本研究之試驗灌區為彰化縣溪州鄉三條圳幹線灌區位於臺灣中部濁水溪沖積扇上游，灌溉面積約209公頃，灌區內之灌溉水源除了雨水及渠道灌溉供水外，農民也會抽取地下水進行灌溉。本研究利用2015年及2016年一期作水文資料，如降雨、蒸發量、作物生長時程等，透過渠道供水量優先之模式以水旱作面積比0.7:0.3情形下進行灌溉用水模擬，模擬結果顯示2015年一期作試驗灌區之平均地下水抽取量為757.75mm，佔總流入量(降雨量、渠道取水量及地下水抽取量之總和)之35.91%。以輪區而言，地下水抽取量由第一輪區逐漸增加，於第五輪區超過50%的總入流量。結果顯示當渠道灌溉供水不足時，自第三輪區起地下水成為灌溉之主要水源；2016年一期作之地下水抽取量為118.3mm，佔總流入量之8.12%，其主因為2016年之降雨分布較為平均且渠道灌溉供水充足供應至下游輪區，故對地下水之依賴程度較2015年一期作低。然2015年一期作期間試驗灌區觀測資料顯示入流量僅為出流量之50%，表示此模式於估算地下水抽取量處於較低估之狀態。 假設水稻田在無降雨天皆進行抽水灌溉的情況下，利用2015年一期作(2015年3月1日至2015年6月30日)之實測入流值及水文資料進行模擬。結果顯示當水稻田於無降雨天逕行抽0.5小時之地下水灌溉時，模擬之一期作總出流量與實測出流量相近。以此模式模擬結果推估，實際上本試驗灌區為產水區產出之水量為7.39百萬公噸，提供灌溉餘水予下游灌區使用。 ;In this study, a system dynamic model (VENSIM) was applied to establish an irrigation water management model for mixed cropping system in Central Taiwan. The experimental site of 209 ha located at the upstream of Choushui river consists of paddy rice and upland mixed crops. There is a conjective use of channel water and groundwater for irrigation purpose in addition with rainwater. The cropping area consist of 70% paddy rice and 30% upland crops. VENSIM model was simulated for the first cropping season of years 2015 and 2016. The input data comprised of hydrological data (such as rainfall, evaporation, etc.) channel water, automatic groundwater pumping rate and crop period. The simulation results of 2015 for the 1st crop season indicated that the average groundwater pumping depth was 757.75mm (35.91%) of total inflow (2110.19mm). The contribution of groundwater gradually increases from the first rotation zone and reach upto one-half (50%) of the total inflow depth in the fifth rotation zone. These results presented the dependency on groundwater usage as major source of irrigation for 3rd to 5th zones, due to insufficient channel water. The depth of groundwater pumping was less in 2016 as compared to 2015 due to more uniform distribution of rainfall. Average groundwater simulated depth was 118.3mm (8.12%) of the total inflow due to sufficient channel water supply and rainfall distribution. However, the observed data of the experimental irrigation area during the first phase of 2015 showed that the inflow depth was only 50% of the outflow depth, indicating that model is in a state of underestimation of groundwater pumping.
The system dynamic model was also used as test scenario for groundwater pumping in paddy field for March 1, 2015 to June 30, 2015 depends on rainfall occurance. When there is a no rainfall (P = 0mm), the test scenario results showed that the paddy field firstly irrigated with groundwater under 0.5hour pumping, as a result the simulated total outflow depth almost as same as measured outflow depth. The overall efficiency of model simulation indicated that the study area can be used as a water yield area, and the excess irrigation water can be supplied to downstream of the study area.
