Ensemble data assimilation and model validation studies using cloud and water vapor sensitive infrared brightness temperatures

Jason A. Otkin
University of Wisconsin-Madison 
Noon March 10 in Room 2155

Infrared brightness temperatures provide valuable information about atmospheric water vapor, cloud cover, and surface properties, and thus are very useful for a wide variety of research purposes. In this presentation, results will be shown for several prior and ongoing studies that have used satellite observations to examine the accuracy of high-resolution numerical model simulations and to improve the analysis and forecast accuracy of regional forecast models through use of ensemble data assimilation systems.

The impact of assimilating both clear and cloud-affected infrared brightness temperatures was assessed using regional-scale Observing System Simulation Experiments (OSSEs) with simulated GOES-R Advanced Baseline Imager (ABI) observations and through real data experiments using observations from the SEVIRI sensor onboard the Meteosat Second Generation satellite. Overall, the OSSE results showed that the assimilation of cloud-affected brightness temperatures had a large positive impact on the cloud and moisture analyses, with lesser or no impact on the wind and temperature analyses. Short range precipitation forecasts for a high-impact weather event were greatly improved when cloudy observations were assimilated. For the real data experiments, observations from the SEVIRI 6.2 μm band that is sensitive to clouds and water vapor in the upper troposphere were assimilated. A bias correction scheme based on the observed and simulated cloud top pressure was also developed. The results showed that the biases were greatly reduced when the SEVIRI infrared observations were assimilated using the cloud-dependent bias correction scheme. Similar to the OSSE results, the largest improvements at the end of the assimilation period were observed in the relative humidity field, with lesser impact on the temperature and wind fields.

Preliminary results will also be shown from two new model validation studies that will be used to examine the accuracy of the simulated cloud and water vapor fields in the operational High Resolution Rapid Refresh (HRRR) and Hurricane WRF (HWRF) models. A real-time GOES-based verification system is being developed for the HRRR model that will be used to provide forecasters objective guidance concerning which of the many overlapping forecasts at a given time is the most accurate based on quantitative comparison of the observed and simulated satellite imagery. Similar retrospective studies are being performed for different configurations of the HWRF model to assess the accuracy of the cloud and moisture fields.