Addressing predictability in the AROME-Arctic ensemble using Stochastically Pertrubed Parameters (SPP) and Ensemble Data Assimilation (EDA)

This report describes work done as part of the Alertness product to address uncertainties in the AROME-Arctic ensemble using Ensemble Data Assimilation (EDA) and Stochastically Perturbed Parameters (SPP). Forecasts using EDA and SPP were compared against reference forecasts using a more basic implementation of the AROME-Arctic ensemble. In general both SPP and EDA led to increased spread in the ensemble than the reference without impacting the errors in a negative way. However, SPP introduced some large biases, especially for 2m temperature and 10m wind speed, and EDA resulted in stringer biases for total cloud cover. It is recommended that SPP should not be taken into use until the biases are rectified, but EDA can be taken into use since overall it leads to a better representation of the uncertainties in the forecasts