Heavy rain intensity distributions on varying time scales and at different temperatures

On the basis of observational precipitation data at a temporal resolution of 5 min from six stations in Germany we obtain scaling relations of the probability distributions of precipitation intensity with temperature and time scale. Each station record contains an approximately 30 year time series of data. By producing a cascade of averaging intervals, we obtain the behavior of precipitation intensity from the instantaneous to the daily resolution. While the intensity distribution of the shortest time scale displays a strict power law tail, it acquires a more elaborate scaling when temperatures are distinguished or when precipitation and dry periods are mixed at longer averaging intervals. The coefficient of increase with temperature is a continuously and strongly varying function of temperature and percentile and does not show an abrupt increase as noted in previous work. Conversely, when considering precipitation events, we find that the temperature dependence is reduced when the amount, not the intensity, of total precipitation produced is considered. As temperature increases, event duration decreases and reduces the accumulated precipitation yield. We caution that the Clausius-Clapeyron relation may not provide an accurate estimate of the temperature relationship of precipitation at any temporal resolution.