河水平均通過時間和新水比例與地景特性的關聯

Abstract

河水平均通過時間 (MTT) 和新水比例 (Qe%) 可分別用以描述集水區中的水文途徑和逕流來源。然而，MTT和Qe% 的估算卻少被用來探討於副熱帶山區，尤其是極端降雨的事件尺度下。本研究應用轉換函數水文分離模式(TRANSEP) 來估算颱風降雨事件中台灣北部坪林集水區 (集水面積: 5-195 km2) 的MTT和Qe%，並測試TRANSEP模式中的四種機率分布模式 (TTD) (珈瑪分布、指數型分布、線性水庫分布、離散型分布)，結果發現 (1)由流量和δ18O追蹤劑的模擬結果中，指數型分布(EXP)的模擬表現分別可高達NSE:0.90和NSE:0.87，指數型分布(EXP)意味著此地區的水文作用主要是均勻混合的水流特性。(2) MTT結果大約是2.8至5.5小時且Qe%範圍在19%至29% 之間，表示著雨水轉換成為逕流的過程相當迅速。(3) 平均高程、集水區周長、平均坡度為控制台灣坪林集水區的MTT (R2: 0.99)的重要地景特性，但還未呼應到全球尺度的集水區，可能受到水流過程的複雜性所致。(4) 無論是地區或全球尺度下，Qe%的模擬結果可良好的被地景特性解釋(R2= 0.83)，並且再次的驗證地形起伏越大的地景特性下，較不利於事件水的產生。

Mean transit time (MTT) and event water ratio (Qe%) are determinants of biogeochemical cycle. However, MTT and Qe% in subtropical montane catchments are rarely discussed, particularly during typhoon invasions. In this study, the transfer function hydrograph separation model (TRANSEP) embedded with 4 TTDs (GM, EXP, TPLR, DM) was applied to estimate MTTs and Qe% in 5 catchments (5-195 km2) in Ping-Lin, Taiwan during a storm event. Results showed that (1) the EXP model outperformed other three TTDs, which can simulate runoff and δ18O tracer with high NSE of 0.90 and 0.87, indicating a well-mixed behavior during typhoon invasion; (2) the MTTs (2.8-5.5 hours) and the Qe% (19-29%) demonstrated quick conversion from rainfall to streamflow; (3) locally, the mean elevation, perimeter and mean slope can estimate the MTT well (R2=0.99), but not for global scale, probably due to the complexity of water transport dynamic; (4) both locally and globally, Qe% can be well estimated by landscape characteristics supporting the steep landscape is unfavorable with event water generation.