Long-term Increases in Oxygen Depletion in the Bottom Waters of Boulder…

69

Lake and Reservoir Management 23:69-82, 2007© Copyright by the North American Lake Management Society 2007

Long-term Increases in Oxygen Depletion in the Bottom Waters of Boulder Basin, Lake Mead, Nevada-

Arizona, USA

James F. LaBounty and Noel M. Burns*

Southern Nevada Water Authority 1900 East Flamingo Road, Suite 255

Las Vegas, Nevada 89119 (702) 822-3357, [email protected]

*Lakes Consulting 42 Seabreeze Rd.

Devonport, New Zealand 0624 (09) 445-7561, [email protected]

Abstract

LaBounty, J.F. and Burns, N.M. 2007. Long-term increases in oxygen depletion in the bottom waters of Boulder Basin, Lake Mead, Nevada-Arizona, USA. Lake Reserv. Manage. 23:69-82.

Long-term changes in the hypolimnetic volumetric oxygen demand (HVOD) of Boulder Basin, Lake Mead were determined from dissolved oxygen profiles collected from 1991 to 2007. HVOD is the rate at which oxygen in a deep layer in contact with the sediments is depleted during the period of thermal and/or chemical stratification. Generally, the rate at which oxygen is depleted is correlated to the amount of organic debris in the hypolimnion and sediments. The sediment oxygen demand reflects historical organic loading, while HVOD is a measure of productivity because of the organic particles settling from above. The lower hypolimnion in Boulder Basin remains relatively stable during the stratification period, enabling the calculation of HVOD in the near-bottom water layer. Small increases and/or decreases that occur in temperature and dissolved oxygen concentrations are detectable. Boulder Basin fully destratifies every other year on average, but mixes only partially in the spring (before May) of the remaining years. The HVOD rates after partial and complete destratification have been assessed separately for 1995-2005. The an-nual HVOD rate is generally lower the year after partial destratification than after complete destratification due to greater downward transport of oxygen into the hypolimnion. The HVOD of Boulder Basin is variable depending on loading of nutrients and water into the Basin. The rate dropped significantly following commencement of advanced wastewater treatment practices in 1994. The rates then increased 1996-2006 at a rate of approximately 0.75 mg DO/m3/day per year, or about 7% annually. During those years the inputs of nutrients steadily increased. Rates have been dropping from 2005 to present (2007) following further reduction of phosphorus input. A multiple regression analysis revealed that HVOD is significantly positive related to the total phosphorus concentration in Las Vegas Bay, but significantly negative to inflows of Colorado River water. That means HVOD was highest when reservoir water was nutrient-rich and flow rates were low. HVOD should be considered a major tool for monitoring trophic state changes in Boulder Basin.

Key Words: applied limnology, dissolved oxygen, HVOD, hypolimnion, Lake Mead, phosphorus, reservoir ecology, temperature

A fraction of the organic material produced by primary pro-duction in the epilimnion of lakes settles into the hypolimnion where it decomposes and causes a hypolimnetic oxygen deficit. Changes in the hypolimnetic oxygen depletion rates over long periods can be indicative of overall changes of productivity and lake trophic level (Lasenby 1975, Walker

1979, Wetzel 2001, Beutel 2003). Development of the theory of oxygen deficits as a measure of lake health has a long his-tory since the fundamental work by Birge and Juday (1911), Thienemann (1928), and Juday and Birge (1932). The relative areal deficit (AHOD) was introduced by Strøm (1931) and modified by Hutchinson (1938, 1957), and was recently used

LaBounty and Burns

70

by Matthews and Effler (2006). Because the AHOD is the sum of the oxygen consumed under unit area of hypolimnion, it is sensitive to changes in hypolimnion thickness. In contrast, the volumetric oxygen deficit, the rate of disappearance of oxygen from unit volume of the hypolimnion, or hypolimnetic volumetric oxygen demand (HVOD), is less dependent on hypolimnion thickness. The HVOD is an indirect measure of epilimnetic primary production, provided no significant primary production is occurring in the hypolimnion (Burns et al. 2005). HVOD rates can also be modified by oxygen transported downward into the hypolimnion. Significant and variable amounts of DO transferred from the thermocline into the hypolimnion must be determined if HVOD rates are to be meaningful indicators of change in trophic level. The occurrence of these processes has been documented in Lake Erie (Rosa and Burns 1987, Burns et al. 2005). HVOD has been used repeatedly to investigate changes in the health of lakes and reservoirs (Burns 1995, Burns et al. 2005, Charlton 1980, Cornett and Rigler 1979, Hieskary and Wilson 2005, Lasenby 1975, Vincent et al. 1984).

The primary focus of this investigation was to evaluate long-term trends of HVOD, phosphorus concentrations and inflows in Boulder Basin, and determine the causes of variability. The most complete sets of annual data are from May 2000 to 2007. A previous analysis of data from 2000 to 2004 was reported in LaBounty and Burns (2005). The oxygen data collected during that time have now been combined with additional data from 1991 to 2007 to calculate 16 annual HVOD rates. Drastic changes in annual HVOD rates required arranging the rates into three groups. Rates from 1995 to 2006 are further used to elucidate potential consequences when various sce-narios of environmental conditions occur. Downward mixing of oxygen into the hypolimnion was estimated for each of the eleven stratified seasons from 1995 to 2005, and the effects on the observed HVOD rates determined.

Materials and MethodsDescription of Study AreaLake Mead is a large mainstream Colorado River reservoir in the Mohave Desert, Arizona-Nevada (Fig. 1). Its lower end is 15 km east of Las Vegas, Nevada. Lake Mead, formed in 1935 following construction of Hoover Dam, is the largest reservoir in the United States by volume (36.7 × 109 m3, 2.975 × 107 ac-ft)1, and is second only to Lake Powell in terms of surface area (660 km2; Lara and Sanders 1970). At full pool (reservoir elevation 374 m a.s.l.), Lake Mead extends 106 km from Black Canyon (Hoover Dam) to Pearce Ferry. Its greatest width is 15 km, and the highly irregular shoreline is 885 km in length. Lake Mead has 4 large sub-basins: Boulder,

Virgin, Temple, and Gregg. Between these basins are 4 nar-row canyons: Black, Boulder, Virgin, and Iceberg (Fig. 1).

Retention time in the reservoir ranges between approximately 1 to 3 years, depending on release and inflow patterns as well as reservoir volume at any one time. The Colorado River contributes about 97% of the annual inflow to Lake Mead; the Virgin and Muddy rivers and Las Vegas Wash provide the remainder. Inflow to Lake Mead is controlled by the amount of water released from Glen Canyon Dam. The minimum release from Glen Canyon dam is 10.2 × 109 m3 yr-1, the 10-year average is <18.5 × 109 m3 yr-1. Annual inflow via Las Vegas Wash is currently about 1.9 × 108 m3, about 1.5% of total inflow to Lake Mead. Discharge from Hoover Dam is hypolimnetic and occurs from 2 intakes, one at elevation 272.8 m and the other at 318.5 m a.s.l. Annual discharge is approximately 9 × 109 m3. During the time of this investigation the 2 intakes for the Southern Nevada Water System (SNWS) were located at 304.8 m and 320 m a.s.l until mid-2004, when they were combined at 304.8 m a.s.l. Annual withdrawal through the SNWS in Boulder Basin is currently about 5.5 × 108 m3.

LaBounty and Horn (1997), LaBounty (2005), LaBounty and Burns (2005) and Holdren et al. (2006) describe the hydro-dynamics and relationships between various limnological variables of Boulder Basin. As with other reservoirs, dam operations exert a great influence on the water quality and ecology of the system (Thornton 1990). The hydrodynamics of such large reservoirs are complex and require analysis of large data sets to understand them. Each basin within Lake Mead is ecologically unique and therefore responds differ-ently to the inflow-outflow regime. Furthermore, the different sources of water entering Lake Mead, as in other reservoirs,

1 1 ac-ft = 3.259 × 105 gal

nisaB redluoB

mrA notrevO

nisaB nigriV

nisaB elpmeT

ynaC redluoB nohT“ e ”sworraN

nisaB ggerG

geV saL a hsaW s

reviR nigriVreviR apaoM

eviR odaroloC r

ynaC kcalB( maD revooH )no

hsaW dnarG

reviR odaroloC

becI ynaC gre no

noynaC nigriV

F ecraeP e yrr

Figure 1.-Location map of Lake Mead, Arizona-Nevada, with an insert of a map of the Colorado River Watershed in the southwestern U.S.

Long-term Increases in Oxygen Depletion in the Bottom Waters of Boulder Basin, Lake Mead, Nevada-Arizona, USA

71

often retain their identity and influence for substantial dis-tances into the reservoir and do not necessarily mix com-pletely with the rest of the water column (Ford 1990). This spatial heterogeneity can lead to significant underestimates of water retention time and uncertainties in the conveyance and fate of materials transported into the reservoir.

Boulder Basin (Fig. 2) is the most downstream basin and collects the combined flows from the reservoir’s 2 main arms. Additionally, it receives all drainage from the Las Vegas Valley via Las Vegas Wash into Las Vegas Bay. Boulder Basin is 15-km wide from Boulder Canyon to Hoover Dam (Black Canyon), and the distance from the confluence of Las Vegas Wash to Hoover Dam is approxi-mately 16 km. The historical Colorado River channel lies along the eastern side of the basin. Lake Mead is mostly mesotrophic and monomictic (LaBounty and Burns 2005). Boulder Basin destratifies completely about every other year. Stratification lasts for approximately 265 days from 1 May to 15 January during years when complete destratification occurs. During the stable summer stratification period the epilimnion, metalimnion and hypolimnion are about 15, 20, and up to 120 m thick respectively, depending on the depth of the reservoir and particular location. The hypolimnion is the thickest and most stable layer, with an average water temperature of 12°C. The deepest portion of Boulder Basin does not completely destratify each year, leaving a remnant portion of the hypolimnion >20 m above the bottom in about 40% of the years (LaBounty and Burns 2005, Holdren et al. 2006). Ecological dynamics until early 2007 are described in this paper. Significant populations of the quagga mussel (Dreissena bugensis) were discovered throughout Boulder Basin in January 2007. As populations of this invasive species explode, the ecological dynamics of Boulder Basin could be drastically altered.

Profile data used for the reported analyses were collected by 3 agencies: the City of Las Vegas, Bureau of Reclamation, and the Southern Nevada Water System. HVOD analysis was performed using data from the hypolimnetic portion of 8 deep-water sampling locations within the Colorado River thalweg of Boulder Basin from the Narrows to Hoover Dam (Fig. 2). Data from 1991 through April 2000 are from 2 deep-water sampling sites, CR346.4 and CR342.25. These locations were sampled twice monthly from March through November and at least monthly from November through April. Data collected after May 2000 through early 2007 are from all stations shown in Fig. 2. Each location is sampled at least once per month; many are sampled once or twice weekly by one or more of the participating agencies.

Field and Analytical Methods, and Data AnalysisEach profile was sampled in a similar manner. Data were collected from the following depths: surface-30 m every 1 m; 30-60 m every 2 m; 60-100 m every 5 m; 100-5 m from the bottom every 10 m; and from 5 m-bottom at least every 1 m (Fig. 3). Each profile includes temperature and DO. Profile data were collected using profiling sondes manufactured by Hydrolab Corporation (1995-early 2007) and Eureka Corpo-ration (mid-2005-early 2007). Repeatability of DO measure-ments is within 0.2% and instrument accuracy = 0.1 mg/L for readings <8 mg/L, 0.2 for readings >8 mg/L. Air calibration is electronic, utilizing the barometric reading at the beginning of the survey. Quality control was accomplished by annual comparison of up to 11 probes from participating data col-lection groups at the same time and place.

- L/gm OD - D 01/taS% O98765432

°( erutarepmeT )C220281614121

0noinmilipE

Dep

th (m

)

001

08

06

04

021

02

noinmilateM

cissalCyH noinmilop

erutarepmeT01/ % ODL/gm OD

yH noinmilopdesu ni siht yduts

Figure 3.-Temperature and dissolved oxygen profiles from sampling location CR346.4 on 20 October 2005.

Figure 2.-Map of Boulder Basin, Lake Mead, Arizona-Nevada. The sampling locations for this investigation are depicted.

-erP snoitacol elpmas 0002 tsoP dnasnoitacol elpmas 0002 tsoP

Las Vegas Bay

hsaW sageV saL

Colo

rado

Rive

r Flo

wWash Flow

maD revooH

LaBounty and Burns

72

Water samples were collected weekly for analysis of nutrients including total phosphorus (TP). Samples were collected us-ing a Van Dorn sampler at all depths except from the surface, where grab samples were collected. Two samples were taken within the epilimnion when it was ≤15 m thick; 3-4 samples within the thermocline, including one at peak conductivity; and up to 4 samples within the hypolimnion. Analyses for TP and orthophosphate were by the USBR laboratories in Den-ver, Colorado and Boulder City, Nevada, the Clark County Water Reclamation District Water Quality Laboratory, Las Vegas, Nevada, and Montgomery Watson Harza Laborato-ries, Monrovia, California. Standard Method SM4500P-E, colorimetric analysis was used for most analyses. Detection limits were 10 µg P/L for TP. All values that were nondetect-able were halved for calculations of averages.

Processing of data, statistical analysis and development of graphical presentations were completed using Excel, PowerPoint (www.microsoft.com), and LakeWatch (www.earthsoft.com; www.lakewatch.net).

Hypolimnetic Volumetric Oxygen Depletion (HVOD) CalculationsThe HVOD rates for Boulder Basin were determined using the HVOD module in the LakeWatch software. To determine the most adequate method, HVOD analysis of the bottom layer was attempted using 3 methods. In the first, the hypo-limnion layer was selected from the normal lower inflexion point on the temperature/depth profile. In the second and third methods, the bottom layer depths were selected as being 5 m and 20 m, respectively, above the bottom. The first method was rejected because the hypolimnion water mass under analysis changes as the selected hypolimnion depth deepens as the season progresses. The bottom 5-m layer did not yield enough observations to give stable values. Consequently, the bottom 20-m layer was chosen for analysis because it gave stable values from an unchanging water mass and contained water with the lowest DO concentrations in the water column. These values could thus show how low the oxygen concentrations became at the end of each stratified season. The average DO content for the bottom layer was calculated for each sampling station on each sampling date. In the remainder of this study, the term hypolimnion refers to the 20-m layer above the bottom (Fig. 3).

HVOD rates were determined according to the stratification conditions prevailing each year. The numbers of days the reservoir remained stratified each year was determined by examining water temperature and DO from the initiation of stratification (usually 1 May) to the point of its interruption (usually 15 January, or 265 days later). The stratification period lasted from 190 to 280 days. The annual oxygen deple-tion rate was calculated by plotting the average hypolimnetic DO concentration sampled at each station as a function of

time (days), and then these data were fitted to a straight line by linear regression analysis. This annual rate of DO loss (HVOD) is expressed in mg/m3/day. Regressions were cal-culated at the observed water temperatures and then corrected to a standard water temperature of 12°C for all years. This method was chosen because the HVOD rate is presumed to double for every 10°C rise in temperature, as rates do in many biochemical reactions. Standardization is necessary for the comparison of data from different years because they have different average temperatures. The temperature of 12°C was chosen as the standard temperature because it was close to the average temperature observed for all years. The annual HVOD rates were plotted separately for 1995-2006, and linear regressions were obtained for each plot for 2 types of years: Type 1 year follows complete destratification; Type 2 follows partial destratification (as described in the follow-ing section). The slopes of the regression lines gave the rate of change of HVOD in Boulder Basin in mg DO/m3/day per year. The strength of relationships between HVOD and other reservoir characteristics was determined by simple and multiple regression analysis.

ResultsVertical MixingWhile temperature and DO of the epilimnetic layer are sea-sonally variable, within the hypolimnion they change slowly throughout the stratified period. During fall the epilimnion thickens due to seasonal cooling. From May to early October the epilimnion layer of Boulder Basin is generally 11-13 m. In early October cooling begins to thicken this layer. For example, on 20 October 2005 the epilimnion was 25 m (Fig. 3). Despite cooling temperatures in fall 2005, neither the water temperature nor the DO profiles indicated strong vertical mixing of the hypolimnion. A typical metalimnetic DO minimum was prominent, while the concentrations within the hypolimnion steadily decreased with depth until there was a sudden drop of >2 mg/L within 5 m of the bottom. Below 90 m, temperature cools with depth into the hypo-limnion by approximately 0.1°C per 20 m. This temperature gradient is maintained from May to December indicating a very stable hypolimnion. At the same time the bottom 20 m water temperatures increased by about 0.3°C throughout the stratified season, indicating a small, steady downward transfer of heat (Fig. 4).

There was a typical annual pattern of temperature and DO within the bottom 20 m of the hypolimnion (Fig. 4) indicating 4 phases of vertical mixing:

Phase 1. During the first 220 days of stratification (normally 1 May to about 6 December) the average water temperature of the bottom 20 m of the hypolimnion gradually but steadily warmed from 11.9°C to 12.2°C, or 0.3°C/220 days. At the

Long-term Increases in Oxygen Depletion in the Bottom Waters of Boulder Basin, Lake Mead, Nevada-Arizona, USA

73

same time the average DO concentration steadily declined from 5.2 mg/L to 2.1 mg/L, or -3.1 mg/L/220 days.

Phase 2. From day 220 to day 300 (about 6 December to 24 February), warming of the bottom 20 m of the hypolimnion accelerated, and partial vertical mixing occurred as indicated by the reoxygenation from 2.1 to 8.5 mg/L in the early part of Phase 3. Water temperature increased nearly 0.3°C dur-ing this phase (12.2°C to about 12.5°C). The surface water temperature at the end of February 2006 was 13.2°C, so when the surface water mixed downward there was relatively little increase in temperature because the surface water tempera-ture was close to that of the deep water, while at the same time there was considerable downward transfer of DO by vertical mixing. However, complete vertical mixing was not achieved because a bottom water temperature of 13.2°C was not obtained. Complete vertical mixing on this date proceeded only to a depth of 46 m.

Phase 3. After day 300 (about 24 February) intrusion of low temperature, low conductance, higher DO Colorado River water from Boulder Canyon was observed, lowering the water temperature suddenly from 12.4 to 12.1°C.

Phase 4. After 1 May (day 1) the cycle began as stratifica-tion strengthens. Oxygen uptake in the bottom 20 m of the hypolimnion occurred.

Fully Destratified (Type 1) Versus Partially Destratified Years (Type 2)Lake Mead is a monomictic lake with turnover (destratifica-tion) occurring generally from 1 March-1 May (LaBounty and Burns 2005). Since stratification is usually complete by 1 May, the “lake year” for every period of 365-days is de-fined as extending from 1 May to 30 April, and thus spans 2 calendar years. Lake years are named from the second of the 2 years spanned. The Type 1 group are those years follow-ing the complete destratification of the reservoir (lake years 1997, 1999, 2001, 2002, 2004, and 2006). The Type 2 group includes those years following a year when the reservoir did not fully destratify (lake years 1995, 1996, 1998, 2000, 2003 and 2005). The pair of years plotted in Fig. 5 (May 2002 through April 2004) are Type 1 and 2, respectively. The Type 2 group of years have higher hypolimnion temperatures and usually lower DO concentrations than the Type 1 years (Figs. 6 and 7).

Study of numerous profiles of temperature, DO and specific conductivity show that Boulder Basin has normal vertical temperature and DO structure for most of the year, with the thermocline deepening after mid-November. In some years the downward mixing progresses only to a depth of 70 m to give a partially destratified year (Type 2 year). In other years, downward mixing proceeds to the lake bottom to give

erutarepmeT ylhtnoM egarevA

Tem

pera

ture

(°C

) 5.21

21

5.11

11

5.01

600250024002300220021002000299918991799169915991

laitraP gniwolloFnoitacifitartseD

epyt 2 raeY

lluF gniwolloFdeifitartseDepyt 1 raeY

negyxO devlossiD ylhtnoM egarevA60025002400230022002100200028991799169915991 9991

7002

7002

Dis

solv

ed O

xyge

n (m

g/L) 21

01

8

6

4

2

0

Figure 6.-Average daily water temperature and dissolved oxygen (DO) concentration of the bottom 20 m of Boulder Basin, 1995-2007.

noitacifitartseD

L/gm OD

C° erutarepmeT

003002001

4.21

21

6.11

2.11

003002001

9876543

Dis

solv

ed O

xyge

n (m

g/L)

3002 40021 yaM 1 naJ 1 naJ1 yaM

20021 yaM

Temperature °C

noitanegyxoer emoS

Figure 5.-Average daily temperature and dissolved oxygen (DO) concentration in the bottom 20 m of the hypolimnion of a partially destratified year (Type 2) followed by a fully stratified year (Type 1), 1 May 2002 to 1 May 2004.

Figure 4.-Average daily water temperature and dissolved oxygen (DO) concentrations from the bottom 20 m of the hypolimnion, 1 May 2005 to 30 April 2006.

06302308204200206102108040

98765432

Temperature ºC

4.21

3.21

2.21

1.21

21

Dis

solv

ed O

xyge

n (m

g/L)

1 yaM morf syaD fo rebmuN1 yaM 03 lirpA

1

2

3

L/gm OD

Cº pmeT

4

6 ceD 42 beF

1 w wolS a – gnimr W trev kae i htiw gnixim lac refsnart OD dna taeh drawnwod thgils

2 – gnimraw retsaF P tra i refsnart OD emoS ;gnixim la

3 ;retaw reviR odaroloC fo noisurtnI tarutas %08 ot noitanegyxoeR i no

4 Stratification strengthens-Oxygen uptake recommences

LaBounty and Burns

74

menced with a Colorado River intrusion of 11.1°C and 980 µS/cm water on 12 April 2004.

Average monthly temperatures of the bottom 20 m of the hypolimnion ranged between 11 and 12.6°C from 1995 to early 2007 (Fig. 6). Holdren et al. (2006) stated that mixing occurs in Boulder Basin when early spring hypolimnetic temperatures are above 12°C; mixing does not occur below 12°C. When destratification is not complete between strati-fied periods (black arrows, Fig. 6), the water temperature continues to rise until destratification does occur (red arrows, Fig. 6). Full mixing that results in bottom waters with the same temperatures as the surface waters occurs only after the temperature difference between the surface and bottom waters is <1°C. Once destratification occurs, its onset is sud-den and lasts 30-60 days. The pattern for DO concentrations reflects that of temperature. In years when destratification was not complete (black arrows), DO concentrations remained generally below 6 mg/L. Upon destratification, however, reoxygenation to saturation near 80-85% was achieved (red arrows). A similar pattern, but the reverse of that of tem-perature, existed for average DO concentration from 1 May 2002 to 30 April 2004 (Fig. 5). Average DO concentrations were 8.2 mg/L on 1 May 2002 and 5.1 mg/L 265 days later in mid-March 2003, a loss of 3.1 mg/L DO in the bottom 20 m of the hypolimnion. DO depletion continued after some oxygenation in late April and early May 2003. The average DO concentration was 5.8 mg/L in May 2003 and ended at 3.6 mg/L in mid-January 2004, about 245 days later, a loss

a fully destratified year (Type 1 year). But Boulder Basin has a unique mixing pattern in one respect. In a normal lake, the temperature at the bottom of the lake at the end of the isothermal period is the lowest temperature observed in the water column that year. However, in Boulder Basin, when the surface water begins to warm after the isothermal period, the hypolimnion from a depth of about 95 m can continue to get colder. This is attributed to a large pool of cool water from the Colorado River slowly moving into Boulder Basin into the bottom of the deepest part of the Basin and usually results in the deeper hypolimnion having a temperature of ≤12°C. The process of an intrusion of Colorado River water into Boulder Basin takes 2 months to arrive from the Colorado River inflow (Dr. Imad Hannoun, Flow Science, Harrisburg, VA., pers. comm.). The intrusion (interflow) of Colorado River water forming the deeper hypolimnion is usually complete by 1 May and sets up a Type 1 year. During the remainder of the year, until the end of Novem-ber when the Colorado River interflow is found within the upper portion of the hypolimnion or in the metalimnion, the temperature in the bottom hypolimnion layer increases slowly as the DO concentrations decrease (Fig. 4 and 5). A Type 1 year is usually followed by a Type 2 year with some downward mixing and/or some Colorado River water being added to the deep hypolimnion between February and May, with DO concentration increasing but with little change in temperature (Fig. 6). At the end of the Type 2 year the hy-polimnion temperature is usually above 12°C enabling full vertical mixing (Holdren et al. 2006), illustrated by the data on a biennial period (Fig 5).

The water in Boulder Basin has a higher conductance than the water from the Colorado River because of evaporation and the higher conductivity water from Las Vegas Wash and the Virgin and Muddy Rivers. The first year (1 May 2002-30 April 2003) of a pair of years (Fig. 5) follows a complete destratification during February-April 2002. At Station CR346.4 on 12 February 2002, the water was isothermal at 11.9°C with a conductance of 935 µS/cm, but by March the Colorado River intrusion water had changed the tempera-ture to 11.1°C with a conductance of 919 µS/cm to set up a Type 1 year. The second year (1 May 2003-30 April 2004) followed partial destratification. Stratification weakened between January and April 2003 but remained present. It was weakest on 4 March 2003 with a surface temperature of 12.8°C and a bottom temperature of 11.6°C. Average bottom water temperature began at about 11.1°C on 1 May 2002 and was 11.8°C on 30 April 2003, a rise of 0.7°C in 365 days. From 1 May 2003 to about 1 February 2003 (275 days) the temperature rose from 11.8°C to about 12.4°C, or 0.6°C. The temperature rose at a rate 12.1% greater in the second year. After 1 February 2004 (within the second year) destratification, or complete mixing, occurred with the water column isothermal at 12.2°C and conductance between 998 and 1000 µS/cm on 13 February 2004. A Type 1 year com-

noitelpeD ODnoissergeR

erutarepmeTnoissergeR

8

7

6

5

4

8.117.116.115.114.113.112.11

– 4002 etaR noitelpeD OD )1 epyT( 5002

DO

(mg/

L)

Temperature C

1 yaM 6 luJ 01 tpeS 12 naJ51 voN0420120810510210906030

Temperature C

DO

(mg/

L)

2.21

1.21

0120810510210906030

5

4

3 21

1 yaM 6 luJ 01 tpeS 51 voN

– 5002 etaR noitelpeD OD )2 epyT( 6002

0

Figure 7.-Dissolved oxygen depletion (HVOD) rate (regression = -0.01375 mg DO/L/day; p < 0.0001) and temperature increase (regression = +0.00112°C/day; p < 0.0001) for (top panel) the first 210 days of 2005 (beginning 1 May 2005), and (bottom panel) the first 265 days 2004 (beginning 1 May 2004) dissolved oxygen depletion (HVOD) rate (regression = -0.01471 mg DO/L/day; p < 0.0001) and temperature increase (regression = +0.00112°C/day; p < 0.0001).

Long-term Increases in Oxygen Depletion in the Bottom Waters of Boulder Basin, Lake Mead, Nevada-Arizona, USA

75

of 2.2 mg/L. The depletion rate for DO was 23.2% greater during the first year than the second.

Annual HVOD RatesThe flow into Las Vegas Bay of Boulder Basin consists mostly of treated wastewater from the Las Vegas Valley. Since 1994, several improvements to waste water treatment (i.e., tertiary treatment of phosphorus and nitrogen) were made, affecting Las Vegas Bay. As a result of this and other external factors (such as above vs. below average inflow to Lake Mead), certain patterns form individual groups of years, each having one or more major events that cause them to be considered a separate group no matter which years are included in other groups. Accordingly, annual depletion rates for lake years 1991-2006 are in 3 distinct groups, each with a major event related to changes in treatment of nutrients released into the Las Vegas Wash and thus into Las Vegas Bay (Table 1; Fig. 8).

Group 1, 1991-1996, includes the initiation of advanced wastewater treatment in 1994. The early HVOD rates were the highest of the 15 years (15.1-18.1 mg/m3/day)2. Group 2, 1996-2004, begins with the lowest annual rates (7.5-8.1 mg/m3/day), and ends with an annual rate of 14.7 mg/m3/day. This group includes the commencement of tertiary treatment of nutrients in discharged wastewater as well as a period of abnormally high inflow to Lake Mead (1995-1999). While average outflow from Glen Canyon Reservoir was always <335 m3/day in 1991-1994 and 2000-2006, average outflow in 1995-1999 was >400 m3/day (peaking at about 600 m3/day in 1997). There was ever-increasing organic loading during this time period due to the population surge in the watershed leading to steadily increasing flows of treated wastewater via Las Vegas Wash. Group 3, 2005-2007, shows declining annual rates (14.7-11.6 mg/m3/day). Advanced treatment of phosphorus in the discharged wastewater effluent is not officially regulated from October through March. However, beginning in winter 2002 treatment of phosphorus in the wastewater outflows from all facilities occurred much more frequently from October to April. Additionally, in 2005 enhanced treatment for phosphorus occurred that further reduced the total load of organics to Boulder Basin. All these efforts reduced the current TP load by 97-98% from levels from the 1970s and 1980s (Dr. Doug Drury, Clark County Water Reclamation District, Las Vegas, Nev., pers. comm.).

Annual HVOD Rate for Years of Type 1 and 2Data for 1995-2005 were selected for analysis and clas-sification into Type 1 and 2 years, because even though the

2 HVOD rates at standard temperature of 12°C

Table 1.-Annual dissolved oxygen depletion rates for 16 years, 1991–2006, Boulder Basin, Lake Mead, Nevada-Arizona.

DO on Ending Annual Lake Type May 1 DO Rate2 Year1 of Year (mg/L) (mg/L) (mg/m3/day)

1991 1 8.8 5.6 18.11992 1 8.4 4.9 17.21993 1 9.0 5.1 15.11994 1 8.8 5.1 17.41995 2 5.2 3.6 8.11996 2 5.5 4.1 7.51997 1 8.4 6.3 9.31998 2 8.2 6 9.61999 1 8.1 5.7 10.72000 2 7.1 4.8 112001 1 8.2 5.2 11.82002 1 8.2 5.1 14.12003 2 5.8 3.6 11.72004 1 7.8 4.2 14.72005 2 5.1 2.7 13.72006 1 8.8 4.8 11.6

Average 1 8.5 5.2 14.0Std Dev 1 ±0.4 ±0.5 ±2.9

Average 2 6.2 4.1 10.3Std Dev 2 ±1.2 ±1.1 ±2.3

1 May of named year to April of the next year2 The HVOD at standard temperature (12°C) regression presented in this

table

Figure 8.-Rate of HVOD change in the bottom 20 m of the hypolimnion of Boulder Basin, 1991-2006. Group 1 (1991-1996): high HVOD rates early (during period of high organic loading), dramatically lower rates later (following implementation of advanced wastewater treatment technology); Group 2 (1996-2004): progressively increasing HVOD rates due to rapid population expansion within the watershed leading to continuous increase in organic loading; Group 3 (2004-2006): beginning of a trend of decreasing HVOD rates due to decreased organic loading.

5002300210029991799139911991

DO

Dep

letio

n R

ate

(mg/

m3 /d

ay)

81

61

41

21

01

8

decnavdAtnemtaerT

snigeB

retniWtnemtaerT

snigeB

cnahnE detnemtaerT

P rofsnigeB

DOVHoissergeR n

pmeT dtS ta DOVHgeR r isse on

2 puorG

3 puorG

1 puorG

5991

G egarevA l ne noynaC ftuO wol > m 004 3 yad/

LaBounty and Burns

76

wastewater treatment remained the same, loading increased with population surges. Dissolved oxygen depletion pro-gresses at a steady rate, but there is a measurable difference between Type 1 and Type 2 years (Fig. 7). Disruption of the upward temperature trend and the downward DO depletion trend happens sometime after day 180 (Fig. 4, a temporal extension of the top panel of Fig. 7). In 2005, the pattern began falling apart after day 210 (6 December 2005). After complete destratification, the stable period lasted at least 265 days to 21 January.

Dissolved oxygen concentrations at the beginning and end of annual stratification were different for the two categories of years (Table 1). Each Type 1 year started and finished the stratified season with higher DO concentrations than Type 2 years. The HVOD rates for the Type 1 group (14.0 ± 2.9 mg/m3/day) are higher than those of Type 2 (10.3 ± 2.3 mg/m3/day). The annual DO depletion rates for 1995-2005 were plotted separately for Type 1 and Type 2 years (Fig. 9). Regressions were calculated and corrected to HVOD at 12.0°C and illustrate increasing trends since 1995 for both types of years. The annual rates, corrected to 12.0°C, increased by 0.91 mg/m3/day per year for the Type 1 years and 0.59 mg/m3/day per year for the Type 2 years. Each of these trends is statistically significant (p < 0.01).

HVOD Change Based on Type 1 and Type 2 YearsThe mixing period of larger lakes is usually annual if the lake is monomictic. Finding a biennial mixing period was unexpected, but even more unexpected was the discovery of different HVOD rates for the different types of years. A number of scenarios were investigated to determine the cause for the different HVOD rates. The first possibility was that different mixing regimes resulted in different organic carbon concentrations in the hypolimnia of the 2 different types of years, thus causing different depletion rates; this was found not to be the case, however; organic carbon concentrations were similar in the pairs of years (Type 1 years 2.94 ± 0.31 mgTOC/m3; Type 2 years 3.00 ± 0.30 mgTOC/m3).

Varying inflow rates and water temperature from Lake Pow-ell (Glen Canyon Dam outflow, Fig. 10) were investigated as possible causes of the different HVOD rates. The inflow from Lake Powell via Boulder Canyon dominates the water mass in Boulder Basin, and the volume and temperature of this inflow is quite variable. Both the volume/inflow rate and water temperature depend on the outflow from Glen Canyon Dam (Lake Powell). In general but not always, the temperature of Lake Powell outflow depends on the depth of Lake Powell. At lower lake levels, outflow temperatures are higher because water is removed from a shallower layer. The magnitude of the flow did not affect the type of year because in the 1995-1999 high flow period there were 2 Type 1 years

(Fig. 6; 1997, 1999) and 3 Type 2 years (Fig. 6; 1995, 1996, 1998), which means the distribution of year types was similar to the low flow period.

The possibility of variable downward transport of oxygen was next investigated. The full and partially destratified years tend to occur in pairs (Fig. 6). However, as Holdren et al. (2006) pointed out, and as we have shown, winter destratification was not completed for the 3-year period 1994-1996. Inflow of colder water into the hypolimnion as an interflow from the Colorado River occurred in those 3 years but also in the other years when destratification occurred. The resistance

DO

Dep

letio

n R

ate

(mg/

m3 /d

ay)

40023002200210020002999189917991

41

21

01

8

noitacifitartsed etelpmoc gniwollof sraeY :1 epyT

50024002300220021002000299918991799169915991

41

21

01

8

DO

Dep

letio

n R

ate

(mg/

m3 /d

ay)

DOVHnoissergeR

ta DOVH T dtS e pmnoissergeR

noitacifitartsed laitrap gniwollof sraeY :2 epyT

DOVHnoissergeR

ta DOVH T dtS e pmnoissergeR

Figure 9.-Rate of HVOD change. Top panel: Type 1 Category years; years that followed complete destratification (regressions: HVOD = +0.803 mg/m3/day per year, p < 0.01; HVOD at 12°C = +0.906 mg/m3/day per year, p < 0.05). Bottom panel: Type 2 category years; years that followed a short period of partial destratification (regressions: HVOD = +0.581 mg/m3/day per year, p < 0.01; HVOD at 12°C = +0.586 mg/m3/day per year, p < 0.01).

Figure 10.-Average daily outflow from Glen Canyon Dam (Lake Powell), 1991-2006. Outflow from Glen Canyon Dam travels through the Grand Canyon into Lake Mead unimpeded. Circled values are years when average flow from Glen Canyon Dam was >average.

maD noynaC nelG ,wolftuO yliaD egarevA

052

003

053

004

054

005

055

006

056

1991 3991 5991 7991 91 99 1002 3002 5002 7002

99 & ,89 ,79 ,69 ,5991> swolF ,egareva

riovreser m 11 esor

m3 /d

ay

Long-term Increases in Oxygen Depletion in the Bottom Waters of Boulder Basin, Lake Mead, Nevada-Arizona, USA

77

to vertical destratification is strongest in the first year of a pair because the water temperature of the hypolimnion is 1-2 degrees cooler than in the second year of the pair. When the water temperature rises during the second year of the pair, resistance to full vertical mixing weakens and full vertical mixing can occur.

From 1995 to 2005, the average HVOD rate for Type 1 years was 2.0 mg/m3/day higher than Type 2 years (Table 2). Further, the rate of increase in the HVOD rate was less for Type 2 years. We investigated the cause of the differences by determining whether there is a difference in the downward transfer of heat and DO between the two types of years.

Downward transfer of heat occurs by conduction from warmer, upper waters to cooler, deeper waters. Similarly, oxygen can move downward by molecular diffusion to the deeper waters containing less oxygen, 2 independent mecha-nisms. The conduction of heat through water is calculated by using the thermal conductivity coefficient of water (0.0014 cal/cm/sec) and the vertical temperature gradient, which at the depth of 20 m above the bottom is approximately 0.6 × 10-4°C/cm in Boulder Basin. This produces a warming in the 20-m thick hypolimnion of approximately 0.0017°C from 8 May-28 December of this particular year. In 2000, the bot-tom waters at the central sampling site (CR346.4) warmed by 0.3°C during this period. Thus the downward transfer of heat by conduction is of little importance in the warming of the hypolimnion.

The molecular diffusion of oxygen through the surface of the hypolimnion can be estimated by using the diffusion coef-ficient of oxygen in water of 2 × 10-9 m2/sec (Pilson 1998) and the average DO concentration gradient at the hypolimnion boundary of 0.20 g/m4. This yields a downward flux of DO of 3.5 × 10-3 mg/m2/day, which would increase the concen-tration of DO in a 20-m thick hypolimnion by 0.17 × 10-3 mg/m3/day. This is much smaller than the observed HVOD rates ranging from 8.1 to 14.7 mg/m3/day. Molecular diffu-sion of oxygen would be of little importance in explaining the difference of 2.0 mg/m3/day between the Type 1 and Type 2 HVOD rates.

During the stratified season, heat and DO mostly move downward in lakes and reservoirs by a vertical mixing process known as vertical eddy diffusion, a process found to be sig-nificant in Onondaga Lake, New York (Mathews and Effler 2006). In eddy diffusion, heat and DO are moved downward in the same water mass; thus, the increase of temperature in the hypolimnion is an indication of the quantity of DO transferred downward in the water. However, the increase in DO concentrations will not be observed if oxygen is being consumed via the degradation of organic matter at a greater rate than its downward displacement. If the eddy diffusion is not extensive and happens to a similar degree from the thermocline downward, it is not easily discernable except for

a small rise in temperature of the water column. The transfer of DO into the bottom 20 m of Boulder Basin by eddy dif-fusion was estimated by an equation successfully developed and applied in Lake Erie (Rosa and Burns 1987):

∆O = (Ot - Oh)/(Tt - Th) × ∆T eq. (1)

where ∆O = increase in DO concentrations if DO was con-served; ∆T = increase rate in hypolimnion temperature; Ot = average thermocline DO concentration for stratified period; Oh = average hypolimnion DO concentration for stratified period; Tt = average thermocline temperature for stratified period; and Th = average hypolimnion temperature for strati-fied period. The temperature increase rates in the bottom 20 m are small but can explain substantial transfer of DO into the hypolimnion (Table 2).

The temperature increase rate is similar for both Type 1 and Type 2 (Table 3), but because of the greater Ot - Oh gradi-ent in Type 2 years, eddy diffusion is able to transfer more oxygen downward than in Type 1 years. The extra oxygen added to the hypolimnion in Type 2 years significantly lowers the observed HVOD rate in those years. Also, the true (as adjusted) oxygen uptake rates are significantly higher than the observed HVOD rates. The rates of HVOD change are 0.59 mg/m3/day per year for Type 2 years and 0.91 mg/m3/day per year for Type 1 years. The increased downward diffusion also explains why Type 2 HVOD rates are increasing more slowly than Type 1 rates. As hypolimnetic DO concentrations become lower, the DO concentration gradient between the hypolimnion and the water above it becomes greater; there-fore, eddy diffusion transports more oxygen downward.

Factors other than hypolimnetic temperature prevent or enhance complete mixing after the second year of a pair of years: winter climatic factors such as ambient air tempera-ture, wind speed and duration; and as discussed, erratic high and low flow patterns to and from Lake Mead (Holdren et al. 2006). These factors altered the pattern of full biennial restratification, caused complete mixing in 2 subsequent years in 2000 and 2001 (Fig. 4), and prevented complete mixing for 3 other years (1994-1996).

Causes of Change in HVODThe period of relatively low HVOD rates from 1995 to 2000 occurred just after the commencement of advanced treatment of the wastewater entering the Las Vegas Wash (Fig. 8), which indicates that the supply of phosphorus to Boulder Basin probably has an effect on the HVOD rates. However, this period of low HVOD rates also coincides partially (1995-1999) with the period of high flows from the Glen Canyon Dam, which also tend to lower the HVOD rates (Fig. 10 and 11). Thus both the Glen Canyon Dam outflow and TP con-centrations in Las Vegas Bay were investigated as affecting

LaBounty and Burns

78

Table 2.-Estim

ate of downw

ard transport of dissolved oxygen (DO

) into the hypolimnion of B

oulder Basin 1995-2005 by eddy diffusion.

A

v A

v

Av

Av

Tem

p

Dow

nward

Therm

o H

ypo

Thermo

Hypo

Tt - T

h increase

Dow

nward

HV

OD

H

VO

D

DO

flux as Y

ear Type D

O

DO

O

t - Oh

Temp

Temp

Temp

rate D

O flux

Obs.

Adj.

% of H

VO

D

and Year

(g/m3)

(g/m3)

(g/m3)

(°C)

(°C)

(°C)

(°C/day)

(g/m3/day)

(g/m3/day)

(g/m3/day)

Adj.

Type 1 – Y

ears After F

ull Destratification

1997 7.39

7.03 0.36

12.091 12.012

0.079 0.0009

0.004 0.009

0.013 29

1999 7.40

7.03 0.36

12.062 11.987

0.075 0.0011

0.005 0.011

0.016 33

2001 7.13

6.81 0.32

11.876 11.827

0.049 0.0012

0.008 0.012

0.020 39

2002 7.25

6.69 0.55

11.405 11.328

0.077 0.0015

0.011 0.014

0.025 43

2004 6.76

6.09 0.67

11.463 11.372

0.091 0.0018

0.014 0.015

0.028 48

Averages

7.18 6.73

0.45 11.779

11.705 0.074

0.0013 0.008

0.012 0.020

40

Type 2 - Y

ears After P

artial Destratification

1995 5.21

4.74 0.47

12.003 11.894

0.109 0.0017

0.007 0.008

0.015 47

1996 5.62

4.87 0.75

12.653 12.545

0.108 0.0011

0.007 0.007

0.014 52

1998 7.53

7.09 0.44

12.252 12.220

0.032 0.0016

0.022 0.010

0.032 70

2000 7.01

6.28 0.73

12.390 12.341

0.049 0.0013

0.019 0.011

0.030 63

2003 5.71

4.73 0.99

12.132 12.007

0.125 0.0015

0.012 0.012

0.024 50

2005 5.31

3.95 1.37

12.202 12.079

0.123 0.0016

0.017 0.014

0.031 56

Averages

6.07 5.27

0.79 12.272

12.181 0.091

0.0014 0.014

0.010 0.024

58

Long-term Increases in Oxygen Depletion in the Bottom Waters of Boulder Basin, Lake Mead, Nevada-Arizona, USA

79

the HVOD in Boulder Basin. Both these variables have an effect on HVOD (Fig. 11). HVOD is positively correlated with TP in Las Vegas Bay (Fig. 11, bottom: r2 = 0.36; p = 0.014) and negatively correlated with Glen Canyon Outflow (Fig. 11, top: r2 = 0.40; p = 0.08). These simple regressions are barely significant, because both variables are changing simultaneously so that a multiple regression is more adequate.

Indeed, the multiple regression of HVOD on TP-City of Las Vegas and Glenn Canyon Flow is highly significant:

HVOD = 18.2 + 0.082TP − 0.023Flow (R2 = 0.68); partial-p (TP) = 0.005, partial-p (Flow) = 0.003; over-all-p <0.001

This relationship provides evidence that TP in Las Vegas Bay has a positive and Flow a negative significant effect on HVOD.

During the low-flow years average daily flow was nearly constant (Fig. 10), so that the regression of HVOD on TP concentrations in Las Vegas Bay was not influenced by any changes in the Glen Canyon outflow. This relationship is highly significant (n = 11, R2 = 0.80, p < 0.001; Fig. 12) and supports a strong positive relationship between HVOD and TP in Las Vegas Bay. Similar relationships were observed in a Snake River reservoir, where high TP concentration increased anoxia and hypoxia, but flow diminished it (Nürn-berg 2002).

Future TrendsHVOD rates increased from 1995 through 2005. If this pat-tern continued, the rates would eventually be large enough to result in the hypolimnion of Boulder Basin becoming anoxic. Fortunately, changes have occurred that have interrupted this trend (decreased phosphorus loading). Average daily flows from Lake Powell (Glen Canyon Dam) >400 m3/day in 1995 decreased the oxygen depletion rate for that year. Decrease of organic loading (= decreased phosphorus loading) and increase of inflows combine to decrease the oxygen depletion rate of Boulder Basin.

When the 11-year data set from 1995 through 2005 is consid-ered alone, the conclusion is that Boulder Basin was changing to become anoxic in the lower portion of the hypolimnion. The analysis of the years to anoxia assumes that phospho-rus loadings were increasing as total discharge of treated

Table 3.-Annual dissolved oxygen (DO) depletion rates and calculated years to anoxia for 11 years, 1995-2005, Boulder Basin, Lake Mead, Nevada-Arizona. * indicates those lake years that began fully destratified (Type 1).

Year (May to Apr Annual Rate Years to of Named Year) (mg/m3/day) Anoxia

1995 8.1 201996 6.8 241997* 9.3 251998 9.6 241999* 10.7 222000 11 272001* 11.8 212002* 14.1 192003 11.7 172004* 14.7 162005 13.7 9

Figure 11.-HVOD vs. flow and TP. Top panel: HVOD vs. average daily flow from Glen Canyon Dam, 1991-2006 (regression: y = -0.025 + 21.87, R2 = 0.405, p < 0.05). Circled years indicate average daily flows >400 m3/day). Average daily flow for the other years <335 m3/day. Bottom panel: HVOD vs. TP in Las Vegas Bay, 1991-2006 (regression: y = 0.093 + 9.5, R2 = 0.360, p < 0.01). Circled values are years when average daily flows were >400 m3/day. Average daily flow for the other years <335 m3/day.

yaB sageV saL ni PT .sv DOVH

m/gµ( PT launnA egarevA 3)

HVO

D (m

g/m

3 /day

)

wolftuO noynaC nelG .sv DOVH

m( wolF yliaD egarevA 3 )s/

HVO

D (m

g/m

3 /day

)

5

7

9

11

31

51

71

91

12

0 10 2 30 40 0 5 60 70 0 80 90

VLC PT .sv DOVHniL e H( ra )PT VLC .sv DOV

89 & ,79 ,69 ,5991wolF gareva � s ,e

ovreser i m 11 esor r

579113151719112

572 3 52 573 4 52 574 5 52 575 6 52

79 ,69 ,5991 99 & ,89 ,wolF gareva � s ,e

ovreser i m 11 esor r

uO .sv DOVH wolftniL e H( ra )wolftuO .sv DOV

Figure 12.-HVOD vs. TP. Average annual TP in Las Vegas Bay for the years when the flow from Glen Canyon Dam <335m3/day (regression: y = 8.562 + 86.3, R2 = 0.80, p < 0.001).

ylnO wolfnI woL fo sraeY—PT susrev DOVH

0

02

04

06

08

01 21 41 61 81 02

PT .sv DOVH)PT .sv DOVH( raeniLA

vera

ge A

nnua

l TP

(µg/

m3 )

m/gm( DOVH 3 )yad/

LaBounty and Burns

80

wastewater effluent increased, and that average daily inflow from Glen Canyon is not >400 m3/day. Based on calculations (Appendix A), the number of years from 2005 to when anoxic conditions might be expected to occur is predicted separately for the 2 different stratification scenarios for Type 1 and 2 years. Using the most recently collected data from Lake Years 2004-2006 (Fig. 7), calculations show that anoxia of Boulder Basin would be expected by day 265 within about 18 years for years following complete destratification (Type 1 years) and within 16 years for Type 2 years.

The large loading of organic material into Las Vegas Bay prior to 1994 (1,300 kgP/day in 1978; Dr. Doug Drury, Clark County Water Reclamation District, pers. comm.) explains the high HVOD rates in those early years. The >400 m3/day average flows from Glen Canyon Dam from 1995 through 1999, along with the incorporation of tertiary treatment process in 1994, caused rapid decline of HVOD in 1995. The increasing rate of organic contribution to the hypolimnion of Boulder Basin from 1994 to 2004 due to a greatly expanding population resulted in increasing organic and nutrient loading, which caused accelerated annual rate of hypolimnetic oxygen depletion between 1996 and 2004. From 1994 through 2005 TP increased in the hypolimnion at a rate of 1.6 mg P/m3/yr, indicating a probable increase in organic matter in the hypolimnion. In this regard, Secchi depth showed a decrease and total organic carbon showed an increase in Boulder Basin from 2000 to 2004 (LaBounty and Burns 2005). The years to anoxia (Table 3), calculated from actual start-of-season DO values versus those calculated using average start-of-season DO concentrations (Appendix A), more accurately display the trends during this time period. The data can be generally interpreted to mean that from 1995 to 2000, the time to anoxia was 20-27 years; but from 2000 to 2005, the time to anoxia diminished steadily, largely because of decreasing start-of-season DO concentrations (Table 1).

The potential occurrence of a series of ≥10 years when in-flows are below normal and organic loading is not decreased is always possible for Lake Mead based on hydrology and water use patterns in the Colorado River Basin. If drought conditions were to persist along with unchecked organic loading, conditions could lead to the onset of years when anoxia begins prior to destratification. Anoxia in a lake is deleterious for the lake because it results in a large increase in the internal loading of phosphorus, ammonia, iron, and manganese from the sediments to the overlying water (Burns and Ross 1972, Nürnberg 2004). Usually the increased inter-nal load of phosphorus enters the life cycle of the lake and can cause even larger releases of phosphorus in subsequent years. This is a particularly undesirable situation for Boulder Basin because of the strong phosphorus limitation to algal growth in the Basin. This situation may be exacerbated by the presence of quagga mussels (discovered throughout Boulder Basin in early January 2007). If the population

explodes as it has in the Great Lakes, it could cause changes in the HVOD regime of Boulder Basin. Nevertheless, if the trend to decreasing HVOD rates shown by the Group 3 years (2004-2006) continues, the progress shown by the Group 2 years (1996-2004; Fig. 8) toward the onset of anoxic condi-tions could be averted.

This study demonstrates that hypolimnetic oxygen deple-tion rates can provide a useful tool for assessing long-term changes in the metabolism of stratified lakes and reservoirs. However, observed depletion rates must be adjusted to re-flect system-specific characteristics, such as hypolimnetic temperature, the downward flux of oxygen to the hypolim-nion and the volume of inflows. Vertical mixing inputs of oxygen were found to be important in this study, probably accounting for 29-70% of the oxygen consumed annually in the hypolimnion and increasing as the DO concentrations become lower. Failure to account for this source of oxygen can result in underestimation of oxygen depletion rates and inaccurate representation of long-term trends.

SummaryThe pattern of thermal stratification of Boulder Basin is dependent on several climatic and hydrologic factors such as the annual flow volumes from the Glen Canyon Dam and the temperature of the inflow that varies from year to year. Temperatures of water entering the upper portion Lake Mead from the Colorado River are normally 9-11°C dur-ing winter months (Holdren et al. 2006), while the surface water of Boulder Basin remains ≥12°C. This temperature difference provides more than enough stability to resist vertical mixing. However, the timing and magnitude of the inflow and the average water temperature of the inflow from the Colorado River determine whether Boulder Basin will destratify completely.

Complete destratification most likely occurs when the aver-age water temperature of the lower portion of the hypolim-nion is >12.5°C and ambient conditions in January cool the epilimnion to an average of <12.5°C. Additionally, arrival of the cooler Colorado River underflow is delayed until at least mid-February. This is usually a short period of time (<1 month). Once the cooler Colorado River underflow arrives, stratification instantly begins. New water from the Colo-rado River forms a layer of colder water at the bottom and becomes the deep hypolimnion for the next 2 years. Under normal or below average inflow from the Colorado River, stratification generally sets up for 2 years in Boulder Basin with a small degree of vertical mixing at the end of the first year. The onset of stratification is so strong that 2 years of warming the bottom waters are usually required to enable complete destratification.

Long-term Increases in Oxygen Depletion in the Bottom Waters of Boulder Basin, Lake Mead, Nevada-Arizona, USA

81

Complete destratification does not occur in years when inflow volumes are greatly below normal, and/or water temperature of the Colorado River inflow during fall and winter are >12°C, and/or Colorado River interflow arrives in Boulder Basin before the average water temperature of the epilimnion >12.8°C.

This leads to 2 types of stratification regimes that affect an-nual HVOD rates; those following complete destratification and those following partial destratification. The hypolimnetic DO concentrations in the second year of a pair of years are lower that those of the first year of a pair. This difference enables vertical eddy diffusion to transport more oxygen downward in the Type 2 years, leading to lower observed HVOD rates in these years. When eddy diffusion is taken into consideration, both Type 1 and Type 2 years probably have similar HVOD rates. The downward diffusion of oxy-gen increases as the hypolimnetic oxygen concentrations decrease; however, this does not prevent the progress toward anoxic conditions, unless the loading of organic carbon to the hypolimnion decreases.

The HVOD rates for the years following full destratification (Type 1 years) averaged 12.1 mg DO/m3/day from 1995 to 2005 and increased annually at the rate of 0.91 mg DO/m3/day per year. The HVOD rates for these years following partial destratification (Type 2 years) averaged 10.1 mg DO/m3/day and increased at the rate of 0.59 mg DO m3/day per year. The increase in the HVOD rates resulted from increased organic and nutrient inputs. These results demonstrate quite clearly that higher TP concentrations in Las Vegas Bay lead to higher HVOD rates. A positive relationship between HVOD and the previous year’s phosphorus loading has also been observed in Lake Erie (Burns et al. 2005). Unless the trend of increased loading with time is stopped, Boulder Basin will become anoxic by the end of the stratified season in about 20 years from 2006. Once anoxia occurs, the release rate of dissolved phosphorus from sediments and the hypolimnetic phosphorus concentrations will increase dramatically. This would be disastrous for Boulder Basin. Fortunately, the 2005 and 2006 HVOD rates are lower, possibly in response to the enhanced phosphorus removal commenced in 2002, but this response still needs to be effectively established.

AcknowledgmentsThis project was sponsored by the Southern Nevada Water Authority (SNWA). We greatly acknowledge the support of Mr. Ron Zegers, Director of the SNWA Southern Nevada Water System (SNWS), and Ms Keiba Crear and Ms. Peggy Roefer, SNWA Watershed Division. Numerous individu-als from the SNWS Field and Laboratory units collected samples, performed analyses, and provided organized data files. We acknowledge their dedication and hard work. We acknowledge the following agencies who supplied data used

for this study: U.S. Bureau of Reclamation, Clark County Water Reclamation District (CCWRD), and the City of Las Vegas Water Pollution Control Facility. We especially thank Dr. Doug Drury, CCWRD, for retrieving some of the oldest data and providing guidance on the water treatment history. We acknowledge the anonymous reviewers of this manu-script. The authors greatly appreciate the editorial efforts of Dr. Gertrud Nürnberg, Freshwater Research, Baysville, Ontario. Dr. Nürnberg handled reviews and she provided her own professional advice. Her input was essential to this endeavor.

ReferencesBirge, E.A. and C. Juday. 1911. The inland lakes of Wisconsin: The

dissolved gases and their biological significance. Bull. Wisc. Geol. Nat Hist. Surv. Number 22. 259 p.

Beutel, M.W. 2003. Hypolimnetic anoxia and sediment oxygen demand in California drinking water reservoirs. Lake Reserv. Manage. 19:208-221.

Burns, N.M. 1995. Using hypolimnetic dissolved oxygen depletion rates for monitoring lakes. N.Z. J. Mar. Freshw. Res. 29:1-11.

Burns, N.M. and C. Ross. 1972. Oxygen-Nutrient relationships within the Central Basin of Lake Erie. In Project Hypo. Canada Centre for Inland Waters, Paper No. 6; USEPA Technical Report, TS-05-71-208-24:85-117.

Burns, N.M., D.C. Rockwell, P.E. Bertram, D.D. Dolan and J.J. Ciborowski. 2005. Trends in temperature, Secchi depth and dissolved oxygen depletion rates in the Central Basin of Lake Erie, 1983-2002. J. Gt. Lakes Res. 31:35-49.

Charlton, M.N. 1980. Hypolimnion oxygen consumption in lakes: Discussion of productivity and morphometry effects. Can. J. Fish. Aquat. Sci. 37:1531-1539.

Cornett, R.J, and F.H. Rigler. 1979. Hypolimnion oxygen deficits: their prediction and interpretation. Science 205:580-581.

Ford, D.E. 1990. Reservoir transport processes. P. 15-42. In K.W. Thornton, B.L. Kimmel and F.E. Payne (eds). Reservoir Limnology: Ecological Perspectives. John Wiley and Sons, New York. 256 p.

Heiskary, S.A. and C.B. Wilson. 2005. Minnesota Lake Water Quality Assessment Report: Developing Nutrient Criteria. Third Edition. Minnesota Pollution Control Agaency, St. Paul, Minn. 176 p.

Holdren, G.C., M.J. Horn and D.M. Lieberman. 2006. The limnology of Lake Mead, Arizona-Nevada, 2005 Report of Findings. Bureau of Reclamation 8220-06-15, Denver, Col. 68 p.

Hutchinson, G.E. 1938. On the relation between the oxygen deficit and the productivity and typology of lakes. Int. Rev. Hydrobiol. 36:336-355.

Hutchinson, G.E. 1957. A Treatise on Limnology. Volume 1: Geogr. Physics Chem.. John Wiley and Sons, Inc. N.Y. 1015 p.

Juday, C. and E.A. Birge. 1932. Dissolved oxygen and oxygen consumed in the lake waters of northeastern Wisconsin. Trans. Wisc. Acad. Sci Arts Lett. 27:415-486.

LaBounty, J.F. 2005. Limnological status of Boulder Basin, Lake Mead, Nevada-Arizona, 2000-2005. Southern Nevada Water Authority. Las Vegas, Nev. 125 p.

LaBounty and Burns

82

Appendix AThe “years to anoxia” can be estimated for each year by using the starting DO concentration for that year, its HVOD rate and the appropriate increase in HVOD rate from Fig. 9.

1) Type 1 example (i.e., following full destratification); Lake Year 2004-2005;

a) Average start of summer concentration = 8.1 mg/L (see Table 1)b) What rate is required for anoxia to set in after 265 days?c) Rate to cause anoxia = 8.1/265 = 0.031 mg/L/day = 31.0 mg/m3/dayd) 2004-2005 HVOD rate = 0.0147 mg/L/day = 14.7 mg/m3/day (see Fig 7: bottom panel)e) HVOD rate increase to reach anoxic HVOD rate = 31.0 - 14.7 = 16.3 mg/m3/dayf) HVOD increase rate = 0.9062 mg/m3/day per year (see Fig. 8: top panel)g) Number of years to anoxia = 16.3 mg/m3/day ÷ 0.9062 mg/m3/yrh) = 18 years

2) Type 2 example (i.e., following partial destratification); Lake Year 2005-2006.

a) Average start of summer concentration = 6.2 mg/L (see Table 1)b) What rate is required for anoxia to set in after 265 days?c) Rate to cause anoxia = 6.2/265 = 0.0233 mg/L/day = 23.3 mg/m3/dayd) 2005-2006 HVOD rate = 0.0138 mg/L/day = 13.8 mg/m3/day (see Fig. 7: top panel)e) HVOD rate increase to reach anoxia at day 265 = 23.3 - 13.8 = 9.5 mg/m3/dayf) HVOD increase rate = 0.5863 mg/m3/day per year (see Fig. 9: top panel)g) Number of years to anoxia = 9.5 mg/m3/day ÷ 0.5863 mg/m3/yrh) = 16.2 years

LaBounty, J.F. and M.J. Horn. 1997. The influence of drainage from the Las Vegas Valley on the limnology of Boulder Basin, Lake Mead, Arizona-Nevada. Lake Reserv. Manage. 13:95-108.

LaBounty, J.F. and N.M. Burns. 2005. Characterization of Boulder Basin, Lake Mead, Nevada-Arizona, USA – Based on analysis of 34 limnological parameters. Lake Reserv. Manage. 21:277-307.

Lara, J.M. and J.I. Sanders. 1970. The 1963-64 Lake Mead survey. Bureau of Reclamation Report. Denver, Col. 172 p.

Lasenby, D.C. 1975. Development of oxygen deficits in 14 southern Ontario lakes. Limnol. Oceanogr. 20:993-999.

Matthews, D.A. and S.W. Effler. 2006. Long-term changes in the areal hypolimnetic oxygen deficit (AHOD) of Onondaga Lake: Evidence of sediment feedback. Limnol. Oceanogr. 51:702-714.

Nürnberg, G.K. 2002. Quantification of oxygen depletion in lakes and reservoirs with the hypoxic factor. Lake Reserv. Manage. 18:298-305.

Nürnberg, G.K. 2004. Quantified hypoxia and anoxia in lakes and reservoirs. TheScientificWorld 4:42-54.

Pilson, M.E.Q. 1998. An Introduction to the Chemistry of the Sea. Prentice Hall.

Rosa, F. and N.M. Burns. 1987. Lake Erie Central Basin oxygen depletion changes from 1929-1980. J. Gt. Lakes Res. 13:684-696.

Strøm, K.M. 1931. Feforvatn. A physiograhic and biological study of a mountain lake. Arch. Hydrobiol. 22:491-536.

Thornton, K.W. 1990. Perspectives on reservoir limnology. P. 1-14. In K.W. Thornton, B.L. Kimmel and F.E. Payne (eds.). Reservoir Limnology: Ecological Perspectives. John Wiley and Sons, N.Y.

Thienemann, A. 1928. Der Sauerstoff im eutrophen und oligottophen Seen. Die Binnengewässer, 4. Stuttgart, Schweizerbartsche Verlagsbuchhandlung. 175 p.

Vincent, W.F., M.M. Gibbs and S.J. Dryden. 1984. Accelerated eutrophication in a New Zealand lake: Lake Rotoiti, central North Island. N.Z. J. Mar. Freshw. Res. 18:431-440.

Walker, W.W. 1979. Use of hypolimnetic oxygen depletion rate as a trophic state index for lakes. Water Resour. Res. 15:1463-1470.

Wetzel, R.G. 2001. Limnology, 3rd Edition, Academic Press. San Diego, Calif. 1006 p.