geology 72h - university of north carolina at chapel hillgeology 72h, december 2011 1 determining...

28
GEOLOGY 72H GE O L O G I C A L S C I E N C E S DECEMBER 2011 VOL 1 UNC GEOLOGICAL SCIENCES INSIDE: Sierra Nevada glacial moraines: hardness and shape Water chemistry and sediment character analysis Comparison of glacial and desert lake water deposits Salt precipitates at desert lake

Upload: others

Post on 23-Mar-2020

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72HG

EOLO

GIC

AL

SCIE

NCE

S

DECEMBER 2011 • VOL 1UNC GEOLOGICAL

SCIENCES

INSIDE: • Sierra Nevada glacial moraines: hardness and shape• Water chemistry and sediment character analysis• Comparison of glacial and desert lake water• deposits• Salt precipitates at desert lake

Page 2: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72HDECEMBER 2011•VOLUME 1

Determining the correlation between moraine age and steepness of slope in Quaternary moraines of eastern CaliforniaSarah Cooley, Alison Domonoske, Erin Moore, and Sagar Shukla

Boulder hardness as an indication of relative age of moraines in eastern CaliforniaAnnie Jin, Tyler Kress, Rachel Medlin, and Camille Morgan

Mineral composition of Deep Springs Valley playa, Inyo County, CaliforniaJonathan W. Villanueva, Morgan L. Johnson, Radhika H. Ghodasara, Ademide A. Adelekun

between glacier-fed and snow-fed lakes near Mt. Conness, Sierra Nevada, CaliforniaCayce Dorrier, Katie Jordan, Andrew Koo, Vinaya Polamreddi

Vining Creek, CaliforniaEmily J Auerbach, Mauricio A. Barreto, Sarah E. Graves, and Sloane K. Miller

River Gorge, CaliforniaGrace A. Blair, Carlos S. Floyd, Olivia L. Karas, and Kai T. Shin

1

5

9

13

17

21

The papers in this volume are the results of original research done by the students of Geol 72H, Field Geology of Eastern California, during the fall semester of 2011. This honors class is part of the First-Year Seminar program at the University of North Carolina at Chapel Hill. Beginning in 2009, students performed original research as part of this class, and this volume represents the first compilation of such research.

Field Geology of Eastern California is an expensive class, and we would like to thank all of those who have made this course possible. In particular, generous gifts from the Anadarko Petroleum Company have supported this class since its inception a decade ago, and in recent years the Honors Program at UNC has contributed significant support. The Graduate Research Consultant Program of the Office of Undergraduate Research and the Department of Geological Sciences both provided graduate student support so that the ratio of instructors to students is high. In 2011 this class was designated a Howard Hughes Medical Institute Collaborative First-Year Seminar.

I would like to particularly thank Prof. Drew Coleman, who joined the class as a second faculty member, graduate students Roger Putnam and Adam Curry, and undergraduate Siobhan Kenney for providing excellent field instruction to the students as well as setting good examples for how research is done and how to thrive in the field. The Sierra Nevada Aquatic Research Laboratory and the White Mountain Research Station provided excellent base camps for our efforts, and we all are grateful to the staffs there for their help and hospitality, in particular Dan Dawson, Denise Waterbury, and Tim Forsell.

The research presented herein is quite exciting, and it is sometimes difficult for me to believe that these students did this while in their first semester of college.

––Allen Glazner, Professor of Geological Sciences and Geol 72H instructor, Fall 2011

Cover: View of glacial lake at base of Mt. Dana in the eastern Sierra Nevada. Two glacial lakes and two glacial moraines occur in the canyon below the still-active glacier. As the glacier receded, deep basins eroded by the glacier filled with water to form the lakes, and glacial till formed the moraines. See Determining a correlation between moraine age and steepness of slope in Quaternary moraines of eastern California (p. 1) and Boulder hardness as an indication of relative age of moraines in eastern California (p. 5). Photo by Sagar Shukla; cover designs by Erin Moore and Alison Domonoske.

Page 3: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 1

Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines of eastern California Sarah Cooley, Alison Domonoske, Erin Moore, and Sagar Shukla Department of Geological Sciences, University of North Carolina, Chapel Hill, North Carolina 27599-3315, USA

ABSTRACT The steepness and slope angle of a glacial moraine

correlates with its age as well as with which side faces the glacier and whether it is a terminal or lateral mo-raine. In the Sierra Nevada Range of California, mo-raines of the Tahoe, Tioga, and Little Ice Age glacia-tions are well preserved, and a strong correlation exists between moraine age and relative steepness. In addi-tion, we compared the slopes of the sides facing the glac-ier to those facing away from it to study the effects of contact with the moving glacier. We found that the sides of the lateral moraines that faced the glacier were steeper than the sides that faced away from the glacier. However, the sides of the terminal moraines that faced the glacier were less steep than the sides that faced away from it. Our research demonstrates a method of differ-entiating moraines based on observing the extent to which erosional processes and weathering affect mo-raine slopes over time.

INTRODUCTION Glacial moraines serve as geologic “footprints,” pre-

serving evidence of former glaciers that advanced and re-treated thousands of years ago. As with other landforms, the steepness of a glacial moraine decreases with time and with exposure to the environment; thus, observing moraines to determine past instances of climate change requires an understanding of moraine shape. One method that analyzes how a moraine has weathered and eroded over time is semi-quantitative differentiation (Sharp, 1969). This technique classifies moraines temporally by analyzing surface boulder frequency and variations in grain sizes of the material in the moraine. In this study we examined the slopes on each side of several moraines. Analysis of published moraine slope angles has shown that these angles decrease over time, and a model has been developed to predict this degradation (Putkonen et al., 2007).

Our research aimed to examine various factors affect-ing moraine shape and degradation. While in the field, we

Figure 1. Map showing locations of surveyed moraines, eastern California. Inset shows location in California.

Page 4: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 2

worked alongside another research group whose project focused on examining patterns in boulder hardness on the surfaces of moraines. By contrasting the slopes of moraines from various glaciations in the Sierra Nevada region (Blackwelder, 1931) – including the Tahoe glaciation (~140,000 to ~80,000 years ago), the Tioga glaciation (26,000 to 18,000 years ago), and the Little Ice Age glacia-tion (1350 to 1850 A.D; all ages from Glazner and Sharp, 2010, p. 26) – we hoped to determine the extent to which a correlation exists between moraine shape and age of glacia-tion. We hypothesized that the older Tahoe moraines, which have experienced increased weathering and erosion, should exhibit gentler slopes than the younger Tioga and Little Ice Age moraines.

METHODS We surveyed 7 different moraines, 3 of Tahoe age, 3 of

Tioga-age, and 1 from the Little Ice Age, in McGee Can-yon, along Convict Creek, and in Glacier Canyon below Mt. Dana (Fig. 1). At each moraine location, we used a GPS receiver to record positions (WGS 84 datum) with which to locate our traverses. We conducted a single tran-sect on each moraine and selected the route by hiking to the crest of the moraine and determining a path free of major obstructions. To measure the horizontal angle of the tran-sect, we used a Laser Technology TruPulse 360 laser rangefinder, accurate to <1 m in horizontal distance. The laser was mounted on a tripod at the moraine crest and fired at a hand-held target on the flanks of the moraine. The tar-get holder walked along the transect route while attempting to stay within one degree of the desired horizontal angle, which was chosen to be perpendicular to the local moraine crest. At each measurement point, we shot the TruPulse laser at the target and recorded the horizontal and vertical distance to the target. The space between measurement points did not remain constant; however, we took enough data points to reconstruct the moraine shapes, averaging 4-8 m between shot points. Once we plotted the cross-section of the moraine and used a linear regression to find the mo-raine’s slope, we compared the angles from the horizontal by taking the inverse tangent of the slopes. We used maps in Wilkerson et al. (2007) to determine moraine ages and configurations in McGee and Convict canyons.

RESULTS Our data yielded complete transects of 6 moraines and

one side of a seventh. For each transect we collected an average of about 35 data points, ranging from 16 to 60 points per moraine. The sizes of the moraines varied signif-

Figure 2. Complete transects of the three steepest moraines, one from each period of glaciation. Each point represents a data point we took in the field. MD = Mount Dana, LIA = Little Ice Age, and MC = McGee Creek.

Figure 3. The angles of the moraines mapped along a unit circle. Each line represents the angle from the horizontal of each side of each mo-raine, and the lines are colored by the period of glaciation. Blue = Little Ice Age, green = Tioga, and purple = Tahoe. The legend is ordered by decreasing steepness. LIA represents Little Ice Age, MD = Mount Dana, MC = McGee Creek, T = terminal and L = lateral.

Figure 4. The slope angle of the three steepest moraines versus the approximate range in the number of years since the glaciation that formed them, using the age limits given in Glazner and Stock (2010). The legend is order by decreasing steepness. MD = Mount Dana, MC = McGee Creek.

Page 5: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 3

icantly, ranging from a vertical height of 12 meters to a height of 228 meters, and the slopes also varied greatly (Fig. 2). Many of the moraines are highly asymmetric (Fig. 2). We used these data to determine the angle from the horizontal of each side of the moraine (Figs. 3, 5, 6), utiliz-ing a linear regression to find the slope and then angle of each side. As most of the moraines are asymmetric, for data analysis we looked at the angles of each side of the moraine separately.

DISCUSSION

Age To examine the correlation between moraine slope and

age, we compared the angles of the slopes for three Tahoe-age moraines, three Tioga-age moraines and one Little Ice Age moraine. These data show that steeper moraine slopes correspond to younger glaciation periods. For example, the non-ice side of the youngest moraine, the Mount Dana Little Ice Age, has the steepest angle, 33.0°, of the mo-raines surveyed (Figs. 2-5). The other side of this moraine has the shallowest angle, 7.8°, which may have resulted from the glacier’s retreat (see below). Furthermore, the Tahoe-age moraines generally exhibit less steep angles than those of the Tioga moraines. The 6 slope angle values of the Tioga-age moraines fall between 17.8° and 31.7°, whereas the 5 slope angle values corresponding to the Ta-hoe-age moraines fall between 12.6° and 26.2° (Fig. 3). Thus, the relatively low Tahoe slope angle range shows that older moraines generally exhibit less steep slopes than their younger counterparts. We attribute this pattern to erosion and weathering, resulting in an overall smoothing and deg-radation of the moraine surface over time (Fig. 4) (Putko-nen et al., 2007).

Models of moraine degradation assume that the initial cross section of a moraine has a slope near the angle of repose, 34° (Putkonen et al., 2007). We found that the Lit-tle Ice Age moraine, the youngest and steepest moraine surveyed with a non-ice side angle of 33.0°, also has the angle closest to the angle of repose. The next-steepest mo-raine, the Convict Tioga, has an ice side angle of 31.7°, followed by the McGee Creek Tahoe, which has an angle of 26.2°. The angles of the three steepest moraines sur-veyed decrease with increasing age, and their relative spac-ing is comparable to the spacing of the glaciations in time (Fig. 4). This observation supports the assumption that a moraine has an initial angle of repose which continues to decrease as the moraine degrades over time.

Ice Sides By comparing ice sides and non-ice sides for each mo-

raine, we observed that terminal moraines are steeper on the non-ice sides whereas lateral moraines generally have steeper ice sides. (Figs. 2, 4, 5, 6). The non-ice sides of the terminal moraines are on average 14.2° steeper than the ice sides, whereas the ice sides of the lateral moraines are on average 7.3° degrees steeper than the non-ice sides (Figs. 5, 6). Of the four lateral moraines for which we surveyed both sides, only the McGee Creek Tioga moraine has a non-ice side steeper than its ice side, with an angle difference of only 0.2°.

The discontinuous recession of ice and its effect on the buildup of rocks may explain the less steep ice sides of

terminal moraines. A glacier “dumps” rather than “bulldoz-es” rock debris when it moves forward (Glazner and Sharp, 2010, p. 26); therefore, the repeated advance and retreat of a glacier may have resulted in a series of terminal moraine deposits up the mountain as the glacier alternated between melting and gaining volume. Unlike terminal moraines, lateral moraines form on either side of the glacier by the rocks pushed outward and dumped on the side as the glaci-er moves forward, meaning glaciers can gradually slide by the same moraines for thousands of years. Glaciers tend to pull rocks with them as they slowly move along their own lateral moraines. This movement exerts force on the piles of debris and causes them to steepen as the glaciers slowly shave rocks off on the ice sides, which may explain why our data show that lateral moraines have steeper ice sides than non-ice sides. We also observed creeks or drainages on the ice sides of each lateral moraine we surveyed, poten-

Figure 5. The angles of the terminal moraines mapped along a unit circle. Each line represents the angle from the horizontal of each side of each moraine, and the lines are colored by ice side vs. non-ice side. Red = non-ice side and blue = ice side. The legend is ordered by decreasing steep-ness. LIA = Little Ice Age, MD = Mount Dana.

Figure 6. The angles of the lateral moraines mapped along a unit circle. Each line represents the angle from the horizontal of each side of each moraine, and the lines are colored by ice side vs. non-ice side. Red = non-ice side and blue = ice side. The legend is ordered by decreasing steepness. MC = McGee Creek.

Page 6: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 4

tially affecting the steepness of the slopes because of in-creased erosion due to the presence of water. Furthermore, the McGee Creek Tioga moraine (the only lateral moraine steeper on the non-ice side) was the only moraine with a drainage on the non-ice side as well as the ice side, an ob-servation that certainly could help to explain this anomaly.

CONCLUSIONS Younger moraines have steeper maximum angles from

the horizontal than older moraines, supporting our hypothe-sis that erosion and weathering cause slopes to decrease as the age of the moraine increases. It is also likely that the initial angle from the horizontal of a moraine is close to the angle of repose, and over time erosion and gradual settling of the rocks causes the moraines to form angles less than the angle of repose. We also found that the ice sides of lateral moraines tend to be steeper whereas the non-ice sides of terminal moraines tend to have greater angles from the horizontal. This disparity is most likely caused by the differences in the way each moraine is formed; lateral mo-raines are formed by the rocks pushed out to the side by the glacier and terminal moraines are formed by the dumping of rocks at each moraine’s terminus. Thus, the ice side of the lateral moraine is steeper because the glacier carves out the ice sides as it passes the moraines, whereas the non-ice side of the terminal moraines is steeper because it is formed essentially by the dumping of rocks at the moraines’ termi-nus as thus is closer to the angle of repose. A limiting fac-tor in our study was the sample size; therefore, future stud-ies would include surveying a much larger selection of moraines across the Owens Valley and Yosemite region as well as other locations with glacial deposits around the world to determine if the correlation holds. Our studies have revealed other interesting trends which should be investigated further such as the variation of slopes from the ice side to the non-ice side and the effect of waterways on weathering.

ACKNOWLEDGMENTS This research was made possible through generous funding

from The First Year Seminar Program, Anadarko Petroleum, The James M. Johnston Center for Undergraduate Excellence and the Honors Program, The Office of Undergraduate Research Graduate Research Consultant Program, and the Department of Geological Sciences. Special thanks to Dr. Allen Glazner, Dr. Drew Coleman, Adam Curry, Roger Putnam and Siobhan Kenney for taking us to California, assisting in the field and back in North Carolina and overall making our research possible.

REFERENCES CITED Blackwelder, E., 1931, Pleistocene glaciation in the Sierra Nevada

and basin ranges: Geological Society of America Bulletin, v. 42, p. 865-922.

Glazner, A. F., and Stock, G. M., 2010, Geology Underfoot in Yosemite National Park: Missoula, Montana, Mountain Press Publishing Company,304 p.

Putkonen J., Connolly J., and Orloff T., 2007, Landscape evolution degrades the geologic signature of past glaciations: Geomor-phology, v. 97, p. 208-217.

Sharp R. P., 1969, Semiquantitative differentiation of glacial mo-raines near Convict Lake, Sierra Nevada, California: Journal of Geology, v. 77, p. 68-91.

Wilkerson, G., Milliken, M., Saint-Amand, P., and Saint-Amand, D., 2007, Roadside geology and mining history Owens Valley and Mono Basin: U.S. Bureau of Land Management,

http://www.blm.gov/pgdata/etc/medialib/blm/ca/pdf/bakersfield/geology.Par.56332.File.dat/ovm07_guidebook.pdf, p. 3-131.

Page 7: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 5

Boulder hardness as an indication of relative age of moraines in eastern California Annie Jin, Tyler Kress, Rachel Medlin, and Camille Morgan Department of Geological Sciences, University of North Carolina, Chapel Hill, North Carolina 27599-3315, USA

ABSTRACT Weathering causes the hardness of rocks in a mo-

raine to decrease as the moraine ages. Schmidt Hammer rebound (R) values measure relative hardness of rocks, which should decrease as moraines age. Weakening of rocks due to weathering produces lower and more vari-able R values. Variability of Schmidt Hammer readings requires that multiple measurements be taken to obtain an accurate representation of the average hardness of a rock. Moraines from the Little Ice Age, Tioga, and Ta-hoe glacial periods were studied. Sample sizes for Little Ice Age, Tahoe, and Tioga were n = 199, n = 529, and n = 270 respectively, where n is the number of R values taken. The mean R values of each glacial period de-creased with increasing age: Little Ice Age = 59.9, Tioga = 53.0, Tahoe = 49.2. The 95% confidence intervals of the mean R values did not overlap: Little Ice Age = (58.7, 61.1), Tioga = (51.9, 54.1), Tahoe = (47.9, 50.6). Approximately 100 measurements are needed to obtain a mean R value that is representative of the moraine. This allows for extrapolation of the relative age of a moraine from an average of R values taken in the field.

INTRODUCTION Understanding relative ages of moraines provides in-

formation about past climate and glaciation patterns. This information can be used to understand current climate pat-terns and predict future changes (Matthews, 2005). Mo-raines are ridges or mounds of till deposited by a glacier while at a stable position (Glazner and Stock, 2010, p. 278). Weathering weakens the boulders in a moraine, which de-creases their hardness (Putkonen et al., 2006). We used a Schmidt Hammer to measure the relative hardness of rocks (Goudie, 2006). Originally designed for measuring concrete hardness, the Schmidt Hammer is a handheld device that measures the rebound of a spring-fired rod after it strikes a surface. The reported value, R, is a measure of the force of the rebound. The Schmidt hammer is a useful instrument for measuring hardness in the field because of its portabil-ity, affordability, and simplicity. Temperature does not affect its performance significantly, and it can be easily calibrated (Goudie, 2006). Overall, the hammer is a con-venient means to establish a comparison of rock hardness while in the field, but its greatest limitation is its low preci-sion.

We studied seven different moraines in eastern Califor-nia in three different regions: Convict Creek, McGee Creek, and in Glacier Canyon below Mt. Dana (Figs. 1, 2). Moraines in these areas were formed during three different glacial periods: Tahoe, Tioga, and Little Ice Age. The Ta-hoe glaciation occurred 140,000-80,000 years ago, fol-lowed by the Tioga glaciation 26,000-18,000 years ago, and the Little Ice Age glaciation 700-200 years ago

(Glazner and Stock, 2010). The purpose of this experiment was to test whether the hardness of rocks in a moraine is related to the relative age of the moraine. We hypothesized that hardness of rocks in a moraine decreases as the mo-raine ages. Our study also sought to establish that the

Figure 1. Top: Base of the McGee Tioga moraine, showing a significant concentration of large boulders. Middle: Dana Little Ice Age Moraine. There were many more boulders in the Dana moraines than in the McGee or Convict Creek moraines, which were composed mostly of sediment and covered in vegetation. Bottom: Convict Tahoe Moraine, showing fewer boulders and more vegetation.

Page 8: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 6

Schmidt Hammer is a viable instrument to assign a moraine of unknown age to a given glacial episode.

METHODS We formulated a procedure using rocks in the rock gar-

den at UNC to refine sampling and data collection methods, and then applied this procedure to moraines in California. In the rock garden, we sampled a variety of rock types, sizes, and surfaces. These included granite, schist, slate, shale, and gneiss, ranging in volume from <0.1 m3 to >3 m3. This testing showed that variability of Schmidt Ham-mer R values increases as rock size decreases and as sur-faces became uneven or distal (McCarroll, 1989). The in-strument was calibrated on a steel anvil before we collected data.

To the extent possible, we avoided irregular surfaces, joints, edges, loose rocks, and rocks <0.1 m3 in volume. Goudie (2006) showed that rocks less than 25 kg in mass can yield inaccurate R values. We selected rocks from all elevation levels within each moraine and set a minimal distance of 3 m between rocks. These restrictions were imposed in order to obtain a representative sample of the moraine. Beyond these limitations, our selection of rocks throughout the moraines was random.

We tested the reproducibility of the R values by col-lecting readings from one spot on both a flat concrete sur-face and the surface of a boulder in the field. R values from the concrete surface had a range of 6, compared to the R values from the boulder surface that had a range of 20. We decided to take 5-10 readings per rock. The Schmidt Ham-mer was perpendicularly oriented against the rock faces for measurement consistency. We discarded measurements in the field when the hammer caused the rock to shift or frac-ture (Goudie, 2006).

RESULTS

Relative Age We analyzed our data by comparing the distributions of

R values of the rocks as grouped by their relative ages (Figs. 3, 4). The means of the distributions (Little Ice Age

= 59.9, Tioga = 53.0, Tahoe = 49.2) decrease with increas-ing age. The 95% confidence intervals of the three means do not overlap, strengthening the conclusion that the means of the distributions reflect the age of the moraine [Little Ice Age = (58.7, 61.1), Tioga = (51.9, 54.1), Tahoe = (47.9, 50.6)].

In Figure 4, the histograms display a left-skewed trend for the Tioga and Little Ice Age distributions. Low-value outliers account for this trend, and also lower the mean value relative to the median. The Tahoe distribution is more Gaussian; its mean and median are closer in value. All three distributions have wide ranges with standard devia-tions that are a large fraction of the mean (Little Ice Age = 8.52, Tioga = 12.86, and Tahoe = 11.05).

Rock Type We also evaluated the relationship between hardness

and rock type for the three Tioga moraines (Fig. 5). The Convict moraine showed a difference in the mean values between granite and other types of rock whereas the McGee moraine showed little difference in mean values. The Dana moraine was composed only of metavolcanic rock. The values from this moraine are higher than those from the McGee and Convict moraines.

DISCUSSION Figure 3 shows that mean hardness values decrease

with increasing moraine age. Because of the overlap among the distributions, an extrapolation of relative age based on a small number of measurements cannot be made. Although the means of the hardness values of the rocks decrease with age, the low precision of the measurements means that a

20

30

40

50

60

70

80

R V

alue

s

A. Little Ice Age B. Tioga C. Tahoe

Moraine Age

Figure 2. Locations of the seven moraines we sampled in Califor-nia. The Convict and McGee moraines are on the southern side of the Long Valley caldera and the Dana moraines are below Mt. Dana, which borders the eastern side of Yosemite National Park. Inset from http://www.50states.com/californ.htm

Figure 3. Box-and-whisker plots comparing the distributions of R values of moraines from three ages: Little Ice Age, Tioga, and Ta-hoe. The green lines represent the means of the sample; Little Ice Age = 59.9, Tioga = 53.0, Tahoe = 49.2. This shows a trend of de-creasing mean hardness with increasing age. The boundaries of the box represent the interquartile range (which contains 50% of the data). The red line in the middle of the box represents the median of the data (Little Ice Age = 62, Tioga = 55, Tahoe = 49). The whiskers are drawn to the endpoints (maximum and minimum). 95% confi-dence intervals for the means of each of the distributions were calculated: Little Ice Age = (58.7, 61.1), Tioga = (51.9, 54.1), Tahoe = (47.9, 50.6).

Page 9: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 7

large number of measurements must be taken to obtain a reliable mean.

We further analyzed the means using a bootstrap meth-od called case resampling. This method is useful in deter-mining the distribution of a statistic of interest (in our case, the mean) when the distribution is unknown. Case resampling utilizes an existing data set by taking random samples and calculating the statistic of interest for this

resample (Wu, 1986). We selected random samples of size m = 1, 2,…, n from within each data set, where n is the total number of measurements in a dataset, and calculated the cumulative averages of the samples. We used this method five times for each glaciation period and plotted these resamples as cumulative averages. At sample sizes of ~100 and greater, the cumulative averages stabilized around the means of the original data sets (Fig. 6). This analysis showed that the cumulative averages of R values stabilize around the discrete mean for each glaciation peri-od.

Rock Type We also looked at rock type as a factor that could con-

found the relationship between rock hardness and moraine age. The rock type varied among the moraines that we stud-ied. The Convict and McGee moraine source areas are largely composed of granodiorite and metasedimentary rocks, which we observed in the field (Rinehart and Ross, 1964). The Dana moraines are composed only of metavol-canic rocks (Kistler, 1966). To analyze whether rock type affected hardness readings, we split the hardness readings of the Tioga moraines into two categories—“granitic” and “other”. We chose the Tioga moraines to study because we had samples from moraines of that age from three moraine complexes, Convict, McGee, and Dana.

After dividing the data in this way, we concluded that there is not a clear correlation between R value of granitic rocks and the R value of non-granitic rocks. Among the three moraines, the granitic rocks produced neither clearly higher nor lower hardness readings than the non-granitic rocks (Fig. 5). The metavolcanic rocks of the Dana moraine were harder than the rocks of the other two moraines, gran-ite or otherwise.

Further Analysis In the field we were limited by time and accessibility of

the rocks. We spent between one and two hours on each moraine; due to this limitation, the number of rocks we could survey, along with the number of samples per rock, was smaller than ideal.

20

40

60

Count

20 30 40 50 60 70 80

1020304050

Count

20 30 40 50 60 70 80

20

30

40

50

60

70

80

R-V

alue

s

Con

vict

Gra

nite

Con

vict

Oth

er

Dan

a O

ther

McG

eeG

rani

te

McG

ee O

ther

Rock Type

Figure 6. Plot of cumulative average R vs. n comparing Little Ice Age, Tahoe, and Tioga moraines. Ran 5 trials for each moraine group and saw that around a sample size of n = 100, the cumulative average R value stayed consistently around the average R value of the total sample size.

Figure 4. Three histograms displaying distributions of the R val-ues for Little Ice Age, Tioga, and Tahoe moraines (with uniform scaling). The distributions of Little Ice Age and Tioga are skewed to the left (skewness value for Little Ice Age = -0.85, Tioga = -0.54), and Tahoe has a relatively normal distribution (skewness value = -0.07). Means, medians, and standard deviations of the distribu-tions are described in Figure 3.

Figure 5. Graph of box-and-whisker plots comparing the distributions of R values of the three Tioga moraines separated by rock type: granite and other. Green lines represent means (Convict granite = 50.9, Convict other = 55.9, Dana other = 61.7, McGee granite = 41.9, McGee other = 43.2).

25

50

75

Count

20 30 40 50 60 70 80

Reviewer
Typewritten Text
Little Ice Age
Reviewer
Typewritten Text
Tioga
Reviewer
Typewritten Text
Tahoe
Page 10: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 8

One question we still have based on the data is what might produce an outlier high in value. Outliers below the expected or actual minimum can be explained by a variety of factors: an unseen crack behind the tested surface that would absorb some of the rebound; a slipping of the ham-mer while taking a reading; or a reading taken on a particu-larly uneven or weak surface. We have not found an answer to how a reading could be higher than normal, other than variability of the instrument.

Alternative Explanation We examined why the distributions of R values for the

Tahoe and Tioga moraines were similar. One possibility is that rocks measured in the older moraines are largely re-sistant corestones (e.g., Glazner and Stock, 2010, p. 230-233). As rocks degrade over time, the outer and weaker surfaces weather; this leaves behind the harder core parts of rocks. Sediments, including dirt, sand, and soil, fill in the spaces between boulders. The rocks that remain are the harder, core rocks that have occasional exposed surfaces. Thus, samples do not always reflect the weathering that has taken place over time. Other methods that have been used to examine relative moraine age include boulder abun-dance, granitic weathering, grain-size distribution, and variations in color and pH (Sharp, 1969).

CONCLUSIONS Our data suggest that an average R value with a large

enough number of measurements reflects the true average of the hardness of rocks of a relative age. Extrapolation of a small number of R values is not reliable, but the average of at least 100 readings reflects an accurate average of a gla-cial period. Our hypothesis that boulder hardness decreases as a moraine ages was supported as our mean R values decreased with moraine age. Ultimately, the Schmidt Hammer can be used as an effective instrument in deter-mining the relative age of a moraine.

ACKNOWLEDGMENTS This research was supported by the Department of Geological

Sciences at the University of North Carolina at Chapel Hill, the Anadarko Petroleum Corporation, the Howard Hughes Medical Institute, the First-Year Seminar Program, and the Office of Un-dergraduate Research Graduate Research Consultant Program. We would like to thank Dr. Allen Glazner, Chair of the Department of Geological Sciences, for his work in organizing the trip and advis-ing us in this project. We would also like to thank Dr. Drew Cole-man for his help and support in the field, and Dr. Jason Barnes for his help in interpreting our data. We also would like to thank grad-uate students Adam Curry, Roger Putnam, Jonathan Syrek, and undergraduate student Siobhan Kenney their help throughout our project.

REFERENCES CITED Glazner, A., and Stock, G., 2010, Geology Underfoot in Yosemite

National Park: Missoula, Montana, Mountain Press Publishing Company, 304 p.

Goudie, A., 2006, The Schmidt Hammer in geomorphological research: Progress in Physical Geography, v. 30, p. 703-718.

Kistler, R., 1966, Geologic map of the Mono Craters Quadrangle, Mono and Tuolumne Counties, California, Geologic Quadran-gle Map - U. S. Geological Survey, Report: GQ-0462.

Matthews, J., 2005, ‘Little Ice Age’ glacier variations in Jotunhei-men, southern Norway: a study in regionally controlled li-chenometric dating of recessional moraines with implications for climate and lichen growth rates, The Holocene, v. 15, p. 1-19.

McCarroll, D., 1989, Potential and limitations of the Schmidt Hammer for relative-age dating: field tests on neoglacial mo-raines, Jotunheimen, Southern Norway: Arctic and Alpine Re-search, v. 21, p. 268-275.

Putkonen, J., Connolly, J., and Orloff, T., 2006, Landscape evolu-tion degrades the geologic signature of past glaciations: Geo-morphology, v. 97, p. 208-217.

Rinehart, C. D., and Ross, D. C., 1964, Geology and mineral de-posits of the Mount Morrison quadrangle, Sierra Nevada, Cali-fornia: United States Geological Survey Professional Paper, v. 385, 106 p.

Sharp, R., 1969, Semiquantitative differentiation of glacial mo-raines near Convict Lake, Sierra Nevada, California: Journal of Geology, v. 77, p. 68-91.

Wu, C., 1986, Jackknife, bootstrap and other resampling methods in regression analysis (with discussions): Annals of Statistics, v. 14, p. 1261-1350.

Page 11: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 9

Mineral composition of Deep Springs Valley playa, Inyo County, California Jonathan W. Villanueva, Morgan L. Johnson, Radhika H. Ghodasara, Ademide A. Adelekun Department of Geological Sciences, University of North Carolina, Chapel Hill, North Carolina 27599-3315, USA

ABSTRACT Deep Springs Lake, which once covered over half of

Deep Springs Valley, has undergone extensive desicca-tion, resulting in the formation of a playa. There is visible stratification of the lacustrine deposits separated into regions of saline crusts, carbonate muds, and fine-grained, well-sorted silt and sand from the outside of the playa inward. Jones (1965) found a systematic pro-gression of minerals as one goes from the outer rings to the interior of the basin. To test these observations, we took samples from salt deposits along a trajectory in the playa surrounding Deep Springs Lake. X-ray diffrac-tion and scanning electron microscopy were utilized to analyze the mineral composition of each sample. The most prevalent minerals were calcite, thenardite, burkeite, halite, and aphthitalite, with a weak pattern of mineral progression. Gaylussite was completely absent from our results, and dolomite was only found in one sample. Gaylussite and dolomite were repeatedly men-tioned in findings by Jones (1965) and Guyen and Kerr (1966).

INTRODUCTION

Playa Formation A playa is a deposit of clay and salts that results from

desiccation of a lake. In order for a playa to be sustained, there must be drainage to a zone with rates of evaporation exceeding the rate of inflow. Playas are flat-bottomed de-pression, located in the central basins of arid and semi-arid regions (Eugster, 1979). As the body of water evaporates, there is significant deposition of salt, sand and mud along the bottom and around the edges of the depression. Deep Springs Lake (Fig. 1) is a small salt lake in a completely closed basin in Inyo County of eastern California at an altitude of around 1500 meters (Eugster, 1980).    

Significant climate change and recent faulting has re-sulted in a shift in channel flow that led to significant des-iccation of the lake and created conditions favorable for playa formation. As recently as 1000 years ago, Deep Springs Lake occupied more than half of Deep Springs Valley. This area is now covered by lacustrine deposits of the playa, spanning as far as eight kilometers from the cen-ter of the playa (Fig. 2; Jones, 1965).

Lacustrine Deposits The lacustrine deposits of the playa comprise three

primary categories. From the center of the playa outward these are saline crusts, carbonate muds, and fine-grained, well-sorted silts and sands. We expected to find a systemat-ic sequence of minerals going from the outer rings of evap-orative basins to the inner rings, with solubility increasing progressively (Table 1). We expected that these would be comparable to the areal patterns of mineral composition observed by Jones (1965; Fig. 2).

Figure 1. Aerial image (from Google Maps) showing Deep Springs Playa and its general location in eastern California. Each red dot marks a sample location. We collected 22 samples, two of which were in the same location on the map. Once we reached the loca-tion where we collected sample 21, the playa surface was too muddy for continued sampling. Sample 1 is not shown and was taken ~1000 m north of the north edge of the map. Concentric zonation of salts and clays is evident.

Figure 2: Map showing zonation of lacustrine minerals in Deep Springs Valley playa, from Jones (1965). The red dashed line shows our sampling traverse. From the outer edge inward, we expected to find calcite and aragonite, dolomite, gaylussite, and perhaps thenardite.

Page 12: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 10

Table 1: The table above shows the solubility of various minerals that were expected to be found in Deep Springs Valley playa. Ac-cording to our hypothesis, we expected that less soluble minerals would be more present on the outer lacustrine rings. As one goes further towards the center of the playa, we expected that the mineral solubility would increase.

METHODS

Data Collection Sampling was limited by the logistics of road condi-

tions and playa muddiness. Our trajectory was intended to be a straight line toward the center of the playa, leaving from Highway 168 about 2 km east of the mouth of Payson Canyon (Fig. 1). Muddy conditions forced us to change this plan because there were areas of the playa that could not be traversed. Our final decision was to work our way to the center of the playa by walking along the curving path, which allowed us to walk part way into the saline crust portion of the playa.

Figure 3: Location of sample 6, looking southeast, showing evaporite rings around Deep Springs Lake.

Figure 4: Sample 21 was very brittle and fractured in thin sheets. Not all of the samples had the same crystal form. Some were soft and powdery whereas others were thick and flaky.

Figure 5: This photograph, taken where sample 7 was collected, displays an area where the salts change in grain size. The bottom left of the photo shows flaky coarse-grained salts whereas the area on the top right of the photo shows thinner, finer-grained sheets of salt. The salts were precipitated on desiccated brown mud.

Figure 6: In contrast to Figure 5, this photograph, taken where sam-ple 11 was collected, shows more evenly dispersed precipitate. Such changes in the deposition of salts allowed us to choose where to take samples.

Samples were collected at sites where we noticed obvi-ous changes in vegetation or the structure of crystals seen (Fig. 1). Twenty-two samples were acquired as we walked to the center of the playa from the northwest. Our trajectory was from the northwest edge of Deep Springs Valley, walk-ing south from sampling site 1 to site 18, then southeast from sampling site 18 to 22. In Figure 1, there are visible rings where high concentrations of salt are present.

At each ring we encountered, we collected salt samples with a spoon, placing 10-15 g of salt precipitate from the surface layer of the playa into a small plastic zip sample bag that was tightly sealed and labeled. Sample locations were documented using a GPS and the WGS 84 datum. Throughout working in the field, notes were taken on the surroundings and the relative location of sample collection. After returning to camp, all samples bags were sealed in a 1-gallon plastic zip bag with 500 ml rice to keep out mois-ture.

Data analysis The salt samples were crushed into dust with a mortar

and pestle in order to run analysis with X-ray diffraction on a Rigaku MiniFlex II diffractometer. We used double-sided tape to mount powders on glass slides. Samples were placed into the X-ray diffractometer and scanned at 2°/minute at a step size of 0.02° (Fig. 7). Mineral identifi-cation was aided by Rigaku’s PDXL program. To limit error, we made sure that minerals selected during XRD analysis were evaporites expected in this area. Frequent reference was made to Ralp and Chau’s (1993b) list of expected minerals in Deep Spring Valley Playa.

Mineral Name Chemical Formula Solubility, mol/l Calcite CaCO3 0.0006170 (in

20 °C water) Thenardite Na2SO4 19.5 (in 20 °C

water) Burkeite Na6(CO3)(SO4)2 Soluble in

water Halite NaCl 35.89 (in 20 °C

water) Aphthitalite (K,Na)3Na(SO4)2 Soluble in

water Dolomite CaMg(CO3)2 Soluble in

water Gaylussite Na2Ca(CO3)2·5H2O Soluble in

water Aragonite CaCO3 0.0006170 (in

20 °C water)

Page 13: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 11

Figure 7: The image above displays the XRD diffractogram data from sample 10, plotted as 2θ vs. intensity of the diffracted X-rays. Each crystallize phase has a unique pattern, and unknown samples can be matched against a library of mineral patterns. In this sample we recognized thenardite and halite.

RESULTS

X-Ray Diffraction The most prevalent minerals were calcite, thenardite,

burkeite, halite, and aphthitalite (Figure 8). Gaylussite was not identified in any samples, and dolomite was present in only one sample. Calcite was present near the start our trajectory at sampling points 1, 5, 6, and 7, with the last appearance of calcite at 11. Thenardite appears in 15 out of 22 of the samples. Burkeite first appeared in 7 and became more common close to the center of our sampling range. Halite first appeared in sample 6 and was present in almost all of our samples from the rest of the traverse. Aphthitalite was present in two samples near the edge of the playa and then became more prevalent in samples between sampling sites 7 and 15. The remaining minerals found only appeared once or twice and showed no spatial pattern.

Scanning Electron Microscope We examined samples 10 and 17 using a scanning elec-

tron microscope to observe the texture and structure of the crystals. Figure 9 shows sample 10’s halite crystals (repre-sented in white) and thenardite (represented by the darker black regions). The smooth gray surfaces are melted halite that hardened into a smoother form. When we looked at the crystal in the yellow circle, we noticed an interesting com-bination of elements that contained elements of both the-nardite and halite. None of the other minerals we recorded contained this combination of elements. We determined that the mineral hanksite a good candidate for this mystery crystal due to its chemical composition (Ralph and Chau, 1993a).

Figure 8: Minerals present in samples 2 through 21 as determined by X-ray diffraction. A marker signifies the presence of a given mineral. There were numerous minerals observed in samples that were ap-proximately 400 m to 800 m from sample 2. Aphthitalite is observed sparsely in samples 0 m to 100 m from sample 2, then consistently in samples 475 m to 700 m from sample 2. Halite is consistently present in samples 440 m to 800 m from sample 2, then seen sporadically to the southeast. Burkeite is present regularly in samples 500 m to 800 m from sample 2, then seen once approximately 350 m further into the traverse. Thenardite is observed from 400 m to 800 m with rela-tive consistency.

Figure 9: This SEM image of sample 10 shows an abundance of thenardite and halite. The smooth gray patches, denoted by a red cross, are halite that melted and then re-hardened, but with a smooth surface instead of the crystal structure of halite which is typically seen in the white areas of the image, denoted by the blue cross. The darker black regions, denoted by a green cross, are thenardite crys-tals that probably served as a foundation for the halite crystals to grow. The area inside of the yellow circle is probably hanksite (see text) that grew on top of the other crystals. We concluded that the-nardite formed first, followed by halite and then finally hanksite.

0 500 1000 1500 Distance from Sampling Site 2 (m)

Mineral Composition of Samples 2 through 21

Calcite Thenardite Burkeite Halite Aphthilite

Page 14: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 12

Figure 10: The backscatter SEM image of sample 17 above is sensi-tive to differences in mean atomic number and can therefore show differences in the chemical make-up. The white areas are halite whereas the surrounding darker areas are thenardite. Backscatter imaging shows clearer distinctions between halite and thenardite.

We also examined a backscatter image of sample 17 (Fig. 10) and found thenardite and halite.

DISCUSSION Jones (1965) reported a consistent zonation of mineral

composition in lacustrine deposits of Deep Springs Valley playa. Listed from edge of the playa to the center, this zo-nation is aragonite, calcite, dolomite, gaylussite, thenardite, and burkeite. We expected solubility to play a key role in the deposition of precipitates, with the least soluble miner-als located at outer rings For example, calcite is composed of CaCO3, a compound that has very low solubility (Table 1). Jones (1965) reported that calcite was found in the outermost rings of the playa, followed inward by dolomite, a slightly more soluble mineral (Table 1).We hypothesized that a similar trend would be found for our samples.

Although our results did show a weak pattern of miner-al zonation, they did not reflect the expected progression of minerals as we moved toward the center of the playa. We found little dolomite and no gaylussite, two minerals that were repeatedly mentioned by Jones (1965) and Guyen and Kerr (1966). Gaylussite was completely absent from our results, and dolomite was only found in sample 5. SEM images of 10 and 17 show that halite grew on top of the thenardite. In 10, we found hanksite growing on top of what looked to be halite. This suggests that the thenardite formed first, the halite next, and hanksite finally formed on top of halite. Hanksite is known to grow in areas with an abundance of halite, and forms from the halite, which would support the fact that we found hanksite forming on top of halite (Ralph and Chau.1993a).

Multiple sources of error could have contributed to the deviations in our results and those of Jones. One major aspect of our methodology that hindered us was the inabil-ity of the XRD method to provide information about the relative amounts of each mineral that was found to exist in a sample; the XRD data only informed us of the presence or absence of a mineral based on the pattern of peaks rather than quantified proportions of specific minerals. Addition-ally, much time has passed since Jones’s study, and perhaps

climatology and other factors have affected the distribution of evaporites since then.

CONCLUSIONS Aerial photographs of the Deep Springs Lake clearly

show the presence of concentric patterns of salts on the playa surface. Previous research predicted the following sequence of minerals as one goes from the outer rim of the playa towards its center: aragonite, calcite, dolomite, gay-lussite, thenardite, and burkeite (Jones, 1965). We collected samples from one traverse on the playa and analyzed the samples using X-ray diffraction and scanning electron mi-croscopy. Our results showed a prevalence of calcite, the-nardite, burkeite, halite, and aphthitalite; however, we did not observe a clear zonation of these minerals.

We expected changes in salt deposition, crystalline structure, and areas of vegetation to correspond with changes in mineral composition in our data. The anomaly that there was a weak pattern in mineral progression could be explained by the use of XRD in data analysis. Although this method gave notification of mineral presence and ab-sence, there was no indication of the proportions in which these minerals were present.

ACKNOWLEDGMENTS This research was supported by the Department of Geological

Sciences at the University of North Carolina at Chapel Hill, the Anadarko Petroleum Corporation, the Howard Hughes Medical Institute, the First-Year Seminar Program, and the Office of Un-dergraduate Research Graduate Research Consultant Program. Thank you to all of the teaching faculty who have given us a solid geological background and offered us support through this journey, especially our professor, Dr. Allen Glazner. We would also like to give a special thanks to Adam Curry and Roger Putnam for aiding in data collection on the field and analysis in the XRD lab.

REFERENCES CITED Eugster, H. P., 1980, Geochemistry of evaporitic lacustrine depos-

its: Annual Review of Earth and Planetary Sciences, v. 8, p. 35-63.

Eugster H. P., and Jones, B. F., 1979, Behavior of major solutes during closed-basin brine evolution: American Journal of Sci-ence, v. 279, p. 609-631.

Guven, N., and Kerr, P. F., 1966, Selected Great Basin playa clays: The American Mineralogist, v. 51, p. 1056-1067.

Jones, B. F., 1965, The hydrology and mineralogy of Deep Springs Lake Inyo County, California: Geological Survey Professional Paper 502-A, 56 p.

Ralph J., and Chau I. 1993a, Classification of hanksite. http://www.mindat.org/min-1815.html.

Ralph J., and Chau I. 1993b, Deep Springs Lake (Deep Springs Playa), Deep Springs Valley, Inyo Co., California, USA. http://www.mindat.org/locdana-3451.html (November 2011).

Page 15: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 13

Geochemical differences between glacier-fed and snow-fed lakes near Mt. Conness, Sierra Nevada, California Cayce Dorrier, Katie Jordan, Andrew Koo, Vinaya Polamreddi Department of Geological Sciences, University of North Carolina, Chapel Hill, North Carolina 27599-3315, USA

ABSTRACT The four Conness Lakes have long been character-

ized by their discrepancy in color, with the two lakes fed by glacial meltwater having a noticeably lighter tur-quoise color than the two lakes fed by snow melt. We examined 2 glacier-fed lakes (GF1 and GF2), 2 snow-fed lakes (SF1 and SF2), , an unnamed small lake composed almost completely of glacial meltwater (SGL), Green-stone Lake (GS), which is fed by streams from glacier-fed and snow-fed lakes, and 2 streams in between the lakes (S1 and S2). Utilizing X-ray diffraction, scanning electron microscopy, inductively coupled plasma mass spectrometry, and a Eureka Manta water multi-probe, we found differences between the lakes in turbidity, concentrations of cations, sediment size, and diatom presence; however, we did not find the one difference we expected to find: the discrepancy in the color be-tween the lakes. There was not a noticeable color differ-ence between the snow-fed and lakes when we visited them in October 2011. Further research could include looking into whether seasonal changes affect the amount of glacial flour in the lakes.

Figure 1. Map of Conness Lakes (2 glacier-fed: GF1, GF2, and 2 snow-fed: SF1, SF2), Greenstone Lake (GS), an unnamed, small glacier-melt lake (SGL) near the Conness glacier, and two streams (S1 and S2) in between the lakes. S1 connects SF2 and GF2, and S2 connects GL2 and GS. The red dots represent the locations where we collected water and sediment samples.

INTRODUCTION

Some of the most enjoyable parts of scenic alpine mountain hikes are the beautiful turquoise glacier-fed lakes. The striking turquoise color of the glacier-fed lakes, so unlike most other lakes we see, is caused by glacial flour suspended in these lakes. Glacial flour is sediment that has been finely ground by glaciers into much smaller sizes than sediment ground by fluvial processes (Dreidger, 1986, p.

80). One of the best examples of these turquoise lakes is the glacier-fed Conness Lakes (Fig. 1). Located on Mount Conness (Guyton, 1998, p. 136) just east of the eastern boundary of Yosemite National Park, the Conness Lakes are a set of 4 lakes, 2 of which are glacier-fed and turquoise in color during part of the year and 2 of which are snow-fed and blue in color (Fig. 2.a). Intrigued by the contrast, we set out to explore the differences between these two lakes in water chemistry, turbidity, color-dissolved organic mat-ter, and sediment composition.

2.a

2.b

Figure 2. (top) Photograph demonstrating the contrast between the GF and SF lakes in August of 2005. The differences in water color are due to the active mechanical production of glacial flour at this time due to fairly high temperatures, which promote sliding (T. Pavelsky, personal communication, 2011). (Photo courtesy of Tamlin Pavelsky) (bottom). Photograph of the GF lakes in October of 2011 showing similar color to the SF lakes in top picture. The small differences in lake color are likely due to the inactive production of glacial flour while Conness Glacier is frozen to its bedrock due to fairly low tem-peratures (T. Pavelsky, personal communication, 2011). (Photo cour-tesy of Jonathan Villanueva)

METHODS

Water and Sediment Samples We collected samples from GF1, GF2, SF1, SF2, the

SGL, GS, S1, and S2 (Fig. 1). We took a water and sedi-ment sample from each of the lakes and only water samples from S1 and S2. At the site of collection, we placed the sample bottles on the surface of the water at an angle and

Page 16: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 14

allowed water to fill the bottle. We then removed the bot-tles and sealed them with tape to prevent leakage. To col-lect sediment samples, we inserted a 2.5 cm PVC pipe into the lake floor around 0.5 m from the edge of the lake in 0.5-1 m of water. Once the pipe was adequately pressed about 5 cm into the sediment we placed a hand on the top opening of the pipe and pulled it out of the water. Just above the surface of the lake, we emptied the PVC pipe into a Ziploc bag.

Manta Multi-Probe We tested the water in each lake using a Eureka Manta

multi-probe (Eureka Environmental, 2011). The probe test-ed for temperature, turbidity, chlorophyll, color dissolved organic matter (CDOM), and specific conductivity. We lowered the probe into the water off a rock around 0.5-1.0 m into the lake to ensure that the water collected was repre-sentative of the rest of the water source. The Manta measures turbidity, which is caused by suspended matter such as clay, silt, finely divided organic and inorganic mat-ter, and plankton and other microscopic organisms, using the light-transmission scattering method, in which scattered light resulting from the suspended substances in the sample and the light which passes through the sample are meas-ured. The unit for turbidity is NTU (nephelometric turbidity unit; Missouri, 2007). We also took several precautions to ensure the validity of the turbidity readings; we made sure that the water was undisturbed prior to testing and did not allow the probe to touch the bottom of the lake. We then put the probe into the water to collect data for 3 minutes.

ICP-MS To investigate the composition of various common cat-

ions in the different bodies of water, we ran the samples through an inductively coupled plasma mass spectrometer (ICP-MS). We used a micropipette to fill 25 mL test tubes with 5.0 mL of our sample water. We then spiked each of the test tubes with 250 mL of nitric acid to help with signal stability and act as a matching matrix for all samples. To standardize the machine we ran 8 standards that contained all elements of interest in concentrations of 25 parts per billion (ppb), 50 ppb, 100 ppb, 200 ppb, 400 ppb, 600 ppb, 800 ppb and 1000 ppb, and compared the reported concen-trations of each element to the expected concentrations to see how accurately the machine was measuring that par-ticular element. We also cleaned the lines with 2% nitric acid to avoid contamination. Inside the ICP-MS the sample was aspirated with argon, forming a mist that is vaporized in the plasma torch. The torch ionizes the atoms, which are then removed from the plasma and separated based on their mass to charge ratio (Worley and Kvech, 2011). After around 3 minutes, the ICP-MS reported the concentrations of various cations, including Mg, Si, P, and K, in ppb.

Scanning Electron Microscope Analysis To distinguish the different sizes of sediment from our

various water samples more accurately, we observed our sediment samples under a scanning electron microscope. To prepare the samples for analysis, we took 3 g aliquots of each sample of sediment and dried them at 50°C for 1.5 hours using an oven. We placed each sample in a vial and then added equal volumes of distilled water and vigorously shook the bottles. Sediments were allowed to settle for 2

minutes, at which time a drop of water from the top of the column was collected with a pipette. These drops were placed on a glass slide and then evaporated in a drying oven.

RESULTS

3.a

3.b Figure 3. Graphs of the concentrations of various common cations in the SGL, GF lakes, SF lakes, the streams, and Greenstone Lake (GS). 3.a is a graph of the concentrations of P vs. K in ppb. It shows that the SGL had the highest concentrations of many of the cations that we tested, likely owing to the presence of glacial flour. The SF lakes generally had the lowest concentrations of cations, and in this graph they have no measurable concentrations of K and low concentra-tions of P. The streams and GS had concentrations in between that of the SF lakes and GS. 3.b is a graph of the concentrations of Mg vs. Si in ppb. This graph also demonstrates the high concentrations in the SGL. The SF lakes again have lower concentrations of the cati-ons, probably owing to the absence of glacial flour, followed by the streams, the GF lakes, and GS.

Solute Concentrations Glacier-fed lakes have higher cation concentrations

than snow-fed lakes due to the weathering of rocks by glac-iers (Mitchell et al., 2001). Likewise, the SGL had the highest concentrations of a majority of the cations we test-ed, followed by the GF lakes. Figures 3.a and 3.b show the concentration of four cations, P, K, Mg, and Si, in the water sources that we tested. The SGL had significantly higher

Page 17: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 15

concentrations of these cations than any of the other water sources. The GF lakes had higher concentrations of these cations than the SF lakes, which consistently had the lowest concentrations. The concentrations of these cations in the streams were in between those of the SF and GF lakes.

Turbidity and Dissolved Organic Matter Although the SGL had consistent turbidity readings

above 0, the GF lakes did not have any measurable turbidi-ty (Fig. 4). It is normal for glacier-fed lakes to have a high turbidity owing to the presence of glacial flour (USGS, 2011), so the absence of turbidity in the GF lakes is surpris-ing. The SF lakes and GS also had no measurable turbidity, which was less surprising because these are not glacier-fed lakes with glacial flour. GS and the SF lakes had the high-est CDOM concentrations, with the SGL having a higher = CDOM concentration than the GF lakes (Fig. 5).

Figure 4. Graph of turbidity in the water samples in NTU. The values are an average of the numbers collected by the Manta over a 3-minute period. The SGL was the only water source tested with a measurable turbidity, which we ascribe to the presence of glacial flour.

Figure 5. Graph of CDOM concentrations in the water samples in micrograms per liter. The values are an average of the numbers collected by the Manta over a 3-minute period. GS and the SF lakes had the highest CDOM concentrations, which we connected to the presence of diatoms, which decompose into CDOM. The SGL had a higher average CDOM concentration than the GF lakes, but it was less than half the concentrations of the SF lakes and GS.

Sediment Size Analysis The sediment size evaluation was difficult owing to

clumping of sediments; however, upon close inspection, it is clear that GF1 and GF2 (Fig 6.a) contains a higher quan-tity of sediments <5 µm in largest dimension than SF1 (Fig 6.b) and SF2. This is probably a result of the same process that results in glacial flour, leading to a higher abundance of the smallest sediments found in the GF lakes compared to the SF lakes.

Figure 6. SEM Images of sediment from GF2 (6.a), SF1 (6.b), GS (6.c), and enlarged image of diatom from SF1 (6.d). Diatoms can be seen in SF and GS samples. There are more sediments <5 µm in the glacier-fed lakes than in the snow-fed lakes..

Diatoms In our analysis of the SEM images, we noticed oval

structures with comb-like protrusions in samples from SF1

Page 18: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 16

(Figs. 6.b,d), SF2, and GS (Fig. 6c). These proved to be diatoms, unicellular algae that are a common type of phy-toplankton (Wikipedia:File:Diatomeas-Haeckel), based on comparison with published photos and consultation with Dr. Pat Gensel (personal communication, 2011). In SF1, there are about 10 diatoms per 200 µm by 150 µm sample area (Fig. 6.b); however, we found no diatoms in GF1 and only one in GF2. These observations suggest that some characteristic of the snow-fed lakes, non-existent in the glacier-fed lakes, allows these diatoms to thrive in their environment. Color dissolved organic matter (CDOM) is higher in SF1 and SF2 than in most other samples (Fig. 5), likely owing to decomposition of diatoms (Kamatani, 1982). GS had a higher CDOM concentration than either of the snow-fed lakes, so we analyzed GS for diatom pres-ence, and it contained about 35 diatoms per 350 µm by 250 µm (Fig. 5.e). There were 3.33 per mm 2 diatoms in the snow-fed lakes and 400 per mm2 in Greenstone Lake in our images.

DISCUSSION One difference that we expected to see but did not was

a difference in color between the glacier-fed and snow-fed lakes (Fig. 2b). Dr. Tamlin Pavelsky (personal communica-tion, 2011) suggested that this phenomenon could be due to the limited glacial flour caused by seasonal changes. One possible explanation is that the Conness glacier froze to its bedrock during autumn temperatures while we were there in October 2011. In August 2005, the glacier-fed and snow-fed lakes exhibited a much starker difference in color (Fig. 2.a). If this difference in color is a result of seasonal chang-es in glacier sliding, color differences could be used to tell when glaciers start and stop melting. Such differences may be resolvable from orbital imagery, allowing seasonal changes in glacier behavior to be monitored by remote sensing (T. Pavelsky, personal communication, 2011). An-other area of possible further research is our hypothesis that glacial flour may inhibit diatom growth made from our findings of diatoms not being present in glacier-fed lakes

CONCLUSION In our comparison of glacier-fed lakes and snow-fed

lakes in the Sierra Nevada Range, we noted differences in turbidity, concentration of cations, sediment size, and dia-tom presence. The SGL had high concentrations of various common cations, which are likely a product of the glacial flour that results from the mechanical action of ice on bed-rock. The SGL also had the only measurable turbidity level, most likely also caused by glacial flour. These data show a correlation between turbidity and cation concentration in glacier-fed lakes. Compared to the glacier-fed lakes, the high CDOM in the snow-fed lakes can be explained by the presence of diatoms that are only abundant in sediments in the snow-fed lakes. Further research is needed to identify the difference between snow-fed and glacier-fed lakes that fosters the growth of diatoms in the snow-fed lakes and inhibits them in glacier-fed lakes. Our analysis of sediment sizes in the two types of lakes showed that there is a higher quantity of sediments <15 µm in glacier-fed lakes than snow-fed lakes. This provides further evidence for the claim that the mechanical interactions between glaciers and bedrock result in the grinding of rocks into fine-grained sediments that are found in abundance in glacier-fed lakes.

ACKNOWLEDGMENTS This research was supported by the Department of Geological

Sciences at the University of North Carolina at Chapel Hill, the Anadarko Petroleum Corporation, the Howard Hughes Medical Institute, the First-Year Seminar Program, and the Office of Un-dergraduate Research Graduate Research Consultant Program. We would like to thank Roger Putnam for helping with data collection and for carrying our broken Cayce down Conness Mountain, Dr. Sorhab Habibi for his help with the ICP-MS, and Dr. Tamlin Pavelsky for his help with the Manta and data analysis. We would also like to thank Adam Curry for helping with SEM analysis and other testing. Finally, we want to thank Allen Glazner and Drew Coleman for making this trip to California possible and giving us the opportunity to do this research project. REFERENCES CITED Diatom-Wikipedia. [Internet]. 2011. [cited 2011 Nov 17] Available

from: http://en.wikipedia.org/wiki/Diatom. Driedger, C., 1986, A visitor's guide to Mount Rainier glaciers,

Seattle, Washington, Northwest Interpretive Association,[cited November 2011 ]. Available from: http://vulcan.wr.usgs.gov/

Volcanoes/Rainier/Publications/PNNPFA-Driedger86/glacier_features.html.

Eureka Environmental. [Internet] 2011. [cited 2011 Nov 22] Available from: http://eurekaenvironmental.com/.

Guyton, B., 1998, Glaciers of California, Berkeley, University of California Press, 197 p.

Kamatani, A., 1982, Dissolution rates of silica from diatoms de-composing at various temperatures: Marine Biology, v. 68, p. 91-96.

Mitchell, A., Brown, G., and Fuge, R., 2001, Minor and trace element export from a glacierized alpine headwater catchment (Haut Glacier d'Arolla, Switzerland): Hydrological Processes, v. 15 p. 3499-3524.

Standard Operating Procedure for: Eureka Amphibian and Manta Water Quality Multiprobe for Multiple Location Parameter Measurement. [Internet]. 22. Missouri: Missouri State Univer-sity. [cited 22 Nov ] Available from: http://oewri.missouristate.edu/assets/OEWRI/1200R02_Eureka_Snapshot.pdf.

U.S. Geological Survey. [Internet]. 2007. [cited 2011 Nov 18] Available from: http://vulcan.wr.usgs.gov/Glossary/ Glaci-ers/glacier_terminology.html.

Wikipedia:File:Diatomeas-Haeckel. [Internet]. 2011. [cited 2011 Nov 17] Available from: http://en.wikipedia.org/wiki/ File:Diatomeas-Haeckel.jpg.

Worley J, Kvech S. 2011. ICP-MS. [Internet]. [cited 2011 Nov 11] Available from: http://www.cee.vt.edu/ewr/environmental/ teach/smprimer/icpms/icpms.htm.

Page 19: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

17

Influences on water chemistry in Lee Vining Creek, California Emily J. Auerbach, Mauricio A. Barreto, Sarah E. Graves, and Sloane K. Miller Department of Geological Sciences, University of North Carolina, Chapel Hill, North Carolina 27599-3315, USA

ABSTRACT In this study we examined the effects of human ac-

tivity, elevation, and changing geology upon water chemistry. We collected water samples at 15 sites along Lee Vining Canyon, California, beginning at Upper Conness Lake (elevation: 3258 m) and ending just be-fore Mono Lake (elevation: 1965 m). We chemically analyzed these samples using an inductively coupled plasma mass spectroscopy machine. These data indicate that temperature changes caused by differences in ele-vation, increases in stream flow, and human activities have a more significant impact upon water chemistry than varying bedrock composition. INTRODUCTION

Rivers and streams provide habitats, build ecosystems, generate power, furnish irrigation, and most importantly, are a vital source of drinking water for humans and other organisms. Rivers and streams are fed from two main sources: surface runoff and groundwater. When the chemis-try of either of these sources is disrupted, both local ecosys-tems and human populations are affected. This paper traces downstream changes in water chemistry in Lee Vining Creek over a 28 km stretch (Fig. 1), from its glacial head-waters to a lake delta, in order to evaluate whether chang-ing bedrock in the downstream direction or various human and meteorological influences have a greater impact on water chemistry and composition.

Fig. 1. Geologic map of Lee Vining Creek traced in red showing bedrock composition by color with sampling sites marked by pins. Geologic base maps from Bateman et al. (1983) and Kistler (1966). See Appendix for key to maps.

METHODS

Preparation We chose to sample a drainage from the source (Upper

Conness Lake) to its lowest point (Mono Lake) to evaluate changes in water chemistry along the drainage, using data from 15 different sampling sites. We selected each site to evaluate a specific source of chemical change, such as hu-man, environmental, and geologic impact (Fig. 2).

To prepare our 250 ml sample bottles for water collec-tion, we cleaned them by soaking each bottle in 2% nitric acid for 48 hours. We then rinsed each bottle in doubly deionized water twice and labeled them with numbers cor-responding to the sites we planned to sample. This prepara-tion took place in the University of North Carolina at Chapel Hill geochemistry labs.

Fig. 2. Possible geologic and human influences on water chemistry with position and elevation along Lee Vining Creek.

Fieldwork We navigated each site by foot carrying maps, a GPS, a

camera, field notebooks, and our 250 ml sample bottles contained in a large zip-lock bag. At each site, we recorded the GPS location in UTM coordinates (WGS 84 datum), recorded general observations, and took two photographs for reference. We used the previously labeled bottles to collect water about 0.5 meters away from the shoreline (there was no exact depth). For sampling we tilted the plas-tic bottle to about 45° and lowered it until it was complete-ly submerged about 3 cm below the surface. Once halfway full we lifted the bottle out of the water and used plastic tape to seal on the lid and avoid contamination. We repeat-ed this fieldwork method for all 15 sites (Fig. 3).

Fig. 3. A typical sampling site--Lower Saddlebag Lake.

Page 20: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

18

CHEMICAL ANALYSIS

Sample Preparation To prepare the samples for analysis, we pipetted 5 ml

of each sample into 15 ml conical screw cap tubes and spiked each with 250 𝜇l of approximately 80% concentrat-ed nitric acid for a 1:20 dilution of acid to sample. The addition of the nitric acid stabilized the signal of the argon plasma and provided a matching matrix for all of the sam-ples (S. Habibi, personal communication, 2011). We then used a vortex machine, which vibrates the tube at a high frequency to mix the samples.

Calibration In order to calibrate the ICP-MS, we ran 8 standards

that contained all elements of interest in concentrations of 25 parts per billion (ppb), 50 ppb, 100 ppb, 200 ppb, 400 ppb, 600 ppb, 800 ppb and 1000 ppb. We chose our range of standard concentrations based on water-quality data collected from the Tuolumne Meadows and Mono Craters quadrangles on the United States Geological Survey web-site (USGS). By running standards with known concentra-tions, we determined how precisely the machine was meas-uring the concentration of each element. We examined the curves and discarded elements that were measured inaccu-rately. We adjusted other elements’ calibration curves by removing the standards that had the greatest deviation so that the coefficient of determination (r2 value) of the line of best fit was close to 1 (S. Habibi, personal communication, 2011). The elements measured were Na, Mg, Al, Si, P, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Sr, Y, Zr, Nb, Mo, Ru, Rh, Pd, Ag, Sn, Pt, Au, Hg, and Pb. Experimentation

First, we ran 2% nitric acid through the lines to clean the machine and avoid contamination. Then we manually fed the line into each sample and ran the program for all the elements we had specified. To analyze the chemical com-position of a sample, the ICP-MS mixed the liquid sample with Ar gas, which was aspirated into a radio frequency generator. This created an Ar plasma “flame” at approxi-mately 8000 K that broke the solute down into its atoms and ionized it. The machine separated these ions by their mass to charge ratio and used the previously run standards to determine each ion’s concentration in the sample (Wor-ley and Kveck, 2011).

Process of Data Editing The ICP-MS yielded data for all 35 elements we includ-

ed in our program. However, many elements were either not present at all or present in such low concentrations that they were below detection limits. We discarded all of these elements and analyzed the data for Na, K, Mg, Ca, Sr, Fe, Mo, Mn, Cr, Cu, Au, V, As, Si, and P.

RESULTS

Trends The alkali and alkaline earth metals showed consistent

increases in concentration as distance from the source in-creased. The transition metals followed this same trend to some degree, but the concentrations detected were low enough to make these results unreliable. The metalloids generally exhibited this trend, and non-metals showed a weak negative correlation between elevation and concentra-

tion. For example, as elevation decreased, Na increased by a factor of 10, Mg increased by a factor of 8, V increased by a factor of 4, and Si increased by a factor of 6. Silicon, an abundant component of granodiorite such as that found in the Conness Lakes region (Bateman and Chappell, 1979), also increases as elevation decreases (Fig. 7).

Fig. 4. Correlation between Mg and Sr. This demonstrates that ele-ments from the same group correlate highly and that the rate at which the concentrations increase as elevation decreases is similar.

Anomalies One unexpected result in our data was the marked

presence of As in Shell Lake Tributary, where concentra-tion spiked from approximately 1 ppb at all other sites to 4.2 ppb (Fig. 5).

Our results also had uncharacteristically low levels of Fe. Although the levels followed a consistent positive trend, our range was from 0 ppb to 43 ppb (Fig. 6). Normal Fe levels for rivers are 0.5-1 parts per million, approxi-mately 20 times that of the levels we found (Iron).

We found Si in lower levels than the 4000 ppb ex-pected in rivers in all of our samples except in the Lower Canyon sample (Silicon). The last four sites on the traverse had significantly higher levels of Si, ranging from 2500 ppb to 4300 ppb (Fig. 7).

Fig. 5. Shell Lake tributary, which is downstream from an abandoned silver mine (Whitehill et. al, 2003), has a much higher As concentra-tion than the other sampling sites. The legal limit of As in drinking water is 10 ppb (Arsenic).

Page 21: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

19

Fig. 6. Although Fe follows the general trend, it is at a level far below the average amount for rivers around that area.

Fig. 7. Si is also found in very low concentrations, and follows the general trend of increasing elemental concentration as elevation decreases.

DISCUSSION

Trends There was a 1300 m elevation change between Upper

Conness Lake and Mono Lake. According to data collected by the Western Regional Climate Center, the mean annual air temperatures at Ellery Lake, one of our highest sample sites, and at Mono Lake, the end of our drainage, vary by 5° C (Ellery Lake). This negative correlation can be at-tributed to the cooling of air as it rises.

This is significant because altitude and the resulting changes in temperature have a significant impact on water chemistry. Drever and Zobrist (1991) found that the con-centrations of major cations and silica in surface waters of the Alps decrease exponentially with altitude. This trend was similar to the data we found both with the alkaline and alkaline earth metals, including Ca, Mg, K, and Si. This occurs because rates of chemical weathering increase as temperature increases, meaning that minerals and the ele-ments they contain (specifically, Si and cations) are more likely to be eroded at the higher temperatures which occur at lower elevations.

Another possible cause of this positive trend is an in-crease in the volume of stream discharge. As the drainage progresses from Greenstone Lake to Ellery Lake, the amount of water discharge increases from an annual mean of 2.59 m3 per second to 3.26 m3 per second (USGS Wa-ter). While rates of bank erosion are positively related to temperature, these rates are also correlated with the amount of stream discharge (Shrestha and Tamraker, 2007). As the volume and velocity of the water in Lee Vining Creek in-

creases, the amount of dissolved minerals increase, result-ing in the positive trend seen throughout our data.

These three factors (stream discharge, elevation, and temperature) also explain the trend for Si. The sites with higher elevations contained more silicate-rich minerals than sites with lower elevations, yet this was not mirrored in the positive trend for Si concentrations (Fig. 1 and Appendix). This demonstrates that bedrock was less influential in water chemistry than increased erosion rates.

Anomalies The area surrounding Shell Lake used to contain a sil-

ver mine (Whitehill et. al, 2003), and as a 2007 study demonstrated, both gold and silver mines are significant sources of As contamination in waterways (Rytuba et al., 2007). Therefore, it is clear that in the case of As, human activity had a significant impact upon the chemical compo-sition of this waterway. It is, however, worth noting that once the tributary joins with the Tioga Pass tributary and the main waterway As levels drop back down to 1 ppb, showing that this anomaly does not have a large effect on the body of water as a whole.

Weathering of magnetite, hematite, goethite and sider-ite contribute to normal Fe levels in rivers. Our uncharac-teristically low levels of Fe can be explained by a relative lack of any of these minerals in the bedrock composition of our water flow (granite only contains low levels of magnet-ite). CONCLUSIONS

The data in this study reveal that as distance down-stream increases, the concentrations of various elements increase significantly. Elevation, temperature, and stream discharge all contribute to this trend because each of these factors strongly correlates to chemical weathering rates, meaning that as the drainage progresses, more minerals are dissolved into Lee Vining Creek, increasing elemental con-centrations. In addition to this trend, we discovered a hu-man-caused anomaly in the As concentrations and an anomaly due to changing bedrock in the Fe concentrations. Lastly, we saw that bedrock composition has very little influence on water chemistry, as the Si trend demonstrates. Our research supports the idea that while meteorological, geologic, and human processes all affect water chemistry, chemical weathering due to increased temperatures at lower elevations is the most significant factor in determining the concentration of a given element.

ACKNOWLEDGEMENTS This research was supported by the Department of Geological

Sciences at the University of North Carolina at Chapel Hill, the Anadarko Petroleum Corporation, the Howard Hughes Medical Institute, the First-Year Seminar Program, and the Office of Un-dergraduate Research Graduate Research Consultant Program. We would like to thank Adam Curry, Roger Putnam, and Siobhan Kenney for helping us coordinate field days and brainstorm geo-logical processes, Dr. Sohrab Habibi for training us in the ICP-MS and for giving his time helping us throughout the experimentation process, Dr. Drew Coleman for his sunny disposition, and Dr. Allen Glazner for leading the trip and guiding our analyses, providing additional resources and references, and driving us around all day.

Page 22: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

20

REFERENCES CITED Arsenic in Drinking Water: United States Environmental Protec-

tion Agency: http://water.epa.gov/lawsregs/rulesregs/sdwa/arsenic/index.cfm (November 2011).

Bateman, P.C., and Chappell, B.W., 1979, Crystallization, frac-tionation, and solidification of the Tuolumne Intrusive Series, Yosemite National Park: California Geological Society of America Bulletin, v. 90, p. 465-482.

Bateman, P.C., and Kistler, R.W., and Peck, D.L., and Busacca, A.J., 1983, Geologic map of the Tuolumne Meadows Quad-rangle, Yosemite National Park, California: Geologic quad-rangle map.

Drever, J.I., and Zobrist, J., 1991, Chemical weathering of silicate rocks as a function of elevation in the southern Swiss alps: Geochimica et Cosmochimica Acta, v. 56, p. 3209-3216

Ellery Lake, California: Period of record monthly climate sum-mary: Western Region Climate Center: http://www.wrcc.dri.edu/cgi-bin/cliMAIN.pl?ca2756 (No-vember 2011).

Iron: Lenntech water treatment solutions: http://www.lenntech.com/periodic/elements/fe.htm (Novem-ber 2011).

Kistler, R.W., 1966, Geologic map of the Mono Crater Quadran-gle, Mono and Tuolumne counties, California: Geologic quad-rangle map U.S. Geological Survey Report.

Rytuba, J. J., Foster, A., Kim, C. S., Slowey, A., Lawler, D., and Forester, R. (2007). Arsenic contamination from the Kelly sil-ver and Yellow Aster gold mine tailings, California: a poten-tial health concern in the north-central Mojave Desert. Paper presented at the , 39(6) 31-31. Retrieved from http://search.proquest.com/docview/50641934?accountid=14244.

Shrestha, P., and Tamraker, N.K., 2007, Bank erosion process and bank material loss potential in Manahara: Bulletin of the de-partment of Geology Tribhuban University, v.10, p.33-44.

Silicon: Lenntech water treatment solutions: http://www.lenntech.com/periodic/elements/si.htm (November 2011).

USGS Water Data for the Nation: http://wdr.water.usgs.gov/nwisgmap (November 2011).

USGS 11274730 Budd C NR Tuolumne meadows CA: United States Geological Survey: http://waterdata.usgs.gov/nwis/inventory?agency_code=USGS&site_no=11274730 (November 2011).

Whitehill, K., Shepherd, S., Whitehill, T., and Wozniak, O., 2003, Best Short Hikes in California’s North Sierra: The Mountain-eers Books, 97 p. Retrieved from http://books.google.com.

Worley, J., and Kvech, S., ICP-MS: The Charles Edward Via, Jr. Department of Civil and Environmental Engineer-ing:http://www.cee.vt.edu/ewr/environmental/teach/smprimer/icpms/icpms.htm (November 2011).

APPENDIX Lithologic key corresponding to Figure 1, from Bateman et al. (1983) and Kistler (1966).

Page 23: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 21

Density of Bishop Tuff in Owens River Gorge, California Grace A. Blair, Carlos S. Floyd, Olivia L. Karas, and Kai T. Shin Department of Geological Sciences, University of North Carolina, Chapel Hill, North Carolina 27599-3315, USA ABSTRACT

Density measurements from the Bishop Tuff near Long Valley, California, have been determined by Jolly balance and weight and volume measurements. The density shows a maximum that corresponds to a cliff of densely welded tuff and a minimum at the canyon rim about 80 m above the densely welded zone. Our data are consistent with the Bishop Tuff in Owens River Gorge being a single cooling unit, but the presence of a second, smaller cliff approximately 65 m below the main cliff indicates that a second cooling unit may be present. Our data were not precise enough to detect this second cliff.

INTRODUCTION The Bishop Tuff was erupted 760,000 years ago during

collapse of the Long Valley caldera (Hildreth and Wilson, 2007). Approximately 350 km3 of Bishop Tuff settled in the caldera and 150 km3 in the surrounding area (Bailey et al., 1976). A stratigraphic section of the tuff ranging from the base through the densely welded zone and up into poor-ly welded material (sillar) is exposed in Owens River Gorge in Mono County, California. Sheridan (1968) suggested that the tuff is a compound cool-ing unit, whereas other studies have suggested that it is a simple cooling unit (Ragan and Sheridan, 1972; Snow and Yund, 1985; Wilson and Hildreth, 2003; Sheridan and Wang, 2005). These later studies suggested that in the Bishop Tuff the densest material is found in the lower cen-tral part of the flow, and that density decreases both upward and downward from the central portion. There will theoret-ically be three zones formed: a densely welded zone (DWZ), a partial welding zone, and zone of no welding (Ragan and Sheridan, 1972). Ragan and Sheridan’s (1972)

data are shown in Figure 2. Based on these studies, we hypothesized that three sections of varying density would be present with the lowest densities occurring at the surface and at the base of the tuff, at or near its contact with the underlying rocks, which are Triassic granite (Bateman, 1992).

METHODS We obtained samples of the Bishop Tuff at 20 locations

along a paved road that goes from Gorge Road to a hydroe-lectric powerhouse at the bottom of the gorge. At each sampling location, 2 pieces of tuff were collected from

Figure 1. An aerial Google Earth photograph of our sampling location within the Owens River Gorge and an inset showing the location of the Long Valley area in California. The densely welded zone (DWZ) is highlighted in red and our sampling began at the granite contact point, indicated in blue. Our sampling locations are marked with the purple symbol.

Figure 2. A scatterplot showing the results of Ragan and Sheridan’s 1972 study, where the x-axis represents the density of their samples and the y-axis represents the sample’s elevation within the Owens River Gorge. The maximum density of 2.44 g/cm3 occurs at an eleva-tion of 59.7 meters with the density decreasing sharply both above and below this point. This trend illustrates the basis for our hypothe-sis that peak density within the tuff would not occur at the base but rather at some middle level in the form of a DWZ.

Page 24: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 22

outcrops with a rock hammer and GPS coordinates (WGS 84 datum) were recorded for each location.

Densities of non-porous samples weighing 10-30 g were measured using a Jolly Balance (Hutchinson, 1974, p. 236-237), using distilled water as the reference fluid. The Jolly Balance uses Archimedes’ principle to measure spe-cific gravities of samples, which is their density relative to the fluid in which they are submerged. Because the density of water at room temperature is 1 g/cm3 the difference be-tween density and specific gravity is negligible compared to the precision of our measurements and we will refer to these measurements as densities for brevity. Our Jolly Bal-ance measurements have a standard deviation of ±0.005 g/cm3, found through multiple measurements of the same sample.

This process is not conducive to finding the density of porous samples, so we cut such samples into rectangular prisms of measurable volume to calculate their densities after weighing.

Since we had two samples from each location, we cre-ated two data sets to increase the confidence in our density measurements. Where the difference between the data sets was not greater than 0.03 g/cm3 we averaged the measure-ments. When the difference was greater, we re-measured.

Tilt Correction In order to quantify this correlation between elevation

and density, we had to correct for post-eruption tectonic tilt of the area, which left the Bishop Tuff sloping downward to the south (Putnam, 1960). This incline rendered the pre-sent-day elevations of our sample points misleading as each value simply gave each sample’s elevation above sea level regardless of its position within the layering of the pyro-clastic flow.

We corrected for this variable by measuring each sam-ple’s vertical distance from the densely welded zone. Using a Google Earth mosaic of the Owens River Gorge area, we measured the distance between our sample locations and a point approximately in the center of the DWZ and then calculated the difference in elevation of those two points. We assumed that the tilt of the surface that the tuff was deposited on would be reflected by the DWZ. We also presumed that tectonic activity would have shifted this zone in the same way as the rest of the surrounding landscape. Thus, by finding the vertical distance between our samples

and the DWZ, we were effectively finding the original, eruption-era elevation of our samples. The elevations of Figure 4 are therefore in reference to a baseline elevation of the DWZ.

RESULTS Figure 4 shows the relationship between the density of

the Owens Gorge samples and their depth within the Bish-op Tuff. Measured densities range from 1.31 g/cm3 to 2.41 g/cm3, with the least-dense samples occurring at the canyon rim. The samples farthest below the DWZ are of slightly lower density than those within this zone. Our sample clos-est to the base, at 117 meters below the DWZ, has a density of 2.31 g/cm3, whereas a mid-DWZ sample has a density of 2.41 g/cm3. Our lowest density samples were those farthest above the DWZ, with the lowest being 1.30 g/cm3 at 79 m above the DWZ. These data support the results of Ragan and Sheridan (1972), which show a trend similar to ours. Density of the tuff increases as the samples approach the DWZ (about 82 meters up through the tuff, according to Sheridan and Ragan) and are lowest on the surface, at the very top of the tuff. This pattern is indicative of a single cooling unit with three density partitions.

DISCUSSION Although the trend that we discovered generally corre-

sponds with that of Ragan and Sheridan (1972), it is im-portant to note the significant differences. Figure 5 shows the difference in the decrease in density below the DWZ. Ragan and Sheridan’s data decreases at a much more rapid rate, reaching a density measurement nearly as low as that at the top of the tuff. The discrepancy in densities is ex-tremely odd. Ragan and Sheridan's 1.17 g/cm3 samples were supposedly found at the very bottom of the tuff, but this measurement is very similar to the top samples, which, on observation, were extremely porous and chalky. It is odd that the tuff could retain this density under the compression of the rest of the cliff.

The significant discrepancy in rate of density change is also unusual, as we sampled from the very bottom of the tuff (at the granite contact). Ragan and Sheridan claimed to

Figure 3. Sampling location 8 along Gorge Road in Owens River Gorge. Samples of tuff were collected with a rock hammer and GPS coordinates were recorded.

Figure 4. A scatterplot showing the results of our data measurements. The x-axis represents density of each sample while the y-axis shows its elevation relative to the DWZ. Our graph confirms that peak density occurs at some central level and decreases both above and below this peak. The peak density is 2.41 g/cm3, closely mirroring the findings of Ragan and Sheridan (1972).

Page 25: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 23

have reached the bottom of the tuff as well. In addition, their sample locations are poorly defined and do not corre-spond to the Owens River Gorge. It is possible, though unlikely, that the surveying on Ragan and Sheridan's older maps was different and therefore accounts for the unusual data. Despite these differences, both Ragan and Sheridan’s and our data provide evidence of a single cooling unit and can be used as a model for determining the cooling unit nature of other pyroclastic flows.

It should also be noted that a second, smaller cliff ex-ists approximately 65 meters below the main cliff, indicat-ing a second cooling unit. Our data, however, was not pre-cise enough to detect this second cliff.

CONCLUSIONS Our conclusions are consistent with studies completed

by Ragan and Sheridan (1972). From this density profile it is evident that there are three distinct welding zones. The maximum degree of compaction of this ash-flow tuff oc-curs in the lower central region of the tuff exposed in the gorge. This confirmed our hypothesis about the single cool-ing unit nature of the tuff in Owens River Gorge.

ACKNOWLEDGMENTS This research was supported by the Department of Geological Sciences at the University of North Carolina at Chapel Hill, the Anadarko Petroleum Corporation, the Howard Hughes Medical Institute, the First-Year Seminar Program, and the Office of Un-dergraduate Research Graduate Research Consultant Program. This study would not have been possible without the guidance of Dr. Allen Glazner and Dr. Drew Coleman. We would like to espe-cially thank Adam Curry, Roger Putnam and Siobhan Kenney for their aid in our data collection and analysis (and for exposing us to Kanye West and Jay-Z). Thanks to the entire 72H class of 2011 for making the trip unforgettable.

Figure 5. This scatterplot shows a compilation of the results of our data measurements and Ragan and Sheridan’s 1972 measure-ments. The densest tuff was found at a similar elevation; however, our findings did not show the same sharp decrease in density below the DWZ depicted in Ragan and Sheridan’s results. Addi-tionally, our measurement of the total depth of the Owens River Gorge is considerably larger than theirs.

Figure 6. A view of the cliffs from the side of the river opposite our sampling locations. The double cliff landscape seems to indicate two welded zones, for which we did not find evidence, but Sheridan (1968) did.

Page 26: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines

GEOLOGY 72H, December 2011 24

REFERENCES CITED Bailey, R.A., Dalrymple, W.A., and Lanphere, M. A., 1976, Vol-

canism, structure, and geochronology of Long Valley caldera, Mono County, California: Journal of Geophysical Research, v. 81, p. 725-744.

Bateman, P. C., 1992, Plutonism in the central part of the Sierra Nevada batholith, California: U.S. Geological Survey Profes-sional Paper 1483, 186 p.

Hildreth, W., and Wilson, C. J. N., 2007, Compositional zoning of the Bishop Tuff: Journal of Petrology, v. 48, p. 951-999.

Hutchinson, C. S., 1974, Laboratory Handbook of Petrographic Techniques, 527 p.

Putnam, W., 1960, Origin of Rock Creek and Owens River Gorg-es, Mono County, California: University of California Publica-tions in Geological Sciences, v. 34, p. 221-279.

Ragan, D. M., and Sheridan, M. F., 1972, Compaction of the Bish-op Tuff, California: Geological Society of America Bulletin, v. 83, p. 95-106.

Sheridan, M. F., 1968, Double cooling-unit nature of the Bishop Tuff in Owens Gorge, California [abs.]: Geological Society of America Special Paper 115, p. 351.

Sheridan, M. F., and Wang, Y., 2005, Cooling and welding history of the Bishop Tuff in Adobe Valley and Chidago Canyon, Cal-ifornia: Journal of Volcanology and Geothermal Research, v. 142, p. 119-144.

Snow, E., and Yund, R.A., 1985, Thermal history of a Bishop Tuff section as determined from the width of cryptoperthite lamel-lae: Geology, v. 13, p. 50-53.

Wilson, C. J. N., and Hildreth, W., 2003, Assembling an ignim-brite; mechanical and thermal building blocks in the Bishop Tuff, California Journal of Geology, v. 111, p. 653-6.

Page 27: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines
Page 28: GEOLOGY 72H - University of North Carolina at Chapel HillGEOLOGY 72H, December 2011 1 Determining the correlation between moraine age and steep-ness of slope in Quaternary moraines