83rd Annual

Wednesday, 12 February 2003
Wind and sea surface pressure fields from the SeaWinds scatterometer and their impacts in NWP
Shannon R. Davis, Florida State Univ., Tallahassee, FL; and M. A. Bourassa, R. M. Atlas, J. Ardizzone, E. Brin, D. Bungato, and J. J. O'Brien
Poster PDF (201.7 kB)
The impacts of the assimilation of scatterometer-derived pressures and wind into an atmospheric model are investigated. Spaceborne scatterometers have already demonstrated a positive impact on model forecasts through the assimilation of ocean surface vector wind measurements. The present NASA scatterometer, Seawinds aboard the QuikSCAT satellite provides near-realtime data, which are suitable for forecasts and assimilation into numerical models in a timely operational sense. The focus here is a comparison of the impact of assimilation of surface pressure fields derived from the QuikSCAT wind data to the impacts of assimilating surface winds. Validation of the scatterometer-derived pressures is demonstrated with co-located in situ observations taken by research vessels and buoys. For the purposes of forecast impact assessment, the winds and QuikSCAT-derived pressures are separately assimilated at the NASA Data Assimilation Office. The NWP package is the NASA/NCAR GCM, with the CORE data assimilation system (DAS), and the Physical-space Statistical Analysis System (PSAS). Separate model runs are compared using assimilated QuikSCAT pressures, assimilated QuikSCAT winds, and a control run without QuikSCAT data. Analyses and 5 day forecasts are examined. The globally averaged differences between the analyses (winds, pressures, and control) are small; however, in areas of strong pressure gradients (e.g., fronts and low pressure systems) the differences can be very large. The difference between wind and pressure based forecasts are smaller than seen in the comparison of analyses. The impacts if winds and pressures on forecasts are similar, and the average impact of the wind assimilation is slightly better than the pressure assimilation.

Supplementary URL: