1 Department of Civil Engineering, The Faculty of Engineering and Science, Aalborg University, VBN2 The Faculty of Engineering and Science, Aalborg University, VBN3 Division of Water and Soil, The Faculty of Engineering and Science, Aalborg University, VBN4 Water and Environment Research Group, The Faculty of Engineering and Science, Aalborg University, VBN5 Urban Water and Environment Research Group, The Faculty of Engineering and Science, Aalborg University, VBN6 Centre for Meteorological Models, Research and Development Department, Danish Meteorological Institute, Copenhagen
A new approach for assimilation of 2D precipitation in numerical weather prediction models is presented and tested in a case with convective, heavy precipitation. In the scheme a nudging term is added to the horizontal velocity divergence tendency equation. In case of underproduction of precipitation, the strength of the nudging is proportional to the offset between observed and modelled precipitation, leading to increased moisture convergence. If the model over-predicts precipitation, the low level moisture source is reduced, and in-cloud moisture is nudged towards environmental values. The method was implemented in the Danish Meteorological Institute numerical weather prediction (DMI NWP) nowcasting system, running with hourly cycles, performing a surface analysis and 3D variational analysis for upper air assimilation at each cycle restart, followed by nudging assimilation of precipitation and then a free forecast. The precipitation fields are based on a 2D composite CAPPI (constant altitude plan position indicator) field made from observations with the DMI weather radars, and have a 10 min time resolution. The results obtained in this study indicate that the new method implies fast adjustment of the dynamical state of the model to facilitate precipitation release when the model precipitation intensity is too low. Removal of precipitation is shown to be of importance and the position of the model precipitation cells becomes skilful even at the smallest scales (∼3 km). Bias is reduced for low and extreme precipitation rates. In this meteorological case, the usage of the nudging procedure has been shown to improve the prediction of heavy precipitation substantially.
Meteorological Applications, 2015, Vol 22, Issue 1, p. 48-59