Spatial and temporal structure of normal and extreme rainfall

The space time behaviour of precipitation is very complex. The knowledge of the dependence structures in space and time is very important for the assessment of flood risks. In this contribution the dependence structures of normal and extreme events are compared. Both rain gauges with high temporal resolution and radar images are investigated. Spatial and temporal copulas are used for this investigation. Due to the large number of zero observations, especially for short temporal aggregations an indicator approach is used to detect structural differences. The results show, that the temporal dependence structure of rainfall gradually changes with increasing intensity. Similar behaviour can be detected for the spatial structure with the addition of advection related differences in both ranges and angles of anisotropy. The findings indicate that metagaussian approaches which only consider spatial and temporal correlations are not appropriate for the description and the simulation of rainfall extremes. Finally a new structural simulation method using non-Gaussian dependence is presented.