摘要
In this paper, the Global Satellite Mapping of Precipitation Moving Vector with Kalman filter (GSMaP_MVK) was evaluated and corrected at daily time scales with a spatial resolution of 0.1°;latitude/longitude. The reference data came from thirty-four rain gauges on Kyushu Island, Japan. This study focused on the GSMaP_MVK’s ability to detect heavy rainfall patterns that may lead to flooding. Statistical analysis was used to evaluate the GSMaP_MVK data both quantitatively and qualitatively. The statistical analysis included the relative bias (B), the mean error (E), the Nash-Sutcliffe coefficient (CNS), the Root Mean Square Error (RMSE) and the correlation coefficient (r). In addition, Generalized Additive Models (GAMs) were used to conduct GSMaP_MVK data correction. The results of these analyses indicate that GSMaP_MVK data have lower values than observed data and may be significantly underestimated during heavy rainfall. By applying GAM to bias correction, GSMaP_MVK’s ability to detect heavy rainfall was improved. In addition, GAM for bias correction could effectively be applied for significant underestimates of GSMaP_ MVK (i.e., bias of more than 55%). GAM is a new approach to predict rainfall amount for flood and landslide monitoring of satellite base precipitation, especially in areas where rain gauge data are limited.
In this paper, the Global Satellite Mapping of Precipitation Moving Vector with Kalman filter (GSMaP_MVK) was evaluated and corrected at daily time scales with a spatial resolution of 0.1°;latitude/longitude. The reference data came from thirty-four rain gauges on Kyushu Island, Japan. This study focused on the GSMaP_MVK’s ability to detect heavy rainfall patterns that may lead to flooding. Statistical analysis was used to evaluate the GSMaP_MVK data both quantitatively and qualitatively. The statistical analysis included the relative bias (B), the mean error (E), the Nash-Sutcliffe coefficient (CNS), the Root Mean Square Error (RMSE) and the correlation coefficient (r). In addition, Generalized Additive Models (GAMs) were used to conduct GSMaP_MVK data correction. The results of these analyses indicate that GSMaP_MVK data have lower values than observed data and may be significantly underestimated during heavy rainfall. By applying GAM to bias correction, GSMaP_MVK’s ability to detect heavy rainfall was improved. In addition, GAM for bias correction could effectively be applied for significant underestimates of GSMaP_ MVK (i.e., bias of more than 55%). GAM is a new approach to predict rainfall amount for flood and landslide monitoring of satellite base precipitation, especially in areas where rain gauge data are limited.
作者
Martiwi Diah Setiawati
Fusanori Miura
Martiwi Diah Setiawati;Fusanori Miura(Integrated Research System for Sustainability Science (IR3S), The University of Tokyo, Bunkyo, Japan;Graduate School of Science and Engineering, Yamaguchi University, Ube, Japan)
