| 摘要: | 近年來受西北太平洋副熱帶高壓增強的影響，台灣乾旱發生的機率變得頻繁。為了解決因乾旱而造成的經濟與社會損失，合理並永續利用相對穩定的地下水資源成為重要議題。從觀測資料與過去的文獻中發現兩點，一是地下水補充會因含水層的飽和而減緩，二是台灣周邊存在海底地下水排放的現象，造成淡水資源的浪費。因此本研究運用系統動態模型模擬地下水側流的日資料，與建立貝氏網路模型探討研究區域的地下水管理方法，有效發現過去常被人們忽視的地下水管理問題。
有別於以往文獻，僅模擬地下水的季尺度或月尺度資料，本研究利用基於水筒模型建立的系統動態模型，模擬出地下水側流的日資料，並利用2014年至2022年的平均地下蓄水量水文歷線圖驗證地下水側流日資料的合理性。最後將地下水側流的模擬值與其他觀測資料輸入進貝氏網路模型中，結果顯示當地下水位越高時，地下水的側向淨流入量越少，證明了地下水補充確實會因為含水層的飽和而逐漸減緩。
另外，在貝氏網路的結果中，假設在乾季與中等降雨的情況下，地下水保持在中等水位，與高等及低等相比，會有較低的流出機率和較高流入機率。而假設在雨季與中等降雨的情況下，地下水保持在低等水位，會使得地下水補充有更好的表現。因此，本研究建議在雨季來臨前將地下水保持在低等水位，讓雨季時的降雨可以更好地補充到地下水，而在雨季結束時將地下水保持在中等水位。這種方法不同於以往雨季時先使用地表水，直到乾季時才尋求地下水的水資源管理方式。本研究建議在雨季來臨時先使用地下水資源，直到乾季地下水達到中等水位時，再使用地表水，讓淡水資源得到更好地利用。
;In recent years, the frequency of droughts in Taiwan has increased due to the strengthening of the subtropical high pressure over the Northwest Pacific. To address the economic and social losses caused by droughts, the rational and sustainable use of relatively stable groundwater resources has become an important issue. From observation data and past literature, two points have been identified: first, groundwater recharge slows down due to the saturation of aquifers, and second, there is a phenomenon of submarine groundwater discharge around Taiwan, leading to the wastage of freshwater resources. Therefore, this study employs a system dynamics model to simulate daily data of groundwater lateral flow and establishes a Bayesian Network model to explore groundwater management methods in the study area, focusing on aspects often overlooked in past groundwater management.
Unlike previous studies that only simulated seasonal or monthly scale groundwater data, this study uses a System Dynamics model based on the tank model to simulate daily data of groundwater lateral flow. The reasonableness of the daily groundwater lateral flow data is verified using the observation average groundwater storage hydrograph from 2014 to 2022. Finally, the simulated values of groundwater lateral flow and other observation data are input into the Bayesian Network model. The results show that the higher the groundwater level, the less the net inflow of lateral groundwater, confirming that groundwater recharge indeed slows down due to the saturation of aquifers.
Additionally, in the results of the Bayesian network, assuming the groundwater level is maintained at a medium level during the dry season with moderate rainfall, compared to a low or high level, has a lower probability of outflow and a higher probability of inflow. Conversely, assuming the groundwater level is maintained at a low level during the rainy season with moderate rainfall, groundwater recharge performs better. Therefore, this study suggests maintaining a low groundwater level before the rainy season arrives to maximize groundwater recharge during the rainy season. After the rainy season ends, the groundwater level should be maintained at a medium level. In other words, this approach differs from the traditional water resource management method, which uses surface water first during the rainy season and then seeks groundwater during the dry season. This study recommends using groundwater resources first when the rainy season arrives. Once the groundwater reaches a moderate level during the dry season, surface water should then be utilized to ensure better use of freshwater resources. |
