圖書典藏及影音數位平台

本署已陸續完成各地區製作大量且高精度地形數值之淹水潛勢資料製作與更新，並以空間不均勻分佈之不同降雨重現時距組合產製龐大之淹水潛勢歷程資料，為延伸淹水潛勢資料之應用範圍，以提升淹水潛勢資料在災害防救計畫之參考價值，本計畫規劃研發方法與技術建置區域型內水淹水預警與推估模式，並滿足區域型淹水之線上即時預測，提供洪災防救應變之重要參考。

With high resolution requirements, the simulations of flood inundation extent need a huge amount of data and computing time by using a series of numerical models. It is really difficult to provide an on-line simulation and forecast of the inundation extent. For extending the applications of the flood inundation maps, this project proposed a methodology of building a regional flood inundation forecasting system based on the flood inundation maps and provided an on-line forecasting system during the period of typhoons and storms. For extending the applications of the flood inundation maps, this project integrated the artificial neural networks, data base, geographic information system, and web interface to establish a regional flood inundation warning system for on-line prediction of regional flood depths in various villages and towns of Yilan County during typhoon or storm events. This project proposed a clustering-based hybrid inundation model to establish a regional flood inundation forecasting system in Yilan County. The modeling process is divided into the data pre-process stage and model construction stage. In data pre-process, the inundation regions were selected and clustered into a number of groups by the K-mean clustering method, then further divided into a total of 53 sub-groups based on the geographic relation and flooded characteristics, and selected the control point and representative point for each sub-classification. In modeling stage, the backpropagation neural networks are built for control points and representative points. The networks are trained and validated by using the inundation maps generated from the designed rainfall patters and typhoon events. The proposed regional inundation forecast system would only take 90 seconds to make one- to three-hours flood forecast for a 24-hour rainfall pattern, while the conventional two-dimensional numerical model would take at least 12 hours to generate 24-hour hydrograph of flood depths. The efficiency has been greatly improved and the proposed system is ready for on-line forecast in the whole region of Yilan County. The on-line regional inundation forecast system can be easily executed through the database to access the data sets and to update the parameters and information obtained from model outputs simultaneously.

本研究針對宜蘭縣，開發區域型內水淹水警示技術與自動化即時預報系統。為克服傳統二維水理模式耗時、難以即時預警的限制，計畫利用淹水潛勢資料，結合 K-Means 聚類分析將淹水特性分群，並建立倒傳遞類神經網路預測模式。 該模式經訓練後，能將 24 小時降雨事件的淹水預測時間從傳統模式的 12 小時大幅縮短至 1.5 分鐘。最終建置的自動化系統，可每小時穩定輸出宜蘭縣各鄉鎮村里未來 1-3 小時的最大淹水深度與位置，並整合 Google Map API，有效協助防災單位即時掌握區域淹水情況，提升災害應變效率。