Evaluation of the Hydrologic System and Selected Water-Management Alternatives in the Owens Valley, California
Download the contents of this page as a pdf file (1.5MB)
Download the entire report as a pdf file (8.3MB)
EVALUATION OF SELECTED WATER-MANAGEMENT ALTERNATIVES
An evaluation of alternative methods of water management involves an appraisal of the present (1988) operating conditions and the physical and social constraints that restrict changes in operations. This evaluation recognizes the social constraints, but focuses on the hydrologic constraints, recognizing that although social constraints might seem to be more encumbering, they often are far less static than the physical constraints presented by precipitation, streamflows, and the aquifer system. Much of the evaluation relies on simulation results from the valleywide ground-water flow model to quantify the likely effects of different management alternatives.
General Water-Management Considerations
Water management of the Owens Valley involves a complex array of conflicting needs and desires. The residents of the Owens Valley need water for local uses such as ranching and domestic supply. Many of the residents desire that water be used for the aesthetic aspects of the valley such as flowing streams and to provide the water needs of native vegetation. The Los Angeles Department of Water and Power, although recognizing these local needs and desires, has continuing needs to export water to Los Angeles. As regional water supplies dwindle and the population of southern California increases, Los Angeles may desire to export additional high-quality water from the Owens Valley. In the difficult task of balancing conflicting needs and desires, the emotional side of water-management issues often tends to take precedence over otherwise purely technical issues.
The goals of water management in the Owens Valley consist of fulfilling both needs and desires. The primary goals include supplying sufficient water for local domestic, ranching, and municipal uses; for native vegetation and aesthetics; and for export to Los Angeles. Secondary goals include mitigation of pumping effects on native vegetation in the immediate area of wells and enhancement of selected areas of the valley. Inherent in achieving these secondary goals, if other water-management practices are continued, is an acceptance of a likely overall decrease in the quantity of native vegetation in other areas of the valley. An ongoing management goal since 1970 has been to decrease consumptive use of water on ranches and lands leased by the Los Angeles Department of Water and Power and to use water more efficiently throughout the valley. Achievement of each of these goals is limited by a variety of considerations that constrain water management in the Owens Valley. The major considerations are described below.
Regional water supplies.—The Owens Valley is part of a much larger network of water supplies, transport, and use. In southern California, water is obtained from a limited number of sources, primarily from northern California, the Colorado River, and the Owens Valley. The use and export of water from the Owens Valley must be viewed within the larger issues of water supply and demand within the arid Southwest, particularly southern California.
Export of surface and ground water.—Water-gathering activities along the aqueduct, primarily north of the Owens Valley in the Mono Basin and the Long Valley, contribute to the total export of water to Los Angeles. A series of reservoirs and ground-water basins along the aqueduct system between the Mono Basin and Los Angeles are used to regulate flow and to store water from one year to the next. Because these storage capacities, in general, are limited, a nearly constant export of water from the Owens Valley is desired. Since 1970, ground-water withdrawals from the Owens Valley have been used to augment surfacewater diversions. In an average-runoff year, some ground water typically is exported; however, in a below-average runoff year, the quantity of ground-water exported out of the valley is increased significantly to make up for the shortage in surface water.
Antecedent conditions from the previous water year affect the quantity of export desired by the Los Angeles Department of Water and Power. If antecedent conditions are dry, then less water is stored in reservoirs and ground-water basins along the aqueduct system, and more water is needed from the Owens Valley. As shown in figure 18, the antecedent conditions in turn affect the quantity of ground water that is pumped. If the preceding year has had average or above-average runoff, then ground-water pumpage is less.
The exportation of water from the Owens Valley to Los Angeles has been the subject of many controversies and lawsuits. Historically, California water law has been interpreted to require maximum beneficial use of water (State of California, 1992). In the early 1900's, beneficial use was nearly synonymous with reclamation of the land for farming and for industrial and municipal use. Since about 1970, the historical beneficial uses of water have been constrained by various environmental issues, such as preservation of phreatophytic vegetation in the Owens Valley and the maintenance of lake levels in the Mono Basin for wildlife habitat. Complying with environmental constraints and satisfying requirements of the California Environmental Quality Act (CEQA) play an increasingly critical role in the export of water from the Owens Valley.
Local use of water. Water use within the Owens Valley includes commitments of water to each of the four major towns, four Indian reservations, three fish hatcheries, and many ranches (figure 1, pl. 3, and table 11; Hollett and others, 1991, figure 5). More recently, additional surface and ground water has been committed to maintain several enhancement and mitigation projects. These relatively high-water-use projects are scattered throughout the valley and provide maintenance of pastureland, wildlife habitat, and riparian vegetation.
Water management in the Owens Valley also has been affected by litigation, particularly the "Hillside Decree" (Los Angeles and Inyo County, 1990a, p. 5—16). This legal injunction required that ground-water pumpage in the Bishop area be used locally within an area extending from north of Bishop to just north of Klondike Lake (figure 11). Within this area, which is referred to as the "Hillside area" or "Bishop Cone," no ground-water pumpage can be exported to other areas of the valley, or out of the valley to Los Angeles. Although the injunction protects the Bishop area, it severely constrains water-management options for the valley as a whole. The Bishop area has the most abundant native water supplies of any area of the valley as indicated by the large discharge of Bishop Creek (average annual discharge is more than 90 ft3/s). Even if local residents, the Inyo County water managers, and the Los Angeles Department of Water and Power should agree on extracting additional ground water from the Bishop area to compensate for reducing ground-water pumpage from another area of the valley, the injunction prevents this reallocation of water.
Hydrologic considerations.—Water management within the Owens Valley also is constrained by physical limitations. Streamflow varies within each year, as well as from year to year. During some highflow periods, not all streamflow can be captured for export or recharged to the ground-water system. During drier periods, minimum flows in the tributary streams may be required to maintain fish populations, and ground-water-recharge operations may be restricted. Some tributary streams, such as Oak Creek, have a large discharge, but a relatively small alluvial fan to be used for ground-water recharge. Other streams, such as Shepherd Creek, have a small discharge and a large alluvial fan.
Antecedent conditions affect the saturated ground-water system. As much as a 3- to 12-month delay occurs in the effect of an above-average runoff year on ground-water levels and discharge rates (well 1T, pl. 1; spring discharge, figure 21). This means that above-average runoff will mitigate some of the adverse effects of a drought that occurs the following year. Ground-water levels beneath the valley floor will tend to rise at the same time as there is a need for additional ground water by native vegetation. The adverse effects of an extended dry period, however, will not be counteracted immediately by an above-average runoff year; the delay in recharge essentially extends the drought for an additional 3 to 12 months.
Antecedent conditions for the unsaturated zone are equally important in water management, as determined during the cooperative vegetation studies (Groeneveld and others, 1986a). In particular, the quantity of water in the unsaturated zone that is carried over from one year to the next is a primary indicator of whether native vegetation will remain healthy (Groeneveld and others, 1986b; Sorenson and others, 1991). As a result of this finding, past water-management practices may need to be altered. For example, ground-water pumpage could be restricted whenever antecedent soil-moisture conditions are too dry.
Simulation of Selected Water-Management Alternatives
The valleywide ground-water flow model was used to evaluate selected water-management alternatives for the Owens Valley. The specific alternatives described in table 14 were chosen after discussion with the technical staffs of Inyo County and the Los Angeles Department of Water and Power. The primary items of concern to valley residents and water managers were the long-term effects of continuing present (1988) operations (alternative 1); the effects of less runoff resulting from long-term climatic cycles or change in climate (alternative 2); the effects of long-term variations in average pumpage (alternative 3); and the ways to mitigate effects of a severe drought and to take advantage of unusually wet conditions (alternative 4). The first three alternatives were simulated with steady-state conditions; the fourth alternative was a 9-year transient simulation.
Because water management in the Owens Valley is exceptionally intricate—involving more than 40 streams, 30 canals, 600 gaging stations, and 200 production wells—the alternatives were designed to simulate general valleywide conditions in order to illustrate how the overall system responds. More detailed site-specific investigations, such as predicting the effects of managing selected wells or streams, are being conducted as part of ongoing water-management activities by Inyo County and the Los Angeles Department of Water and Power.
Alternative 1: Continue 1988 Operations
Alternative 1 addresses the question, "What will happen if present (1988) operations are continued?" That is, what will be the average condition (steady state) of the aquifer system if operations as of 1988 are continued for a long time, probably tens of years? To aid in defining 1988 operations and in evaluating the difference between present and past water-management practices, general water use in the Owens Valley since about 1900 was summarized. Periods with relatively similar characteristics of water use, and therefore relatively similar operation of the surfacewater and ground-water systems, were identified (table 4). Results of this analysis were used in selecting appropriate time periods to calibrate and verify the ground-water flow model, as well as in identifying how 1988 conditions were different from past operations, even those as recent as the early 1980's.
Changes in water-management operations undoubtedly will be made as the hydrologic system and native vegetation of the Owens Valley are more fully understood. An important caveat in viewing the "1988 conditions," as defined in this report, is that the study period was a time of considerable change, or proposed change, in water-management practices. Wide-ranging discussions between Inyo County and the Los Angeles Department of Water and Power typify the process of developing a joint water-management plan for the valley. Possible changes in water management being discussed include discharging a small quantity of water down the lower Owens River to maintain wildlife habitats along the river; installing new wells or using surface-water diversions to provide water for additional enhancement and mitigation sites; and installing new production wells with perforations only in the lower zones of the aquifer system (hydrogeologic unit 3)—not in hydrogeologic unit 1 where effects on the water table and native vegetation are more direct. Additional pumpage for enhancement and mitigation projects may prompt a reduction in pumpage for other uses, including export. Thus, the 1988 conditions as defined in this report likely will evolve over time as understanding of the hydrology of the Owens Valley improves and negotiations between Inyo County and the Los Angeles Department of Water and Power continue. Nevertheless, the 1988 conditions as defined in this report represent the best estimates of future operations based on information available in 1988, and most results based on this definition will not be changed significantly by minor changes in local operations.
Average 1988 conditions in the Owens Valley were defined using a combination of long-term historical data (water years 1935—84) and selected recent data (water years 1985—88) that reflect recent water-management practices (tables 4 and 11). The selection of specific values for the ground-water flow model can be grouped into four categories depending on how static each item has been.
Long-term average relations.—A long-term average period, water years 1935—84, was used to define average-runoff conditions. The relations of runoff to ground-water recharge for tributary streams (figure 13) and for ungaged areas (table 11), both of which were used to simulate ground-water conditions during water years 1963—88, were assumed to remain valid for future conditions.
Long-term constant values.—Underflow and recharge from precipitation were held constant as they had been during simulation of water years 1963—88 (table 11).
Recent constant values.—Recharge from irrigated areas was the same as the constant values used during simulation of water years 1970—88. This period reflects the change in water use that occurred about 1970 (table 4). The maximum evapotranspiration rate was the same as that used to simulate water years 1978—88.
Recent average values.—A recent period (water years 1985, 1986, and 1988) was selected to represent average conditions for those items that were recently added or changed. The selection of these specific years included an evaluation of the probability of different percent-runoff years (figure 12) and of the effect of antecedent conditions on pumpage (figure 18). The selected period includes a wet water year (1986), an average water year (1985), and a dry water year (1988). This period was used to determine recharge from miscellaneous operations, recharge from water use on Indian lands, recharge from canals and ditches, and discharge from pumping. Pumpage from enhancement and mitigation wells, which were being installed during water years 1985—88, was planned to provide a virtually constant supply regardless of runoff conditions (R.G. Wilson, Los Angeles Department of Water and Power, oral commun., 1988). As a result, average pumpage for enhancement and mitigation wells was defined as the values for water year 1988. An important assumption regarding pumpage was that average pumpage for enhancement and mitigation projects was in addition to average pumpage for export.
These values of recharge and discharge defined for average 1988 conditions were used in the calibrated ground-water flow model to determine a steady-state solution of simulated heads, recharge, and discharge (table 11). The simulated change in water-table altitude between water year 1984 (figure 19 and pl. 1) and 1988 steady-state conditions is shown in figure 26. Water year 1984 was chosen for comparison because ground-water levels were relatively high over most of the basin, most springs had resumed some discharge, and the ground-water basin was nearly as "full" as it had been prior to 1970 (Hollett and others, 1991). A comparison of water-budget components for the 1988 steady-state period with those for water years 1963—69 and water years 1970—84 is shown in figure 27. These three periods represent the main changes in the Owens Valley hydrologic system (table 4) since the early 1900's.
On the basis of the model simulations, changes in the 1984 water-table altitude and in recharge and discharge will occur if the 1988 operating conditions, as defined above, are continued. Most of the predicted water-table changes occur in the alluvial fan areas, particularly in the Taboose—Aberdeen and Independence areas (sections C—C' and D—D', figure 26). A large difference also is predicted in the Laws area and near Big Pine. The valley floor exhibits somewhat less change in the water table, as expected because of hydraulic buffers. Decreases in evapotranspiration and changes in the ground-water flow rate to or from the river-aqueduct system and the lower Owens River tend to minimize fluctuations in heads. On the valley floor, changes are characterized primarily by differences in recharge and discharge, as indicated by the simulated decrease in evapotranspiration (figure 27 and table 11). Interestingly, total ground-water inflow is greater in the 1988 simulation (figure 27) because a lower water table induces additional recharge from surface-water features. On the basis of observations made during calibration and verification of the ground-water flow model and during testing of water-management alternative 4, described later, reaching new steady-state conditions may require as much as from 10 to 20 years of similar operations (figure 21 and pl. 1).
Although some uncertainty is present in the assumptions of this simulated steady-state condition, the general conclusions are not altered by slightly different assumptions about specific recharge or discharge components. The main difference between the 1988 steady-state values of recharge and discharge and previous values is the marked increase in ground-water pumpage, especially pumpage from enhancement and mitigation wells (table 11). An additional difference is that the long-term average runoff (100 percent of average runoff) assumed for the 1988 steady-state period is somewhat lower than that during water years 1963—84 (107 percent of average runoff).
The large increase in pumpage that occurred during water years 1970—84 was offset partially by a decrease in springflow, which helped to minimize changes in the water-table altitude. By 1984, total spring discharge was significantly less than it was prior to 1970, and the buffering effect on the water table was largely gone (figure 21 and table 11). The further increase in pumpage assumed for the 1988 steady-state period combined with the slight decrease in average runoff resulted in a further decline of the water table in comparison with 1984 conditions (figure 26).
During the initial part of this study, the 1984 water year was perceived to represent a return to relatively average conditions—water levels had returned to near the 1970 levels in most parts of the valley. However, this condition was highly contingent on the large runoff quantities of the late 1970's and early 1980's (figure 12 and table 7) and the relatively lower pumpage (figure 18). In contrast, the 1988 steady-state conditions assume long-term average runoff and a much higher quantity of average pumpage (table 15), albeit for various uses other than export out of the valley. If these assumptions remain valid, then the basin, as of 1988, is in the midst of another transition, one prompted largely by the increased pumpage from the enhancement and mitigation wells (table 11).
In general, the water-table decline is greatest in the alluvial fans, and least in the areas of seeps, drains, and surface-water bodies (hydraulic buffers) that are in contact with the ground-water system. The significant water-table decline in the alluvial fans will have no effect on overlying vegetation because the water table is many tens or hundreds of feet beneath the land surface of the fans, except in highly faulted areas, such as near Red Mountain or immediately north of the Alabama Hills (figures 3 and 14). The water-table decline in the alluvial fans, however, will reduce the ground-water flow rate toward the valley floor, which in turn will reduce ground-water discharge, primarily transpiration from native vegetation on the valley floor. Plant stress similar to that observed by Sorenson and others (1991) can be expected to occur in areas near the toes of the fans and in parts of the valley floor near Big Pine and Laws if 1988 conditions are continued. It is important to note that there may be only a slight change in water-table altitude beneath these plants as a result of changes in plant transpiration and changes in flow to nearby seeps, drains, and surface-water bodies. This is a characteristic response of a ground-water system modulated by hydraulic buffers.
Changes in water management can offset some of the adverse effects implied in figure 26. Increased recharge of surface water during wet years, especially in or upgradient from areas likely to have decreased transpiration by native vegetation, would help to minimize a long-term reduction in native vegetation on the valley floor. In contrast to other nearby basins, however, the recharged water is not retained for an extended period of time (Danskin, 1990). The relatively high transmissivity of sand and gravel deposits and the exceptionally high transmissivity of volcanic materials tend to dissipate recharged water relatively fast (within a few years). In order to successfully mitigate the effects implied in figure 26, recharge needs to be increased above historical averages (figures 21 and 27; tables 10 and 11) and pumpage probably needs to be decreased in selected areas where recharge cannot be increased.
Alternative 2: Continue 1988 Operations with Long-Term Changes in Climate
Alternative 2 addresses the question, "What if climatic cycles or long-term climatic change cause average basinwide runoff to be slightly less, or more?" The time period, water years 1935—84, that was used to analyze the surface-water system and develop runoff-recharge relations (figure 13 and table 11), despite being 50 years long, may not be representative of average runoff conditions for the next 25 to 50 years. Normal variations in climate could produce a change of a few percent in long-term average runoff. In addition, possible climatic change caused by human activities, although a highly controversial and largely unresearched topic (Danskin, 1990), is a recent global concern. The specific effects of induced climatic change are unknown; however, changes in the average annual runoff in basins in the Southwestern United States, including the Owens Valley, have been suggested (Revelle and Waggoner, 1983; Lins and others, 1988; Lettenmaier and Sheer, 1991). It also is possible that an induced climatic change may alter runoff conditions even more within individual years (Wigley and Jones, 1985; Moss and Lins, 1989), but this highly speculative aspect was not addressed in this study.
Simulation of alternative 2 used the 1988 steady-state conditions (alternative 1) with variations of plus or minus 10 percent in the average percent of runoff. This relatively small deviation reflects the generally well-known and stable condition of long-term average runoff. Also, the runoff-recharge relations are likely to remain valid for small changes in runoff. Analysis of a greater change in average runoff, which might result from more substantial changes in climate, would require a reinterpretation of precipitation patterns and amounts (figure 7) and streamflow relations (figure 13). In the present analysis, the quantities of ground-water recharge affected by the change in percent runoff include recharge from tributary streams, from mountain-front runoff between tributary streams, and from local runoff from bedrock outcrops within the valley fill (table 10). Recharge from precipitation was assumed to occur primarily during extremely wet years and was not changed. All other quantities of ground-water recharge and discharge were the same as those defined for alternative 1.
Results from alternative 2 are shown in figure 28 for representative sections across the valley. Sections B—B', C—C', D—D', and E—E' in figure 28 correspond closely with hydrogeologic sections B—B', D—D', E—E', and F—F', respectively, of Hollett and others (1991, pl. 1 and 2). Also shown on the sections in figure 28 are simulated water tables for water year 1984 and for average runoff conditions (1988 steady-state simulation, figure 26) and the range in simulated water tables for water years 1963—88. Only the simulated heads for the upper model layer (water table) are shown because they are most important in predicting effects on native vegetation; simulated heads for the lower model layer show a similar pattern, but with some vertical offset from heads for the upper model layer.
Most obvious in figure 28 is the difference between simulated steady-state conditions for 1988 (100 percent runoff) and simulated conditions for water years 1963—88. By comparison, variations of 10 percent in average basinwide runoff produced less difference in the water table in most areas of the basin, except along the western edge of the valley from Independence to Lone Pine (sections D—D' and E—E' in figure 28). As expected, water-table differences resulting from variations in runoff are most pronounced in the recharge areas, particularly under the western alluvial fans. The river-aqueduct system, the lower Owens River, and native vegetation act as hydraulic buffers and help to reduce water-table changes near the valley floor.
Variations in runoff have less effect in the Bishop and the Laws areas than in the Taboose and the Independence areas. In the Lone Pine area, the marked change in the water table west of the Alabama Hills is largely a result of low transmissivities associated with the thin alluvial fan deposits and probably is not a major concern. The Alabama Hills effectively isolates the fan area to the west from the valley floor and related native vegetation to the east. In the Taboose and the Independence areas, however, the change in the water table beneath the alluvial fans translates to a significant decrease in the rate of ground-water movement toward the valley floor and a consequent decrease in evapotranspiration from the valley floor. Long-term monitoring of ground-water levels beneath the alluvial fans and valley floor and of evapotranspiration by native vegetation on the valley floor would identify such a long-term trend. In the Lone Pine area just west of the Owens River, the simulated water table for 1988 is higher than that for 1984 because of additional recharge from a new enhancement and mitigation project started in 1988.
Also of importance in figure 28 is a change in the river-aqueduct system in section C—C'. Simulation of 1988 steady-state conditions and variations in runoff of 10 percent indicate that under these conditions the river-aqueduct loses water to the Taboose—Aberdeen well field to the west. This change in flow direction could be verified with detailed water-level monitoring and water-quality sampling of the river-aqueduct and aquifer systems.
One management technique to minimize the effect of a long-term decrease in runoff is to increase the recharge from streams that have relatively low loss rates (figure 13 and table 11). These streams include Bishop, Big Pine, Birch, Shepherd, and Lone Pine Creeks. Indeed, on the basis of results from alternative 1, increasing the recharge from streams is indicated even if long-term runoff does not decrease. Because past management efforts have pursued this option, it is unclear how much more water can be recharged on the alluvial fans in the critical areas of Taboose and Independence. An alternative management technique is to selectively decrease pumpage in sensitive areas.
The effects of a slightly different long-term average runoff, such as might occur as a result of climatic variations in precipitation, are less than those induced by human water-management decisions. Long-term variations in climate that produce slightly different annual quantities of runoff, assuming that stream-loss relations (figure 13) continue to be valid, will not markedly affect the valley.
Alternative 3: Increase or Decrease Long-Term Average Pumpage
Alternative 3 addresses the question, "What will happen if average pumpage is increased or decreased from 1988 steady-state conditions?" One of the few aspects of the hydrologic system of the Owens Valley that can be altered readily is the quantity of pumpage. Over the past 20 years, pumpage has increased (figure 17; tables 10 and 15) and has been the primary cause of change in the Owens Valley aquifer system during that time. Alternative 3 simulates scaling average annual basinwide pumpage up or down.
The design of alternative 3 was similar to that of alternative 2. Steady-state conditions for 1988 were assumed for all ground-water recharge and discharge, except pumpage. The value of pumpage at each well was scaled to 25, 50, 75, 100, and 125 percent of the 1988 steady-state value (table 9). The 100-percent pumpage simulation is identical to the 100-percent runoff simulation (alternative 2), which is identical to the 1988 steady-state simulation (alternative 1).
Although future pumpage in the valley is likely to be somewhat different from past pumpage because old wells occasionally are replaced with new wells, this difference is probably minimal for steady-state conditions, such as those simulated in alternative 3. Replacement wells usually are right next to the original well and are designed to extract water directly from hydrogeologic unit 3 (lower model layer) in order to delay the effects of pumpage on the water table. Given sufficient time, however, these effects will be transmitted to hydrogeologic unit 1 (upper model layer). The change in well design is recognized as an important management technique for shorter time periods, but it will become less valuable over time as the entire aquifer system equilibrates. Also, the valleywide ground-water flow model, as demonstrated during calibration, is relatively insensitive to withdrawing a greater percentage of pumpage from the lower layer.
Results from simulating alternative 3 are shown in figure 29 for the same sections shown in figure 28. The variations in pumpage are shown in 25-percent increments of the assumed 1988 steady-state pumpage. The increments are arbitrary, but they are within the confidence limits of the calibration model. Also shown is the simulated water table for water year 1984 in order to aid in correlating with figure 28 and plate 1.
As was true of figure 28, the most notable feature shown in figure 29 is the significant difference between the simulated water table for water year 1984 and that for 1988 steady-state conditions (100 percent pumpage) (figure 26). This difference illustrates the large quantity of pumpage assumed for 1988 steady-state conditions—a quantity that combines average pumpage for export and new pumpage for enhancement and mitigation projects. In order to approximate the 1984 levels, average pumpage needs to be decreased significantly, to about 50 percent of the value assumed for the 1988 steady-state conditions, or to about 75,000 acre-ft/yr (figure 29 and table 15).
The general linearity of pumpage effects is shown by an approximately even change in water-table altitude for each 25-percent increment. This feature is to be expected for a model using constant transmissivities and operating within the linear range of headdependent recharge and discharge relations (table 13). A marked change in water-table altitude, however, is visible in the Taboose area (section C—C' in figure 29) for the 125-percent increment. This result indicates that the simulated water table in the surrounding area has dropped below the zone of linearity of the head-dependent evapotranspiration and stream-recharge relations (refer to McDonald and Harbaugh, 1988, p. 10—3 and 6—9). When this occurs, the hydraulic buffering action is no longer effective, and the water table declines at a more rapid rate.
Different parts of the basin respond very differently to reductions in pumpage. The greatest change in the water table occurs near pumped wells, near bedrock boundaries, and away from headdependent sources of recharge, such as the river-aqueduct system. As a result, a large change in the water table occurs on the west side of the valley, and relatively little change occurs on the east side of the valley across the Owens Valley Fault where there are few pumped wells (figures 14 and 17). As noted in the discussion of alternative 2, wide variations in water-table altitude beneath the alluvial fans (such as those shown in section D—D in figure 28) do not affect overlying vegetation but do change the hydraulic gradient toward the discharge areas, and thereby decrease evapotranspiration rates for native vegetation some distance away on the valley floor.
Changes in the water table in the Bishop Basin occur mostly in the Laws area (section A—A' in figure 29). Because head-dependent recharge along the eastern edge of the basin near Laws is minimal, no additional source of water is available except ground-water storage, and the simulated water table rises and falls dramatically with changes in pumpage. A similar response has been observed in measured ground-water levels (pl. 1). If some sources of recharge in the Laws area, such as the McNally Canals (figures 11 and 29), act in a head-dependent way rather than as defined quantities of recharge as simulated in the model, then the use of head-dependent relations (table 13) to simulate these features will lessen the simulated fluctuations in the water table near Laws (figure 29). Gaging of discharge in the canals and ditches, in addition to monitoring local ground-water levels, will aid in better defining these surface-water/ground-water relations.
The simulated water table in the area just south of Bishop is as unaffected by changes in pumpage as by changes in recharge (compare figures 28 and 29). This lack of response results primarily because the area historically has had little recharge or pumpage, and, therefore, little was simulated in the model. A similarly static response was found in measured ground-water levels for well 335T (pl. 1) during water years 1963—88, a period of large variations in pumpage and recharge.
A decrease in evapotranspiration from the valley floor in the area south of Bishop may occur, however, even when the water table changes as little as 2 to 3 ft (Sorenson and others, 1991, p. G33). This decrease in evapotranspiration coincides with a decrease in the biomass of the native vegetation, as noted by Griepentrog and Groeneveld (1981, map 2) and by Sorenson and others (1991, figure 24). Therefore, caution is required in interpreting simulation results even in areas that appear to have a minimal change in watertable altitude.
In the Owens Lake Basin, the primary effects of simulated changes in pumpage occur between Taboose and Independence Creeks (figure 29). There is an indication in the Taboose area, as well as in the Laws area (section A—A' in figure 29), that pumpage in excess of the 1988 steady-state quantity may cause hydraulic separation of the Owens River from the adjacent water table, creating a partially saturated zone beneath the river. This separation as simulated in the model causes a precipitous lowering of the water table, as discussed previously and as shown by the 125-percent increment.
In summary, results of model simulations suggest that the water table will continue to decline for some time if recharge and pumpage remain at the assumed 1988 steady-state values. This water-table decline will result in a decrease in evapotranspiration and a decrease in the biomass of native vegetation. Results of simulations indicate that to maintain the water table at an altitude similar to that of 1984, total pumpage needs to be about 75,000 acre-ft/yr, or about 50 percent of the assumed 1988 steady-state value.
Alternative 4: Manage Periodic Variations in Runoff and Pumpage
Alternative 4 addresses the question, "How can a sequence of dry and wet years be managed?" For example, which areas of the valley are likely to be affected most by a severe drought, which least, and how fast do the different areas recover? Which areas need help in recovering to pre-drought conditions? The Owens Valley hydrologic system historically has cycled between droughts and periods of abundant water (table 7). Because of the multiplicity of and constant change in water-management operations, such as during water years 1970—88, it is difficult to identify the effects of a typical cycle using historical data. Simulation of alternative 4 attempts to clarify these effects with a simple, but typical, management scenario.
A schematic of the 9-year transient simulation used for alternative 4 is shown in figure 30. The 9-year simulation period has similarities to drought, average-runoff, and above-average-runoff conditions experienced during the 1970's and 80's. Initial conditions for alternative 4 were assumed to be alternative 1 (1988 steady-state) conditions. The first 3-year period (I) represents drought conditions and simulates 70 percent of average runoff and maximum pumpage. Maximum pumpage is defined as the maximum annual pumpage recorded at each well during water years 1985—88; maximum pumpage for enhancement and mitigation wells is the value recorded for water year 1988 (table 11). The implicit water-management goal during the first 3-year period is to maximize export of ground water to compensate for decreased export of surface water. The second 3-year simulation period (II) represents a return to average conditions and simulates 100 percent of average runoff and the same value of pumpage as the initial (1988 steady-state) conditions. The management question during the second period is, "How fast does the system return to normal?" The third 3-year simulation period (III) represents wet conditions and simulates 130 percent of average runoff and the same average pumpage as during the second 3-year period. Actual pumpage during a wet cycle most likely will be somewhat less than average, particularly after a couple of wet years (figure 18). This decrease, however, is poorly quantified for future conditions and was not incorporated in the simulation. Results from the third period identify areas of the valley in which the simulated heads have not recovered to initial conditions even after 3 years of average conditions and 3 years of wet conditions. Specific values of recharge and discharge are given in table 11.
The simulated change in water-table altitude at the end of each 3-year period (drought, average, and wet) with respect to initial conditions is shown in figures 31, 32, and 33, respectively. Because no site-specific water-management techniques were incorporated in the simulation, the results identify those stressed areas of the valley that require additional monitoring and possibly additional manipulations of ground-water recharge and discharge.
The areas of the valley that show the greatest effects at the end of a 3-year drought marked by lesser runoff and greater pumpage are identified in figure 31. Clearly, the effect of drought is widespread. Much of the decline in the water table occurs beneath the alluvial fans and volcanic deposits, as in other simulations (figures 23, 26, 28, and 29). Areas with the most dramatic changes are those in abundant recharge areas (Bishop and Oak Creeks). Other areas with significant water-table decline are near the well fields (Laws, Big Pine, Taboose—Aberdeen, and Independence—Oak) (figure 17). As determined during sensitivity analysis of the ground-water flow model, the effect of lower runoff near well fields is minimal in comparison with the effect of nearby pumping.
Some areas on the valley floor that have a simulated decline in water-table altitude greater than 10 ft are areas that are covered with native vegetation identified as susceptible to stress from pumping (R.H. Rawson, Los Angeles Department of Water and Power, written commun., 1988; Sorenson and others, 1991). The significant water-table decline in these areas decreases evapotranspiration, prompts native vegetation to drop leaves, and reduces total biomass on the valley floor. Some species, such as rabbitbrush (Sorenson and others, 1991, p. G35) may die during a 3-year drought if the plants cannot grow additional roots deep enough and fast enough.
Areas of the valley floor that are isolated from recharge and pumping effects, such as between Bishop and Big Pine and east of the Owens River, have a simulated decline in water-table altitude of only a foot or two. Although some decrease in evapotranspiration is likely, the effects on native vegetation are much less than effects near recharge areas and well fields. Because these isolated areas have few monitoring wells, simulation results need to be viewed cautiously.
The Taboose—Aberdeen area exhibits a broad areal change in water-table altitude, broader than in most other areas of the valley. The many springs in the area historically acted as hydraulic buffers and dampened the effects of pumping on water-table fluctuations. That capacity, however, now is largely gone (figures 17 and 21), and, with changes in pumpage, the water-table fluctuations are greater (pl. 1). Neither the Owens Valley Fault nor the unnamed fault near the aqueduct (figure 14) is an effective barrier to ground-water flow in this part of the Owens Lake Basin. Cones of depression in the water table created by pumping in well fields (figure 17) propagate unimpeded eastward across the valley.
In the southern part of the Bishop Basin, cones of depression are transmitted even more effectively through hydrogeologic unit 3 to the east side of the valley because of the presence of the relatively impermeable blue-green clay (Hollett and others, 1991, pl. 1). This thick clay layer effectively restricts the vertical flow of water from hydrogeologic unit 1 to hydrogeologic unit 3 in the center of the valley. Release of water from hydrogeologic unit 3 is derived mostly from elastic expansion of water and compression of the aquifer, which results in a storage coefficient that is much smaller than specific yield. As a result of these conditions, the cone of depression expands to cover a large area. The highly transmissive sand and gravel beds in hydrogeologic unit 3 aid in propagating the cone of depression horizontally. On the east side of the valley, the alluvial fan deposits have a greater vertical hydraulic conductivity than does the blue-green clay, and ground water can readily flow from hydrogeologic unit 1 to hydrogeologic unit 3. In this way, the water table along the east side of the valley responds to pumping on the west side. The net result is that most of the nearby area north and south of the Tinemaha Reservoir exhibits a significant decline in the simulated water table. Associated adverse effects on nearby native vegetation are likely, particularly in areas distant from surface-water features, which are a source of recharge.
Historical water-management operations in the Owens Valley have tended to create feast or famine conditions for native vegetation. For example, the recent (1984) rise in the water table near Laws and Independence (figure 23) resulted from an abundance of recharge in these areas, primarily as a result of water-spreading activities by the Los Angeles Department of Water and Power (pls. 1 and 3; table 11), and from a temporary reduction in pumpage (figure 17). Native vegetation responds to increased water availability by increasing leaf growth or plant density, which results in a commensurate increase in evapotranspiration (Groeneveld and others, 1987). A subsequent period of drought and increased pumpage, such as during water years 1987—88 (pl. 1) or as simulated during the first 3-year period of alternative 4 (figures 30 and 31), results in a declining water table and a decrease in plant leaf area and evapotranspiration. The declining water table then prompts a water-management decision to decrease pumpage and implement water-spreading efforts to increase recharge when water is again abundant. This cyclic pattern of response by the aquifer system and native vegetation to alternating drought and high runoff, accentuated by water-management decisions that increase pumpage during droughts and then increase artificial recharge during periods of high runoff, typifies a more highly managed Owens Valley.
One attribute of a more highly managed aquifer system is that native vegetation will be less evenly distributed. The natural flow of the aquifer system tends to smooth out ground-water levels, recharge, and discharge. Human changes in the aquifer system tend to focus recharge and discharge into smaller areas. As the valley becomes more controlled, it will become more pod-like, with pods of thriving vegetation near enhancement and mitigation projects and pods of highly stressed vegetation near wells. In between, native vegetation will be using less water than it had been using prior to the increase in water development.
A water-management goal for most ground-water basins is the same as for a surface-water reservoir. Empty the reservoir when water is scarce; fill it when water is plentiful. The paradox in managing the Owens Valley is that if the water table beneath the valley floor fluctuates too much, native vegetation is adversely affected. Therefore, the reservoir must be kept virtually full.
Alternative water-management techniques to lessen the effect of pumping on the water table and nearby native vegetation are limited in many ways, as discussed in the section "General Water-Management Considerations." From a long-term, valleywide perspective, the water table is affected most by the quantity of water pumped, not by the particular location of pumping in the valley (figure 26). Nevertheless, locations with pumped wells have greater fluctuations in the water table and a greater likelihood of having native vegetation adversely affected by water-table fluctuations (compare figures 17 and 31). Locating pumping on alluvial fans away from the valley floor will lessen the decline of the water table near sensitive vegetation. Pumping from high on the western alluvial fans, in particular in areas of abundant recharge, will lessen the immediate effects on the valley floor.
However, past experiences of drilling on the western alluvial fans (well 1T, pl. 1) showed that installation of wells has been difficult or nearly impossible because of massive rock and boulders (M.L. Blevins, Los Angeles Department of Water and Power, oral commun., 1987). Also, transmissivities of the alluvial fans and related well yields are significantly less than in transition-zone or volcanic deposits (figure 15). Electrical usage is higher in order to lift water the greater distance to land surface. Similar difficulties might be encountered in installing new wells on the eastern alluvial fans. In addition, the eastern alluvial fans are areas of limited recharge and, possibly, poorer quality ground water with a higher concentration of dissolved solids.
Pumping from high on the Bishop Creek alluvial fan (Bishop Cone), although now limited by the Hillside Decree, probably would produce minimal effects on the valley floor, especially if pumping were limited to short-term supply during a drought. This broad, gently sloping fan is characterized by abundant recharge from Bishop Creek. The fan has additional recharge potential through the use of spreading basins, and it might be easier to drill through this fan than through the steep, rocky fans near Independence.
Much of the valley floor in the Bishop and Big Pine areas is urban or irrigated land that is not affected by a decline in the water table. Additional pumping from within these areas probably will have less effect on native vegetation than pumping from other areas of the valley floor.
Pumping only from lower zones of the aquifer system, beneath hydrogeologic unit 1, reduces the immediate decline of the water table. The amount of this reduction is unknown, but it could be approximated using detailed, site-specific ground-water flow models of individual well fields, or possibly by field testing a single pumped well surrounded by several, multiple-depth monitoring wells (Driscoll, 1986, p. 719—728). The benefit of pumping from lower zones, however, decreases the longer the wells are pumped continuously. Hydrogeologic boundary conditions and vertical leakage through hydrogeologic unit 2 and alluvial fan deposits eventually will transmit the effects of pumping from lower zones to hydrogeologic unit 1, lowering the water table and decreasing evapotranspiration from areas where the water table is within 15 ft of land surface (table 5).
Differences in the simulated water-table altitude following 3 years of drought and 3 years of average conditions are shown in figure 32. The areas of residual decline in the water table are similar to those in figure 31, but the magnitude is less. Areas where the decline is greater than 10 ft indicate locations in the valley that need careful monitoring of the water table, soil-moisture zone, and native vegetation. Results from simulating alternative 4 also suggest that monitoring the effects of a drought need to be continued for several years following the end of the drought—much longer than previously thought necessary.
Differences in the simulated water-table altitude following 3 years of drought, 3 years of average conditions, and 3 years of 130-percent runoff are shown in figure 33. As expected, recharge areas show a considerable rise in the water table, as do areas of focused artificial recharge, such as near Laws and Independence (figure 33 and pl. 3). Somewhat surprising, however, is that 6 years after a drought and immediately following 3 years of above-average runoff, the water table in many areas of the valley still shows signs of the drought and coincident pumpage. Minor residual drawdown is present over most of the valley floor, and an isolated area of declines greater than 10 ft still is present beneath the alluvial fans east of Big Pine. This result demonstrates the slowness of recovery in areas away from abundant recharge.
The period of recovery for the water table is much longer than was hypothesized at the beginning of the modeling studies. This characteristic of the aquifer system, however, agrees well with the tentative conclusion that the aquifer system and native vegetation were still in transition in the mid-1980's from the effects of increased pumping in the early 1970's and the drought conditions in 1976—77.
The water-table decline simulated in alternative 4 can be reduced by focusing artificial-recharge efforts in areas of greatest decline and concentrated pumping (figures 17 and 31). Localized recharge efforts may need to be continued for as long as 6 years after the end of a 3-year drought in order to compensate for the decline in water table. Areas of abundant water and lush vegetation induced by artificial recharge likely will become areas of stressed vegetation in future drought conditions (compare figures 31 and 33).
Because of the limitations associated with the valleywide ground-water flow model and the unique characteristics of a particular drought, ongoing monitoring of the aquifer system, soil-moisture zone, and native vegetation needs to be continued, particularly in areas simulated in alternative 4 as having water-table declines greater than 10 ft (figures 31, 32, and 33).
Optimal Operation of Well Fields
An extensive body of literature deals with the general topic of mathematical optimization of physical systems (Gorelick, 1983; Rogers and Fiering, 1986), and a few applications have been made to combined surface-water and ground-water systems (Young and Bredehoeft, 1972; Bredehoeft and Young, 1970, 1983; Danskin and Gorelick, 1985). Although use of these techniques was proposed initially as a promising method of evaluating water management in the Owens Valley, detailed appraisals during the 6-year study identified several numerical limitations. The mathematical dimensions (m × n matrix) required by a realistic optimization model for the Owens Valley are very large. There are more than 40 streams, 9 well fields, 200 production wells, 800 observation wells, and 600 surface-water gaging stations—as well as a multitude of decision points in the basin, such as whether or not to divert a stream. Also, the optimization problem is moderately nonlinear as a result of the piecewise-linear relations used to approximate some recharge and discharge components in the ground-water flow model (table 13). The large dimensionality and nonlinearities would require considerable computer time to solve even a relatively simple problem in a mathematically rigorous way. As computer capabilities increase and costs diminish, a basinwide optimization study may prove to be more tractable. The approach presented in this report uses the basics of the mathematical optimization techniques and could serve as the foundation of a simple optimization model.
The actual operation of individual well fields is a complex and iterative process, dependent on many factors—including those general concerns presented in the section entitled "General Water-Management Considerations," as well as day-to-day concerns of mechanical efficiency, repair and maintenance, and personnel requirements. Optimal operation probably involves meeting several different objectives, which makes the mathematical problem even more complex and makes a simple, instructive version of the water-management system difficult to define.
For this evaluation, however, optimal operation of well fields was defined in a semi-quantitative way to be the most pumpage for the least adverse effect on native vegetation. The ground-water flow model was used to determine the effect of pumpage from each well field. The model response, referred to in optimization literature as a "response function," is the change in head, recharge, and discharge in response to a defined increase in pumpage. A unit increase in pumpage produces a "unit response." Those well fields that produce the least adverse effects on native vegetation (least water-table decline under vegetation that relies on ground water) are considered the optimal well fields to use. Well fields with a greater water-table decline are less desirable, or less optimal.
Two similar analyses were done to determine the effect of pumpage from each well field. Each analysis involved simulating the response to pumpage at individual well fields. The simulation timeframe was 1 year with constant stresses. Initial conditions for each simulation were the 1988 steady-state conditions (alternative 1). To simplify the analysis, the Independence—Oak, the Symmes—Shepherd, and the Bairs—George well fields (figure 17) were grouped together and are referred to as the "Independence south" well field. The Lone Pine well field was not included in the first analysis because of its limited capacity, the presence near the well field of relatively fine-grained and less transmissive aquifer materials (figures 15 and 16), and the abundance of nearby en echelon faults that limit production (figure 4).
The first analysis involved increasing pumpage at each well field (table 11 and 15) by 10,000 acre-ft/yr more than the 1988 steady-state simulation (alternative 1). Pumpage for an individual well was increased in proportion to its 1988 steadystate value (table 11). After 1 year of simulation, the decline in water-table altitude was noted and is shown in figure 34. From this analysis, the well field having the greatest effect on native vegetation is readily discernible as the one producing the greatest watertable decline under the largest area of native vegetation dependent on the water table. This technique of using a unit stress (10,000 acre-ft/yr of pumpage) to observe the "unit response" (drawdown surrounding each well field) is a dominant feature in most hydraulic optimization techniques (Gorelick, 1983). For comparison, the combined effect of 10,000 acre-ft of additional pumpage at each of the six well fields is shown in figure 34D.
The approximate area of native vegetation dependent on the water table is indicated by the boundary of alluvial fans (compare figures 4 and 34). Detailed mapping by the Los Angeles Department of Water and Power (R.H. Rawson, written commun., 1988) identified a few isolated parts of the valley floor, primarily east of the lower Owens River, where native vegetation may not be dependent on ground water. Vegetation in these areas of the valley floor presumably is isolated from the effects of pumpage
All well fields produce approximately the same areal effect (figure 34). Cones of depression in the water table extend to the edge of the Owens Valley aquifer system, even within a single year. The cones of depression extend somewhat farther up and down the valley because of boundary effects along the edges of the valley and the linearity of hydrogeologic units (figure 5). All well fields except the Bishop produce greater than 5 ft of drawdown beneath the valley floor, but the magnitude of drawdown is somewhat more concentrated in well fields that have fewer, higher production wells, such as the Big Pine and the Thibaut—Sawmill well fields. The combined pumpage of an additional 60,000 acre-ft/yr (figure 34D) indicates that cones of depression from individual well fields merge and extend over most of the valley.
The most surprising result of this first "unit response" analysis is the similarity of response from each of the well fields. No obviously better place to extract water is evident despite the spatial differences in hydraulic properties of the aquifer system, the distribution of wells, the locations of surface-water features, or the presence of faults that retard ground-water movement. The Bishop well field probably produces the least effect on native vegetation, but water from this well field cannot be used for export, as stipulated by the Hillside Decree. The optimal management of well fields favors producing a large volume of water from a small area, such as from the Thibaut—Sawmill well field. The resulting drawdown is greater, but the area of significant drawdown is more localized.
Extraction of water from the large alluvial fan near Bishop in lieu of other areas of the valley is a favorable management alternative, as discussed in the preceding section (p. 122), except for the restrictions imposed by the Hillside Decree. Vegetation covering most of the fan is not dependent on ground water because the water table is tens or hundreds of feet beneath land surface. The present distribution of wells (figure 17) indicates that the fan is not used extensively for production. Increasing production uniformly (figure 34B) produces a small area with greater than 5 ft of drawdown near the edge of the fan. By distributing production farther up the fan, the area of greatest drawdown will be reduced in size, and any increased drawdown will occur beneath vegetation that does not subsist on ground water. An important caveat, however, is that sustained pumping from alluvial fan areas eventually decreases ground-water flow rates toward the valley floor area and will cause some change in native vegetation, even if the water table beneath the valley floor remains relatively unaffected. Although pumping from other alluvial fans will yield similar beneficial results, the benefits will be limited by problems of lesser recharge and technical difficulties in installing wells.
The second analysis involved increasing 1988 steady-state pumpage at each well field to the maximum annual value measured at each well during water years 1985—88 (table 11 and 15). This analysis is designed to optimally distribute present pumping capacity in excess of the 1988 steady-state quantity (alternative 1). Water-table decline after the 1-year simulation is shown in figure 35. For some well fields, the increase is approximately 10,000 acre-ft/yr and the drawdown in figure 35 resembles that in figure 34.
Most of the pumpage from the Bishop and the Thibaut—Sawmill well fields is used for ongoing commitments of water (figure 17 and table 11), and little pumping capacity above the 1988 steady-state values is available (table 15). Some flexibility exists in managing pumpage from Laws, Big Pine, Taboose, and Independence south well fields. None of these well fields, however, creates a pattern of drawdown that is markedly better with respect to native vegetation than the others (figures 34 and 35). An ideal pattern from the simulation is zero drawdown beneath native vegetation on the valley floor. The area surrounding the Big Pine well field, because of the large area of irrigated lands and sparsely vegetated volcanic flows, is probably least affected and closest to the ideal. The Laws well field, because of its great distance from a large alluvial fan that acts as a storage reservoir, seems to affect the largest area of the valley floor and is the poorest choice. Consequently, mitigation measures need to be more intensive in that area—as they have been in recent years—than in other parts of the valley.
The simulated water-table decline after 1 year of maximum pumpage at the six well fields, in comparison with 1988 steady-state conditions, is shown in figure 35D. As with the simulation of unit responses (figure 34D), the cones of depression from the individual well fields overlap, but not to a significant degree. Pumping from the small Lone Pine well field, which has limited extra capacity (table 15), has a minimal effect on the rest of the valley (figure 35E).
One feature that is interesting to note is an unaffected area south of Bishop. This area, near Collins Road and vegetation sites C and D (figure 2), shows no decline in the simulated water table after 1 year of maximum pumpage (figure 35E). Coincidentally, native vegetation in that area was observed to remain greener than in other parts of the valley during 1982—88, a period of wide variations in precipitation, recharge, and pumpage. This observation, paired with the simulated results presented in figures 34D, 35D, and 35E, helps to confirm the reasonableness of the ground-water flow model in that part of the valley. The primary reasons the area remains unaffected by changes elsewhere in the valley are the lack of nearby pumping (figure 17) and the effectiveness of hydraulic buffering of the water table by native vegetation and the Owens River.
In summary, optimal water management of the well fields—with the objective of minimizing declines in the water table—is relatively insensitive to pumpage from a specific well field. The areal extent of greatest drawdown in the water table is similar for each of the six well fields, both from the standpoint of installing new production wells (figure 34) and of using existing capacity (figure 35). If pumpage can be increased at one or two well fields for only a single year or part of a year, then drawdown and any adverse effects on native vegetation will be restricted to a small, more manageable area. Rotating pumpage from one well field to another may facilitate this result, and may be an optimal way to manage the well fields during times of below-average runoff.
Reliability of Results
The reliability of this evaluation of water management in the Owens Valley depends on three critical assumptions: first, that the aquifer system and native vegetation are conceptualized correctly; second, that the aquifer system is numerically approximated with only minor, recognized errors; and third, that the selected water-management alternatives are a realistic representation of possible future conditions.
The conceptualization of the aquifer system and native vegetation was the focus of related studies by Groeneveld and others (1985, 1986a); Hutchison (1986b); Dileanis and Groeneveld (1989); Sorenson and others (1989, 1991), Duell (1990), and Hollett and others (1991). Although not all aspects of the aquifer system and native vegetation are well understood, the important role of the aquifer system in providing water for the long-term health of native vegetation on the valley floor is well documented. The primary difficulty in predicting the response of native vegetation to a change in water availability is that a decline in the water table does not always result in an immediate adverse effect on native vegetation (Sorenson and others, 1991, p. G35). For example, if precipitation on the valley floor is well above average, native vegetation can survive, even prosper, for 1 to 3 years with no water supplied via capillarity from hydrogeologic unit 1.
Because precipitation on the valley floor and valleywide runoff from the surrounding mountains are not well correlated, it is possible to have precipitation on the valley floor and thus an increase in soil moisture, which promotes additional plant growth, and at the same time have reduced runoff from the mountains, which prompts an increase in pumpage and results in a lowering of the water table. Under these conditions, the native vegetation remains healthy, but the water table declines. However, if the extra pumpage continues through a period of below-average precipitation on the valley floor, then plants will begin dropping leaves to conserve water and the overall health of native vegetation is jeopardized. During the evaluation of different water-management alternatives, this variability of response was recognized, but an assumption was made that the plants were not aided by a short-term increase in precipitation.
The numerical approximation of the aquifer system was made using a ground-water flow model that incorporates most of the major concepts of the aquifer system as well as the use of ground water by native vegetation. The limitations of ground-water flow models in general, and the valleywide model in particular, are discussed extensively in a previous section, entitled "Use, Limitations, and Future Revisions." The reliability of the ground-water flow model is affected most by those limitations. For example, two areas of the basin—west of Bishop and near Lone Pine—are either poorly understood or poorly simulated. Results in these areas are less reliable than those in other parts of the basin. During development of the valleywide model, several other ground-water flow models of parts of the Owens Valley were developed by a number of different organizations and individual researchers (figure 2 and table 2). Each of the models tends to show similar results. Although it is possible that all the models are incorrect, this uniformity gives additional credibility to the modeling approach and results.
Use of the ground-water flow model to identify areas where native vegetation is likely to be affected adversely by pumping is based on the assumption that a hydraulic stress (decline in water-table altitude) equates to a vegetative stress (decrease in biomass). As discussed above, this is not always true. For longer periods of time, however, such as the period of steady-state conditions simulated in three of the four alternatives evaluated, the assumption becomes more reliable. The benefits of a short-term increase in precipitation on the valley floor are outweighed by long-term water requirements for transpiration. More reliable results might be produced by using another type of model that explicitly incorporates vegetative growth, precipitation, and use of ground water and is linked to a valley-wide ground-water flow model. For the present study, however, such a model was deemed to be numerically too large and to have too many poorly quantified parameters.
Changes in simulated recharge and discharge in the valleywide ground-water flow model that were required to evaluate different water-management alternatives were well within the range of values used during calibration and verification of the model. This minimal modification of the model increases the reliability of results—particularly, if the results are viewed in a general, semi-quantitative way. In analyzing the different water-management alternatives, the simulated drawdown seems to be somewhat greater than what might actually occur. A simulated 30-ft decline might represent an actual decline of 20 ft; a simulated 10-ft decline, an actual decline of 6 ft; and so forth. The reason for the deviation is not known, but it may result from greater delayed drainage of hydrogeologic unit 1 or more effective action of hydraulic buffers, such as evapotranspiration. Because the ground-water flow model uses generalized model zones of aquifer properties and localized recharge and discharge, the spatial pattern and relative magnitude of drawdown probably are more reliable than the specific value of drawdown.
The selection of water-management alternatives was based on what was considered a realistic representation of possible future conditions. Because of the extremely wide-ranging nature of negotiations between Inyo County and the Los Angeles Department of Water and Power in designing a water-management plan for the Owens Valley, the definition of realistic is somewhat subjective. For example, the assumption that 1988 steady-state pumpage is the sum of average historical pumpage and new enhancement and mitigation pumpage was an arbitrary choice reflecting one possible agreement. The choice of some lesser quantity of pumpage would have been an equally valid assumption. Choice of a greater quantity of pumpage did not seem politically plausible. The use of 0, 25, 50, 75, 100, and 125 percent of 1988 steady-state pumpage for alternative 3 brackets the range of what was deemed realistic.
Many of the choices in defining future conditions were much less subjective. Several were based on long-term hydrologic conditions, such as runoff for water years 1935—84 or land use for water years 1970—88. Values of recharge and discharge based on past long-term conditions are probably reliable indicators of future long-term conditions.
Only a few choices were based on recent changes in water management, primarily the addition of enhancement and mitigation pumpage and related recharge. Both hydrologically and politically, the recently altered recharge and discharge are much less certain than long-term values. Additional changes in water management, such as reestablishing the lower Owens River as a perennial stream or establishing alfalfa fields near well fields, seem likely and will affect localized areas of the valley. The evolving water management of the Owens Valley prompted by the requirement of a court-accepted EIR and joint water-management plan for the valley creates the greatest uncertainty in future conditions and is probably the most important caveat in assessing the reliability of results presented in this report.
Potential Changes in Operation
The following is a summary of potential changes in water-management operations designed to protect native vegetation as well as to provide water for export to Los Angeles. The options involve changes in recharge, changes in pumpage, and changes in mitigation measures.
Increase tributary stream recharge.—An increase in recharge from tributary streams is limited by the timing and quantity of runoff from the Sierra Nevada. Some tributary streams have a lower loss rate (figure 13 and table 9) than others, depending on characteristics of the surficial deposits and length of the stream channel. Estimates of evapotranspiration for vegetation along tributary stream channels indicate that most of the loss actually seeps into the ground and recharges the aquifer system. An increase in the recharge rate of selected streams, therefore, can compensate for an increase in ground-water pumpage, depending on the timing of recharge and pumping.
Most tributary streamflow that does not seep into the ground is exported out of the valley. Increasing the recharge rate in years of average or below-average runoff probably is not productive, as a reduction in streamflow means that additional ground water likely will be pumped from other parts of the valley to make up the difference. If the total quantity of water exported in average-runoff years could be reduced, then increasing recharge from some tributary streams, in particular Taboose and Bishop Creeks, can provide additional ground water in future years. A further increase in recharge for these or other tributary streams may be possible through modifications of the diversion operations near the base of the mountains or use of a different configuration of diversion channels on the alluvial fans. Increasing recharge during years of above-average runoff may be advantageous, but this general operating policy has been in effect since the early 1970's. Also, some of the recharge, particularly during wet periods, will be lost to increased evapotranspiration and gain of water by the river-aqueduct system.
Increase artificial recharge on the valley floor.—Artificial recharge of surface water on the valley floor is being done in the Bishop and the Laws areas, and to a lesser extent, in the Big Pine area (table 11 and pl. 3). The purpose of the recharge is to replenish ground-water storage that has been depleted by pumping and to enhance recovery of the water table in order to protect native vegetation. Expansion of these efforts may be possible to further reduce the adverse effects of pumping on native vegetation.
Artificial recharge in most parts of the valley floor is limited by the presence of fine-grained deposits and the horizontal layering of the aquifer system (figures 5 and 14). Although unlined surface-water features are an important source of local recharge, direct irrigation of the native vegetation has been discounted as an option because of likely problems with salinity and disruption of the soil horizon (D.P. Groeneveld, Inyo County Water Department, oral commun., 1987). Direct recharge through wells, however, may be a water-management option—particularly, as new wells are installed with perforations only in the lower zones. Use of recharge wells can help repressurize the production zone after large extractions have been made, such as during a drought, or whenever extra surface water is available. Repressurizing a confined zone results in a moderate increase in ground-water storage—much less than if the zone is unconfined—and an important recovery of ground-water levels and gradients. Evaluation of the likely changes in ground-water quality resulting from direct recharge of surface water will require additional water-quality data.
Recharge surface water on the east side of the valley.—Artificial-recharge efforts on the east side of the valley during periods of above-average runoff will provide some additional storage of ground water. Because natural runoff on the east side of the valley is scant, recharge efforts probably will require diversion of surface water from the river-aqueduct system into those areas. As indicated by simulations using the valleywide ground-water flow model (figures 34 and 35), drawdown cones from well fields reach to the bedrock sides of the valley. Recharge along the sides of the valley, even the east side, will help to reduce the effects of pumping. However, recharged water that is not captured by pumping may eventually seep into the river-aqueduct system or the lower Owens River, and may induce more growth of vegetation between the recharge and discharge points.
Recharge on the east side of the Bishop Basin, particularly east of the Big Pine well field, might help minimize the areal effects of pumping in the Big Pine area, as well as provide some additional ground-water storage, particularly beneath the blue-green clay. In contrast, recharge east of the Owens Valley Fault in the Owens Lake Basin has little effect on the western well fields. The Owens Valley Fault tends to channel recharge water down the east side of the basin, allowing only small quantities of flow westward across the fault.
Extract ground water from the Bishop Creek alluvial fan.—Extraction of water in the Owens Valley is a highly charged topic that does not lend itself to purely scientific assessments. Nevertheless, one of the premier places to extract water and have little effect on native vegetation seems to be near Bishop, particularly the Bishop Creek alluvial fan (Bishop Cone). The great depth to water over much of the fan, abundance of recharge, prevalence of urban land and irrigated vegetation, and large number of canals and ditches criss-crossing the fan make it an area with higher recharge and production potential and fewer adverse effects on native vegetation than most other areas of the valley. Uncertainties about the aquifer system west of Bishop do not alter this conclusion. However, additional understanding of how the Bishop Tuff, the Coyote Warp, and valley-fill faults (figure 4) affect the aquifer system will be most helpful in planning any changes in water management.
Extract ground water from the Owens Lake area.—Additional extraction of ground water from the area south of the Alabama Hills and surrounding the Owens Lake may be possible. Although drilling and lithologic data are sparse for that part of the valley, depositional concepts indicate that the alluvial fan deposits along the western side of the basin probably grade into a narrow band of moderately transmissive transition-zone deposits. Extraction of a significant quantity of ground water near the Owens Lake probably will require additional recharge in order to minimize the migration of poorer quality (higher dissolved-solids concentration) ground water from beneath the lakebed toward the production wells. South of the valleywide model area, Cottonwood Creek (Hollett and others, 1991, figure 16) has a greater discharge than any other tributary stream in the Owens Valley except Bishop and Big Pine Creeks. If recharge from Cottonwood Creek could be increased, especially by utilizing its large alluvial fan, then additional ground-water extractions from that area might increase water-management flexibility. Ground-water pumpage in that area likely will affect a narrow band of native vegetation near the springline and edge of the lakebed (figures 1 and 3). Additional drilling, aquifer tests, water-level and water-quality monitoring, and possibly small-scale simulation studies will be required to further document and evaluate this option.
Extract ground water from the east side of the Owens Valley.—Extraction from the east side of the Owens Valley is not as efficient as extraction from the west side. Aquifer materials on the east side are finer and probably less transmissive. If the depositional models are correct for that side of the basin, then a narrow band of transition-zone deposits should be present as suggested on plate 2. The most transmissive deposits and greatest quantity of transition-zone deposits probably are near the alluvial fans of Waucoba and Mazourka Canyons (figure 4). Because of the apparent symmetry of the basin and aquifer materials, the pattern and extent of drawdown from pumping on the east side of the valley probably will be similar to that of drawdown from pumping on the west side of the valley (figure 34).
A major limitation of pumpage from the east side of the basin is the meager quantity of natural recharge. Without additional recharge near proposed wells, ground-water storage will be depleted rapidly. This depletion is accentuated by the restriction to ground-water flow caused by the Owens Valley Fault. Both the quality of ground water along the eastern side of the basin and the probable changes in ground-water quality resulting from recharge and extraction in that area are unknown. Despite these considerable limitations, extraction from the east side of the valley should be hydrogeologically feasible and might offer some flexibility in future water management.
Extract ground water from the Lone Pine area.—The Lone Pine area is characterized by finer-grained materials, lower transmissivities, more en echelon faulting, and possibly poorer water quality than in many other parts of the basin. These characteristics alone do not make it a particularly desirable place to develop additional well production. A more complete assessment requires a better understanding and simulation of ground-water flow in that part of the valley.
Pump from selected well fields.—A shift of pumping to selected well fields may provide protection for native vegetation in other areas. For example, the prevalence of irrigated lands near the Big Pine well field makes widespread, adverse effects on native vegetation less likely than at other well fields such as the Taboose—Aberdeen or the Independence—Oak (figure 17). Also, localized pumping from highly transmissive volcanic deposits at the Thibaut—Sawmill well field restricts the areal extent of the adverse effects on native vegetation (figure 34). Extraction from similar well fields or parts of the valley will require less mitigation for native vegetation than will extraction at other locations.
Rotate pumpage among well fields.—As indicated in figures 25, 34, and 35, rotational pumpage may have some advantage over continual extraction from a single well field. A key to the health of native vegetation is the water availability within the rooting zone of the plants (Groeneveld, 1986; Sorenson and others, 1991). Cycling pumpage from one well field to another can enable the water table near the wells to recover and soil moisture in the overlying unsaturated zone to be replenished via capillarity. Although recovery of the water table occurs fairly rapidly, replenishment of soil moisture is much slower (Groeneveld and others, 1986a, 1986b). Field data and modeling results suggest that a few weeks or months are needed to replenish soil moisture (Groeneveld and others, 1986a, p. 86; Welch, 1988). Although the valleywide model can give some semi-quantitative guidance, water management using rotational pumpage needs to rely on monitoring of multiple-depth wells and soil-moisture sites in the vicinity of well fields, and possibly on results from unsaturated-saturated flow models.
Seal upper perforations of existing wells.—Sealing of perforations adjacent to the unconfined zone in existing production wells was investigated during this study and was found to be marginally successful. Continuation of this effort will limit the immediate effect of production wells on the unconfined zone and the related adverse effects on nearby native vegetation (figure 25). Sealing of abandoned wells limits the short-circuiting of flow that occurs through a casing that is open to multiple strata. Installation of new production wells with perforations only in the lower zones (hydrogeologic unit 3) of the aquifer system will reduce the effects of pumping on the water table and native vegetation. Adverse effects on native vegetation, however, still will occur if a large quantity of water is pumped for an extended period of time, possibly 1 to 3 years (figure 25; Sorenson and others, 1991, p. G35).
Utilize other ground-water basins.—Additional recharge and extraction facilities in other basins along the route of the dual-aqueduct system might provide additional flexibility in the water management of the Owens Valley (Danskin, 1990). For example, the Indian Wells Valley, just south of the Owens Valley, is having ground-water storage depletion and related ground-water-quality problems (Berenbrock and Martin, 1991; Berenbrock and Schroeder, 1994) that might be mitigated by additional recharge. During periods of above-average runoff in the Sierra Nevada or during a period of lesser demand in Los Angeles for water from the Owens Valley, surplus water could be conveyed via the Los Angeles Aqueduct to the Indian Wells Valley, and recharged there. Conversely, during drier periods, ground-water production from the Indian Wells Valley could be increased to augment flow in the Los Angeles Aqueduct, thereby reducing the quantity of water needed from the Owens Valley. Other desert basins between the Owens Valley and Los Angeles, such as in the Mojave Desert, the Antelope Valley, and the Coachella Valley, have a large potential for ground-waterstorage (California Department of Water Resources, 1964, 1967a; the Antelope Valley—East Kern Water Agency, 1965; Reichard and Meadows, 1992). These basins, which are connected to the extensive system of water delivery in southern California (California Department of Water Resources, 1987), could provide additional water-banking opportunities.
Questions about Owens Valley Hydrogeology? Please contact Wes Danskin (email or address). 619.225.6132
California Water Science Center Home Page
