[AAS04-P01] Validation of GOSAT cloud judgement by using Himawari-8 data
Keywords:GOSAT, Himawari-8, cloud judgement, validation
In the retrieval processing of greenhouse gases from the Thermal and Near Infrared Sensor for Carbon Observation (TANSO)−Fourier Transform Spectrometer (FTS) on board Greenhouse Gases Observing Satellite (GOSAT), cloud contamination in the field of views (FOVs) of TANSO−FTS has been judged by TANSO−Cloud and Aerosol Imager (CAI) in the daytime and by the thermal infrared (TIR) band of TANSO−FTS in the nighttime. This study has compared the cloud detections in the TANSO−FTS FOVs by TANSO−CAI or TANSO−FTS TIR band with cloud detections by the Advanced Himawari Imager (AHI) on board Himawari−8.
For each of all TANSO−FTS observations over the ocean in January and July in 2016, we first selected all coincident Himawari−8 data when their center positions were located in the TANSO−FTS FOVs and they were obtained within one minute before or after the TANSO−FTS observations. Next, we conducted cloud detection tests for each of the selected Himawari−8 data based on their reflectance and brightness temperature values, and then judged each of TANSO−FTS FOVs as a clear or cloudy scene by combining the cloud detection tests of the Himawari−8 data in the TANSO−FTS FOVs.
About 90% of cloud judgements by TANSO−CAI and Himawari−8 agreed with each other. TANSO−CAI used only reflectance data in visible bands for their cloud detections, whereas Himawari−8 used both reflectance and brightness temperature data in visible and infrared bands, respectively, which may cause their ~10% disagreements. In contrast, agreements in cloud judgements between TANSO−FTS TIR band and Himawari−8 got worse by 13−20% compared to the TANSO−CAI and Himawari−8 cloud judgements. This is because that the FOVs of TANSO−FTS (10.5 km) is larger than those of TANSO−CAI and Himawari−8 (0.5 km and 2 km), and cloud detections by TANSO−FTS TIR band used only brightness temperature data.
In addition, GOSAT cloud judgements (TANSO−CAI or TANSO−FTS TIR band) tended to judge TANSO−FTS FOVs as “clear”, whereas Himawari−8 judged the same TANSO−FTS FOVs as “cloud” in most cases of disagreements in their cloud judgements. This is partly because of “clear conservativeness” of our cloud detection algorithm by Himawari−8. As for TANSO−FTS FOVs judged as “clear” by TANSO−FTS TIR band and “cloud” by Himawari−8, small-scale clouds partly existed in the TANSO−FTS FOVs judging from relatively low brightness temperature values of the coincident Himawari−8 data. This suggests the possibility that TANSO−FTS TIR band cannot detect such small-scale clouds.
For each of all TANSO−FTS observations over the ocean in January and July in 2016, we first selected all coincident Himawari−8 data when their center positions were located in the TANSO−FTS FOVs and they were obtained within one minute before or after the TANSO−FTS observations. Next, we conducted cloud detection tests for each of the selected Himawari−8 data based on their reflectance and brightness temperature values, and then judged each of TANSO−FTS FOVs as a clear or cloudy scene by combining the cloud detection tests of the Himawari−8 data in the TANSO−FTS FOVs.
About 90% of cloud judgements by TANSO−CAI and Himawari−8 agreed with each other. TANSO−CAI used only reflectance data in visible bands for their cloud detections, whereas Himawari−8 used both reflectance and brightness temperature data in visible and infrared bands, respectively, which may cause their ~10% disagreements. In contrast, agreements in cloud judgements between TANSO−FTS TIR band and Himawari−8 got worse by 13−20% compared to the TANSO−CAI and Himawari−8 cloud judgements. This is because that the FOVs of TANSO−FTS (10.5 km) is larger than those of TANSO−CAI and Himawari−8 (0.5 km and 2 km), and cloud detections by TANSO−FTS TIR band used only brightness temperature data.
In addition, GOSAT cloud judgements (TANSO−CAI or TANSO−FTS TIR band) tended to judge TANSO−FTS FOVs as “clear”, whereas Himawari−8 judged the same TANSO−FTS FOVs as “cloud” in most cases of disagreements in their cloud judgements. This is partly because of “clear conservativeness” of our cloud detection algorithm by Himawari−8. As for TANSO−FTS FOVs judged as “clear” by TANSO−FTS TIR band and “cloud” by Himawari−8, small-scale clouds partly existed in the TANSO−FTS FOVs judging from relatively low brightness temperature values of the coincident Himawari−8 data. This suggests the possibility that TANSO−FTS TIR band cannot detect such small-scale clouds.