. R. Soc. Lond. Ser. B (1994) 343:41–49.
Schrum C., Alekseeva I., St John M. ECOSMO, a coupled ecosystem model of the North Sea and Baltic Sea. Part I: model description and simulation of annual characteristics in the North Sea. J. Mar. Syst. (2006) 61:79–99.[CrossRef]
Sheldon R. W., Parsons T. R. A continuous size spectrum for particulate matter in the sea. J. Fish. Res. Board Can. (1967) 24:909–915.
Sheldon R. W., Prakash A., Sutcliff W. H. Jr. The size distribution of particles in the ocean. Limnol. Oceanogr. (1972) 17:327–339.
Werner F., Quinlan J., Lough R., et al. Spatially explicit individual based modeling of marine populations: a review of the advances in 1990’s. Sarsia (2001) 86:411–421.[ISI]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||