Two-Look Polarimetric (2LP) Microwave Radiometers for Ocean Vector Wind Retrieval

Friday, 19 December 2014
Frank J Wentz1, Kyle A Hilburn1, Thomas Meissner1 and Shannon E Brown2, (1)Remote Sensing Systems, Santa Rosa, CA, United States, (2)NASA Jet Propulsion Laboratory, Pasadena, CA, United States
This talk discusses the future utilization of two-look polarimetric (2LP) microwave radiometers for measuring the ocean surface wind vector. Potentially, these 2LP satellite radiometers offer two advantages over conventional scatterometers: unambiguous wind vector retrievals and low-cost. One concept for a 2LP radiometer is being developed by JPL and is called the Compact Ocean Wind Vector Radiometer (COWVR). A space demonstration of COWVR is planned for 2016 timeframe.

To explore the potential of 2LP radiometers, we use the 11 years of WindSat observations as a testbed. We only use that portion of the WindSat swath that has both fore and aft observations. WindSat provides fully polarimetric observations (all four Stokes parameters) at 11, 19, and 37 GHz. This represents 12 independent channels for each of the two azimuth directions. A wind vector retrieval algorithm is developed to fully utilize this wide assortment of information. Since this analysis is based on actual observations, it provides a realistic picture of what to expect from future 2LP radiometers. To our knowledge, this is the first time that the combination of WindSat’s fore and aft observations has been fully utilized for wind vector retrievals.

 In our talk we compare the 2LP wind vector retrieval performance with that of single-look polarimetric radiometers (i.e., WindSat standard product) and scatterometers. We provide basic statistics from a triple collocation between winds from WindSat, QuikScat, and NDBC/PMEL ocean moored buoys. The statistics include the standard deviation of the first ranked ambiguity direction, skill rate, and number of ambiguities. All available data from the common period of operation between WindSat and QuikScat (2003-2009) are used. We characterize the wind direction accuracy as a function of wind speed, and show how 2LP retrievals are able to extend the wind vector accuracy to lower wind speeds than previously considered possible for radiometers.