MIDAS – Mesoscale Improved Data Assimilation of Scatterometer Winds
MIDAS is a collaborative project between KNMI and IPMA that aims to provide EUMETSAT users and the NWP community relevant information on the use of scatterometers in mesoscale DA. We investigated the use of scatterometer wind data in the non-hydrostatic convective-scale-resolving model HARMONIE-AROME. A 4D-Var data assimilation configuration was evaluated against a 3D-Var formulation still used in operations in Numerical Weather Prediction (NWP) centres using HARMONIE-AROME. The best performance of 4D-Var was shown for 10-m winds over the ocean and over land. The added value of including scatterometer winds in the assimilation system is also demonstrated. A number of observing system experiments (OSE) were designed to identify the optimal scatterometer winds usage in HARMONIE-AROME 4D-Var. Thinning procedures, observation error prescription and superobbing of observations were tested. While the added value of using the 4D-Var configuration and of including scatterometer observations in the assimilation system was clearly shown, the different approaches tested on the use of scatterometer winds showed marginal differences. On the other hand, it is shown that, over the ocean, HARMONIE-AROME has systematic biases both in speed and direction when compared with ScatSat and HSCAT derived winds. Model winds present clockwise rotation against observations and a systematic speed overestimation. Although the bias is partially corrected by 4D-Var and ASCAT winds assimilation, the model returns quickly to the biased atmospheric state. The study demonstrates the need to further understand the causes of these biases in HARMONIE-AROME in order to address them. Furthermore, bias adjustment is needed to optimally exploit the observational information.
