On the use of Radar data to verify the long term performance of the UK Met Office Mesoscale Model precipitation forecasts

So far, little use has been made of the high spatial and temporal resolution of radar data to objectively verify operational model forecasts of precipitation. First results of a multiyear comparison of radar data with precipitation forecasts from the UK Met. Office mesoscale model will be presented.

Emphasis will be put on the objective determination of the spatio-temporal scales on which the observed and modelled precipitation agree best. In order to do so, different measures will be computed, e.g. unconditional and conditional distributions, skill scores, wavelet spectra, timing and duration of precipitation events. Spatial and temporal aggregation will be used to determine the optimal match and to address the need of customers to know the verification characteristics on, e.g. a catchment scale or in mountainous areas. As far as possible, rigorous statistical testing will be applied to determine the significance of differences found in the above mentioned measures between observations and forecasts and between the performance in different weather situations, e.g. convective versus frontal rain and strong versus light rain.