It is difficult to gauge the accuracy of the estimates from satellite measurements particulary when no similar dataset exists. The next closest thing to a true verification is a comparison from output of numerical models that incorporate observations into the model system. This is done in the figure to the right where model data was convoluted to simulate the satellite estimate in the vertical. That is, the height of the model value is determined by simulating the satellite measured brightness temperature using the model thermodynamic profile as a surrogate atmosphere. Next the height of the brightness temperature is assumed the mean layer height of both the winds and humidity.

   By clicking on the figure to the right, time series of three other parameters are shown. It is readily apparent that the models tend to agree best with the satellite data on the zonal (or east-west) component of the wind. Trends are captured quite well by both data sets; particularly the Boreal spring 1988 transition to the cold phase of the Southern Oscillation. Diagreement is also observed particularly in the meridional (or north-south) component of the wind. This is an area of continuing research as the satellite estimates may be superior to model estimates particularly over the southern oceans where conventional measurements aren't available. However, with the satellite-derived winds and humidity, certain regions (e.g., very dry clear conditions or high convective clouds) present challenges in not only obtaining a water vapor wind vector, but determining at which height the wind is representative of.

Last updated on: November 2, 1999