Analysis of 1997–2008 groundwater level changes in the upper Deschutes Basin, Central Oregon

Abstract

Groundwater-level monitoring in the upper Deschutes Basin of central Oregon from 1997 to 2008 shows water-level declines in some places that are larger than might be expected from climate variations alone, raising questions regarding the influence of groundwater pumping, canal lining (which decreases recharge), and other human influences. Between the mid-1990s and mid-2000s, water levels in the central part of the basin near Redmond steadily declined as much as 14 feet. Water levels in the Cascade Range, in contrast, rose more than 20 feet from the mid-1990s to about 2000, and then declined into the mid-2000s, with little or no net change. An existing U.S. Geological Survey regional groundwater-flow model was used to gain insights into groundwater-level changes from 1997 to 2008, and to determine the relative influence of climate, groundwater pumping, and irrigation canal lining on observed water-level trends. To utilize the model, input datasets had to be extended to include post-1997 changes in groundwater pumping, changes in recharge from precipitation, irrigation canal leakage, and deep percolation of applied irrigation water (also known as on-farm loss). Mean annual groundwater recharge from precipitation during the 1999–2008 period was 25 percent less than during the 1979–88 period because of drying climate conditions. This decrease in groundwater recharge is consistent with measured decreases in streamflow and discharge to springs. For example, the mean annual discharge of Fall River, which is a spring-fed stream, decreased 12 percent between the 1979–88 and 1999–2008 periods. Between the mid-1990s and late 2000s, groundwater pumping for public-supply and irrigation uses increased from about 32,500 to 52,000 acre-feet per year, partially because of population growth. Between 1997 and 2008, the rate of recharge from leaking irrigation canals decreased by about 58,000 acre-feet per year as a result of lining and piping of canals. Decreases in recharge from on-farm losses over the past decade were relatively small, approaching an estimated 1,000 acre-feet per year by the late 2000s. All these changes in the hydrologic budget contributed to declines in groundwater levels. Groundwater flow model simulations indicate that climate variations have the largest influence on groundwater levels throughout the upper Deschutes Basin, and that impacts from pumping and canal lining also contribute but are largely restricted to the central part of the basin that extends north from near Benham Falls to Lower Bridge, and east from Sisters to the community of Powell Butte. Outside of this central area, the water-level response from changes in pumping and irrigation canal leakage cannot be discerned from the larger response to climate-driven changes in recharge. Within this central area, where measured water-level declines have generally ranged from about 5 to 14 feet since the mid-1990s, climate variations are still the dominant factor influencing groundwater levels, accounting for approximately 60–70 percent of the measured declines. Post-1994 increases in groundwater pumping account for about 20–30 percent of the measured declines in the central part of the basin, depending on location, and decreases in recharge due to canal lining account for about 10 percent of the measured declines. Decreases in recharge from on-farm losses were simulated, but the effects were negligible compared to climate influences, groundwater pumping, and the effects of canal lining and piping. Observation well data and model simulation results indicate that water levels in the Cascade Range rose and declined tens of feet in response to wet and dry climate cycles over the past two decades. Water levels in the central part of the basin, in contrast, steadily declined during the same period, with the rate of decline lessening during wet periods. This difference is because the water-level response from recharge is damped as water moves (diffuses) from the principal recharge area in the Cascade Range to discharge points along the main stems of the Deschutes, Crooked, and Metolius Rivers in the central part of the basin. Water levels in the central part of the basin respond more to multi-decadal climate trends than shorter term changes. Groundwater-flow simulations show that the effects from increased pumping and decreased irrigation canal leakage extend south into the Bend area. However, the only wells presently monitored in the Bend area are heavily influenced by the Deschutes River, which dampens any response of water levels to external stresses such as groundwater pumping, changes in canal leakage, or climate variations.