Radiance-based retrieval of total water vapor content from sentinel-3A OLCI NIR channels using ground-based GPS measurements

We calibrated the integrated water vapor (IWV) data retrieved from near-infrared (NIR) channels of the Ocean and Land Color Instrument (OLCI) onboard the Sentinel-3A satellite using in-situ GPS-sensed IWV observations. Unlike conventional water vapor retrieval methodologies relying upon radiative transfer code, this method utilized a regression equation to empirically estimate GPS IWV from two NIR absorption channels at 900 nm and 940 nm. The GPS IWV data were used as reference to define the relationship between the measured radiance ratio and IWV. We collected IWV data from June 1, 2016 to May 31, 2019 from 453 GPS stations situated in the inland and coastal areas of Australia. The retrieval approach was analyzed by using different sample sizes and training datasets. The evaluation results between June 1, 2019 and May 31, 2020 in Australia indicated that the algorithm could reduce the root-mean-square error (RMSE) of the operational OLCI IWV products by 12.91% from 3.114 to 2.712 mm at the O19 channel, by 10.69% to 2.781 mm at the O20 channel, and by 11.75% to 2.748 mm for the weighted mean IWV when compared with GPS reference IWV data. When compared to European Centre for Medium-Range Weather Forecasts reference IWV data, the RMSE was reduced by 12.94% from 3.154 to 2.745 mm, by 11.04% to 2.805 mm, and by 11.93% to 2.777 mm, at the O19 channel, O20 channel, and the weighted mean, respectively. The spatiotemporal performance of the OLCI IWV measurements was improved in both station-scale and daily-scale after applying the new empirical regression retrieval method. The seasonal and land-surface-type dependence of the retrieval approach was also discussed in this research.