Remote sensing of inhomogeneous clouds with MSG/SEVIRI

Abstract

Clouds are inherently inhomogeneous on all spatial and temporal scales. The neglect of this fact in standard remote sensing techniques introduces uncertainties in form of biases and noise. The quantification of deviations caused by cloud inhomogeneity and possible solutions are of major concern, considering the potential of remote sensing for a global characterisation of clouds and their interaction with the radiation field. We found that, on average, reliable cloud properties are retrieved by a standard method from MSG/SEVIRI data for overcast pixels. However, larger systematic errors occur for partly cloudy pixels. For individual pixels the uncertainties can be considerable. The high resolution visible (HRV) channel of SEVIRI has the capability to improve the normal/low resolution situation, because it provides some sub-pixel inhomogeneity information. In a EUMETSAT case study on the quantitative exploitation of the HRV in synergy with low resolution spectral channels we studied possible improvements for cloud detection and microphysical property retrievals. The accuracy of the derived optical thickness and effective radius were significantly improved when HRV sub-pixel information was included. Future work is planned on more general cloud situations derived from numerical weather model data.