Article:

Gyalistras, D. & Fischlin, A., 1999. Towards a general method to
        construct regional climatic scenarios for model-based impacts 
        assessments. Petermanns geographische Mitteilungen, 143(4): 251-264.

Abstract:

Studies that use simulation models to assess possible regional impacts of climatic change have very diverse and demanding requirements for climatic input data. This paper presents a general method to construct climatic scenarios for such studies. It was developed in the context of several case studies dealing with possible climatic impacts on forest succession, grasslands, and snowpack/run-off in the European Alps. The following set of requirements was identified from the case studies and other, independently formulated scenario needs: The method should provide physically consistent, spatially and temporally extended scenarios at a high spatial and temporal resolution, be based upon a statistically accurate description of local weather and climate, and be robust, flexible, formally defined, and efficient. Based on our case studies and an evaluation of existing methods we derived the following general procedure: (1) Describe weather at the locations of interest as a stochastic process; (2) Estimate the process/climatic parameters for present climate from measurements; if no measurements are available, use a spatial interpolation procedure to estimate the parameters from nearby climatological stations; (3) Apply statistical downscaling to the output of a climate model to estimate time-dependent changes in selected climatic parameters; (4) Use the results from downscaling to adjust the parameters of the stochastic process, and generate weather sequences by means of stochastic simulation. The paper describes the current implementation of the individual components of the method (downscaling, stochastic weather generation, interpolation) and their combination within an overall framework for scenario construction. The proposed method is then compared to alternative approaches and discussed in light of our case studies. It is concluded that the method satisfies most of the above formulated requirements and thus provides a generally useful technique for model-based impact assessments. A main limitation presents the extensive use of statistical-descriptive models, which may not necessarily hold under a future climate. However, the proposed method supports extensive sensitivity studies, and thanks to its modular structure enhancements of the individual components can be easily incorporated as soon as they become available.

Full text (2.4MB)   (lower quality, but smaller 1.5MB)