Ensemble radar precipitation estimation — a new topic on the radar horizon

The uncertainty in radar precipitation estimates is the superposition of errors of many different error sources such as ground clutter, hardware instability, beam shielding, beam broadening, variability in the hydrometeor phase and size distribution, and signal attenuation by water on the radome or in the atmosphere. In the past decade MeteoSwiss developed and implemented a series of sophisticated algorithms to correct for several of the above errors. In spite of significant improvements, for hydrological applications the residual uncertainty is still relatively large. An elegant way to express this residual uncertainty is the generation of an ensemble of radar precipitation fields using stochastic simulation and knowledge of errors. The ensemble represents the uncertainty in radar precipitation estimates by introducing perturbations with the correct spacetime variances and auto-covariances. The variance and autocovariance of errors strongly depend on the location, in particular in a mountainous region such as Switzerland. This dependence on location needs to be taken into account in the stochastic simulation. The paper presents a prototype ensemble generator that allows full flexibility with respect to the radar error structure.