Ensemble precipitation forecasts made with quantile regression forests and deterministic harmonie-arome inputs

A gridded 50-member ensemble of precipitation forecasts that are created using a tree-based machine learning method, quantile regression forests, and inputs from the deterministic Harmonie-Arome (HA) forecasts. The target data set is rain-gauge-adjusted radar data that is upscaled by taking 3x3 km means and then a maximum is taken in a 7.5 x 7.5 km box. Inputs to the machine learning model include HA precipitation, and indices of atmospheric instability. Spatial and temporal dependencies are restored using the Schaake Shuffle. Forecasts are available during the extended summer period (mid-April to mid-October). Hourly forecasts are issued 4 times per day (00, 06, 12 en 18 UTC) for 48-hours into the future.