WindSat passive microwave ocean wind and precipitation retrieval in hurricanes

Ocean surface wind retrieval from spaceborne passive microwave observations is a mature capability for wind speeds of 25 m/s or less and for clear to lightly precipitating conditions. A more robust retrieval algorithm has been developed which relaxes both of these limitations so that retrievals are possible in tropical cyclone conditions. The algorithm is comprised of an atmospheric radiative transfer model which includes both emission and scattering in the rain and ice columns, differentiating between stratiform and convective precipitation as needed. By properly accounting for atmospheric effects, the underlying ocean surface microwave emission can be resolved even in high precipitation conditions. Sensitivity of the microwave emissivity to wind is demonstrated for wind speeds of up to 60 m/s (Category 4 cyclone).

The atmospheric radiative transfer model uses Eddington's second approximation for propagation through the ice and rain layers and full Mie expressions for absorption and scattering assuming spherical hydrometeors. A Sekhon and Srivastava drop size distribution is assumed, with number density adjusted as a free parameter to determine the precipitation rate. The surface emission model is derived empirically using a two stage process. In the first stage, column integrated water vapor, liquid water content and ice water path are estimated using brightness temperature observations at higher frequency, vertically polarized channels which are more sensitive to atmospheric effects and less sensitive to the dependence of surface emissivity on near surface wind speed. In the second stage, the optical depth and atmospheric upwelling and downwelling brightness are predicted at all observing frequencies (including the lowest frequencies where surface sensitivity is largest) so that those measurements can be properly corrected and the surface emissivity can be derived.

The retrieval algorithm has been developed for use with the WindSat radiometer channel set. It has been validated using WindSat overpasses of several tropical cyclones for which NOAA-HRD wind fields are available using their H*WIND interpolator of aircraft and ground truth observations. Surface wind retrievals are reliably produced in moderate to heavy precipitating conditions. The precipitation classification portion of the algorithm is able to differentiate between stratiform and convective cases based on a chi-squared residual error metric that is compared for two radiative transfer models that assume either potential type of precipitation. Rain rate retrievals, derived as part of the atmospheric clearing algorithm, are shown to properly resolve the eye, eye wall, and rain band features of the hurricane.