A latent heat nudging scheme for the assimilation of precipitation data into an operational mesoscale model

Abstract

Surface precipitation-rate estimates derived from radar data are potentially of considerable value to high-resolution Numerical Weather Prediction (NWP) models. This paper describes a scheme developed to assimilate precipitation rates derived from the UK weather radar network into the UK Met. Office Mesoscale Model, with the aim of improving the analysis and forecast of precipitation. It is based on 'latent heat nudging', in which the model profiles of latent heating are scaled by the ratio of observed and model precipitation rates. This causes the model to adjust so that the diagnosed precipitation rate agrees more closely with observations.

The assimilation algorithm is outlined, and the results of a trial of the scheme are described. The scheme brings an increase in forecast skill for precipitation distribution in the first six to nine hours of the forecast, a conclusion supported both by objective verification against radar data and subjective assessment of 14 forecasts. The main benefit was found to occur in frontal cases. The scheme was implemented operationally on 16 April 1996.