Stimulated Effects of Irrigation on Salinity in the Arkansas River Valley in CO

Home Browse Search

aquifer are largest during seasonal high flows in otherwise rela- tively dry years. The temporal nature of the gain -loss relation is evident in the model simulations for the dry period from 1974 -78 (Figure 10). During wet years, more streamflow is available for diversion into the coral and the hydraulic gradient between the aquifer and the river increases because recharge increases due to larger surface water applications and canal leakage. Therefore, losses from the river to the aquifer are not as large during the seasonally high flows in relatively wet years. The temporal nature of the gain -loss relation is evident in the model simulations for a relatively wet period in 1982- 87 (Figure 10). A streamflow gauge at the downstream end of the study area would improve the understanding of stream- aquifer interac- tions, which might improve the model calibration. The streamflow gauge that was in place at the downstream end of the study site dur- ing the original 1971 -72 study (Konikow and Bredehoeft 1974a and 1974b) had been removed at the end of that study. Several potential sources of uncertainty in the model might account for errors in model simulations of water levels, salinity, and stream - aquifer interaction. In addition to errors inherent to the selection of model parameters, these sources include the estimations of canal flow, evapotranspiration, and ground water recharge. Dash (1994) estimated that discharge in the Fort Lyon Canal, which is measured using a 12 in Parshall flume, was accurate within 6% of actual discharge during freeflow conditions in the flume. During periods of backwater conditions in the flume, measured discharge was only accurate within 40% of the actual discharge. Backwater conditions at flows larger than 8.4 m /s typically occur during late spring and early summer (Dash 1994). Because applied surface water was estimated to equal 5.5 % of total canal diversions, there might be substantial error in these estimates during periods of backwater conditions. Estimates of evapotranspiration are another source of model error. Although the Blaney - Criddle method of evapotranspiration estimation is well accepted and widely used, it is a rather simplistic model of a complicated process and is diffi- cult to verify. The same is true for the determination of aquifer recharge. The amount of water recharged to the aquifer is a critical component of the model; however, it is difficult to verify because it varies in space and time and is a complex function of many physical and climatological variables. The model recharge parameter used in this study was estimated by comparing simulated and mea- sured ground water levels and salinity. Potential Impact of Water Use Changes As previously discussed, two water supply issues in the Arkancac River va11P�� rnrrPntl >>Pfllltere�t' (l�the p Atentlal yal- ley-wide decrease in ground water withdrawals due to stricter enforcement of state pumping regulations; and (2) the current and future decrease in irrigated acreage resulting from the transfer of water from agricultural to municipal use. Decreased ground water withdrawals or decreases in irrigated acreage might result in changes in ground water recharge, a decrease in the evapoconcen- tration of salts, and a change in the gain -loss relation between the stream and aquifer. The calibrated model was used to estimate the effects of these changes in the study area. Two categories of irri- gation management were simulated: (1) a decrease in ground water withdrawals for irrigation, and (2) the cessation of all irrigation from ground water and surface water sources. Individual simulation scenarios in the two management categories were performed with the three sub -areas (Figure 2) affected individually and collec- tively by the potential changes in irrigation practices. We assume that crop types, farming practices, and irrigation methodology remain essentially unchanged, as it is beyond the scope of this study to examine the effects of such potential changes. Decreased Ground Water Withdrawals for Irrigation During the 24 year study period, the average total applied irrigation water was 1.2 m/yr, of which about 0.5 m/yr was applied from ground water sources. In the modeled scenarios of decreased pumping, the total amount of applied water was decreased by decreasing ground water withdrawals. Surface water applications remained unchanged; they were not increased to offset the decrease in total applications. Three scenarios of decreased historical ground water with- drawals were simulated: (1) 25% decrease in pumping; (2) 50% decrease in pumping; and (3) 100% decrease in pumping. The simulated decrease in ground water salinity that resulted from decreased ground water withdrawals ranged from 0.5 to 6.9% (Table 2) and was related to the decrease in ground water with- drawals. The largest change in ground water salinity resulted from the complete cessation of pumping. In this scenario, the average monthly ground water salinity decreased from 2180 to 2030 mg/L. The relative difference between average ground water salinity for the 24 year baseline and the decreased pumping scenarios varied 82 Table 2 Model Results for Scenarios of Decreased Ground Water Pumpage for Irrigation Alluvial Aquifer Arkansas River Average Monthly Average Monthly Water Average Monthly Average Monthly Model Run Salinity (mg/L) Level (m above mean sea level) Salinity (mg/L) Streamflow Gains (m Base Condition (annual avg. of 1.6 of /acre pumped) 2180 1224.42 1810 0.176 Study Area -Wide Decreases in Ground Water Pumpage 25% Reduction (annual avg. of 1.2 of /acre pumped) 2170 (- 0.5 %) 1224.42 1810(0.0%) 0.176 (0.0 %) 50% Reduction (annual avg. of 0.8 of /acre pumped) 2160 ( -0.9 %) 1224.42 1810(0.0%) 0.179 (1.6 %) 100% Reduction (annual avg. of 0.0 of /acre pumped) 2030 (— 6.9 %) 1224.48 1800 (-0.6%) 0.196(11%) [ Numbers in parentheses indicate the percent difference between the modeled base condition and the reduced - pumpage scenarios. aquifer are largest during seasonal high flows in otherwise rela- tively dry years. The temporal nature of the gain -loss relation is evident in the model simulations for the dry period from 1974 -78 (Figure 10). During wet years, more streamflow is available for diversion into the coral and the hydraulic gradient between the aquifer and the river increases because recharge increases due to larger surface water applications and canal leakage. Therefore, losses from the river to the aquifer are not as large during the seasonally high flows in relatively wet years. The temporal nature of the gain -loss relation is evident in the model simulations for a relatively wet period in 1982- 87 (Figure 10). A streamflow gauge at the downstream end of the study area would improve the understanding of stream- aquifer interac- tions, which might improve the model calibration. The streamflow gauge that was in place at the downstream end of the study site dur- ing the original 1971 -72 study (Konikow and Bredehoeft 1974a and 1974b) had been removed at the end of that study. Several potential sources of uncertainty in the model might account for errors in model simulations of water levels, salinity, and stream - aquifer interaction. In addition to errors inherent to the selection of model parameters, these sources include the estimations of canal flow, evapotranspiration, and ground water recharge. Dash (1994) estimated that discharge in the Fort Lyon Canal, which is measured using a 12 in Parshall flume, was accurate within 6% of actual discharge during freeflow conditions in the flume. During periods of backwater conditions in the flume, measured discharge was only accurate within 40% of the actual discharge. Backwater conditions at flows larger than 8.4 m /s typically occur during late spring and early summer (Dash 1994). Because applied surface water was estimated to equal 5.5 % of total canal diversions, there might be substantial error in these estimates during periods of backwater conditions. Estimates of evapotranspiration are another source of model error. Although the Blaney - Criddle method of evapotranspiration estimation is well accepted and widely used, it is a rather simplistic model of a complicated process and is diffi- cult to verify. The same is true for the determination of aquifer recharge. The amount of water recharged to the aquifer is a critical component of the model; however, it is difficult to verify because it varies in space and time and is a complex function of many physical and climatological variables. The model recharge parameter used in this study was estimated by comparing simulated and mea- sured ground water levels and salinity. Potential Impact of Water Use Changes As previously discussed, two water supply issues in the Arkancac River va11P�� rnrrPntl >>Pfllltere�t' (l�the p Atentlal yal- ley-wide decrease in ground water withdrawals due to stricter enforcement of state pumping regulations; and (2) the current and future decrease in irrigated acreage resulting from the transfer of water from agricultural to municipal use. Decreased ground water withdrawals or decreases in irrigated acreage might result in changes in ground water recharge, a decrease in the evapoconcen- tration of salts, and a change in the gain -loss relation between the stream and aquifer. The calibrated model was used to estimate the effects of these changes in the study area. Two categories of irri- gation management were simulated: (1) a decrease in ground water withdrawals for irrigation, and (2) the cessation of all irrigation from ground water and surface water sources. Individual simulation scenarios in the two management categories were performed with the three sub -areas (Figure 2) affected individually and collec- tively by the potential changes in irrigation practices. We assume that crop types, farming practices, and irrigation methodology remain essentially unchanged, as it is beyond the scope of this study to examine the effects of such potential changes. Decreased Ground Water Withdrawals for Irrigation During the 24 year study period, the average total applied irrigation water was 1.2 m/yr, of which about 0.5 m/yr was applied from ground water sources. In the modeled scenarios of decreased pumping, the total amount of applied water was decreased by decreasing ground water withdrawals. Surface water applications remained unchanged; they were not increased to offset the decrease in total applications. Three scenarios of decreased historical ground water with- drawals were simulated: (1) 25% decrease in pumping; (2) 50% decrease in pumping; and (3) 100% decrease in pumping. The simulated decrease in ground water salinity that resulted from decreased ground water withdrawals ranged from 0.5 to 6.9% (Table 2) and was related to the decrease in ground water with- drawals. The largest change in ground water salinity resulted from the complete cessation of pumping. In this scenario, the average monthly ground water salinity decreased from 2180 to 2030 mg/L. The relative difference between average ground water salinity for the 24 year baseline and the decreased pumping scenarios varied 82