Linking GCM-simulated climatic changes to ecosystem models: case studies of statistical downscaling in the Alps

Abstract

Based on the statistical approach proposed by von Storch et al. (1993; J. C h . 6: 1161-1171], a general, flexible method to assess changes in the local climatic inputs of ecosystem models from largescale clunatic changes as simulated by general circulation lnodels (GCMs) was developed, verified and applied at 5 Swiss locations for the summer and winter seasons. According to the requirements of various ecosystem models, at each location 17 seasonal statistics related to daily temperatures, precipitation, sunshine duration, air humidity and wind speed were considered. Year-to-year variations of the local variables were linked by means of Canonical Correlation Analysis to simultaneous anomalies in the North AtlanticEuropean sea-level pressure and near-surface temperature fields. The analysis was performed for the period 1901 to 1940, separately for each season and location. In all cases, physically plausible statistical models were found which quantified the local effects of changes in major circulation patterns. such as the strength of westerly flow in winter and of large-scale subsidence in summer. In the venfication interval 1941 to 1980, most variables were better reconstructed in winter than in summer, and better at the 3 north alpine than at the 2 south alpine locations. The best-reconstructed variables were seasonal mean daily temperatures and daily temperature extremes, for which on average 42 to 75 % of the total variances over the verification interval at the different locations could be explained Explained variances for precipitation totals were 29 to 55% in winter and 10 to 28% in summer, and for mean daily relative sunshine durations in both seasons ca 12 to 53%. Seasonal mean relative humidities, mean wind speeds, and within-summer standard deviations of daily variables were generally poorly reproduced. It was fo.und that the procedure of von Storch et al. can be generally improved by using, in addition to sea-level pressure, the near-surface temperature as a large-scale predictor. Improvement was strongest for temperaturerelated variables, and for the summer season. The established statistical relationships were applied to anomaly fields as simulated by the Hamburg fully coupled atmospheric/oceanic ECHAMULSG GCM under increasing atmospheric greenhouse gas concentrations. The procedure yields time-dependent, internally consistent, and regionally strongly differentiated climatic change estimates for several important ecosystem inputs, at a spatial resolution far above the resolution of present GCMs.