單森華

　　(四創科技有限公司，福建 福州 350100)

　　摘 要：為減輕臺風災害風險，輔助開展臺風風險防控，基于1974—2019年臺風路徑、風力和降雨等數據對我國臺風危險性進行分析。將我國劃分為多個地理網格，采用分布式數據庫Hbase構建網格、臺風和影響分值列簇，基于網格對歷史臺風路徑及風力、降雨再分析數據進行預處理,建立臺風影響范圍擬合和影響力衰減等函數，利用分布式計算引擎MapReduce計算臺風對各網格造成的影響力分值，并計算不同年份、年段臺風危險性。結果表明：1)本方法數據依賴性較低，計算效率較高，評估結果更客觀，推廣性較強;2)研究過程中采用歸一化法對危險性進行分級，導致不同時間段評估結果出現較大差異，須進一步研究改進;3)我國臺風危險區總體呈現為極高、高、中、低4條危險帶，臺風危險性呈現增強趨勢，并持續向東北、中西部延伸，未來防臺風形勢依然嚴峻。

　　關鍵詞：臺風危險性;評估方法;空間網格;風力;降雨 ;風雨數據

　　Research and implementation of typhoon risk assessment method

　　based on wind and rain data

　　SHAN Senhua

　　(IStrong Technology Co.,Ltd., Fuzhou 350100, China)

　　Abstract：In order to reduce typhoon disaster risk and assist in typhoon risk control management, the typhoon risk analysis in China is conducted based on the data of typhoon path, wind force and rainfall from 1974 to 2019. China is divided into several geospatial grids, and the grid, typhoon and impact score column clusters are constructed by using the distributed database HBase. The historical typhoon path, wind force and rainfall reanalysis data are preprocessed based on the grid. Then the typhoon influence range fitting function and influence attenuation function are established. The influence score of typhoon on each geospatial grid is calculated by distributed computing engine MapReduce.. which calculates the typhoon risk of different years and periods as well. The results show that :1) this method has lower data dependence, higher computational efficiency, more objective assessment results and strong popularization; 2) the normalization method is used to classify the risk, which leads to great differences in the assessment results in different periods, and needs further research and improvement; 3) the typhoon risk areas in China generally present four risk zones: extremely high, high, medium and low. The risk of typhoon in China is increasing and extending to the northeast and western regions. The situation of typhoon risk prevention is still serious in the future.

　　Key words：typhoon risk;assessment method;spatial grid;wind force;rainfall;wind and rain data