Dependence of radar quantitative precipitation estimation error on the rain intensity in the Cévennes region, France

Abstract

Radar quantitative precipitation estimates (QPEs) were assessed using reference values established by means of a geostatistical approach. The reference values were estimated from raingauge data using the block kriging technique, and the reference meshes were selected on the basis of the kriging estimation variance. Agreement between radar QPEs and reference rain amounts was shown to increase slightly with the space-time scales. The statistical distributions of the errors were modelled conditionally with respect to several factors using the GAMLSS approach. The conditional bias of the errors presents a complex structure that depends on the space-time scales and the considered geographical sub-domains, while the standard deviation of the errors has a more homogeneous behaviour. The estimation standard deviation of the reference rainfall and the standard deviation of the errors between radar and reference rainfall were found to have the same magnitude, indicating the limitations of the available network in terms of providing accurate reference values for the spatial scales considered (5-100 km 2 ). © 2014 IAHS Press.