• 作者： 張領孝; 蔣志才; 蔣為民; 李南文
• 作者服務機構： 中央氣象局
• 中文摘要： 大甲溪流域上游集水區，位在中央山脈?山之中，山?標高均達三千公尺以上
  ，河川坡度1：16.7，溪流蜿蜓曲折，兩岸均為懸崖絕壁，地勢峻險，每在颱風大
  量降水，造成急勢逕流，而沿溪流建有德基、達見、青山、谷關、天輪等水壩與發
  電所，難以適度控制水位，若能預測颱風降水量，則有助於保障各水壩安全，增進
  水資源利用之效益，與減免下游居民之洪患。
  依據資料分析，颱風降水量之多寡與降水強度，均與颱風路徑有關，以颱風中
  心在宜蘭與花蓮間登陸者為最多，在南巴士海峽以南經過之較遠處颱風為最少。至
  於氣象預報因子在24個模式中，合計採用15項，其重要性如下：1.臺北（桃園）
  700mb之風速，被採用15次為最常用。2.馬公地面風速、臺中地面氣壓與氣溫，均
  被採用12次為次常用。3.馬公露點與颱風中心12小時位移，各被採用11次為再次常
  用。4.700mb風速、500mb風速與馬公氣溫，均被採用10次，其他均少於10次。
• 英文摘要： This investigation collected the dataof 77 typhoons invading Taiwan or thevicinity during the years from 1966 to1982. they can be divided into eighttracks of typhoon as figure 3. All themonthly precipitation amount of sevenstations along the Ta-Chia River havebeen calculated according to ThiessenMethod (6,8) as tablel 1. By the tablel,we found the most monthly averageprecipitation amount was in June and thenext was in August. But in Taiwan,June is still in raining season (Huang-Mei), and just 6 typhoons affected thisisland among 77 cases, therefore, themost precipitation in June was by thefront. The next amount of precipitationwas in August, it should be really aff-ected by approaching typhoon because21 typhoons influenced this island among77 cases. Analyzing the average arealprecipitation distribution along the Ta-Chia River, we found the rightsidestation A (fig. 2) was the most andthe central area of river watershed wasthe least. As for the hourly precipi-tation distribution, the location and theintensity of the peak rainfall dependon eight typhoon tracks are as figures4.5 & 6. It's helpful to estimate therainfall intensity and the location ofpeak rainfall during the typhoon appr-oaching. We chose sixty-seven predictorsand calculated correlation coefficientswith the predictants of Typhoon preci-pitation amount applying the RegressionMethods to develop 24 precipitationforecasting models, shown in table 6,for the 3 valid times of 12, 24, 48hours and the 8 Typhoon tracks. Thesemodels have been verified by F-Test &T-Test with the actual Typhoon preci-pitation data, and we compared theprecipitation of the forecasting modelsand the actual precipitation, the resultsare shown in table 7. If more Typhoondata were used to modify these objectiveforecasting models, these models shouldbe more acceptable and useful in preci-pitation amount prediction during theTyphoon period.
• 中文關鍵字： --
• 英文關鍵字： --