P 4.51 ANALYSIS OF UNCERTAINTIES IN SEVIRI CLOUD PROPERTY RETRIEVALS FOR CLIMATE MONITORING Robert A. Roebeling, N. A. J. Schutgens and A. J. Feijt 1) Royal Netherlands Meteorological Institute, De Bilt, The Netherlands 1. INTRODUCTION The representation of clouds in climate models is one of the biggest sources of uncertainty in climate studies. To reduce these uncertainties satellites can provide valuable information on cloud properties and their spatial and temporal variation. Satellite observed spectral radiances can be used to retrieve cloud properties because the radiative behaviour of clouds is related to properties such as: thermodynamic phase, optical thickness and droplet effective radius. The accuracy of satellite retrievals is estimated through comparison with ground based observations. This comparison encounters various uncertainties that are not related to retrieval errors, but to the spatial and temporal variability of clouds. In order to assess the accuracy of cloud property retrievals from satellite these uncertainties need to be quantified. Various methods have been developed to retrieve Cloud Optical Thickness (COT), cloud particle size and Cloud Liquid Water Path (CLWP) from satellite radiances (Nakajima and Nakajima, 1995, Han et al., 1994 and Watts et al., 1998). The principle of these methods is that the reflection of clouds at the non-absorbing visible channel (0.6 or 0.8 μm) is primarily a function of the cloud optical thickness, while the reflection at a water (or ice) absorbing near-infrared channel (1.6 or 3.7 μm) is primarily a function of cloud particle size. Within the Climate Satellite Application Facility (CM-SAF) of the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT) Royal Netherlands Meteorological Institute (KNMI) (Feijt et al. 2004 and Roebeling et al. 2006) developed an algorithm to retrieve COT and CLWP from visible (0.6μm) and near-infrared (1.6μm) reflectances of the Spinning Enhanced Visible and Infrared Imager (SEVIRI) onboard the Meteosat Second Generation (MSG). The SEVIRI- CLWP retrievals are compared with ground based microwave radiometer LWP (MW-LWP) values to estimate the retrieval accuracy. Although the comparison aims to estimate the accuracy of the SEVIRI-CLWP retrievals, it should be realized that part of the differences observed between ground based and SEVIRI retrievals of LWP result from uncertainties due to: • Microwave radiometer derived LWP accuracy, • Co-location of satellite and ground observations, The co-location error may be split into several independent contributions • Ground site not coinciding with SEVIRI pixel centre, • Different Field Of Views (FOV) for ground- and satellite observation, • Incorrect cloud location for the satellite due to the parallax effect. Note that the accuracy of MW-LWP retrievals is well known to vary between 20 and 30 g m -2 (Dong et al., 2000 and Crewell and Löhnert , 2003). While all other uncertainties show larger variations, and reduce to zero for extended homogeneous clouds. The impact of cloud in-homogeneity is therefore an important consideration for the present study. This paper aims to quantify the differences in CLWP retrievals due to the uncertainties related to comparing satellite and ground based observations. To estimate the expected differences between ground based and satellite retrieved LWP values one month of SEVIRI-CLWP and MW-LWP retrievals are compared. MODIS AQUA data are used to simulate ground based and SEVIRI LWP fields, which are then compared to quantify the differences due to above described uncertainties. We will discuss how we plan to use these results to develop an improved sampling method for comparing ground based observed and satellite retrieved cloud properties. The outline of the paper is as follows. The methods used for the uncertainty analysis are presented in section 2. In section 3, the study procedure is described. The results are presented in section 4. Finally, in section 5, the results are summarized and conclusions are drawn. 2. METHODS In this section we present the methods that are used for the analysis of uncertainties in the comparison between ground based and SEVIRI retrievals of LWP. The Cloud Physical Properties algorithm (CPP) is based on reflectances at visible (0.6 μm) and near-infrared (1.6 μm) wavelengths. The COT and particle size are retrieved for cloudy pixels in an iterative manner, by simultaneously comparing satellite observed reflectances at visible and near- infrared wavelengths to Look Up Tables (LUTs) of simulated reflectances for given optical thicknesses and particle sizes (Roebeling et al., 2006). The CLWP is computed from the retrieved cloud optical thickness at wavelength and droplet effective radius.