Limnol. Oceanogr., 42(2), 1997, 230-238 0 1997, by the American Society of Limnology and Oceanography, Inc.

Stable isotope studies on Mono Lake (California). 1. al80 in lake sediments as proxy for climatic change during the last 150 years

Hong-Chun Li, Teh-Lung Ku, and Lowell D. Stott Department of Earth Sciences, University of Southern California, Los Angeles, California 90089-0740

Robert F. Anderson Lamont-Doherty Earth Observatory, Columbia University, Palisades, New York 10964

Abstract Mono Lake is a hypersaline, alkaline lake in the Mono Basin located #at the eastern base of the Californian Sierra

Nevada. Its lake-level history since 1912 has been recorded instrumentally, showing the decline of lake-surface elevations initiated by the 1941 artificial diversion of stream inflow. We have made high-resolution oxygen isotopic measurements on the total carbonate fraction of lake sediments and shown that the S’*O record parallels the lake- level fluctuations rather well. The measurements were carried out on sediments that had been leached with deionized water to isolate the isotopic signals of the calcium carbonate from those of pore water and water-soluble carbonate salts in the sediment. Extending the S’“O record back in time, we found that lake level changed markedly during the past 150 yr, reflecting climatic variations and resultant runoff fluctuiitions. Lake levels were high around 1845, 1880, and 19 15 and low around 1860, 1900, and 1933. This study demonstrates that closed-basin lake sediment S180 provides an effective means of probing past precipitation variations in arid to semiarid regions such as the Great Basin in the western United States.

Numerous attempts have been made to reconstruct paleo- lake levels on the basis of geological and stable isotopic evidence from which paleohydrological conditions have been inferred. Studies on closed-basin lakes in semiarid and tropical zones have contributed greatly to paleoclimatic re- search, particularly for the Late Pleistocene and the Holo- cene (e.g. Street-Perrott and Harrison 1985; Chivas et al. 1986; Gasse et al. 1991). Water entering closed-basin lakes via rain and runoff leaves by evaporation. Hence, fluctua- tions in lake level and volume should serve to indicate changes in the moisture budget associated with climatic change. Mono Lake, located in eastern California about 480 km north of Los Angeles (Fig. l), is the only terminal lake in the Mono Basin. Its present ratio of surface area (150 km2) to catchment area (2,072 km2) is -7 : 100, in contrast to the ratio of -33 : 100 for Lake Russell during the last glacial period (Russell 1889; Lajoie 1968; Mono Lake Com- mittee 1981). Since 1912, the levels of Mono Lake have been measured by the U.S. Geological Survey (USGS) and the Los Angeles Department of Water and Power (LADWP) (Blevins et al. 1987). Since its historic high in 1915, the water level within the lake has fallen as a result of climatic changes. The lake level fall has been exacerbated by the diversion of inflow to the lake by LADWP ‘since 1941 and has caused a further increase in salinity and alkalinity. Dur-

Acknowledgments We thank D. E. Hammond and S. P. Lund for insightful discus-

sions. D. R. Cayan kindly made available to us his unpublished data on Nevada City precipitation and streamflow. Comments made by L. V. Benson and an anonymous reviewer greatly helped our pre- sentation.

This work was supported by NSF grants ATM 93-03587 (T.-LX.) and OCE 93-14192 (L.D.S.), and by NOAA grant NA16-RC-0084 (R.EA.).

ing the diversion period the lake level fell - 13 m (from an elevation of 1,956 m to its historic low stand of 1,942 m) between 194 1 and 1982. During the high runoff years be- tween 1982 and 1986 (a wet period caused by strong 1982- 1983 El Nifio), the lake level rose by -2.4 m to 1,944.6 m before falling again.

The fluctuations in lake level and volume of a closed- basin lake such as Mono Lake may be recorded in the 6’*0 of authigenic carbonates deposited within the lake sedi- ments. This is feasible because the lake S180 monitors the hydrologic balance between the isotopically distinct inflow and outflow. For example, Benson (1994) monitored the 6’*0 of lake waiter from 1985 to 1992 in Pyramid Lake, Nevada, another closed-basin lake in the Great Basin. His results showed that although the surface-water &*O exhibits sea- sonal variations, reflecting varying amounts of fresh water discharged to and lost from the lake, the yearly S180 trend over the study period correlates well with that of the lake- volume change (lake-water oxygen isotopic values become lighter as volume increases). This observation suggests that when the lake undergoes a volume change, the hydrological balance is the most important factor influencing the lake 6’*0, although other factors such as stratification, relative humidity, ,air temperature, and the atmospheric @*O may also play a role. If carbonate minerals form in a lake in isotopic equilibrium with lake water, their 6’*0 values can be used toI reconstruct lake volume fluctuations. This ap- proach to lake history investigations forms the basis of sev- eral previous studies (Lister et al. 1991; Phillips et al. 1992; Johnson et al. 1991). However, it would be reassuring if a comparison could be made between the reconstructed lake- level changes with instrumental records. Such a comparison can be conducted for Mono Lake because, as mentioned above, its lake-level record is available for the past 80 yr through the measurements of LADWP and USGS (Blevins

230

Stable isotopes in Mono Luke 231

Fig. 1. Map showing Mono Basin and sample locations (adapt- ed from the Mono Lake Committee 198 1).

et al. 1987). In this paper, we describe our efforts in the use of the sediment S’*O record to reconstruct lake-level fluc- tuation and climatic change in the Mono Basin during the last 150 yr.

Mono Lake hydrology

The moisture input to the hydrologically closed Mono Ba- sin is principally controlled by the amount of precipitation it receives, which, in a larger scheme, is tied to the mete- orology of the Great Basin (Houghton 1969; National Re- search Council 1987). Precipitation in the Great Basin is dominated by Pacific winter storms. Winter snows and rain- falls feed the stream runoff from the Sierra Nevada, which constitutes -70% of the water input to Mono Lake. The remaining 30% is roughly divided between direct precipi- tation onto the lake surface and groundwater seepage (Blev- ins et al. 1987). As a result, surface runoff and rainfall in the Mono Basin assume a linear relationship (Fig. 2a). Be- cause output of the lake water is by evaporation only and the annual rate of evaporation is relatively constant (Blevins et al. 1987), lake volume also varies linearly with runoff to the lake (Fig. 2b). Figure 2 shows that closed-basin lake- volume fluctuations could effectively reflect changes in bas- inwide precipitation, because evaporation, being highly tem- perature dependent, remains rather constant on annual-to- decadal time scales. The relationships shown in Fig. 2 pro- vide a framework to probe paleoprecipitation and paleohydrology in a closed basin. Namely, if a means can be found to reconstruct the paleolake volume (or level), then past changes in the surface runoff and precipitation may be inferred from the reconstruction. According to the measure- ments of volume and elevation of Mono Lake (Blevins et al. 1987), there is a strong linear correlation between lake volume and lake level: lake level (m) = 1,929.9 + 4.8515

lake volume (km3) (R2 = 0.99). Therefore, volume change can be directly read from lake-level change. We describe

‘lake-level fluctuations reconstructed from S180 records of carbonates deposited within the lake. From these records we infer lake-volume fluctuations.

Methods

Sample collection and core chronologies-Core ML9 1 FC3 was collected in 1991 from the deepest part of Mono Lake at 39 m (Fig. 1) with a rectangular freeze corer similar to the type described by Crusius and Anderson (1991). Sed- iments were kept frozen before returning to the laboratory for processing. The core is 76 cm long and exhibits fine laminations throughout (Fig. 3). Its chronology can be de- termined from the distributions of 210Pb and 239+240Pu in an adjacent freeze core (ML91 FC6) taken on the same day and at the same location. The laminar features of the two cores are correlative and indicate similar sedimentation histories. The average sedimentation rate of the cores as determined from 210Pb and 239+240Pu is 0.65 cm yr-’ (Fig. 4a,b). There is a change in the sedimentation rate beginning at -35 cm depth in the core (Fig. 4a), from -0.75 cm yr-1 above to 0.4 cm yr-1 below. The 35 cm horizon corresponds to the time when substantial changes in runoff discharge to the lake occurred due to the stream diversion by LADWF? The 210Pb profile does not allow annual resolution nor does it give overall chronology to better than + 10%. X-ray diffraction (XRD) and scanning electron microscope (SEM) analyses show that major minerals of the sediment consist of calcite, halite, and quartz. Downcore XRD analyses (at l-cm inter- vals) showed that the carbonate phase in the sediments of the core is mostly calcite, except near 39 cm where aragonite is the dominant phase. As Li (1995) has shown, calcite and aragonite are the primary mineral precipitates in Mono Lake sediments, with aragonite mostly occurring in surface sedi- ments from water depths shallower than 16 m in the lake. Because FC3 core was taken from the deepest part of Mono Lake at 39 m, aragonite content in the sediments of this core should be very low.

Stable isotope measurements-Core ML91 FC3 was sec- tioned at 0.4-cm intervals under frozen conditions and - 100-400 pg of the sediment from each interval was sam- pled for 6180 analyses. The material, after oven-drying over- night at 7O”C, was converted to CO, in 100% phosphoric acid at 9O”C, using an “autocarb” device attached to a VG Prism II triple-collecting mass spectrometer. Replicate anal- yses of a calcite standards run together with the samples gave reproducibilities of t0.26%0 for 6’*0. The method of S1*O analysis for water samples is the CO,-H,O equilibrium method (Epstein et al. 1953). The isotope ratios are reported as the per mil deviation from the PDB standard for carbonate material and from the SMOW standard for water samples. The measurements were first made on CO, released from acidification of 182 bulk sediment samples. Although the 6180 evolution trend of these unwashed, bulk samples cor- relates well with the historic lake volume changes, the al80 values cannot be explained by calcite precipitation in iso- topic equilibrium with the lake water. Therefore, we then

Li et al.

0 20 40 60 80 100 120

Rainfall in Mono Basin (cm yi’)

. = -0.1345 + 1.0811x R2 = 0.83 Y

-,.,2J-J-LJ,,

0 0.1 0.2 0.3

Runoff to Mono lake (km3 yr-‘)

Fig. 2. [a.] Correlation between rainfall and stream runoff in the Mono Basin from 1935 to 1985, at the Gem Lake Station lo- cated 20 km south of Mono Lake in the Sierra Nevada. [b.] Cor- relation between stream runoff to Mono Lake and Mono Lake vol- ume change from 1935 to 1985, based on measurements made by LADWP; the volume change refers to the volume difference be- tween two consecutive years. Data source: Blevins et al. 1987.

made 61x0 analyses on 97 sediment samples (selected from the 182 samples) after they had been leached with deionized water buffered at a pH of 7 in order to dissolve the interstitial salts. The leaching was done as follows. About 25 ml of water was mixed with a known quantity (-0.5 g) of the sample. The mixture was shaken, centrifuged, and the su- pernatant discarded. The leaching procedure was repeated three times until pH of the leachate was close to 7. The

washed salmples were oven-dried overnight at 70°C and weighed again to determine the salt content. Table 1 shows the S180 vailues of the washed samples.

Relative to the washed samples (8180 from +0.3 to -6.0%0), the unwashed bulk samples show systematically higher 618Cl values (+ 13.0 to -4.5%0), the differences being well beyond the analytical errors. Heavy S180 values of the unwashed samples may be caused by Na,CO, that formed as the samples were dried. The aI80 of the rapidly forming Na,CO, is expected to be close to that of dissolved bicar- bonate ions whose 6180 values are -8%0 heavier relative to the equilibrated water (Craig 1965; O’Neil et al. 1969). Li (1995) has noted that the 6180 offset between the washed and unwashed samples becomes larger when the Na,CO, concentration increases and when the CaCO, concentration decreases in the bulk sediments. The NqCO, concentration reflects roughly the carbonate alkalinity of porewater and the CaCO, concentration is related to the input of Ca2+; both are functions of lake volume. This explains the anticorrelation between cV0 and lake-volume change, even though the S180 of unwashed samples does not reflect that of lake water.

Results and discussion

Carbonate precipitation in sur$ace water-Authigenic carbonates in lake sediments can be inorganic or biogenic. Although direct experimental data are lacking, available em- pirical evidence suggests that oxygen isotopic equilibrium between lake water and lacustrine carbonates exists (e.g. Gasse et al. 1987; Turner et al. 1983; Johnson et al. 1991). The carbonate aI80 is a function of both temperature and lake-water S180. The 6180 of the washed sample (Scalcltc) in the topmost layer (representing the deposit of summer 1991) of core ML91 FC3 is - 1.05%0. XRD analyses showed that the carbonate phase in this sample is calcite.

We have measured the S180 of Mono Lake surface water (a,,,,,) collected in summer 1991 to be -O.l%, (SMOW) (Table 2). Assuming water-carbonate isotopic equilibrium, we calculate a temperature of 22°C for the lake carbonate formation. ‘The paleotemperature equation used, ~=calcite - &,,,,,, = 3.95 - 0.232T (“C), is an average of the rather tight re- sults of previous work on the calcite-water system (Epstein et al. 1953; O’Neil et al. 1969; Grossman and Ku 1986). The calculated temperature indicates that the calcite precipitated from Mono Lake surface waters during warm seasons be- cause only during June-October do surface waters reach temperatures of 22°C (R. Jellison unpubl. data 1991). This observation is consistent with studies carried out on Walker Lake (Newton and Grossman 1988), Lake Ontario (Schelske and Hodell 1991), and several other lakes (Kelts and Hsii 1978; McKenzie 1985) all showing a summer precipitation record. In #spring, streams supply Ca2+ to the lake. Precipi- tation of CaCO, may occur in the summer and autumn as a result of increase of Ca2+ supply, loss of CO, from the epi- limnion by photosynthesis and/or gas exchange with the at- mosphere, increase of water temperature, and increase of nu- cleation surface due to enhanced biomass.

PO in authigenic carbonates as proxy for lake-level jluc- tuation-For relatively small lakes such as Mono Lake, the

,

Fig 3. Photograph of the freeze core ML91 FC3 showing the laminated nature of rhc bedmxma. Top of the core is at upper left. Scales shown on the sediments are in centime!ers.

lake-water 61a0 depends largely on the difference in volume and rS”O between the input and output waters, the summer temperature remaining relatively constant. The lake 61x0 of a closed-basin lake progressively gets heavier as evaporation proceeds until an isotopic steady state is achieved. Thus 6180 of input waters (surface runoff + direct precipitation) is al- ways lighter than that of the lake water. For example, the mean 6’80 of input waters to Pyramid Lake is about - 1 0%0 @MOW) as opposed to -OS%0 for the lake water (Benson 1994). In Walker Lake, Nevada, the input water has an av- erage 6’“O of ~ 14.1%0, whereas the lake SIX0 is 2X%:, (New- ton and Grossman 1988). The larger the amount of fresh- water inflow relative to evaporation, the lighter the lake SIXO

will become. We have measured the 6’“O of mput warerb (streams, snow melt, and springs) to Mono Lake and ob- tained an average of 14.6 -+ 1.0%~ (Table 2). The following estimation shows the response of Mono Lake water S’80 to changes in the lake level. The 1982-1983 El Niiio event caused Mono Lake to rise 2 m, increasing the lake storage (volume) by -12% (Blevins et al. 1987). It is assumed that prior to the event Mono Lake water had a ,YxO value of -O.l%o. Mixing this water with the input water that had a PO value of 14.6%0 would lower the lake 6180 by about

1.7%-a signal that is readily detectable. Our estimation ignores the vapor exchange across the

lake-atmosphere interface, which can play a very important

234 Li et al.

Excess 2’aPb (dpm g-l) 1

10 -t

20 -t

30

40

50 E

(a)

60 -+- f

700

239+24(& (dpm g“)

30F-,i-- +++I- Fig. 4. [a.] Excess 210Pb distribution in core ML91 FC6, col-

lected on the same day from the site where core ML91 FC3 was retrieved. Average sedimentation rate estimated by fitting data from the top 35 cm is 0.75 cm yr-I, and that by fitting data from 35 to 56 cm is 0.4 cm yr-I. [b.] 279+240Pu profile on ML91 FC6. The peak at 19.5 cm is assumed to correspond to year 1963, yielding an average sedimentation rate of 0.7 cm yr-I for the top 20 cm of the

role in affecting lake al80 and forcing it to reach a steady- state value, as 618O of the atmospheric vapor has a relatively constant value. For a stabilized lake level or when lake vol- ume changes slowly, lake S180 is mainly controlled by vapor exchange and there will be no correlation between #*O and lake volume. Only when the lake level experiences relatively rapid fluctuations does the vapor exchange effect diminish, so that S180 and lake volume covary (Lister et al. 1991; Cat 1995; Li 1995). Mono Lake experienced large, rapid changes in volume during the period under study allowing us to use the S180 record to reconstruct lake volume changes.

In Fig. 5 we compare the sediment al80 record from Mono Lake with the historic lake-level record. The historic Mono

Lake levels from 19 12 to 199 1 were recorded by LADWF and USGS (Blevins et al. 1987; USGS Water Resour. Data- Calif. V. l-5, 1986-1991). In Fig. 5, the relative scales for the two ordinates are set arbitrarily to facilitate the compar- ison that shows a general anticorrelation between lake sed- iment 6180 and lake level. However, we emphasize that one cannot read the lake volume directly from the aI80 record, which has been pointed out by Benson (1994) in his Pyramid Lake study. Hydrological balance in a closed basin may per- sist for several years leading to a near-stable lake level. As Lister et al. (1991) and Benson et al. (1996) have stated, the S180 of a closed-basin lake depends on the rate of change in lake volume. When the rate of change in lake volume is small and the lake level is stabilized, the S180 of lake water will approach a steady-state value, the magnitude of which is controlled by vapor exchange across the lake-atmosphere interface and the S180 of input water but is independent of lake volume. Furthermore, the rate of change in lake volume is also a function of initial lake volume. For a given input flux, the smaller the initial lake volume, the larger the rate of lake-volume change will be. Therefore, a return to a for- mer lake level due to changes in climatic conditions does not mean the al80 should return to the former value asso- ciated with. the former lake level (Benson 1994).

Quantitative reconstruction of lake volume (level) must be obtained from a dynamic isotope mass-balance model. Nevertheless, we use al80 change to indicate lake-level fluc- tuation, that is, a sharp (say a change of >1%0 over 10 yr) decrease in. aI80 reflects a rapid rise in lake level, and a sharp increase in S18O represents a strong decline of lake level. A period (> 3 0 yr) of relatively constant S180 may result from achievement of a hydrologic and isotopic steady state within the closed-basin lake. In addition, aI80 values of individual data points may be affected by seasonal lake stratification and inhomogeneity of water temperature and S180. As Ben- son (1994) pointed out in his Pyramid Lake study, freshwater input during high-streamflow years (e.g. 1986) lowered the S180 in surface water (epilimnion) by 1.0%0 relative to that in deep water (hypolimnion). In dry years (e.g. 1991), how- ever, al80 of the epilimnion was enriched by 0.5%0 due to surface evaporation. Hence, it is better to use a 5-yr running average aI80 curve to infer lake-level fluctuations.

Large offsets between the two curves exist around 1933 and the interval 1972-1980-two time periods each near the end of a prolonged drop in lake level. In Fig. 5, a marked decrease in S180 from 1900 to 1910 implies a strong trans- gression in lake level that was previously demonstrated by Harding (unpubl. rep. 1962). The al80 values reached a min- imum value at the historic high stand near 19 15. A strong decline of lake level between 1920 and 1933 is reflected in a sharp increase in the S180 values. In addition to evapora- tion, those heavy S’*O values may also have been caused by intensified stratification of aI80 in the lake water. The lake- level drops reflected in the aI80 values prior to 1933 (Fig. 5) signify a prolonged period of dry climate in the Mono Basin, which could have led to the higher epilimnion 6180 values that are recorded by the carbonates. Lake level in- creased from 1933 to 1935 and then remained relatively sta- ble, corresponding to a wet climatic regime from around 1935 to 1943 (see Fig. 7). The 1.8%0 decrease in Si80 be-

Stable isotopes in Mono Lake 235

Table 1. The measured SIR0 data of washed samples in core ML91 FC3.

Sample Depth (mm) Year

6’80 (%o PDB) Sample

Depth (mm> Year

6’80 (%o PDB)

FC3-4 10.0 1989.4 -2.24 FC3-5 12.0 1989.1 -1.71 FC3-6 14.5 1988.6 - 1.45 FC3-7 18.0 1987.9 -1.59 FC3-8 23.5 1987.1 -2.29 FC3-9 29.5 1986.4 -1.96 FC3-10 34.5 1985.7 -1.63 FC3-11 39.5 1985.1 -1.70 FC3-12 44.5 1984.4 -2.76 FC3-13 49.5 1983.7 -1.54 FC3-14 54.5 1983.1 -2.07 FC3-15 59.5 1982.4 -1.81 FC3-16 64.5 1981.7 -1.05 FC3-17 69.5 1981.1 -1.72 FC3-23 99.0 1976.8 -2.03 FC3-24 106.5 1975.8 -1.58 FC3-25 114.0 1974.3 -0.37 FC3-27 125.0 1973.8 -0.85 FC3-28 129.0 1973.3 0.30 FC3-29 133.0 1970.9 -0.50 FC3-34 151.0 1970.5 -3.00 FC3-35 154.0 1969.8 -2.50 FC3-36 159.0 1969.3 -1.78 FC3-37 163.0 1968.7 -1.57 FC3-38 167.0 1968.2 -1.26 FC3-39 171.0 1967.7 - 1.46 FC3-42 183.0 1966.1 -0.50 FC3-44 191.0 1965.0 - 1.20 FC3-46 199.0 1963.9 -0.58 FC3-47 203.0 1963.4 -0.14 FC3-48 207.0 1962.9 -0.85 FC3-49 211 .o 1962.3 -0.58 FC3-53 227.0 1960.2 -1.50 FC3-57 243.0 1958.1 -1.39 FC3-58 247.0 1957.5 -2.13 FC3-59 251.0 1957.0 -0.85 FC3-61 259.0 1955.9 -2.00 FC3-63 267.0 1954.9 -2.80 FC3-66 279.0 1953.3 -3.15 FC3-67 283.0 1952.7 -3.37 FC3-68 287.0 1952.2 - 1.70 FC3-69 291.0 1951.7 -2.65 FC3-73 307.0 1949.5 -2.94 FC3-74 311.0 1949.0 -2.25 FC3-75 315.0 1948.5 -2.01 FC3-76 319.0 1947.9 -2.10 FC3-78 327.0 1946.9 -3.30 FC3-80 335.0 1945.3 -3.50

FC3-82 343.0 1943.3 -4.50 FC3-86 359.0 1939.3 -2.01 FC3-87 363.0 1938.3 -2.58 FC3-88 367.0 1936.6 -2.59 FC3-89 374.0 1935.0 -1.37 FC3-91 388.0 1931.3 -1.59 FC3-96 411.0 1926.0 -4.28 FC3-97 418.0 1924.0 -3.46 FC3-100 433.0 1920.7 -4.50 FC3- 105 460.0 1914.0 -5.99 FC3-106 470.0 1912.0 -5.89 FC3-108 481.0 1908.8 -4.51 FC3-109 485.0 1907.8 -4.59 FC3-111 493.0 1905.8 -4.30 FC3-113 501.0 1903.8 -3.80 FC3-115 509.0 1901.8 -3.50 FC3-116 513.0 1900.8 -2.92 FC3-117 517.0 1899.8 -2.82 FC3-118 521.0 1898.8 -2.44 FC3-119 525.0 1897.8 -2.75 FC3-120 529.0 1896.8 -2.87 FC3-124 542.5 1893.7 -3.70 FC3-126 548.5 1892.2 -2.87 FC3-128 554.5 1890.7 -3.65 FC3- 129 557.5 1890.0 -3.96 FC3-134 573.0 1886.0 -3.28 FC3-137 582.5 1883.7 -3.50 FC3-138 585.5 1883.0 -4.34 FC3-141 595.0 1880.5 -5.50 FC3-145 610.0 1876.6 -5.16 FC3-146 614.0 1875.6 -3.59 FC3-147 618.0 1874.6 -3.21 FC3-151 634.0 1870.6 -3.95 FC3-152 638.0 1869.6 -3.43 FC3-155 650.0 1866.6 -2.96 FC3-157 658.0 1864.6 -2.67 FC3- 159 666.0 1862.6 -2.50 FC3-161 674.0 1860.6 -3.00 FC3-163 682.0 1858.6 -3.50 FC3-165 690.0 1856.6 -4.54 FC3-166 694.0 1855.6 -3.54 FC3-167 698.0 1854.6 -3.17 FC3-168 702.0 1853.6 -3.82 FC3-171 714.0 1850.6 -4.27 FC3-173 722.0 1848.6 -4.50 FC3-176 734.0 1845.6 -5.70 FC3-177 738.0 1844.6 -3.72 FC3-181 754.0 1840.6 -4.37 FC3-182 758.0 1839.6 -4.36

tween 1933 and 1945 may be caused by a combination of three processes: increased freshwater discharge between 1933 and 1935; vertical mixing of isotopically light hypo- limnion water with heavy epilimnion water inherited from a previous period of lake stratification; and approaching a lighter steady-state S80 value after 1935 when the lake level was stabilized. Beginning in 1941 when the stream diversion started in the Mono Basin, the Mono Lake had experienced an almost continuous decrease in water level until 1982. The

6180 value continuously increased from 1945 to 1963 due to evaporation prevailing over input. Decreases in PO at the intervals of 1965-1969 and 1981-1986 are attributed to in- creases in lake level during these two periods. An apparent offset between the two curves exists around the interval 1972-1980, i.e. lake level decreases while the PO also de- creases. The cause for the offset in the interval 1972-1980 remains unclear.

The Fig. 5 comparison indicates that the WO of authi-

236 Li et al.

Table 2. S’*O measurements in waters from the Mono Lake Basin.

Sample Distance* Elevation S’“0 No. Location Sample type M-0 W-0 (%o, SMOW)

M-l M-2 M-3 M-4 M-5 M-6 M-7 M-8 M-9 M-10 M-11 M-12 M-13 M-14 M-15 M-16 M-17 M-18 M-19B

Ellery Lake Saddle Bag Lake Saddle Bag Lake Lee Vining Creek Hwy 395 Lundy Lake Mill Creek Wilson Creek Hwy 167 Wilson Creek lower level Grant Lake Rush Creek Hwy 395 Walker Creek Hwy 395 Rush reek lower level Mono County Park Mono County Park Mono County Park Old Marina Old Marina Navy Beach Navy Beach

Lake water Lake water Snow Creek water Lake water Creek water Creek water Creek water Lake water Creek water Creek water Creek water Spring water Spring water Spring water Spring water Spring water Warm spring water Mono lake water

16 2,907 - 14.4 22 3,048 - 14.3 22 3,048 -12.1

3.8 2,067 - 14.3 13 2,714 - 14.7 10 2,256 -15.8 5.2 2,048 -14.6 3.5 2,013 -15.8

12.5 2,173 -13.7 8.5 2,09 1 -13.8 7.3 2,073 -16.0 2.2 1,964 -15.3 0.02 1,946 -16.2 0.02 1,946 -14.4 0.2 1,946 -17.5 0.03 1,946 -14.5 0.03 1,946 -14.0 0.05 1,946 -14.1 0 1,945 -0.1

* From the Mono Lake shore in 1991, at 1,944.6 m elevation.

genie carbonates from a closed-basin lake can be used as a proxy for lake-level fluctuations in the past. The S80 record suggests that for the past 150 yr or so, Mono Lake has had three major cycles of lake-level fluctuations before the trib- utary diversion begun in 1941, with low stands occurring around 1860, 1900, and 1933 and high stands around 1845, 1880, and 19 15. After 1940, pulses of high input of fresh- water occurred around 1943, 1953, 1969, and 1983, as reg- istered by the light S80 peaks superimposed on the contin- uous decline of lake level due to the stream diversion.

----OS- measured 6”O - 5-yr running ave. - lake level

0

, I /I I ,_I

1840 1860 1880 1900 1920 1940 1960 1980 2000

Year (A.D.)

Fig. 5. Comparison of measured SIX0 and their 5-yr running averages in washed samples of core ML91 FC3 with the historically measured Mono Lake level. Scales for the two ordinates are set arbitrarily to facilitate the comparison. S180 decrease corresponds to lake level increase. Solid arrows indicate periods of high runoff and high lake stands. Open arrows indicate periods of low runoff and low lake stands. Note that the lake level continuously declined between 1941 and 1981 largely due to the stream diversion, with secondary peaks denoting high runoff periods during this interval.

Paleo-lake level, paleoprecipitation, and streamjlow his- tory---We think the good correlation between S180 changes and historic lake-level variations strongly support the notion that sediment proxies can be used to extend the lake-level histories back in time. However, we would like to know whether lake-level variations on time scales of 10°.5-102 yr reflect primarily climate signals of precipitation (P) or of evaporation (E). Variations in the size of a closed-basin lake are mainly controlled by changes in the effective wetness (P - E). Although climatic variables such as precipitation, tem- perature, humidity, cloudiness, and wind speed all affect the moisture budget, it can be argued that precipitation is the dominant factor. This argument gains support from the monthly records of precipitation, runoff, and evaporation shown in Fig. 6 for a closed-basin lake in Oregon (Hostetler 1995). The records show that on an annual basis, river runoff correlates well with precipitation (with slight delays in the “peaking”’ of runoff due partly to melting of winter snows) but not with evaporation, which remains rather constant. As evaporation is highly temperature dependent, one may sur- mise that relatively constant average annual temperatures impart little to runoff variations. Another reason for rela- tively constant evaporation in the Great Basin is probably due to the cloudless skies that prevail in warm seasons when evaporation occurs (Benson 1981). Hence, we submit that for the Mono Basin changes in the effective wetness mainly reflect changes in precipitation, at least on interannual-to- decadal time scales. In the Mono Basin, relationships be- tween lake volume, precipitation, and surface runoff, as il- lustrated in Fig. 2, also delineate this situation. These rela- tionships are generally characteristic of arid and semiarid regions where groundwater discharge is small compared to streamflow. In summary, when precipitation rate increases in a closed basin, runoff and lake storage in the basin will increase and the lake S1*O will be depleted.

Stable isotopes in Moio Lake 237

6 Silvies

4 c

6 Evaporation

4 c

1980 1981 1982 1983 1984 1985 1986 1987 YEAR

Fig. 6. Monthly records of precipitation, runoff, and evapora- tion in Harney-Malheur Lake, Oregon, 1980-1986. Vertical dashed lines are for demarcations of water years. The surface elevation graph shows both the simulated (solid line) and measured (triangles) lake level for the same period (after Hostetler 1995).

In Fig. 7 we plotted the records of Mono Lake sediment S’*O, the precipitation anomalies of Los Angeles and the Mono Basin, and the precipitation and streamflow anomalies of Nevada City (located at the western slope of the Sierra Nevada, -200 km northwest of Mono Lake; see Fig. 1). As we mentioned earlier, precipitation in the Great Basin is dominated by Pacific winter storms. Li (1995) has shown that the rainfall pattern in the Mono Basin is similar to those of the longer records kept for Nevada City and Los Angeles. The mean precipitations at these locations differ: 37.6 cm yr- 1 in Los Angeles (1878-1991), 55 cm yr-1 in the Mono Basin (1926-1985), and 135.7 cm yr-I in Nevada City (1864-1994). However, the precipitation anomalies, i.e. de- viations from the mean values, show a similar pattern that can be matched by the 6180 variability in Mono Lake sedi- ments (Fig. 7). The sediment aI80 also records the stream diversion effect, hence the post-1941 values are distinctly heavier. There appears to exist a phase shift of a few years, with the precipitation curves leading the S’*O curve, due probably to the water residence time of about 25 yr in the

600

400

b 1 c 200

Ed

;z "zi

0

-200 .o z gy -400 :$ g

JP -600

-800

i -10 -8 -6 -4 -2

s

04 0 6 F F i 3' 8

&

1840 1860 1880 1900 1920 1940 1960 1980 2000 Year (A.D.)

Fig. 7. [a.] Comparison of SIX0 variations in Mono Lake sedi- ments (washed) with precipitation and streamflow anomalies at Ne- vada City, California. [b.] Comparison of SIX0 variations in Mono Lake sediments (washed) with precipitation anomalies in Los An- geles and Mono Basin. Data sources: Nevada City precipitation and streamflow anomalies from D. R. Cayan pers. comm.; Mono Basin precipitation anomalies from Blevins et al. 1987; Los Angeles pre- cipitation anomalies from data on file at the National Climatic Data Center, Asheville, North Carolina and compiled by the National Oceanic and Atmospheric Administration.

lake (Blevins et al. 1987). The matching of the cycles tends to break down prior to 1900. Whether this breakdown, as in the case of the phase lag, is a consequence of age uncer- tainties associated with the sediment record remains to be ascertained.

Conclusion

This study shows that the oxygen isotopic composition of carbonate sediments in closed-basin saline lakes provides a semiquantitative means of probing past precipitation and sur- face runoff variations in arid to semiarid regions such as the Great Basin. Most of the authigenic carbonates deposited in Mono Lake are found to form in surface water during warm seasons of the year. In view of the small changes in inter- annual summer temperatures in the region, the effect of tem- perature on the carbonate 6180 is relatively constant. The hydrological history of a hypersaline closed-basin lake (e.g. Mono Lake) can be well deciphered by the #*O record ob- tained from the lake’s inorganically precipitated calcium car- bonates. On annual-to-decadal time scales, the effective wet- ness (P - E) of a closed-basin lake is mainly reflected by

238 Li et al.

changes in precipitation. Therefore, a detailed study on sed- imentary S180 records of closed-basin lakes offers a new approach to acquiring high-resolution time series records of precipitation variability. The feasibility of the approach has been demonstrated in the S180 record of Mono Lake sedi- ment, showing three major cycles of climatically induced lake-level fluctuations during the past 150 yr.

References

HOUGHTON, J. G. 1969. Characteristics of rainfall in the Great Ba- sin. Ph.D. thesis, Univ. Oregon. 292 p.

JOHNSON, T. C., J. D. HALFMAN, AND W. J. SHOWERS. 199 1. Pa- leoclimate of the past 4000 years as Lake Turkana, Kenya, based on the isotopic composition of authigenic calcite. Pa- laeogealgr. Palaeoclimatol. Palaeoecol. 85: 189-l 98.

KELTS, K., ,%ND K. J. Hsii. 1978. Freshwater carbonate sedimen- tation, p. 295-323. In A. Lerman [ed.], Lakes: Chemistry, ge- ology, physics. Springer.

BENSON, L. V. 1981. Paleoclimatic significance of lake-level fluc- tuations in the Lahontan Basin. Quat. Res. 16: 390-403.

-. 1994. Behavior of the stable isotopes of oxygen and hy- drogen in the Tiuckee River-Pyramid Lake surface-water sys- tem. 1. Data analysis and extraction of paleoclimatic infor- mation. Limnol. Oceanogr. 39: 344-355.

- L. D. WHITE, AND R. RYE. 1996. Carbonate deposition, Pyiamid Lake Subbasin, Nevada: 4. Comparison of the stable isotope values of carbonate deposits (tufas) and the Lahontan lake-level record. Palaeogeogr. Palaeoclimatol. Palaeoecol. 122: 45-76.

LAJOIE, K. R. 1968. Lake Quaternary stratigraphy and geologic history of Mono Basin, eastern California. Ph.D. thesis, Univ. California, Berkeley. 379 p.

BLEVINS, M. L., E. L. COUFAL, L. M. WATANABE, M. R. GARCIA, AND C. J. FOLEY. 1987. Mono Basin geology and hydrology. LADWP, Aqueduct Div.-Hydrol. Sect. 108 p.

CHIVAS, A. R., I? DE DECKKER, M. NIND, D. THIRIET, AND G. WAT- SON. 1986. The Pleistocene paleoenvironmental record of lake Buchanan: An atypical Australian playa. Palaeogeogr. Palaeo- climatol. Palaeoecol. 54: 21-43.

LI, H.-C. 1995. Isotope geochemistry of Mono Basin, California: Applications to paleoclimate and paleohydrology. Ph.D. thesis, Univ. Southern California, Los Angeles. 244 p.

LISTER, G. S., K. KELTS, K.-Z. CHEN, J.-Q. Yu, AND E NIESSEN. 1991. Lake Qinghai, China: Closed-basin lake levels and the oxygen isotope record for ostracoda since the latest Pleisto- cene. P.aleogeogr. Palaeoclimatol. Palaeoecol. 84: 141-162.

MCKENZIE, J. A. 1985. Carbon isotopes and productivity in the lacustrine and marine environment, p. 99-118. In W. Stumm [ed.], Chemical processes in lakes. Wiley.

MONO LAKE: COMMITTEE. 1981. Mono Lake guidebook. Kutsavi Press. 113 p.

CRAIG, H. 1965. The measure of oxygen isotope paleotempera- tures, p. 9-130. In E. Tongiorgi [ed], Stable isotopes in ocean- ographic studies and paleotemperatures. CNR, Italy.

CRUSIUS, J., AND R. E ANDERSON. 1991. Core compression and surficial sediment loss of lake sediments of high porosity caused by gravity coring. Limnol. Oceanogr. 36: 1021-1031.

EPSTEIN, S., R. BUCHSBAUM, H. A. LOWENSTAM, AND H. C. UREY. 1953. Revised carbonate-water isotopic temperature scale. Bull. Geol. Sot. Am. 64: 1315-1326.

GASSE, E, AND OTHERS. 1987. Biological remains, geochemistry and stable isotopes for the reconstruction of environmental and hydrological changes in the Holocene lakes from North Sahara. Palaeogeogr. Palaeoclimatol. Palaeoecol. 60: l-46.

AND OTHERS. 199 1. A 13,000-year climate record from wektern Tibet. Nature 353: 742-745.

GAT, J. R. 1995. Stable isotopes of fresh and saline lakes, p. 139- 165. In A. Lerman et al. [eds.], Physics and chemistry of lakes. Springer.

NATIONAI, R.ESEARCH COUNCIL. 1987. The Mono Basin ecosystem: Effects of changing lake level. NRC.

NEWTON, M’. S., AND E. L. GROSSMAN. 1988. Late Quaternary chronology of tufa deposits, Walker Lake, Nevada. J. Geol. 96: 417-434.

O’NEK, J. R.., R. N. CLAYTON, AND T. K. MAYEDA. 1969. Oxygen isotope fractionation in divalent metal carbonates. J. Chem. Phys. 51: 5547-5558.

PHIIXPS, E M., AND OTHERS. 1992. A reconstruction of the rc- sponse of the water balance in Western United States lake ba- sins to climatic change. New Mexico Water Resour. Res. Inst. WRRI IRep. 26. V. 1.

GROSSMAN, E. L., AND T-L. Ku. 1986. Oxygen and carbon isotope fractionation in biogenic aragonite: Temperature effects. Chem. Geol. 59: 59-74.

RUSSELL, I. C. 1889. Quaternary history of Mono Valley: Califor- nia. U.S. Geol. Surv. 8th Annu. Rep.

SCHELSKE, C. L., AND D. A. HODELL. 1991. Recent changes in productivity and climate of Lake Ontario detected by isotopic analysis of sediments. Limnol. Oceanogr. 36: 961-975.

STREET-PERI~OTT, E A., AND S. F! HARRISON. 1985. Lake levels and climate reconstruction, p. 291-340. In A. D. Hecht [ed.], Paleocli mate analysis and modeling. Wiley.

TURNER, J. V., I? FRITZ, P E KARROW, AND B. G. WARNER. 1983. Isotopic: and geochemical composition of marl lake waters and implications for radiocarbon dating of marl lake sediments. Can. J. Earth Sci. 20: 599-615.

HOSTETLER, S. 1995. Hydrological and thermal response of lakes Received: 13 July 1995 to climate: Description and modeling, p. 63-82. In A. Lerman Accepted: 25 June 1996 et al. [eds.], Physics and chemistry of lakes. Springer. Amended: 23 October 1996