USGS Open-File Report 2005-1278
By Wesley R. Danskin, Kelly R. McPherson, and Linda R. Woolfenden
Sacramento, California 2006
The San Bernardino area of southern California has complex water-management issues. As an aid to local water managers, this report provides an integrated analysis of the surface-water and ground-water systems, documents ground-water flow and constrained optimization models, and provides seven examples using the models to better understand and manage water resources of the area. As an aid to investigators and water managers in other areas, this report provides an expanded description of constrained optimization techniques and how to use them to better understand the local hydrogeology and to evaluate inter-related water-management problems.
In this report, the hydrology of the San Bernardino area, defined as the Bunker Hill and Lytle Creek basins, is described and quantified for calendar years 1945–98. The major components of the surface-water system are identified, and a routing diagram of flow through these components is provided. Annual surface-water inflow and outflow for the area are tabulated using gaged measurements and estimated values derived from linear-regression equations. Average inflow for the 54-year period (1945–98) was 146,452 acre-feet per year; average outflow was 67,931 acre-feet per year. The probability of exceedance for annual surface-water inflow is calculated using a Log Pearson Type III analysis. Cumulative surface-water inflow and outflow and ground-water-level measurements indicate that the relation between the surface-water system and the ground-water system changed in about 1951, in about 1979, and again in about 1992. Higher ground-water levels prior to 1951 and between 1979 and 1992 induced ground-water discharge to Warm Creek. This discharge was quantified using streamflow measurements and can be estimated for other time periods using ground-water levels from a monitoring well (1S/4W–3Q1) and a logarithmic-regression equation. Annual wastewater discharge from the area is tabulated for the major sewage and power-plant facilities.
The ground-water system consists of a valley-fill aquifer and a much less permeable bedrock aquifer. The valley-fill aquifer, which is the focus of this study, is composed primarily of highly transmissive unconsolidated and poorly-consolidated deposits. The bedrock aquifer is composed of faulted and fractured igneous and metamorphic rock. The valley-fill aquifer is underlain by the bedrock aquifer and and is bounded laterally by the bedrock aquifer and by faults with varying capabilities to transmit ground water. Some underflow occurs across faults in the valley-fill sediment, particularly beneath the Santa Ana River. Essentially no underflow occurs from the surrounding bedrock. Hydrogeologic units were defined for the valley-fill aquifer using driller’s logs, geophysical logs, and hydrographs from multiple-depth piezometers. These units are shown on a detailed hydrogeologic section constructed along Waterman Canyon Creek. A large-scale aquifer test demonstrated the continuity of the hydrogeologic units and their hydraulic properties in the center of the Bunker Hill basin. Gross annual pumpage from the valley-fill aquifer for 1945–98 was compiled from reported and estimated data, and then used to estimate extraction from the upper and lower layers of the valley-fill aquifer and return flow to the upper layer. Annual values of the major components of recharge and discharge for the valley-fill aquifer are calculated for 1945–98. Average recharge occurs primarily from gaged streamflow (67 percent), ungaged mountain-front runoff (9 percent), and pumpage return flow (16 percent); average discharge occurs primarily as pumpage (88 percent).
Computer models, including a ground-water flow model and a constrained optimization model, are described. The ground-water flow model includes the Bunker Hill and Lytle Creek basins and simulates three-dimensional ground-water flow in the valley-fill aquifer using finite-difference techniques. The model consists of an upper layer representing the upper unconfined/semi-confined hydrogeologic unit and a lower layer representing a combination of several lower confined hydrogeologic units. The vertical connection between the model layers is approximated by Darcian flow. The flow-impeding effect of faults within the valley-fill aquifer is simulated by a horizontal flow-barrier package. The model also includes a streamflow-routing package that simulates the interaction of a complex network of streams with the valley-fill aquifer. Calibration of the flow model was for 1945–98.
The constrained optimization model uses linear programming to calculate the minimum quantity of recharge from imported water and pumpage from wells necessary to solve various water-management problems. A description of linear-programming techniques and a simplified example problem are provided. The optimal quantity of recharge or pumpage is determined by their availability and by constraints on ground-water levels and ground-water quality. The response of ground-water levels to recharge and pumpage is calculated by the ground-water flow model. The mathematically optimal solutions derived from the optimization model, in concert with field data and hydrogeologic concepts, can be used to guide water-management decisions.
Selected water-management alternatives for the San Bernardino area were evaluated with the aid of the ground-water flow and constrained optimization models. Seven scenarios were designed to answer specific water-management questions and to demonstrate key hydrogeologic characteristics of the area. A 32-year simulation period, 1999–2030, with annual values of recharge and pumpage, was used for each scenario. The scenarios include: (1) average historical conditions; (2) annually varying historical conditions; (3) additional artificial recharge provided by construction of Seven Oaks Dam; (4) increased ground-water pumpage; (5) optimal pumpage from barrier wells designed to prevent further spread of contamination from the Newmark U.S. Environmental Protection Agency superfund site; (6) optimal pumpage needed to control ground-water levels in an area with potential liquefaction and land subsidence; and (7) optimal recharge and pumpage to control ground-water levels and to prevent migration of the Newmark contamination.
Results of the evaluation include the following conclusions. Additional pumpage in the vicinity of the former marshland is needed to prevent a reoccurrence of dangerously high ground-water levels similar to those experienced in 1945 and 1980. High ground-water levels are a water-management concern because they indicate soil is saturated near land surface and is susceptible to liquefaction during an earthquake. The optimal location of additional pumpage is near Warm Creek in an area of historically rising ground water. The high ground-water levels occur primarily during short periods following abundant natural recharge. About 15,000 acre-feet per year of additional, areally distributed pumpage are needed to control the high ground-water levels.
Demand for water in the San Bernardino area is projected to increase during the next 25 years by as much as 50,000 acre-feet per year. Part of this increased demand can be met by additional pumpage needed to prevent a rise in ground-water levels (15,000 acre-feet per year) and part by increased local supply (3,000 acre-feet per year) resulting from construction of a conservation pool behind Seven Oaks Dam. As much as 70,000 acre-feet of additional pumpage is theoretically available from existing wells using excess pumping capacity. However, additional pumpage greater than about 15,000 acre-feet per year likely will result in a longterm decline in ground-water levels such as occurred during the 1960’s, a decline which prompted land subsidence. To meet future demand for water in excess of about 15,000 acre-feet per year and to prevent a reoccurrence of land subsidence, imported water probably will need to be used, either for direct delivery or for recharge of the ground-water system.
Much of the recharge to the valley-fill aquifer occurs during years with unusually abundant runoff, which occurs on average about once every 5 to 10 years. Maintaining and enhancing capabilities to artificially recharge native runoff are likely to be necessary to meet increased demand for water from the valley-fill aquifer. Since 1945, significant fluctuations in ground-water storage have become common because of the abundant, but highly variable, recharge combined with relatively constant ground-water pumpage. Annual fluctuations in ground-water storage from 50,000 to 100,000 acre-feet are common. Cumulative fluctuations in ground-water storage greater than 500,000 acre-feet in a 10-year period also are common, but this magnitude will not significantly affect the availability of ground water, as long as historic recharge capacities are maintained or enhanced.
Hydraulic control of the Newmark contamination site is unlikely to occur using only the five planned extraction (barrier) wells; another four wells may be needed, each with a capacity of about 3.5 cubic feet per second. Without additional extraction wells, contaminated ground water tends to migrate around the barrier wells, especially to the west. Minimum total pumpage at nine barrier well sites, located across the leading edge of the contamination, needs to be at least 14,000 acre-feet per year, based on results from the optimization model.
Control of maximum and minimum ground-water levels in the vicinity of the former marshland does not require significant additional recharge of imported water, but does require additional pumpage. Additional pumpage from as many as 29 potential production wells located along the proposed eastern and southern extensions of the Baseline feeder pipeline is unlikely to sufficiently control high ground-water levels. Additional pumpage needs to be areally distributed in the vicinity of Warm Creek, just north of the San Jacinto fault. Extraction needs to occur from the highly permeable deposits of the upper water-bearing unit to prevent an upward hydraulic gradient and from the less permeable near-surface deposits that remain saturated even as hydraulic head in the underlying production zone is declining.
The high probability of a major earthquake on either the San Jacinto fault or San Andreas fault in the San Bernardino area makes control of high ground-water levels a pressing economic concern. Significant mitigation of this threat by additional extraction of ground water is possible, especially if a use can be found for the surplus water. Reoccurrence of land subsidence is a continuing concern and can be monitored with multiple-depth piezometers, extensometers, and satellite-borne interferometry, particularly if pumpage near the former marshland is increased. Contamination of the valley-fill aquifer is widespread both areally and vertically. Plans for clean up will be aided by continued mapping of the hydrogeologic units, which strongly influence ground-water flow paths. The large magnitude of proposed ground-water extraction to clean up several areas of contamination suggests that these plans need to be coordinated with plans to prevent liquefaction and land subsidence.
Purpose and Scope
Description of Study Area
Sources of Inflow and Outflow
Ground-Water Discharge into Warm Creek
Description of Valley-Fill Aquifer
Recharge and Natural Discharge
Recharge from Gaged Streamflow
Recharge from Ungaged Runoff
Recharge from Direct Precipitation
Recharge from Local Runoff
Artifical Recharge from Imported Water
Return Flow from Pumpage
Ground-Water Discharge into Warm Creek
Ground-Water Flow Model
Relation to Previous Flow Model
Design and Discretization
Recharge and Discharge
Use and Limitations
Constrained Optimization Model
Representation of the Ground-Water Flow System
Mathematical Form of Objectives and Constraints
|Water Supply and Water Distribution
Use and Limitations
Evaluation of Selected Water-Management Alternatives
Description of Water-Management Alternatives
Selection and Design of Scenarios
Scenario 1: Average Recharge and Discharge, 1999–2030
Scenario 2: Annual Variations in Recharge and Discharge, 1999–2030
Scenario 3: Increased Recharge Made Possible by Seven Oaks Dam, 1999–2030
Scenario 4: Increased Pumpage Using Existing Wells, 1999–2030
Scenario 5: Optimal Hydraulic Containment of Contaminated Ground Water in the Newmark Area
Scenario 6: Optimal Pumpage Using New Wells to Control Ground-Water Levels in the Former Marshland
Suggestions for Future Work
Summary and Conclusions
High Ground-Water Levels
Meeting Future Demand for Water
Document Accessibility: Adobe Systems Incorporated has information about PDFs and the visually impaired. This information provides tools to help make PDF files accessible. These tools convert Adobe PDF documents into HTML or ASCII text, which then can be read by a number of common screen-reading programs that synthesize text as audible speech. In addition, an accessible version of Acrobat Reader 5.0 for Windows (English only), which contains support for screen readers, is available. These tools and the accessible reader may be obtained free from Adobe at Adobe Access.