Extended Abstract (2.5M)P5.1
Variational assimilation of SSM/I observations in clear skies at the Meteorological Service of Canada
David Anselmo, MSC, Dorval, PQ, Canada; and G. Deblonde
In recent years, the assimilation of satellite data has become a vital component of the global and regional assimilation systems at the Canadian Meteorological Centre (CMC). Specifically, the direct assimilation of satellite radiances from the Advance Microwave Sounding Unit (AMSU-A and AMSU-B), and the Geostationary Operational Environmental Satellite (GOES) water vapor channel has resulted in notable improvements in the short and medium range CMC forecasts. This has been demonstrated in Observation System Experiments conducted by CMC.
In preparation for the operational assimilation of Special Sensor Microwave Imager (SSM/I) brightness temperatures in the four-dimensional variational (4D-Var) global assimilation cycle at CMC, two 3D-Var experiments are conducted. In the first experiment, observations from the seven SSM/I channels are added to the operational configuration of the CMC's global analysis system. In the second experiment, SSM/I data is added together with a more rigorous quality control of the AMSU data currently assimilated. The objective of the enhanced filtering of AMSU data is to prevent the assimilation of observations affected by non-precipitating clouds. These experiments are conducted for summer and winter conditions.
In all the experiments, improvements are evident in the analyzed integrated water vapour and surface wind speed fields, and in mean daily precipitation rates when compared against independent observation sets. Furthermore, small gains are realized in the forecasts, when validated against radiosonde data. Other indicators such as anomaly correlation, RMSE, and Quantitative Precipitation Forecast (QPF) scores show a net positive effect to the modifications. Overall, the experiment with enhanced filtering of AMSU data shows better results than that with the addition of SSM/I by itself.
Poster Session 5, Data Assimilation
Thursday, 2 February 2006, 9:45 AM-9:45 AM, Exhibit Hall A2
Next paper