Introduction


The upper Klamath Basin occupies a broad volcanic plateau in the transition between the Cascade Range and the Basin and Range geologic provinces in south-central Oregon and northern California (fig. 1). Although the region is high desert, the basin receives considerable recharge from the Cascade Range on the western margin and the volcanic uplands on the eastern margin. As a result, the permeable volcanic bedrock in the basin contains a robust regional groundwater system that contributes much of the flow to the streams and lakes that serve as the principal water supply for irrigated agriculture and habitat for endangered freshwater and anadromous fishes (Gannett and others, 2007; National Marine Fisheries Service and U.S. Fish and Wildlife Service, 2013). 


Water use and water needs in the upper Klamath Basin are closely related to agriculture, which is a significant component of the basin economy. The agricultural economy provides about 10 percent of the jobs in the upper Klamath Basin and about 15 percent of the economic output (National Research Council, 2004). More than 500,000 acres in the basin are irrigated, about 190,000 acres of which are within the Bureau of Reclamation Klamath Irrigation Project (Project) (Burt and Freeman, 2003; National Resources Conservation Service, 2004). The primary source of irrigation water for the Project is Upper Klamath Lake and the primary crops grown on Project land are alfalfa, irrigated pasture, and grain. Between 2005 and 2009, crops grown on Project land provided an average annual revenue of $148.6 million
(U.S. Department of the Interior and others, 2012). 


Prior to 2001, annual diversions by the Project from Upper Klamath Lake and the Klamath River were as much as about 490,000 acre-ft (McFarland and others, 2005). In 2001, water-management priorities in the basin shifted to protect aquatic habitat. A series of Endangered Species Act biological opinions now require the Project to limit surface-water diversions to protect habitat for Lost River and shortnose suckers and coho salmon (National Marine Fisheries Service and U.S. Fish and Wildlife Service, 2013). This realignment of water supply and demand has resulted in reductions in surface water diverted for agriculture and increased demand for groundwater, particularly in drought years. 


In order to balance the benefits of water for aquatic wildlife and agriculture, stakeholders in the basin developed the Klamath Basin Restoration Agreement (KBRA; 2010), a framework for water-resources management jointly developed by Federal, state, and county agencies, Indian tribes, irrigation communities, and conservation and fishing groups. The primary goals of the KBRA are to restore historical fish habitat and populations and establish reliable water supplies for irrigation. To achieve these goals, the KBRA includes elements related to water diversion and use that substantially change water-resources-management practices in the basin. Among those is the Water Resources Program (Klamath Basin Restoration Agreement, 2010, section 15), which establishes a permanent limit on the amount of water that can be diverted to the Project from Upper Klamath Lake and the Klamath River. A comparison of historical diversion amounts to the maximum diversion amounts permitted by the KBRA indicates the Project could experience irrigation deficits of about 100,000 acre-ft in the driest years.


The shift in water-management priorities is expected to create a sustained demand for groundwater to supplement reduced surface-water diversions to the Project. There is concern that the effects of increased groundwater use by the Project could spread beyond the Project pumping centers to other parts of the basin, with the possibility for adverse effects on groundwater users outside the Project (off-Project water users) and surface-water resources that are critical to aquatic wildlife. The competing demands on the groundwater resources of the upper Klamath Basin are creating a need for improved techniques to better understand and manage the groundwater development in the area. 


Purpose and Scope


Sound management of the water resources of the upper Klamath Basin requires understanding the complex and interconnected groundwater and surface-water systems of the region. The U.S. Geological Survey (USGS) completed studies to characterize the regional groundwater system in the basin (Gannett and others, 2007) and develop groundwater-simulation and management models of the region (Gannett and others, 2012). In order to understand the broad and long-term effects of groundwater-development decisions, the USGS, in cooperation with the Klamath Water and Power Agency (KWAPA) and the Oregon Water Resources Department (OWRD), began a follow-up study to develop an improved framework for testing and designing groundwater-management strategies for the Project. The improved framework uses techniques of groundwater simulation and optimization to enhance the utility of the simulation model by directly incorporating management goals and constraints into the modeling process. Of particular interest was the ability of the groundwater system to meet the increased groundwater demand of the Project foreseen under the KBRA while avoiding the adverse effects of groundwater development that concern resource-management agencies. This report provides an overview of the groundwater development issues in the upper Klamath Basin, describes the development of the combined simulation and optimization groundwater-management model, and provides applications of the model to evaluate alternative groundwater-management strategies. The report is intended to provide a basic understanding of the techniques of groundwater-management models and help water users and resource-management agencies understand the capacity for, and limitations of, supplemental groundwater use by the Project. 


Description of Study Area


The upper Klamath Basin comprises the entire drainage basin upstream of Iron Gate Dam, including the internally drained Lost River and Butte Valley subbasins, and encompasses about 8,000 mi² (fig. 1). The demarcation between the upper and lower basins near Iron Gate Dam corresponds with the transition from a geologic terrane dominated by permeable volcanic rock to a terrane dominated by older rock with much lower permeability. Hence, there is negligible groundwater flow between the upper and lower basins.


The upper Klamath Basin occupies a broad, faulted volcanic plateau that spans the boundary between the Cascade Range and the Basin and Range geologic provinces. The basin is bounded by the volcanic arc of the Cascade Range on the west, the Deschutes River Basin to the north, internally drained basins to the east, and the Pit River Basin to the south. The elevation of the Cascade Range along the western margin ranges from 5,000 to 7,000 ft with major peaks exceeding 9,000 ft. The interior parts of the basin are dominated by northwest-trending, fault-bounded basins, typically several miles wide, with intervening uplands. Basin floors range in elevation from roughly 4,000 to 4,500 ft, and adjoining fault-block upland elevations range from 4,500 to more than 5,000 ft. The northern and eastern parts of the upper Klamath Basin consist of a volcanic plateau with numerous eruptive centers, some of which reach elevations exceeding 8,000 ft. The southeastern margin of the upper Klamath Basin consists of a broad, rugged, volcanic upland known as the Modoc Plateau, where the land-surface elevation ranges from 4,500 to 5,000 ft. The southern margin of the basin is marked by the broad shield of Medicine Lake Volcano, which reaches an elevation of 7,913 ft.


The upper Klamath Basin is semiarid because the Cascade Range intercepts much of the moisture from the eastward moving Pacific weather systems. Mean annual precipitation (1971–2000) ranges from 67 in. at Crater Lake National Park in the Cascade Range to 13.5 in. at the city of Klamath Falls, Oregon (Western Regional Climate Center, 2013). November through March precipitation accounts for 71 percent of the total at Crater Lake and 64 percent of the total at Klamath Falls. Most precipitation falls as snow at high elevations. The interior parts of the basin are dry during the spring and summer; mean monthly precipitation at Klamath Falls generally is less than 1 in. from April through October. Winters generally are cold, with January mean minimum and maximum temperatures at Klamath Falls of 21.9 and 39.9°F, respectively. Summers, in contrast, are warm, with July mean minimum and maximum temperatures at Klamath Falls of 51.9 and 85.7°F, respectively.


The 250-mi-long Klamath River begins at the outlet of Upper Klamath Lake. Major drainages upstream of the lake include the Williamson, Sprague, and Wood Rivers (fig. 1). The Williamson River drains much of the northern and eastern parts of the basin and flows directly into Upper Klamath Lake; the Sprague River, tributary to the Williamson, drains part of the eastern side of the basin. The Wood River and several smaller tributaries drain the valley north of the lake and the adjacent part of the Cascade Range. The Lost River system, also included in the upper Klamath Basin, is internally drained. It originates at Clear Lake in the southeastern part of the basin and originally terminated in Tule Lake. Tule Lake has now mostly been drained, and Lost River is now diverted just downstream of Olene through a channel to the Klamath River. Generally, little water from the Lost River drainage upstream of the diversion channel flows to the Tule Lake subbasin. The Tule Lake sumps, remnants of the original lake, now collect return flow from the Project (fig. 1). 


Upper Klamath Lake is the largest lake in the basin with a surface area between 100 and 140 mi² (including non-drained fringe wetlands) depending on stage (Hubbard, 1970; Snyder and Morace, 1997). Lake stage and outflow are controlled by a diversion dam at the lake outlet. About 1 mi downstream of the diversion dam, the river flows into a 20-mi-long narrow reservoir behind a dam at Keno. John C. Boyle Reservoir and its dam are about 10 mi downstream of Keno. Downstream of the John C. Boyle Reservoir, the river enters a narrow canyon and flows freely about 20 mi to Copco Lake (a reservoir) and immediately downstream of that, Iron Gate Reservoir. Iron Gate Dam, which impounds Iron Gate Reservoir, marks the downstream boundary of the upper Klamath Basin. No impoundments are on the Klamath River downstream of Iron Gate Dam.


The surface-water hydrology of the upper Klamath Basin has been extensively modified by drainage of lakes and wetlands for agriculture, and the diversion and routing of water for irrigation. Prior to development, the Tule Lake and Lower Klamath Lake subbasins contained large lakes fringed by extensive wetlands. At high stage (about 4,060 ft elevation) Tule Lake covered an area exceeding 150 mi² (La Rue, 1922). Historical accounts indicate that at high stage Tule Lake drained into the lava flows along the southern margin. Because of subsurface drainage, Tule Lake was not an evaporative sink and was not saline (La Rue, 1922). Since 1942, excess water from the Tule Lake sumps has been pumped using a tunnel through Sheepy Ridge west into the Lower Klamath Lake subbasin. The Lower Klamath Lake subbasin once held a large lake-marsh complex that covered about 88,000 acres, about 58,000 acres of which were marginal wetlands with the remaining 30,000 acres open water (La Rue, 1922). Lower Klamath Lake was connected to the Klamath River through a channel known as the Klamath Strait, and probably through the expansive wetland that separated the lake from the river elsewhere. In the early 1900s, an earth-fill railroad bed was constructed across the northwestern margin of the Lower Klamath Lake subbasin, cutting off flow between the lake and river except at the Klamath Strait. In 1917, the control structure built into this impoundment at the Klamath Strait was closed, cutting off flow to the lake. As a result, Lower Klamath Lake is now largely drained, with much of the former lakebed and fringe wetlands under cultivation. Areas of open water remain in the Lower Klamath Wildlife Refuge in the southern part of the subbasin in California.


About 500,000 acres of agricultural land are irrigated in the upper Klamath Basin, roughly 190,000 of which are included in the Project (Carlson and Todd, 2003; Natural Resources Conservation Service, 2004). This total does not include wildlife refuge areas within the Project.


Irrigation water comes from various sources in the upper Klamath Basin. Upstream of Upper Klamath Lake, in the Williamson, Sprague, and Wood River drainages, private (off-Project) irrigation water primarily comes from diversion of surface water from the main stem streams or tributaries. A smaller amount of irrigation water is pumped from wells, particularly in the Sprague River Valley and Klamath Marsh areas (fig. 1). In the Langell and Yonna Valleys of the upper Lost River subbasin, irrigation water comes from Clear Lake and Gerber Reservoirs. Irrigators use groundwater and some surface water in Swan Lake Valley. Groundwater is used for irrigation in areas not served by irrigation districts and to supplement surface-water supplies throughout the area. 


Upper Klamath Lake is the largest single source of irrigation water in the basin. Water is stored in and diverted from the lake to irrigate the Project, which encompasses land south of Klamath Falls, including the Klamath Valley, Poe Valley (in the Lost River subbasin), and the Lower Klamath and Tule Lake subbasins. Water diverted from the Klamath River several miles downstream of the lake also is used for irrigation and refuges in the Lower Klamath Lake subbasin. Return flow from the Lower Klamath Lake subbasin is routed through a series of pumping stations back to the Klamath River.


A certain amount of groundwater is used for irrigation on land surrounding the Project upslope of the major canals. Primary areas of groundwater use surrounding the Project include the southern end of the Klamath Hills, parts of the Klamath Valley, and the northern and eastern margins of the Tule Lake subbasin (fig. 1). Some groundwater traditionally has been used for supplemental irrigation in the Project area. 


Previous Studies


The modeling work described in this report builds on efforts dating back several decades. Newcomb and Hart (1958) described the geologic framework of the groundwater system, inventoried wells, and quantified groundwater discharge to major springs and streams in the basin. They also demonstrated the relation between spring discharge and long-term precipitation patterns. At the time of their study, total groundwater pumping in the Oregon part of the basin was about 30,000 acre-ft. Leonard and Harris (1974) refined the assessment of groundwater in the upper basin, inventoried additional wells, and quantified geographic distribution of groundwater discharge along the Sprague and Lost Rivers. At the time of the Leonard and Harris study, total groundwater pumping in the Oregon part of the basin was about 61,000 acre-ft. Both these early studies were restricted to Oregon.


One of the first detailed groundwater studies in the California part of the upper Klamath Basin was by Wood (1960), who mapped the geology and described the groundwater hydrology of the Butte Valley area. About 21,000 acre-ft of groundwater was pumped for irrigation in Butte valley in 1953. Groundwater continues to be a major source of irrigation water in Butte Valley. Hotchkiss (1968) described the geologic and hydrologic conditions in and around Lava Beds National Monument adjacent to the Project. The primary purpose of the Hotchkiss study was to identify water sources for the Monument. The California Department of Water Resources (CDWR; 1998) published an investigation into well interference in Butte Valley, which includes results of aquifer tests and estimates of aquifer hydraulic characteristics. None of these early studies focused specifically on the area of the Project.


In response to a marked increase in groundwater pumping in 2001, California Department of Water Resources (2003) investigated the hydrogeology and groundwater conditions in the Tule Lake subbasin. That report includes an inventory of wells and pumping in the area, a description of the hydrogeology, hydrographs of several wells, a map of the top of bedrock beneath the basin, and maps of the water-table surface at different times. 


The Oregon Water Resources Department (OWRD) conducted an investigation of groundwater conditions in the upper Lost River area in response to increased groundwater demand and the potential for interference with major springs discharging to the river (Grondin, 2004). For that study, OWRD inventoried and measured water levels in hundreds of wells, measured groundwater discharge to the upper Lost River, described the geologic framework, and reported results of several aquifer tests conducted in the area. Grondin (2004) includes a comprehensive dataset, detailed analysis, and includes maps of the top of bedrock and the water-table elevation in the area.


The present (2013) understanding of groundwater-flow directions, the influence of climate, and the effects of pumping, is the result of many decades of well inventory and water level monitoring by multiple agencies. CDWR has collected extensive data in the California part of the basin, particularly the Butte Valley and Tule Lake areas. Several long-term monitoring wells operated by OWRD in the upper Klamath Basin have records extending back to the 1960s, and extensive data from the upper Lost River subbasin and around the Oregon parts of the Project. The USGS has collected quarterly water-level data on 50–70 wells since the late 1990s, focusing primarily on upland areas to characterize climatic influences (Gannett and others, 2007). 


Building on these decades of data collection and interpretive studies, Gannett and others (2007) described and synthesized the groundwater hydrology of the entire upper Klamath Basin and developed a regional water budget, a map and table that describe groundwater discharge to streams, and a regional water-table map. That effort formed the basis for development of the regional groundwater-flow model and coupled the groundwater-management model (Gannett and others, 2012) used for the simulations described in this report.


Groundwater-Simulation Model


The groundwater hydrology and groundwater-simulation model of the upper Klamath Basin are described in detail in Gannett and others (2007, 2012). The discussion in this section is largely from those reports.


The upper Klamath Basin has a substantial regional groundwater system that is used for irrigation in certain areas of the basin and supports aquatic habitat for endangered species. Tertiary and Quaternary volcanic rocks and sedimentary deposits underlie the region and compose eight regional-scale hydrogeologic units. The late Tertiary and Quaternary volcanic rocks in the basin are generally permeable and comprise a system of interconnected aquifers that are interbedded with fine-grained lake deposits and basin-filling sediments. The regional groundwater system is underlain and bounded on the east and west by older, less permeable Tertiary rocks that act as barriers to groundwater flow. Boundaries between the regional groundwater system and the basins to the north and south are defined by surface-water divides. 


Groundwater originates as recharge in the Cascade Range and upland areas in the basin interior and eastern margins, and flows toward stream valleys and interior subbasins. Groundwater discharges to streams throughout the basin, with large amounts of groundwater discharging to streams in the Wood River subbasin, the lower Williamson River, and along the margins of the Cascade Range. Groundwater also discharges directly to Upper Klamath Lake, and a substantial amount of the total inflow to the lake originates as groundwater discharge to streams and spring complexes within roughly 10 mi of the lake. Groundwater discharge areas also are in the eastern part of the basin (for example, in the upper Williamson, Sprague River, and Lost River subbasins). 


The groundwater system in the upper Klamath Basin responds to external stresses, such as climate cycles and groundwater pumping. Basin-wide, decadal-scale climate cycles are the largest factors controlling groundwater-level and discharge fluctuations. Water-table fluctuations of 5 ft or more are common throughout the basin as a result of decadal-scale climate cycles, and groundwater discharge varies by a factor of two or more as a result of wet and dry periods. The response of the groundwater system to pumping generally is largest in areas with irrigation pumping. The effects of recent increases in pumping by the Project are apparent in areas within and adjacent to the Project where the water table has declined more than 30 ft over areas covering tens of square miles (Gall, 2011). The effects of pumping on streams and springs, however, generally are diffuse and difficult to measure. 


The groundwater-simulation model of the upper Klamath Basin was developed using the USGS MODFLOW groundwater-modeling code (McDonald and Harbaugh, 1988). The simulation model has 285 rows, 210 columns, and 3 layers; model cells have dimensions of 2,500 by 2,500 ft and model layers have thicknesses ranging from 5 to 3,600 ft. The hydraulic characteristics of the regional groundwater system are defined with 18 model-parameter zones that represent large-scale geologic conditions. Boundaries with adjacent basins and underlying low permeability Tertiary strata are defined as zero-flux boundaries in most areas. Head-dependent boundaries define streams, lakes and reservoirs, agricultural drains, evapotranspiration in areas of shallow groundwater, and boundaries with adjacent basins in select areas. Groundwater recharge and pumping at more than 1,000 wells are simulated as time-varying fluxes defined for quarterly stress periods. The groundwater model was calibrated to transient conditions from 1989 to 2004 using groundwater-level observations at 663 wells and groundwater-discharge measurements for 10 stream reaches or spring complexes. The calibrated groundwater model simulates climate-driven decadal variations in groundwater levels and discharge in the upper Klamath Basin, as well as the observed seasonal and year-to-year groundwater-level fluctuations in response to groundwater pumping.


The groundwater-simulation model has the capability to calculate the effects of alternative pumping strategies on groundwater levels and discharge to streams, lakes, and drains. Example model simulations that show the various responses of groundwater levels and discharge to pumping depending on the pumping location are shown in Gannett and others (2012). Simulations show that the groundwater-discharge features most affected by pumping in the area of the Project are agricultural drains, and effects on other surface-water features are relatively small. When coupled with optimization, the capability of the groundwater-simulation model is enhanced. The coupled simulation-optimization model can evaluate the effects of a wide range of potential pumping scenarios on groundwater levels and discharge, and identify the best pumping strategy to meet the goals of water users and resource-management agencies. 


First posted April 23, 2014