Evaluation of Marine Surface Wind Speed Observations From AMSR2 on GCOM-W Satellite

• Published in: IEEE journal of selected topics in applied earth observations and remote sensing Vol. 10; no. 9; pp. 3955 - 3962
• Main Author: Ebuchi, Naoto
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.09.2017; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Advanced microwave scanning radiometer 2 (AMSR2); Bias; Buoys; Economic forecasting; Errors; Evaluation; global change observation mission-water satellite (GCOM-W); marine surface wind; microwave radiometer; Microwave radiometry; Ocean temperature; Offshore; Radiometers; Random errors; Satellite observation; Satellites; Scatterometers; Sea measurements; Sea surface; Surface wind; Tornadoes; validation; Weather forecasting; Wind data; Wind direction; Wind speed
• ISSN: 1939-1404; 2151-1535
• DOI: 10.1109/JSTARS.2017.2685432

Observations of marine surface scalar wind speeds from the advanced microwave scanning radiometer 2 (AMSR2), onboard the global change observation mission-water satellite (GCOM-W), were evaluated by comparisons with offshore moored buoy measurements, output from the European Centre for Medium Range Weather Forecast Reanalysis Interim, and observations by the RapidScat (RSCAT) scatterometer onboard the International Space Station. In general, the AMSR2 wind speeds agreed well with the reference data. The root mean square difference between the AMSR2 and buoy measurements was 1.09 ms -1 , which is slightly larger than the mission goal of 1 ms -1 . Underestimation at low wind speeds (<;5 ms -1 ) was found in the comparisons. The AMSR2 wind speeds were found to contain a slight scan bias; namely, wind speeds in the right swath are higher than those in the left swath by 0.2-0.3 ms -1 . Systematic dependence of the wind speed bias on the wind direction relative to the AMSR2 looking direction was found in a residual analysis. Results of the triple collocation analysis suggest that the random errors in the AMSR2 wind speed are less than 1 ms -1 and are smaller than those in the outputs from the numerical weather prediction (NWP) models, if random errors in the reference wind data (buoy, NWP, and RSCAT) are considered explicitly.