Clouds as Seen by Satellite Sounders (3I) and Imagers (ISCCP). Part II: A New Approach for Cloud Parameter Determination in the 3I Algorithms

First comparisons of improved initialization inversion (3I) cloud parameters determined from TIROS-N Operational Vertical Sounder observations with time–space-collocated clouds from the recently reprocessed International Satellite Cloud Climatology Project (ISCCP) dataset have shown a reasonable agreement between all cloud types, with exception of the stratocumulus regions off the western coasts. Here, 3I clouds were found systematically thinner and higher than ISCCP clouds. These results have initiated a careful investigation of the methods used to convert measurements from IR sounders into cloud parameters. All existing methods get very sensitive to the chosen temperature profile toward lower cloud heights, due to a denominator approaching zero. This leads to a bias like the one seen in the comparison with ISCCP. Therefore, a new 3I cloud scheme has been developed, based on a weighted-χ² method, which calculates the effective cloud amount from the CO₂-band radiances, but weighted differently according to the effect of the brightness temperature uncertainty within an air mass on these radiances at the various cloud levels. This physically much more correct method led to unbiased 3I cloud parameters for homogeneous cloud types. The ISCCP comparison agrees much better now, especially in the stratocumulus regions where the cloud type matching improved from about 50% to 75%. In 1° grid boxes covered uniformly with the same ISCCP cloud type, the matching reaches even 87%. Remaining discrepancies in cloud classification can be explained by partly cloudy fields and differences in temperature profiles and cloud detection. The weighted-χ² method can be used in other IR sounder inversion algorithms, if the empirical weights, taking care of the effect of temperature profile uncertainties on the difference between clear sky and cloudy radiances for different cloud levels and spectral channels, have been reevaluated so that they can be calculated automatically by the corresponding inversion algorithm.