engleski

Estimation of precipitation extremes in ungauged areas

Very often there is a demand from users for estimates of the maximum precipitation relative to different return periods at locations where there are no measurements, or limited precipitation records. This analysis aims to respond to this problem.In order to classify estimated maximum daily precipitation for Croatia, the pairs of parameters of the Jenkinson GEV distribution R0, k and a were grouped by means of cluster analysis. Clustering of parameters k and  did not indicate on existence of larger homogeneous areas. Therefore, the further regional analysis has been based on spatial characteristics of parameter R0 only.The standardized values RT/R0 (growth factor), being only a function of R0, have been calculated for all stations in each group individually and then their probability distribution for each group. RT/R0 have been calculated for selected percentiles: median, 80th percentile, 90th percentile, 95th percentile, 98th percentile, and 99th percentile. Obtained growth factors are applicable to any observed location defining at first its R0. It can be calculated for locations having even short data series. For locations without measurements, values of R0 from the closest, climatologically similar station can be associated. According to these R0 values the dependence of the location to one of clusters is determined, in order to apply respective growth factors for calculation of expected annual RT for certain return period T in regional normal climate of short-term precipitation.

daily precipitation maximum; Jenkinson GEV distribution; growth factors

nije evidentirano