Improvement of WBGT estimation method from MTSAT by combining bio-geophysical parameters

摘要

The thermal environment in living spaces has been worse because of global warming and urban heat islands, and the risk of heat disorders has also increased recently. In order to prevent heat disorders, Wet-Bulb Globe Temperature (WBGT) were used as the index of thermal environment condition. In Japan, Ministry of the Environment offers the hourly WBGT on the web site, which is calculated from meteorological data at each meteorological observation station. Therefore, the WBGT data is not spatially continuous data. In this study, we developed the method for estimating WBGT using remote sensing data by combining bio-geophysical parameters in order to obtain continuous spatial distribution of WBGT in Japan. This method used the Multi-functional Transport Satellite-2 (MTSAT-2) data, clear-sky solar radiation data, normalized difference vegetation index data, and elevation data as the variables of WBGT estimate formula. The estimation accuracy was evaluated by RMSE, an overall accuracy, and a kappa coefficient. The overall accuracy and kappa coefficients were calculated using confusion matrix when the observed and estimated WBGTs were classified into several warning levels of heat disorders based on WBGT criteria of heat disorder prevention guidelines. The RMSE was approximately 2.7 degree Celsius, the overall accuracy was over 64%, and kappa coefficient was 0.49. Thus, it was revealed that it was difficult to use estimated WBGT values directly for heat disorder prevention, but the maps of heat disorder warning levels could be created using MTSAT-2 data with reasonable accuracy.