Evaluation of models capacity to predict size spectra parameters in ecosystems under stress

- Size spectra models are capable of predicting effects of stress on average.
- Empirical studies confirm a decreasing trend for the slope of stressed ecosystems.
- SS's Intercept and deviation from linear model are positively related to specific stressors.
- Data ranging 74 stress situations, 19 orders of magnitude in size and 20 stressors.

Size spectra are important indicators of biomass distribution, for fundamental understanding as well as for application in management. While many size spectra models have been developed, it is unclear how response to stress varies across models and in comparison to data. In the present study, we therefor collected papers on empirical size spectra in stressed aquatic ecosystems and compared the results to theoretical models' predictions from literature. A general decrease of the slope, as suggested by most of the studied models, is supported by the data found. The investigated empirical studies revealed a strong trend for the slope: 40 out of 57 studies reported a decrease with stress. Deviations from the linear model increased in 10 out of 20 cases and the intercept became larger in 25 out of 43 cases. The data concerning specific impact drivers is generally consistent and supportive of current expert opinions. Our results support that an integration of deviation from the linear model and intercept in the size spectra models could provide better coupling between modeling and field surveys for some stressors.

93