Purpose and Scope

The purpose of this report is to describe an updated hydrogeologic framework for the Big Lost River Basin to help inform current water-management decisions. The hydrogeologic framework includes three major parts: (1) a conceptual description of the hydrogeologic units, (2) a three-dimensional hydrogeologic framework model of the hydrogeologic units and subunits, and (3) a description of groundwater occurrence and movement within the basin. This framework is based primarily on a compilation and interpretation of existing data and previous studies, including well-driller reports, geologic maps, and groundwater-level data collected by IDWR, as well as borehole geophysical data collected during this study. The three-dimensional hydrogeologic framework model and supporting hydrogeologic framework well database are available in a companion data release to this report (Zinsser, 2021). This hydrogeologic framework is the first of three planned reports on water resources of the Big Lost River Basin; subsequent reports are intended to present a water budget for the basin and to describe the results of a series of measurement events quantifying gains to and losses from surface water in the Big Lost River (Dudunake and Zinsser, 2021).

Study Area

The Big Lost River Basin encompasses mountainous uplands and an irrigated agricultural valley in south-central Idaho (fig. 1). The mountainous areas include the western face of the Lost River Range to the east, the White Knob Mountains to the west, part of the Pioneer Mountains to the far west, and part of the Boulder Mountains to the northwest. Most of the basin is mountainous rangeland; about 8 percent of the land area is agricultural land and 18 percent is forest. Altitude in the basin ranges from a maximum of 12,667 feet (ft) on Mount Borah, the highest point in Idaho, to a minimum of 5,240 ft near Arco, Idaho. Mean annual precipitation in the basin is 18.2 inches (U.S. Geological Survey, 2016).

The main geographic feature in the Big Lost River Basin is a long, narrow, northwest-trending, intermontane valley stretching from the headwaters of Thousand Springs Creek to south of Arco (fig. 1). The valley is fault-bounded on the eastern range front and contains extensive irrigation infrastructure, including the Mackay Dam and Reservoir and a system of canals. Landmarks and incorporated communities in the valley from north to south include Chilly, Mackay, Leslie, Darlington, Moore and Arco. The Big Lost River Valley specifically refers to the valley south of Mackay Reservoir. The drainage area contributing to U.S. Geological Survey (USGS) streamgage 13132500, located at the mouth of the Big Lost River Valley south of Arco, is 1,410 square miles (U.S. Geological Survey, 2016). Downstream from this streamgage, the Big Lost River arcs south and east through the INL until reaching the Big Lost River Sinks, where the surface water is lost to evaporation or percolates into the subsurface and recharges the eastern Snake River Plain aquifer. Streamflow through this full channel length is intermittent depending on precipitation and upstream irrigation demands. The eastern Snake River Plain aquifer is highly productive and consists of volcanic and sedimentary rocks. It is recharged through infiltration of precipitation, applied irrigation water, streamflow (including the Big Lost River), and groundwater inflow from tributary mountain basins (including the Big Lost River Basin) (Ackerman and others, 2006).

Irrigation is vital to agriculture and ranching in the Big Lost River Basin. The first water right was filed in 1879 by an early settler for a ranch north of Mackay. Copper mining boomed in the White Knob Mountains and small-scale, irrigated agriculture began in the valley between Mackay and Arco in the 1880s (Bottolfsen, 1926). Extensive irrigation infrastructure, including the construction of Mackay Dam and Reservoir, was developed in the Big Lost River Basin during the early 1900s under the Carey Act (Lovin, 1987). By 1931, Livingston (1931) attributed the “present economic distress” to the overdevelopment of irrigated lands during the relatively wet period from 1904 to 1918. To supplement inadequate surface-water supply, extensive groundwater-resource development occurred in the basin from 1950 to 1977 (Sukow, 2017). Surface-water and groundwater irrigation use continues as the predominant water use for counties in the study area (U.S. Geological Survey, 2021a). Thus, water-resource management for irrigation, streamflow, and other economic benefits remains a vital concern in the Big Lost River Basin (Bernal and Broadie, 2017).

Although the Big Lost River Basin is areally extensive, this hydrogeologic framework focuses on the alluvial aquifer in the Big Lost River Valley, with less attention paid to the alluvial aquifer in the valley north of Mackay Reservoir, in the upper Big Lost River valley (the primary drainage basin west of Chilly, hereinafter referred to with lowercase “valley” as that area is not a formally designated valley in the U.S. Board on Geographic Names), in the Antelope Creek valley, and the bedrock aquifer (fig. 1). This focus is largely driven by predominant water-use and water-management questions associated with the alluvial aquifer in the Big Lost River Valley. Additionally, few hydrologic data are available for bedrock, tributaries, and upland areas. This study follows the convention of Crosthwaite and others (1970a) by using “alluvial aquifer” to refer to the water-bearing, unconsolidated basin-fill sediments and using “major bedrock aquifer” to refer to water-bearing sedimentary rocks. The Big Lost River Valley as far upstream as the Mackay Dam is within the boundaries of the Enhanced Snake Plain Aquifer Model Version 2.1 (fig. 1; Idaho Department of Water Resources, 2013), and Quaternary basalts constitute the main aquifer in the eastern Snake River Plain with alluvium in tributary drainage basins (Whitehead, 1992). Thus, although groundwater in the Big Lost River Basin is part of and tributary to the eastern Snake River Plain aquifer, this study follows the nomenclature convention of Crosthwaite and others (1970a) to describe aquifer resources within the Big Lost River Basin.

Previous Work

Early work in the Big Lost River Basin focused on improving water delivery for irrigation and, therefore, necessarily considered patterns of surface-water losses and gains. Debler and others (1931) described major losses of water from the Big Lost River in sinks near Chilly, near Darlington, below the Moore Diversion, and below Arco. Losses also were reported in Antelope Creek upstream from its confluence with the Big Lost River. Water returns to the Big Lost River were described upstream from Mackay Reservoir, in the river upstream from Mackay, and upstream from the Moore Diversion (fig. 1). Notably, Debler and others (1931, p. 3) concluded that “the waters lost by Big Lost River in the Chilly sinks is returned to the stream above Mackay, and largely so above Mackay Reservoir” and that “waters lost in the Darlington Sinks merely pass through the Antelope Creek delta and return in part to the river in the vicinity of the Moore dam [diversion].” Debler and others (1931) suggested using by-pass channels or canals to avoid losses of surface water to groundwater in the informally named Chilly sinks below the Moore Diversion and across the Antelope Creek fan but advised against a canal to bypass water around the Darlington Sinks because of cost and uncertainties regarding patterns of loss and gain in this area. Early work by Stearns and others (1938) provided a relatively detailed description of surface-water and groundwater resources in the Big Lost River Basin as a part of a comprehensive volume describing the water resources of the Snake River Plain. This work included one map showing the depth to water table with depths less than 10 ft coinciding with areas of gaining surface water, and a second more detailed map of groundwater surface altitudes near the Moore Diversion (Stearns and others, 1938).

The most comprehensive characterization of water resources in the Big Lost River Basin was made by Crosthwaite and others (1970a). The study provided a hydrogeologic framework and water budget by drawing on a wide range of new and existing data, including precipitation data; streamflow records and measurements; geologic maps; groundwater levels in approximately 200 existing wells; geologic and hydrologic interpretation of six new wells; and seismic, gravity and resistivity surveys. Crosthwaite and others (1970a) defined five hydrogeologic units: (1) carbonate rocks, (2) noncarbonate rocks, (3) cemented alluvial deposits, (4) unconsolidated alluvial deposits, and (5) basalt. Carbonate rocks represented the major bedrock aquifer, and basalt locally yielded substantial quantities of water, but the principal aquifer was the unconsolidated alluvial fill in the valley. Geophysical data were used to define basin fill geometry; basin fill was interpreted to be 2,000–3,000 ft in the Thousand Springs Creek area, relatively thin (200–400 ft) near the Mackay Dam, over 5,000 ft east of Mackay, and progressively thinning to about 2,500 ft near Arco. Test wells described multiple water-bearing zones near Arco; groundwater generally was unconfined but there were some areas of perched water and some artesian wells. Variability in groundwater gradient, indicated by areas of gaining versus losing streamflows, was attributed to variations in the cross-sectional area of the valley, permeability of the alluvial fill, and the quantity of water in the aquifer. Like previous reports, Crosthwaite and others (1970a, p. 34) recognized the unusually extensive connection between surface water and groundwater, stating that “[s]urface and ground water are so closely related that neither can be considered as a separate source of supply.”

Water use in the Big Lost River Basin has been a vital concern for almost a century (Debler and others, 1931; Livingston, 1931). Although early reports focused on the efficiency and viability of surface-water infrastructure such as using canals to move water past the various sinks, by 1970 attention had turned to the potential effects of groundwater resource development. Crosthwaite and others (1970b) showed the value of conjunctive management of surface water and groundwater, suggesting that groundwater could be more heavily used in dry years but that use would need to be minimized in wet years to promote recharge because of the likelihood that increased groundwater withdrawals could interfere with streamflows and hence affect surface-water rights. Nonetheless, this era included substantial development of groundwater resources, with groundwater appropriations in the basin increasing from about 1 to 80 percent of the current total rights from 1950 to 1977 (Sukow, 2017). Correspondingly, reports in the 1980s and 1990s reflect an increasing need to understand the nature of the groundwater resource. Szczepanowski (1989) summarized previous work and advocated for conjunctive water-use management, Bassick and Jones (1992) completed an aquifer test and described groundwater flow and pumpage in the Big Lost River Valley, and Johnson and others (1991) quantified the detrimental impact to surface-water rights from increased groundwater use. A series of reports in the 1980s also documented the hydrologic and other effects of the 1983 Borah Peak earthquake (Stein and Bucknam, 1985; Wood and others, 1985). In a recent review of groundwater levels in the Big Lost River Basin, Sukow (2017) reported that many wells had statistically significant declining trends in groundwater levels since 1977. These trends were attributed to a combination of more frequently occurring dry years and increased use of groundwater for crop irrigation and were concurrent with decreased surface-water diversion use.

Well Data

During the development of the hydrogeologic framework, a digital hydrogeologic framework well database of 621 wells in the Big Lost River Basin was created in RockWorks17™; the database includes well identifiers, location, total depth, borehole lithology, well construction, and other information. The primary source of this information was the IDWR database of scanned well-driller reports for wells in Idaho (Idaho Department of Water Resources, 2021a). Well drillers were not required to obtain a permit to drill a well or submit a report to IDWR prior to 1987; thus, the IDWR database mostly contains reports since 1987 and should not be considered comprehensive. About 1,800 reported wells from within the Big Lost River Basin initially were identified using the IDWR “Wells” geographic information system (GIS) dataset (Idaho Department of Water Resources, 2018). These wells were winnowed down to about 600 for inclusion in this analysis. Wells were selected foremost to achieve good spatial coverage across the basin; if multiple wells were available in an area, the well with maximum depth or most complete lithology and well construction descriptions was selected for inclusion. Additionally, groundwater-level monitoring wells in the Big Lost River Basin that were not represented in the IDWR well construction database were added to the hydrogeologic framework well database. When possible, construction and lithology information was retrieved from USGS well data archives and entered in RockWorks17™. However, 10 wells without complete lithology and well-construction information were included nonetheless in the hydrogeologic framework well database because these wells were used to monitor groundwater levels. The digital hydrogeologic framework well database is available in the companion data release (Zinsser, 2021).

Three-Dimensional Hydrogeologic-Framework Modeling

A three-dimensional hydrogeologic-framework model was developed to assist in visualizing and understanding hydrogeologic units in the basin. Well lithologies, interpreted from well-driller reports and entered in the digital hydrogeologic-framework well database, were assigned to hydrogeologic units and used to develop the three-dimensional model that represents the subsurface thickness and distribution of hydrogeologic units. Well locations were obtained from the IDWR “wells” GIS dataset (Idaho Department of Water Resources, 2018). Historically, well locations were reported to the 160-, 40-, or 10-acre tract level using the Public Land Survey System (PLSS); therefore, well locations may be approximate, although more recent wells tend to be more precisely located using Global Positioning System. Well land-surface altitudes were assigned using the USGS National Elevation Dataset 1/3 arc-second grid (approximately 33 ft or 10 meters [m]) digital elevation model; U. S. Geological Survey, 2020).

For modeling purposes, well lithologies were simplified to predominant type; for example, sandy clay was entered with clay as the primary lithology and sand as the secondary lithology. Where wells intercepted bedrock, the rock was interpreted based on the well-driller description, published geologic maps, and professional judgment. For example, “lava” was interpreted as basalt when found in the southern end of the Big Lost River Valley, and descriptions such as “sandstone,” “limestone,” and “shale” were simplified to the more generic term “sedimentary rock.” Each well lithology was then assigned to an overarching hydrogeologic unit.

There are 621 wells in the hydrogeologic-framework well database, of which 604 are real wells and 17 are synthetic wells. The 17 synthetic wells were generated to improve and constrain model representation of key bedrock features. Of these 17 synthetic wells, 13 are duplicates of real wells for which well-driller reports showed termination in several feet of bedrock; in the duplicate record, bedrock was extended to 25 ft total at the bottom of the well to improve modeling. For 4 of the 17 synthetic wells, 300-ft-deep sedimentary rock boreholes were generated in Leslie Butte to constrain model representation of this bedrock feature where no actual wells were located. In total, 608 wells (17 of which were simulated) were used to generate the three-dimensional hydrogeologic framework model (the 13 real wells that were duplicated with extended bedrock were not used to generate the model).

The three-dimensional hydrogeologic framework model was developed in RockWorks17™ using the highest probability algorithm with a horizontal search distance of 6,562 ft (2,000 m), a vertical search distance of 50 ft (15.2 m), and closest point declustering. Both feet and meters are reported because the model projection, Idaho Transverse Mercator, uses meters for horizontal location and well-driller reports use feet for depth. A polygon shapefile representing the alluvial aquifer surface trace was applied as a filter in RockWorks17™ prior to model processing such that the model was only developed for the area inside the polygon. A grid file representing the surface topography—with 656-ft (200-m) horizontal spacing and altitude derived from the National Elevation Dataset 1 arc-second grid (approximately 98 ft or 30 m) digital elevation model (U.S. Geological Survey, 2020)—was applied as a superface filter in RockWorks17™ to force the model to adhere to surface topography. Voxel size, where one voxel constitutes one model point as represented in three-dimensional space, matched the surface grid and was 656 ft by 656 ft by 25-ft deep (200 m by 200 m by 7.6-m deep).

Groundwater-Level Measurements

IDWR currently maintains a network of groundwater-level monitoring wells in the Big Lost River Basin, although the number and location of wells monitored have varied over time, and many of the wells previously were monitored by the USGS (table 1). Groundwater-level measurements therefore are available on the USGS website for older data (U.S. Geological Survey, 2021a), and on the IDWR website for older and newer data (Idaho Department of Water Resources, 2021b). These data have several uncertainties. Groundwater levels were not barometrically corrected prior to October 2019, but unvented transducer data have been corrected since. The land-surface elevations of the wells vary in provenance and thus the vertical datum cannot be verified for most of them. However, because groundwater-level altitudes vary by more than 1,000 ft from the upper to lower parts of the study area, the influence of uncertainties attributed to barometric effects and potential discrepancies between vertical datums likely are small relative to the overall gradient described. Groundwater levels within wells were not directly compared between corrected and uncorrected time periods.

Groundwater-level data were used to simulate potentiometric surface altitude by kriging in ArcGIS^™. Because wells varied in instrumentation and the frequency of measurements, manual measurement events were preferentially selected to provide maximum well coverage. For each year selected, the groundwater-level measurements recorded closest to April 1 or November 1, plus or minus 3 weeks, were used. Thus, data associated with an “April” surface for a given year may have been collected in March or April. Groundwater-level data also were used to evaluate vertical gradients in a series of shallow, well clusters, installed by IDWR in late 2019 as part of this project. Well cluster locations were selected by IDWR based on access and prior knowledge about areas where vertical gradients likely were important to understanding groundwater movement.

Geophysical and Borehole Video Log Data

Borehole geophysical data and borehole video files were collected from 10 groundwater-level monitoring wells using wireline logging tools. Geophysical logs were collected only in cased borehole sections because of the concerns of neutron and gamma-gamma logging tools being ensnared in uncased (open) wells. For each well, the measured depth to water, logging date, logging depth, and log type are identified in table 2. Wireline geophysical and borehole video equipment are owned and operated by the USGS INL Project Office. The USGS calibrates geophysical logging equipment and sensors annually or on an as-needed basis.

Borehole video files were used to confirm well depth, well-casing integrity, well-casing size, and well obstructions prior to collecting geophysical data. Natural gamma, neutron, gamma-gamma dual density, and fluid logs were collected and used to examine lithology and to approximate location and thickness of unconsolidated sediment and rock units. The natural gamma signal was used to identify changes in lithology by looking at signal changes. Neutron logs are a general indicator of hydrogen content and, when combined with natural gamma logs to determine lithology location, can be used to identify perched groundwater. The gamma-gamma dual density log was used to examine bulk density changes in the rock and unconsolidated sediment layers. Fluid specific conductance and temperature provide a general indicator for changing water chemistry when run in a vertical fluid column.

Well-driller reports, when available, were used initially to identify general borehole lithology for each well. However, this information was general or incomplete so geophysical logs were used to update lithology by estimating the location of unconsolidated sediment and rock contacts and changes in unconsolidated sediment layering where reasonable. Where well-driller reports were not available, well lithology was approximated based on well-driller reports for nearby wells (2 miles or less) and geophysical data. Wells without well-driller reports presented a higher degree of uncertainty when interpreting lithology. A more detailed description of the geophysical methods is presented in appendix 1.

Geologic History

The geologic history of the Big Lost River Basin is complex, and much of it is tangential to the movement and occurrence of water in the basin. As such, the geologic history is presented briefly here (fig. 2) and the reader is referred to other authors for more extensive treatment. The Big Lost River Basin is a part of the Cordilleran orogenic belt and the Basin-and-Range province; it lies north of the Snake River Plain-Yellowstone hot spot track. The oldest rocks present are the Archean to Paleoproterozoic metamorphic rocks exposed in the Pioneer core complex at the western boundary of the basin (Link and Janecke, 1999; Lewis and others, 2012).

Most of the sedimentary rocks exposed in the White Knob, Pioneer, and Lost River Ranges were deposited in a series of successive, passive continental margin basins during the Paleozoic. During the late Neoproterozoic and Cambrian, siliciclastic rocks of the Windermere Supergroup were deposited southward into a passive basin; these rocks are exposed in the pebbly sandstones of the Wilbert Formation at the base of the Borah Peak horst, and on the western flank of the Pioneer core complex. An unconformable contact between these rocks and overlying Lower Ordovician sandstones of the Summerhouse Formation suggest an uplift event described as the Salmon River arch or Lemhi arch. Subsequent tectonic stability and depositional continuity began in the Middle Ordovician, represented geologically by the Kinnikinic Quartzite, and extended through the Devonian. Extensive carbonate banks developed during the Silurian, manifested in the dolomitic rocks of the Fish Haven and Laketown Formations. These rocks are exposed along much of the western face of the Lost River Range (Link and Janecke, 1999; Lewis and others, 2012). The Antler orogeny, characterized by presumed uplift to the west, occurred in the Late Devonian and Early Mississippian and is represented in the clastic foreland basin deposits of the McGowan Creek Formation argillite and conglomerates of the Copper Basin Group. In the Late Mississippian and Pennsylvanian, large carbonate bank complexes developed east of the Antler trough, eventually prograding to the west over the former trough. These carbonate bank deposits are extensive and include the White Knob Limestone and the Big Snowy, Surrett Canyon, South Creek, and Scott Peak Formations. These rocks are exposed most prominently in the upper Big Lost River, in the White Knob Mountains near Mackay (including abutting the Mackay Dam), near Leslie, along Antelope Creek, and at the southern end of the Lost River Range east of Arco (Link and Janecke, 1999; Lewis and others, 2012). Fine-grained siliciclastic and carbonate rocks deposited in the basin during the Pennsylvanian and Permian include the Snaky Canyon and Arco Hills Formations and are exposed most prominently at the mouth of Antelope Creek and in the Arco Pass Basin (Skipp and others, 1979; Link and Janecke, 1999; Lewis and others, 2012).

The development of the Cordilleran orogenic belt in the Mesozoic caused crustal deformation and shortening expressed in widespread folding and faulting. Regional thrust faults in the Pioneer Mountains in the western part of the Big Lost River Basin developed during this time, including the Hawley Creek-Copper Basin thrust plate that comprises the rocks exposed in the White Knob Mountains, the Lost River Range, and other ranges beyond the study area (Link and Janecke, 1999; Lewis and others, 2012). Extension and faulting occurred in the Tertiary prior to, continued synchronously with, and post-dated the Eocene Challis magmatic event. The Challis magmatic event was relatively short-lived but intense, producing lava flows, tuffs and intrusive rocks. The resulting Challis Volcanic Group includes some intrusive rocks that outcrop in the White Knob Mountains east of Mackay and minor sedimentary rocks throughout the Big Lost River Basin, but most of the rocks associated with the magmatic event are andesitic and dacitic lava flows and tuffs. These volcanic rocks unconformably overlie the Paleozoic sedimentary rocks and are prominent throughout the western side of the Big Lost River Basin in the Pioneer, Boulder, and White Knob Mountains (Moye and others, 1988; Link and Janecke, 1999; Lewis and others, 2012).

The eastern Snake River Plain represents the track of volcanism formed by the migration of the Yellowstone hot spot as the North American Continental Plate moved westward over a mantle plume during the late Tertiary and into the Quaternary (Pierce and Morgan, 1992). Basaltic shield volcanoes likely began as fissure eruptions and lasted as long as several months at a central vent or vent complex. These features define volcanic rift zones, which are linear arrays of volcanic landforms and structures. The volcanic rift zones developed roughly parallel to the basin-and-range structures at the boundaries of the eastern Snake River Plain (Kuntz and others, 1992). Craters of the Moon National Monument and Preserve is centered on the Great Rift, and the Arco-Big Southern Butte rift extends southeast from approximately Arco; flows from these rift zones mostly date to the Holocene (Kuntz and others, 1992; Skipp and others, 2009). Basaltic flows from the Craters of the Moon lava field are present in the far southwestern end of the study area, whereas the Lost River butte (the basalt knob west of Arco) and most of the basalts at the south end of the study area are associated with the Arco-Big Southern Butte rift (Kuntz and others, 1992; Skipp and others, 2009); these rocks are more broadly described as the Snake River Group. Uplift occurring with the Arco-Big Southern Butte rift created a new drainage divide and forced the west-flowing Pliocene Big Lost River east to its current course (Hodges and others, 2009).

Extension associated with Basin-and-Range faulting began in the late Tertiary, continues to the present, and generated the Lost River Range. Faults associated with this event generally strike northwest (Link and Janecke, 1999). The Lost River fault most recently ruptured in 1983 with the magnitude 7.3 (Richter scale) Borah Peak earthquake, originating at a depth of about 13 kilometers (km) and manifesting at the land surface with a fault scarp that visibly offsets alluvial deposits by more than 4 m. The Lost River fault is a normal fault for which the movement is down to the west (that is, the main valley is down-dropped relative to the Lost River Range) and has an estimated 2.7 km of structural relief (Stein and Bucknam, 1985). Thus, the present geometry of the main valley developed in the late Tertiary and is defined by Basin-and-Range extension and the Snake River Plain volcanism.

Less comprehensive geologic work has been done on the alluvial deposits of the Big Lost River Basin, but the timing of basin development suggests that alluvial fill in the main valley has been accumulating since the late Tertiary. Small cirque and valley glaciers were present in the Lost River Range and in the Pioneer and White Knob Mountains during the Pleistocene (Pierce and Scott, 1982). Pleistocene glacial deposits form most of the alluvial valley fill in the East and North Forks of the Big Lost River and in parts of the main valley northeast and southwest of Chilly (Lewis and others, 2012). The extensive river valley and alluvial fan gravels were deposited largely during the Pleistocene under markedly different climatic conditions and greater streamflow in both glaciated and non-glaciated drainages (Pierce and Scott, 1982). Although cementation occurs in some older alluvial deposits (Crosthwaite and others, 1970a), limited cementation has been described in more recently mapped alluvial fan surfaces and vertical sections (Patterson, 2006; Kenworthy and others, 2014). Alluvial deposits in the valley thus consist of glacial deposits in the upper Big Lost River and tributaries, extensive alluvial fans along the range fronts in the main valley, and fluvial stream deposits along the main axis of the valley. The cut and fill nature of streams and the subsequent reworking of glacial and alluvial deposits by stream migration has generated a highly spatially heterogenous distribution of sediments in the valley.

Adding to the complexity of the basin fill, basalt flows of the Snake River Group encroached from the south in the Quaternary, creating an area of interbedded basalt rocks and unconsolidated sediments in the southern end of the valley. Several researchers have suggested that basalt flows from the south have intermittently dammed the Big Lost River and produced a laterally continuous, fine-grained deposit in the subsurface near Arco (Stearns and others, 1938; Crosthwaite and others, 1970a). Detailed geologic mapping of the area shows at least three laterally extensive basalt flows that potentially could have blocked the Big Lost River for a period of time (Skipp and others, 2009).

Hydrogeologic Units

Given the complex geologic history of the Big Lost River Basin, the hydrogeologic units are defined broadly to simplify hydrogeologic interpretation and modeling, and to accord with the level of information available in the well-driller reports. The hydrogeologic units are shown in figures 2–3.

Geophysical Results

The results from the borehole geophysics were difficult to interpret in detail for many wells because multiple casing strings and large-diameter wells muted the response of the geophysical tools. Additionally, unknown construction methods and limited or entirely missing lithology descriptions further complicated interpretation. The primary goal for the borehole geophysics was to help determine and define, if applicable, the presence of a laterally continuous, fine-grained confining layer. Based on the results of the geophysics, as well as the three-dimensional hydrogeologic framework model, confining layers in the Big Lost River Basin aquifer generally are spatially discontinuous. Thus, the borehole geophysical results primarily are helpful in resolving local-scale lithology and are discussed in sections, “Three-Dimensional Hydrogeologic Framework Model” and “Groundwater Occurrence and Movement,” as pertinent. The full borehole geophysical results are presented in appendix 1 or can be accessed online through the GeoLog Locator web page (U.S. Geological Survey, 2021b).

Three-Dimensional Hydrogeologic Framework Model

As described in section, “Approach,” the three-dimensional hydrogeologic framework model was developed with lithology data from 608 wells (fig. 4). The hydrogeologic framework well database and three-dimensional hydrogeologic framework model are available in the companion data release (Zinsser, 2021). The hydrogeologic framework model is used primarily to describe the spatial distribution of hydrogeologic units in the Big Lost River Basin. An overview of the model boundaries, model zones, well locations, and geologic section lines is provided in figure 4. Model zones were delineated to simplify display and discussion of the model; they do not imply or represent hydrologic boundaries. Geologic sections parallel to the long-axis of a valley are referred to herein as long sections; sections that cut across a valley (perpendicular to the long-axis) are referred to herein as cross sections. The geologic sections illustrate slices of the three-dimensional hydrogeologic framework model with intercepted wells shown to illustrate how accurately the model represented well lithology. However, the model slice depicted in the section was developed from multiple proximal wells according to the modeling algorithm search distances (6,562 ft horizontal, 50 ft vertical); thus, hydrogeologic units and subunits represented in section may or may not occur in the intercepted wells. The model also contains interpolation gaps in some areas owing to the sparseness of underlying well data; these gaps are depicted as blank areas in the geologic sections.

Upper Big Lost River and Tributaries

The upper Big Lost River and tributaries primarily consist of upland areas dominated by the Paleozoic sedimentary rocks and Tertiary volcanic rocks units and narrow valleys with glacial and alluvial deposits constituting the thin Quaternary unconsolidated sediments unit. Correspondingly, the hydrogeologic framework model in the upper Big Lost River and Antelope Creek valleys has a greater volume of the Paleozoic sedimentary rocks and Tertiary volcanic rocks units (44 percent total), and more boulders (12 percent) in the Quaternary unconsolidated sediments unit, than in the main valley (table 4; fig. 5). Well data are sparse, but most wells are relatively shallow and are completed in the Quaternary unconsolidated sediments unit, although some deeper bedrock wells are completed in Paleozoic sedimentary rocks unit (fig. 6A–6B). In both the upper Big Lost River and Antelope Creek, the Quaternary unconsolidated sediments generally thicken and coarsen towards the mouth of the valley (fig. 6A–6B); this pattern is coincident with documented streamflow losses in the Chilly sinks and the Antelope Creek fan (Crosthwaite and others, 1970a; Clark, 2022).

Table 4.    

Hydrogeologic unit volumes in each hydrogeologic framework model zone, Big Lost River Basin, south-central Idaho.

[Hydrogeologic unit volumes by model zone are derived from the three-dimensional hydrogeologic framework model. Model zone boundaries are shown in figure 4. All numbers in table are percentages]

Table 4.    Hydrogeologic unit volumes in each hydrogeologic framework model zone, Big Lost River Basin, south-central Idaho.

Model zone	Paleozoic sedimentary rocks	Tertiary volcanic rocks	Quaternary basalt rocks	Quaternary unconsolidated sediments
				Boulders	Clay	Cemented sediments	Gravel	Sand	Soil
Upper Big Lost River and tributaries	22	22	0	12	13	2	16	12	2
Valley above Mackay Dam	15	0	0	1	18	4	24	35	2
Big Lost River Valley between Mackay Dam and Moore Diversion	16	8	0	3	14	10	29	19	1
Big Lost River Valley south of Moore Diversion	2	0	45	0	25	4	9	14	1

Valley Above Mackay Dam

The valley above Mackay Dam represents an alluvial fill, structurally controlled extensional basin with the Paleozoic sedimentary rocks unit in the east and northwest range fronts and Tertiary volcanic rocks unit to the west. The Quaternary unconsolidated sediments unit is variable in depth and subunits are laterally discontinuous with variable grain size. This is shown in the model, with deeper Quaternary unconsolidated sediments with more clay north of Chilly; deeper and coarser sediments near the Chilly sinks; and shallower and coarser sediments approaching the Mackay Reservoir (figs. 4, 7). The area with fine subunits (clay) shown north of Chilly (figs. 4, 7A–7B) aligns with geophysical results (app. 1, well IDWR-1) and well-driller reports of confining conditions. Areas with coarse sediment subunits (sand and gravel) align well with documented streamflow losses in the Chilly sinks (figs. 4, 7A–7C) and streamflow gains upstream from the Mackay Reservoir (Crosthwaite and others, 1970a). Some deep wells in the Paleozoic sedimentary rocks unit occur on the eastern valley margin and near Chilly (figs. 4, 7A), indicating the importance of this unit to water production at the margins of the Quaternary unconsolidated sediments unit. Overall, the Quaternary unconsolidated sediments unit in this part of the model is predominantly coarse (59 percent gravel and sand subunits), although the clay subunit (18 percent) and Paleozoic sedimentary rocks unit (15 percent) also represent appreciable model volumes (table 4; fig. 5).

Big Lost River Valley Between Mackay Dam and Moore Diversion

Below Mackay Dam, the Big Lost River Valley narrows substantially, with Paleozoic sedimentary rocks outcropping in the east and west, Tertiary volcanic rocks outcropping in the west, and thin Quaternary unconsolidated sediments in the valley bottom. The structurally controlled valley widens and deepens between Alder Creek and the Moore Diversion, with a thicker Quaternary unconsolidated sediment unit dominated by coarse subunits (gravel and sand). Prominent alluvial fans occur on the eastern range front throughout this part of the valley. The model represents the narrow, thin Quaternary unconsolidated sediments as shown in the sections through Mackay, with the Tertiary volcanic rocks unit on the west and Paleozoic sedimentary rocks unit on the east (fig. 8A–8B). The thin Quaternary unconsolidated sediments unit near Mackay has been associated historically with reported streamflow gains (Stearns and others, 1938; Crosthwaite and others, 1970a). The western Tertiary volcanic rocks unit and eastern Paleozoic sedimentary rocks unit shown in figure 8B. broadly agree with regional geology, but the modeled contact between the units is inconsistent with expected stratigraphy and was derived solely from the model extrapolation algorithm; it is unlikely to be an accurate representation of the subsurface geology.

Paleozoic sedimentary rocks also outcrop in Leslie Butte and are represented in the model at the head of the Darlington Sinks (figs. 4, 8A, 8C). The Darlington Sinks, known for large historical and contemporary streamflow losses (Crosthwaite and others, 1970a; Rice and Boyd, 2008; Dudunake and Zinsser, 2021), align with coarse subunits (gravel and sand) interpolated within the Quaternary unconsolidated sediments unit (figs. 8A, 8C, 9A). Fine subunits (clay) occur near the Moore Diversion (figs. 4, 9A–9B) where streamflow gains have been described historically (Debler and others, 1931; Crosthwaite and others, 1970a). Overall, this zone of the model has the highest proportion of the gravel subunit (29 percent) and lesser amounts of the sand (19 percent) and clay (14 percent) subunits within the Quaternary unconsolidated sediments unit, and the Paleozoic sedimentary rocks unit represents 16 percent (table 4; fig. 5). This zone also has the highest proportion of cemented sediments subunit (10 percent), which likely owes to the occurrence of wells in the large alluvial fans on the eastern range front.

Big Lost River Valley South of Moore Diversion

The Big Lost River Valley south of the Moore Diversion is characterized by laterally discontinuous subunits of the Quaternary unconsolidated sediments unit and interbedding of this and the Quaternary basalt rocks unit as the valley opens onto the volcanic rift zones of the eastern Snake River Plain. Correspondingly, from Moore to just north of the Lost River butte, the model represents only the Quaternary unconsolidated sediments unit with predominantly coarser subunits (gravel and sand) in the long section and cross section (figs. 4, 10, 11A). Well-driller reports describe multiple water-bearing and non-water-bearing zones near Moore and long open intervals occur in wells in this area (fig. 10). The cross section through Moore (fig. 11A) also shows some clay on the western valley margin with wells open below the clay subunit; this is consistent with reports of a confining layer in this area (Bassick and Jones, 1992) and consistent with the conceptual understanding of coarse-grained sediments in the valley axis and fine-grained sediments at the margins (Crosthwaite and others, 1970a).

From approximately the Lost River butte to south of Arco, the hydrogeologic framework model represents interbedded Quaternary unconsolidated sediments and Quaternary basalt rocks units (figs. 4, 10). Basalt is first intercepted in the model at well IDWR-22 north of the Lost River butte; from that point south, basalt becomes increasingly predominant in the model. Wells are open in multiple units at various depths; this is visible in the long section (fig. 10) and cross section (figs. 11B–11C, 12) and is consistent with previous descriptions of water-bearing unconsolidated sediments and basalts in this area (Crosthwaite and others, 1970a). For example, wells are open in two distinct basalt flows and multiple separate sand and gravel subunits in the long section through Arco (fig. 10), and are open in at least three basalt flows and several sand and clay subunits in the cross section south of Arco (fig. 12). Clay subunits also become notably more dominant in the Arco area, which is congruent with previous reports of a fine-grained confining layer in this area (Stearns and others, 1938; Crosthwaite and others, 1970a). Overall, this model zone represents 45-percent Quaternary basalts rocks unit and 53 percent Quaternary unconsolidated sediments unit, with major subunits clay (25 percent), sand (14 percent) and gravel (9 percent gravel) (table 4; fig. 5).

Groundwater Occurrence and Movement

Groundwater Occurrence

Based on the lithologies and depths of the wells in the hydrogeologic framework well database, groundwater resources are most heavily used in the Quaternary unconsolidated sediments unit within 250 ft of the land surface; this unit comprises the alluvial aquifer. The Paleozoic sedimentary rocks and Tertiary volcanic rocks units are important to water production and recharge at the margins of the alluvial aquifer, and the Quaternary basalt rocks unit is important to water production at the southern end of the Big Lost River Valley, but less data are available for these units than for the Quaternary unconsolidated sediments unit. Thus, most of the discussion of groundwater occurrence and movement focuses on the alluvial aquifer of the Quaternary unconsolidated sediment unit.

Although groundwater-level monitoring data exist for the Big Lost River Basin and can be used to describe regional groundwater movement in some parts of the valley (see section, “Groundwater Altitude and Flow Direction”), the coverage of these data is relatively sparse for the scale of the basin (fig. 1). Therefore, the depth to the top of the screened, perforated, or open interval (described herein as open or opening) for wells in the hydrogeologic framework well database was used to describe the general occurrence of water-bearing zones across the study area (fig. 13). Although the well opening data provided the best spatial coverage of water-bearing zones, these data have limitations and should not be understood as the top of the aquifer, nor as a potentiometric surface. For example, well-driller reports describe water found tens of feet before screen or perforations were installed in some wells, other wells are open solely at the bottom, and still other wells are open over long intervals (greater than 100 ft). The depth to screen therefore is deeper than the top of the aquifer in some places, may be shallower than a seasonally fluctuating water table in other places, and is a poor proxy for the potentiometric surface where wells are screened below confining layers. Additionally, domestic wells frequently are shallower with shorter screen intervals than irrigation wells, and wells often are drilled deeper over time, but well types and development date were reasonably evenly distributed across the basin and thus not expected to overly influence the results. Therefore, although it is an imperfect proxy, depth to well opening for the 604 real wells in the hydrogeologic framework well database was used to provide some insight into the general depth to water-bearing zones at a spatial resolution not possible using the groundwater-level monitoring wells.

Although wells are limited in the tributaries to the Big Lost River (East Fork and North Fork Big Lost River, Wildhorse Creek, Antelope Creek) and the valley above Chilly, the existing data suggest that groundwater is relatively shallow in these areas, with well intervals open less than 105 ft below land surface (fig. 13). Other areas with relatively shallow open intervals agree with historical accounts of areas with gaining streamflows, including above the Mackay Reservoir, between the towns of Mackay and Leslie, above the Moore Diversion, and near Arco. Similarly, areas of historical sinks align with moderate depth well openings (105–260 ft) in the Darlington Sinks but not in the Chilly sinks, where well openings were shallower (less than 105 ft). Some areas with deep well openings (261–662 ft) include wells northwest of Chilly, and wells open in bedrock at the valley margins south of Mackay, near Pass Creek, in the narrow part of Antelope Creek valley coincident with several mapped faults (Skipp and others, 2009), and south of Arco (fig. 13).

Groundwater Altitude and Flow Direction

Groundwater movement through an aquifer occurs as water flows from areas with higher hydraulic head to areas with lower hydraulic head; these head distributions are described by potentiometric surface maps. The IDWR maintains a network of groundwater-level monitoring wells in the Big Lost River Basin (table 1); these groundwater-level data were used to develop potentiometric surface maps to describe general groundwater movement between approximately Mackay and Arco. Because of the existence of confining layers in some parts of the valley, potentiometric surfaces are used here to be inclusive of both confining and non-confining conditions. The groundwater-level data, and hence the potentiometric surfaces, have some important limitations. For example, only about 24 wells have been semi-regularly measured over time, only three monitoring wells are located north of Mackay, and well completion depths and construction methods vary. Furthermore, wells may be open to multiple permeable zones, or screened in a locally confined layer. Several wells in the long-term monitoring network were removed from this analysis because they are screened in deep basalt (well IDWR-32) or in deep Quaternary unconsolidated sediments below basalt (wells IDWR-22 and IDWR-27) or indicate reservoir surface-water levels (well IDWR-3). Because the relation between groundwater and surface water is not known for all river sections, and because where known, relations can vary seasonally and over time (Dudunake and Zinsser, 2021), contours were generated without respect to surface-water altitudes.

Groundwater flow generally mimics topography, with flow overall down-valley to the south (fig. 14). Consistent with previous reports (Crosthwaite and others, 1970a; Bassick and Jones, 1992), the groundwater gradient flattens, around 15 feet per mile (ft/mi), between Darlington and Moore as the valley widens. The groundwater gradient steepens, around 65 ft/mi, at the mouth of the Big Lost River Valley as the alluvial aquifer transitions to the deeper, basalt aquifer in the eastern Snake River Plain (Stearns and others, 1938; Whitehead, 1992; Ackerman and others, 2010). Despite inconsistencies in the monitoring well network over time, 50-ft contour intervals of the potentiometric surface altitude are similar between April 1968 and April 2019, although the overall potentiometric surface altitude has lowered (that is, lines of equal potential occurred farther upgradient in 2019 than in 1968) from approximately Leslie through Moore (fig. 14).

Time-series data showing potentiometric surface altitude for select wells in the Big Lost River Valley show interannual fluctuations that likely respond to annual precipitation, recharge and irrigation demand, and decadal variations that likely are influenced by climatic cycles (fig. 15). Consistent with the potentiometric surface altitude map (fig. 14), the wells near Mackay (well IDWR-5) and south of Arco (well IDWR-33) show dampened fluctuations (11- and 23-ft maximum change, respectively), whereas wells between the Moore Diversion and Arco (IDWR-16 to IDWR-21) show larger fluctuations (71–109 ft maximum change; fig. 15). Wells south of Darlington show similar responses over time, suggesting that despite the presence of locally confining conditions in some areas, at least the upper 200 ft of the Quaternary unconsolidated sediments unit is hydraulically well connected overall.

Fluctuations in potentiometric surface altitudes also are common seasonally, although the direction of change can vary. For example, 2017 was a wet year for the basin (Natural Resources Conservation Service, 2017; Clark, 2022), and potentiometric surface altitudes in the basin showed increases (that is, lines of equal potential moved downgradient) during April–November 2017 (fig. 16). An increase in the potentiometric surface is particularly apparent between Leslie and Arco, and suggests that natural recharge, incidental recharge from irrigation, and (or) intentional aquifer recharge activities occurring during irrigation season can increase storage in the aquifer over months. Alternately, 2020 was a dry year for the basin (Natural Resources Conservation Service, 2020), and potentiometric surface altitudes declined during April–November 2020 (fig. 17), particularly between Leslie and Arco. These fluctuations suggest that drawdown on the aquifer can be substantial over the course of an irrigation season and that hydrogeologic units in this area are highly transmissive and well-connected. Substantial changes in groundwater storage in this part of the basin are consistent with results from Clark (2022) that indicate incidental recharge is important to the groundwater budget in this area. Notably, potentiometric surface altitudes around Mackay and south of Arco changed relatively little seasonally or interannually; this may be a function of less groundwater demand in these areas and (or) aquifer characteristics such as volume, permeability, and zonation.

Information about vertical hydraulic gradients in the Big Lost River Basin is limited. To address this data gap, a series of shallow (20-, 40-, and 60-ft) well clusters were installed by IDWR in late 2019 (fig. 1); the data therefore have a short period of record but nonetheless are helpful in describing vertical gradients in key locations. At well cluster site IDWR-8, located immediately north of Leslie Butte and near the Darlington Sinks, water consistently was present only in the mid-depth (40-ft) well, which suggests limited groundwater adjacent to Leslie Butte in 2019 and 2020 (fig. 18). Also occurring in the Darlington Sinks, the shallow (20-ft) well at well cluster site IDWR-10 due east of Leslie Butte was dry in late 2019 and 2020, but potentiometric surface altitudes in the 40- and 60-ft wells indicated a large downward gradient, although responses were somewhat decoupled between the two wells (fig. 18). At well cluster sites IDWR-14, IDWR-23, and IDWR-29, potentiometric surface altitudes at each depth are within inches of each other and responses are similar, suggesting no or little vertical gradient in these areas during late 2019 and 2020 and limited attenuation within 60 ft of the land surface (fig. 18). The 20- and 40-ft wells at well cluster site IDWR-24 have nearly identical potentiometric surface altitudes, whereas the 60-ft well potentiometric surface altitude is lower than the shallow wells; this indicates a downward vertical gradient and some attenuation at the lowest well (fig. 18). Notably, the southernmost well cluster sites (IDWR-23, IDWR-24 and IDWR-29) show rapid increases in potentiometric surface altitude in March 2020; this rapid recharge was synchronous with increased streamflow in the Big Lost River below the Moore Diversion (streamgage 13132100), and the resumption of streamflow through the Big Lost River near Arco (streamgage 13132373) and south of Arco (streamgage 13132500) (fig. 18). This response suggests that potentiometric surface altitudes at depths of as much as 60 ft can be rapidly influenced by the resumption of streamflow in the Big Lost River. Thus, although the well cluster data are limited in temporal scale, they support existing understandings of vertical gradients in shallow groundwater (less than 60 ft) in the Big Lost River Valley.

Deeper vertical gradients (greater than 60 ft) and connectivity of water-bearing zones in the southern end of the Big Lost River Valley cannot be fully described with existing data. Wells in this part of the basin are variable in depth (fig. 13) and open to Quaternary unconsolidated sediments and Quaternary basalt rocks hydrogeologic units (figs. 11C, 12), but it is unclear if the multiple water-bearing zones are hydraulically connected. Detailed logging for the deep (760-ft) groundwater-level monitoring well, IDWR-22, noted five discrete water-bearing zones and a “cavernous” basalt that “would take more than 500 [gallons per minute] without difficulty” (Crosthwaite and others, 1970a, p. 74). This well is screened below the basalt and open to Quaternary unconsolidated sediments at 656–661 ft (altitude 4,734–4,729 ft) and 685–690 ft (altitude 4,705–4,700 ft). The potentiometric surface altitude (around 4,800 ft) in well IDWR-22 is lower than the altitude in nearby wells (IDWR-21 around 5,400 ft and IDWR-25 around 5,300 ft, fig. 19; and IDWR-23 around 5,370 ft and IDWR-24 around 5,350 ft, fig. 18). Hydrographs of the potentiometric surface altitude over the last 20 years show similar multiyear fluctuations in the deep well (IDWR-22) and the shallower wells (IDWR-21 and IDWR-25) but the shallow wells show more annual variability (fig. 19). This suggests that both shallow and deep water-bearing zones are influenced by similar long-term impacts such as climate variability but does not resolve local hydraulic connection, particularly given the water-taking basalt flow between the water-bearing zones. Doing an aquifer test in a well screened within 100–200 ft of the land surface near Arco and measuring the response in the deep well (IDWR-22) and nearby well clusters (IDWR-23 and IDWR-24) could help to determine hydraulic connection between the shallow (less than 60 ft), middle (100–200 ft) and deep (760 ft) water-bearing zones.

Hydrogeologic Controls on Groundwater and Surface Water

The three-dimensional hydrogeologic framework model shows general patterns in subsurface hydrogeologic units and subunits that are consistent with the geologic history and support some previous hypotheses regarding controls on gaining and losing river reaches and the extent of confining subunits. Coarse-grained, highly transmissive subunits occur in the subsurface along the main axis of the valley, and coarse-grained subunits occur on major tributary alluvial fans such as Alder Creek (fig. 8C) and Antelope Creek (fig. 6B), and near Chilly where the Big Lost River enters the main valley (fig. 6A). This pattern is consistent with well-sorted and fluvially reworked sediments expected proximal to the Big Lost River and major tributaries. In contrast, more clay subunits and confining conditions occur at the margins of the valleys, consistent with the occurrence of alluvial fans with more poorly sorted sediment and distal deposition of finer-grained fluvial sediment. Groundwater-level monitoring wells between Darlington and Arco (IDWR-11 to IDWR-21) show the greatest fluctuations in groundwater levels (fig. 15), indicating that highly transmissive subunits in this part of the Big Lost River Valley are susceptible to variations in recharge and groundwater use, and suggesting that declining groundwater levels in this part of the valley could affect streamflows (Dudunake and Zinsser, 2021). These findings are consistent with observations of previous researchers (Crosthwaite and others, 1970a; Johnson and others, 1991; Bassick and Jones, 1992), but the three-dimensional hydrogeologic framework model provides a spatial interpretation of the hydrogeologic units that can be used in the development of a potential future groundwater model.

The three-dimensional hydrogeologic framework model also provides insights into patterns of surface-water losses to and gains from groundwater in the Big Lost River. Previous researchers attributed these patterns to changes in aquifer width proportionate to valley width (Livingston, 1931; Stearns and others, 1938; Crosthwaite and others, 1970a), changes in porosity (Livingston, 1931; Crosthwaite and others, 1970b; Rice and Boyd, 2008), and changes in water quantity (Crosthwaite and others, 1970a). The thickness of the Quaternary unconsolidated sediments is critical for understanding spatial variations in aquifer volume but remains poorly defined. Thicknesses interpreted from geophysical surveys range from about 100 ft near Mackay to 5,000 ft between Mackay and Leslie, and average around 2,000 ft (Crosthwaite and others, 1970a), whereas changes in valley width are on the order of several miles. Thus, changes in valley width likely exert primary control on the volume of the Quaternary unconsolidated sediments and subsequently affect river gains and losses, although depth may be more important to aquifer geometry in the narrow valley near Mackay and fault structures may also play a role throughout the basin.

Historical Big Lost River streamflow losses and gains are associated with changes in the alluvial aquifer geometry and porosity. Losses in the Chilly sinks occur in coarse-grained sediments (fig. 6A) where the valley widens as the upper Big Lost River flows into the main valley. Streamflow gains upstream and downstream from Mackay Reservoir occur in a relatively narrow part of the valley with shallow bedrock (fig. 8B). Losses in the Darlington Sinks occur in a wide part of the valley where the aquifer is coarse-grained (figs. 8A, 9A). Structural controls in this fault-bounded basin also may play a role in aquifer geometry and streamflow losses but could not be thoroughly explored in this study. Streamflow gains occur as the valley narrows upstream from the Moore Diversion and near Arco (fig. 1), but other factors also likely influence streamflow in these areas.

Although the valley narrows near the Moore Diversion, Antelope Creek also enters the Big Lost River Valley in this area. Antelope Creek is diverted for irrigation use and lost to the subsurface (Crosthwaite and others, 1970a; Clark, 2022), thereby contributing to aquifer recharge. Potentiometric surface altitudes also are influenced by interannual variations in precipitation and seasonal irrigation demands (figs. 15–16). High potentiometric surface altitudes are associated with large surface-water gains and low potentiometric surface altitudes are associated with surface-water losses (Dudunake and Zinsser, 2021). Thus, aquifer geometry, groundwater recharge by Antelope Creek, and variations in groundwater use all likely affect the occurrence and magnitude of surface-water gains in the Big Lost River above the Moore Diversion.

Historical streamflow gains in the Big Lost River near Arco likely are influenced by local porosity, aquifer zonation, and changing irrigation practices. Historical accounts described a shallow water formation that fluctuated with the river level (Stearns and others, 1938); more recently, Owsley (2013) reported that the timing of streamflow gains coincided with irrigation season. The three-dimensional hydrogeologic framework model shows a laterally continuous clay subunit in the shallow subsurface (about 50–100 ft below the surface) near Arco (figs. 10, 11C), and shallow (20-ft to 60-ft) wells at well cluster site IDWR-24 show a rapid response to changes in surface-water management (that is, resumption of flows in the Big Lost River through Arco; fig. 18). This response suggests that the application of irrigation water and leakage from the Big Lost River and canals can cause rapid recharge of the alluvial aquifer above the confining layer near Arco, in turn producing streamflow gains downstream in the Big Lost River. Diminishment of streamflow gains over time may be the result of decreased potentiometric surface altitude in the uppermost unconfined zone although the hydraulic connection between this and the underlying confined zone is unclear. If the upper zone is perched, decreased groundwater levels could be influenced primarily by localized effects like decreased incidental recharge (Johnson and others, 1991) due to increased efficiency of irrigation methods (such as canal lining and conversion of flood irrigation to center pivots). If the upper zone is hydraulically connected to the underlying zone, decreased potentiometric altitudes could be affected by decreases in local incidental recharge and regional changes in water use and climate that have caused statistically significant groundwater-level declines throughout the Big Lost River Basin (Sukow, 2017). Given the spatial heterogeneity of subunits within the Quaternary unconsolidated sediments, the upper unconfined zone and lower confined zone likely are hydraulically connected at least at the valley scale. Continued monitoring of groundwater levels in IDWR-24 (in the upper unconfined zone) and IDWR-25 (in the lower confined zone) and (or) doing an aquifer test in IDWR-25 would improve understanding of hydraulic connectivity.

Data Gaps

One of the major limitations of this hydrogeologic framework is that wide spacing between long-term groundwater monitoring wells limits detailed interpretation of the potentiometric surface. A synoptic groundwater-level measurement event that captured upwards of 200 wells, similar to the networks in Crosthwaite and others (1970a) and Bassick and Jones (1992), would provide a clearer understanding of current potentiometric surface altitude and flow directions and potentially would illustrate changes since the 1968 and 1991 measurement events. Likewise, additional data collection at the well clusters would continue to yield insights about interannual and seasonal variability in vertical groundwater gradients. Aquifer tests near Arco could improve understanding of hydraulic connectivity between the shallow (60-ft), middle (100–200-ft) and deep (700-ft) water-bearing zones. Continued monitoring and trend analysis of groundwater levels in irrigated and non-irrigated areas would help in understanding whether declines in groundwater levels are continuing, and if so, whether declines are driven by either climate or water use, or both. In the three-dimensional hydrogeologic framework model, an improved geologic representation that incorporates bedrock contacts, geometry, and faults would improve the understanding of recharge at the valley margins. Additional geophysical work to refine the Quaternary unconsolidated sediments unit geometry could help define aquifer volumes.