Use of a lidar forward model to perform global cloud fraction comparisons between the ICESat lidar and the ECMWF model

There have been a number of studies of cloud representation in operational models using active ground-based instruments (radar and lidar), but these observations are only representative of a single point on the globe.

ICESat is a spacebourne lidar mission to remotely sense the height of the ice sheets and as a secondary objective is able to sense clouds on a global scale. However, as the laser signal passes through ice cloud it attenuates and in the case of very thick ice cloud is completely extinguished. Therefore, any comparison between operational model and ICESat must take account of the attenuation.

We present a solution to this problem by taking the model output of ice water content and liquid water content and using the model assumptions run a lidar forward model to predict the backscatter that would be received by ICESat, allowing a much fairer comparison between model and instrument.The results have been used to provide statistics on how well the ECMWF model cloud fraction performs as a function of latitude, height and surface type. Comparisons of model simulated backscatter with ICESat backscatter show that the modelled backscatter is too high, particularly at ice cloud altitudes, highlighting errors in the way radiation interacts with cloud within the model.