Currently, global warming due to increasing concentrations of CO, and other greenhouse gases is widely accepted. Climate is an important forcing parameter of landslides and, hence, implications of climate change for landslide activity are of high concern for geomorphological research. The present paper offers a method for assessing climate change impacts on landslide frequency based on general circulation models (GCM). GCM results are downscaled with an empirical-statistical technique to derive local precipitation scenarios. These scenarios are used as input to a simple slope hydrological and stability model. The landslide is defined 'active' if simulated groundwater levels exceed a critical level established with the stability model. Recurrence intervals for landslide activity are obtained by applying a Gumbel regression to the simulated annual maximum groundwater levels. Furthermore, it is shown that indirect climate change impacts as well as changing non-climatic parameters can be important for future landslide frequencies too. The use of three different GCM experiments for the assessment of the activity of a small landslide in SE France did not show a consistent picture of future landslide frequencies. This is due to differences between the GCM experiments but might be enhanced by the limited ability of the applied downscaling technique to carry climate change signals. Finally, some possibilities of improving the approach are outlined and the need for better GCM experiments, which provide the basic input of the approach, is addressed. (C) 1999 Elsevier Science B.V. All rights reserved.
