Improved modelling and analysis of variable resolution IR and MW satellite observations using convective scale numerical weather prediction

The accurate and timely forecasting of extreme rainfall events is of the utmost importance to both life and property. Satellite observations from multi-spectral, multi-sensor observations provide information that aid the forecasting of such events. To advance the accuracy of these forecasts at the necessary temporal and spatial resolutions, further research is necessary to address the optimal use of such information. Satellite observations exist at a range of temporal and spatial resolutions, and different wavelengths sense different depths within weather systems. These effects lead to different errors in modelling the observations arising from, for example, heterogeneous field of views, and representativeness leading both to correlated and uncorrelated components of error. It is not well understood how these errors propagate into the analysis using variational data assimilation. The UK Meteorological Office operates a number of numerical weather prediction models, of which the 1.5km convective scale model provides very high spatial and temporal resolution short-range forecasts of weather systems. However, the satellite observations assimilated into the high-resolution model have generally much coarser resolutions (both spatially and temporally). The impact of this scale mis-match will assessed quantitatively through the comparison with the results from the Meteorological Office's North Atlantic and European model (12km resolution). A number of UK weather events will be selected for this study covering both high-profile events (such as the Morpeth floods of September 2008) and more representative weather situations. The output of the NWP simulations will be used to simulate the expected satellite observations through radiative transfer models, which in turn will be compared with the actual satellite observations. The overall aim of this project is to investigate the scale-induced errors associated with assimilation of satellite radiances into numerical weather prediction models. The objectives are: 1. To characterise the errors in radiative transfer modelling arising from the computationally affordable treatment of fine scale structure in clouds and precipitation; 2. To improve our understanding of the information availability in different channels with varying resolutions; 3. To assess the different assumptions within the 1D analysis scheme for a range of meteorological conditions; and, 4. To evaluate the effects of beam-filling inherent in low-resolution satellite observations. The project will be based in the School of Geography, Earth and Environmental Sciences at the University of Birmingham, with supervision provided by Dr Chris Kidd and Dr Xiaoming Cai. The Satellite Radiance Assimilation Group of the UK Meteorological Office in Exeter will provide the necessary 'industry' support through the expertise of Stephen English. Both project partners will be supported by members of their research teams to provide the necessary research and training environment. Dr Chris Kidd will provide expertise in satellite observations and retievals supported by Dr Xiaoming Cai with expertise in atmospheric modelling. Dr Stephen English will provide expertise in microwave radiative transfer modelling, data assimilation and applications of infrared and microwave sounding data. Both organisations have complementary experience in satellite meteorology and data assimilation that can be exploited in the proposed studentship. The proposed research is important to both parties in that it addresses advancing the effective use of cloud and precipitation information from satellites observations in high resolution NWP. On-going work on the assimilation radar observations is essentially restricted to land-areas, consequently the proposed research will complement this by advancing our ability to assimilate available information over ocean areas.