|
Abstract:
|
"Social-ecological systems (SESs) are undoubtedly complex. Their complexity presents a major obstacle in research, in particular for the question which system attributes have to be in place to make ecological success probable. To overcome the many problems due to complexity, we suggest several solutions, which in their combination enable us to develop a quantitative model of the ecological success of SESs. First, the limitations of single case studies may be overcome by analysing larger data sets from one source. Among the advantages are the comparability of case studies and a broad quantitative basis for statistical analyses. Second, due to the complex and nonlinear relationships between system attributes, a method is required that is able to cope with these problems. We use artificial neural networks (NN), a non-parametric statistical method. Third, a review of existing frameworks and the literature ensure that almost all relevant system attributes are actually taken into account, not only a fraction. The resulting model, although still limited by a small case number (n = 122), explains the ecological success of irrigation and fisheries dealing with CPR-problems satisfactorily (R² = 0.68), mean standard error (MSE) = 0.02). Such a model could be used for policy-advice e.g. by NGOs or national governments trying to support community-based resource management."
|