AMV Retrieval Using Brightness Temperature Difference Image

Brightness temperature difference (BTD) between 11µm channel and 12µm channel (BTD11-12=TBB11-TBB12) is a good indicator of total precipitable water over cloud free ocean. This BTD11-12 is also effective in identifying optically thin cirrus cloud that consists of ice crystal, and optically thick high/low cloud. Utilizing the BTD11-12, one of the authors developed a simple cloud type classification algorithm that can be classified cirrus type cloud and low/middle/high-level cumulus type cloud. Cloud types classified by the algorithm show reasonable agreement with CALIOP observations in cloud top height. Thus the BTD11-12 image represents cloud height (low/middle/high) over cloud region and spatial distribution of moisture over cloud free ocean. Himawari-8 carries 10 infrared channels including 11 and 12µm. We studied the feasibility and effectiveness of AMV retrieval using the BTD11-12 image observed by Himawari-8. In this study, combination of four channels of 8.6, 10, 11 and 12µm are also studied by constructing BTD image. The advantage of using the BTD image is seamless retrieval of AMV both day and night time. Further BTD enhances the characteristics of cloud. For example, the BTD between 8.6 and 12 µm enhances the low-level cloud against surrounding ocean area. This clear contrast of low-level cloud against the ocean is effective in retrieval of low-level cloud motion comparing with low contrast of brightness temperature between low-level cloud and surrounding ocean in single infrared image. The BTD pattern appeared over cloud free ocean can be tracked as well. The movement of the BTD pattern is consistent with the movement of near-by low-level cloud. Thus AMV can be retrieved over cloud free ocean area, too. Height of this AMV can be assigned as low-level (say 850 hPa), since water vapor is generally concentrated at low-level of the atmosphere by downdraft associated with high pressure system. Therefore, we propose the use of BTD image in AMV retrieval.