Measurement of Ocean Surface Wind by AMSR2 on GCOM-W1

The Japan Aerospace Exploration Agency (JAXA) launched the first satellite of Global Change Observation Mission–W series (GCOM-W1) in May 2012. The GCOM-W1 carried the Advanced Microwave Scanning Radiometer 2 (AMSR2), which is a successor of AMSR-E on Aqua and AMSR on ADEOS-II. The science data have been continuously obtained by AMSR2 since July 2012. As the standard data product, integrated water vapor, liquid cloud water, marine surface wind speed, sea surface temperature (SST), precipitation, sea ice concentration, soil moisture and snow depths has been provided together with other research and experimental data products. In this paper, the marine surface scalar wind speed, which is one of the essential physical parameters in the air-sea boundary processes, were evaluated and characteristics of the measurement error is discussed.

Wind speed globally observed by AMSR2 were compared with in situ data from offshore moored buoys, vector wind data from RapidScat (RSCAT) on International Space Station (ISS), and outputs from the ERA-interim reanalysis. Several versions of AMSR2 wind data products from JAXA and the Remote Sensing Systems (RSS) were utilized in the comparisons.

Data from offshore moored buoys operated by the National Data Buoy Center (NDBC), Tropical Atmosphere and Ocean (TAO), Prediction and Research Moored Array in the Atlantic (PIRATA), and the Research Moored Array for African–Asian–Australian Monsoon Analysis and Prediction (RAMA) projects were utilized in the comparison. Spatial and temporal separations between the AMSR2 and buoy observations were limited less than 12.5 km and 5 min. Height and stability corrections were made to the buoy data using a method proposed by Liu and Tang (1996). Wind speed contained in the JAXA's standard product (ver. 2.1) showed good agreement with the buoy data. The bias and root-mean-squared difference were -0.22 m/s and 1.09 m/s, respectively. The wind data retrieved and distributed by the RSS showed better agreement. In the JAXA's wind, a trend of slight underestimation was discernible in low wind speed range.

Comparisons with ISS/RSCAT wind data and outputs from the ERA-interim reanalysis confirmed these results. Triple collocation analysis (Stoffelen, 1998) using the collocation among AMSR2-buoy-ERA and AMSR2-RSCAT-ERA showed that the substantial random error in AMSR2 wind is estimated to be approximately 1.0 m/s, while the buoy and RSCAT indicated slightly lower values. Results of these analyses exhibited that the JAXA's latest version of AMSR2 wind is of quality for scientific studies concerning the air-sea boundary processes. Spatial resolution of 25 km, swath of 1400 km, and simultaneous observation with other parameters (SST, water vapor…) may also contribute to those studies. The data are available for science users from the JAXA's website (https://gcom-w1.jaxa.jp/auth.html).