Assessment of AMV Quality Using Cloud Information and Tracking Parameters
AMV quality is mostly assessed by looking at their background departures after filtering based on the supplied quality indicator values. Following this, selection for assimilation involves excluding AMVs over several geographic areas. A new possibility for selecting AMVs to assimilate is to consider how the quality varies with the properties of the tracked cloud features, or with parameters from the derivation algorithm.
In this talk I will show a quality assessment of AMVs based on filtering by the cloud type and tracking consistency information supplied by NESDIS with their GOES and VIIRS AMVs. The idea of using such information to accept or exclude AMVs, rather than geographic exclusions, will be considered. The possibility of assigning AMV observation errors using the new information will also be explored. I also show the result of filtering applied by EUMETSAT where some pixels are rejected from the AMV processing based on their cloud properties.
