Recent improvements to LWP retrievals from AVHRR

In the framework of the Cloud Liquid Water Network (CLIWA-NET) project, an attempt is made to reconstruct the spatial distribution of liquid water path (LWP) by combining data from a network of ground-based microwave radiometers with imagery of the Advanced Very High Resolution Radiometer (AVHRR). In order to provide atmospheric modelers sufficient high-quality reference data for model evaluation and improvement, measurements from three intensive measurement campaigns are analyzed. This paper describes the AVHRR quantitative cloud analysis environment that was developed and tested at the Royal Netherlands Meteorological Institute (KNMI) during the project. The environment is named KLAROS (KNMI Local Implementation of APOLLO in an Operational System). KLAROS includes retrieval of a suite of physical cloud parameters. In order to evaluate the liquid water path retrieval, ground-based radiometer measurements are used. A new technique, called the Russian doll method, is introduced to assess the relation between ground and satellite measurements. This relation depends on the variability in time and space of the observed cloud fields and shows up clearly in the Russian doll diagrams. First, results from the comparisons give confidence that KLAROS gives plausible liquid water path values. We find that AVHHR data from the NOAA-16 platform is much better suited for quantitative cloud analysis than its predecessors. In particular, the availability of a 1.6-mum channel and the good calibration contribute to the improved accuracies. The retrieval method for liquid water path will be implemented in the framework of the Satellite Application Facility of Climate Monitoring (CM-SAF). (C) 2004 Published by Elsevier B.V.