Integrating simulation models into environmental information systems - model analysis

Although environmental information systems have become very popular, the use of simulation models within such systems has not increased at the same rate. Mostly, new and specialized models will be implemented rather than integrating existing ones. This mainly results from the absence of helpful documentations and knowledge about the admissible spatial and temporal scales of the models. This article mainly deals with the necessary steps to analyze simulation models. In particular, the application of long-term simulations (e.g. Global Change) demands precise mathematical knowledge about a models' specific behaviors. similar to signal processing, various methods for analyzing and filtering are necessary to protect model systems against intrinsic conflicts, e.g. interfering scales. In this text we focus on model analysis with the Wavelet transformation. Utilizing this technique, the detection of periodicities within model inputs and outputs is possible. Similar to the Fourier transformation, the Wavelet analysis effects a scale-by-scale decomposition of time series or spatial data. But, in contrast to spectral analysis, the Wavelet transformation retains location information. Concluding with an example, this article shows that the Wavelet approach meets the requirements of model analysis.