Reconstructing historical marine ecosystems using food web models: Northern British Columbia from Pre-European contact to present
Introduction
For thousands of years, humans have been exploiting the seas for food (Bar-Yosef, 2004, Yellen et al., 1995, Fiore et al., 2004). Paleoecological and archaeological evidence records the significant impacts we have had (Jackson et al., 2001). However, it was not until the development of industrial fisheries, less than 200 years ago, that the major depletion of marine systems began (e.g., Myers and Worm, 2003, Pauly et al., 2005). Only recently has the magnitude of historic declines become apparent (e.g., Christensen et al., 2003, Myers and Worm, 2003, Rosenberg et al., 2005, Ward and Myers, 2005).
Human activities may have contributed in other ways to fundamental changes in the structure and functioning of marine ecosystems. Changes in ecological life-histories caused by over-exploitation (Odum's ratchet), and mal-adaptive economic investment (Ludwig's ratchet) are two ratchet-like depletion processes (Pitcher, 2001). A third ratchet is provided by Pauly's (1995) shifting baseline syndrome. He suggested that one's concept of abundance is based on a mental benchmark set at the beginning of one's career. As the ecosystem is slowly degraded, each generation accepts a lower standard as normal. Our view then of what a healthy ecosystem should look like may be biased and under-achieving.
If we could explicitly quantify changes that have occurred over long time scales it could help combat the shifting baseline syndrome, inform science behind ecosystem-based management (EBM: FAO, 1995, Garcia et al., 2003, Hall and Mainprize, 2004, UNCLOS, 2005), and help us design intelligent restoration targets (Ainsworth and Pitcher, 2008). Useful tools for EBM are still under development (Fulton et al., 2007), but ecosystem models are playing an increasingly important role because of the holistic picture they can provide.
In this article, we assemble the available information from scientific sources and traditional and expert knowledge to represent four historical states of the Northern British Columbia (BC) marine ecosystem over the last 250 years. Our focus is on the marine food web and we use Ecopath with Ecosim models (EwE: Christensen and Pauly, 1992, Walters et al., 1997) to represent a ‘best guess’ of what the historic ecosystems may have looked like. Four time periods are chosen to represent distinct eras in the development of west coast fisheries: prior to European contact (c. 1750), before the introduction of steam trawlers (c. 1900), during the peak of the Pacific salmon fishery (1950), and in the present day (2000). This structure helps us frame the historical impacts of humans on the marine ecosystem. For discussion on the history of this system and its fisheries, see Pitcher et al. (2002); also, Ainsworth et al. (2002) and Alcock et al. (2007).
To move from the static representation offered by Ecopath to the dynamic representation used by Ecosim we are required to tune parameters that are poorly supported by data such as Ecosim's predator–prey ‘vulnerabilities’. These critical parameters describe the density-dependent functional response regulating predation mortality in trophic interactions. The process of tuning the model to observational data allows us to hone in on a plausible set of parameters. However, in building models of the distant past, as well as models of the present, a long time series of observational data is often lacking. We introduce a new technique here to parameterize vulnerabilities of these models. The parameter set for the 1950 model is fitted to observational data and then rescaled and applied to the other time periods by assuming stationarity in the density-dependent foraging tactics of species.
The recent history of climate fluctuations in northern BC may have been recorded in the population dynamics of the ecosystem. By analyzing the reconstructed dynamics of the 1950 model we can improve our understanding of recent ecosystem functioning. For example, by making assumptions on the loss of system energy it is possible for Ecosim to back-calculate the amount of primary production required to explain observed dynamics.
The reconstruction of production trends can teach us what impact climate fluctuations may have had on the ecosystem, so we compare calculated anomaly patterns with regional climate series. If we can assume that climate variations in the future will be similar to the past, then we can apply the inverse technique to calculate depletion risks for predictive policy forecasts. Policy forecasts using this technique are made elsewhere: previous authors have used this method to conduct an ecosystem-level population viability analysis and quantify risk associated with policy scenarios (Pitcher et al., 2005, Ainsworth et al., 2008b). Ainsworth (2006) did this with the models described here. A similar technique using production forcing patterns could be used to explore consequences of directional climate change and climate change scenarios; however, EwE is limited to a relatively simple representation of effects.
Driving the 1950 model forward 50 years should produce a new ecosystem structure that is similar to the 2000 model if the dynamics are being driven by historical fishing mortalities and climate forcing. We compare two models of the 2000 ecosystem, one that is built on current scientific data (the ‘proper’ model) and one that represents the end-state of a 50-year simulation using the 1950 model (the ‘derived’ model). The static condition of the derived model is compared to the proper model, and the dynamic responses of the two models are compared using predictive forecasts to 2050. In the forward simulation, we eliminate fishing and watch the systems recover. Since the models use related vulnerability matrices we expect the carrying capacity of functional groups to be similar, but compounding effects of species interactions still have the potential to create alternate stable ecosystem states.
In this contribution, we compare four Ecopath models for northern BC time periods: 1750, 1900, 1950 and 2000. We tune the 1950 model to time series data, comparing the calculated climate anomaly trends for primary production and herring recruitment to known environmental indices, and introduce a technique to recreate accurate variability in species biomass for depletion risk scenarios. We transfer the fitted vulnerability matrix of the 1950 model to the 1900 and 2000 models, assuming stationarity in foraging tactics of species, and perform various tests with both models to validate the new method. We also comment on significant findings of the dynamic simulation from 1950 to 2000, including predicted population dynamics, trophic control of the ecosystem, and the effects of regional climate. Finally, we comment on the approach of modelling historic ecosystems as a means to improve forward-looking predictions.
Section snippets
Ecopath with Ecosim
EwE was invented by Polovina (1984) and advanced by Christensen and Pauly, 1992, Christensen and Pauly, 1993, Walters et al., 1997, Walters et al., 1998, Walters et al., 2000 and Christensen and Walters (2005) among others. The suite of programs includes Ecopath, which provides a static representation or snapshot of the ecosystem, and Ecosim, which allows dynamic simulation. The trophic model Ecopath acts as a thermodynamic accounting system as it summarizes the instantaneous flows of matter
Comparison of static models
Fig. 1 shows the decline in calculated commercial fish biomass in northern BC from 1750 to 2000 from Ecopath, and the increase in fisheries catch throughout the same period. Fig. 2 shows the fraction of system primary production required to sustain capture fisheries in these periods. Prior to European contact, First Nations took a small annual catch of mammals by our estimation. However, the exploitation increased during the first half of the 20th century due to the industries of Russians and
Population dynamics
The population dynamics of seals and sea lions could not be reconciled with the available catch and biomass trends. In developing a model from the same roots but independent of ours, Preikshot (2005) encountered a similar problem using the same data series for observed pinniped biomass in BC, which suggests that the anomaly may derive from the data. The data represents an average of biomass change throughout the whole BC coast (harbour seals: Olesiuk, 1999; sea lions: Bigg, 1985), so one likely
Conclusion
Ecosystem models have the ability to describe and predict long timescale patterns in species abundance and food web interactions. They will become a more important part of EBM and ecosystem science because they offer a kind insight that is unavailable from classical models. They can be used to create a coherent view of the ecosystem based only on piecemeal information, by using fundamental assumptions on ecosystem functioning. Observational data places further constraints on the functioning of
Acknowledgements
The authors would like to thank Dr. Carl Walters, Dr. Villy Christensen and William Cheung at the University of British Columbia Fisheries Centre for their input. We also acknowledge support from the University of British Columbia Graduate Fellowship fund.
References (95)
- et al.
Estimating illegal, unreported and unregulated catch in British Columbia's marine fisheries
Fisheries Research
(2005) - et al.
Evidence of fishery depletions and shifting cognitive baselines in Eastern Indonesia
Biological Conservation
(2008) - et al.
Ecosystem simulations supporting Ecosystem Based Fisheries Management in the Coral Triangle, Indonesia
Ecological Modelling
(2008) - et al.
Linking oceanic food webs to coastal production and growth rates of Pacific salmon (Onchorhynchus spp.), using models on three scales
Deep-Sea Research II
(2005) - et al.
Trade-offs between conservation and socioeconomic objectives in managing a tropical marine ecosystem
Ecological Economics
(2008) - et al.
ECOPATH II—a software for balancing steady-state models and calculating network characteristics
Ecological Modeling
(1992) - et al.
Ecopath with Ecosim: methods, capabilities and limitations
Ecological Modeling
(2004) - et al.
Empirical evidence for North Pacific regime shifts in 1977 and 1989
Progress in Oceanography
(2000) - et al.
Have there been recent changes to climate? Ask the fish
Progress in Oceanography
(2000) Anecdotes and the shifting baseline syndrome of fisheries
Trends in Ecology and Evolution
(1995)