On the development of a scatterometer-based correction for NWP wind forcing systematic errors: Impact of satellite sampling

Local systematic differences between scatterometer and global numerical weather prediction (NWP) model stress equivalent winds (SEW) are due to unresolved geophysical processes by the model, e.g., ocean currents and moist convection. A scatterometer-based correction, which contains the mesoscale informationpresent in the Advanced Scatterometer (ASCAT) observations, sets the grounds for a high-resolution ocean forcing product. To assess the effectiveness of such correction, a Monte Carlo simulation procedure is applied to NWP SEW. It allows for a thorough evaluation of the NWP error reduction, which depends on the scatterometer sampling. The local NWP biases are reduced at the cost of a somewhat increased variance, and the total error mitigation is constrained to regions covered by the scatterometer at least 3 times over 5 days. Despite the limited sampling in the tropics, the real NWP corrected SEW over the West African coast show areas of increased wind variability associated to moist convection.