| 摘要: | 受全球氣候變遷影響下，水文因子不確定性提高，其中地下水占可用水資源中約30%，故水資源的調度更顯重要。在台灣水資源的需求以農業用水為最高，故本研究從農業灌區著手，以桃園大圳第十支線作為研究區域，探討該地區的地下水水位變化。埤塘灌溉系統為桃園地區內特有的農業灌溉方式，埤塘作用不僅能對於田區供水灌溉，更能在降雨時提供雨水蓄存的空間。為了探討此區域地下水水位受埤塘灌溉系統作用影響，本研究運用小波分析，以灌區上游(頭洲)及灌區下游(東明)兩區域地下水觀測站資料，分析降雨事件及長天期無降雨事件對兩區域地下水水位造成的影響。利用系統動力模式建立埤塘灌溉系統，運用實測之大圳引水量及降雨量進行模擬，模擬該區域田區的供水狀況，以2015年及2017年間的一期作與二期作為模擬時間，觀察乾旱年(2015)及豐水年(2017)埤塘操作下之田間湛水深對地下水水位的變化影響。
根據結果顯示，上游地區(頭洲)地下水因降雨事件隨之上升，無降雨而呈現洩降趨勢；下游地區(東明)地下水也受降雨事件上升，但在無降雨事件中洩降趨勢較緩，推論為下游地區受降雨停留在灌溉田區的湛水深，持續提供地下水補注，地下水水位的變化因為灌溉而相對不顯著。
;Due to climate change impact, the uncertainty of water resources system is getting worse, which enhance the importance of water resources planning and management on regional and global scale. In Taoyuan, a system dynamic model was used to simulate a unique agricultural irrigation method called pond irrigation system. This irrigation method was not only to assist the irrigation water supply, but also used for water storage during rainfall events. In this study, two observation stations, upstream and downstream of the irrigation area, were compared and used as examples to discuss the rise in groundwater level caused by rainfall, and the decrease in long-term non-rainfall period. Then simulation is performed for the pond irrigation system with a system dynamic model by utilized the measurement of water intake and rainfall. The changes of field water depth and the groundwater level were compared during the irrigation operation for two different crop period in 2015 and 2017.
According to the simulation results, the groundwater level in the upstream area rose rapidly during rainfall events, and fell sharply when rainfall ceased. On the other hand, the groundwater level in the downstream area experienced little impact of irrigation water due to that irrigation maintained stable field water supply. In the agricultural field, groundwater level is affected by rainfall kept in the field, which continues to supply groundwater recharge. Based on the theory of the recharge of the groundwater level in the agricultural field, a regression equation is introduced by utilized the data collected in 2014 to 2017, which relating the groundwater level rose with the cumulative residual water in the site. As a result, when the rainfall is small, the next day groundwater level would be approximate to the current observation, due to minor changes in the groundwater level. However, in the heavy rainfall event, although there may be some estimation overrated, the simulation results effectively predict the changes in the groundwater level. |
