An Adaptive Nonlinear MOS Scheme for Precipitation Forecasts Using Neural Networks

Abstract

A novel neural network (NN)?based scheme performs nonlinear model output statistics (MOS) for generating precipitation forecasts from numerical weather prediction (NWP) model output. Data records from the past few weeks are sufficient for establishing an initial MOS connection, which then adapts itself to the ongoing changes and modifications in the NWP model. The technical feasibility of the algorithm is demonstrated in three numerical experiments using the NCEP reanalysis data in the Alaskan panhandle and the coastal region of British Columbia. Its performance is compared with that of a conventional NN-based nonadaptive scheme. When the new adaptive method is employed, the degradation in the precipitation forecast skills due to changes in the NWP model is small and is much less than the degradation in the performance of the conventional nonadaptive scheme.