dc.description.abstract | Typhoons usually pose a serious threat to the livelihood of people. Therefore, making accurate rainfall predictions has been a key focus in weather forecasting. Kidder et al. (2005) introduced a method called the Tropical Rainfall Potential Technique (TRaP) to predict the accumulated rainfall. However, the method didn’t consider the cyclone intensity variations and rotation, which may cause significant errors when estimating the rainfall potential. The intensity decay model (Kaplan et al., 1995) is used in this study to estimate the landing cyclone intensity variations. The different landing typhoon intensity, seasons, and intensity plus seasons, were categorized and tested in this model, the results show that the consideration of intensity had the best result.
Based on the TRaP method, the study considered the cyclone rotations (Liu et al., 2008) and intensity variations in estimating the rainfall potential in the southern China coastal area. Due to the limitation of the rain gauge temporal resolution, only 6 and 24 hours accumulated rainfall were analyzed in this study. For the 6 hrs rainfall potential, results revealed that the correlation was 0.74 when only the shift-motion was considered. Yet, the correlation rose to 0.78 when the tropical cyclones’ intensity and rotation changes were also taken into account. In terms of the 24 hrs rainfall potential prediction, the correlation increased from 0.51 to 0.60 and the RMSE decreased from 37.1 to 13.8 mm/24hr when intensity variations and rotation change were factored in. In general, the results showed that the additional considerations in this study could notably improve the TRaP method accuracy.
| en_US |