Traditional methods for deriving wind vectors from sequential geostationary satellite imagery involve the tracking of coherent clouds and moisture features in single channels (spectral bands). While this data source has proven to be important to global wind analyses, the approach is limited in two major ways: 1) The heights assigned to the vectors are not precise, leading to problems in data assimilation, and 2) Vertical profiles of the wind at a given geo-location are not provided, adding further stress to objective data assimilation (difficulty with single-level observations).

A new approach to deriving winds from sequential satellite observations is introduced in this paper. The method utilizes the same basic automated tracking code developed at CIMSS, however the input to the algorithm is in the form of constant-level moisture analyses derived from hyperspectral sounding information. Since the altitude of the features being tracked are already determined by the soundings/analyses, the height assignment ambiguities associated with the traditional approaches are ameliorated. Furthermore, the hyperspectral infrared (IR) information provides detailed vertical profiles of moisture where there are no clouds. This allows analyses of moisture at multiple vertical levels, which can then be used to attempt winds tracking, and potential vertical profiles of wind.

To date, the new scheme has been trialed on simulated data from GIFTS, and on one case of real data from airborne observations provided by the NAST-I instrument. From these first attempts, the “proof of concept” is successfully illustrated, and will be shown at the conference.