99-4087Cover_5.fm Page 1 Friday, January 31, 2003 3:55 PM

In cooperation with the Lower Colorado River Authority and the City of Austin

Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92

Water-Resources Investigations Report 99–4087

Lake Buchanan

Lake LBJ Inks Lake

Lake Travis Lake Marble Falls

Lake Austin

Town Lake U.S. Department of the Interior U.S. Geological Survey Cover photographs of dams at Lake Buchanan, Inks Lake, Lake LBJ, Lake Marble Falls, Lake Travis, and Lake Austin courtesy of Lower Colorado River Authority. Photo of dam at Town Lake used with permission of the Austin History Center, Austin Public Library (photo # Pica 20992). U.S. Department of the Interior U.S. Geological Survey

Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92

By Timothy H. Raines and Walter Rast

U.S. GEOLOGICAL SURVEY Water-Resources Investigations Report 99–4087

In cooperation with the Lower Colorado River Authority and the City of Austin

Austin, Texas 1999 U.S. DEPARTMENT OF THE INTERIOR Gale A. Norton, Secretary

U.S. GEOLOGICAL SURVEY Charles G. Groat, Director

Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

For additional information write to

District Chief U.S. Geological Survey 8027 Exchange Dr. Austin, TX 78754–4733 E-mail: [email protected]

Copies of this report can be purchased from

U.S. Geological Survey Information Services Box 25286 Denver, CO 80225–0286 E-mail: [email protected] ii CONTENTS

Abstract ...... 1 Introduction ...... 1 Purpose and Scope ...... 2 Description of Study Area ...... 2 Description of Mass-Balance Model ...... 4 Acknowledgments ...... 5 Characterization of the Quantity and Quality of the Highland Lakes ...... 5 Simulation of the Quantity and Quality of the Highland Lakes ...... 12 Model Development ...... 14 Model Testing ...... 19 Model Simulation ...... 19 Summary ...... 43 References Cited ...... 44

FIGURES

1. Map showing location of study area ...... 3 2. Graphs showing measured monthly streamflow and concentrations of dissolved solids, chloride, and sulfate at Colorado River near San Saba, Texas, 1983–92 ...... 6 3–9. Graphs showing measured monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at: 3. Lake Buchanan, Texas, 1983–92 ...... 7 4. Inks Lake, Texas, 1983–92 ...... 8 5. Lake Lyndon B. Johnson, Texas, 1983–92 ...... 9 6. Lake Marble Falls, Texas, 1983–92 ...... 10 7. Lake Travis, Texas, 1983–92 ...... 11 8. Lake Austin, Texas, 1983–92 ...... 12 9. Town Lake, Texas, 1983–92 ...... 13 10. Map showing location of selected streamflow-gaging stations near the Highland Lakes, Texas ...... 15 11–13. Graphs showing daily mean streamflow in relation to dissolved solids, chloride, and sulfate loads at: 11. Colorado River near San Saba, Texas ...... 20 12. Llano River at Llano, Texas ...... 21 13. Pedernales River near Johnson City, Texas ...... 22 14–15. Graphs showing measured monthly streamflow and estimated monthly concentrations of dissolved solids, chloride, and sulfate at: 14. Llano River at Llano, Texas, 1983–92 ...... 23 15. Pedernales River near Johnson City, Texas, 1983–92 ...... 24 16–22. Graphs showing measured and simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at: 16. Lake Buchanan, Texas, 1983–92 ...... 25 17. Inks Lake, Texas, 1983–92 ...... 26 18. Lake Lyndon B. Johnson, Texas, 1983–92 ...... 27 19. Lake Marble Falls, Texas, 1983–92 ...... 28 20. Lake Travis, Texas, 1983–92 ...... 29 21. Lake Austin, Texas, 1983–92 ...... 30 22. Town Lake, Texas, 1983–92 ...... 31 23. Graphs showing measured and simulated monthly streamflow and concentrations of dissolved solids, chloride, and sulfate at Colorado River at Austin, Texas, 1983–92 ...... 32

CONTENTS iii 24–26. Graphs showing simulated monthly streamflow and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at: 24. Colorado River above Lake Buchanan, Texas, 1987–96 ...... 33 25. Llano River above Lake Lyndon B. Johnson, Texas, 1987–96 ...... 34 26. Pedernales River above Lake Travis, Texas, 1987–96 ...... 35 27–33. Graphs showing simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at: 27. Lake Buchanan, Texas, 1987–96 ...... 36 28. Inks Lake, Texas, 1987–96 ...... 37 29. Lake Lyndon B. Johnson, Texas, 1987–96 ...... 38 30. Lake Marble Falls, Texas, 1987–96 ...... 39 31. Lake Travis, Texas, 1987–96 ...... 40 32. Lake Austin, Texas, 1987–96 ...... 41 33. Town Lake, Texas, 1987–96 ...... 42 34. Graphs showing simulated monthly streamflow and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Colorado River at Austin, Texas, 1987–96 ...... 43

TABLES

1. Selected physical and hydrologic characteristics of the Highland Lakes, Texas ...... 4 2. Applicable water-quality standards for each tributary and reservoir of the Highland Lakes, Texas ...... 5 3. Measured concentrations of dissolved solids, chloride, and sulfate in the Highland Lakes, Texas, 1983–92 ...... 14 4. Characteristics of selected tributary and outlet streams of the Highland Lakes, Texas ...... 16 5. Measured and simulated water gains and losses of the Highland Lakes, Texas ...... 17 6. Regression equations for estimating concentrations of dissolved solids, chloride, and sulfate for selected tributaries of the Highland Lakes, Texas ...... 18 7. Absolute error and error between measured and simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate in the Highland Lakes, Texas, 1983–92 ...... 45 8. Monthly demand distribution for Lake Travis ...... 45 9. Selected simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate in the Highland Lakes, Texas, for wet, average, and dry hydrologic conditions, 1987–96 ...... 46

iv Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92

By Timothy H. Raines and Walter Rast

Abstract From the initial increase in constituent concentra- tions in Lake Buchanan (summer 1987) in response The Highland Lakes, located in central to the saline inflow, the high-salinity water passed Texas, are a series of seven reservoirs on the through the entire Highland Lakes in about 3.5 Colorado River (Lake Buchanan, Inks Lake, Lake years. Lyndon B. Johnson, Lake Marble Falls, Lake A mathematical mass-balance model was Travis, Lake Austin, and Town Lake). The reser- used to simulate the input and movement of high- voirs provide hydroelectric power for the area. In salinity water through the Highland Lakes and addition, Lake Austin and Town Lake also provide to estimate monthly mean concentrations of dis- the public water supply for the Austin metropolitan solved solids, chloride, and sulfate for wet, aver- area. Saline water released from Natural Dam Salt age, and dry hydrologic conditions. The simulated Lake during 1987–89 caused increased concern median monthly concentrations during the 10-year among water managers that high-salinity water simulation period for each reservoir generally are entering the Highland Lakes could result in water- larger for the average condition than for the wet quality problems, necessitating additional treat- condition and generally are larger for the dry ment of the water. condition than for the average condition. The The maximum dissolved solids concentra- simulated concentrations of dissolved solids, tions for the reservoirs after the saline inflow were chloride, and sulfate decreased to levels less than about two to three times the average concentrations the concentrations of the applicable water-quality before the inflow. The maximum concentrations standards in about 2 to 5 years after the saline water of chloride and sulfate after the inflow were about inflow of 1987–89 was simulated for the three three to five times the average concentrations hydrologic conditions. before the inflow. The concentrations of dissolved Results from the simulations indicate that solids, chloride, and sulfate in Lake Buchanan, saline inflows to the Highland Lakes similar to Inks Lake, Lake Lyndon B. Johnson, and Lake those of the releases from Natural Dam Salt Marble Falls were less than the concentrations of Lake during 1987–89 are unlikely to cause large the applicable water-quality standards by the end increases in future concentrations of dissolved of 1990. Concentrations of these constituents in solids, chloride, and sulfate in the Highland Lakes. Lake Travis, Lake Austin, and Town Lake did The results also indicate that high-salinity water not decrease to previous levels, which were less will continue to be diluted as it is transported than the concentrations of the applicable water- downstream through the Highland Lakes, even quality standards, until the end of 1991. Constitu- during extended dry periods. ent concentrations for Lake Buchanan and Inks Lake; for Lake Lyndon B. Johnson and Lake INTRODUCTION Marble Falls; and for Lake Travis, Lake Austin, The headwaters of the Colorado River Basin and Town Lake were similar because of the relative contain natural salt deposits and include active and inac- storage capacities and location of tributary inflows. tive oil and gas fields. Runoff from some of these areas

Abstract 1 containing large concentrations of dissolved solids, chloride, and sulfate were compared to applicable chloride, and sulfate drain into Natural Dam Salt Lake water-quality standards. A mass-balance model was (fig. 1). Natural Dam Salt Lake is a small but extremely used to simulate 10 years of input and movement of saline reservoir (saline water is defined for purposes of water through the seven Highland Lakes and to estimate this report as water having dissolved solids concentra- monthly mean concentrations of dissolved solids, tions greater than 1,500 milligrams per liter (mg/L)). chloride, and sulfate for wet, average, and dry hydro- Water-quality data for Natural Dam Salt Lake indicate a logic conditions. Measured and estimated streamflow long-term increase in dissolved solids concentrations and water-quality data for the Colorado River upstream because water primarily is lost from evaporation. Sub- of Lake Buchanan, the Llano River, and the Pedernales stantial amounts of precipitation during 1986–87 in River were used for inflows to the model. parts of the contributing watersheds of Natural Dam Salt Lake and E.V. Spence Reservoir resulted in releases Description of Study Area from these reservoirs containing thousands of tons of dissolved solids into the Colorado River and ultimately The drainage basin of the Highland Lakes is char- into the Highland Lakes. The Highland Lakes, a series acterized by rolling hills with stony soils and areas of of seven reservoirs on the Colorado River upstream exposed granite and limestone. Soils generally are thin, from and in Austin, Tex., provide considerable natural overlie hard limestone, and primarily are clay or silty and economic benefit to the area, including water sup- clay with low permeability. More detailed geologic ply, flood control, hydroelectric power generation, and information is provided by Garner and Young (1976). recreation. The reservoirs, from upstream to down- The predominant land cover is rangeland, consisting stream, are Lake Buchanan, Inks Lake, Lake Lyndon B. mostly of grasses and shrubs. The local climate is char- Johnson (LBJ), Lake Marble Falls, Lake Travis, Lake acterized by short, mild winters; long, moderately hot Austin, and Town Lake. Saline water has the potential summers; moderately high humidity; and southerly pre- to impart undesirable tastes to drinking water, interfere vailing winds. The mean annual precipitation in Austin with fish production, reduce irrigated crop yields, and during 1928–92 is about 32 inches (in.) and ranges from interfere with industrial water uses. Water managers are about 11 to 51 in. with an average of about 80 storms concerned that the large concentrations of dissolved sol- annually (Miertschin and Armstrong, 1986, p. 175). The ids, chloride, and sulfate entering the Highland Lakes greatest precipitation typically occurs in the spring and could result in water-quality problems, necessitating early fall from frontal storms. Widely scattered thunder- additional treatment of the water prior to its use for storms produce precipitation during the summer. Major municipal and industrial purposes. tributaries of the Highland Lakes include the Llano and In 1990, the U.S. Geological Survey (USGS) Pedernales Rivers. Other smaller tributaries include began an investigation in cooperation with the Lower Sandy, Bull, Barton, and Shoal Creeks (fig. 1). Colorado River Authority (LCRA) and the City of E.V. Spence Reservoir is on the Colorado Austin to characterize and simulate the quantity and River about 315 miles (mi) upstream from Lake quality of the Highland Lakes. The objectives of the Buchanan (fig. 1). Natural Dam Salt Lake is located on study were to (1) characterize the recent quantity and Beals Creek, a tributary to the Colorado River upstream quality of water in the Highland Lakes and (2) use a from E.V. Spence Reservoir, and has a drainage area 2 mass-balance model to simulate the quantity and quality of about 500 square miles (mi ) that normally receives of water in the Highland Lakes for wet, average, and dry an average of about 18 in. of precipitation annually. hydrologic conditions for comparison. Evaporation is the primary means of water loss from Natural Dam Salt Lake. During 1986–87, parts of the Purpose and Scope upper Colorado River Basin received heavy rains that resulted in releases of water that contained hundreds This report describes the quantity and quality of of thousands of tons of salt from Natural Dam Salt water in the Highland Lakes from January 1983 to Lake (and subsequently from E.V. Spence Reservoir) December 1992 and presents the results from simula- into the Colorado River. The chloride concentration in tion of the quantity and quality of water in the Highland Natural Dam Salt Lake before the 1986–87 precipita- Lakes for three hydrologic conditions: wet, average, tion period was estimated to be as large as 6,000 mg/L and dry. Measured concentrations of dissolved solids, (C. Wingert, Colorado River Municipal Water District,

2 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 ADO R R E O V OL I C R

BASTROP

Creek

Shoal Shoal

Creek Bull CALDWELL Austin LAKE TRAVIS

WILLIAMSON ek

o e

r 97 C TRAVIS

on t HAYS

r

a

TOWN LAKE B LAKE AUSTIN Marble Falls INKS LAKE LAKE MARBLE FALLS LAMPASAS LAKE BUCHANAN BURNET Johnson City

er

R er iv

E v o IV i R

R R 30

BLANCO

Creek

es l

Sandy a

MILLS O n

D Llano

A er

LAKE LYNDON B. JOHNSON

R d

e o

O

P n

L la

O LLANO L San Saba C o 98 SAN SABA BROWN GILLESPIE MASON MC CULLOCH o O.H. IVIE RESERVOIR 31 COLEMAN o 0 20 40 60 MILES 99 CONCHO RUNNELS

R

E

V

I San Angelo

R NOLAN COKE

O D

A SPENCE E.V. RESERVOIR

R TOM GREEN LO CO MITCHELL o

k

e

100 e r C Study area

s

l STERLING a e B LOCATION MAP TEXAS Big Spring Location of study area. HOWARD NATURAL DAM NATURAL LAKE SALT GLASSCOCK Digital base from U.S. Geological Survey Scale 1:2,000,000 Albers equal-area projection based on Standard parallels 45.5 and 29.5 degrees o 32 Figure 1. Figure

INTRODUCTION 3 Table 1. Selected physical and hydrologic characteristics of the Highland Lakes, Texas

Reservoir altitude, top Reservoir Reservoir Mean of operating range1 surface area, top storage, top of retention Reservoir (feet above of operating range1 operating range1 time2 mean sea level) (acres) (acre-feet) (months) Lake Buchanan 1,020 28,400 895,000 11–130 Inks Lake 888 793 17,300 .2–2 Lake Lyndon B. Johnson 825 6,380 138,000 1–4 Lake Marble Falls 737 650 8,050 .06–.2 Lake Travis 681 18,900 1,170,000 6–19 Lake Austin 493 1,780 21,200 .1–.4 Town Lake 428 420 3,490 .02–.06

1 Source (except for Town Lake): C.A. Riley, Lower Colorado River Authority, written commun., 1999; Town Lake data from Miertschin and Armstrong (1986). 2 Retention times vary based on inflow and reservoir storage volumes. written commun., 1992). By comparison, a typical Description of Mass-Balance Model chloride concentration in the downstream E.V. Spence Reservoir prior to the 1986–87 precipitation was about The Reservoir Operating and Quality Routing 700 mg/L. program (RESOP II), developed by the Texas Water Development Board (Browder, 1978), provides detailed The storage capacities for Lakes Buchanan, LBJ, and Travis (table 1) are appreciably larger than the stor- analyses of the annual water yield of individual age capacities of the other four reservoirs, Inks Lake, reservoirs and contains a water-quality component Lake Marble Falls, Lake Austin, and Town Lake. Lakes for modeling conservative water-quality constituents. Buchanan and Travis provide water storage to meet RESOP II can be used in an iterative manner to simulate downstream water demands. The five other reservoirs multiple reservoirs in a series.The model performs a are operated as generally constant-level reservoirs. water mass balance for each month using the monthly Because of the smaller storage capacities of the reser- mean storage (S), monthly inflow (I), monthly demand voirs immediately downstream of Lakes Buchanan, (D), monthly releases (R), and monthly evaporation (E). LBJ, and Travis, inflows with large concentrations of The next month ()n1+ storage is computed from dissolved solids, chloride, and sulfate will more quickly the current month (n) values using the following affect the resulting concentrations in the immediately equation: downstream reservoirs. The rate of mixing is repre- sented by the mean retention time for each reservoir. Sn1+ = Sn + In – Dn – Rn – En .(1) The maximum retention time for the four smaller reser- voirs is 2 months compared to 130 months for the three RESOP II also computes concentrations of larger reservoirs. The retention time for a reservoir is a conservative constituents (dissolved solids, chloride, function of the inflow and storage volumes. The travel and sulfate) assuming uniform mixing of the reservoir. time required for water to pass through all the Highland The concentrations of the reservoir demands and Lakes (assuming all the inflow displaces all the storage) releases are assumed to be equal to the concentrations ranges from about 18 to 160 months with an average of of the reservoir storage. The concentration of evapora- about 44 months (Miertschin and Armstrong, 1986, tion is assumed to be zero. For any constituent, the next () p. 51). month reservoir concentration Cn1+ is a function of

4 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 () 1 the current month reservoir concentration Cn and Table 2. Applicable water-quality standards for each () inflow concentration CIn : tributary and reservoir of the Highland Lakes, Texas ()[in milligrams per liter] SnCn + InCIn – DnCn – RnCn Cn1+ = ------.(2) Sn1+ Total Reservoir dissolved Chloride Sulfate Using these variables, the model simulates solids monthly reservoir operation for a desired time interval Colorado River above Lake 875 200 155 (maximum of 50 years) and calculates the storage of the Buchanan reservoir and concentrations of the constituents. Lake Buchanan 525 145 95 A limitation of the original version of RESOP II Pedernales River 525 105 50 was the use of constant annual or monthly reservoir Llano River 300 45 25 demands throughout the entire period of simulation. In Inks Lake 525 135 95 reality, the measured demand varies monthly. RESOP II Lake Lyndon B. Johnson 450 115 70 was modified to accept variable monthly demands Lake Marble Falls 450 115 70 from each reservoir, as measured by the LCRA, City Lake Travis 375 85 60 of Austin, or USGS, to improve model simulation Lake Austin 375 85 60 results. Town Lake 375 75 60 Acknowledgments Colorado River below Town 425 90 60 Lake Lewis Browder, Texas Water Development 1 Board, provided considerable assistance in refining Texas Natural Resource Conservation Commission (1997). and updating the RESOP II model. Jeannie Wigington, Environmental Conservation Section, Water and solved solids (C. Dvorsky, Lower Colorado River Wastewater Department, City of Austin, compiled Authority, written commun., 1992). As a result, the water-quantity and water-quality data for Town Lake. dissolved solids load flowing into Lake Buchanan dur- Michelle Creasy and John Wedig, Water Resources ing 1987–89 represented about 40 percent of that Section, LCRA, provided the water-quantity and water- released from Natural Dam Salt Lake during the period quality data for the Highland Lakes. Jeff Saunders, (Slade and Buszka, 1994). LCRA, provided valuable comments during review of Monthly mean streamflow and concentrations of this report. dissolved solids, chloride, and sulfate at the Colorado River near San Saba during 1983–92 are shown in CHARACTERIZATION OF THE QUANTITY figure 2. The dissolved solids concentration in the Col- AND QUALITY OF THE HIGHLAND LAKES orado River increased from a mean of about 340 mg/L during 1986 to a mean of more than 1,200 mg/L During 1986–87, E.V. Spence Reservoir received during 1988–89. During this period, concentrations runoff from substantial amounts of precipitation, of dissolved solids exceeded the applicable water- resulting in massive inflows to the reservoir, including quality standard for the Colorado River segment above inflows from Natural Dam Salt Lake. Between July Lake Buchanan as established by the Texas Natural 1986 and July 1988, about 70,000 acre-feet (acre-ft) Resource Conservation Commission (1997). Similarly, of water was released from Natural Dam Salt Lake, concentrations of chloride and sulfate exceeded the including major releases in July 1986 and May 1987 applicable water-quality standard during this period. (J. Pickle, Colorado River Municipal Water District, The applicable water-quality standards are different for oral commun., 1992). The inflows to E.V. Spence each tributary and reservoir of the Highland Lakes and Reservoir from Natural Dam Salt Lake contained hun- are listed in table 2. dreds of thousands of tons of dissolved solids. About 225,000 acre-ft of water was released from E.V. Spence Monthly mean reservoir storage and concentra- Reservoir downstream into the Colorado River and was tions of dissolved solids, chloride, and sulfate for estimated to contain more than 1.4 million tons of dis- the Highland Lakes during 1983–92 are shown in

CHARACTERIZATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 5 700,000 5,000

600,000 4,000

500,000

3,000 400,000

300,000 2,000

200,000 IN MILLIGRAMS PER LITER STREAMFLOW, IN ACRE-FEET STREAMFLOW,

DISSOLVED SOLIDS CONCENTRATION, DISSOLVED 1,000 100,000

0 0

1,400 1,200

1,200 1,000

1,000 800

800 600 600

400

IN MILLIGRAMS PER LITER 400 IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE CHLORIDE CONCENTRATION,

200 200

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION

Water-quality standard—Texas Natural Resource Conservation Commission (1997)

Figure 2. Measured monthly streamflow and concentrations of dissolved solids, chloride, and sulfate at Colorado River near San Saba, Texas, 1983–92. figures 3–9. The concentrations of dissolved solids, very similar to those of Lake Buchanan. In both reser- chloride, and sulfate in Lake Buchanan began increas- voirs, the concentrations peaked during 1989–90 and ing during summer 1987. The concentrations in the subsequently decreased to previous levels in 1991. The immediately downstream, smaller Inks Lake are concentrations in Lake LBJ and in the immediately

6 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 1,000,000 1,000

800,000 800

600,000 600

400,000 400 IN MILLIGRAMS PER LITER

200,000 200 DISSOLVED SOLIDS CONCENTRATION, DISSOLVED RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR

0 0

350 250

300 200

250

150 200

150 100

100 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 50 50

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997)

Figure 3. Measured monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Lake Buchanan, Texas, 1983–92. downstream, smaller Lake Marble Falls began after to previous levels in 1991. The concentrations in increasing in early 1988, about 1 year later than the Lake Travis and in downstream Lake Austin and Town concentrations in Lake Buchanan and Inks Lake. The Lake began increasing by the end of 1988, more concentrations peaked in late 1989, decreasing there- gradually than the upstream reservoir groups. The

CHARACTERIZATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 7 22,000 1,000

20,000

18,000 800 16,000

14,000 600 12,000

10,000 400 8,000

6,000 IN MILLIGRAMS PER LITER 200 4,000 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR DISSOLVED SOLIDS CONCENTRATION, DISSOLVED 2,000

0 0

300 250

250 200

200 150

150

100 100 IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE IN MILLIGRAMS PER LITER

CHLORIDE CONCENTRATION, 150 150

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997)

Figure 4. Measured monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Inks Lake, Texas, 1983–92. concentrations peaked in late 1990 or early 1991 and the Highland Lakes from upstream to downstream is the decreased thereafter. The time required for the high- result of dilution by inflows from the tributaries with salinity water to move through all the Highland Lakes lower salinity concentrations than those of water in the was about 42 months.The decrease in concentrations in reservoirs during the high-salinity period.

8 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 160,000 800

140,000 700

120,000 600

100,000 500

80,000 400

60,000 300 IN MILLIGRAMS PER LITER 40,000 200 DISSOLVED SOLIDS CONCENTRATION, DISSOLVED

RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 20,000 100

0 0

250 200

180

200 160

140

150 120

100

100 80

60 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER CHLORIDE CONCENTRATION, SULFATE CONCENTRATION, SULFATE 50 40

20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997)

Figure 5. Measured monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Lake Lyndon B. Johnson, Texas, 1983–92.

In downstream order, the changes in the con- centrations in the reservoirs after the saline inflows centrations of dissolved solids, chloride, and sulfate were about two to three times the mean concentrations in the Highland Lakes resulting from the saline inflows before the saline inflows. The maximum concentrations are listed in table 3. The maximum dissolved solids con- of chloride and sulfate after the saline inflows were

CHARACTERIZATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 9 10,000 800

700 8,000 600

500 6,000

400

4,000 300

IN MILLIGRAMS PER LITER 200 2,000 DISSOLVED SOLIDS CONCENTRATION, DISSOLVED RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 100

0 0

250 180

160

200 140

120 150 100

80 100 60 IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE IN MILLIGRAMS PER LITER 40 CHLORIDE CONCENTRATION, 50

20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997)

Figure 6. Measured monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Lake Marble Falls, Texas, 1983–92. about three to five times the mean concentrations before mean concentrations at Lake Buchanan, and the maxi- the saline inflows. The mean concentrations of the three mum concentrations (after the saline inflows) in the six constituents (before the saline inflows) in the six down- downstream reservoirs ranged from 43 to 102 percent of stream reservoirs ranged from 50 to 109 percent of the the maximum concentrations at Lake Buchanan.

10 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 1,800,000 600

1,600,000 500 1,400,000

1,200,000 400

1,000,000 300 800,000

600,000 200

400,000 IN MILLIGRAMS PER LITER 100 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 200,000 SOLIDS CONCENTRATION, DISSOLVED

0 0

160 120

140 100

120

80 100

80 60

60 40 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER

40 CONCENTRATION, SULFATE CHLORIDE CONCENTRATION,

20 20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION

Water-quality standard—Texas Natural Resource Conservation Commission (1997)

Figure 7. Measured monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Lake Travis, Texas, 1983–92.

The concentrations of dissolved solids, chloride, from 25 to 42 months during 1987–91. The concentra- and sulfate in the Highland Lakes exceeded the concen- tions of dissolved solids, chloride, and sulfate in Lake trations of the applicable water-quality standards (Texas Buchanan, Inks Lake, Lake LBJ, and Lake Marble Natural Resource Conservation Commission, 1997) Falls decreased to less than the concentrations of the

CHARACTERIZATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 11 25,000 600

500 20,000

400 15,000

300

10,000 200 IN MILLIGRAMS PER LITER 5,000 100 DISSOLVED SOLIDS CONCENTRATION, DISSOLVED RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR

0 0

160 120

140 100

120

80 100

80 60

60 40 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER

40 CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 20 20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997)

Figure 8. Measured monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Lake Austin, Texas, 1983–92. applicable water-quality standards by the end of 1990. SIMULATION OF THE QUANTITY AND The concentrations in Lake Travis, Lake Austin, and QUALITY OF THE HIGHLAND LAKES Town Lake did not decrease to previous levels, which were less than the concentrations of the applicable The saline inflows were simulated for three dif- water-quality standards, until the end of 1991. ferent hydrologic conditions—wet, average, and dry—

12 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 4,000 500

3,500

400 3,000

2,500 300

2,000

200 1,500

1,000 IN MILLIGRAMS PER LITER 100 DISSOLVED SOLIDS CONCENTRATION, DISSOLVED RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 500

0 0

140 120

120 100

100 80

80 60 60

40 40 IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE IN MILLIGRAMS PER LITER CHLORIDE CONCENTRATION, 20 20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997)

Figure 9. Measured monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Town Lake, Texas, 1983–92. to determine a range in monthly concentrations of dis- Because reservoir inflows are proportional to solved solids, chloride, and sulfate for each of the three precipitation in each drainage basin, the period of conditions in the Highland Lakes. record for each major tributary to the Highland Lakes

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 13 Table 3. Measured concentrations of dissolved solids, chloride, and sulfate in the Highland Lakes, Texas, 1983–92

[mg/L, milligrams per liter; --, not applicable]

Number of Period in 1986 Maximum Ratio of Ratio of months which Ratio of monthly concen- mean con- maximum exceeding water- maximum mean concen- tration centration concentration water- quality Reservoir concentration tration before after saline to Lake to Lake quality standards to mean 1 saline inflow inflow Buchanan Buchanan standard exceeded concentration (mg/L) (mg/L) concentration concentration during from saline 1983–92 inflow Dissolved solids Lake Buchanan 350 965 2.8 -- -- 35 9/87–7/90 Inks Lake 370 950 2.6 1.06 0.98 37 9/87–9/90 Lake Lyndon B. Johnson 275 755 2.7 .79 .78 28 9/88–9/90 Lake Marble Falls 285 770 2.7 .81 .80 29 9/88–9/90 Lake Travis 255 520 2.0 .73 .54 36 1/89–12/91 Lake Austin 265 515 1.9 .76 .53 36 2/89–12/91 Town Lake 270 495 1.8 .77 .51 34 2/89–11/91 Chloride Lake Buchanan 80 305 3.8 -- -- 39 9/87–7/90 Inks Lake 85 295 3.5 1.06 .97 42 9/87–10/90 Lake Lyndon B. Johnson 50 240 4.8 .62 .79 25 9/88–9/90 Lake Marble Falls 55 225 4.1 .69 .74 29 9/88–9/90 Lake Travis 45 140 3.1 .56 .46 41 1/89–12/91 Lake Austin 45 145 3.2 .56 .48 39 1/89–12/91 Town Lake 40 130 3.2 .50 .43 26 10/89–11/91 Sulfate Lake Buchanan 55 225 4.1 -- -- 38 9/87–7/90 Inks Lake 60 230 3.8 1.09 1.02 38 8/87–9/90 Lake Lyndon B. Johnson 35 180 5.1 .64 .80 31 11/88–11/90 Lake Marble Falls 35 170 4.9 .64 .76 36 9/88–11/90 Lake Travis 30 112 4.2 .55 .56 42 8/89–12/91 Lake Austin 30 100 3.3 .55 .44 36 11/88–12/91 Town Lake 30 100 3.3 .55 .44 26 10/89–11/91

1 Water-quality standard established by the Texas Natural Resource Conservation Commission (1997). was analyzed for the concurrent period of record, hydrologic condition. Available water-quality data were 1944–95, to compute the maximum, mean, and mini- used to estimate inflow concentrations. Available reser- mum 7-year period of cumulative streamflow. The voir demands, releases, and evaporation data were used consecutive 7-year period with the maximum storage for the simulations. was considered to represent the wet hydrologic condi- tion, the consecutive 7-year period with the storage Model Development most comparable to the mean storage for 1944–95 was considered to represent the average hydrologic condi- The major tributaries for the Highland Lakes tion, and the consecutive 7-year period with the mini- are the Colorado, Llano, and Pedernales Rivers, and mum storage was considered to represent the dry additional tributaries include Sandy, Bull, Barton, and

14 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 MILLS

98o COLORADO EXPLANATION SAN SABA 08154700 Streamflow-gaging station LAMPASA and number RIVER 97o 08154700

31o BURNET MASON LLANO LAKE BUCHANAN Llano INKS WILLIAMSON Llano River LAKE 08151500 LAKE LYNDON LAKE MARBLE FALLS B. JOHNSON Marble Sandy Falls Creek TRAVIS 08152000 08154700 GILLESPIE BLANCO Bull Creek Shoal Creek River LAKE TRAVIS BASTROP Creek Pedernales arton 08153500 B Austin LAKE AUSTIN COLORADO 08155300 TOWN LAKE 08156800 08158000 RIVER Digital base from U.S. Geological Survey HAYS Scale 1:2,000,000 30o Albers equal-area projection based on Standard parallels 45.5 and 29.5 degrees CALDWELL

0 20 40 60 MILES

Figure 10. Location of selected streamflow-gaging stations near the Highland Lakes, Texas.

Shoal Creeks (fig. 10). Drainage areas of these tributar- Of the 25,400 mi2 gaged area, about 24,900 mi2 or 98 ies account for about 26,800 of the 27,600 mi2 or about percent of the area is gaged at the Colorado River near 97 percent of the contributing drainage area to Town San Saba, Llano River at Llano, and Pedernales River Lake (table 4). The remaining area comprises interven- near Johnson City streamflow-gaging stations. More ing drainage from local tributaries to the Highland than 50 years of data are available from each of these Lakes. Of the 26,800 mi2, about 25,400 mi2 or 95 per- stations to define the hydrologic conditions for the three cent is measured at USGS streamflow-gaging stations. major tributaries as compared to about 20 years or less

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 15 ------1.03 1.06 1.13 1.42 1.42 1.03 1.11 1.09 1.00 areas stream Ratio of Ratio drainage to station ------5,450 53,700 11,100 42,700 mean 621,000 308,000 186,000 annual 1983–92 1,228,000 1,550,000 (acre-feet) streamflow ------record Period of Period 1941–present 1939–present 1966–present 1940–present 1979–present 1975–present 1984–present 1898–present ) 22.3 12.3 2 346 901 116 ------4,192 area (mi 19,819 25,409 27,606 Station drainage contributing ------Station location Station near San Saba at Llano Kingsland near City Johnson near Austin near 360 at Loop Austin 360, at Loop Austin Street, at 12th at Austin ------Station number 08147000 08151500 08152000 08153500 08154700 08155300 08156800 08158000 1 ) 2 31.6 13.6 391 120 (mi area 4,450 1,281 Stream 20,512 20,552 25,523 25,605 27,352 27,443 27,577 27,600 drainage contributing Stream Characteristics of selected tributary and outlet streams of the Highland Texas Lakes, Tovar and Maldonado, 1981. Maldonado, and Tovar 1 , square miles; --, not available] miles; square available] not , --, B. Johnson 2 Tributary Buchanan at Lake River Colorado Lake Inks at River Colorado Llano River Sandy Creek Lyndon Lake at River Colorado Falls at Marble Colorado River River Pedernales Travis Lake at River Colorado Bull Creek Austin Lake at River Colorado Barton Creek Shoal Creek Total Outlet Lake Town at River Colorado Table 4. [mi

16 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 Table 5. Measured and simulated water gains and losses of the Highland Lakes, Texas

Measured Simulated Reservoir Water gains Water losses Water gains Water losses

Lake Buchanan Colorado River Evaporation; downstream Colorado River Evaporation; downstream releases to Inks Lake releases to Inks Lake Inks Lake Lake Buchanan releases; Evaporation; downstream Lake Buchanan Evaporation; downstream local tributaries releases to Lake Lyndon releases releases to Lake Lyndon B. Johnson B. Johnson Lake Lyndon B. Inks Lake releases; Llano Evaporation; downstream Inks Lake releases; Evaporation; downstream Johnson River; local tributaries releases to Lake Marble Llano River releases to Lake Marble Falls Falls Lake Marble Falls Lake Lyndon B. Johnson Evaporation; downstream Lake Lyndon B. Evaporation; downstream releases; local releases to Lake Travis Johnson releases releases to Lake Travis tributaries Lake Travis Lake Marble Falls Evaporation; downstream Lake Marble Falls Evaporation; downstream releases; Pedernales releases to Lake Austin releases; releases to Lake Austin River; local tributaries Pedernales River Lake Austin Lake Travis releases; Bull Evaporation; downstream Lake Travis releases Evaporation; downstream Creek; local tributaries releases to Town Lake; releases to Town Lake; public water supply public water supply withdrawals by City of withdrawals by City of Austin Austin Town Lake Lake Austin releases; Evaporation; downstream Lake Austin releases Evaporation; downstream Barton Creek; Shoal releases to lower releases to lower Colorado Creek; local tributaries Colorado River; public River; public water supply water supply withdrawals withdrawals by City of by City of Austin Austin for the smaller tributaries. The three major tributaries three major tributaries provide most of the annual account for about 91 percent of the 1983–92 mean inflow to the Highland Lakes and have an adequate annual inflow to the Highland Lakes but only about 72 amount of data to represent the wet, average, and dry percent of the 1983–92 annual streamflow released periods for simulations. from the Highland Lakes measured at Colorado River at Because only periodic water-quality data were Austin. The net change in storage from 1983–92 was a available for the stations on the Llano and Pedernales decrease of about 27,000 acre-ft in Lake Buchanan and Rivers and monthly water-quality data were not avail- an increase of about 210,000 acre-ft in Lake Travis. able for the Colorado River at San Saba prior to 1980, A list of the sources of the measured and simu- regression equations were developed from the available lated water gains and losses for each reservoir of the water-quality data and daily mean streamflows to esti- Highland Lakes is shown in table 5. The water gains mate loads for dissolved solids, chloride, and sulfate for for the Highland Lakes include tributary inflows and the Colorado, Llano, and Pedernales Rivers using ordi- releases from upstream reservoirs. The water losses nary least-squares regression (Helsel and Hirsch, 1992). include evaporation, reservoir releases, and water The regression equations were log-transformed to make supply withdrawals. The primary difference between the variance of the regression residuals approximately the measured and simulated water gains and losses is constant (constant variance is a necessary assumption that the simulated water gains include only tributary of least-squares regression). The equations to estimate inflows from the Colorado, Llano, and Pedernales loads of dissolved solids, chloride, and sulfate as a Rivers with no additional tributary contributions. The function of streamflow (table 6) contain bias correction

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 17 Number of Number observations Highland Lakes, (SE) error in log units log Standard ; CONC, concentration in milligrams 1 ) 2 (R Coefficient of Coefficient determination , bias correction factor correction , bias 2 ) = LOAD/2.446Q CONC ) 0.93 = LOAD/2.446Q CONC 0.390) .98 = LOAD/2.446Q CONC ) 81 = LOAD/2.446Q CONC .207) .89 = LOAD/2.446Q CONC .94 54 .514) .90 = LOAD/2.446Q CONC .223 81 .359) .86 46 = LOAD/2.446Q CONC 54 .608 .93 81 .323 54 2 2 2 2 2 2 2 ) = LOAD/2.446Q CONC .98 .129 46 ) = LOAD/2.446Q CONC .96 .224 46 2 2 Sulfate Chloride Dissolved solids Dissolved 8; Q, streamflow in cubic feet per second; 0.5 SE 0.5 second; feet per in cubic 8; Q, streamflow Tributary Equation Regression equations estimating for concentrations of dissolved solids, chloride, and selected sulfate for tributaries of the Helsel and Hersch (1992). Hersch Helsel and 1 Pedernales River near Johnson City near Johnson River Pedernales SE + 0.5 + 0.904Q EXP(7.07 = LOAD City near Johnson River Pedernales SE + 0.5 + 0.851Q EXP(4.96 = LOAD Llano River at Llano River Llano SE 0.5 + + 1.01Q EXP(6.23 = LOAD Llano at River Llano City near Johnson River Pedernales SE + 0.5 + 0.756Q EXP(5.84 = LOAD SE + 0.5 + 0.894Q EXP(4.59 = LOAD at Llano River Llano SE 0.5 + + 1.11Q EXP(3.06 = LOAD Colorado River near San Saba San near River Colorado SE + 0.5 + 0.909Q EXP(7.35 = LOAD Saba San near River Colorado SE + 0.5 + 0.922Q EXP(5.82 = LOAD Saba San near River Colorado SE + 0.5 + 0.936Q EXP(5.37 = LOAD per liter] Tex as [LOAD,load in kilograms e = 2.7182 per day; EXP, Table 6.

18 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 factors (Helsel and Hirsch, 1992, p. 256). Bias correc- chloride concentrations, and sulfate concentrations are tion factors were used in the equations because the listed in table 7 (at end of report). The mean absolute estimated loads from log-transformed equations, when errors of monthly reservoir storage in the Highland retransformed back to original units, are biased low and Lakes range from 0.9 to 4.7 percent with a minimum thus underestimate the actual loads. The data related to error of 0 to a maximum error of 55.0 percent. The mean the saline inflow during 1987–89 were not used in the errors of monthly reservoir storage range from –2.8 to regression analysis for the Colorado River. The coeffi- 2.2 percent with a minimum error of –32.2 percent to a cients of determination ranged from 0.86 to 0.98, and maximum error of 55.0 percent. the standard errors ranged from 0.129 to 0.608 log The mean absolute errors of monthly concentra- units for the three major tributaries (table 6). The daily tions of dissolved solids in the Highland Lakes range mean streamflow in relation to the load for each of from 14.1 to 35.7 percent, and the mean errors range the three constituents in the Colorado, Llano, and from 11.3 to 35.7 percent. The mean absolute errors of Pedernales Rivers are shown in figures 11–13. The esti- monthly concentrations of chloride range from 16.0 to mated monthly mean concentration is computed using 63.5 percent, and the mean errors range from 13.1 to the equations in table 6 and the monthly mean stream- 63.5 percent. The mean absolute errors of monthly con- flow. centrations of sulfate range from 20.3 to 51.5 percent, The monthly mean streamflow and estimated and the mean errors range from 18.0 to 51.5 percent. monthly mean concentrations of dissolved solids, The absolute errors and errors of monthly concen- chloride, and sulfate at the Llano River at Llano and trations of dissolved solids, chloride, and sulfate gener- the Pedernales River near Johnson City streamflow- ally increase in downstream order, which probably is gaging stations are shown in figures 14–15. The the result of intervening, unaccounted inflow from other measured streamflows for those two streams and for tributaries. The errors accumulate as the water moves the Colorado River near San Saba were adjusted for downstream. The difference between the total measured simulation by the corresponding ratio of stream to sta- and simulated streamflow at Colorado River at Austin is tion drainage area (table 4) to account for the interven- –33.7 percent (fig. 23). The negative difference between ing drainages between the streamflow-gaging stations the measured and simulated streamflow indicates that and the reservoirs. an insufficient amount of water (only about two-thirds of the actual amount) is added to the system from the Model Testing three major tributaries to balance the water quantity. As a result, the error between measured and simulated Measured inflow quantity and estimated water quality increases as the water moves through the quality data for the Colorado River, Llano River, and Highland Lakes because of the lack of water (in the sim- Pedernales River for 1983–92; measured evaporation ulated system) to dilute the concentrations of dissolved data; actual reservoir demands and releases; and initial solids, chloride, and sulfate. reservoir capacities and concentrations were used as The model errors could be reduced if the quantity input to RESOP II to determine if the simulated results and quality of water from minor tributaries and other are comparable to measured values. The measured and intervening inflow to the Highland Lakes were simulated monthly reservoir storage and concentrations accounted for in the simulations. An adequate amount of dissolved solids, chloride, and sulfate for the High- of water quantity and quality data are not available for land Lakes are shown in figures 16–22. the minor tributaries to represent the three hydrologic conditions. However, the results of the model testing are The simulated reservoir storages are similar to judged acceptable to determine the relative differences the measured reservoir storages. The simulated dis- in the quantity and quality of the Highland Lakes from solved solids, chloride, and sulfate concentrations simulations of wet, average, and dry conditions after the generally are larger than the measured concentrations, saline inflow. especially in the downstream reservoirs. The pattern of the simulated concentrations tracks the pattern of the Model Simulation measured concentrations fairly well. The absolute error and the error between the measured and simulated From an analysis of the annual streamflow reservoir storage, dissolved solids concentrations, measured at the Colorado River near San Saba, Llano

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 19 17 15

16 14

15 13

14 12

13 11

12 10

11 9

10 8

9 IN KILOGRAMS PER DAY LOG OF CHLORIDE LOAD, 7 LOG OF DISSOLVED SOLIDS LOAD, IN KILOGRAMS PER DAY SOLIDS LOAD, LOG OF DISSOLVED

8 6 0 24 6 810 0 24 6 810 LOG OF STREAMFLOW, IN CUBIC FEET PER SECOND LOG OF STREAMFLOW, IN CUBIC FEET PER SECOND

15

14

13

12

11

10

9

8 LOG OF SULFATE LOAD, IN KILOGRAMS PER DAY LOAD, LOG OF SULFATE 7

6 0 24 6 810 LOG OF STREAMFLOW, IN CUBIC FEET PER SECOND

Figure 11. Daily mean streamflow in relation to dissolved solids, chloride, and sulfate loads at Colorado River near San Saba, Texas.

20 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 14 12

13 11

12 10

11 9

10 8

9 7 LOG OF CHLORIDE LOAD, IN KILOGRAMS PER DAY LOG OF CHLORIDE LOAD, LOG OF DISSOLVED SOLIDS LOAD, IN KILOGRAMS PER DAY SOLIDS LOAD, LOG OF DISSOLVED 8 6 2 345 6 7 8 2 345 6 7 8 LOG OF STREAMFLOW, IN CUBIC FEET PER SECOND LOG OF STREAMFLOW, IN CUBIC FEET PER SECOND

12

11

10

9

8

7

LOG OF SULFATE LOAD, IN KILOGRAMS PER DAY LOAD, LOG OF SULFATE 6

5 2 345 6 7 8 LOG OF STREAMFLOW, IN CUBIC FEET PER SECOND

Figure 12. Daily mean streamflow in relation to dissolved solids, chloride, and sulfate loads at Llano River at Llano, Texas.

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 21 15 13

14 12

13 11

12 10

11 9

10 8

9 7

8 IN KILOGRAMS PER DAY LOG OF CHLORIDE LOAD, 6 LOG OF DISSOLVED SOLIDS LOAD, IN KILOGRAMS PER DAY SOLIDS LOAD, LOG OF DISSOLVED

7 5 0 24 6 810 0 24 6 810 LOG OF STREAMFLOW, IN CUBIC FEET PER SECOND LOG OF STREAMFLOW, IN CUBIC FEET PER SECOND

13

12

11

10

9

8

7

LOG OF SULFATE LOAD, IN KILOGRAMS PER DAY LOAD, LOG OF SULFATE 6

5 0 24 6 810 LOG OF STREAMFLOW, IN CUBIC FEET PER SECOND

Figure 13. Daily mean streamflow in relation to dissolved solids, chloride, and sulfate loads at Pedernales River near Johnson City, Texas.

22 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 250,000 350

300 200,000

250

150,000 200

150 100,000

100 IN MILLIGRAMS PER LITER

STREAMFLOW, IN ACRE-FEET STREAMFLOW, 50,000

DISSOLVED SOLIDS CONCENTRATION, DISSOLVED 50

0 0

50 30

25 40

20 30

15

20 10 IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE IN MILLIGRAMS PER LITER CHLORIDE CONCENTRATION, 10 5

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997)

Figure 14. Measured monthly streamflow and estimated monthly concentrations of dissolved solids, chloride, and sulfate at Llano River at Llano, Texas, 1983–92.

River at Llano, and Pedernales River near Johnston City period with the maximum average annual storage, about during 1944–95, the wet, average, and dry 7-year peri- 1,570,000 acre-ft during 1955–61, represents the wet ods were selected from a 7-year moving average of the hydrologic condition, the 7-year period with the storage cumulative annual streamflow volumes. The 7-year most comparable to the mean of the average annual

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 23 200,000 600

180,000 500 160,000

140,000 400 120,000

100,000 300

80,000 200 60,000 IN MILLIGRAMS PER LITER

STREAMFLOW, IN ACRE-FEET STREAMFLOW, 40,000 100 DISSOLVED SOLIDS CONCENTRATION, DISSOLVED 20,000

0 0

120 60

100 50

80 40

60 30

40 20 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 20 10

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997)

Figure 15. Measured monthly streamflow and estimated monthly concentrations of dissolved solids, chloride, and sulfate at Pedernales River near Johnson City, Texas, 1983–92. storages, about 1,010,000 acre-ft during 1972–78, rep- The time series of streamflow for model input resents the average hydrologic condition, and the 7-year for 1987–96 for the Colorado, Llano, and Pedernales period with the minimum average annual storage, about Rivers for each of the three hydrologic conditions con- 682,000 acre-ft during 1978–84, represents the dry sisted of the measured streamflow for the period 1987– hydrologic condition. 89, adjusted by the corresponding ratio listed in table 4

24 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 1,200,000 1,200

1,000,000 1,000

800,000 800

600,000 600

400,000 400 IN MILLIGRAMS PER LITER

200,000 SOLIDS CONCENTRATION, DISSOLVED 200 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR

0 0

350 250

300 200

250

150 200

150 100

100 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER CHLORIDE CONCENTRATION, SULFATE CONCENTRATION, SULFATE 50 50

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Measured Simulated

Figure 16. Measured and simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Lake Buchanan, Texas, 1983–92. to account for additional intervening inflow between the The simulated monthly streamflow and concentrations streamflow-gaging station and the reservoir, followed of dissolved solids, chloride, and sulfate for major by the representative 7-year wet-period streamflow, inflows to the Highland Lakes are shown in figures 24– average-period streamflow, or dry-period streamflow. 26. The concentrations of dissolved solids, chloride, and

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 25 25,000 1,200

1,000 20,000

800 15,000

600

10,000 400

5,000 IN MILLIGRAMS PER LITER 200 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR DISSOLVED SOLIDS CONCENTRATION, DISSOLVED

0 0

350 250

300 200

250

150 200

150 100

100 IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE IN MILLIGRAMS PER LITER

CHLORIDE CONCENTRATION, 50 50

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Measured Simulated Figure 17. Measured and simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Inks Lake, Texas, 1983–92. sulfate were estimated from the regression equations the annual demand and monthly distribution factors (table 6). (percentages of annual demand) for Lake Travis were Because actual releases and demands were needed for input to the simulation model (table 8 at end not available for the simulated hydrologic conditions, of report). The Highland Lakes are operated as a system

26 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 160,000 1,000

140,000

800 120,000

100,000 600

80,000

400 60,000

40,000 IN MILLIGRAMS PER LITER 200 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 20,000 SOLIDS CONCENTRATION, DISSOLVED

0 0

250 200

180

200 160

140

150 120

100

100 80

60 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 50 40

20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Measured Simulated Figure 18. Measured and simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Lake Lyndon B. Johnson, Texas, 1983–92. with demand being met by Lake Buchanan and Lake the demand is met by Lake Travis. When the storage of Travis (Lower Colorado River Authority, 1993). Until Lake Travis is between 850,000 and 700,000 acre-ft, the storage of Lake Travis falls below 850,000 acre-ft, 35 percent of the demand is met by Lake Buchanan and

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 27 10,000 1,000

8,000 800

6,000 600

4,000 400 IN MILLIGRAMS PER LITER 2,000 200 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR DISSOLVED SOLIDS CONCENTRATION, DISSOLVED

0 0

250 200

180

200 160

140

150 120

100

100 80

60 IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE IN MILLIGRAMS PER LITER

CHLORIDE CONCENTRATION, 50 40

20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Measured Simulated Figure 19. Measured and simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Lake Marble Falls, Texas, 1983–92.

65 percent by Lake Travis. When the storage of Lake Lake Buchanan is below 150,000 acre-ft. When the Travis is below 700,000 acre-ft, 90 percent of the storage of Lake Buchanan is between 50,000 and demand is met by Lake Buchanan until the storage of 150,000 acre-ft, 35 percent of the demand is met by

28 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 1,800,000 700

1,600,000 600 1,400,000 500 1,200,000

1,000,000 400

800,000 300

600,000 200 IN MILLIGRAMS PER LITER 400,000 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR

DISSOLVED SOLIDS CONCENTRATION, DISSOLVED 100 200,000

0 0

160 120

140 100

120

80 100

80 60

60 40 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER 40 CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 20 20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Measured Simulated

Figure 20. Measured and simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Lake Travis, Texas, 1983–92.

Lake Buchanan. Finally, when the storage of Lake Simulated monthly reservoir storage and concen- Buchanan is below 50,000 acre-ft, all downstream trations of dissolved solids, chloride, and sulfate for demands are met by Lake Travis. wet, average, and dry conditions in the Highland Lakes

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 29 25,000 700

600 20,000

500

15,000 400

300 10,000

200 IN MILLIGRAMS PER LITER 5,000 RESERVOIR STORAGE, IN ACRE FEET IN ACRE STORAGE, RESERVOIR DISSOLVED SOLIDS CONCENTRATION, DISSOLVED 100

0 0

160 120

140 100

120

80 100

80 60

60 40 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER

40 CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 20 20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Measured Simulated Figure 21. Measured and simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Lake Austin, Texas, 1983–92. are shown in figures 27–33. For Lake Buchanan, the range during much of the simulation period (fig. 27). simulated reservoir storage for the wet condition was at Simulated storage in Lake Buchanan for the average or near the reservoir storage at the top of operating and dry conditions was less than storage at the top of

30 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 5,000 700

600 4,000

500

3,000 400

300 2,000

200 IN MILLIGRAMS PER LITER 1,000 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR

DISSOLVED SOLIDS CONCENTRATION, DISSOLVED 100

0 0

160 120

140 100

120

80 100

80 60

60 40 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER CHLORIDE CONCENTRATION, 40 CONCENTRATION, SULFATE

20 20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Measured Simulated Figure 22. Measured and simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate at Town Lake, Texas, 1983–92. operating range. For Lake Travis, the simulated reser- part of the simulation period (fig. 31). Simulated storage voir storage for the wet condition was greater than in Lake Travis for the average and dry conditions did the reservoir storage at the top of operating range, but not reach storage at the top of operating range. Simu- less than the reservoir storage at spillway altitude, for lated storages for all three conditions at the five other

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 31 1,600,000 700

1,400,000 600

1,200,000 500

1,000,000 400 800,000 300 600,000

200

400,000 IN MILLIGRAMS PER LITER

RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 100 200,000 SOLIDS CONCENTRATION, DISSOLVED

0 0

160 120

140 100

120

80 100

80 60

60 40 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER 40 CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 20 20

0 0 1984 1986 1988 1990 1992 1984 1986 1988 1990 1992 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Measured Simulated

Figure 23. Measured and simulated monthly streamflow and concentrations of dissolved solids, chloride, and sulfate at Colorado River at Austin, Texas, 1983–92. reservoirs were equal to their respective reservoir The simulated concentrations of dissolved solids, storages at the top of operating range (figs. 28–30, chloride, and sulfate in Lake Buchanan (fig. 27) and 32–33). Inks Lake (fig. 28) decreased to pre-saline-inflow levels

32 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 1,600,000 5,000

1,400,000 4,000 1,200,000

1,000,000 3,000

800,000

2,000 600,000 STREAMFLOW, IN ACRE-FEET STREAMFLOW, 400,000 IN MILLIGRAMS PER LITER 1,000

200,000 SOLIDS CONCENTRATION, DISSOLVED

0 0

2,000 2,000

1,800 1,800

1,600 1,600

1,400 1,400

1,200 1,200

1,000 1,000

800 800

600 600 IN MILLIGRAMS PER LITER CHLORIDE CONCENTRATION, IN MILLIGRAMS PER LITER 400 CONCENTRATION, SULFATE 400

200 200

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry Figure 24. Simulated monthly streamflow and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Colorado River above Lake Buchanan, Texas, 1987–96. less than or near the concentrations of the applicable for the dry condition simulation, the simulated water-quality standards during 1990–91 for the wet con- concentrations increased. The simulated concentrations dition and in 1992 for the average and dry conditions. of dissolved solids, chloride, and sulfate in Lake LBJ As the reservoir storage in Lake Buchanan neared zero (fig. 29) and Lake Marble Falls (fig. 30) decreased to

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 33 250,000 350

300 200,000

250

150,000 200

150 100,000

100 IN MILLIGRAMS PER LITER

STREAMFLOW, IN ACRE-FEET STREAMFLOW, 50,000

DISSOLVED SOLIDS CONCENTRATION, DISSOLVED 50

0 0

50 30

25 40

20 30

15

20 10 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 10 5

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry Figure 25. Simulated monthly streamflow and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Llano River above Lake Lyndon B. Johnson, Texas, 1987–96. levels less than or near the concentrations of the appli- Lake Austin (fig. 32), and Town Lake (fig. 33) peaked cable water-quality standards in 1993 for the three con- during 1990–92. The simulated concentrations for the ditions. The simulated concentrations of dissolved wet condition peaked first, followed by the average and solids, chloride, and sulfate in Lake Travis (fig. 31), then dry conditions. The simulated concentrations for

34 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 300,000 600

250,000 500

200,000 400

150,000 300

100,000 200 IN MILLIGRAMS PER LITER STREAMFLOW, IN ACRE-FEET STREAMFLOW,

50,000 100 DISSOLVED SOLIDS CONCENTRATION, DISSOLVED

0 0

250 80

70 200 60

50 150

40

100 30 IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE

IN MILLIGRAMS PER LITER 20

CHLORIDE CONCENTRATION, 50 10

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry Figure 26. Simulated monthly streamflow and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Pedernales River above Lake Travis, Texas, 1987–96. the three conditions decreased to less than or near the wet, average, and dry conditions in the Highland Lakes pre-saline-inflow levels during 1993–94. are presented in table 9 (at end of report). The reservoir Selected results of simulated storages and con- storage in Lake Buchanan ranged from 442,000 to centrations of dissolved solids, sulfate, and chloride for 992,000 acre-ft for the wet simulation, from 39,900 to

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 35 1,200,000 1,200

Top of operating range 1,000,000 1,000

800,000 800

600,000 600

400,000 400 IN MILLIGRAMS PER LITER

RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 200,000 200 DISSOLVED SOLIDS CONCENTRATION, DISSOLVED

0 0

400 300

350 250

300

200 250

200 150

150 100 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER

100 CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 50 50

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry

Figure 27. Simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Lake Buchanan, Texas, 1987–96.

992,000 acre-ft for the average simulation, and from to 1,270,000 acre-ft for the average simulation, and 0 to 992,000 acre-ft for the dry simulation. The reser- from 0 to 1,270,000 acre-ft for the dry simulation. voir storage in Lake Travis ranged from 644,000 to The simulated reservoir storage in Lake Buchanan and 1,350,000 acre-ft for the wet simulation, from 30,800 Lake Travis during the dry condition decreased to zero

36 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 20,000 1,200

18,000 Top of operating range 1,000 16,000 Simulated storage for wet, average, and dry conditions equal to storage 14,000 at top of operating range 800 12,000

10,000 600

8,000 400 6,000 IN MILLIGRAMS PER LITER 4,000

RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 200 DISSOLVED SOLIDS CONCENRATION, DISSOLVED 2,000

0 0

400 350

350 300

300 250

250 200 200 150 150

100 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER

100 CONCENTRATION, SULFATE CHLORIDE CONCENTRATION,

50 50

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry Figure 28. Simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Inks Lake, Texas, 1987–96. during the 10-year simulation because the annual Highland Lakes from the three major tributaries demand exceeded the annual inflow from the three accounted only for about 72 percent of the measured major tributaries and the storage in Lake Buchanan streamflow released during 1983–92, the simulated and Lake Travis. Because the measured inflow to the inflow from the three major tributaries probably

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 37 200,000 1,000

180,000

160,000 800 Top of operating range 140,000 Simulated storage for wet, average, 120,000 and dry conditions equal to storage 600 at top of operating range 100,000

80,000 400

60,000 IN MILLIGRAMS PER LITER 40,000 200 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 20,000 SOLIDS CONCENTRATION, DISSOLVED

0 0

350 300

300 250

250 200

200 150 150

100 100 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE CHLORIDE CONCENTRATION,

50 50

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry Figure 29. Simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Lake Lyndon B. Johnson, Texas, 1987–96. accounts for a similar fraction of total inflow to the The median monthly concentrations of dissolved Highland Lakes. Therefore, it is unlikely that Lake solids, chloride, and sulfate during the 10-year simula- Buchanan and Lake Travis would be depleted during tion period for each reservoir generally are larger for the the 7-year period. average condition than for the wet condition and are

38 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 10,000 1,000

Top of operating range 8,000 800 Simulated storage for wet, average, and dry conditions equal to storage at top of operating range

6,000 600

4,000 400 IN MILLIGRAMS PER LITER 2,000 200 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR DISSOLVED SOLIDS CONCENTRATION, DISSOLVED

0 0

350 300

300 250

250 200

200 150 150

100 100 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 50 50

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry

Figure 30. Simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Lake Marble Falls, Texas, 1987–96. larger for the dry condition than for the average condi- and sulfate in all reservoirs for the three simulations tion. The maximum concentrations for the reservoirs are are smaller than simulated maximum concentrations. comparable for the three simulations. The simulated The results indicate that the continuing inputs of large ending concentrations of dissolved solids, chloride, volumes of water (even with large constituent loads)

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 39 1,400,000 800

1,200,000 700

Top of operating range 600 1,000,000

500 800,000 400 600,000 300

400,000 200 IN MILLIGRAMS PER LITER Storage when

RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 200,000 reservoir at spillway altitude: 1,950,000 SOLIDS CONCENTRATION, DISSOLVED 100 acre-feet 0 0

250 200

180

200 160

140

150 120

100

100 80

60 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 50 40

20

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry Figure 31. Simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Lake Travis, Texas, 1987–96. after the saline inflow during 1987–89 tend to have a The simulated monthly streamflow and concen- diluting effect on the constituent concentrations in the trations of dissolved solids, sulfate, and chloride for reservoirs. Generally, the larger the inflow volume, the wet, average, and dry hydrologic conditions at the less time required to dilute the increased concentrations Colorado River at Austin, downstream of Town Lake, in the reservoirs. are shown in figure 34. The simulated concentrations

40 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 22,000 800

20,000 Top of operating range 700 18,000 600 16,000

14,000 500 12,000 Simulated storage for wet, average, 400 10,000 and dry conditions equal to storage at top of operating range 8,000 300

6,000 IN MILLIGRAMS PER LITER 200 4,000 RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR DISSOLVED SOLIDS CONCENTRATION, DISSOLVED 100 2,000

0 0

250 200

180

200 160

140

150 120

100

100 80

60 IN MILLIGRAMS PER LITER CHLORIDE CONCENTRATION, IN MILLIGRAMS PER LITER 50 CONCENTRATION, SULFATE 40

20

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry Figure 32. Simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Lake Austin, Texas, 1987–96. decreased to levels less than the concentrations of the concentrations of the water-quality standards during applicable water-quality standards during 1992–93 1996 because the amount of water stored in Lake for the three hydrologic conditions but exceeded the Buchanan and Lake Travis and the amount of tributary standards again during 1996 for the dry condition. Sim- inflow were insufficient to dilute the concentrations ulated concentrations for the dry condition exceeded before the water was released downstream.

SIMULATION OF THE QUANTITY AND QUALITY OF THE HIGHLAND LAKES 41 5,000 800

700 4,000 Top of operating range 600

Simulated storage for wet, average, 500 3,000 and dry conditions equal to storage at top of operating range 400

2,000 300

IN MILLIGRAMS PER LITER 200 1,000 DISSOLVED SOLIDS CONCENTRATION, DISSOLVED RESERVOIR STORAGE, IN ACRE-FEET STORAGE, RESERVOIR 100

0 0

250 200

180

200 160

140

150 120

100

100 80

60 IN MILLIGRAMS PER LITER IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE CHLORIDE CONCENTRATION, 50 40

20

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry Figure 33. Simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Town Lake, Texas, 1987–96.

Results from the simulations indicate that saline sulfate in the Highland Lakes. The results also indicate inflows to the Highland Lakes similar to those during that saline water will continue to be diluted as it is 1987–89 are unlikely to cause extremely large increases transported downstream through the Highland Lakes, in concentrations of dissolved solids, chloride, and even during extended dry periods.

42 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 1,000,000 800

700 800,000 600

500 600,000

400

400,000 300

IN MILLIGRAMS PER LITER 200 200,000

DISSOLVED SOLIDS CONCENTRATION, DISSOLVED 100 STREAMFLOW, IN CUBIC FEET PER SECOND STREAMFLOW,

0 0

250 200

180

200 160

140

150 120

100

100 80

60 IN MILLIGRAMS PER LITER SULFATE CONCENTRATION, SULFATE IN MILLIGRAMS PER LITER

CHLORIDE CONCENTRATION, 50 40

20

0 0 1988 1990 1992 1994 1996 1988 1990 1992 1994 1996 YEAR YEAR

EXPLANATION Water-quality standard—Texas Natural Resource Conservation Commission (1997) Wet Average Dry Figure 34. Simulated monthly streamflow and concentrations of dissolved solids, chloride, and sulfate for wet, average, and dry hydrologic conditions at Colorado River at Austin, Texas, 1987–96. SUMMARY the public water supply for the Austin metropolitan area. The saline water released from Natural Dam Salt The Highland Lakes are a series of seven reser- Lake into the Highland Lakes during 1987–89 created voirs on the Colorado River in Central Texas. The the potential for water-quality problems that would reservoirs provide hydroelectric power and also provide necessitate additional treatment of the water.

SUMMARY 43 The maximum dissolved solids concentrations for for the wet condition and generally are larger for the the reservoirs after the saline inflow were about two to dry condition than for the average condition. The simu- three times the average concentrations before the lated concentrations of dissolved solids, chloride, and inflow. The maximum concentrations of chloride and sulfate decreased to less than the concentrations of the sulfate after the inflow were about three to five times the applicable water-quality standards in about 2 to 5 years average concentrations before the inflow. The concen- after the saline-water inflow for the three hydrologic trations of dissolved solids, chloride, and sulfate after conditions. the saline inflow decreased from dilution by tributary Results from the simulations indicate that saline inflow. The concentrations of dissolved solids, chloride, inflows to the Highland Lakes similar to those of 1987– and sulfate in Lake Buchanan, Inks Lake, Lake LBJ, 89 are unlikely to cause large increases in concentra- and Lake Marble Falls decreased to less than the con- tions of dissolved solids, chloride, and sulfate in the centrations of the applicable water-quality standards by Highland Lakes. The results also indicate that high- the end of 1990. The concentrations in Lake Travis, salinity water will continue to be diluted as it is trans- Lake Austin, and Town Lake did not decrease to previ- ported downstream through the Highland Lakes, even ous levels, which were less than the concentrations of during extended dry periods. the applicable water-quality standards until the end of 1991. Constituent concentrations for Lake Buchanan REFERENCES CITED and Inks Lake, Lake LBJ and Lake Marble Falls, and Lake Travis, Lake Austin, and Town Lake were similar Browder, L.E., 1978, RESOP II Reservoir Operating and because of the relative storage capacities and location of Quality Routing program—Program documentation and tributary inflows. From the initial increase in concentra- users manual: Texas Department of Water Resources tions in Lake Buchanan (summer 1987) in response to Users Manual, 35 p. the period of saline inflow, the high-salinity water Garner, L.E., and Young, K.P., 1976, Environmental geology passed through the entire Highland Lakes series in of the Austin area—An aid to urban planning: Austin, about 3.5 years. University of Texas, Bureau of Economic Geology A mathematical mass-balance model was used to Report of Investigations 86, 39 p. simulate 10 years of inflow and movement of water Helsel, D.R., and Hirsch, R.M., 1992, Studies in environmen- through the Highland Lakes and to estimate the concen- tal science 49—Statistical methods in water resources: trations of dissolved solids, chloride, and sulfate for Amsterdam, Elsevier, 522 p. each of three hydrologic conditions—wet, average, and Lower Colorado River Authority, 1993, Water management dry. Simulated tributary inflow comprised the Colorado plan for the lower Colorado River Basin: Austin, Lower River, Llano River, and Pedernales River. The model Colorado River Authority, 128 p. was tested with 1983–92 data for each of the Highland Miertschin, J.D., and Armstrong, N.E., 1986, Evaluation of Lakes to determine the adequacy for simulating the the effects of point and nonpoint source phosphorus three hydrologic conditions. The errors increased in loads upon water quality in the Highland Lakes: Austin, downstream order for each reservoir, but the model was University of Texas, Center for Research in Water judged to be acceptable for comparing differences in Resources Technical Report CRWR 215, 449 p. simulated concentrations of dissolved solids, chloride, Slade, R.M. Jr., and Buszka, P.M., 1994, Characteristics of and sulfate during the three conditions. streams and aquifers and processes affecting the salinity The model results could be improved by account- of water in the upper Colorado River Basin, Texas: ing for the additional quantity and quality of the minor U.S. Geological Survey Water-Resources Investigations tributaries and intervening areas to the Highland Lakes. Report 94–4036, 81 p. The mass balance of the water quantity would be Texas Natural Resource Conservation Commission, improved from an increase in the simulated water added 1997, Texas surface water quality standards, chapter to the system, which would provide more water to dilute 307: Austin, Texas Natural Resource Conservation the simulated concentrations of dissolved solids, chlo- Commission, Texas Administrative Code, Title 30, ride, and sulfate. Part I, 307.1–307.10. The simulated median monthly concentrations Tovar, F.H., and Maldonado, B.N., 1981, Drainage areas of during the 10-year simulation period for each reservoir Texas streams, Colorado River Basin: Texas Department generally are larger for the average condition than of Water Resources LP–145, 36 p.

44 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92 Table 7. Absolute error and error between measured and simulated monthly reservoir storage and concentrations ofTable dissolved 7 solids, chloride, and sulfate in the Highland Lakes, Texas, 1983–92 [absolute error = |(simulated – measured)/measured|*100; error = (simulated – measured)/measured)*100] Absolute error Error Reservoir (percent) (percent) Mean Median Minimum Maximum Mean Median Minimum Maximum Reservoir storage Lake Buchanan 3.0 2.7 0 32.4 –0.3 –0.6 –10.6 32.4 Inks Lake 1.8 .9 0 18.4 –1.0 –.3 –18.4 3.9 Lake Lyndon B. Johnson .9 .5 0 5.7 .6 .4 –1.3 5.7 Lake Marble Falls 1.6 1.1 0 6.2 .1 –.1 –5.3 6.2 Lake Travis 4.7 2.9 0 55.0 –2.8 –2.5 –32.2 55.0 Lake Austin 2.4 2.1 0 10.8 2.2 2.1 –2.0 10.8 Town Lake 1.6 1.3 0 4.6 1.0 .9 –2.9 4.6 Dissolved solids Lake Buchanan 14.1 14.2 .4 50.9 11.3 12.2 –27.0 50.9 Inks Lake 17.6 16.0 .1 59.4 17.1 16.0 –11.0 59.9 Lake Lyndon B. Johnson 21.5 16.3 .8 131 21.0 16.3 –9.6 131 Lake Marble Falls 21.3 16.3 .3 147 20.7 16.1 –18.6 147 Lake Travis 27.0 27.5 .1 69.0 25.8 27.5 –16.7 69.0 Lake Austin 34.2 31.2 6.2 104 34.2 31.2 6.2 104 Town Lake 35.7 34.3 .4 74.6 35.7 34.3 .4 74.6 Chloride Lake Buchanan 16.0 13.6 .4 61.3 13.1 11.3 –31.8 61.3 Inks Lake 18.7 15.8 .1 113 16.3 13.6 –27.0 113 Lake Lyndon B. Johnson 27.2 18.3 .3 281 23.8 17.4 –18.7 281 Lake Marble Falls 23.0 16.1 .2 132 19.9 13.8 –22.4 132 Lake Travis 28.5 24.4 0 107 25.6 23.0 –36.1 107 Lake Austin 40.1 33.9 1.0 161 38.5 33.9 –24.3 161 Town Lake 63.5 64.3 .6 128 63.5 64.3 .6 128 Sulfate Lake Buchanan 20.3 9.8 .1 159 18.0 6.7 –16.8 159 Inks Lake 22.6 13.6 0 163 19.9 11.4 –23.2 163 Lake Lyndon B. Johnson 38.9 26.8 .4 361 36.7 26.8 –15.0 361 Lake Marble Falls 30.9 21.4 .1 138 28.9 23.3 –38.5 138 Lake Travis 32.4 25.0 0 139 28.9 23.3 –38.5 139 Lake Austin 43.5 35.2 .5 185 42.2 35.2 –28.4 185 Town Lake 51.5 51.5 .5 101 51.5 51.5 .5 101

Table 8. Monthly demand distribution for Lake Travis1 [acre-ft,Table acre-feet] 8 Demand Demand Month Month Percent of annual demand Volume (acre-ft) Percent of annual demand Volume (acre-ft) Jan. 6.4 54,900 Aug. 12.4 106,300 Feb. 6.1 52,300 Sept. 9.4 80,600 Mar. 6.8 58,300 Oct. 8.2 70,300 Apr. 7.6 65,200 Nov. 6.8 58,300 May 8.1 69,400 Dec. 6.7 57,500 June 9.5 81,400 July 12.0 102,900 Annual 100 857,400 1 Modified from Lower Colorado River Authority, 1993.

Table 8 45 kes, Texas, Texas, kes, Volume and concentration and Volume Sulfate (mg/L) Chloride (mg/L) Chloride Dissolved solids (mg/L) solidsDissolved Reservoir storage Reservoir (acre-ft) Wet Average Dry Min Median Max Ending Min Median Max Ending Min Median Max Ending Initial Selected simulated monthly reservoir storage and concentrations of dissolved solids, chloride, and sulfate in the Highland La Reservoir Table 9 Table 9 Lake Buchanan LakeInks B. JohnsonLyndon Lake Lake Marble Falls 138,500Lake Travis 963,000 138,500 442,000 138,500Lake Austin 922,000 LakeTown 138,500 17,100 992,000 8,760 138,500 987,000 17,100Lake Buchanan 138,500 8,760 1,290,000 17,100Inks Lake 138,500 644,000 39,900 8,760 981,000 B. JohnsonLyndon Lake 457,000 20,600 138,500 17,100 1,350,000Lake Marble Falls 138,500 20,600 992,000 824,000 17,100 3,490 8,760Lake Travis 101,000 20,600 332 138,500Lake Austin 3,490 424 30,800 17,100 138,500 8,760 LakeTown 405,000 20,600 17,100 315 3,490 138,500 1,270,000 306 20,600 8,760 326 138,500 0 217,000 420Lake Buchanan 17,100 368,000 408 3,490 8,760Inks 20,600 Lake 17,100 440 318 992,000 311 B. JohnsonLyndon Lake 20,600 266 3,490 17,100 47,300 Lake Marble 8,760 0 Falls 943 280 1,010 414 345,000 446 17,100 20,600Lake Travis 269 3,490 281 8,760 1,270,000 81 20,600Lake Austin 415 272 410 107 17,100 3,490 1,020 LakeTown 946 394 273 8,760 17,100 20,600 397 75 0 80 338 20,600 412 286 102 85 3,490 8,760 419Lake Buchanan 398 735 LakeInks 3,490 175 99 20,600 518 741 566 B. JohnsonLyndon Lake 8,760 82 76 308 332 48 20,600 414 740 Falls Marble Lake 3,490 8,760 48 415 1,010 TravisLake 116 100 344 314 887 564 532 45 3,490 416 49 48 AustinLake 269 68 49 340 LakeTown 103 101 342 272 1,020 362 315 3,490 47 898 90 47 464 273 46 61 90 58 469 3,490 230 323 104 102 321 44 58 347 90 470 75 225 756 47 78 716 62 30 543 226 762 151 250 125 31 50 59 338 375 227 762 92 300 75 1,010 34 79 379 31 285 716 92 867 153 126 345 548 288 31 381 92 269 236 76 259 31 49 1,020 67 77 432 300 272 60 58 49 68 362 290 870 484 273 103 250 47 76 68 491 43 78 182 102 32 456 489 54 59 183 41 56 732 103 228 88 183 741 68 113 44 36 230 232 137 68 366 740 46 62 229 383 66 68 235 88 113 286 67 31 378 230 136 345 281 31 68 43 37 43 237 31 300 71 42 96 49 362 281 71 123 47 44 71 39 44 124 183 23 102 46 185 124 203 183 99 185 48 47 27 205 49 205 68 101 181 50 286 230 74 27 31 72 230 300 24 65 31 91 93 70 27 93 160 161 162 47 51 50 for wet, average, and dry wet, average, hydrologicfor conditions, 1987–96 acre-feet;per milligrams mg/L, [acre-ft, liter] Table 9.

46 Characterization and Simulation of the Quantity and Quality of Water in the Highland Lakes, Texas, 1983–92