Combining statistical and ensemble methods for probabilistic wind field prediction

Abstract

Probabilistic wind forecasting is a methodology to deal with uncertainties in numerical weather prediction models (NWP). In this work, we present a combination of statistical and ensemble methods designed for the downscaling wind model Wind3D coupled with the HARMONIE-AROME mesoscale model. The ensemble method is based on a Monte Carlo approach, and the statistical method is the linear model in quantile regression (LMQR). The combination of both results give a probabilistic forecasting.