Sea-Level Pressure Retrieval from SAR Images of Tropical Cyclones

Abstract

Validation and calibration of wind vector retrievals from synthetic aperture radar images of tropical cyclones remains a serious challenge. The basic wind vector measurements in tropical cyclones come from the approximately ten to twenty drop sonde profiles that are obtained during reconnaissance flights. In the highly turbulent tropical cyclone boundary layer, any given drop sonde profile represents a single realization of the virtual ensemble that must be averaged to estimate a mean surface wind vector. This is the quantity of interest because geophysical model functions are calibrated in terms of mean surface wind speeds. Even if a mean surface wind vector can be estimated from a single sonde profile, it can only be compared to the nearest wind retrieval. Furthermore, in the high wind regime geophysical model functions can be extremely sensitive to small errors in either backscatter measurements or assumed wind direction. Drop sondes do provide very accurate atmospheric pressure data, which has the beneficial property of being a scalar mean flow quantity. A method for calculating surface pressure fields from synthetic aperture radar images of tropical cyclones is presented. These fields are very accurate, with an RMS error of about 3 mb. Importantly, the surface pressure field represents an integral of the full wind vector field. Hence, comparing the pair-wise (between drop sonde splash locations) aircraft-and satellite-derived pressure differences provides a means by which the overall quality of the wind vector retrievals can be assessed. Finally, wind vector fields derived from the surface pressure fields can provide significantly improved estimates in regions of the image where objective quality flags reject the raw winds.