Sudden regime shifts and ecosystem collapses are likely to occur in stressed ecosystems. A better understanding of the relative importance of top-down (e.g. overfishing) versus bottom-up (e.g. increased nutrient input causing eutrophication) controls is essential and can only be achieved through modeling (Daskalov 2002). Catastrophic regime shifts have been related to alternative stable states which can be linked to a critical threshold (Scheffner and Carpenter 2003, Scheffner et al . 2001). Based on this hypothesis, a gradual increase of one driver has little influence until a threshold is reached at which a large shift is observed that is difficult to reverse.
Ecotoxicological studies have shown that toxic contaminants affect the mortality and the reproduction rate of populations. On the other hand, the increased nutrient load may raise the primary production of phytoplankton at the bottom of the food chain. Thus, the joint effect of pollutants and nutrients on a coastal ecosystem is difficult to evaluate, since many direct and indirect effects have to be considered.
In this study we try to represent the effect of contaminants on population dynamics on a theoretical basis, by means of bifurcation analysis, which assesses the long-term behavior of non-linear systems for different parameter regimes. Assuming that the toxicity of contaminants affects only the mortality of one population, the asymptotic behavior of the
food chain model is calculated for different mortality rate values.
When studying the dynamics of simple food chain and food web models it is important to bear in mind that the response might depend on the complexity of the represented system. Chaotic dynamics seem to be more frequent in simple model ecosystem or in models with a high number of trophic levels (Fussmann and Herber 2002). In this report
we compare the dynamics of three different food-chain models for an increased mortality rate caused by the presence of contaminants and increased nutrient loads. Focusing on the long-term dynamical behavior of tri-trophic food-chains, the objective of the study is to test if oscillatory and chaotic dynamics are likely to occur under biologically meaningful parameter ranges and if these dynamics are observed across different models representing similar system with different grade of detail. Furthermore, this kind of study may point out some counterintuitive indirect effects.