Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah, Updated to Conditions through 2011, with New Projections and Groundwater Management Simulations
In cooperation with the Southern Utah Valley Municipal Water Association
Open-File Report 2013–1171
U.S. Department of the Interior U.S. Geological Survey Cover. Southern Utah Valley, view northwest from Santaquin area. Photo credit: Hailee Cox, U.S. Geological Survey. Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah, Updated to Conditions through 2011, with New Projections and Groundwater Management Simulations
By Lynette E. Brooks
In cooperation with the Southern Utah Valley Municipal Water Association
Open-File Report 2013–1171
U.S. Department of the Interior U.S. Geological Survey U.S. Department of the Interior SALLY JEWELL, Secretary U.S. Geological Survey Suzette M. Kimball, Acting Director
U.S. Geological Survey, Reston, Virginia: 2013
For more information on the USGS—the Federal source for science about the Earth, its natural and living resources, natural hazards, and the environment, visit http://www.usgs.gov or call 1–888–ASK–USGS. For an overview of USGS information products, including maps, imagery, and publications, visit http://www.usgs.gov/pubprod. To order this and other USGS information products, visit http://store.usgs.gov.
Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. Although this information product, for the most part, is in the public domain, it also may contain copyrighted materials as noted in the text. Permission to reproduce copyrighted items must be secured from the copyright owner.
Suggested citation: Brooks, L.E, 2013, Evaluation of the groundwater flow model for southern Utah and Goshen Valleys, Utah, updated to conditions through 2011, with new projections and groundwater management simulations: U.S. Geological Survey Open-File Report 2013–1171, 35 p. iii
Contents
Abstract...... 1 Introduction...... 1 Purpose and Scope...... 2 Description of Study Area...... 2 Groundwater Hydrology ...... 2 Description of the 1995 Southern Utah and Goshen Valleys Groundwater Flow Model ...... 5 Recharge...... 5 Discharge...... 6 Update of the Model to Conditions Through 2011...... 9 Groundwater Withdrawals ...... 9 Recharge...... 10 Performance of the Updated Model...... 13 Appropriate Uses of the Model...... 17 Model Projections 2012 to 2050...... 17 Example Groundwater Management Simulations...... 22 Potential Revisions and New Data to Improve the Updated Model ...... 29 Summary ...... 29 References ...... 30 Appendix 1: Supplemental Groundwater Management Simulation...... 31 Decision Variables and Constraints...... 31 Results ...... 31
Figures
1. Map showing location of southern Utah and Goshen Valleys study area, Utah ...... 3 2. Generalized block diagram showing the basin-fill groundwater system in southern Utah and Goshen Valleys, Utah ...... 4 3. Numerical model grid, southern Utah and Goshen Valleys, Utah ...... 5 4. Map showing average recharge simulated using the Recharge Package in the numerical model from 1949 to 1990, southern Utah and Goshen Valleys, Utah...... 7 5. Map showing location of head-dependent boundaries in the numerical model, southern Utah and Goshen Valleys, Utah ...... 8 6. Graph showing irrigation, municipal, and industrial groundwater withdrawals from 1949 to 2011, southern Utah and Goshen Valleys, Utah ...... 10 7. Map showing location and magnitude of groundwater withdrawals in the numerical model, southern Utah and Goshen Valleys, Utah...... 11 8. Graph showing simulated groundwater budget components for A, southern Utah Valley and B, Goshen Valley, Utah ...... 12 9. Map showing location and model layer of groundwater-level measurements, southern Utah and Goshen Valleys, Utah ...... 13 10. Graphs showing measured and simulated water levels, 1949 to 2012, southern Utah and Goshen Valleys, Utah ...... 15 iv
11. Graph showing withdrawals simulated in projections 2, 3, and 4, and in groundwater management example 1, 2012 to 2050, southern Utah and Goshen Valleys, Utah ...... 17 12. Map showing areas and amounts of projected managed aquifer recharge, southern Utah Valley, Utah ...... 18 13. Graphs showing projected simulated water-level altitude in selected wells, southern Utah Valley, Utah ...... 19 14. Graphs showing projected simulated selected groundwater-budget components, southern Utah Valley, Utah ...... 20 15. Maps showing drawdown compared to base projection for southern Utah and Goshen Valleys, Utah, in A, projection 2; B, projection 3; and C, projection 4 ...... 21 16. Maps showing location of well decision variables and managed aquifer recharge decision variables for optimization examples, southern Utah and Goshen Valleys, Utah ...... 23 17. Maps showing locations in southern Utah and Goshen Valleys, Utah, of A, areas in which to maximize water levels and spring discharge and B, water-level and spring-discharge constraints for optimization examples ...... 24 18. Maps showing drawdown compared to base projection for southern Utah and Goshen Valleys, Utah, in A, projection 3; B, optimization example 1; and C, optimization example 2 ...... 25 19. Graphs showing selected simulated groundwater-budget components for optimization examples, southern Utah Valley, Utah ...... 26 20. Graphs showing amounts of groundwater withdrawals by area for optimization examples, southern Utah and Goshen Valleys, Utah ...... 27 21. Maps showing drawdown compared to base projection for southern Utah and Goshen Valleys, Utah, in A, projection 4 and B, optimization example 3 ...... 28 A1-1. Maps showing location in southern Utah and Goshen Valleys, Utah, of A, managed aquifer recharge decision variables and B, water-level and spring- discharge constraints ...... 32 A1-2. Map showing drawdown compared to base projection, southern Utah and Goshen Valleys, Utah ...... 33 A1-3. Graphs showing selected simulated groundwater-budget components, southern Utah Valley, Utah ...... 34 A1-4. Graph showing amounts of optimized managed aquifer recharge by area, southern Utah Valley, Utah ...... 35
Tables
1. Conceptual groundwater budget for the main groundwater system in the basin fill, southern Utah and Goshen Valleys, Utah, 1990 ...... 4 2. Comparison of simulated stream-aquifer interactions for varying flow rates in the original model, southern Utah Valley, Utah ...... 9 3. Description of water applied to irrigated areas, southern Utah and Goshen Valleys, Utah ...... 12 4. Summary statistics of model fit to 25 water-level observations, southern Utah and Goshen Valleys, Utah ...... 14 v
Conversion Factors and Datums
Inch/Pound to SI Multiply By To obtain Length foot (ft) 0.3048 meter (m) mile (mi) 1.609 kilometer (km) Area acre 4,047 square meter (m2) acre 0.4047 hectare (ha) square mile (mi2) 2.590 square kilometer (km2) Volume acre-foot (acre-ft) 1,233 cubic meter (m3) acre-foot (acre-ft) 0.001233 cubic hectometer (hm3) Flow rate acre-foot per year (acre-ft/yr) 1,233 cubic meter per year (m3/yr) acre-foot per year (acre-ft/yr) 0.001233 cubic hectometer per year (hm3/yr) foot per year (ft/yr) 0.3048 meter per year (m/yr) gallon per minute (gal/min) 0.06309 liter per second (L/s)
Vertical coordinate information is referenced to the National Geodetic Vertical Datum of 1929 (NGVD 29). Horizontal coordinate information is referenced to the North American Datum of 1983 (NAD 83). Altitude, as used in this report, refers to distance above the vertical datum. vi
Wells by the Cadastral System of Land Subdivision The well-numbering system used in Utah is based on the Cadastral system of land subdivision. The well-numbering system is familiar to most water users in Utah, and the well number shows the location of the well by quadrant, township, range, section, and position within the section. Well numbers for most of the State are derived from the Salt Lake Base Line and Meridian.
Sections within a township Tracts within a section R. 3 E. Section 18
3456 2 1 b a 7 8 9 10 11 12 b a b a T. dc 131415161718 c d 8 S. 22212019 23 24
252627282930 c d
34333231 3635
6 miles 1 mile 9.7 kilometers 1.6 kilometers (D-8-3)18bdc-1
B A Salt Lake Base Line
Salt Lake City
T. 8 S., R. 3 E. Meridian C D Lake Salt
Surface-Water Sites— Downstream Order and Station Number Since October 1, 1950, hydrologic-station records in U.S. Geological Survey reports have been listed in order of downstream direction along the main stream. All stations on a tributary entering upstream from a main-stream station are listed before that station. A station on a tributary entering between two main-stream stations is listed between those stations. As an added means of identification, each hydrologic station and partial-record station has been assigned a station number. These station numbers are in the same downstream order used in this report. In assigning a station number, no distinction is made between partial-record stations and other stations; therefore, the station number for a partial-record station indicates downstream-order position in a list composed of both types of stations. Gaps are consecutive. The complete 8-digit (or 10-digit) number for each station such as 09004100, which appears just to the left of the station name, includes a 2-digit part number “09” plus the 6-digit (or 8-digit) downstream order number “004100.” In areas of high station density, an additional two digits may be added to the station identification number to yield a 10-digit number. The stations are numbered in downstream order as described above between stations of consecutive 8-digit numbers. Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah, Updated to Conditions through 2011, with New Projections and Groundwater Management Simulations
By Lynette E. Brooks
Abstract withdrawals and managed aquifer recharge in southern Utah The U.S. Geological Survey (USGS), in cooperation Valley, but rapid changes in withdrawals and increasing with- with the Southern Utah Valley Municipal Water Association, drawals dramatically may reduce the accuracy of the predicted updated an existing USGS model of southern Utah and Gos- water-level and groundwater-budget changes. The model hen Valleys for hydrologic and climatic conditions from 1991 should not be used for projections in Goshen Valley until to 2011 and used the model for projection and groundwater additional withdrawal and discharge data are collected and the management simulations. All model files used in the transient model is recalibrated if necessary. Model projections indicate model were updated to be compatible with MODFLOW-2005 large drawdowns of up to 400 feet and complete cessation of and with the additional stress periods. The well and recharge natural discharge in some areas with potential future increases files had the most extensive changes. Discharge to pumping in water use. Simulated managed aquifer recharge counteracts wells in southern Utah and Goshen Valleys was estimated and those effects. Groundwater management examples indicate simulated on an annual basis from 1991 to 2011. Recharge that drawdown could be less, and discharge at selected springs estimates for 1991 to 2011 were included in the updated model could be greater, with optimized groundwater withdrawals and by using precipitation, streamflow, canal diversions, and irri- managed aquifer recharge than without optimization. Recali- gation groundwater withdrawals for each year. The model was bration to more recent stresses and seasonal stress periods, and evaluated to determine how well it simulates groundwater con- collection of new withdrawal, stream, land-use, and discharge ditions during recent increased withdrawals and drought, and data could improve the model fit to water-level changes and to determine if the model is adequate for use in future plan- the accuracy of predictions. ning. In southern Utah Valley, the magnitude and direction of annual water-level fluctuation simulated by the updated model reasonably match measured water-level changes, but they do Introduction not simulate as much decline as was measured in some loca- tions from 2000 to 2002. Both the rapid increase in ground- Groundwater is the primary source of drinking water in water withdrawals and the total groundwater withdrawals in southern Utah Valley and the primary source of drinking water southern Utah Valley during this period exceed the variations and irrigation water in Goshen Valley. Reported municipal and and magnitudes simulated during the 1949 to 1990 calibration industrial withdrawals (Utah Division of Water Rights, 2012a) period. It is possible that hydraulic properties may be locally in southern Utah Valley doubled from 1999 to 2000, but did incorrect or that changes, such as land use or irrigation diver- not continue increasing. The average amount of withdraw- sions, occurred that are not simulated. In the northern part als from 2001 to 2011 was about the same as withdrawals in of Goshen Valley, simulated water-level changes reasonably 2000. Increased withdrawals and below-average precipita- match measured changes. Farther south, however, simulated tion and streamflow from 2000 to 2003 caused water levels declines are much less than measured declines. Land-use in many wells in the area to decline to their lowest recorded changes indicate that groundwater withdrawals in Goshen levels by 2005 (Burden and others, 2012, fig. 14). Water-level Valley are possibly greater than estimated and simulated. It is declines could affect the ability to withdraw water from wells, also possible that irrigation methods, amount of diversions, and could affect discharge to springs, drains, streams, flowing or other factors have changed that are not simulated or that wells, and Utah Lake. aquifer properties are incorrectly simulated. The model can be used for projections about the effects of future groundwater 2 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
The U.S. Geological Survey (USGS), in cooperation The population of incorporated areas of southern Utah with the Southern Utah Valley Municipal Water Associa- Valley increased from about 47,000 to 112,000 from 1991 to tion, updated an existing USGS model of southern Utah and 2011. Land is being converted from agricultural to urban and Goshen Valleys (Brooks and Stolp, 1995) for hydrologic and suburban use to accommodate this growth, and the locations climatic conditions from 1991 through 2011. The existing and amounts of water use and applied water may be chang- model simulated conditions from 1949 to 1990. Numerical ing. The changes in land use and water use have the potential models are constructed on the basis of available information to affect groundwater levels; recharge; discharge to springs, and data; if new data become available, testing the model with drains, and streams; and water quality. To incorporate land-use those data will result in a better understanding of the model changes in the numerical model and to evaluate the possible and the groundwater system (Konikow and Bredehoft, 1992). effects of land-use and associated water-use changes would The model was evaluated to determine how well it simulates have required additional data collection and a recalibration of groundwater conditions during recent increased withdrawals the model, which were beyond the scope of this project. and drought, and to determine if the model is adequate for use in future planning. Groundwater flow models can be used to help understand the effects of increased withdrawals, for Groundwater Hydrology analyses of managed aquifer recharge, or other scenarios. The The groundwater system simulated by this model includes model was used to simulate the effects of projected ground- the upper 1,000 ft of basin-fill deposits in the valleys. At least water withdrawals and managed aquifer recharge through one well has recently been completed in the consolidated rock 2050 and to demonstrate uses of groundwater management in the southeast part of Utah Valley, but the basin fill is consid- simulations. ered the principal aquifer in the study area. Lacustrine, allu- vial, and colluvial processes deposited and sorted the basin-fill Purpose and Scope deposits according to the level of valley lakes and location of streams at the time of deposition (Brooks and Stolp, 1995, p. This report evaluates an extended simulation period of a 13). These depositional environments create alternating and numerical model of the groundwater flow system in southern interfingering layers and lenses, causing vertical and horizon- Utah and Goshen Valleys, Utah, that originally simulated con- tal heterogeneity in the basin fill. ditions from 1949 to 1990 (Brooks and Stolp, 1995). Estimates Groundwater occurs in the basin-fill deposits under uncon- of annual recharge to the system and discharge from wells fined and confined conditions. Groundwater is unconfined in from 1991 to 2011 were added to the model, and the model the poorly sorted deposits near the mountains, but becomes performance was evaluated. Land-use changes between 1991 confined toward the center of the valleys as clay lenses and 2011 were not simulated in the model. This report also become more prominent in the deposits (Brooks and Stolp, presents projection simulations that have increased withdraw- 1995, p. 15). Unconfined conditions exist in about the upper als and managed aquifer recharge, and example groundwater 50 ft of deposits throughout the two valleys. Both confined management simulations to demonstrate the utility of optimi- and unconfined water are considered to be part of the main zation modeling for the period 2012 to 2050. groundwater system (fig. 2). Recharge to the groundwater system is from subsurface inflow from consolidated rocks surrounding the valleys, Description of Study Area streams and canals, irrigation, and precipitation on the valley (table 1). Groundwater discharge is to springs and drains, The study area covers about 390 mi2 in southern Utah and evapotranspiration, wells, streams and canals, Utah Lake, and Goshen Valleys in the north-central part of Utah and corre- sewer systems that act as drains in some locations. sponds to the extent of the unconsolidated basin-fill deposits The only part of the study area with a continuous clay layer (fig.1). Utah Lake occupies much of the northern part of the is the Mapleton Bench, which is near the mountains between study area and covers about 75 mi2 of the study area. The Hobble Creek and Spanish Fork River (fig. 1). The Mapleton study area is bounded by the Wasatch Range on the east and Bench is underlain by at least one thick, continuous layer of south, by the East Tintic Mountains on the west, and by an clay, locally mixed with sand and silt (Brooks and Stolp, 1995, arbitrary divide on the north. The two valleys are separated p. 15). The clay isolates the unconfined groundwater system by Utah Lake and West Mountain, but are hydraulically con- in this area from the main groundwater system (Brooks and nected beneath Utah Lake and south of West Mountain. Stolp, 1995, fig. 6). Recharge from precipitation, irrigation, Land-surface altitude in the study area, excluding West streams, and canals in this area becomes discharge to springs Mountain, ranges from 4,489 ft at Utah Lake to about 5,200 ft along the outer margins of the bench (Richardson, 1906, p. 53) at the southeastern edge of Utah Valley. Altitudes in the and to Hobble Creek and Mill Race Canal (Brooks and Stolp, Wasatch Range east of the study area exceed 10,000 ft, the 1995, p. 31). The perched system on the Mapleton Bench is highest altitude in the East Tintic Mountains west of the study not included in the groundwater flow model; more detail about area is about 6,400 ft, and the altitude of West Mountain is the groundwater budget in the Mapleton Bench area can be about 6,900 ft. found in Brooks and Stolp (1995). Introduction 3
111°35'
15
80 Salt Lake City 40°30' Study area 40 112°00'
JordanJordan U T A H
70 RiverRiver 15 15
Utah LakeUtah Lake
HobbleHobble
CreekCreek SpringvilleSpringville SpanishSpanishMillMill MapeltonMapelton SouthSouth RaceRace DitchDitch Spanish Fork Southern SouthSouthCanalCanal ForkFork
Utah FieldField SalemSalem 111°10' West Valley CanalCanal
Mountain PaysonPayson CanalCanal
PeteetneetPeteetneetHighlineHighline West DitchWest Ditch
Goshen berryberry
CreekCreek Valley SantaquinSantaquin StrawStraw WarmWarm 40 ElbertaElberta CurrantCurrant SummitSummit SpringsSpringsDitchDitch
Wasatch Range CurrantCurrant
CreekCreek CreekCreek CanalCanal CreekCreek
15 East Tintic Mountains East Tintic
39°45' EXPLANATION Boundary of active model cells
0 5 10 15 20 Miles
0 5 10 15 20 Kilometers
Base from U.S. Geological Survey Digital Line Graph data Hillshade from U.S. Geological Survey 10-meter National Elevation Dataset, 1999–2005 Universal Transverse Mercator Projection, Zone 12, North American Datum of 1983
Figure 1. Location of southern Utah and Goshen Valleys study area, Utah. 4 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
EXPLANATION Direction of groundwater movement
Recharge area
Discharge area
Utah Perched water Lake on Mapleton Deeper bench unconfined aquifer
Deeper confined aquifer
Fine-grained deposits
Water level in shallow unconfined aquifer
Consolidated rock Water level in deeper confined aquifer Water level in deeper unconfined Confining Coarse-grained aquifer layer deposits Modified from Hely and others, 1971
Figure 2. Generalized block diagram showing the basin-fill groundwater system in southern Utah and Goshen Valleys, Utah.
Table 1. Conceptual groundwater budget for the main groundwater system in the basin fill, southern Utah and Goshen Valleys, Utah, 1990. [Modified from Brooks and Stolp, 1995, table 4. All flows in acre-feet per year. —, not applicable]
Flow, in acre-feet Flow, in acre-feet Budget Element Southern Goshen Budget Element Southern Goshen Study area Study area Utah Valley Valley Utah Valley Valley Recharge Discharge Subsurface inflow 1 65,000 1 13,000 1, 2 79,000 Springs and drains 42,700 0 42,700 Perennial streams and major canals 33,400 8,100 41,500 Evapotranspiration 26,000 14,000 40,000 Irrigation and precipitation 14,900 400 15,300 Pumped wells 14,000 13,500 27,500 Intermittent and ephemeral runoff 6,400 400 6,800 Perennial streams and major canals 20,700 2,200 22,900 Intervalley flow3 0 7,800 — Utah Lake 9,600 3,600 13,200 Total recharge (rounded) 120,000 30,000 3 143,000 Sewer systems 5,000 0 5,000 Flowing wells 4,400 0 4,400 Intervalley flow3 7,800 0 — Total discharge (rounded) 130,000 33,000 3 156,000
Water going into (+) or out of -9,800 -3,400 -13,200 (-) storage4 (rounded)
1 Calculated as a residual of the discharge minus all other forms of recharge. 2 Total for study area does not equal sum of two valleys because of rounding error. 3 Intervalley flow not used for study area total. 4 Water going into (+) storage is considered to be discharge and should be added to total discharge; water going out of (-) storage is considered to be recharge and should be added to total recharge. Description of the 1995 Southern Utah and Goshen Valleys Groundwater Flow Model 5
Description of the 1995 Southern Utah Recharge and Goshen Valleys Groundwater The model simulates inflow from consolidated rock, recharge from mine-water dispersion ponds in Goshen Valley, Flow Model irrigation and precipitation, and streams and canals. The model uses the Recharge Package (McDonald and Harbaugh, 1988, The USGS developed a numerical model in 1995 to simu- p. 7–1), the Well Package (McDonald and Harbaugh, 1988, late the groundwater system in the basin fill in southern Utah p. 8–1), and the Stream Package (Prudic, 1989) to simulate and Goshen Valleys (Brooks and Stolp, 1995). Model con- these processes. struction, discretization, recharge, discharge, and MODFLOW Inflow to both valleys from consolidated rock of the sur- packages used are described in Brooks and Stolp (1995) and rounding mountains was simulated with injection wells in are summarized here. The model was constructed using the model layers 1 and 2 along the boundary of the active cells. MODFLOW program (McDonald and Harbaugh, 1988) and The location and amounts of inflow were determined during consists of 45 rows, 103 columns, and 5 layers (fig. 3). Cell size ranges from 0.03 mi2 to 0.9 mi2 (Brooks and Stolp, 1995, p. 54). The model simulates recharge from consolidated rock, irrigation and precipitation on the basin fill, and streams and Column103 canals. The model simulates groundwater discharge to evapo- transpiration, pumping wells, springs, drains, flowing wells, streams, Utah Lake, and sewer systems. Impermeable bound- aries are assumed below 1,000 ft in the basin fill and at the contact between basin fill and consolidated rock at the edges of the valleys.
Column 1
1 Row
EXPLANATION Inactive cell Active cell Active cell, constant head in layer 1
0 2 4 86 10 Miles
0 2 4 6 8 10 Kilometers
Row 45
Figure 3. Numerical model grid, southern Utah and Goshen Valleys, Utah. 6 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah model calibration (Brooks and Stolp, 1995, p. 63). The amount Perennial streams and major canals are simulated by using of inflow from consolidated rock does not vary during the the Stream Package (Prudic, 1989) if the amount of recharge transient simulations. Mine water from the East Tintic Moun- or discharge was estimated to be dependent on the ground- tains was dispersed in ponds in Goshen Valley from 1969 to water level near the stream or canal (fig. 5). Other perennial 1976; this is simulated with a single injection well in model streams and canals were simulated by using the Recharge layer 1. Package because the groundwater level is as much as 400 ft Recharge from precipitation and irrigation is simulated below the stream or canal (Brooks and Stolp, 1995, p. 57). as a specified-flux boundary by using the Recharge Package Recharge from canals simulated with the Recharge Package (McDonald and Harbaugh, 1988, p. 7–1) and applied to the was constant during the transient simulation; recharge from highest active cell. Recharge from irrigation and precipita- West Ditch, Summit Creek, and ephemeral streams varied tion on cropland is dependent on the amount of water applied, annually on the basis of the natural flow in Spanish Fork River soil type, method of application, and crop type. Geographic near Castilla (U.S. Geological Survey, 2012). information system (GIS) methods were used to divide the study area into zones representing these differences. Diver- sions from surface water were tabulated, and the amount of Discharge surface water and groundwater applied to each irrigation area Discharge from the groundwater system is simulated to was determined annually. Summer precipitation was included springs, drains, evapotranspiration, streams, wells, Utah Lake, in the consumptive-use estimates, and winter precipitation and sewer systems. These processes are simulated by the (November through March) was considered to be applied use of the Drain Package (McDonald and Harbaugh, 1988, water. Recharge in each area was calculated annually on the p. 9–1), the Evapotranspiration Package (McDonald and basis of equation 1 (Brooks and Stolp, 1995, table 6): Harbaugh, 1988, p. 10–1), the Stream Package, the Well Pack- age, and the Time-Variant Specified Head (CHD) Package Recharge = [(applied surface water + applied (Leake and Prudic, 1991, Appendix C). Except for pumped groundwater + winter precipitation) × infiltration wells, all discharge is simulated as head-dependent boundar- coefficient] - consumptive use (1) ies (Brooks and Stolp, 1995, p. 63–64) and fluctuates in the transient model in response to changes in simulated water Recharge from precipitation on undeveloped land was levels. Annual withdrawals for pumped irrigation, municipal, assumed to be 10 percent of annual precipitation near the and industrial wells are simulated. Domestic wells and small mountains in southern Utah Valley and 5 percent of annual flowing wells are not simulated. precipitation in the middle of the valley in southern Utah Val- ley (Brooks and Stolp, 1995, fig. 8). Recharge from precipita- tion on undeveloped land in Goshen Valley is assumed to be negligible. Recharge from all sources simulated by using the Recharge Package is shown on figure 4. Description of the 1995 Southern Utah and Goshen Valleys Groundwater Flow Model 7
Utah Lake
Southern Mapleton Utah Bench Valley West Mountain West
Wasatch Range
Goshen Valley EXPLANATION Recharge, in feet per year
0.01 – 0.20
0.21 – 0.48
0.49 – 0.80
0.81 – 1.13
1.14 – 1.46
1.47 – 1.87
1.88 – 2.50
2.51 – 4.03
East Tintic Mountains East Tintic 0 2 4 6 8 10 Miles 4.04 – 6.45
0 2 4 6 8 10 Kilometers 6.46 – 12.67
Boundary of active model cells
Figure 4. Average recharge simulated using the Recharge Package in the numerical model from 1949 to 1990, southern Utah and Goshen Valleys, Utah. 8 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
Utah Lake HobbleHobble
CreekCreek
MillMill SouthSouth DitchDitch RaceRaceSouthern Utah
CanalCanalValley
SouthSouth SpanishSpanish FieldField CanalCanal ForkFork West Mountain West
West DitchWest PeteetneetPeteetneetDitch Wasatch Range
CreekCreek
EXPLANATION WarmWarm Head dependent cell Evapotranspiration
Currant Creek Springs, drains, flowing wells, sewer systems Goshen DitchDitch SummitSummit SpringsSprings Valley Streams CreekCreek Boundary of active model cells U.S. Geological Survery stream gage Spanish
CurrantCurrant Fork at Castilla
0 2 4 6 8 10 Miles
CreekCreek 0 2 4 6 8 10 Kilometers East Tintic Mountains East Tintic
Figure 5. Location of head-dependent boundaries in the numerical model, southern Utah and Goshen Valleys, Utah. Update of the Model to Conditions Through 2011 9
Update of the Model to Conditions and losses in the rivers (table 2) in the original model indi- cates little difference in the net recharge with flow in Hobble Through 2011 Creek, and an inverse difference in the net recharge with flow in Spanish Fork River. Higher flows in Spanish Fork River The model files used in the transient model (Brooks and appear to be associated with higher water levels and more dis- Stolp, 1995) were updated to be compatible with MOD- charge to the river; the amount of flow in the river may have FLOW-2005 (Harbaugh, 2005). GIS methods were used dur- little influence on the amount of recharge and discharge. Hold- ing this process to ensure accurate spatial control of physical ing these stresses constant during the updated model period features and the finite-difference model grid. Model arrays is not expected to have a significant effect on model perfor- were imported to the GIS model grid to enable visualization of mance. The exception would be if decreased water levels aquifer properties, heads, drawdowns, and model features. cause enough stream loss to deplete the streams during periods For the update period, the well and recharge files had the of low flow, but data are not available for this analysis. most extensive changes, and these changes are discussed in Hydraulic properties of the groundwater system were not more detail in following sections. Minor changes were made changed. The distribution and amount of inflow from con- to other files to enable the updated transient model to simulate solidated rock simulated by using the Well Package were 63 stress periods (1949 to 2011) instead of the original 42 kept constant throughout the 63 stress periods as they were in stress periods (1949 to 1990). The altitude of Utah Lake was the original model. The distribution and properties of head- changed during the new stress periods on the basis of lake dependent boundaries to simulate discharge to evapotranspira- altitude or volume on March 15 of each year (Utah Division tion, springs, drains, flowing wells, and sewer systems were of Water Rights, 2012b); the lake level also varied annually in kept constant throughout the 63 stress periods. The recharge the original model (Brooks and Stolp, 1995, p. 64). or discharge at head-dependent boundaries changes during the The average streamflow simulated from 1949 to 1990 was original and extended transient model in response to changes used for the new 21 stress periods in the Stream Package. in simulated water levels. In the original model, the discharge of selected streams was varied annually; this is no longer possible because data are not being collected for the two river systems (Spanish Fork River Groundwater Withdrawals and Hobble Creek) that were simulated with varying flow rates Discharge to pumping wells in southern Utah and Goshen in the original model. No flow data are being collected on the Valleys was estimated and simulated on an annual basis from amount of water being diverted from Spanish Fork River to 1991 to 2011. Annual withdrawals from irrigation wells were the Power Canal just to the east of the study area, or on how estimated from unpublished data in the files of the U.S. Geo- much water returns to the river downstream from the Power logical Survey Utah Water Science Center office in Salt Lake Canal. No flow data are being collected on Hobble Creek. City, Utah. Annual withdrawals from municipal and industrial An analysis of flow rates and associated simulated gains wells were obtained from the Utah Division of Water Rights (2012a). Annual withdrawals from 1991 to 1999 were of similar magnitude to previous years, with some fluctuation for climatic variation (fig. 6). Reported municipal withdrawals Table 2. Comparison of simulated stream-aquifer interactions increased dramatically from 6,300 acre-ft in 1999 to 13,800 for varying flow rates in the original model, southern Utah Valley, acre-ft in 2000. Since 2000, withdrawals have remained at the Utah. higher levels, but have not increased other than fluctuation for [Modified from Brooks and Stolp, 1995, table 18. All flows in acre-feet per year] climatic variation. New wells (since 1990) are concentrated in Year 1990 1964 1949 1983 the Spanish Fork and Springville area, the Santaquin area, and Hobble Creek central Goshen Valley (fig. 7). The increase in municipal with- drawals from 1999 to 2000 is significant and was examined in Starting flow 9,200 15,000 26,000 84,000 more detail. A corresponding drop in water levels in a number Recharge from 4,400 4,800 5,000 5,700 of wells was also observed in the 1999 to 2004 period, which stream provides corroborative evidence for this sizeable increase in Discharge to stream 1,200 1,700 2,200 2,900 withdrawals. The likely explanation for this large year-to- Net recharge 3,200 3,100 2,800 2,800 year increase is a unique combination of climate effects on Spanish Fork River available spring discharge, increased growth in municipal Starting flow 15,000 25,000 26,000 147,000 demands, and the need for groundwater supplies to meet these demands. It is also possible that water use was under-reported Recharge from 10,300 8,100 6,400 4,600 stream in 1999 and that the increase was more gradual than appears by reported numbers. Discharge to stream 2,600 2,900 4,700 5,000 Net recharge 7,700 5,200 1,700 -400 10 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
50,000
45,000 Goshen Valley irrigation Southern Utah Valley irrigation Municipal and industrial 40,000 Goshen Valley total Southern Utah Valley total 35,000
30,000
25,000
20,000
Withdrawals, in acre-feet 15,000
10,000
5,000
0
1949 1954 1959 1964 1969 1974 1979 1984 1989 1994 1999 2004 2009 2014 Figure 6. Irrigation, municipal, and industrial groundwater withdrawals from 1949 to 2011, southern Utah and Goshen Valleys, Utah.
Recharge 1 and by 0.94 to estimate precipitation on the undeveloped areas. Recharge estimates for 1991 to 2011 were developed using Additional data needed to estimate recharge from irrigation methods described by Brooks and Stolp (1995) by using pre- included estimating annual applied surface water for select cipitation, streamflow, canal diversions, and irrigation ground- irrigated areas (table 3), average applied surface water for water withdrawals for each year. Recharge from irrigation was select irrigated areas, assignment of average municipal water determined by using equation 1, but minor changes were made applied to irrigated areas (including lawns and gardens), and from the previous estimates. First, the previous study changed assignment of groundwater withdrawals applied to irrigated consumptive use of each crop type on an annual basis. The areas. All areas with variable applied surface water received update does not account for this because the small variation in water from canals reported by the Spanish Fork River Com- annual consumptive use is negligible compared to unknown missioner (Utah Division of Water Rights, 2012b). changes in land use, crop type, and irrigation method. The Recharge from West Ditch, Summit Creek, and ephemeral previously determined average annual consumptive use from streams was varied during the updated stress periods on the 1949 to 1990 for each crop type in each valley was used basis of annual “natural” flow in the Spanish River at Castilla for 1991 to 2011. Second, precipitation at Payson (Western (USGS gaging station 10150500). The amount of water deliv- Regional Climate Center, 2012a) was previously used, but ered from Strawberry Reservoir to the Spanish Fork River that station has no data after June 1999. For this report, the (John Mendenhall, Spanish Fork River Commissioner, written precipitation at Payson is used through 1998 (with the excep- commun., August 2012) was subtracted from the gaged flow tion of 1996, which is missing some data) and the precipita- at Castilla (U.S. Geological Survey, 2012) for each year to tion at the Santaquin chlorinator (Western Regional Climate determine the natural flow in the river. The original model Center, 2012b) is used for 1996 and 1999 to 2011. During the varied these sources of recharge by the use of a multiplier of period when data are available for both stations, Payson had the 1949–90 average annual natural flow (Brooks and Stolp, an average April–March precipitation that was 0.94 times the 1995, fig. 7), but values were not reported by Brooks and April–March precipitation at the Santaquin chlorinator and an Stolp (1995) for the 1949–90 average natural flow or for the average November–March precipitation that was 1.03 times 1990 natural flow. Brooks and Stolp (1995, p. 11) state that in the precipitation at the Santaquin chlorinator. For consistency 1990, natural flow in Spanish Fork at Castilla was 48 percent with the previous report and model, precipitation from the of the 1949–90 average annual flow. Calculations using data Santaquin chlorinator is multiplied by 1.03 for use in equation from U.S Geological Survey (2012) and John Mendenhall Update of the Model to Conditions Through 2011 11
Utah Lake HobbleHobble Springville
CreekCreek
MillMill SouthSouth Spanish Fork DitchDitch RaceRace
Southern CanalCanal Utah SouthSouthSpanishSpanish Valley FieldField CanalCanal ForkFork Salem West Mountain West Payson
West DitchWest PeteetneetPeteetneetDitch Wasatch Range
Goshen CreekCreek Valley
WarmWarm
Santaquin EXPLANATION Simulated withdrawal, in acre-feet per year
Elberta Currant Creek 2,001 – 2,450 DitchDitch SummitSummit SpringsSprings 1,001 – 2,000 CreekCreek 501 – 1,000 101 – 500 0 – 100 CurrantCurrant Red, withdrawal in 2011, no withdrawal in 1990 Green, withdrawal in 2011, had withdrawal in 1990 Purple, withdrawal in 1990, no withdrawal in 2011 CreekCreek Boundary of active model cells
East Tintic Mountains East Tintic 0 2 4 6 8 10 Miles
0 2 4 6 8 10 Kilometers
Figure 7. Location and magnitude of groundwater withdrawals in the numerical model, southern Utah and Goshen Valleys, Utah.
(Spanish Fork River Commissioner, written commun., August the natural flow in the Spanish Fork River, and recharge from 2012) yield a value of natural flow in 1990 of 47,400 acre-feet. West Ditch, Summit Creek, and ephemeral streams, averaged The 1949–90 average was calculated on the basis of calendar about 25 percent more during 1991 to 2011 than during 1949 year annual flow to be 47,400 divided by 0.48, or about 98,800 to 1990. This is consistent with precipitation at Spanish Fork acre-feet. This value was used as the average annual flow to Power House being greater in the more recent years (Burden determine the multiplier for West Ditch, Summit Creek, and and others, 2012, fig. 14). Average annual recharge simulated ephemeral streams for each year from 1991 to 2011 in the from all sources during the updated stress periods is similar to, updated model periods. The ratio of annual flow to average but less variable than, average annual recharge simulated dur- annual flow in the updated stress periods ranges from 0.38 in ing the original transient simulation (fig. 8). 1994 to 2.46 in 2011. Using these multipliers indicated that 12 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
Table 3. Description of water applied to irrigated areas, A 120,000 southern Utah and Goshen Valleys, Utah. 100,000 Inflow from consolidated rock [Applied surface water, listed as “variable”, “none”, or nonvariable amount, in Recharge from the Recharge 80,000 acre-feet per year; applied groundwater, listed as “variable”, “none”, or nonvari- Package able amount, in acre-feet per year] 60,000 Applied Number of Applied 40,000 Unofficial name of Applied spring water, irrigated surface irrigated area groundwater in acre-feet 20,000 area1 water per year 30,000 Southern Utah Valley 20,000 10,000 1 Springville City 4,460 none 481 0 3 Springville High- 1,908 none 0 -10,000 line Withdrawal from pumping wells -20,000 Change in storage 5 Sage 660 none 172 -30,000 80,000 6 Swenson 2,447 variable 0 Evapotranspiration 70,000 Springs, drains, and flowing wells 7 Dry Creek none variable 4,700 60,000 10 Westfield variable variable 0 50,000 Original model Updated model 11 Mill Race 2,400 none 0 40,000 30,000 12 Lake Shore variable none 0 Groundwater budget component, in acre-feet 20,000 13 Spanish Fork City variable 2 1,600 0 14,000 12,000 Discharge to Utah Lake 14 South Field variable none 0 10,000 Net recharge from streams 15 Salem variable variable 0 8,000 6,000 16 Salem City 1,000 2 320 0 4,000 2,000 17 Strawberry-One variable variable 0 0 18 Strawberry-Two variable variable 60 -2,000 16,000 19 Strawberry-Three variable variable 170 B Inflow from consolidated rock and dispersion ponds 20 Payson City 3,084 2 1,650 0 14,000 Recharge from the Recharge Package 21 Summit variable variable 0 12,000
Goshen Valley 10,000 22 Strawberry-Four variable none 0 8,000 23 Bateman- none variable 0 25,000 Howlett 20,000 Withdrawal from pumping wells 15,000 Change in storage 24 LDS North none variable 0 10,000 5,000 25 LDS Central none variable 0 0 26 LDS Orchards none variable 0 -5,000 -10,000 27 Goshen 7,153 none 0 -15,000 20,000 29 Goshen Mouth 1,400 none 0 29 Currant 3 6,390 variable 0 18,000 Evapotranspiration 30 Lunceford-Ekin none variable 0 16,000 1 Number refers to Brooks and Stolp (1995, table 6 and figure 8).Areas that recharge only the Mapleton Bench are not included. 14,000 2 Municipal supply used for irrigation. Groundwater budget cmoponent, in acre-feet 12,000 3 Was variable in transient model, but flow records were not found for 10,000 1991–2011. Unpublished U.S. Geological Survey files indicate this amount was 8,000 Discharge to Utah Lake commonly used in original model. Net recharge from streams 6,000 4,000 2,000 0 -2,000
1/1/1949 1/1/1954 1/1/1959 1/1/1964 1/1/1969 1/1/1974 1/1/1979 1/1/1984 1/1/1989 1/1/1994 1/1/1999 1/1/2004 1/1/2009 1/1/2014
Figure 8. Simulated groundwater budget components for A, southern Utah Valley and B, Goshen Valley, Utah. Performance of the Updated Model 13
Performance of the Updated Model available from 1991 to 2011. Water-level measurements made in March of multiple years between 1991 and 2012 exist for The ability of the updated transient model to match hydro- 18 wells in southern Utah Valley and 13 wells in Goshen Val- logic conditions determined for 1991 to 2011 was evaluated ley (fig. 9). March measurements collected throughout Utah by comparing water-level changes measured in wells to are generally considered to represent the effects of the previ- changes simulated by the model. The match or mismatch of ous year’s hydrologic conditions on the groundwater system levels before 1991 was evaluated as part of model calibration and are used to examine long-term trends. The main recharge in Brooks and Stolp (1995) and was not considered a factor in April–June and the increased withdrawals from May–Sep- in this evaluation of the ability of the model to simulate more tember of the previous year have the least local effect in recent conditions. Model performance was not evaluated for March. changes in natural discharges because few observations were
2 Utah Lake HobbleHobble Springville 1 3 CreekCreek 4 8
MillMill 9 SouthSouth Spanish Fork DitchDitch RaceRace
6 5 CanalCanal SouthSouth SpanishSpanish Southern 10 Utah 7 Valley FieldField CanalCanal ForkFork 16 20 Salem 21 17 West Mountain West Payson 11 18
12 PeteetneetPeteetneet Wasatch Range West DitchWest Ditch19 22 24 23 Goshen 13 14 CreekCreek Valley 25 15 WarmWarm 26 Santaquin EXPLANATION 27 Well long-term water level locations—Color of well Currant Creek symbol indicates deepest model layer of well. DitchDitch SummitSummit Elberta Number refers to hydrograph in figure 10 28 SpringsSprings 28 Model layer 1 CreekCreek 15 29 Model layer 2 29 Model layer 3 CurrantCurrant 31 Model layer 4 31 27 30 Model layer 5 Boundary of active model cells CreekCreek
East Tintic Mountains East Tintic 0 2 4 6 8 10 Miles
0 2 4 6 8 10 Kilometers
Figure 9. Location and model layer of groundwater-level measurements, southern Utah and Goshen Valleys, Utah. 14 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
In southern Utah Valley, the magnitude and direction of the model, the simulated increase in withdrawals results in annual water-level fluctuation simulated by the model reason- decrease in storage, decrease in discharge to evapotrans- ably match measured water-level changes (hydrographs 1–10 piration, decrease in discharge to drains (springs, drains, and 14–19 on fig. 10). The similar response indicates that the and flowing wells), and increase in recharge from streams variation in annual recharge and annual withdrawals simulated (fig. 8). In the groundwater system, observed water-level in the model are generally good approximations of the actual changes indicate that the decrease in storage is greater than changes in the groundwater system. The similar response also in the model (observed water-level declines are greater than indicates that land-use changes have not had a major influence simulated water-level declines). It is possible that increased on groundwater recharge and that the hydraulic properties recharge from streams or decreased discharge to evapotrans- simulated in the model are appropriate. Measured and simu- piration, springs, drains, and flowing wells may be less in the lated water levels generally rise from the end of the original groundwater system than in the model; measurements are transient simulated period to about 1999, decline from 1999 not available to examine the possible changes in the ground- to about 2005, and rise again after 2005. Although the hydro- water budget. Both the rapid increase in withdrawals and the graphs indicate a reasonable fit, an analysis of model fit to the total withdrawals in southern Utah Valley exceed changes water level in 25 wells measured in multiple years (table 4) and amounts that were simulated for the pre-1991 calibration indicates that the model does not match those levels as well period; it is possible that simulated hydraulic properties may as it matched 1991 levels. In Goshen Valley, simulated water be locally incorrect resulting in simulated changes in natural levels do not decline as much as observed levels from 1991 discharge that are faster than occur in the groundwater system. to 2012. In southern Utah Valley, simulated water levels rise It is also possible that other changes in the system that are not more from 1991 to 1999 than observed levels, but then more simulated contributed to the steep groundwater-level decline; closely match the overall decline from 1999 to 2005 and the these could include difference in streamflow, differences in overall rise from 2005 to 2012. land use and irrigation, or other causes. The model reasonably In several areas, however, the model does not simulate as simulates water-level changes after 2002, indicating that the much decline as was measured from 2000 to 2002 (hydro- rapid increase in withdrawals, not the amount of withdrawals, graphs 3, 5, 9, 12, 13, and 16 on fig. 10). The steep measured may be causing the discrepancy between the simulated water- water-level declines in southern Utah Valley were probably level changes and the measured changes. the result of a rapid increase in withdrawals combined with In the northern part of Goshen Valley, simulated water-level below-average recharge (fig. 8A). Recharge from irrigation changes reasonably match measured changes (hydrographs and precipitation during the early 2000s was below average, 20–22 on fig. 10). Farther south, however, simulated declines but was not as low as during other periods. The increase in are much less than measured declines (hydrographs 23–28 and municipal well withdrawals from 1999 to 2000 in south- 30 on fig. 10), even though in the transient calibration period, ern Utah Valley, however, was unprecedented (fig. 6). In simulated changes reasonably matched measured changes. The pattern of measured declines in Goshen Valley indicates that groundwater withdrawals exceed recharge and the ability Table 4. Summary statistics of model fit to 25 water-level of the groundwater system to reduce natural discharge, so observations, southern Utah and Goshen Valleys, Utah. water withdrawn from wells is continuing to decrease stor- [All water levels in feet] age in the groundwater system. The simulated water levels Average Average Sum of Average indicates that groundwater withdrawals are better balanced Year observed simulated squared residual value value residuals with recharge and the reduction in natural discharge, so less change in storage occurs. Land-use changes indicate that more Southern Utah Valley land is being irrigated than during the previous study, but 1991 4,598 4,605 -6 1,512 withdrawal records indicate similar amounts of groundwater 1999 4,605 4,615 -10 2,244 were withdrawn recently and in the 1970s. It is possible that 2000 4,603 4,613 -10 2,434 groundwater withdrawals in Goshen Valley are greater than 2002 4,595 4,607 -12 2,851 simulated. It is also possible that unknown changes in recharge from Currant Creek, Currant Creek Canal, and the irrigation 2005 4,592 4,603 -11 3,313 from Warm Springs have occurred that are not simulated. As a 2012 4,601 4,611 -10 2,503 test of the possibility of less recharge, recharge from Currant Goshen Valley Creek Canal and the Warm Springs distribution system was 1991 4,524 4,529 -4 822 reduced in some simulations; simulated drawdowns increased 1999 4,516 4,528 -12 2,513 only slightly. It is also possible that groundwater withdrawal 2000 4,514 4,526 -12 2,239 during the calibration period did not cause enough stress to determine accurate hydraulic properties of the aquifer and that 2002 4,511 4,526 -15 2,610 they are not simulated correctly. 2005 4,511 4,528 -17 7,811 2012 4,497 4,518 -21 6,666 Performance of the Updated Model 15
4,530 4,615 4,525 1 9 4,605 No data 4,520 Observed Model 4,595 4,515 4,510 4,585 (D-7-2)34dcd-1 Original model Updated model (D-8-3)14acc-1 Original model Updated model 4,505 4,575 4,600 4,690 4,590 2 4,680 10 4,670 4,580 4,660 4,570 4,650 (D-8-3)34bab-1 4,560 (D-7-3)33baa-6 4,640 4,550 4,630 4,580 4,575 3 4,590 11
Water-level altitude, in feet Water-level 4,570 No data 4,580 4,565 4,570
4,560 altitude, in feet Water-level 4,560 4,555 (D-7-3)33ccc-5 (D-9-1)14aad-1 4,550 4,550 4,520 4,670 4,515 4 (D-9-1)23ada-1 12 4,510 4,650 4,505 4,630 4,500 4,495 4,610 4,490 (D-8-1)2ccd-1 4,485 4,590 4,610 4,545 5 4,540 4,600 13 4,535 4,590 4,530 4,525 4,580 4,520 (D-8-2)25dac-3 4,515 (D-9-1)32bbd-1 4,570 4,510 4,560 4,750 (D-9-1)36bbc-1 4,555 6 14 4,550 4,730 4,545 4,710 4,540 4,690 4,535 (D-8-2)29aaa-7 4,530 4,670 4,535 4,780 (D-9-1)36cdd-1 4,530 7 4,760 15 4,525 4,740 4,520 Water-level altitude, in feet Water-level No data 4,720
Water-level altitude, in feet Water-level 4,515 4,700 (D-8-2)31cdb-1 4,510 4,680 4,590 4,605 8 4,600 4,580 4,595 16 4,570 4,590 4,585 4,560 4,580 4,550 4,575 (D-8-3)8abd-1 4,570 (D-9-2)2dad-2 4,540 4,565
1/1/1984 1/1/1989 1/1/1994 1/1/1999 1/1/2004 1/1/2009 1/1/2014 1/1/1949 1/1/1954 1/1/1959 1/1/1964 1/1/1969 1/1/1974 1/1/1979
1/1/1984 1/1/1989 1/1/1994 1/1/1999 1/1/2004 1/1/2009 1/1/2014 1/1/1949 1/1/1954 1/1/1959 1/1/1964 1/1/1969 1/1/1974 1/1/1979
Figure 10. Measured and simulated water levels, 1949 to 2012, southern Utah and Goshen Valleys, Utah. 16 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
4,600 4,550 4,595 Observed 17 24 4,590 Model 4,530 4,585 4,580 4,510 4,575 (D-9-2)9bac-1 4,570 4,490 Original model Updated model (C-9-1)29acc-1 Original model Updated model 4,565 4,470 4,630 4,560 18 4,550 25 4,620 4,540 4,610 4,530 4,520 4,600 4,510 4,590 (D-9-2)17ada-1 4,500 (C-10-1)4bbb-1 4,490 4,905 4,560 19 4,550 26 4,895 4,540 (D-9-2)22cad-1 4,530 4,885 4,520
Water-level altitude, in feet Water-level Water-level altitude, in feet Water-level 4,510 4,875 4,500 (C-10-1)4cbb-1 4,490 4,530 4,560 4,520 20 4,550 27 4,510 4,540 4,500 4,530 4,490 4,480 4,520 4,470 (C-9-1)4ccc-1 4,510 (C-10-1)17bba-1 4,460 4,500 4,530 4,600 4,520 21 28 4,580 4,510 4,500 4,560 4,490 4,540 4,480 (C-9-1)4ddc-1 (C-10-1)18ccc-1 4,470 4,520 4,540 4,570 4,530 22 4,560 29 4,520 4,550 4,510 4,540 4,500 4,490 4,530 4,520 4,480 (C-9-1)20ddd-1 (C-10-1)24ddc-1 4,470 4,510 4,550 4,590
Water-level altitude, in feet Water-level 4,540 23 4,570 30 4,530 4,520 4,550 4,510 4,530 4,500 altitude, in feet Water-level 4,510 4,490 (C-10-1)28ccb-1 (C-10-1)31cdd-1 4,480 4,490 4,580 31
1/1/1984 1/1/1989 1/1/1994 1/1/1999 1/1/2004 1/1/2009 1/1/2014 1/1/1949 1/1/1954 1/1/1959 1/1/1964 1/1/1969 1/1/1974 1/1/1979 4,570 4,560 4,550 4,540 (C-10-1)33cbb-1
1/1/1984 1/1/1989 1/1/1994 1/1/1999 1/1/2004 1/1/2009 1/1/2014 1/1/1949 1/1/1954 1/1/1959 1/1/1964 1/1/1969 1/1/1974 1/1/1979
Figure 10. Measured and simulated water levels, 1949 to 2012, southern Utah and Goshen Valleys, Utah.—Continued Model Projections 2012 to 2050 17
Appropriate Uses of the Model Strawberry Highline Canal is piped and does not recharge the groundwater system, and (4) the same stresses as in (3) with Because the updated model reasonably represents condi- 30,000 acre-ft/yr of MAR. tions in southern Utah Valley, the model can be used for Average recharge was determined by using the 2010 projection and groundwater management simulations of future applied surface-water and groundwater amounts, a multiplier groundwater withdrawals and managed aquifer recharge. of 1.09 for recharge from ephemeral streams based on natural Rapid changes in the system in response to greatly increased flow in the Spanish Fork River from 1949 to 2011, and multi- withdrawals may not be simulated correctly, and increasing pliers for winter precipitation and total precipitation at Payson withdrawals greatly over withdrawals simulated during the that represent the average from 1949 to 2011. Precipitation calibration period may reduce the accuracy of the projections. was used to calculate recharge as explained in the “Recharge” The model can also be used for particle tracking in southern sections of this report. For the projections with increased Utah Valley to identify the source of water to wells and the withdrawals, municipal withdrawals were increased a total of possible flowpaths of recharged water if water-quality issues about 55,000 acre-ft/yr by 2050 (Mark Chandler; Caldwell, are identified. The model should not be used for projections Richards, and Sorensen (CRS) Engineers, written com- in Goshen Valley until additional withdrawal and discharge mun., October 3, 2012); earlier withdrawals are less (fig. 11). data are collected and the model is recalibrated, if necessary, MAR was applied at locations and in amounts provided by to improve the match of simulated water-level fluctuations to Mark Chandler (CRS Engineers, written commun., March 8, observed water-level fluctuations. 2012). The availability of water and land for recharge and the capability of infiltration in these areas (fig. 12) have not been verified by the USGS. Both the completion of piping of the Strawberry Highline Canal and the beginning of MAR were Model Projections 2012 to 2050 simulated to occur starting in the stress period representing 2017 in the projections. The updated groundwater model was used to estimate Because of uncertainty in the accuracy of the updated possible effects on the groundwater system and groundwater model in Goshen Valley and because increased withdrawals levels caused by increased groundwater withdrawals, chang- are simulated mainly in southern Utah Valley, the effects of the ing the Strawberry Highline Canal from an unlined canal to a projected changes on the simulated groundwater budget were piped system, and the addition of managed aquifer recharge analyzed only for southern Utah Valley, and projected simu- (MAR). One-year stress periods were used to simulate the lated water-level hydrographs are presented only for southern following stress conditions: (1) average recharge and the Utah Valley. With average recharge and 2010 withdrawal rates same withdrawals as in 2010 projected from 2012 to 2050, (projection 1), the system reaches a new steady-state condi- (2) average recharge and increasing municipal withdrawals tion in about 10 years with minimal water-level declines and through 2050, (3) the same stresses as in (2) and assuming the minimal changes to natural discharge or recharge from streams
100,000
90,000 Projected withdrawal Specified withdrawal for groundwater- management simulations 80,000
70,000
60,000
50,000
40,000
30,000
Simulated withdrawals, in acre-feet 20,000
10,000
0
2050 2048 2046 2044 2042 2040 2038 2036 2034 2032 2030 2028 2026 2024 2022 2020 2018 2016 2014 2012
Figure 11. Withdrawals simulated in projections 2, 3, and 4, and in groundwater management example 1, 2012 to 2050, southern Utah and Goshen Valleys, Utah. 18 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
Utah Lake HobbleHobble Springville
CreekCreek 5,000
3,000
MillMill SouthSouth Spanish Fork DitchDitch RaceRace
Southern CanalCanal
Utah SpanishSpanish Valley 5,000
ForkFork Salem West Mountain West Payson 5,000
West DitchWest PeteetneetPeteetneetDitch Wasatch Range 5,000
Goshen CreekCreek Valley
WarmWarm
Santaquin EXPLANATION Potential recharge, in acre-feet per year Elberta Currant Creek 7,000
DitchDitch SummitSummit SpringsSprings Boundary of active model cells
CreekCreek
CurrantCurrant 0 2 4 6 8 10 Miles
0 2 4 6 8 10 Kilometers CreekCreek East Tintic Mountains East Tintic
Figure 12. Areas and amounts of projected managed aquifer recharge, southern Utah Valley, Utah.
(figs. 13 and 14). This projection provides the comparison for natural discharge is captured, and a new steady-state condition the next projections and is referred to as the base projection; is not achieved by 2050; water levels continue to decline, nat- drawdown is considered to be the simulated water level at the ural discharge continues to decline, and recharge from streams end of projection 1 minus the simulated water level at the end continues to increase (figs. 13 and 14). A fifth projection of projections 2, 3, and 4. The effects of projection 2 were continued the stresses at the end of projection 3 for another 50 only slightly less than the effects of projection 3 and are not years. Steady-state conditions were achieved within 10 years shown on all the figures. of withdrawals being held steady at the projected 2050 rates. For the three projections with increased withdrawals, the The effects of the fifth projection are similar to the effects of maximum drawdown is about 400 ft (fig. 15), and in projec- projection 3 and are not shown on the figures. Projection 4, tion 3, some areas of natural discharge no longer discharge with MAR, reduces the effects of increased withdrawals and groundwater (fig. 14). Because withdrawals continue to has less area with large water-level declines and less reduction increase, drawdown must continue outward until further in natural discharge than projection 3 (figs. 14 and 15). Model Projections 2012 to 2050 19
4,530 4,620 Updated model Projection model 4,525 1 4,610 9 4,600 4,520 4,590 4,580 Observed 4,515 Projection 1 4,570 Projection 3 4,510 (D-7-2)34dcd-1 4,560 (D-8-3)14acc-1 Projection 4 Updated model Projection model 4,505 4,550 4,600 4,700 4,590 2 4,680 10 4,660 4,580 4,640 4,570 4,620 4,600 4,560 (D-7-3)33baa-6 4,580 (D-8-3)34bab-1 4,550 4,560 4,580 4,600 4,570 3 4,590 11 4,580 4,560 4,570 4,550
Water-level altitude, in feet Water-level Water-level altitude, in feet Water-level 4,560 4,540 (D-7-3)33ccc-5 4,550 (D-9-1)14aad-1 4,530 4,540 4,520 4,670 4,515 12 4 4,650 4,510 4,505 4,630 4,500 4,610 4,495 4,490 (D-8-1)2ccd-1 4,590 (D-9-1)23ada-1 4,485 4,570 4,610 4,545 4,600 5 4,540 13 4,590 4,535 4,530 4,580 4,525 4,570 4,520 4,560 (D-8-2)25dac-3 4,515 (D-9-1)32bbd-1 4,550 4,510 4,570 4,760 4,740 4,560 6 4,720 14 4,550 4,700 4,540 4,680 4,660 4,530 (D-8-2)29aaa-7 4,640 (D-9-1)36bbc-1 4,520 4,620 4,535 4,780 15 4,530 7 4,740 4,525 4,700 4,520
Water-level altitude, in feet Water-level
Water-level altitude, in feet Water-level 4,660 4,515 (D-8-2)31cdb-1 (D-9-1)36cdd-1 4,510 4,620 4,590 4,610 4,580 8 4,600 16 4,590 4,570 4,580 4,560 4,570 4,550 (D-8-3)8abd-1 4,560 (D-9-2)2dad-2 4,540 4,550
1/1/2044 1/1/2039 1/1/2034 1/1/2029 1/1/2024 1/1/2019 1/1/2014 1/1/2009 1/1/2004 1/1/1999 1/1/1994 1/1/1989 1/1/1984 1/1/1979 1/1/1974 1/1/1969 1/1/1964 1/1/1959 1/1/1954 1/1/1949 1/1/2049
1/1/2049 1/1/2044 1/1/2039 1/1/2034 1/1/2029 1/1/2024 1/1/2019 1/1/2014 1/1/2009 1/1/2004 1/1/1999 1/1/1994 1/1/1989 1/1/1984 1/1/1979 1/1/1974 1/1/1969 1/1/1964 1/1/1959 1/1/1954 1/1/1949
Figure 13. Projected simulated water-level altitude in selected wells, southern Utah Valley, Utah. 20 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
4,600 30,000 4,590 20,000 Change in storage Projection 1 17 Projection 3 4,580 10,000 Projection 4 4,570 0 4,560 Observed Projection 1 -10,000 4,550 Projection 3 4,540 (D-9-2)9bac-1 Projection 4 Updated model Projection model -20,000 Updated model Projection model 4,530 -30,000 4,640 45,000 Evapotranspiration 4,620 18 40,000 4,600 35,000 4,580 30,000 4,560 25,000 (D-9-2)17ada-1 4,540 20,000
Water-level altitude, in feet Water-level 4,900 19 5,000 4,880 4,000
4,860 3,000 Discharge to Utah Lake 4,840 2,000 (D-9-2)22cad-1 4,820 1,000 Groundwater budget component, in acre-feet 25,000 20,000 Recharge from streams
1/1/2049 1/1/2044 1/1/2039 1/1/2034 1/1/2029 1/1/2024 1/1/2019 1/1/2014 1/1/2009 1/1/2004 1/1/1999 1/1/1994 1/1/1989 1/1/1984 1/1/1979 1/1/1974 1/1/1969 1/1/1964 1/1/1959 1/1/1954 1/1/1949 15,000 Figure 13. Projected simulated water-level altitude in selected 10,000 wells, southern Utah Valley, Utah.—Continued 5,000 0 -5,000 80,000 70,000 60,000 50,000 40,000 Springs, drains, and flowing wells in southern Utah Valley 30,000 20,000 6,000 5,000 Salem Lake springs 4,000 3,000
2,000 1,000 0 10,000 8,000 6,000 4,000 2,000 Springs, drains, flowing wells, and evapotranspiration near Spring Lake
Groundwater budget component, in acre-feet 0 12,500 12,000 11,500 11,000 10,500 10,000 9,500 9,000 Springville fish hatchery springs 8,500 8,000
1/1/2049 1/1/2044 1/1/2039 1/1/2034 1/1/2029 1/1/2024 1/1/2019 1/1/2014 1/1/2009 1/1/2004 1/1/1999 1/1/1994 1/1/1989 1/1/1984 1/1/1979 1/1/1974 1/1/1969 1/1/1964 1/1/1959 1/1/1954 1/1/1949
Figure 14. Projected simulated selected groundwater-budget components, southern Utah Valley, Utah. Model Projections 2012 to 2050 21
A
HobbleHobble
EXPLANATION Utah Lake CreekCreek Drawdown compared to base projection, layer 3, in feet -25 – 0
MillMill SouthSouth Southern 1 – 5 RaceRace DitchDitch Utah Valley 6 – 10 CanalCanal
SpanishSpanish
11 – 20 ForkFork 21 – 50
51 – 100 Mountain West
101 – 200 Wasatch Range
PeteetneetPeteetneet 201 – 406 West DitchWest Ditch
Model cells with projected CreekCreek increased withdrawals Boundary of active model WarmWarm cells
Currant Creek
Goshen DitchDitch SummitSummit Valley SpringsSprings CreekCreek 0 2 4 86 10 Miles CurrantCurrant 0 2 4 6 8 10 Kilometers
CreekCreek East Tintic Mountains East Tintic
B
HobbleHobble
Utah Lake CreekCreek
MillMill SouthSouth DitchDitch RaceRaceSouthern Utah
CanalCanalValley SpanishSpanish C ForkFork HobbleHobble Utah Lake CreekCreek West Mountain West
Wasatch Range
West DitchWest DitchPeteetneetPeteetneet MillMill SouthSouth Southern DitchDitch RaceRace Utah CreekCreek Valley CanalCanal
WarmWarm SpanishSpanish ForkFork
Currant Creek
Goshen DitchDitch SummitSummit Valley SpringsSprings Mountain West CreekCreek
Wasatch Range
PeteetneetPeteetneet CurrantCurrant West DitchWest Ditch
CreekCreek CreekCreek
WarmWarm East Tintic Mountains East Tintic
Currant Creek
Goshen DitchDitch SummitSummit Valley SpringsSprings CreekCreek
CurrantCurrant
CreekCreek East Tintic Mountains East Tintic
Figure 15. Drawdown compared to base projection for southern Utah and Goshen Valleys, Utah, in A, projection 2; B, projection 3; and C, projection 4. 22 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
Example Groundwater Management maximum withdrawal from each well was constrained at 1,500 gal/min. MAR is allowed at locations discussed in the “Model Simulations Projections” section of this report and along the Strawberry Highline Canal (fig. 16). Maximum infiltration capacity is Groundwater flow models can be combined with optimi- assumed to be 1,000 ft/yr on the basis of MAR infiltration zation techniques to determine water-resource management rates in Arizona (Gorey and Dent, 2007), but the infiltration strategies that best meet a particular set of management capacity in southern Utah Valley has not been analyzed by objectives and constraints (Ahlfeld and others, 2005, p. 2). the USGS. This large capacity gives GWM-2005 the most Optimization techniques are a set of mathematical programs flexibility in optimizing recharge locations. During simula- that seek to find the optimal (or best) allocation of resources to tions, the maximum infiltration capacity at MAR area MAPLE competing uses. In the context of groundwater management, was reduced to about 380 ft/yr and the maximum infiltration the resources are typically the groundwater and surface-water capacity at MAR area PETEET was reduced to about 24 ft/ resources of a basin and the financial resources of the com- yr to reduce simulated water-levels that would be above land munities that depend on the water (Ahlfeld and others, 2005, surface. p. 2). To demonstrate the utility of combining groundwater The three examples of optimization presented in this report models with optimization models in southern Utah Valley, the include constraints to groundwater withdrawals in southern computer program GWM-2005 (Ahlfeld and others, 2005) Utah Valley by minimizing groundwater-level declines and was used to simulate and solve groundwater management minimizing the reduction in discharge to selected springs. scenarios. A groundwater management formulation consists of Water levels were controlled in example 1 by maximizing decision variables, an objective function, and a set of con- simulated water levels at selected cells in four areas (fig. 17A) straints (Ahlfeld and others, 2005, p. 7). Decision variables are and in examples 2 and 3 by constraining simulated drawdown quantifiable controls that are determined by the model, such as at selected model cells (fig. 17B). Additional constraints on the amount of pumping from each well or the amount of MAR water levels were used to prevent high water levels near MAR at each location. The objective function is used to identify the sites in example 3 (fig. 17B). Discharge to springs is con- best solution and is typically phrased as “maximize withdraw- strained in example 1 by maximizing the smallest discharge to als”, “minimize drawdown”, or “minimize MAR”. The con- Salem Lake, Spring Lake, and springs near the fish hatchery in straints impose restrictions on the decision variables or state Springville. Discharge to springs in examples 2 and 3 is con- of the system, such as withdrawal at a well cannot exceed a strained by requiring a specified minimum discharge at each of certain amount, or drawdown cannot exceed a certain amount. those three springs (fig. 17B). The three examples described in this report are provided to The first groundwater management example demonstrates illustrate the possible uses of GWM-2005 and are based on the ability of GWM-2005 to determine withdrawal locations hypothetical and simplified decision variables and constraints. and amounts that allow for the highest possible water levels The water managers in the study area were not involved in set- and spring discharge for a given withdrawal and no MAR. The ting these hypothetical values, nor are financial and engineer- objective function in GWM-2005 was to maximize the sum of ing considerations included in these examples. To enable faster the lowest water level in each of four areas (fig. 17A) and the model simulations, the linear solution of GWM-2005 (Ahlfeld smallest spring discharge. The constraints were that overall and others, 2005, p. 19) was used for these examples. This cre- withdrawals must meet the expected demands in each decade ates some error in the solution because of the non-linearities (fig. 11). The effects of this example can be compared to in the system, especially at head-dependent boundaries such projection 3, where municipal increases were applied at each as springs and streams, but is sufficient for the demonstration well in each municipality. Example 1 indicates that maximum purposes of this report. drawdown can be less than in projection 3 (fig. 18) and that For these examples, decision variables in southern Utah spring discharge can be greater than in projection 3 at Salem Valley include how much water to withdraw from each area Lake and Spring Lake (fig. 19) by withdrawing water mostly and how much, if any, MAR to apply in each area. Pro- in the Springville, Mapleton, and Spanish Fork City areas jected municipal withdrawals were either set at the estimated (fig. 20). demand, or were maximized by GWM-2005 while meeting The second groundwater management example demon- drawdown and spring discharge constraints. To minimize strates the ability of GWM-2005 to maximize withdrawals the number of decision variables, the individual wells were given constraints on drawdown and spring discharge. Con- grouped into areas of municipal withdrawal, and the with- straints were that water levels at selected cells (fig. 17B) be no drawals from each area were used as decision variables (fig. lower than 100 ft below the water level at the end of the base 16). Also, to minimize the number of decision variables, the projection and that discharge to springs at Salem Lake, Spring municipal withdrawals were assumed to increase in each Lake, and springs near the fish hatchery in Springville be at period from 2012 to 2020, 2021 to 2030, 2031 to 2040, and least 65 percent of the discharge at the end of the base projec- 2041 to 2050, instead of annual increases (fig. 11). To provide tion. This example, which can also be compared to projection optimum flexibility, the total projected municipal require- 3, indicates that about 92 percent of the projected demand can ment was used rather than individual city requirements. The be supplied and meet these constraints (figs. 18C, 19, and 20). Example Groundwater Management Simulations 23
A EXPLANATION Utah Lake Well decision variable Springville ELK1 SF2
Southern ELK2 SF3 Utah MAPLE1 SPRVL2 Valley MAPLE2 W18_98 Spanish Fork PAY1 W22_90 SANT1 W25_21 SANT2 W36_69 SF1 WH1
Salem Boundary of active model cells West Mountain West Payson
Wasatch Range
Goshen Valley
Santaquin
B Utah Lake Springville
Southern Utah Valley Spanish Fork
Salem West Mountain West Payson
Wasatch Range
Goshen EXPLANATION Valley
Managed aquifer recharge decision Santaquin variable GOLF SH1 GRAVEL SH2 HOBBLE SH3 MAPLE SH4 PETEET SUMMIT 0 2 4 6 8 10 Miles Boundary of active model cells 0 2 4 6 8 10 Kilometers
Figure 16. Location of well decision variables and managed aquifer recharge decision variables for optimization examples, southern Utah and Goshen Valleys, Utah. 24 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
A EXPLANATION Utah Lake Area in which to maximize water level Springville LOW MAPLE SOUTH Southern WH Utah Spanish Fork Valley Area in which to maximize spring discharge Near fish hatchery Salem Lake Spring Lake Boundary of active model cells
Salem West Mountain West Payson
Wasatch Range
Goshen Valley
Santaquin
B Utah Lake Springville
Southern Utah Spanish Fork Valley
Salem West Mountain West Payson
Wasatch Range
EXPLANATION Goshen Valley Water-level constraint Drawdown for example 2 Additional drawdown for example 3 Santaquin Prevent high level for example 3 Minimum discharge requirement Near fish hatchery Salem Lake Spring Lake 0 2 4 6 8 10 Miles Boundary of active model cells 0 2 4 6 8 10 Kilometers
Figure 17. Locations in southern Utah and Goshen Valleys, Utah, of A, areas in which to maximize water levels and spring discharge and B, water-level and spring-discharge constraints for optimization examples. Example Groundwater Management Simulations 25
A
HobbleHobble
Utah Lake CreekCreek EXPLANATION Drawdown compared to base projection, layer 3, in feet MillMill SouthSouth Southern 1 – 5 RaceRace DitchDitch Utah Valley 6 – 10 CanalCanal
SpanishSpanish
11 – 20 ForkFork 21 – 50
51 – 100 Mountain West
101 – 200 Wasatch Range
PeteetneetPeteetneet 201 – 406 West DitchWest Ditch
Model cells with possible CreekCreek increased withdrawals Boundary of active model WarmWarm cells
Currant Creek
Goshen DitchDitch SummitSummit Valley SpringsSprings CreekCreek 0 2 4 86 10 Miles CurrantCurrant 0 2 4 6 8 10 Kilometers
CreekCreek East Tintic Mountains East Tintic
B
HobbleHobble
Utah Lake CreekCreek
MillMill SouthSouth DitchDitch RaceRaceSouthern Utah
CanalCanalValley SpanishSpanish C ForkFork HobbleHobble Utah Lake CreekCreek West Mountain West
Wasatch Range
West DitchWest DitchPeteetneetPeteetneet MillMill SouthSouth Southern DitchDitch RaceRace Utah CreekCreek Valley CanalCanal
WarmWarm SpanishSpanish ForkFork
Currant Creek
Goshen DitchDitch SummitSummit Valley SpringsSprings Mountain West CreekCreek
Wasatch Range
PeteetneetPeteetneet CurrantCurrant West DitchWest Ditch
CreekCreek CreekCreek
WarmWarm East Tintic Mountains East Tintic
Currant Creek
Goshen DitchDitch SummitSummit Valley SpringsSprings CreekCreek
CurrantCurrant
CreekCreek East Tintic Mountains East Tintic
Figure 18. Drawdown compared to base projection for southern Utah and Goshen Valleys, Utah, in A, projection 3; B, optimization example 1; and C, optimization example 2. 26 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
30,000 Because GWM-2005 was being run in linear mode, the final 20,000 Change in storage values of water levels and spring discharge are slightly less 10,000 than the constraints in some areas. 0 The third groundwater management example illustrates -10,000 combining well and MAR variables to maximize withdraw- -20,000 Updated model Projection model als if 30,000 acre-ft/yr of MAR were available. Constraints -30,000 45,000 are that water levels in select cells (fig. 17B) be no lower than Evapotranspiration 40,000 20 ft below the water level simulated at the end of the base 35,000 projection and that spring discharge at Salem Lake, Spring
Lake, and the springs near the fish hatchery in Springville 30,000 Projection 3 be at least 75 percent of the discharge at the end of the base 25,000 Projection 4 GWM with optimal withdrawal GWM with given withdrawal GWM with managed aquifer recharge projection. To limit rising water levels because of MAR, 20,000 additional constraints are placed at select cells that do not 5,000 allow the simulated water level in layer 1 to be higher than the 4,000 simulated water level at the end of projection 1 or land surface minus 20 ft, whichever was highest (fig. 17B). This example 3,000 Discharge to Utah Lake can be compared to projection 4, and it illustrates that optimi- 2,000 zation of withdrawals and MAR can limit effects of increased 1,000
Groundwater budget component, in acre-feet 25,000 withdrawals more efficiently than MAR without optimiza- 20,000 Recharge from streams tion (figs. 19 and 21). Because GWM-2005 was being run in 15,000 linear mode, the final values of water levels are less than the 10,000 constraints in some areas. The example also illustrates that 5,000 withdrawals can be more evenly distributed throughout the 0 municipalities by using optimization with MAR than by using -5,000 optimization without MAR (fig. 20). 80,000 70,000 60,000 50,000 40,000 Springs, drains, and flowing wells in southern Utah Valley 30,000 20,000 6,000 5,000 Salem Lake springs 4,000 3,000
2,000 1,000 0 12,000 10,000 8,000 6,000 4,000 2,000 0 Springs, drains, flowing wells, and evapotranspiration near Spring Lake
Groundwater budget component, in acre-feet -2,000 13,000 12,000 11,000 10,000 9,000 8,000 Springville fish hatchery springs 7,000 6,000
1/1/2049 1/1/2044 1/1/2039 1/1/2034 1/1/2029 1/1/2024 1/1/2019 1/1/2014 1/1/2009 1/1/2004 1/1/1999 1/1/1994 1/1/1989 1/1/1984 1/1/1979 1/1/1974 1/1/1969 1/1/1964 1/1/1959 1/1/1954 1/1/1949
Figure 19. Selected simulated groundwater-budget components for optimization examples, southern Utah Valley, Utah. Example Groundwater Management Simulations 27
A 20,000 2012–2020 Example 1 16,000 2021–2030 2031–2040 12,000 2041–2050
8,000
4,000
0 20,000
Example 2 16,000
12,000
8,000
4,000
Withdrawal, in acre-feet per year 0 20,000 Example 3 16,000
12,000
8,000
4,000
0
Salem Payson Goshen Elk Ridge Springville Mapleton Santaquin
Woodland Hills Spanish Fork City Area of withdrawal 12,000 B Example 3
8,000
4,000
in acre-feet per year
Managed aquifer recharge, 0
GOLF SH1 SH2 SH3 SH4 MAPLE PETEET HOBBLE GRAVEL SUMMIT Area of recharge
Figure 20. Amounts of groundwater withdrawals by area for optimization examples, southern Utah and Goshen Valleys, Utah. 28 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
HobbleHobble Utah Lake CreekCreek A
MillMill SouthSouth Southern DitchDitch RaceRace Utah Valley CanalCanal
SpanishSpanish
ForkFork West Mountain West
Wasatch Range
West DitchWest DitchPeteetneetPeteetneet
CreekCreek
WarmWarm
Currant Creek
Goshen DitchDitch SummitSummit Valley SpringsSprings CreekCreek
CurrantCurrant
CreekCreek East Tintic Mountains East Tintic
HobbleHobble
Utah Lake CreekCreek
B
MillMill SouthSouth DitchDitch RaceRaceSouthern Utah Valley EXPLANATION CanalCanal SpanishSpanish Drawdown compared to base ForkFork projection, layer 3, in feet -54 – -25
-24 – 0 Mountain West
1 – 5 Wasatch Range 6 – 10 West DitchWest DitchPeteetneetPeteetneet 11 – 20
CreekCreek 21 – 50
51 – 100 WarmWarm 101 – 200
Currant Creek
201 – 406 Goshen DitchDitch SummitSummit Valley SpringsSprings Model cells with possible CreekCreek increased withdrawals Boundary of active model cells CurrantCurrant
CreekCreek East Tintic Mountains East Tintic
0 2 4 86 10 Miles
0 2 4 6 8 10 Kilometers
Figure 21. Drawdown compared to base projection for southern Utah and Goshen Valleys, Utah, in A, projection 4 and B, optimization example 3. Summary 29
Potential Revisions and New Data to Summary
Improve the Updated Model The U.S. Geological Survey (USGS), in cooperation with the Southern Utah Valley Municipal Water Associa- Some revisions could be made to the model that may tion, updated an existing USGS model of southern Utah and improve model fit to water-level changes and improve the Goshen Valleys for hydrologic and climatic conditions from accuracy of predictions without collecting new data. Recali- 1991 to 2011. Increased withdrawals from 1999 to 2000 and bration to new stresses and available data may improve the below-average precipitation and streamflow in 2000 to 2003 confidence in model parameters and reduce the uncertainty caused water levels in many wells in the area to decline to in simulated predictions. The dramatic increase in groundwa- their lowest recorded levels by 2005. The model was updated ter withdrawals in southern Utah Valley from 1999 to 2000 and evaluated to see if it could adequately simulate these caused a stress on the system outside the range of stresses dur- conditions and to determine if the model is adequate for use in ing the calibration period. It is possible that transmissivity and future planning. storage properties could be changed in such a way that would All model files used in the transient model were updated to allow the model to simulate more accurately the drawdown be compatible with MODFLOW-2005 and with the additional from March 2000 to March 2002. Changing from an annual stress periods. The well and recharge files had the most exten- stress period to a seasonal stress period may improve model fit sive changes. Discharge to pumping wells in southern Utah if data are available to understand the system and calibrate the and Goshen Valleys was estimated and simulated on an annual model on a seasonal basis. Many new wells are located close basis from 1991 to 2011. New wells (since 1990) are concen- to Spanish Fork River or Hobble Creek (fig. 7), and the model trated in the Spanish Fork and Springville area, the Santaquin predicts increasing recharge from streams with increased area, and central Goshen Valley. Recharge estimates for 1991 withdrawals. During the period of greatest withdrawals in to 2011 were included in the updated model by using precipi- summer, however, those streams do not have much flow and tation, streamflow, canal diversions, and irrigation groundwa- possibly could not recharge the groundwater system as much ter withdrawals for each year. as is allowed during an annual stress period. In southern Utah Valley, the magnitude and direction of Model calibration could be further improved by collection annual water-level fluctuations simulated by the updated of the following data: the flow in Spanish Fork River below model reasonably match measured water-level changes. The the Power Canal diversion, the flow in Hobble Creek, seep- similar response indicates that the variation in annual recharge age runs on Spanish Fork River and Hobble Creek, updated and annual withdrawals simulated in the model are generally estimate of discharge to evapotranspiration, discharge to good approximations of the actual changes in the groundwater springs, and an updated estimate of discharge to large flow- system. In several areas, however, the model does not simulate ing wells. Additional data about stream-aquifer interactions as much decline as was measured from 2000 to 2002. Both could be obtained with shallow piezometers in the streambeds the rapid increase in withdrawals and the total withdrawals in and measurements of temperature and chemical changes southern Utah Valley during this period exceed changes and along the stream channels. An updated evaluation of land use, amounts in the calibration period; it is possible that hydrau- crop types, irrigation methods, and municipal water applied lic properties may be locally incorrect and allow changes in to lawns and gardens would allow recharge to be better simulated natural discharge that do not happen as rapidly in simulated. the groundwater system. It is also possible that other changes The model probably cannot be improved in Goshen Valley in the system that are not simulated contributed to the steep without additional data collection that includes the following: groundwater-level decline; these could include differences updated withdrawal locations and amounts, updated evalua- in streamflow, differences in land use and irrigation, or other tion of crop types and irrigation methods, updated estimate of causes. The model reasonably simulates water-level changes discharge to evapotranspiration, updated estimates of recharge in southern Utah Valley after 2002. from and discharge to Currant Creek and Currant Creek Canal, In the northern part of Goshen Valley, simulated water-level and updated information on the distribution of water from changes reasonably match measured changes. Farther south, Warm Springs. After collection of the data, the model would however, simulated declines are much less than measured need to be modified to account for the changes found, and declines, even though in the transient calibration period, aquifer properties would need to be recalibrated to represent simulated changes reasonably matched measured changes. the groundwater system more closely. Land-use changes indicate it is possible that groundwater Additional improvements in the model could be made by withdrawals in Goshen Valley are greater than estimated. It is using the methods available with MODFLOW-2005 (Har- also possible that groundwater withdrawal during the calibra- baugh, 2005) and UCODE_2005 (Poeter and others, 2005) tion period did not cause enough stress to determine accurate to estimate model parameters by using regression techniques. hydraulic properties of the aquifer. These techniques also provide a quantitative assessment of The model can be used for projections about the effects of model and prediction uncertainty. future groundwater withdrawals and applied recharge water in southern Utah Valley. Rapid changes may not be simulated 30 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah correctly, and increasing withdrawals dramatically may reduce Leake, S.A., and Prudic, D.E., 1991, Documentation of a com- the accuracy of the predicted water-level and groundwater puter program to simulate aquifer-system compaction using budget changes. The model should not be used for projections the modular finite-difference ground-water flow model: U.S. in Goshen Valley until additional withdrawal and discharge Geological Techniques of Water-Resources Investigation data are collected and the model is recalibrated, if necessary. Book 6, Chapter A2, variously paged. Recalibration to new stresses and seasonal stress periods, and collection of new withdrawal, stream, land-use, and discharge McDonald, M.G., and Harbaugh, A.W., 1988, A modular data could improve the model and the accuracy of predictions. three-dimensional finite-difference ground-water flow For three projections with increased withdrawals, the maxi- model: U.S. Geological Survey Techniques of Water mum drawdown from the base projection with no increase in Resources Investigations, Book 6, Chapter A1, variously withdrawals is about 400 ft. In projection 3, with increased paged. withdrawals and piping of the Strawberry Highline Canal, Poeter, E.P., Hill, M.C., Banta, E.R., Mehl, Steffen, and Chris- some areas of natural discharge no longer discharge ground- tensen, Steen, 2005, UCODE_2005 and six other computer water. Projection 4, with managed aquifer recharge (MAR), codes for universal sensitivity analysis, calibration, and reduces the effects of increased withdrawals and has less area uncertainty evaluation: U.S. Geological Survey Techniques with large water-level declines and less reduction in natural and Methods 6–A11, 283 p. discharge than the other projections with increased withdraw- als. Optimization examples indicate that groundwater-level Prudic, D.E., 1989, Documentation of a computer program to declines can be less and discharge to select springs can be simulate stream-aquifer relations using a modular, finite-dif- greater by optimizing the location of withdrawals and MAR. ference, ground-water flow model: U.S. Geological Survey Open-File Report 88–729, 113 p. Richardson, G.B., 1906, Underground water in the valleys of References Utah Lake and Jordan River, Utah: U.S. Geological Survey Water-Supply and Irrigation Paper No. 157, 81 p. Ahlfeld, D.P., Barlow, P.M., and Mulligan, A.E., 2005, U.S. Geological Survey, 2012, USGS Surface-Water GWM—A ground-water management process for the U.S. Annual Statistics for Utah: accessed on Septem- Geological Survey modular ground-water model (MOD- ber 28, 2012, at http://waterdata.usgs.gov/ut/nwis/ FLOW-2000): U.S. Geological Survey Open-File Report annual?site_no=10150500&agency_cd=USGS&referred_ 2005–1072, 124 p. module=sw&format=sites_selection_links. Brooks, L.E., and Stolp, B.J., 1995, Hydrology and simulation Utah Division of Water Rights, 2012a, Water Records and of ground-water flow in southern Utah and Goshen Valleys, Use Information Viewer: accessed multiple times from Utah: Utah Department of Natural Resources Technical April to September 2012 at http://www.waterrights.utah. Publication No. 111, 96 p. gov/cgi-bin/wuseview.exe?Startup. Burden, C.B., and others, 2012, Groundwater conditions Utah Division of Water Rights, 2012b, River Commissioner in Utah, spring of 2012: Utah Department of Natural Records Viewer: accessed multiple times from April to Resources Cooperative Investigations Report No. 53, 118 p. September 2012 at http://www.waterrights.utah.gov/cgi-bin/ Gorey, Timothy, and Dent, Patrick, 2007, Maintenance dvrtview.exe?Startup. techniques utilized by Central Arizona Project to optimize Western Regional Climate Center, 2012a, Payson, Utah recharge in Fox, Peter, ed., Management of aquifer recharge (426724), Monthly Climate Summary: accessed on May for sustainability: Phoenix, Ariz., Acacia Publishing Inc., p. 9, 2012, at http://www.wrcc.dri.edu/cgi-bin/cliMAIN. 404–411. pl?ut6724. Harbaugh, A.W., 2005, MODFLOW-2005, the U.S. Geologi- Western Regional Climate Center, 2012b, Santaquin Chlorina- cal Survey modular ground-water model—The groundwa- tor, Utah (427686) Monthly Climate Summary: accessed on ter flow process: U.S. Geological Survey Techniques and May 9, 2012, at http://www.wrcc.dri.edu/cgi-bin/cliMAIN. Methods 6–A16, variously p. pl?ut7686. Hely, A.G, Mower, R.W., and Harr, C.A., 1971, Water resources of Salt Lake County, Utah: Utah Department of Natural Resources Technical Publication 31, 240 p. Konikow, L.F., and Bredehoft, J.D., 1992, Ground-water models cannot be validated: Advances in Water Resources, v. 15, No. 1, p. 75–83. Appendix 1: Supplemental Groundwater Management Simulation 31
Appendix 1: Supplemental Groundwater Management Simulation
An additional groundwater management simulation was 1. The constraint for drawdown is 50 ft for most of the completed at the request of the cooperator. Because the report modeled area. Discussions with the cooperator clarified was already in review, this additional simulation is described that this would not be possible in all areas given the in this appendix. The objective of the simulation was to mini- other constraints on withdrawal locations and high water mize managed aquifer recharge (MAR) while meeting specific levels near recharge areas. Drawdown is calculated as constraints. The locations and amounts of MAR, the amount the difference between simulated levels at the end of of increased withdrawals, and the guidelines for drawdown, projection 1 (base case with average recharge with no spring discharge, and high water levels were provided by increase in withdrawals) and levels simulated during the Caldwell, Richards, Sorensen (CRS) Engineers, the represen- optimization. tative of the cooperator. 2. The constraint for discharge to Salem Pond, Spring Lake and surrounding area, and the springs near the fish Decision Variables and Constraints hatchery in Springville are that rates are not be reduced more than 10 percent from the simulated discharge at The only decision variables in this optimization simulation the end of projection 1. Discussions with the cooperator are the locations and amounts of MAR. To provide an esti- clarified that it might not be possible for optimization to mate of when MAR would be required in each area, the MAR obtain a solution that minimizes changes in discharge decision variables were assigned by decade; this is similar in the Spring Lake area while not allowing high levels to how withdrawals were assigned in the other optimization in the recharge area south of Payson (area PTEET). If simulations described in this report. Five locations for possible a solution was not possible, the spring constraint was MAR were provided by the cooperator (fig. A1-1); each can adjusted to a lower amount to allow optimization. have a different recharge rate, up to the maximum at each site, for each of the four decades of the simulation. The maximum 3. Water-level rises are limited to prevent flooding. The amount of MAR is 58,000 acre-ft/yr, but this maximum would following criteria were used to minimize high water only be simulated if the MAR in each area was maximized. levels: Unused MAR at one location would not be available at a. In many parts of the study area and the model, another location. From north to south, the maximum amounts, water levels are within 20 ft of land surface with in acre-ft/yr, are as follows: HBBLE, 4,000; HWY89, 15,000; the current hydrologic conditions. In areas where GRVEL, 15,000; PTEET, 12,000; SMMIT, 12,000. the simulated water level at the end of projection For the example groundwater management simulations pre- 1 is within 20 ft of land surface, the constraint is sented in this report, the municipal demand could be supplied that water levels not rise above that level. from any well in any city up to a maximum amount per well. The cooperator desired that for the groundwater management b. In areas where the simulated water level at the simulation presented in this appendix, the projected demands end of projection 1 is lower than 20 ft below land be kept in each city; as a result of this, the increased withdraw- surface, the constraint is that water levels cannot als were removed as decision variables in this simulation and rise above 20 ft below land surface. are considered as a specified discharge at the location of the existing wells. This is the same method and amount of with- drawal as used in the projection simulations described in this Results report, but it limits the ability of optimization to determine the best withdrawal locations. The maximum total MAR for this simulation is about The objective function in this simulation minimizes the 33,000 acre-ft/yr. MAR in areas HBBLE and GRVEL is at the total amount of MAR applied during the simulation. Because maximum amount allowed in each area in the final decade of of this, constraints are set at the end of each decade to ensure the simulation. MAR is limited in areas HWY89 and PTEET that the simulation does not allow undesirable conditions in because water levels would rise too high with greater recharge early decades and apply enough MAR during the last decade rates. As a result, drawdowns exceed 50 ft in parts of Payson, to meet the constraints only at the end of the simulation. Con- in one cell with a withdrawal well near Mill Race Canal, and straints are used to define the maximum changes in the system in one cell with a withdrawal well near the mouth of Span- that are acceptable to water users and communities in the area. ish Fork Canyon (fig. A1-2). Because MAR is limited by the The following constraints are applied throughout southern high water levels south of Payson, the discharge in the Spring Utah Valley (fig. A1-1): Lake area cannot be maintained at 90 percent of the flow in projection 1 and is reduced to about 70 percent of the flow by the end of the simulation (fig. A1-3). The amounts of recharge at each MAR area generally increase with each decade, but 32 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
A EXPLANATION Managed aquifer recharge Utah Lake decision variable Springville GRVEL HBBLE HWY89 Southern Utah Spanish Fork PTEET Valley SMMIT Boundary of active model cells
Salem West Mountain West Payson
Wasatch Range
Goshen Valley
Santaquin
B Utah Lake Springville
Southern Utah Spanish Fork Valley
Salem West Mountain West Payson
Wasatch Range
EXPLANATION Goshen Valley Head constraint Drawdown Drawdown and flood Santaquin Flood Minimum discharge requirement Near fish hatchery Salem Lake Spring Lake 0 2 4 6 8 10 Miles Boundary of active model cells 0 2 4 6 8 10 Kilometers
Figure A1-1. Location in southern Utah and Goshen Valleys, Utah, of A, managed aquifer recharge decision variables and B, water- level and spring-discharge constraints. Appendix 1: Supplemental Groundwater Management Simulation 33
HobbleHobble Utah Lake CreekCreek
MillMill SouthSouth Southern DitchDitch RaceRace Utah Valley CanalCanal
SpanishSpanish
ForkFork West Mountain West
Wasatch Range
West DitchWest DitchPeteetneetPeteetneet
CreekCreek
WarmWarm EXPLANATION Drawdown compared to base projection, layer 3, in feet Currant Creek -104 – -55 Goshen DitchDitch SummitSummit Valley SpringsSprings -54 – -25 CreekCreek -24 – 0 1 – 5 6 – 10 CurrantCurrant 11 – 20 21 – 50
CreekCreek 51 – 100 101 – 200
East Tintic Mountains East Tintic 201 – 403 Model cells with possible increased withdrawals Boundary of active model cells
0 2 4 6 8 10 Miles
0 2 4 6 8 10 Kilometers
Figure A1-2. Drawdown compared to base projection, southern Utah and Goshen Valleys, Utah. there are exceptions (fig. A1-4). Detailed examination of could provide a unique solution, but was outside the scope of GWM-2005 (Ahlfeld and others, 2005) output files indicates this project. that more than one recharge scenario involving the areas The constraints on high water levels prevent determining south of Payson and near Santaquin will meet the constraints how much MAR would be required to meet the drawdown posed in this problem. To a certain extent, recharge at the area constraints in Payson and near the mouth of Spanish Fork south of Payson can be replaced by recharge near Santaquin Canyon; it may be less than or more than 58,000 acre-ft/yr. and satisfy the constraints. A more detailed analysis involv- Because the objective, constraints, and decision variables are ing preferred recharge areas, costs of recharge for each area, different in this simulation than in the example optimization or different water-level and spring-flow discharge constraints simulations, direct comparisons of the results cannot be made. 34 Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah
30,000 20,000 Change in storage 10,000 0 -10,000
-20,000 Updated model Management model -30,000 45,000 40,000 Evapotranspiration 35,000 30,000 25,000 20,000 5,500 5,000 4,500 Appendix optimization 4,000 3,500 3,000 Discharge to Utah Lake 2,500 2,000 1,500 1,000 25,000 20,000 Recharge from streams 15,000 10,000 5,000 0 -5,000 80,000 70,000 Updated model Management model 60,000 50,000 40,000 Springs, drains, and flowing wells in southern Utah Valley 30,000 Groundwater budget component, in acre-feet 20,000 6,000 5,000 Salem Lake springs 4,000 3,000 2,000 1,000 0 12,000 10,000 8,000 6,000 4,000 2,000 0 Springs, drains, flowing wells, and evapotranspiration near Spring Lake -2,000 13,000 12,000 11,000 10,000 9,000 Springville fish hatchery springs 8,000 7,000 6,000
1/1/2049 1/1/2044 1/1/2039 1/1/2034 1/1/2029 1/1/2024 1/1/2019 1/1/2014 1/1/2009 1/1/2004 1/1/1999 1/1/1994 1/1/1989 1/1/1984 1/1/1979 1/1/1974 1/1/1969 1/1/1964 1/1/1959 1/1/1954 1/1/1949 Date
Figure A1-3. Selected simulated groundwater-budget components, southern Utah Valley, Utah. Appendix 1: Supplemental Groundwater Management Simulation 35
16,000
14,000 2012–2020 12,000 2021–2030 2031–2040 10,000 2041–2050 8,000
6,000
in acre-feet per year 4,000
Managed aquifer recharge, 2,000
0 HBBLE HWY89 GRVEL PTEET SMMIT Managed aquifer recharge area
Figure A1-4. Amounts of optimized managed aquifer recharge by area, southern Utah Valley, Utah.
Brooks—Evaluation of the Groundwater Flow Model for Southern Utah and Goshen Valleys, Utah, Updated to Conditions through 2011, with New Projections and Groundwater Management Simulations—Open-File Report 2013–1171