论文部分内容阅读
Typhoons are among the most devastating of natural disasters.Accurate inundation forecasts are necessary for flood warning and mitigation, and the development of real-time flood forecasting models has been recognized as an important task.Therefore, to improve inundation forecasts is expected to be useful in disaster mitigation.In this paper, an effective forecasting model is proposed to yield 1 to 6-h lead-time inundation maps for early warning systems during typhoons.The model is composed of two modules, point forecasting and spatial expansion.First, all 7-Eleven stores in the study area are collected and inundated and non-inundated areas are determined according to the inundation database.The 7-Eleven stores located in the inundated area are used as the inundation points.Second, the rainfall intensity, cumulative rainfall and inundation depths are employed as input for the construction of the point forecasting module to yield 1 to 6-h lead-time inundation forecasts at each inundation point.Third, according to the point forecasting results and geographic information, the relationship between each grid and inundation points is investigated and then the spatial expansion moduleis developed.The rainfall intensity, cumulative rainfall, geographic information (i.e.elevation and coordinates) of each grid and the point forecasts are used as input to the spatial expansion module to yield 1 to 6-h lead-time inundation maps.An application to Chiayi City in southern Taiwan(Figure 1) is conducted to demonstrate the superiority of the proposed model.For 1 to 3-h ahead forecasts, comparisons of the inundation data with the forecasted depths resulting from the point forecasting module of the proposed model are presented in Figure 2 for Typhoon Krosa.The forecasted results are significantly similar to the inundation data, which indicates that the forecasted inundation depth is in good agreement with the inundation data.Moreover, the superiority of the proposed model for the spatial expansion is also highlighted in depth in this paper.The comparison of the inundation data and corresponding forecasts resulting from the proposed model for Typhoon Krosa is presented in Figure 3.The maps of the forecasted results are similar to that of the inundation data for shortlead times.In addition, the forecasts are in good agreement with the inundation data, especially for the potential inundation area.Furthermore, the pattern of the forecasted maps is similar to that from the inundation data.The results clearly indicate that the proposed model provides reasonable spatial inundation forecasts.The proposed model is able to deal with the nonlinear relationships between inputs and desired output.We have shown that the proposed model is able to provide accurate point forecasts at each inundation point.Moreover, the spatial expansion module is capable of producing accurate spatial inundation forecasts.In conclusion, the proposed forecasting model is suitable and useful for improving the inundation forecasting during typhoons.In the future, actually observed inundation depths could be applied to the proposed model to forecast inundation maps if they could be obtained from the 7-Eleven stores.