ASSIMILATION OF M8-AMV DATA IN THE HIRLAM-NWP MODEL

Atmospheric Motion Vectors (AMV) generated from M8 geostationary satellite images have been used to carry out observation usage experiments with the HIRLAM assimilation and short-range limited-area NWP system. The experiments have been performed at two different operational NWP centres within the HIRLAM consortium, the Danish Meteorological Institute (DMI) and the Spanish National Meteorological Institute (INM). The main target of these experiments has been to test the impact of some new features and enhancements incorporated to AMV data over recent past (e.g., new imagery data) on the quality of weather predictions. The results obtained at DMI indicate that M8-AMV data has a slight positive impact on this quality as measured with surface and upper-air in-situ observations from the EWGLAM stations list. This implies that geostationary AMV wind data can be accommodated in the data assimilation schedule, without necessarily expecting a degradation of the forecast skill scores, together with other polar satellite data (scatterometer sea-surface winds and AMSU-A temperature soundings) currently used at DMI. The results obtained at INM indicate that AMV data has a significant impact on the way the model simulates the Atmospheric Dynamics at middle and low latitudes over the Northern Hemisphere Atlantic Ocean, that this impact is positive and sizeable (about 20 % geopotential error reduction for the +24 h forecast) when measured by verification with analysed fields, and that it is particularly clear at low and middle levels in the Troposphere