Assimilating visible satellite images for convective scale numerical weather prediction
Observations from instruments on geostationary satellites provide a wealth of information about convective activity and are therefore seen as a important type of observation for convective scale data assimilation (DA). In particular the solar channels contain information on the cloud distribution, cloud microphysical properties and cloud structure with high temporal and spatial resolution. However, due to a lack of suitable forward operators for the solar spectral range, where multiple scattering makes radiative transfer (RT) very expensive, currently only thermal channels are directly assimilated at operational centers. Recently, we have developed a look-up table based RT method for the visible spectrum that is orders of magnitude faster than conventional approaches. A forward operator based on this RT method has been implemented to simulate synthetic MSG-SEVIRI images from COSMO-DE model output and is now integrated in the km-scale Ensemble Data Assimilation (KENDA) system of DWD. We demonstrate that assimilating 0.6mu SEVIRI observations improves the cloud distribution in forecasts for several hours without degrading the beneficial impact of conventional observations significantly. Moreover, the assimilation of these observations can also lead to the correction of errors in the precipitation field.
Dr. Leonhard Scheck Hans-Ertel Centre for Weather Research / German Weather Service