Application of SMAP Data for Observations of Extreme Ocean Surface Winds

The NASA Soil Moisture Active Passive (SMAP) mission was launched January 31st, 2015. It is designed to measure the soil moisture over land using a combined active / passive L-band system. SMAP’s low-frequency L-band radiometry allows for estimation of higher ocean surface wind speeds than possible with higher-frequency microwave radiometers and scatterometers, enabling important and detailed features of wind speeds in tropical storms, hurricanes and typhoons to be quantified and mapped.

The heritage of SMAP wind retrieval algorithms, showing that SMAP and Aquarius show excellent agreement in the ocean surface roughness correction, is described. Then an overview of some newly developed algorithms, including a new galaxy correction and land correction enabling wind retrievals up to 40 km from coast, is provided. We discuss recent improvements to the SMAP data processing for version 4.0.

We have validated the SMAP high-wind products using collocations with the NOAA Stepped Frequency Microwave Radiometer (SFMR) winds acquired during tropical cyclones as well as triple-collocations with RapidScat and WindSat. We consider two validation regimes, storm force winds and hurricane force winds. For storm force winds we validate using other space-borne scatterometers and microwave radiometers as well as with SFMR, however, for hurricane force winds we must use SFMR. Comparison of SMAP and SFMR (averaged to SMAP spatial scale) winds in 2015/2016 (27 storms with 184 SFMR tracks) indicates a Root-Mean-Square-Difference (RMSD) of about 3 m/s, ~0.83 correlation and a regression slope of 1.01. We find that the SMAP-SFMR RMSD has no obvious dependence on rain rate, a significant advantage over high frequency (>10 GHz) satellite scatterometer and radiometer for ocean surface wind mapping. Validation of the SMAP winds for gale and storm force winds is performed by comparison with the RapidSCAT winds for rain-free conditions identified by WindSat. There were about 13 million matchups of SMAP, RapidScat and Windsat within 90 minutes of collocation under rain-free conditions identified by WindSat. We find very small speed bias (<0.5 m/s) up to 40 m/s as compared to Rapidscat. The agreement between SMAP and RapidSCAT is ~1.73 m/s in RMSD, 0.9 correlation, and a regression slope of 0.83. The results of validation analysis support the applicability of SMAP data for observations of severe storm.