台灣地區本身季節性之降雨就有差異明顯，受全球性的氣候變遷、極端氣候之影響，引發極端乾旱或洪澇水文事件，使得水資源時空分布及氣候不確定的問題更為嚴重，農業用水為水資源用水為最大宗，缺水容忍度較民生及工業兩者為高，面對水資源短缺及供需風險提升，灌溉期間經灌區流出後未進入水道前之農業排放水具有相當之可利用潛勢，如果透過管理手段，對於水資源增加將可以有相當可觀的幫助。 本研究以桃園灌區及二支線為研究區域，以1990年為分界點來探討氣候變遷條件下前後時段雨量的變化，結果顯示1990年以後無論是降雨量或是降雨天數都是呈現減少的趨勢，平均降雨量減少157.27mm，降雨天數減少約9天，兩者減少百分比約為8.6%。透過Mann-Kendall趨勢檢定結果顯示，乾季旬降雨量有減少的趨勢，除間接增加乾季大圳灌溉用水供應需求風險外，也將影響一期作灌溉的決策；分界年前後降雨趨勢在第一期稻作灌溉期間，旬降雨量由增加趨勢轉為減少趨勢，可以看出氣候變遷確實造成氣象變化的轉折，然而比較全期及1990年之後的旬降雨趨勢分析結果，全期資料認為一期作可以穩定可靠供水，但1990年之後之資料則認為降雨減少趨勢增加，因此面對氣候變遷情況下，資料區間的選取造成數據呈現不同結果，在決策時可能會有低估的風險。 中央氣象局之季長期氣象預報可提供未來三個月之降雨機率的預測，其結果顯示，統計台北站與桃園灌區歷史雨量資料，雨量在三個命中區間內的命中率皆有0.61以上，顯示台北站的季長期天氣預報資料可以套用在桃園灌區使用，然而針對台北站之預報能力分析，顯示未來三個月的雨量預報仍不穩定。 本研究區域以實施輪流灌溉、並配合埤池調蓄方式為供水操作基礎，配合中央氣象局長期降雨預報之條件機率，以氣象合成模式(WGEN)合成未來雨量資料，透過系統動力模式(VENSIM)模擬桃園灌區期作之灌溉情形，透過案例模擬可以發現，在各個水源正常供水的情況下，二支線灌區並不會發生缺水的狀況，下游(C區)一、二期作總計出流量分別0.071CMS及0.042CMS，中游(B區)一、二期作總計出流量分別0.025CMS及0.013CMS，顯示區域仍有相當的流量釋出，透過水資源管理的手段仍有相當大的節水空間。當供水嚴苛發生缺水時，均落在自幼穗分化期至抽穗開花期結束的灌溉用水尖峰時段，總降雨量的多寡並不一定是缺水事件發生的必要條件，與降雨的時空分布有很大的關係，灌區缺水事件對應的累積雨量越高，表示其可能發生缺水事件的機率也越大；當氣候條件更為嚴峻，風險高灌區除應加強預警外，也可提早研擬因應對策。因此，本研究認為透過案例分析可了解灌溉管理策略對農業生產的影響，及農田及水利相關單位在面對用水調配時可能之風險與水量參考。 ;The seasonal rainfall is distinct difference in Taiwan. Under the effect of global climate change, the extreme flood or drought events showed the problem of water resource space-time distribution and climate uncertainty become more and more serious. Agriculture water is the largest demand of water resource and it has better shortage toleration than industry and household demand. Due to the water resource shortage and demand risk increasing, the usage of return flow amount in irrigation area during irrigation period has huge potential with management method. Taoyuan Channel #2 Feeder was selected as the study area. This research discussed the influence of rainfall under climate change, and analyzed rainfall by the Mann-Kendall trend test for the trends in total 50 years, year 1963-1990 and that in 1991-2012. The results showed the trends of rainfall amount and day were both decreased 8.6%. Average rainfall and day decreased 157.27mm and 9 days. The trend of 10-day rainfall was decreased in dry season. It was not only indirect increased the risk of Taoyuan Channel irrigation water supply or demand in dry season, but influenced the crop cultivation irrigation arrangement for the first crop period at next year. The trend of 10-day rainfall during the first crop irrigation period by the increase became reduce between year 1963-1990 and year 1991-2012. The weather change caused by climate change was indeed a turning point. However, compared with the trend of whole year and year 1991-2012, the whole year data predicted the rainfall will apply a reliable water supply during the first crop period, but year 1991-2012’s data prediction indicated a decrease trend of rainfall. Different time interval of data showed distinct results. That caused policy decision may underestimate risk under climate change. The Central Weather Bureau (CWB) issued the seasonal climate forecast every month. The CWB seasonal climate forecast provide for precipitation and temperature as percentage likelihood with lead time of 3 months in 1-month moving windows. To gather statistics of Taipei station and Taoyuan irrigation area historical rainfall data, the rainfall shooting average in three shooting sections (higher/ normal/ lower) are all above 0.61. The result showed that the seasonal climate forecast of Taipei station can also apply to Taoyuan irrigation area. However, the forecast ability analyze of Taipei station for precipitation as percentage likelihood with lead time of 3 months was still unstable. The irrigation supply operation was included rotation and pond irrigation system in study area. The conditional probability of the Central Weather Bureau (CWB) seasonal climate forecast was involved to Weather Generator for future rainfall data. The system dynamic model (Vensim) was used to establish the irrigation system of Taoyuan study area. The simulation case showed the #2 Feeder water shortage situation will not happen under normal supply of various water conditions. The total outflow of the first and second crop irrigation periods were 0.071CMS and 0.042CMS in the downstream, 0.025CMS and 0.013CMS in the midstream. It showed a considerable flow display released in this area. There was still great water saving space through resources management process. When water supply condition was severe, the shortage always occurred in the panicle initiation to heading stage, the peak of irrigation water period. The more or less of total rainfall was not the necessary condition of water shortage events. It was related to the time and space distribution of rainfall. The higher irrigation water accumulated rainfall corresponds to the water shortage events indicated the possible occurrence probability greater. When weather conditions were more severe, the high risk irrigation area should strengthen early warning and making corresponding strategy.
