石門水庫集水區之降雨特性對崩塌及輸砂量的關係

Abstract

本研究主要是利用中央氣象局之劇烈天氣監測系統 (QPESUMS) 所收集的雨量資料，並配合現地撞擊式雨滴譜儀 (JWD) 的觀測資料，來探討2005年瑪莎颱風及2008年辛樂克颱風期間，在石門水庫集水區內的降雨特性相對崩塌及輸砂量的影響關係。研究結果顯示，當降雨強度超過40 mm/hr時，本研究區之降雨動能會趨近於 28.73 J/m2 mm。而白石溪子集水區在兩次颱風事件中，所累積之降雨動能皆超過17000 J/m2，為所有子集水區中分佈最高者，其輸砂量佔了全區的45%以上。另外，從福衛二號衛星影像判釋發現，研究區內的崩塌率，在兩次颱風事件中分別為1.03%及0.96%，沒有明顯的差異，而區域內的澳底層、乾溝層及大桶山層，之崩塌面積佔了全區的80%以上，其中又以大桶山層的33%為最高，這與其岩石強度57.1 MPa，以及單位體積節理數約19.5條/m3有相當之關連性。2005年瑪莎颱風崩塌的新生率74.7%，及重現率31.9%，比2008年辛樂克颱風的66.1%及24.6%高，這可能與瑪莎颱風於44小時內的累積雨量632.5 mm有關。 從兩次颱風事件中的降雨動能與輸砂量對比發現，當降雨動能累積超過2000 J/m2時，河水中的輸砂量會有明顯上升的趨勢，這可能是因為當降雨動能累積到此門檻值時，舊有的崩塌地會開始再次的活動，或是新的崩塌地開始發生，使得這些崩塌材料被大量地表逕流沖刷至河道中。另外，從雙偏極化雷達所估算之降雨量及降雨動能，雖然較劇烈天氣監測系統與現地撞擊式雨滴譜儀觀測資料小，但其有關空間解析度之雨滴粒徑分布估算能力，則有助於了解降雨量及降雨動能的分布對崩塌及輸砂量之影響。

Rainfall properties had been mentioned as an important factor to trigger landslides of watershed in many literatures. In this study, I try to use the rainfall data of the QPSUMS (Quantitative Precipitation Estimation and Segregation Using Multiple Sensor) from Central Weather Bureau and JWD (Joss-Waldvogel Disdrometer) data to figure out the relationship between rainfall, landslides and sediment discharge during the typhoon Matsa and typhoon Sinlaku. Investigation results indicate that the rainfall kinetic energy is approach to the 29 J/m2-mm when the rainfall intensity exceeds 40 mm/hr. During these events, the highest kinetic energy of accumulated rainfall was reached to 17,000 J/m2 around the Siouluan sub-watershed. The sediment discharge of the Bai-shi sub-watershed was more than 2.6 Mt, and occupied around 45% in total of the study area. The landslide ratio of the study area is 1.03% in Matsa typhoon event and 0.96% in Sinlaku typhoon event by using the Formosa II images judgments. The various formations of Aoti, Kankou and Datunshang have more than 80% landslide distribution in total study areas, and the Datunshang Formation have highest distribution 33%. The results on the above mentioned have links with rock strength 57.1 MPa and volumetric joint 19.5. The new generation ratio 74.7% and reactivated ratio 31.9% in typhoon Matsa are higher than typhoon Sinlaku in 66.1% and 24.6%. It seems to have relative connection with accumulated rainfall 632.5 mm in 44 hours during typhoon Matsa. Compared with rainfall kinetic energy and sediment discharge in these events, the sediment discharge increased obviously when the accumulated rainfall kinetic energy exceeds 2000 J/m2. This result show that the accumulated rainfall kinetic energy of 2000 J/m2 is seems a thresholds value in triggering landslide. It means the previous landslides were reactive or new landslides occur when the accumulated rainfall kinetic energy exceeds this thresholds value. And these materials were be flush out by the massive runoff. In this study, I also use the Polarmetric Dopplar Radar to estimate the rainfall rate and kinetic energy. Although the results from Polarmetric Dopplar Radar are smaller than QPESUMS couples with JWD estimation, the high resolution in spaces of the rainfall rate and kinetic energy are very helpful to understanding the rainfall distribution affect the landslide and sediment yields.