Saline Lakes

Loading...
Saline Lakes

Developments in Hydrobiology 162

Series editor H. J. Dumont

Saline Lakes Publications from the 7th International Conference on Salt Lakes, held in Death Valley National Park, California, U.S.A., September 1999

Edited by

John M. Melack1 , Robert Jellison 2 & David B. Herbse 1 Department of Ecology, Evolution and Marine Biology, and Bren School of Environmental Science and Management, University of California, Santa Barbara, U.S.A. 2 Marine Science Institute, University of California, Santa Barbara, U.S.A. 3 Sierra Nevada Aquatic Research Laboratory, University of California, Mammoth Lakes, U.S.A.

Reprinted from Hydrobiologia, volume 466 (2001)

Springer-Science+Susiness Media, SV.

Library of Congress Cataloging-in-Publication Data

A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN 978-90-481-5995-6 DOI 10.1007/978-94-017-2934-5

ISBN 978-94-017-2934-5 (eBook)

Printed on acid-free paper

All Rights reserved

© 2001 Springer Science+Business Media Dordrecht Originally published by Kluwer Academic Publishers in 2001 Softcover reprint of the hardcover 1st edition 2001 No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without written permission from the copyright owner.

v

TABLE OF CONTENTS

Preface

ix

Nitrogen limitation and particulate elemental ratios of seston in hypersaline Mono Lake, California, U.S.A. Robert Jellison, John M. Melack

1-12

Nutrient fluxes from upwelling and enhanced turbulence at the top of the pycnocline in Mono Lake, California Sally Macintyre, Robert Jellison

13-29

Airborne remote sensing of chlorophyll distributions in Mono Lake, California John M. Melack, Mary Gastil

31-38

Re-appearance of rotifers in hypersaline Mono Lake, California, during a period of rising lake levels and decreasing salinity Robert Jellison, Heather Adams, John M. Melack

39-43

Stratification of microbial assemblages in Mono Lake, California, and response to a mixing event James T. Hollibaugh, Patricia S. Wong, Nasreen Bano, Sunny K. Pak, Ellen M. Prager, Cristian Orrego

45-60

The bioenergetic basis for the decrease in metabolic diversity at increasing salt concentrations: implications for the functioning of salt lake ecosystems Aharon Oren

61-72

Comparative metabolic diversity in two solar salterns Carol D. Litchfield, Amy Irby, Tamar Kis-Papo, Aharon Oren

73-80

Polar lipids and pigments as biomarkers for the study of the microbial community structure of solar salterns Carol D. Litchfield, Aharon Oren

81-89

Limnological effects of anthropogenic desiccation of a large, saline lake, Walker Lake, Nevada Marc W. Beutel, Alex J. Horne, James C. Roth, Nicola J. Barratt

91-105

VI

Oxygen consumption and ammonia accumulation in the hypolimnion of Walker Lake, Nevada Marc W. Beutel

107-117

Limnological control of brine shrimp population dynamics and cyst production in the Great Salt Lake, Utah Wayne A. Wurtsbaugh, Z. Maciej Gliwicz

119-132

International study on Artemia LXIII. Field study of the Artemia urmiana (Gunther, 1890) population in Lake Urmiah, Iran Gilbert Van Stappen, Gholamreza Fayazi, Patrick Sorgeloos

133-143

Dispersal of Artemia franciscana Kellogg (Crustacea; Anostraca) populations in the coastal saltworks of Rio Grande do Norte, northeastern Brazil Marcos R. Camara

145-148

Anostracan cysts found in California salt lakes William D. Shepard, Richard E. Hill

149-158

Thermal, mixing, and oxygen regimes of the Salton Sea, California, 1997-1999 James M. Watts, Brandon K. Swan, Mary Ann Tiffany, Stuart H. Hurlbert

159-176

Pleurochrysis pseudoroscoffensis (Prymnesiophyceae) blooms on the surface of the Salton Sea, California Kristen M. Reifel, Michael P. McCoy, Mary Ann Tiffany, Tonie E. Rocke, Charles C. Trees, Steven B. Barlow, D. John Faulkner, Stuart H. Hurlbert

177-185

Chattonella marina (Raphidophyceae), a potentially toxic alga in the Salton Sea, California Mary A. Tiffany, Steven B. Barlow, Victoria E. Matey, Stuart H. Hurlbert

187-194

Parasites of fish from the Salton Sea, California, U.S.A. Boris I. Kuperman, Victoria E. Matey, Stuart H. Hurlbert

195-208

Gradients of salinity stress, environmental stability and water chemistry as a templet for defining habitat types and physiological strategies in inland salt waters David B. Herbst

209-219

Thermal tolerance and heat shock proteins in encysted embryos of Artemia from widely different thermal habitats James S. Clegg, Nguyen Van Hoa, Patrick Sorgeloos

221-229

Land-use influence on stream water quality and diatom communities in Victoria, Australia: a response to secondary salinization Dean W. Blinn, Paul C.E. Bailey

231-244

A study of the Werewilka Inlet of the saline Lake Wyara, Australia - a harbour of biodiversity for a sea of simplicity Brian V. Timms

245-254

Demography and habitat use of the Badwater snail (Assiminea infima), with observations on its conservation status, Death Valley National Park, California, U.S.A. Donald W. Sada

255-265

Vll

Holocene hydrological and climatic changes in the southern Bolivian Altiplano according to diatom assemblages in paleowetlands S. Servant-Vildary, M. Servant, O. Jimenez

267-277

Reconnaissance hydrogeochemistry of economic deposits of sodium sulfate (mirabilite) in saline lakes, Saskatchewan, Canada Lynn I. Kelley, Chris Holmden

279-289

Benthos of a seasonally-astatic, saline, soda lake in Mexico Javier Alcocer, Elva G. Escobar, Alfonso Lugo, L. Maritza Lozano, Luis A. Oseguera

291-297

Phytoplankton dynamics in a deep, tropical, hyposaline lake Ma. Guadalupe Oliva, Alfonso Lugo, Javier Alcocer, Laura Peralta, Ma. del Rosario Sanchez

299-306

Food-web structure in two shallow salt lakes in Los Monegros (NE Spain): energetic vs dynamic constraints Paloma Alcorlo, Angel Baltanas, Carlos Montes

307-316

Avian communities in baylands and artificial salt evaporation ponds of the San Francisco Bay estuary John Y. Takekawa, Corinna T. Lu, Ruth T. Pratt

317-328

Anthropogenic salinisation of inland waters William D. Williams

329-337

On salinology Zheng Mianping

339-347

.... Hydrobiologia 466: ix, 2001. 1.M. Melack, R. lellison & D.B. Herbst (eds), Saline Lakes.

ft

IX

Preface The Seventh International Conference on Salt Lakes was held in Death Valley National Park, California, U.S.A. in September 1999. The conference was sponsored by the International Society for Salt Lake Research, Societas Internationalis Limnologiae, and University of California-Santa Barbara. Since 1979 a series of international symposia on inland saline waters have served to strengthen and expand the scope of limnological research on salt lakes. The seventh conference continued this tradition with a set of plenary talks and oral and poster sessions focusing on promising research directions, including the ecology of microbial communities, the influence of habitat geochemistry on biogeography of flora and fauna, physical and geochemical processes, and the conservation of inland saline waters. Sixty participants from eleven countries participated. The venue of the conference in Death Valley encouraged informal interactions in a striking landscape rich in saline environments. A 5-day, post-conference tour visited a wide variety of saline ecosystems located on the western edge of the North American Great Basin, a region noted for its remarkable ecological diversity and striking beauty. Major stops included Owens, Mono, Walker, and Pyramid lakes. Inland saline waters are threatened worldwide by diversion and pollution of their inflows, introductions of exotic species and economic development of these ecologically valuable habitats. Several sessions at the conference concerned anthropogenic impacts and conservation with special attention paid to Walker Lake, Nevada (U.S.A.), the Salton Sea, Mono and Owens lakes and Death Valley, California (U.S.A.), Siberian salt lakes and salinization in Australia. Continued local, national and international efforts are required to inform the public and decision-makers about the environmental problems faced by saline waters. All manuscripts were critically refereed by well qualified experts, revised by the authors and edited before acceptance. We gratefully thank Death Valley National Park for hosting the conference, Doug Threloff for his assistance with the local arrangements, and the manuscript reviewers for their care and rigor. JOHN M. MELACK ROBERT JELLISON DAVID B. HERBST

Hydrobiologia 466: 1-12,2001. J.M. Melack, R. Jellison & D.E. Herbst (eds), Saline Lakes. © 200 I Kluwer Academic Publishers.

Nitrogen limitation and particulate elemental ratios of seston in hypersaline Mono Lake, California, U.S.A. Robert Jellison 1* & John M. Melack 1,2 I Marine

Science Institute, University of California, Santa Barbara, CA 93106, U.S.A. 2Department of Ecology, Evolution and Marine Biology, University of California, Santa Barbara, CA 93106, U.S.A. Key words: Mono Lake, hypersaline, nutrient limitation, elemental ratios, meromixis

Abstract Particulate elemental ratios (C:N, N:P and C:Chl a) of seston in hypersaline (70-90 g kg-I) Mono Lake, California, were examined over an II-year period (J 990-2000) which included the onset and persistence of a 5-year period of persistent chemical stratification. Following the onset of meromixis in mid-1995, phytoplankton and dissolved inorganic nitrogen were substantially reduced with the absence of a winter period of holomixis. C:N, N:P and C:Chl a ratios ranged from 5 to 18 mol mol-I, 2 to 19 mol mol- 1 and 25 to 150 g g-I, respectively, and had regular seasonal patterns. Deviations from those expected of nutrient-replete phytoplankton indicated strong nutrient limitation in the summer and roughly balanced growth during the winter prior to the onset of meromixis. Following the onset of meromixis, winter ratios were also indicative of modest nutrient limitation. A 3-year trend in C:N and N:P ratios toward more balanced growth beginning in 1998 suggest the impacts of meromixis weakened due to increased upward fluxes of ammonium associated with weakening stratification and entrainment of ammonium-rich monimolimnetic water. A series of nutrient enrichment experiments with natural assemblages of Mono Lake phytoplankton conducted during the onset of a previous episode of meromixis (19821986) confirm the nitrogen will limit phytoplankton before phosphorus or other micronutrients. Particulate ratios of a summer natural assemblage of phytoplankton collected under nitrogen-depleted conditions measured initially, following enrichment, and then after return to a nitrogen-depleted condition followed those expected based on Redfield ratios and laboratory studies.

Introduction Saline lakes are widely recognized as highly productive aquatic habitats, harboring specialized assemblages of species and often supporting large populations of both migrating and breeding birds. Many saline lake ecosystems throughout the world are threatened by decreasing size and increasing salinity due to diversions of freshwater inflows for irrigation and other human uses (Williams, 1993). Because saline lakes primarily occur in endorheic basins, they may be particularly sensitive to global climate change as their size, salinity and annual mixing regimes vary with alterations in their hydrologic budgets (Romero & Melack, 1996; Jellison et al., \998). Determining

* Corresponding author. E-mail: [email protected]

the temporal variation and degree of nutrient limitation is critical to understanding these ecologically valuable aquatic environments. Mono Lake lies in a hydrologically closed, highdesert basin just east of the Sierra Nevada. External inputs of nitrogen including nitrogen fixation (Herbst, 1998; Oremland, 1990) are low and dissolved inorganic N:P ratios very low (

.r=

~ 0.5

u o

U")

11

12

13

14

October

11

12

October

13

14

Figure 6. (A.C,E) 14 °C, 7 °C and 5 °C isotherms filtered to show contributions from different frequencies: 0.07-2.4 cpd (blue), 2.4-19.2 cpd (green), 19.2-192 cpd (red), and 192-600 cpd (turquoise). First frequency range includes basin scale waves; second contains harmonics of these waves, the first vertical first horizontal, and includes waves at critical frequencies for breaking; third range also is within range of critical frequencies; and fourth range is primarily above critical frequencies. (B, D, F). Potential energy per unit mass for the three isotherms in these four frequency bands showing decay over time. Window sizes are 24, 12, 12 and 12 h. Potential energy for the frequency band 19.2-192 cpd has been multiplied by 10; that in the band 192-600 cpd has been multiplied by 100.

Time series plots show that the ammonium maximum was broader and that concentrations were lower within it inshore than offshore (Fig. 8A,B). In particular, the maximum inshore spanned 4 m on the afternoon of 11 October. At both stations, concentrations were 0.3 p,m or less in surface waters. Towards the end of the study, patches of water with higher NH4 concentration appeared at the inshore station. The chlorophyll minimum coincided with the NH4 maximum at both inshore and offshore stations (Fig.

9). Inshore, it tended to be located at the upper boundary of the NH4 maximum; offshore it tracked it more closely. Over the 4 days of measurements, the chlorophyll a values above 8 m depth inshore increased from 2 p,g 1-1 to 4 p,g I- I in the upper 8 m; offshore they increased from 3 p,g I-I to 4.5 p,g 1-1. Similar patterns were observed at station 8, near the far north-eastern shore. A maximum in NH4 also occurred at 10-11 m depth with lower concentrations below. N~ values in the upper mixed layer increased

11 October

1540h

1511h

1151h

1342h

Off

On

Off

II

1340h

• •

1157hl l

12 October

1617h

1414h

On

Off

I 13 October

;; iiiiii

1145h - 1331h 1418h and 1535h - 1556h 1427h I

On

-

10 -6

10 -8

I

10- 10

10 -9

. 10 -7 t-



1

Dissipation (m 2 s-3 )

Figure 7. Rates of dissipation of turbulent kinetic energy at inshore and offshore stations on 11, 12 and 13 October ]995. Dissipations were highest in the upper 10 m 011 II October due to wind. On 12 October, the pycnocline has been energized at the inshore station, but only at the top of the pycnocline offshore, On 13 October, high turbulence is still present at the top of the pycnocline and at its base inshore; turbulence is less offshore in the pycnocline. Only segments with values of E greater than 10- 7 m 2 s-3 are sufficiently energized to support turbulent transport. Black line indicates ZTM. Last two profiles on 12 October were downcasls as were al1 on 13 October. Only two of the offshore casts extended to the bottom of the lake. Navy - not turbulent.

35 IL~~~~L-~~~~~~--~~~~~~~~~~~~--~~~

30

25

o 20

Q)

a.

-

---.c:

E

5

o

N

22 A

11

12

13

B

14

11

October 1995

12

13

14 October 1995

October 1995

Figure 8. (A). N14 time series from II to 14 October at the inshore station showing the elevated concentrations near 10 m depth. an NH4 minimum at 14 m depth, and the nutricline below 15 m depth. Upwelling is indicated by letter A. (B). As in A but at the offshore site. Contour intervals are 0.4 /Lm, O.S /Lm, I /Lm from I to 5 /Lm, and 5 /Lm from 5 to 45 /Lm. Inshore, the upwelling of NH4 after the onset of high winds and the much greater dispersion of N14 both above and below the maximum at 10 m is apparent.

October 1995

Figure 10. (A). As in Figure SA but C:N molar ratios. (B). As in Figure SB but C:N. Ratio is elevated at ZTM; decrease in the upper mixed layer occurs over the four days. Greater decrease below 5 m and above ZTM indicates the upper mixed layer was not fully mixing. Contours are at 0.5 intervals from 4.5 to S with a step change to 10 offshore in panel B.

:§: ~ Q)

:§:

0

~ Q)

0

11 11

12

13

October 1995

14

11

12

13

14

October 1995

Figure 9. (A). As in Figure 8A but chlorophyll a concentrations inshore. The chlorophyll minimum co-occurs with the NH4 maximum; chlorophyll a concentrations in the upper mixed layer increased after the strong wind forcing of 11-12 October. The upwelling on 11 October is illustrated (letter A) as are the increases in chlorophyll a in the mixed layer over the 4 days. (B). As in Figure SB but chlorophyll a concentrations offshore. Contour intervals are 0.5 /Lg 1-1 from 0 to 6 /Lg 1- 1,2 /Lg 1-1 from 6 to 16 /L 1- 1,6 /Lg I-I from 16 to 2S /Lg 1-1, S /Lg I-I from 2S to 60 /Lg 1- ,and 4 /Lg 1-1 from 60 to 70 /Lg I-I.

r

from 0.1 p,m to 1 p,m from 0830 h II October to 0820 h 12 October. Similarly, chlorophyll a increased from 2.8 to 5 p,g I-I. C:N molar ratios at the inshore station were always close to Redfield and not indicative of nutrient limitation, but did show a slight decrease over time (Fig. lOA). C:N molar ratios in the upper mixed layer offshore exceeded 9 at some depths on the first day of

12

13

October 1995

14

11

12

13

14

October 1995

Figure 11. (A). As in Figure SA but C:Chl weight ratios. Upwelling is indicated by letter A. (B). As in Figure SB but C:Chl. Ratio is elevated at ZTM; decrease in the upper mixed layer occurs over the 4 days. Greater decrease below 5 m and above ZTM indicates the upper mixed layer was not fully mixing. Contours are 20 weight weight intervals.

sampling, values indicative of nutrient limitation. The ratio decreased to values near 7 over the four days (Fig. lOB). At both stations, highest values were observed at the temperature maximum; the ratio decreased below this depth. The decreases in the upper mixed layer occurred below 5 m depth. C:Chl weight ratios were higher than 130 and suggestive of nutrient limited phytoplankton in the mixed layer inshore and offshore on the first day of sampling. The ratio decreased over time (Fig. l1A, B). At both stations, the decrease was largely below 5 m and above ZTM. The largest ratios occurred in the temperature maximum at both stations and also decreased over

23 Table I. Abundance of adult A. monica on 14 October 1995 Depth (m) 0-8 9 10 II

Adults m- 3

Standard error

507 797 1033

38 146

797

68

72

time. The ratio was lower below ZTM and values were typical for nutrient replete, light limited cells. Zooplankton A. monica were sampled on 14 October at the offshore

station using a vertical net haul for the upper 8 m and a Schindler trap at 9, 10 and II m (Table 1). The number of adults m- 3 from 9 to II m was 70% higher than in the upper mixed layer. Calculated fluxes of ammonia at the depth of the ammonia maximum For vertical flux to occur, E > 15vN2 (lvey & lmberger, 1991). We determined whether the threshold was exceeded one meter above and one meter below the temperature maximum. Inshore, the percentage of segments in which dissipation rates exceeded the threshold on II, 12 and 13 October were 39%, 72% and 45%, respectively. Offshore, the percentages were 76%, 12% and 20%. These calculations indicate that vertical transport at the top of the pycnocline would have been greatest offshore on the first day and lower thereafter; vertical transport would have been greater onshore on the second and third days. Had microstructure profiling at the inshore site occurred on II October after several hours of wind forcing, percentages may have been more similar to those at the offshore site. We developed two approaches to compute lakewide estimates of the upward flux of ammonia. For the first approach, we assumed fluxes occurred on all days and weighted fluxes by multiplying by the percentage of segments in which the turbulence exceeded the threshold for mixing (Method A). We weighted the inshore and offshore fluxes by 20% and 80% based on the calculations in MacIntyre et al. (1999). For the second approach, we assumed that turbulence was only sufficiently energetic for turbulent transport at the offshore site on 11 October. This assumption was

based on the persistence of a high NH4 gradient at the offshore site and because E only exceeded the critical value for turbulent transport a significant fraction of the time on II October. Similarly, we assume that turbulence was sufficiently energetic for turbulent transport on II and 12 October at the inshore site. This assumption is based on the reduction in the nutrient gradient at the inshore station from II to 12 October, but its increase on the third day, and the higher percentages of E greater than the threshold for turbulent flux on 12 October than 13 October. Consequently, we computed fluxes for 11 October using results from both stations, for 12 October using only the inshore station. (Method B). The areal weighting was as for Method A. The average flux over three days was 1.6 mmoles m- 2 d- 1 using Method A and 1.1 mmoles m- 2 d- 1 using Method B (Table 2).

Discussion Major progress in understanding physical controls on lacustrine ecosystem dynamics will occur when we can relate the onset, spatial variability and intensity of turbulence to wind forcing and the resulting internal wave field. Power spectra from our data clearly show the higher energy content in the wave field after persistent, high winds on 11 October and the subsequent decline in energy (Fig. 5). Rates of dissipation of turbulent kinetic energy were higher at the top of the pycnocline on the first 2 days of the study and were highest at the top and base of the pycnocline at the inshore site on the second day. Both the power spectra, the filtered data, and the potential energies indicate the energy in the wave field was elevated over a broad range of frequencies after wind forcing (Figs 5 and 6). Expansion of the ideas developed in Monismith (1985) may lead to new insights. Monismith showed that the magnitude of shear depends on the phase between waves on different interfaces. The largest shears occur when isopycnal surfaces become level in the transition between the pycnocline's up and downwelling (Monismith, 1985). Such shears may be sufficient to induce turbulence. Monismith's analysis was restricted to a three-layer model of a lake's stratification and to the first two vertical mode internal waves. However, stratification within lakes is much more complex. Not only can higher order vertical wave modes be supported by these stratifications (LaZerte, 1980), but also harmonics of the horizontal modes. In consequence, the movement of adjacent

24 Table 2. Flux of ammonia away from the Nl4 maximum in Mono Lake. F =Kz B[NH41/Bz where Kz is the arithmetic average in I m bin. Calculations were done within the N14 maximum at depths where the concentration gradient was maximal. I indicates inshore data used in the calculation; 0, offshore. Lakewide fluxes are computed when turbulence is sufficiently energetic to cause vertical fluxes. Depth (m)

IlNH4 JLMm- 1

Kz m2 s-I

Inshore II Oct 1410 h 12 Oct 0925 h 13 Oct 1600 h

6-8 9-10 10-11

1.1 0.7 3.3

1.3x 10- 5 1.3x 10-5 1.6x 10- 6

Offshore II Oct 1620 h 12 Oct 1310 h 13 Oct 1445 h

8-10 10-11 10-11

2.5 3.8 5

1.6x 10- 5 5.7x 10- 5 8.7xlO-6

Three-day

Flux mmoles m- 2 d- I

1.2 1.1 0.45

Lakewide Flux (A)

Lakewide Flux

mmoles m- 2 d- I

mmoles m- 2 d- 1

2.2 1.9 0.6

3010 0.2,) 0.0

1.6

1.1

(B)

3.4 18.7 3.7

average flux

isopycnals may frequently be out of phase causing intermittent increases in shear. Consequently, turbulence may be intermittently generated by the mechanism of shear instability. A full exploration of these ideas is beyond the scope of this paper, but phase differences between adjacent isopycnals were present. For example, at ca. 0200 h 13 October, the 6 °C isotherm down wells ca. 0.75 m, while the 7°C isotherm remains nearly level (Fig. 3B). Such differences in movement may lead to shear. Phase differences between waves in different frequency bands are illustrated in Figure 6A,C, E. For instance, while the lowest frequency waves comprising the 7 and 5 °C isopycnals are nearly in phase at 0000 h 13 October, the waves in the 2nd lowest band go through essentially one up and downwelling cycle from 0000 h to 0200 h on the 7 °C isotherm while they go through three such cycles on the 5 °C isopycnal. These differences in frequency lead to the phase differences between isotherms, as observed between the 7 °C and 6 °C isotherms in Figure 3B, and may contribute to steppiness in the pycnocline as observed inshore and to the development of shear and subsequent turbulence. Waves of different frequencies could also be caused by the interaction of basin-scale waves with topography, and their interaction with the ambient density field could lead to enhanced turbulence (Thorpe, 1998). For example, steppiness can be generated in the pycnocline due to the interaction of the ambi-

ent density field with lee waves which are generated by the interaction of larger scale waves with topographic features (Thorpe, 1998). The step-like features in the pycnocline inshore on 12 and 13 October, with enhanced dissipation rates associated with the steps, provide evidence that such interactions may be occurring in Mono Lake. Step-like features may also be due to intrusions as would result from wave breaking inshore and flow of the well mixed water along isopycnals offshore. Dissipation rates were also high at the top of the pycnocline both inshore and offshore. We do not have time series temperature data there to help define the mechanisms likely to have generated the turbulence. The turbulence may have been due to shear instability with shear generated due to the initial upwelling of basin scale waves. The magnitude of the Turner angle indicates that double diffusive convection may also have contributed to turbulence at the top of the pycnocline (Macintyre et aI., 1999). In summary, we observed enhanced turbulence at the top of the pycnocline both inshore and off when internal wave amplitudes increased due to strong wind forcing. Turbulence was enhanced throughout the pycnocline inshore on the second and third days after the initial wind forcing. We now combine these data with our chemical and biological observations to address the question of when and where vertical fluxes of NH4 occurred and the consequences for phytoplankton in Mono Lake.

25 The persistent ammonium maximum Ammonium profiles consistently show that the highest concentration of ammonium occurred within a meter of the temperature maximum and that this feature persisted over a 4-day period despite enhanced turbulence at those depths. Changes in concentration over time depend on sources, sinks and vertical fluxes, and we do not have sufficient data to construct a full budget. For instance, quantifying uptake of N by phytoplankton would have required measurements at the time of the experiment. However, we can evaluate sources and vertical fluxes. If the fluxes due to turbulent mixing were large relative to the sources, we would not expect the NH4 maximum to persist. We can determine whether sufficient ammonium was mixed vertically to affect nutrient limitation by phytoplankton. Sources Sources of NH4 include excretion by A. monica, enhanced diffusive fluxes at the sediment water interface, advection or excretion by microzooplankton. Data on microzooplankton are not available. Excretion rates of NH4 by adultA. monica are 0.6 {lmole d-' (Jellison et aI., 1993). With an average abundance of A. monica of 876 adults m- 3 at the depths of the temperature maximum, the flux into a 1 m3 parcel of water would be 0.5 mmoles m- 3 d-'. JuvenileA. monica were not present in sufficient numbers to contribute to the ammonium pool. Previous lake wide sampling has indicated that the maximum numbers of brine shrimp can range from 500 to 3000 adults m- 3 (Lenz, 1980). Were such a wide range present during our sampling, maximum fluxes ofNH4 could be up to l.8 mmoles m- 3 d-'. Jellison et al. (l993b) estimated diffusive fluxes from the sediments at depths below the nutricline to be 6 mmoles m- 2 d- 1. Data are not available to calculate fluxes from sediments above the nutricline. Were they comparable, the flux into aim layer spanning the whole lake would be 2 {lmoles m- 3 d- 1, a flux three orders of magnitude too low to account for the higher NH4 at those depths. Shear stresses at the sedimentwater interface may enhance the fluxes, but at present, theoretical models indicate at most a three-fold increase due to turbulence at the benthic boundary layer (Dade, 1993). Advection of water with higher NH4 content could contribute to persistence of the feature, but our data do not show large spatial differences in NH4 concentrations. Based on these comparisons, we infer that the high concentrations of NH4 at the top of the pycnocline

resulted from excretion by adult A. monica. The elevated concentrations could only have persisted if the vertical fluxes of NH4 due to turbulence were less than the NH4 excretion rates and if uptake rates by phytoplankton or bacteria were low. Vertical fluxes Several lines of evidence suggest that vertical fluxes of NH4 occurred and that turbulent transport may have been greater at inshore sites than offshore sites. The time series measurements of temperature, NH4, chlorophyll a and C:Chl indicate upwelling occurred inshore on II October but not offshore. This upwelling, coupled with the enhanced turbulence at the depths of upwelling, may have contributed to the greater spread of NH4 in the NH4 maximum and of chlorophyll a in the chlorophyll a minimum (Figs 8 and 9). Increased spread of a solute and slightly lower maximum concentrations are to be expected when mixing has occurred. Chlorophyll a is more dispersed inshore below 12 m depth. For vertical mixing to occur and to cause transport across isopycnals, dissipation rates must be high enough to exceed the damping induced by viscous and buoyancy forcing. Dissipation rates were sufficiently high at the top of the pycnocline 76% of the time offshore but less than 20% of the time on the next 2 days. Inshore, dissipation rates were sufficiently high 72% of the time on the second day and 45% of the time on the third day. Temperature profiles with millimeter scale resolution indicated that temperature inversions with scales of centimeters were present when E was large enough to cause transport. Inversions were on smaller scales or not present when it was less than 15 vN 2 (S. MacIntyre, unpublished data). Consequently, when E exceeded the threshold, turbulent overturns I occurred allowing vertical transport. Calculated daily vertical fluxes of NH4 at the offshore site are higher than likely excretion rates (Table 2). We assume the flux is into a 1 m3 volume for comparison with the A. monica excretion rates. The calculated fluxes were particularly high on days when E was frequently below the threshold for vertical transport and when the NH4 gradient remained high. These results suggest vertical fluxes, calculated as F = Kz 3[NH4]/3z, were much less than calculated. The discrepancy indicates that calculating K z as rEN- 2 (Osborn, 1980) is not always valid. Itsweire et al. (1993) found that the assumptions of Osborn's (1980) model were not met when stratification and viscosity were high relative to the forces

26 that would generate turbulence. In contrast, calculated daily fluxes at the inshore site could be supported by excretion by A. monica. Inshore a greater percentage of the dissipation rates exceeded the threshold for turbulent transport; consequently Osborn's model was applicable for calculating K z . Vertical fluxes and phytoplankton growth In comparison to the quantity of ammonia excreted by A. monica, the calculated lake-wide fluxes of NH4 are reasonable although at the upper end of the range possible (Table 2). As these calculated fluxes take into account the percentage of the time that turbulence is sufficiently high to overcome damping forces, we use them to estimate the possible growth of phytoplankton. Using a C:N molar ratio of 7, the lower of the three day averages, 1.1 mmoles m- 2 d- I , the fluxes would have supported primary productivities of 0.1 g C m- 2 d- I . Daily rates of primary production are between 0.5 and 2 g C m- 2d- 1 during the summer in Mono Lake (Jellison & Melack, 1993a). The fluxes would have provided sufficient N~ to support 5-20% of the daily requirements for growth. Growth will only occur at depths to which the NH4 is mixed. Persistence of thermal stratification in the upper mixed layer and E only consistently exceeding the threshold for vertical mixing in the upper 5 m indicates the upper mixed layer did not fully mix (MacIntyre et aI., 1999, MacIntyre, unpublished data). Consequently, the vertical fluxes of NH4 will be into a limited volume and growth or reduction of nutrient limitation will not occur throughout the entire upper mixed layer. To further determine whether vertical fluxes of NH4 occurred, we detennine whether there was evidence for alleviation of nutrient limitation or for growth. Vertical fluxes of NH4 and changes in nutrient limitation and abundance of chlorophyll C:N ratios and C:Chl ratios decreased at depths above the temperature maximum at all sites over the course of the study (Figs 10 and 11). The decrease in the C:N ratio in the upper mixed layer at the offshore station provided the strongest evidence for a reduction in nutrient limitation. C:N molar ratios above 8.3 indicate waters that are moderately N deficient (Guildford & Hecky, 2000), and offshore waters initially had C:N molar ratios in excess of 9. In contrast, the ratios in the upper mixed layer inshore do not indicate nutrient

limitation during this experiment. Decreases occurred below 5 m depth. Jellison & Melack (2001) illustrate the seasonal variation of C:Chl weight ratios in Mono Lake. Ratios above 200 occur in summer when phytoplankton are most nutrient stressed. Ratios decline in the autumn due to the onset of vertical mixing and have the lowest values in winter. Frenette et al. (1996) attribute a decrease in C:Chl from a range of 100-150 to a range of 50-100 to increased intracellular chlorophyll after a wind event caused resuspension of sediments and nutrient rich pore waters. They reasoned that resuspension of non-algal material would increase the C:Chl ratio, and that the reduction was therefore indicative of increased chlorophyll in phytoplankton cells. An alternate explanation for a decrease in the C:Chl ratio is vertical mixing of dark adapted phytoplankton cells into the upper mixed layer. For instance, for cells grown in cultures at high light the ratio was around 100, at low light, 20 (Cullen & Lewis, 1988). Depending on the time scale for photoadaptation, a decrease in the ratio in surface waters could indicate upward mixing of dark-adapted cells with higher intracellular chlorophyll. Because a minimum in chlorophyll a occurred at the depths where E was highest, we do not believe the decreased ratio we observed was due to transport of dark-adapted cells from deeper in the pycnocline into the upper mixed layer. The decrease reflected either growth of phytoplankton or growth of chloroplasts within individual cells. Again, the decrease in the ratio occurred below 5 m depth and above ZTM·

Over the 4 days, chlorophyll a abundance doubled in the upper mixed layer at the inshore station and increased by 30-50% offshore. Advection could have led to the observed changes in biomass if the gyre (Melack & Gastil, this volume; Melack, in press) in the northeastern section of the lake on 10 October shifted its position due to wind forcing or was dispersed. On 10 October, chlorophyll concentrations in the gyre ranged from 2.8 to 4.3 ug I-I. Station 8, just on the south-eastern edge of the feature, had a chlorophyll a concentration of 2.9 {Lg I-I, and concentrations at stations 5 and 6 were 2.2 {Lg I-I and 1.9 {Lg I-I, respectively. Movement of the gyre could account for the increased biomass observed at our study sites. Inside the gyre the C:N ratio was ~ 12 and C:Chl ratios exceeded 170. These values indicated the phytoplankton in the gyre were nutrient limited. Consequently, the decreases in these ratios at our study sites indicate

27 that alleviation of nutrient limitation was not due to advection but to vertical mixing of NH4. In summary, changes in C:N and C:Chl ratios indicate that vertical fluxes of ammonium occurred. Vertical fluxes of N~ were induced by increased potential energy in the internal wave field due to wind forcing. After the initial wind forcing, fluxes were more likely to occur inshore. While these fluxes were sufficient to cause slight increases in growth, they did lead to suppression of nutrient limitation and an increase in chlorophyll within phytoplankton cells. Consequently, an initial strong wind forcing event reduced nutrient limitation which could permit subsequent growth if fluxes of nutrients continued from vertical mixing or A. monica excretion. The ratios decreased more at the base of the upper mixed layer indicating that it never fully mixed. Enhanced growth would be most likely to occur at those depths. Within the temperature maximum, NH4 values stayed fairly constant, particulate C as well as chlorophyll a was always lower than in the water immediately above and below, and the C:Chl weight ratio decreased. On 14 October, A. monica abundance was highest where NH4 concentrations were highest. Consequently their grazing and excretion may explain the changes in the concentrations of NH4, chlorophyll a and particulate carbon within the temperature maximum. At the beginning of our sampling period, the water was nearly anoxic at 12 m depth, and algal biomass increased at 12 m. Hence, the zooplankton may have initially been abundant at a depth above 12 m where oxygen concentrations were sufficient and abundances of phytoplankton were higher than in overlying waters. The turbulence in the metalimnion would have continued to supply the A. monica with phytoplankton. The decrease in the C:Chl ratio at that depth may reflect the upward mixing of dark adapted phytoplankton cells.

Persistent layering Thin layers of phytoplankton, zooplankton and nutrients that extend for kilometers have recently been observed in coastal waters and fjords (Cowles & Desiderio, 1993; Hanson & Donaghay, 1998; Alldredge et aI., in press.). Bacteria also accumulate in the layers, and layers of particulate organic matter are elevated at the same depths as these layers or slightly below. Typically the layers are associated with density discontinuities (Macintyre et aI., 1995). Turbulent intensities are too low and turbulent eddies are too small

to disrupt them (Alldredge et aI., in press.). In fact, turbulence is often higher on their boundaries than within them, indicating that the layers are flowing between other water masses as intrusions. Our finding a layer with elevated N~ and abundant zooplankton in Mono Lake is another example of persistent layering over space and time. It persisted because the turbulence was intermittent, and did not remain at intensities sufficient to cause vertical transport.

Summary Intermittent transport and mixing events are critical for resupply of nutrients to the euphotic zone in stratified water bodies. Mechanisms of supply include upwelling (Coulter & Spigel, 1991; Ostrovsky et aI., 1996), boundary mixing (Goudsmit et aI., 1997; Macintyre et aI., 1999), shear and static instability of internal waves (Thorpe, 1978, 1994), entrainment caused by wind mixing or heat loss by convection (Macintyre & Melack, 1995), differential heating and cooling (James & Barko, 1991; Nepf & Oldham, 1997), and sediment resuspension with subsequent advection (Robarts et aI., 1998). Our study indicates that the energy supplied by the wind energizes the internal wave field, that energy decays rapidly, and that the resulting turbulence occurs over a limited period of time. Initially the turbulence occurred at the top of the pycnocline inshore and offshore, but later the turbulence was greater inshore. Vertical transport of nutrients occurred when dissipation rates were sufficient to overcome viscous and buoyancy forces. At those times, turbulent mixing caused sufficient nutrient flux to alleviate nutrient limitation and to support modest growth. This improved physiological state could lead to increased growth were fluxes to continue. While strong wind events occur infrequently in the autumn in Mono Lake, frontal systems with high winds are frequent in the spring and a strong diurnal wind pattern sets up in the summer. In these other seasons, wind forcing may lead to injections of limiting nutrients and enhanced growth. Similar scenarios are likely in other lakes. In addition, spatial-temporal variations in vertical mixing and its limited vertical extent abets structuring of biological communities despite strong wind forcing. Rates of biological and chemical processes may differ inshore and off.

28

Acknowledgements We thank Darla Heil, Mike Emory, Pete Kirchner and Kevin Flynn for their help with field work and chemical processing. Manuela Lorenzi-Kayser, Kevin Flynn and Lorenz Moosmann helped with data processing, programming, and graphics. Mary Gastil provided essential information regarding the gyre. We thank William Shaw for assistance with signal processing techniques, G.w. Kling and L. Moosmann for helpful comments, and E.E. McPhee, W. Shaw and lM. Melack for critically reading the manuscript. The Centre for Water Research, University of Western Australia, provided the software for microstructure analysis. We thank Michael Head for his technical support of the microstructure profilers. Financial support was provided by NSF Grants DEB93-17986, DEB97-26932, and OCE99-06924 to S.M. and DEB95-08733 to R.I. and lM. Melack. Logistic support was provided by the Sierra Nevada Aquatic Research Laboratory, University of California Natural Reserve System.

References Alldredge. A. L., T. J. Cowles, S. MacIntyre, J. E. B. Rines, P. L. Donaghay, C. F. Greenlaw, D. V. Holliday, M. M. Dekshenieks, J. M. Sullivan & J. R. V. Zaneveld, in press. Occurrence and mechanisms of formation of a dramatic thin layer of marine snow in a shallow Pacific fjord. Mar. Eco!. Prog. Ser. Coulter, G. W. & R. H. Spigel, 1991. Hydrodynamics. Chapter 3 In Coulter, G. W. (ed.), Lake Tanganyika and its Life. Oxford Univ. Press: 49-75. Cowles, T. J. & R. A. Desiderio, 1993. Resolution of biological microstructure through in situ fluorescence emission spectra. Oceanography 6: lOS-III. Cullen, J. J. & M. R. Lewis, 1988. The kinetics of algal photoadaptation in the context of vertical mixing. 1. Plankton Res. 10: 1039-1063. Dade, W. B., 1993. Near-bed turbulence and hydrodynamic control of diffusional mass transfer at the sea floor. Limno!. Oceanogr. 38: 52-69. DeSilva, I. P. D., J. Imberger & G. N. Ivey, 1997. Localized mixing due to a breaking internal wave ray at a sloping bed. J. Fluid Mech. 350: 1-27. Dillon, T. M. & D. R. Caldwell, 1980. The Batchelor spectrum and dissipation in the upper ocean. J. Geophys. Res. 85: 1910-1916. Etemad-Shahidi, A. & J. Imberger, 200 I. Anatomy of turbulence in thermally stratified lakes. Limno!. Oceanogr. 46: 1158-1170. Frenette, J. J., W. F. Vincent, L. Legendre & T. Nagata, 1996. Size-dependent phytoplankton responses to atmospheric forcing in Lake Biwa. J. Plankton Res. 18: 371-391. Gloor, M., A. Wuest & M. Munnich, 1994. Benthic boundary mixing and resuspension induced by internal seiches. Hydrobiologia 284: 59-68.

Golterman, H. L., 1969. Methods for Chemical Analysis of Fresh Waters. International Biological Program Handbook. No.8. Blackwell Scientific Publications, Oxford: 166 pp. Goudsmit, G.-H., F. Peeters, M. Gloor & A. Wuest, 1997. Boundary versus internal diapycnal mixing in stratified natural waters. J. Geophys. Res. 102: 27903-27914. Guildford, S. J. & R. E. Hecky. 2000.Total nitrogen, total phosphorus, and nutrient limitation in lakes and oceans: is there a common relationship? Limno!. Oceanogr. 45: 1213-1223. Hanson, A. K. & P. L. Donaghay, 1998: Micro- to fine-scale chemical gradients and layers in stratified coastal waters. Oceanography II: 10--17. Imberger, J., 1985. The diurnal mixed layer. Limno!. Oceanogr. 30: 737-770. Imberger, J. & G. Ivey, 1991. On the nature of turbulence in a stratified fluid. Part 2: application to lakes. J. Phys. Oceanogr. 21: 659-680. Imberger, J. & J. Patterson, 1990. Physical limnology. Adv. App!. Mech. 27: 303-475. Itsweire, E. C., J. R. Koseff, D. A. Briggs & J. H. Ferziger, 1993. Turbulence in stratified shear flows - implications for interpreting shear-induced mixing in the ocean. J. Phys. Oceanog. 23: 1508-1522. Ivey, G. & J. Imberger, 1991. On the nature of turbulence in a stratified fluid. Part I: the efficiency of mixing. J. Phys. Oceanogr. 21: 650--658. James, W. F. & J. W. Barko, 1991. Littoral-pelagic phosphorus dynamics during night-time convective circulation. Limno!. Oceanogr. 31: 900--906. Jassby, A. & T. M. Powell, 1975. Vertical patterns of eddy diffusion during stratification in Castle Lake, California. Limno!. Oceanogr. 38: 1008-10 19. Jellison, R. S. & J. M. Melack, 1988. Photosynthetic activity of phytoplankton and its relation to environmental factors in hypersaline Mono Lake, California. Hydrobiologia 158: 69-88. Jellison, R. S. & J. M. Melack, 1993a. Algal photosynthetic activity and its response to meromixis in hypersaline Mono Lake, California. Limno!. Oceanogr. 38: 818-837. Jellison, R. S. & J. M. Melack, 1993b. Meromixis in hypersaline Mono Lake, California. I: Vertical mixing and density stratification during the onset, persistence and breakdown of meromixis. Limno!. Oceanogr.38: 1008-1019. Jellison, R. S. & J. M. Melack, 2001. Nitrogen limitation and particulate elemental ratios of seston in hypersaline Mono Lake, California, U.S.A. Hydrobiologia 466 (Dev. Hydrobio!. 162): 1-12. Jellison, R. S., L. Miller, J. M. Melack & G. L. Dana, 1993. Meromixis in hypersaline Mono Lake, California. 2: Nitrogen fluxes. Limno!. Oceanogr. 38: 1020-1039. Jellison, R., S. MacIntyre & F. 1. Millero, 1999. Density and conductivity properties of Na-C03 -CI-S04 brine from Mono Lake, California. Int. J. Salt Lake Res. 8: 41-53. LaZerte, B., 1980. The dominating higher order vertical modes of the internal seiche in a small lake. Limno!. Oceanogr. 25: 846854. Ledwell, J. R., E. T. Montgomery, K. L. Polzin, L. C. St. Laurent, R.W. Schmitt & J. M. Toole, 2000. Evidence for enhanced mixing over rough topography in the abyssal ocean. Nature 403: 179-182. Lenz, P., 1980. Ecology of an alkali-adapted variety of Artemia from Mono Lake, California, U.S.A. In Persoone, G., P. Sorgeloos, O. Roels & E. Jaspers (eds), The Brine Shrimp Artemia. Vo!. 3. Ecology, Culturing. Use in Aquaculture. Universa Press, Welteren, Belgium: 79-96.

29 Lueck, R. G. & T. D. Mudge, 1997. Topographically induced mixing around a shallow seamount. Science 276: 1831-1833. Macintyre, S., 1993. Vertical mixing in a shallow eutrophic lake: possible consequences for the light climate of phytoplankton. Limnol. Oceanogr. 38: 798-817. Macintyre, S., 1998. Turbulent mixing and resource supply to phytoplankton. In Imberger, J. (ed.), Physical Processes in Lakes and Oceans, Coastal and Estuarine Studies, AGU: 539-567. Macintyre, S., A. L. Alldredge & C. C. Gotschalk, 1995. Accumulation of marine snow at density discontinuities in the water column. Limnol. Oceanogr. 40: 449-468. Macintyre, S. & 1. M. Melack, 1995. Vertical and horizontal transport in lakes: linking littoral, benthic and pelagic habitats. J. n. am. Benthol. Soc. 14: 599-615. Macintyre, S., K. M. Flynn, R. Jellison & 1. R. Romero, 1999. Boundary mixing and nutrient flux in Mono Lake, CA. Limnol. Oceanogr. 44: 512-529. Macintyre, S. & 1. R. Romero, 2000. Predicting upwelling, boundary mixing, and nutrient fluxes in lakes. Verh. int. Ver. Limnol. 27: 246-250. Maxwortby, T., J. Imberger & A. Saggio, 1998. A laboratory demonstration of a mechanism for the production of secondary, internal gravity-waves in a stratified fluid. In Imberger, 1. (ed.), Physical Processes in Lakes and Oceans. Coastal and Estuarine Studies, AGU: 261-270. Melack, J. M., (in press). Ecological dynamics in saline lakes. Verh. int. Ver. Lirnnol. Melack, J. M. & R. Jellison, 1998. Limnological conditions in Mono Lake: Contrasting monomixis and meromixis in the 1990s. Hydrobiologia 384: 21-39. Melack, 1. M. & M. Gastil, 2001. Airborne remote sensing of chlorophyll distributions in Mono Lake, California. Hydrobiologia 466 (Dev. Hydrobiol. 162): 31-38. Monismith, S. G., 1985. Wind-forced motions in stratified lakes and their effect on mixed-layer shear. Limnol. Oceano gr. 30: 771783. Mourn, J. N., D. Hebert, C. A. Paulson & D. R. Caldwell, 1992. Turbulence and internal waves at the equator. Part I. Statistics from towed thermistors and a microstructure profiler. J. Phys. Oceanogr. 22: 1330-1345. Nepf, H. M. & C. E. Oldham, 1997. Exchange dynamics of a shallow contaminated wetland. Aquat. Sci. 59: 193-213. Oakey, N., 1985. Statistics of mixing parameters in the upper ocean during JASIN Phase 2. J. Phys. Oceanogr. IS: 1662-1675. Osboru, T. R., 1980. Estimates of the rate of vertical diffusion from dissipation measurements. J. Phys. Oceanogr. 10: 83-89.

Ostrovsky, I, Y. Z. Yacobi, P. Walline & I. Kalikhman, 1996. Seiche-induced mixing - its impact on lake productivity. Limnol. Oceanogr. 41: 323-332. Patterson, 1. c., 1991. Modelling the effects of motion on primary production in the mixed layer oflakes. Aquat. Sci. 53: 218-238. Polzin, K. L., J. M. Toole, J. R. Ledwell & R. W. Schmidt, 1997. Spatial variability in turbulent mixing in the abyssal ocean. Science 276: 93-96. Robarts, R. D., M. Waiser, O. Hadas, T. Zohary & S. Macintyre, 1998. Contrasting relaxation of phosphorus limitation due to typhoon-induced mixing in two morphologically distinct basins of Lake Biwa, Japan. Limnol. Oceanogr. 43: 1023-1036. Romero, J. R. & J. M. Melack, 1996. Sensitivity of vertical mixing in a large saline lake to variations in runoff. Limnol. Oceanogr. 41: 955-965. Romero, 1. R., R. Jellison & J. M. Me1ack, 1998. Stratification, mixing and ammonia flux in a hypersaline lake. Archiv. Hydrobiol. 142: 283-315. Saggio, A. & J. Imberger, 1998. Internal wave weather in a stratified lake. Limnol. Oceanogr. 1780-1795. Saggio, A. & J. Imberger, 2001. Mixing and turbulent fluxes in the metalimnion of a stratified lake. Limnol. Oceanogr. 46: 392-409. Strickland, 1. S. H. & T. R. Parsons, 1972. A practical handbook of seawater analysis. Bull. Fish. Res. Bd Can. 167: 1-130. Svensson, U., 1978. Examination of the summer stratification. Nordic Hydrology 9(2). Tennekes, H. & J. L. Lumley, 1972. A First Course in Turbulence. MIT Press: 300 pp. Thorpe, S. A., 1978. On the shape and breaking of finite amplitude internal gravity waves in a shear flow. J. Fluid Mech. 85: 7-31. Thorpe, S. A., 1984. A laboratory study of stratified ace1erating shear flow over a rough boundary. 1. Fluid Mech. 138: 185-196. Thorpe, S. A, 1987. On the reflection of a train of finite-amplitude internal waves from a uniform slope. J. Fluid Mech. 178: 279302. Thorpe, S. A., 1989. The distortion of short internal waves produced by a long wave, with applications to ocean boundary mixing. 1. Fluid Mech. 208: 395-415. Thorpe, S. A., 1994. Statically unstable layers produced by overturning internal gravity waves. J. Fluid Mech. 260: 333-350. Thorpe, S. A., 1998. Some dynamical effects of the sloping sides of lakes. In Imberger, J. (ed.), Physical Processes in Lakes and Oceans. Coastal and Estuarine Studies, AGU: 441-460. Wuest, A., G. Piepke & D. C. Van Senden, 2000. Turbulent kinetic energy balance as a tool for estimating eddy diffusivity in windforced stratified waters. Limnol. Oceanogr. 45: 1388-1400.

Hydrobiologia 466: 31-38,2001. J.M. Melack, R. Jellison & D.B. Herbst (eds), Saline Lakes. © 2001 Kluwer Academic Publishers.

31

Airborne remote sensing of chlorophyll distributions in Mono Lake, California John M. Melack1,2 & Mary Gastill 1Institute for Computational Earth System Science, University of California, Santa Barbara, CA 93106, U.S.A. 2Department of Ecology, Evolution and Marine Biology, University of California, Santa Barbara, CA 93106, U.S.A.

Key words: remote sensing, chlorophyll, Mono Lake, saline lakes

Abstract As part of a long-term investigation of seasonal and interannual variations of plankton in Mono Lake (California), we developed a methodology using airborne imaging spectrometry to synoptically measure chlorophyll concentrations. Images of Mono Lake were acquired with NASA's Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) and were atmospherically corrected by applying a version of the radiative transfer model MODTRAN. Using a predictive equation for calculation of chlorophyll based on a band ratio of remote sensing reflectances (R rs ; Rrs 490 nml Rrs 550 nm), spatial distributions of chlorophyll throughout the lake were determined; broad east to west gradients in chlorophyll and gyres are evident.

Introduction Variations in phytoplankton abundance occur over a wide range of temporal and spatial scales as a consequence of interactions among biological, chemical and physical processes (Harris, 1986; Denman, 1994; Neill, 1994). A commonly used measure of the abundance of phytoplankton is the concentration of chlorophyll. Measurement of chlorophyll can be done in the laboratory after collection of discrete samples, in situ by derivation from fluorescence, or by remote sensing, which makes possible synoptic estimates over large areas (e.g., Abbott et aI., 1982; Galat & Verduin, 1989). Remote sensing of chlorophyll depends on detection of sunlight or laser light (Hoge & Swift, 1983; Kirk, 1983) that has interacted with chlorophyll within aquatic plants and is upwelled toward a sensor located at some altitude above the water. Most sensors are mounted on airborne or spaceborne platforms, although hand-held instruments can be used. Recent advances in remote sensing systems have expanded the capabilities and opportunities for applications to the study of inland waters (Dekker et aI., 1995). Our purpose here is to describe a methodology for using a remote sensing technique called imaging spec-

trometry for the determination of chlorophyll in lakes. Airborne imaging spectrometry uses high resolution, continuous reflectance spectra to decipher optical conditions in the water and atmosphere. We illustrate the approach with data from hypersaline Mono Lake, California.

Background on remote sensing of lakes Sunlight incident on a lake is partially reflected from the surface and partially transmitted into the lake where it is scattered and absorbed (Kirk, 1983). The theoretical basis for remote sensing of optical properties of water is well known (Gordon & Morel, 1983; Kirk, 1983; Mobley, 1994). The water-leaving radiance is remotely sensed by inverting the equation describing the radiance measured at high altitude, such as by the Airborne Visible and Infrared Imaging Spectrometer (AVIRIS): L AVIRISAhB

=

LpathAhB

+ LwaterAB TAhe

+LskYAhOB hhe PAiJ

+LsUnAhoBs hhesPAes

(1)

where subscript A is wavelength, h is sensor altitude, ho is lake surface altitude, () is viewing angle (at or

32 near zenith), and es is solar angle from zenith. LAVIRIS is the spectral radiance measured by AVIRIS, Lpath is the diffuse atmospheric radiance (i.e., path radiance), Lwater is the water-leaving radiance, Lsky is sky radiance, Lsun is direct solar radiance, and T is the diffuse transmittance from surface to sensor, and p is the Fresnel reflectance coefficient. After computational correction for path radiance and reflected skylight and by selection of viewing geometry to eliminate direct solar reflectance (i.e., sun glint), water-leaving radiance just above the surface can be obtained. The water-leaving radiance can be related to the upward irradiance (Eu). Eu is a function of downward irradiance (Ed) and the inherent optical properties of the water. The irradiance reflectance ratio (R) is expressed as Eu R=Ed

(2)

and can be related to remote sensing reflectance (Carder et aI., 1986; see below). Irradiance reflectance is not practical to measure at altitudes significantly above the lake; upwelling radiance at the sensor's view angle is what is detected by high altitude sensors such as AVIRIS. With the sky radiance reflected off the lake surface subtracted, the remaining upwelling radiance is the water-leaving radiance; this is used to define remote sensing reflectance, R as R rs

Lwater

= --. Ed

(3)

About 90% of the light scattered back into the atmosphere originates within the depth of water in which the downwelling irradiance is attenuated to about 10% of its subsurface value (Smith, 1981). Hence, remotely sensed chlorophyll includes only the upper portion of the euphotic zone. Moreover, the light scattered from within the water represents only a small fraction (i.e., less than 10%) of the light that reaches a high altitude sensor. Most of the light, if sun glint is avoided, received by the sensor originates from atmospheric Rayleigh and aerosol scattering and is called path radiance. Measurements made from low-flying aircraft or with hand-held radiometers near the water are not complicated by path radiance. Satellite and airborne sensors of light upwelling from lakes have permitted some success in quantitative detection of chlorophyll in lakes (Dekker et aI., 1995). Modest results have been obtained with the Landsat Multispectral Scanner (MSS) and the Thematic Mapper (TM) (e.g., Almanza & Melack, 1985; Lathrop &

Lillesand, 1986; Galat & Verdin, 1989). The sensitivity of Landsat MSS was not adequate, and its spectral bands were too broad and did not have a band in the blue region; the Landsat TM has narrower bands including one in the blue (Dekker et aI., 1992). While the Coastal Zone Color Scanner (CZCS) and Sea-viewing Wide Field-of-view Sensor (SeaWiFS) have proved valuable in mapping pigments in oceans (Smith & Baker, 1982; Gordon et aI., 1983; O'Reilly et aI., 1998), their application to small or moderate-sized lakes is precluded by their large pixel sizes. Recent results from airborne imaging spectrometers have shown much promise for inland waters (Melack & Pilorz, 1990; Curran, 1994; Melack & Gastil, 1994a; Dekker et aI., 1995). These devices have tens to hundreds of narrow spectral bands, and pixels usually less than 20 m on a side. If flown at high altitude to increase spatial coverage, obtaining a large signal to noise ratio is technically difficult because of the small, spectrally narrow pixels. Most algorithms for calculation of chlorophyll from remotely sensed, upwelling light are valid for clear waters with low concentrations of chlorophyll. However, many lakes contain moderate to high amounts of phytoplankton, other particulates, and dissolved organic compounds (Melack 1985). Hence, one critical step in the application of remote sensing is the development and testing of algorithms appropriate to compute chlorophyll concentrations in optically complex waters (Gitelson & Kondratyev, 1991; Mittenzwey, 1992; Melack & Gastil, I 994b ). For a range of chlorophyll concentrations from about I to 50 mg m- 3 Melack & Gastil (1994b) related a reflectance peak near 689 nm, partially attributable to solar induced fluorescence, to chlorophyll concentration. They reported that the peak height, expressed as the difference between the Rrs at 689 and 669 nm, was a better predictor of chlorophyll (r2, 0.94) than the area under the curve that formed the peak at 689 nm. A similar relation was described by Gitelson (1992) as a ratio of the maximum peak height in the red end of the spectrum to the reflectance at the broad maximum around 560 nm. Because the position of the peak at the red end of the spectra shifts to higher wavelengths as chlorophyll concentration increases, an instrument with high spectral resolution in the region from 670 to 720 nm is required. Another approach derives information from the slope of the spectra, using the shape of the spectra instead of single or only a few bands. Novo et al. (1995) found that the first derivative at 557 nm provided the

33 best correlation with chlorophyll in surface waters of Mono Lake. Campbell & Esaias (1983) described another algorithm that uses spectral curvature to calculate chlorophyll. Lee & Carder (2000) compared band-ratio and spectral-curvature algorithms. Study site At an elevation of 1943 m above sea level, Mono Lake (38° N, 119 0 W) is a large (160 km 2), deep (mean depth 17 m), hypersaline lake (salinity about 94 g I-I) located on the western edge of the North American Great Basin. The phytoplankton community has few species and varies widely in abundance seasonally (Jellison & Melack, 1993). During the early 1990s, when our remote sensing studies were done, the lake was monomictic, circulating during the winter (Melack & Jellison, 1998).

Approach to remote sensing and validation Remote sensing system The Airborne Visible and Infrared Imaging Spectrometer (AVIRIS), operated as a NASA Facility Sensor for scientific research and applications, was used to obtain hyperspectral imagery of Mono Lake. AVIRIS has been operational since 1989, and continual improvements in its hardware, calibration and data processing have occurred in conjunction with on-going scientific applications (Vane & Goetz, 1993; Vane et aI., 1993; Green, 1995). AVIRIS utilizes silicon and indium antimonide line array detectors to cover the spectral region from 370 to 2500 nm with 224 contiguous, ca. 10 nm wide bands. When flying at an altitude of 20 krn aboard NASA's ER-2 aircraft, the instrument images a 10-km swath obtained with a cross-track scanning mechanism and with ground instantaneous field of view of 20 m. Between the 1994 and 1995 flight seasons, the focal planes in AVIRIS were upgraded and the digitization increased from 10 to 12 bits per pixel (Sarture et aI., 1995). The AVIRIS digital output is converted to radiance values in units of /LW cm- 2 nm- I sr- I based on calibration in the laboratory, in the field and with an on-board procedure. Noise equivalent delta radiance expressed as /L W cm- 2 nm- I sr- I is between 0.02 and 0.025 from about 700 to about 480 nm and increases to about 0.05 by 400 nm.

Field and laboratory measurements At about the same time as the overflight of AVIRIS, measurements were made on the lake and on the shore to assist in atmospheric correction and validation of the calculation of chlorophyll. Samples obtained for the measurement of chlorophyll were integrated over the depth sensed by AVIRIS at known locations; buoyed stations or hand-held global positioning system (GPS) receivers with or without differential correction were used to determine locations. Underwater light attenuation of photosynthetically available radiation (PAR) was determined with a cosine-corrected underwater quantum sensor (LiCorI9IS). Samples for the measurement of the chlorophyll a (chi a) content of the phytoplankton were filtered through Gelman AlE glass fiber filters, and the filters were kept frozen at -14°C until pigments were analyzed. Except during periods of low biomass, chI a was determined by spectrophotometric analysis with correction for pheopigments (Golterman et aI., 1978) after a 40-min extraction of the macerated filters in 90% acetone at room temperature in the dark. Low chI a concentrations «5 mg m- 3) were measured on a fluorometer calibrated against spectrophotometric measurements with large-volume lake samples. Field measurements of reflectance spectra were made on the lake and on land at a site appropriate for evaluation of atmospheric correction. The FieldSpec™ from Analytical Spectral Devices Inc. (ASD; Boulder, CO) was used. The instrument uses a linear array of photodiodes to record digital counts in 512 spectral bands (350-1050 nm) witn a 1.4-nm sampling interval and 3-nm spectral resolution (full width at half maximum of the detector response functions). The procedure for measuring remote sensing reflectance (Rrs) was adapted from a method described by Hamilton et al. (1993). Spectra were collected from three targets: a 10% reflectance gray card, lake or land surface, and blue sky. The gray card is a 25 x 25-cm slab of Spectralon SRS-Gray, a calibrated Lambertian reflecting material (LabSphere, Inc., NH). The gray card was held horizontal on a gimbal and above obstructions to the full upward hemisphere. The spectrum from the lake surface includes skylight reflected off the water (Fresnel reflectance) and the light upwelling through the surface from below. To avoid effects from the side of a small boat, the fiber optic cable connected to the sensor was held about 1 m from the sunny side of the boat with the tip pointed

34

Figure I. Process for calculating concentration of chlorophyll from AVIRIS data.

vertically down. The tip was pointed vertically up to measure the spectrum of the blue sky at zenith. Three series of card, water, and sky spectra were recorded at each station. Processing

Processing steps are outlined in Fig. 1. The first step in radiometric calibration is subtraction of the dark current followed by conversion of raw digital numbers to radiances using auxiliary data provided by the Jet Propulsion Laboratory. Atmospheric correction of the AVIRIS data were done with MODTRAN, a MODerate resolution atmospheric radiance and TRANsmittance model, that includes the capabilities of the widely used LOWTRAN 7 model (Kneizys et aI., 1988), but incorporates a more sensitive molecular band model. It also has other desirable characteristics such as inclusion of single and multiple scattering (Rayleigh and Mie); modifiable default atmospheric values for gases (H20, C02, 03, 02 and N2), aerosols, and clouds; and an updated incident solar irradiance spectrum (Anderson et aI., 1995). An updated version of the MODTRAN 3 radiative transfer model (Green, 1991; personal communication from R.O. Green, Jet Propulsion Laboratory, Pasadena, CA) was used for atmospheric correction of the AVIRIS data. The multiple scattering mode was used for both up looking and downlooking radiance; downwelling irradiance was calculated in single scattering mode. Measurable inputs included the overflight

time and geographic location, the altitudes of sensor and surface, sensor viewing angle; horizontal visibility was a subjective observation. Default values for profiles of atmospheric gases were used, except a factor of 1.06 was applied to C02 to reflect increased atmospheric C02 since assignment of the default values, and a factor of 0.55 was applied to water vapor because of the dry atmosphere over the site. MODTRAN was run in about 1-nm increments of wavelength, then integrated, using the band centers and response filters of AVIRIS, to match the sensor's bands. Outputs from MODTRAN used in the reflectance retrieval equations (Eqs (l) and (3)) were skylight radiance, diffuse transmission, path radiance and downwelling irradiance. All of these vary with view angle except Ed; MODTRAN outputs were interpolated for all side scan angles. The reflectance calculated by MODTRAN was verified and adjusted using an appropriate site on land. Our land calibration site, Pumice Flats, is about 1 km wide, nearly flat, and covered with white pumice gravel. The roughness results in very little specular reflection, and the flatness results in the fraction of the surface in shadow varying only slightly with sun angle. When dry and snow-free, its reflectance spectrum is nearly time invariant. For acquisitions when Pumice Flats was obscured in the image due to clouds or snow, the atmospheric correction was not as accurate. To georeference the images, ground control points were located on the images, and their coordinates extracted from high-resolution orthophotographs (scale 1:24000) or obtained in the field with GPS receivers. Remote sensing reflectance spectra were extracted from the hyperspectral image at the sampling locations. A bin of lOx 10 pixels produced a sufficiently smooth spectrum. These spectra were used to validate the atmospheric correction by comparison to field spectra. The subtraction of Rrs (800 nm) as a constant from the entire AVIRIS spectrum was found to improve the fit to in situ spectra. These spectra were used to find the band combination which best predicted chlorophyll. The choice of AVIRIS bands to ratio was suggested by work done on the SeaWiFS project and published in O'Reilly et al. (1998). The best fit was found to be a log-linear line fit to a band ratio.

Log 10 (chI a) = [

0.926 -

[Rrs490- RrS800]] Rrs555 - Rrs800

0.378

(4)

35

Figure 2. Enhanced true color image made by combining red (488 nm), green (557 nm) and blue (667 nm) bands.

Figure 3. Distribution map of chlorophyll, Mono Lake, CA.

Using 132 pairs of measured chlorophyll and Rrs ratio from seven AVIRIS acquisition dates, the Pearson's correlation coefficient r2 was 0.70.

The AVIRIS spectra at full resolution are needed in the validation of the atmospheric correction and in the development of the chlorophyll algorithm. However,

36 the equation we developed for chlorophyll only uses three of the 224 bands of AVIRIS and an additional band for masking land. Since the spectral calibration varies each year, the AVIRIS bands used are those with centers closest to 490 and 555 nm (with a nominal 10 nm full width at half maximum). The band closest to 800 nm was subtracted to reduce coherent noise stripes. The 70th band, in the near-IR (1042 nm) was used to mask land by applying a brightness threshold. This algorithm (Eq. (4)) can then be applied to the whole image to produce a distribution map of chlorophyll. Scenes are mosaiced to form a whole-lake chlorophyll map. We smoothed the chlorophyll map with a two-pixel moving average.

Mono Lake example Two AVIRIS overpasses of Mono Lake were obtained just after local solar noon, at 12:07:01 and 12:18:00, on 7 October 1992. The surface of the lake was 1942.7 m above sea level. One optical depth, derived for PAR, was 6 m. The images were atmospherically corrected, combined to create a whole lake scene, and the land was masked. An enhanced true color image was made by combining a red (488 nm), green (557 nm) and blue (667 nm) band (Fig. 2). The two passes did not have same overall brightness. The region inshore of the 6-m contour appears bright because reflection from the bottom is evident. Several smaller gyres (G) are revealed; most appear to have counterclockwise rotation. These gyres were likely formed by persistent winds advecting water across the lake and interacting with topographic features. Foam lines (F) appear as thin white threads and are conspicuous south and southwest of the Paoha Island. These features were likely formed on convergence of downwellings, and are commonly observed while sampling the lake. Differences in Iimnological conditions on opposite sides of the foam lines indicate that different water masses are converging (unpublished data). Application of our algorithm for calculation of chlorophyll from remote sensing reflectance resulted in a distribution map of chlorophyll (Fig. 3). Predicted chlorophyll ranged in concentration from less than 5 to about 50 mg m- 3 . Measurements made at 20 stations on the date of the overflight ranged from II to 21 mg m - 3 , which indicates that the algorithm over and under estimated the probable range of val-

ues. The western side of the lake had higher pigment concentrations than the eastern side. The gyres are less distinct than they are in the enhanced true color image. The most conspicuous gyre is located south of Paoha Island. The limnological challenge is to provide an explanation for these patterns which requires data on physical processes and biological and chemical conditions not available at the time the images were acquired. Ongoing studies at Mono Lake are beginning to provide the kind of information needed to understand the spatial patterns evident in the AVIRIS images. While Iimnological applications of imaging spectrometry have been few (e.g., Hamilton et aI., 1993; Novo et aI., 1995; Schaepman et aI., 1995; Dekker, 1997; Theimann & Kaufmann, 2000), the potential is considerable. Availability of data and of the hardware and software required for image analysis and data processing are improving. Limnologists' awareness of the new technologies offered by remote sensing is becoming more sophisticated. Moreover, the advent of imaging spectrometers on spacecraft will greatly expand the opportunity for obtaining data from lakes throughout the world.

Acknowledgements The AVIRIS field experiments at Mono Lake were a success due to team effort involving M. Embry, G. Dana, D. Heil and R. Jellison and the Sierra Nevada Aquatic Research Laboratory (SNARL). We thank the AVIRIS team at JPL and Edwards Air Force Base. The UCSB Maps and Imagery Laboratory provided access to ERDAS georeferencing software, training, and advice. R. Green provided the interface program for MODTRAN and valuable discussions. D. Roberts reviewed the methods and provided insight on MODTRAN use. The bathymetry contours were overlaid using ENVI and all other image processing was done with IDL, both products of Research Systems, Inc. The work was funded by NASA contracts NAGW-2602, NAGW-5185 and NAS5-51713.

References Abbott. M. R., T. M. Powell & P. 1. Richerson, 1982. The relationship of environmental variability to the spatial patterns of phytoplankton in Lake Tahoe. J. Plankton Res. 4: 927-941. Almanza. E. & J. M. Melack, 1985. Chlorophyll differences in Mono Lake (California) observable on Landsat imagery. Hydrobiologia 122: 13-17.

37 Anderson, G. P., 1. Wang & J. H. Chetwynd, 1995. MODTRAN3: An update and recent validations against airborne high resolution interferometer measurements. In Green, R. O. (ed.), Summaries of the Fifth Annual JPL Airborne Earth Science Workshop, Vol. I. AVIRIS Workshop. JPL Public. 95-1, Jet Propulsion Laboratory, Pasadena, CA: 5-8. Campbell, J. W. & W. E. Esaias, 1983. Basis for spectral curvature algorithms in remote sensing of chlorophyll. Appl. Optics 22: 1084-1093. Carder, K. L., R. G. Stewart, J. H. Paul & G. A. Vargo, 1986. Relationships between chlorophyll and ocean color constituents as they affect remote-sensing reflectance models. Limnol. Oceanogr. 31: 403-412. Carder, K. L, P. Reinersmann, R. F. Chen, F. Muller-Karger, C. O. Davis & M. Hamilton, 1993. AVIRIS calibration and application in coastal oceanic environments. Remote Sens. Environ. 44: 205-216. Curran, P. J., 1994. Imaging spectrometry. Progr. physical Geogr. 18: 247-266. Dekker, A. G., T. 1. Malthus, M. M. Wijnen & E. Seyhan, 1992. The effect of spectral bandwidth and positioning on the spectral signature analysis of inland waters. Remote Sens. Environ. 41: 211-225. Dekker, A. G, T. J. Mallhus & H. J. Hoogenboom, 1995. The remote sensing of inland water quality. In Danson, F. M. & S. E. Plummer (eds), Advances in Environmental Remote Sensing. John Wiley & Sons Ltd., New York: 123-142. Denman, K. L., 1994. Scale-determining biological-physical interactions in oceanic food webs. In Giller, P. S., A. G. Hildrew & D. G. Raffaelli (eds), Aquatic Ecology: Scale, Pattern and Process. Blackwell Sci. Publ., London: 377-402. Galat, D. L. & J. P. Verdin, 1989. Patchiness, collapse and succession of a cyanobacterial bloom evaluated by synoptic sampling and remote sensing. 1. Plankt. Res. II: 925-948. Gastil, M. & 1. M. Melack, Improved atmospheric correction for AVIRIS spectra from inland waters. In Green, R. O. (ed.), Summaries of the Seventh JPL Airborne Earth Science Workshop. JPL Publication 97-21, Vol. 1. Jet Propulsion Laboratory, Pasadena, CA: 141-145. Gitelson, A. A., 1992. The peak near 700 nm on radiance spectra of algae and water: relationships of its magnitude and position with chlorophyll concentration. Int. 1. Remote Sensing 12: 33673373. Gitelson, A. A. & K. Ya. Kondratyev, 1991. Optical models of mesotrophic and eutrophic water bodies. Int. J. Remote Sensing 12: 373-385. Gollerman, H. L., R. S. Clymo & M. A. M. Ohnstad, 1978. Methods for Physical and Chemical Analysis of Fresh Waters. IBP Handbook No.8, 2nd edn. Blackwell Sci. Pub!., Oxford. Gordon, H. R. & A. Morel, 1983. Remote Assessment of Ocean Color for Interpretation of Satellite Visible Imagery: a Review. Springer, New York. Gordon, H. R., D. K. Clark, J. W. Brown, O. B. Brown, R. H. Evans & W. Broenkow, 1983. Phytoplankton pigment concentrations in the Middle Atlantic Bight: Comparison of ship determinations to CZCS estimates. Appl. Optics 22: 20-35. Green, R. 0.,1991. Retrieval ofreflectance from AVIRIS measured radiance using a radiative transfer code. In Proc. Third AVIRIS Workshop, JPL Publication 91-28. Jet Propulsion Laboratory, Pasadena, CA: 200-210. Green, R. O. (ed.), 1995. Summaries of the Fifth Annual JPL Airborne Earth Science Workshop, Vol. I. AVIRIS Workshop. JPL Public. 95-1, Jet PropUlsion Laboratory, Pasadena, CA.

w.

Hamilton, M. K., C. O. Davis, W. 1. Rhea, S. H. Pilorz & K. L. Carder, 1993. Estimating chlorophyll content and bathymetry of Lake Tahoe using AVIRIS data. Remote Sens. Environ. 44: 217230. Harris, G. P., 1986. Phytoplankton Ecology. Chapman and Hall, London: 384 pp. Hoge, F. & R. Swift, 1983. Airborne dual laser excitation and mapping phytoplankton photopigments in a Gulf Stream warm core ring. Appl. Opt. 22: 2272-2281. Jellison, R. & 1. M. Melack, 1993. Algal photosynthetic activity and its response to meromixis in hypersaline Mono Lake, California. Limnol. Oceanogr. 38: 818-837. Johnson, H. K. & R. O. Green, 1995. AVIRIS user' guide. In Green, R. O. (ed.), Summaries of the Fifth Annual JPL Airborne Earth Science Workshop, Vol. 1. AVIRIS Workshop. JPL Publication 95-1, Jet Propulsion Laboratory, Pasadena, CA: 105-108. Kirk, 1. T. 0., 1986. Light and Photosynthesis in Aquatic Ecosystems. Cambridge University Press, Cambridge: 40 I pp. Kneizys, F. X. et aI., 1988. User's guide to LOWTRAN 7, Air Force Geophysics Laboratory. Lathrop, R. G. & T. M. Lillesand, 1986. Use of thematic mapper data to assess water quality in Green Bay and central Lake Michigan. Photogram. Eng. Remote Sens. 52: 671-680. Lee, Z. & K. L. Carder, 2000. Band-ratio or spectral-curvature algorithms for satellite remote sensing? Appl. Opt. 39: 4377-4380. Melack, 1. M., 1985. Interactions of detrital particulates and plankton. Hydrobiologia 125: 209-220. Melack, 1. M. & S. A. Pilorz, 1990. Reflectance spectra from eutrophic Mono Lake, California, measured with the Airborne Visible and Infrared Imaging Spectrometer (AVIRIS). Int. Soc. Opt. Eng. Imaging Spectroscopy Terrestrial Environment 1298: 202-212. Melack, 1. M. & M. Gastil, 1994a. Airborne visible imaging spectrometry applied to limnology: chlorophyll variation in Mono Lake. In Proc. of International Symposium on Spectral Sensing Research, San Diego, CA: 691-695. Melack, 1. M. & M. Gastil, I 994b. Comparison of spectral feature algorithms for remote sensing of chlorophyll in eutrophic lakes. In Proc. Intern. Geosci. Remote Sens. Sympos., Vol. IV: 23632365. Melack, 1. M. & R. Jellison, 1998. Limnological conditions in Mono Lake: contrasting monomixis and meromixis in the 1990s. Hydrobiologia 384: 21-39. Mittenzwey, K. H., 1992. Determination of chlorophyll of inland waters on the basis of spectral reflectance. Limnol. Oceanogr. 37: 147-149. Mobley, C. D., 1994. Light and Water: Radiative Transfer in Natural Waters. Academic Press, New York: 592 pp. Neill, W. E., 1994. Spatial and temporal scaling and the organization of limnetic communities. In Giller, P. S., A. G. Hildrew & D. G. Raffaelli (eds), Aquatic Ecology: Scale, Pattern and Process. Blackwell Sci. PubL London: 189-231. Novo, E .. M. Gastil & 1. M. Melack, 1995. An algorithm for chi orophy II using first difference transformation of AVIRIS reflectance spectra. In Green, R. O. (ed.), Summaries of the Fifth Annual JPL Airborne Earth Science Workshop, Vol. 1. AVIRIS Workshop. JPL Public. 95-1, Jet Propulsion Laboratory, Pasadena, CA: 121-124. O'Reilly, 1. E., S. Maritorena, B. G. Mitchell. D. A. Siegel, K. L. Carder, S. A. Garver, M. Kahru, C. McClain, 1998. Ocean color chlorophyll algorithms for SeaWiFS, J.Geophys. Res. (Oceans) 103: 24937-24953. Sarture, C. M. et aI., 1995. Airborne VisiblelInfrared Imaging Spectrometer (AVIRIS): sensor improvements for 1994 and 1995. In

38 Green, R. 0, (ed.), Summaries of the Fifth Annual JPL Airborne Earth Science Workshop, Vol. 1. AVIRIS Workshop. JPL Public. 95-1, Jet Propulsion Laboratory, Pasadena, CA: 145-148. Schaepman, M. et aI., 1995. Extraction of ozone and chlorophyll-a distribution from AVIRIS data. In Green, R. 0. (ed.), Summaries of the Fifth Annual JPL Airborne Earth Science Workshop, Vol. I. AVIRIS Workshop. JPL Public. 95-1. Jet Propulsion Laboratory, Pasadena, CA: 149-152. Smith, R. c., 1981. Remote sensing and depth distribution of ocean chlorophyll. Mar. Ecol. Prog. Ser. 6: 359-361. Smith, R. C. & K. S. Baker, 1982. Oceanic chlorophyll concentra-

tions as determined by Nimbus 7 Coastal Zone Color Scanner imagery. J. mar. BioI. 66: 269-279. Thiemann, S. & H. Kaufmann, 2000. Determination of chlorophyll content and trophic state of lakes using field spectrometer and IRS-l C satellite data in the Mecklenberg lake district, Germany. Remote Sens. Environ. 73: 227-235. Vane, G. & A. F. H. Goetz, 1993. Terrestrial imaging spectrometry: Current status, future trends. Remote Sens. Environ. 44: 117126. Vane, G. et aI., 1993. The Airborne Visible/Infrared Imaging Spectrometer (AVIRIS). Remote Sens. Environ. 44: 127-143.

Hydrobiologia 466: 39-43,2001. J.M. Melack, R. Jellison & D.B. Herbst (eds), Saline Lakes. © 2001 Kluwer Academic Publishers.

39

Re-appearance of rotifers in hypersaline Mono Lake, California, during a period of rising lake levels and decreasing salinity Robert Jellison l ,*, Heather Adams 2 & John M. Melack l ,3 I Marine Science Institute, University of California, Santa Barbara, CA 93106, U.S.A. 2CoUege of Creative Studies, University of California, Santa Barbara, CA 93106, U.S.A. 3Department of Ecology, Evolution, and Marine Biology, University of California, Santa Barbara, CA 93106, U.S.A.

Key words: Mono Lake, hypersaline, rotifer, Hexarthrajenkinae, Brachionus plicatilis

Abstract The surface elevation of Mono Lake, California, rose "'2 m and mixed-layer salinities declined about 5 g kg- 1 during the 3 years (1995-1997) following the decision to restrict water diversions out of the Mono Basin. Abundant (18000 m- 2) Hexarthrajenkinae de Beauchamp were noted in pelagic samples in October 1997 after three decades of absence or very low abundance. Abundance subsequently increased to 100000 m- 2 in December 1997 before declining to low numbers through 1998 and 1999. The re-appearance of Branchionus plicatilis Muller in pelagic samples occurred in September 1998. B. plicatilis areal abundance increased to'" 15000 m- 2 in OctoberDecember of both 1998 and 1999 but was low throughout the rest of the year. Both rotifers were noted in nearshore ponds, but were only abundant in those with salinities below 53 g kg-I. During 1998-1999 when the salinities of the upper water column were 73-75 g kg-I, less saline shoreline habitats may have been seeding the offshore rotifer populations.

Introduction Zooplankton species diversity is severely reduced in hypersaline (>50 g kg-I) lakes compared to less saline and freshwater lakes (Hammer, 1986). Mono Lake, California, U.S.A., is a large, hypersaline saline, lying in a hydrologically closed high-desert basin just east of the Sierra Nevada within the North American Great Basin. Throughout much of the 20th century, the surface elevation of Mono Lake declined and salinity increased, first due to extended drought early in the century and then due to diversion of streams out of the basin beginning in 1941. Over the next 40 years, the lake's salinity doubled as its volume was halved. During limnological surveys of Mono Lake in 1959, 1963 and 1964, when the salinity was 60-70 g kg-I, Mason (1967), noted the presence of several species of protozoans; the two rotifers, Hexarthra jenkinae and Brachionus plicatilis; and the endemic brine shrimp, Artemia monica. Subsequent to these observations, the

* Corresponding author. E-mail: [email protected]

lake continued to shrink and the salinity increased to ",82 g kg- I in 1976, by which time, rotifers were not reported in plankton samples (Winkler, 1977). During the next 16 years, the surface elevation of the lake continued to decline reaching its historic low of 1942 m during winter 1981-1982, at which time, the salinity was "'92 g kg- I (Jellison & Melack, 1993b). In 1994, the California State Water Resources Control Board decided to raise the surface elevation of Mono Lake to 1948 m as part of a general restoration plan (Decision #1631) by restricting water diversion by the City of Los Angeles. By October 1997, the lake had risen ",2 m to 1945.2 m and the mixolimnetic salinity decreased by 5 g kg- 1 to 76 g kg- 1 when the re-appearance of rotifers in plankton samples was first noted. Here, we describe the pelagic abundance of Hexarthra jenkinae and Brachionus plicatilis during late 1997 through 1999. We also sampled nearshore less saline ponds, some of which contained abundant rotifers.

40 + 6

Mono Lake 1946 m asl

rotifer samples were collected with both 120- and 5011m nets. Both vertical tows extended to 20 m depth, but the 50-11m net was raised at approximately 10 cm s-I. The efficiency of vertical net tows varies as a function of mesh size, shape, ascent rates and the plankton community. Data reported here assume a 70% net efficiency derived from previous comparisons of Artemia abundance collected by vertical net tows and a Schindler-Patalas trap (Lenz, 1980).

Results Om

5,ODom

Re-appearance, individual size and net efficiency

Figure 1. Pelagic sampling stations on Mono Lake (solid circles) and locations of adjacent nearshore ponds (crosses; ST is the South Tufa pond). All ponds were immediately adjacent to the lake except for pond #6.

Methods Zooplankton samples have been collected at approximately monthly intervals throughout the year at 1220 stations distributed around the lake (Fig. I) from 1982 to the present. Samples were taken with a plankton net (I xO.30 m diameter, 120-l1m Nitex mesh) towed vertically through the water column at approximately 0.5 m s-I. Samples were preserved with 5% formalin in lakewater, and counted under a stereo microscope (x6 power). While samples were not explicitly examined for the presence of rotifers, abundant rotifers were first noted in October 1997 samples. Therefore, beginning in October 1997 samples were saved following enumeration of Artemia and examined for thc presence of rotifcrs. Artemia were removed and samples concentrated to 25-70 ml depending on the abundance of rotifers by filtering through a fine sieve and then rinsing individuals into a sample container. Samples were thoroughly mixed and then three or four I-ml subsamples removed for counting under an inverted microscope (Olympus BX40) with an attached video camera and monitor. In October and December 1997 samples, individual lengths were measured using Optimus 5.2 software specifically designed for the microscope and camera. Beginning in May 1998, a 50-11m net lowered to a depth of 5 m was used to collect a separate sample for rotifer enumeration. Low abundances during May and June, prompted a change in sampling protocol. Beginning in July and continuing through mid-1999,

The re-appearance of the rotifer, H. jenkinae, in vertical net tow samples from Mono Lake was first noted on 27 October 1997 when their mean abundance at three pelagic stations (Stations I, 6, 8; Fig. 1) was 18000 m- 2 . The mean lengths of 187 individuals from October and 343 individuals from December were nearly identical on the two sampling dates, 233 and 229 11m, respectively. The frequency distribution of individual lengths indicated 90% of the individuals collected were larger than 180 11m, and thus, we assume samples collected with the 120-11m net capture nearly all individuals. As rotifer sampling is often conducted with a 50-11m net, we sampled with both 50- and 120-l1m mesh nets from July 1998 through 1999. H. jenkinae abundance was too low throughout this period to assess the different net efficiencies, but B. plicatilis appeared in plankton samples in September 1998. A comparison of B. plicatilis abundance in paired samples indicated the 120-l1m net hauled vertically as "-'0.5 In S 1 collected significantly more (Wilcoxon signed runk test with n= 18; P80 mg 1-1) and bacterioplankton abundance (> 1Q lOcells 1-1). Bacterioplankton cells tend to be large (> I /Lm) with many chain-forming organisms, particularly when A. monica is not present. This is also consistent with reports from other soda lakes (Zehr et a1., 1987; Duckworth et ai., 1996). However, relatively little is known about the types of microorganisms dwelling in the lake and their phylogenetic diversity, taxonomy, ecology or ecophysiology. Bacteria isolated from the lake's surface waters grow best at pH 9.7 and are able to use organic osmolytes as carbon sources (Diaz & Taylor, 1996). Alkaliphilic bacteria are of interest from physiological and industrial perspectives (Horikoshi & Akiba, 1982; Grant el ai., 1990), and recent studies have stressed the phylogenetic diversity of bacteria isolated from soda lakes (Duckworth et ai., 1996). However, little work has been done on the in situ phylogenetic diversity of bacterial populations from soda lakes, or in ascribing aspects of the biogeochemical cycles of these lakes to these bacteria. Meromixis is a recurring event in Mono Lake's limnology, and several papers have described the consequences of meromixis and its breakdown to Mono Lake's ecology and geochemistry (Jellison et a1., 1993a,b; Miller et ai., 1993; Jellison et aI., 1996; Melack & Jellison, 1998). A long period of meromixis in the 1980's was initiated by runoff from the EI Niiio winter of 1982-1983. This event persisted until 1988, when drought in California contributed to the weakening of salinity stratification, initiating a period of monomixis that persisted through the dry winter of 1993-1994. Meromixis was re-established in 1995 as a result of changes in water management policy and high runoff during the El Niiio winter of 19941995. A hydrodynamic mixing model (Jellison et a1., 1998) predicts that the lake will exhibit prolonged meromixis. The re-establishment of meromixis in Mono Lake in 1995 afforded us an opportunity to investigate the response of the lake's microbial community to this condition. We were able to characterize the microbial community that had developed during 6 years of monomixis and to compare it with the community that developed within months of the re-establishment of

meromixis. These ana1yses provide a starting point for future studies of the response of the lake's microbia1 community to prolonged meromixis. We also began characterizing the phylogenetic composition and vertical distribution of Mono Lake's microbia1 community.

Materials and methods Sample collection

Water samples were collected at Station 3 [Fig. 1; Station 9 in previous papers on Mono Lake such as Jellison & Melack (1993a), Melack & Jellison (1998) and Joye et al. (1999)], located offshore in 30-m deep water. Water column profiles and samples were taken at midday on July 20, 1994; April 9, 1995; and July 20, 1995. Temperature, pressure, photosynthetically active radiation (PAR at 400-700 nm, Licor 41i sensor) and in vivo chlorophyll fluorescence (relative fluorescence, RF, in arbitrary units; SeaTech fluorometer) were obtained with a SeaBird SeaCat profiler on the July expeditions. Conductivity was not measured because this instrument was not equipped with a high-sa1inity conductivity sensor. Oxygen profiles were taken with a polarographic oxygen sensor (YSI) equipped with a Clarke-type electrode. Beam attenuation (%T) was measured with a Martek XMS transmissometer equipped with a 25-cm path length sensor. Conductivity and temperature data for April 1995 were supplied by R. Jellison and were taken with a SeaBird SeaCat conductivity, temperature and pressure logger (CID) equipped with a special highsalinity conductivity sensor. Vertical distributions of chlorophyll, ammonium, methane, ammonia oxidation rate and methane oxidation rate at this station and on these sampling dates are given in Joye et ai. (1999). Water samples (500-1000 ml) were taken at discrete depths using a Niskin water sampler. The samples were poured into dark plastic bottles and held in a cooler until they were processed (within 4 h of collection). The samples were filtered through either 47 mm diameter Millipore GV filters (0.22 /Lm pore size, 80 kPa vacuum) or Millipore Sterivex filter cartridges (0.22 /Lm pore size, 100 kPa pressure) to collect microbial biomass for subsequent DNA extraction. Flat filters were placed in disposable 15-ml plastic centrifuge tubes (Falcon), bathed with about 3 ml of extraction buffer (40 mM EDTA; 50 mM Tris, pH 8.3; 0.75 M sucrose) and frozen on dry ice. Water

48 remaining in Sterivex filter cartridges was expelled with pressurized air delivered with a syringe; 1.8 ml of extraction buffer was then added to the cartridges, which were capped and frozen on dry ice. All filters were stored at -20°C until processed.

DNA extraction and PCRJDGGE DNA was extracted from the filters and used as the template for polymerase chain reaction (PCR) amplification of an approx.imately 200-bp region of the 16S rRNA gene. The mixed PCR products were then resolved via denaturing gradient gel electrophoresis (DGGE). The procedures used were as described in detail in Murray et al. (1996) and Ferrari & Hollibaugh (1999). Briefly, material retained by filters was digested with lysozyme, proteinase K and 1% sodium dodecyl sulfate at 60°C. The crude extract was purified by phenol/chloroform extraction and ultrafiltration (Centricon) or ethanol precipitation. PCR used I-lOng of purified DNA as template and Taq polymerase with 17-mer primers at positions 340-356 (primer 356f, eubacterial) and 517-533 (primer 517r, universal) of the Escherichia coli gene (Brosius et aI., 1978). These primers (which were previously erroneously reported as at 341-358 and 517-534) flank one of the hypervariable regions of the 16S rRNA gene (the variable 3, or v3, region) and amplify the 16S rRNA gene of most species of Bacteria (Muyzer et aI., 1993; Murray, 1994). A 40-bp GC clamp was added to the 5' end of primer 356f to improve DGGE resolution (Myers et aI., 1985, 1987). PCR used the hot-start and touchdown protocols. Amplifications, in 100-fLl volumes, were run for 30 cycles. Extraction blanks (filters or filter cartridges through which no water had passed) served as negative PCR controls. Genomic DNAs from Clostridium perfringens (Sigma Genomic Ultra Pure, D5l39) and Bacillus thuringiensis (Sigma High Molecular Weight, D3l71) were used as positive PCR controls and as DGGE standards. DGGE was performed using 6.5% polyacrylamide gels containing a 40-70% gradient of deionized formamide-urea. Electrophoresis was carried out at 60 °C for 15 h at 75 V using a CBS Scientific DGGE apparatus (gel format: 14.5 cm wide x 20 cm long x 0.75 mm thick). Approximately 750 ng of PCR product was loaded onto each lane. Standards were run in at least three lanes (sides and middle) of every gel. All of the samples from a given sampling date were run on the same gel in the order of depth so

that vertical distributions of ribotypes (bands in DGGE gels representing different rDNA sequences) could be ascertained more easily. DGGE gels were stained with ethidium bromide. Gel images were displayed and recorded using a UVP Corp. Model 7500 Gel Documentation System. The images (640 pixels wide and 480 pixels along lanes, 256 grayscale) were first adjusted for brightness and contrast, then viewed at different exposure lengths to adjust for differences in staining intensity and properties of the banding patterns and finally recorded digitally.

Processing DGGE bands for sequencing Standards and a subset of samples selected by inspection of the DGGE gels were amplified a second time from the corresponding genomic extracts and run on a DGGE gel as described above, except that PCR product from two reactions was combined and loaded into each lane so that bands contained more DNA. We then sampled DNA from 28 dominant bands in this gel using the peel-off procedure (Wu et aI., 1996) for capturing bands from polyacrylamide gels. This method allows rapid tagging of multiple bands per lane, as typical of a DGGE run, without excessive exposure of the desired material to ultraviolet light from a transilluminator. Lanes were exposed one at a time by placing an exposed X-ray film with a slit cut to accommodate about half of the length of a lane under the gel. A thin plastic film (e.g. Saran wrap) placed between the polyacrylamide gel and the X-ray film permits the gel to slide over the film without damage. Bands were tagged by putting slivers of 3 MM Whatman filter paper cut to size on top of each chosen band. The entire gel was then covered with 3 MM paper and dried (on a gel dryer) under vacuum at 80°C until the plastic wrap could be easily removed from the surface of the gel (at least 1 h). The gel was then archived in a dry place at ambient temperature. Bands were labeled on the gel with a pen next to the paper strips and cross-referenced to a printed image of the stained gel. To sample a band, a flame-sterilized scalpel was used to cut through the dried gel along the edges of the paper sliver; then flame-sterilized tweezers were used to remove the sliver of gel-covered filter paper. DNA was eluted from the gel plus paper strip in a 1.5-ml micro centrifuge tube containing 60-70 fLl of water by heating briefly (30 s) in a microwave oven and then stored at 4 0C. Fragments were purified by reversible immobilization on magnetic particles according to DeAngelis

49 et al. (1995). This method yielded uniformly high recovery (>90%), which is of key importance given the small amounts of DNA contained in many of the DGGE bands. Fragments were eluted from the magnetic beads in 20 ILl of 10 mM Tris-HCl, pH 8, followed by evaporation of the supernatant fluid (after separation from the beads) by centrifugation under vacuum (Speed Vac Plus SCllOA). The product was dissolved in 5-10 ILl of sterile water by heating at 50 °C for 5 min and then stored at 4 0c. Sequencing DGGE bands

Cycle sequencing used primers 356f and 517r and the ABI PRISM™ Ready Reaction dye-terminator cycle sequencing kit (PE Applied Biosystems) for the ABI PRISM™ 377 DNA sequencer from PE Applied Biosystems. Cycling was performed in the PE Applied Biosystems GeneAmp® PCR System 9600 at one-half the volume recommended (PE Applied Biosystems, August 1995) and with 80 cycles to compensate for low template concentrations (R. Zebell and C. Orrego, unpub. observ.). Amplification reaction conditions per cycle were as follows: 10 s at 96°C (denaturation), 5 s at 57°C (annealing) and I min at 70 °C (extension). A ramp to 4 °C and hold until purification followed the last cycle. Products were purified by ethanol precipitation prior to electrophoresis in the ABI PRISM™ 377 sequencer. Sequences were determined from pherograms using DNA Sequence Analysis software version 2.1.2 (PE Applied Biosystems). Sequence editing and alignments were performed using the Sequencher 3.0 software (Gene Codes Corporation). The portion of the pherogram obscured by noise from dye-terminator 'clouds' (PE Applied Biosystems) that tend to become more pronounced at high cycling was not used. Sequence ambiguities identified by the instrument were edited manually observing the rules concerning peak height patterns obtained with AmpliTaq® DNA polymerase FS dye-terminator sequencing. We compared the sequences we obtained to those in the GenBank database with the BLAST program (Altschul et a!., 1990). Phylogenetic analyses were conducted by aligning 16S rDNA sequences from DGGE bands with similar (as indicated by BLAST) sequences from the database using the Genetic Computer Group package (Madison, Wisconsin). Phylogenetic trees were constructed using JukesCantor distances and the neighbor-joining method (PHYLIP package; Felsenstein, 1993). Branching pat-

tern robustness was tested by bootstrap analysis (100 replicates). Picocystis sequences Cultures of Picocystis salinarum strains SSFB (South San Francisco Bay, CA, U.S.A., saltern), L7 (Mono Lake, CA, U.S.A.) and IM214 (alkaline, hypersaline lake in Inner Mongolia, People's Republic of China) were provided by Dr R. A. Lewin (University of California at San Diego). Cells were concentrated by centrifugation at 14000 rpm for 10 min at 4°C, the supernatant was removed, and then the pellet was suspended in lysis buffer and nucleic acids were extracted as described above. DNA was concentrated with Centricon-IOO concentrators (Amicon). Nuclear small subunit (18S) rRNA genes were amplified using the primers (GOl-GlO, Table 1) described by Andersen et al. (1993) and Saunders & Kraft (1994). Internal primers (NI-N5, Table I) were used to join nonoverlapping fragments. Amplification reaction conditions were as follows: initial denaturation for 3 min at 94°C; 30 cycles of 30 s at 94°C, 45 s at 55°C and 10 min at 72 °C; a final extension for 10 min at 72 °C. PCR products were purified using the Wizard® PCR Preps DNA purification system (Promega, Madison, Wisconsin). Cycle sequencing used primers GO I-G 10 and internal primers NI-N5 (Table I) and was performed at the Molecular Genetics Instrument Facility, University of Georgia, using the ABI PRISM™ dyeterminator kit. Sequences were read on ABI 373 or 377 DNA sequencers. Final sequences were obtained by sequencing and aligning up to seven fragments (Table 2). Picocystis small subunit rRNA gene sequences (chloroplast and nuclear) obtained in this study have been deposited in GenBank under accession numbers AF125167, AFI25173, AFI25174, AF153313 and AF153314. Sequences obtained from DGGE gels were not submitted to GenBank because they were short ( 30 lim

~

1995

Figure 4. Seasonal changes in environmental factors and brine shrimp abundance in the Great Salt Lake during 1994-95. (A) Temperature in the mixed layer (2 m). (B) Epilimnetic chlorophyll a levels. Open symbols show the percentage of the chlorophyll that was in the > 30 !Lm size fraction. (C) Transparency measured with a Secchi disk. (D) Mean water colunm densities of Artemia franciscana. Densities from Station 1 on 3 October were not included in the calculation of the mean because of extraordinarily high densities (85 1-1). Had this station been included densities would have been 12.61- 1. Means and S.E. (where shown) of 5-12 stations (average

=9).

exponential phase, and then slowing once the individuals began reproducing. Reproduction began (nauplii produced) in some individuals within 7 d when the mean length of the cohort was near 6 mm. Limnological variables

Whole-lake mean epilimnetic temperatures below the diel thermocline (2 m) were near 21°C at the start of the study in June and rose to 27 °C in August before declining throughout the late summer and fall (Fig. 4A). Temperatures approached 1 °C in December and January. The lake warmed rapidly in the spring and had reached 8 °C by the time we reinitiated sampling in mid-March 1995. By mid-June, temperatures were near 20°C, only slightly cooler than in June 1994. Temperature and oxygen profiles taken near the center of the lake (Station 2) indicated that the lake was thermally and chemically stratified below 7 m between March and early May, but by June it had mixed to the bottom.

Mean salinities in the lake during the study were within a range allowing good growth and reproduction of Artemia (Stephens, 1990). Mean salinity in the integrated water column samples in the main lake was 149 g I-I at the start of the study in June 1994, rose steadily through the summer and fall, and peaked at 160 g I-lin October. By April 1995, spring runoff had decreased salinity to 144 g I-I, and it continued to decline to 132 g I-I by mid-June. Stations 1, 2a and 12 near the 'estuaries' in the lake often had salinities 1520 g 1-1 lower than other sites. Mean total phosphorus concentrations measured in April-June 1995 were 8.2 {Lm. Approximately 52% of the total was in the form of soluble reactive phosphorus with a mean concentration of 4.3 {LM (range 2.5-6.4 {LM). The 0.6--150 {Lm fraction (seston) contained 27% of the phosphorus (2.2 {LM) and the> 150 {Lm fraction (primarily Artemia) contained 21 % (1.7 {Lm). Phytoplankton chlorophyll concentrations in June 1994 were 0.7m

I 0.5-0.7 m

o

>7m

Ii 4-7m

• 4 mg I-I (Fig. 3C). This calculation assumed a maximum depth of IS m, that oxygen levels below 14 m were the same as those at 14 m, and that dissolved oxygen concentration at a given depth nearshore was the same as the mean oxygen concentration of water at the same depth at the mid-lake stations. Sulfide oxidation has potential for deoxygenating the water column during mixing events following stratification. We quantified this using oxygen and sulfide profiles for 31 July and 16 August 1999, a mixing and deoxygenation event apparently having occurred between the two dates. pH on these dates was 7.6 - 8.3, and in this range sulfide is found primarily as sulfide ion (HS-) (~80%, pKA 6.88 for 35 g I-I) with hydrogen sulfide (H2S) making up ~ 20% and S2- making up less than 1% (Snoeyink & Jenkins, 1980; Millero, 1986). Proposed reactions of the oxidation of sulfide have three products; thiosulfate, sulfite and sulfate. With time, both thiosulfate and sulfite are converted to sulfate. In addition, experimental data has shown that when oxygen is the limiting reagent in the sulfide oxidation reaction , as in the Sea on the dates sulfide was measured, greater than 50% of the reaction products of sulfide oxidation are sulfate (Cline & Richards, 1969). Therefore, consumption of dissolved oxygen in the lake by oxidation of sulfide was calculated using the reaction: HS- + 202 -> S04 2- + H+ (Chen & Morris, 1972) Three assumptions were made in this calculation. First, that hydrogen sulfide is converted to sulfide ion before oxidization during a mixing event. This assumption makes our estimate of oxygen consumption conservative as more dissolved oxygen is consumed in the oxidation of hydrogen sulfide than in the oxidation of sulfide ion. Second, that all sulfide is oxidized to sulfate. Third, that the dissolved oxygen and sulfide profiles prior to the mixing event were similar to those of 31 JUly.

164

.I:

0. Q)

o

100

r:: Q) u .... Q)

Q.

80

1997

Percent of lake bottom at less than _ mg02 r1

1999

C

60

I

I

40 20 0

J

F M A M J

J A

D J

F M A M J

J A S

o

N D J

F M A M J

J A SON D

Figure 3. Salton Sea temperature and oxygen conditions, 1997-1999. (A) Water temperature. (B) Dissolved oxygen concentrations. (C) Percent of lake bottom exposed to different dissolved oxygen concentrations. Shaded areas indicate extent of anoxic or near anoxic conditions (> I mg 02 1- 1).

Specific conductance values (mS cm- I at 25 DC) were converted to salinity (sum of major ions, in g 1-1) using the relation: S = 17.102 -0.233K+ 0.013K2 This empirical relation was determined using specific conductance and major ion data for 24 samples from the Sea that ranged in salinity from 20 to 42 g I-I (Setmire et aI., 1993; Hurlbert & Detwiler, unpubl. data). Water column stability is determined by both salinity and temperature gradients. Calculations were performed for S-4 on 23 June 1999 to compare the contribution of the temperature gradient and the salinity gradient to stratification stability on this date. Stability calculations used the equation developed by Schmidt as modified by Idso (1973) and assumed eight 1m thick strata, each isothermal. Density-temperatureconductivity relations have not been developed for Salton Sea water. To perform stability calculations,

standard seawater densities (Kennish, 1989) were used to approximate density changes with temperature. Stability was estimated for three scenarios; first, using only temperature data for S-4 and assuming that salinity was invariate with depth (observed thermal gradient only); second, using an extreme hypothetical situation (35 DC at 0-4 m, 12 DC at 4 - 6 m) and assuming salinity was invariate with depth (hypothetical high thermal gradient only); and third using the observed salinity profiles for S-4 and assuming an isothermal water column (observed salinity gradient only). The second scenario provides a point of reference to illustrate the dominant role of salinity gradients when present; 35 DC approximates the lakes maximum summer temperature and 12 DC its minimum winter temperature Analysis of meteorological data

Measurements of solar insolation, air temperature, wind speed and direction for 1997-1999 were ob-

165 Table 2. Location of California Irrigation Management Infonnation System (CIMIS) weather stations Station Approximate Elevation location (m)

127 128 141 154

Salton City Wister Mecca North Shore

-68 -68 -55 -61

Distance to Geographic coordinates nearest (N latitude; W longitude) shoreline (km) 2

0.5 5 0

33° 33° 33° 33°

19' 38"; I3' 12"; 32' 17"; 32' 56";

115° 115° 115° 115°

57' 00" 34' 48" 59' 30" 54' 58"

tained from four California Irrigation Management Information System (CIMIS) meteorological stations located on the perimeter the Salton Sea (Fig. I, Table 2). These stations are maintained by the Department of Water Resources (DWR) and record meteorological data each minute using an onsite datalogger. Hourly means are calculated onsite and transmitted to a CIMIS public database for long-term storage and public access. Wind speed was measured using an anemometer (accurate to 0.4±0.1 m s-I), mean daily air temperature (thermistor accurate to 0.1 °C), and daily solar insolation (pyranometer accurate to 20±4 Watts m- 2). These measurements, averaged over the four stations, are presented. In the case of wind speed, we present seven-day running means calculated from the daily four station means.

Results Climate and weather

Local climatic conditions, especially wind, air temperature and solar insolation, were important factors that determined mixing and water temperature (Fig. 2A, B, C, E). The region experienced large seasonal differences in air temperature with a monthly mean air temperature of 13.5 °C in January and of 33.5 °C occurring in June or August (Fig. 2B). Solar insolation followed a similar pattern with a maximum occurring one month earlier than the maximum air temperature (Fig. 2A, B). Prevailing winds in the Salton Sea basin were from the north (300°-360°), roughly paralleling the long axis of the lake. Infrequent wind events originated in the opposite direction (120°-180°), also along the long axis of the lake. Spring consistently had the highest 7-day running mean wind speeds (Fig. 2C). Although winter and fall generally had lower 7-day running mean wind speeds, storm events did

occur. Lowest mean wind speed occurred in summer (June-September) in 1997 and 1998 and in fall in 1999. Notable differences in wind speed and air temperature were measured among the three years of the study. Spring (March-May) 1998 had higher 7-day running mean wind speeds than did spring of 1997 or 1999 (Fig. 2C). Between April and June 1998, 7day running mean wind speed never dropped below 2.8 m s-I and had the highest values during the study (4.3 m s-I). In summer 1998 (June-August), 7-day running mean wind speeds were always lower than 3 m s-I, whereas both 1997 and 1999 had summer 7day running mean wind speeds above 3 m s-I. Mean spring air temperature (March-June) in 1997 was 3 °C higher than in 1998 and 2 °C higher than in 1999 (Fig. 2B). Maximum mean air temperature was reached on 3 July in 1998, but not until 6 September in 1997 and 18 August in 1999. Salinity and lake elevation

Salinity and lake elevation varied seasonally and were inversely related. Salinity ranged between 41.0 and 44.7 g I-I with the maximum values occurring in November 1998 and December 1999 (Fig. 2D). Maximum salinity coincided with lake elevation minima in November of each year. Lake elevation varied between -68.8 and -69.2 m below sea level, with a mean of -69.0 m (Fig. 2D). Highest lake elevation occurred in early summer. Evaporation in the region is high then (Blaney, 1955; Hely et aI., 1966), but inflows of agricultural wastewaters to the Sea are highest during the preceding spring months (April-June). Thermal and mixing regimes

As the Salton Sea warmed from January to August, periods of stratification were sporadically interrupted by mixing that increased bottom water temperature throughout the warming period (Fig. 3A) and also distributed oxygen downward to bottom waters (Fig. 3B). Maximum water temperature was reached two months earlier in 1998 than in 1997 or 1999. Maximum mean water column temperatures were 30.8 °C on 9 September 1997, 32.8 °C on 3 July 1998, and 30.1 °C on 28 August 1999 (Fig. 2E). Higher wind speeds in 1998 relative to 1997 or 1999 probably were responsible for the accelerated heating during May-June 1998 even though, with respect to air temperature, it was the coolest spring of the triennium

166 (Fig. 2C, E). By mixing downward of heat taken up by surface waters, wind-generated turbulence can diminish heat loss via backradiation and increase heat gain via conduction. The Sea cooled from September to January (Fig. 3A). Except for daily surface (0-2 m) warming (Fig. 3A), the midlake water column tended to be isothermal during this period. It presumably mixed daily or nearly daily due to convectional circulation driven by conductive and evaporative cooling of surface waters and by periodic windy conditions (Fig. 2E). Seasonal lows for mean water column temperature always occurred in January and were 13.9 °C on 27 January 1997, 13.2 °C on 6 January 1998 and 13.8 °C on 25 January 1999. Dissolved oxygen regime

Dissolved oxygen concentrations ranged from 0 to >20 mg 1-1 with surface waters typically containing >4 mg 02 1-1 during the day (Fig. 3B). When the Sea was stratified in spring and summer, an oxygen gradient existed between surface (6-> 20 mg 1-1 ) and bottom waters (0-3 mg 1-1) (Fig. 3B). Anoxic conditions developed rapidly in bottom waters during periods of stratification. From September to January the entire water column usually had> 3 mg 02 1-1. The entire water column occasionally became anoxic, or nearly so, in late summer. We observed this on 12 September 1998 and 16 August and 25 September 1999 (Fig. 3B). It may also have occurred just prior to September 6 1997. These events occurred during or just after a wind-driven mixing event following a period of stratification. Mixing during the September events was most likely facilitated by reduced stability and increased circulation following cooling of surface waters (Fig. 2B, E). The September 1998 mixing event probably occurred during windy conditions on 9 September 1998, 3 days prior to sampling (mean daily wind speed of 3.6 m s-1 and maximum mean hourly wind speed of 8 m s-1 , 6 September). In August 1999, a wind event occurred two days (mean = 3 m s-1 , max =7.6 m s-l, 14 August) and in September 1999 3 days prior to detection of the hypoxic water column (mean =3.6 m s-l, max = 5.3 m s-l, 22 September). Following these events, photosynthesis and shallow mixing replenished dissolved oxygen in surface waters (Fig. 3B). Dissolved oxygen injected into bottom waters was rapidly consumed and following most spring and summer mixing events, the deeper portions of the lake bottom remained anoxic (Fig. 3B). In fact, large portions (60-100%) of the water column

and lake bottom were anoxic or nearly so on most sampling dates in spring and summer of all years (Fig. 3B, C). Extensive hypoxia or anoxia of the lake bottom was observed as early in the year as March (1997) and persisted until early October of all 3 years. In 1998, a windy year, it developed three months later than it did in 1997 and 2 months later than it did in 1999. Sulfide

Sulfide presumably was usually present in anoxic waters when the Salton Sea was stratified. On 31 July 1999, sulfide concentration was 1.0-3.0 mg 1-1 at 810 m depth and >5 mg 1-1 at 10-14 m depth. On 16 August 1999, it wasO.5-1.0mg 1-1 at 0-6 m depth and 3-5 mg 1-1 at 6-14 m depth. These sulfide concentrations were high enough to deoxygenate the entire water column during a mixing event, at least over the greater part of the lake, where depth > 8-1 0 m. At station S-I on 31 July, we calculated 3.8 mg 02 1-1 would be consumed by complete oxidation of 1.9 mg HS- 1-1, the mean sulfide concentration averaged over the 0-14 m water column. On this date the mean oxygen concentration was 2.8 mg 021-1. This would have been enough to oxidize 1.4 mg HS- 1-1 and to leave a final mean concentraion of 0.50 mg HS - 1-1 in the water column. On 16 August 1999, mean dissolved oxygen was 0.1 mg 1-1 and mean sulfide concentration was 1.3 mg 1-1. Even in surface waters, oxygen concentration was low (mean of 0.2 mg 02 1-1 for 0-3 m) on 16 August. Since the mixing event that initiated this condition occurred at least two days prior to 16 August (see above), the residual levels of sulfide in surface waters on 16 August seemed high. Mid-lake horizontal variations

In general, temperature and dissolved oxygen profiles were similar among S-l, S-2 and S-3 during the 3-year sampling period. However, large differences among stations in temperature and dissolved oxygen profiles were found on a few dates (Fig. 4A, B, D, F, G, I). On 3 July 1998, the northern basin apparently had mixed more recently than the southern basin (Fig. 4B, G). On this date S-l was nearly isothermal below 2 m, whereas S-2 and S-3 were stratified with a difference of 3 °C between 2 m and 12 m (Fig. 4B). All three stations were anoxic below 6 or 8 m (Fig. 4G). The southern basin may have mixed more recently than the northern basin just prior to 19 July 1999. All five stations were thermally stratified on that date (Fig. 4D). However, the water below 6 m was 2-6 °C colder

167 20 22 24 26 28 30

o

------J

24-May1998 ,

A

2

,

,

£6 Co

,

Q)

08

,,

o

o

10

15

24-May1998

F 2

I.'

1

,I

4

4

4

6

6

6

6

8

8

8

8

10

10

12

12

o

12

o

5

15

3-Ju11998 ,

G

2

.J'" ,

/' ,I "

4

10

0

5

10

o

15

H

2

2

ft: ,

'.

4

~

5

2

,

I

U "

6 "

8

8

8

10

10

10

;'

10

12 _...-... __._._..__.-...

12

12

,: ;/

;,

~.-

12

15

16-Aug 1999:

,I.,": "

: I,

I

10

'1 f

6

6

I,

12

o

4

'J

'.

o

J

'/I ,

/I

4

15

}:I

!: I"

10

5

0 +---L-12--D-e-c-19-9-'8

~

16-Aug 1999'

E

"[ir l

;,

10

19-Ju11999,

D

23 25 27 29 31 33

oI

4-

J

., " :.: ,,

-L_~

2

12

5

12-Dec 1998,

c

21 23 25 27 29 31

o 2

.. 1

I---'_~_-l

13 15 17 19 21 23

2

10

I

3-Ju11998 '

B

o

2

I

10· , 12

30 32 34 36 38 40

o

--S-1 -- -- S-2 :

·····-·S-3 : _·_·-S-4 :

_ .. -. S-5 .._.............

Figure 4, Temperature and dissolved oxygen profiles for the midlake sampling stations on five dates in 1997-1999. Temperatures for top 2 m of the water column are not shown as they were strongly influenced by time of day and the sequence in which stations were visited. Temperature ranges shown differ, but each spans 10 °C.

at S-1 than at S-2 and S-3, indicating that the latter two stations had mixed more completely or recently than had S-1. Between 3 July and 19 July, bottom water (> 7 m) temperature showed no increase at S-l but a ~3 °C increase at S-2 and S-3, Dissolved oxygen profiles also support this conclusion as S-2 and S-3 had measurable dissolved oxygen near the lake bottom while station S-1 was anoxic below 8 m (Fig. 41) An unusual unstable condition was observed on 24 May 1998. The water column was almost isothermal at S-I, S-2 and S-3 with a I-2°C difference between 2 and 12 m at each station (Fig. 4A). At all depths, however, the northernmost station (S-1) was 2-3 °C warmer than the middle station (S-2), and that in tum was 2-3 °C warmer than the southern station (S-3). That the three stations had mixed recently to a depth of 10 m was suggested by the dissolved oxygen concentrations at depth (Fig. 4F). This unusual situation may have resulted from a combination of greater windiness and downward mixing of surface waters in the southern part of the lake, and from warmed surface waters

being advected southward by winds. Wind speeds were higher at the southernmost meteorological station prior to 24 May 1998. During 10-24 May 1998, the northernmost two stations (CIMIS 141 and CIMIS 154), and the eastern station (CIMIS 127) had an average wind speed of 3.0-3.3 m s-I. For this same period, the southeastern station (CIMIS 128) had an average wind speed of 4.2 m s-I. Winds were predominantly from the north and northwest (300°-360°) during this period. Variations with distance from shore Nearshore areas differed from offshore areas in their oxygen profiles during summer. On 17 June and 29 August 1999, the stratum of water containing > 1 mg 02 I-I thickened from lake center toward shore (Fig. SA, B, D). In other words, the isopleth for 1 mg 02 I-I lay deeper nearshore than in lake center. An hypoxic stratum (< 1 mg 02 I-I) was not detected in areas shallower than approximately 8 m.

Cl

c. Q)

.c ....

S

Cl

c. Q)

.c ....

~

'"' E

Cl

Q)

~

E .c a.



3

2

.,

_

3

2

6

Bottom

4

4

_ 5

7

5

6

Distance to shore (krn)

3

5

6 7

6

6

_

SJ

A

6

7

C, •

7'____-'

7

/-~3

_

__

,______--7----

,....------e

11

" "---2j

l~/',// ~,L

Bottom

5'>

3

4

__,

Bottom

-!~~6~ary 6, 2000 _

r

a

2

.~~

---..r"7~

J~~~,~st 29, 1999

. -_ _ ,

"-2_

1'00 -1400 i

_w_

5

11

._

'\

==-7~~~

June 17, 1999

.'----

~

V-

~

a.

.c ....

S

......

14

12

10

a

6

4

2

~5

9-

'" 9

-

1130

1100 - 1200

2

2

January 6,

5E

_

r--'

3

3 mg 021-1 (Fig. 5C, E). On the southern transect, nearshore water contained higher dissolved oxygen concentrations (by 1-2 mg I-I) than did midlake. On the northern transect, the nearshore water contained 2-4 mg 02 I-Iless than did mid-lake. We do not present here the temperature profiles along the transects. Those profiles show little variation and do not aid understanding of the oxygen concentration variations. On all three sampling dates, water temperature at any given depth varied only I-2°C between nearshore and offshore waters. Effect of inflows on vertical sUrface salinity gradients and stability

Vertical salinity gradients were detected in the southern part of the lake. On 23 June 1999 after a windless period of 10 days, a salinity gradient was observed in surface waters at S-4, the station located 10 km to the north northwest of the mouth of the Alamo River (Fig. 6). Salinity was 17 g I-I at the surface and 41 g I-I at 1 m and deeper; temperatures were 30.1 °C at 0.15 m, 29.4 °C at 1 m, 28.4 °C at 2 m, 28.2 °C at 3 m, 28.2 °C at 4 m, 28.0 °C at 5 m, 27.9 °C at 6 m, and 24.7 °C at 7 m. No salinity gradients were observed on this date at any of the other four sampling stations. On

other dates after windless periods, salinity differences between 0.15 m and I m ranging from 0.6 to 2.3 g I-I were observed at station S-4. The potential import of such salinity gradients is reflected in some simple calculations of stability, the amount of work that would be required to eliminate the density gradient (Table 3). Salinity differences between the 0.15 m and 1 m contributed to stability more than did the thermal gradient on 23 June 1999 at S-4. Salinity stability was 14 times greater than stability due to the temperature gradient (Table 3). When thermal stability was calculated for the hypothetical condition of 35°C at 0-4 m and 12°C at 4-6 m, it was still only 43% that of salinity stability.

Discussion Wind regime as major driver

Throughout the year, wind events are important for mixing of the Sea. Wind also drives the currents of the Sea that distribute the nutrient rich, low salinity waters flowing into it via the New, Alamo and Whitewater rivers. The frequency, strength and duration of wind events affects currents, temperature, dissolved oxygen, sulfide, nutrient cycling and the distribution and abundance of biota, as in other polymictic lakes (Mitteilung, 1988; MacIntyre, 1993; MacKinnon & Herbert, 1996). Differences among years in spring and summer wind patterns can cause significant variation in lake dynamics among years, making many processes and phenomena very unpredictable in their timing and magnitude. Frequent and strong wind events, as in spring 1998, can repeatedly break down incipient thermal stratification throughout the spring and delay the onset of anoxia and sulfide accumulation in bottom waters. Conversely, when spring wind events are infrequent or weak, thermal stratification and anoxia develop sooner in the year and persist for longer intervals.

170 Convection due to heat loss from surface waters likely causes much of the water column to mix on a daily or near daily basis in fall and winter and surface waters in spring and summer. Heat loss results from conduction and evaporation (lmberger, 1985). As the Salton Sea is isothermal or nearly so from October to January, even low wind speeds can contribute directly to mixing. Heat loss occurs in surface waters even during spring and summer as a result of both high evaporation rates and large day-night air temperature differentials. Differential mixing due to differences in wind conditions and bathymetry of the northern and southern basins probably accounts for differences in their temperature and dissolved oxygen regimes (Fig. 4). Evidence of differential mixing among stations was found almost exclusively during the spring and summer when wind drives mixing, as on 24 May 1998 (Fig. 4A, F). Mean daily wind speed for the 10 days previous to this date were 1-4.5 m s-I higher at the southeastern meteorological station than the northern and eastern stations and may have generated the 4-6 DC temperature gradient that existed along the lake's main axis. Differential mixing would be increased due to bathymetric differences between the two basins. No sharp boundary exists between the northern and southern basins, but a convenient dividing line can be drawn based on circulation data (Arnal, 1961; Cook & Orlob, 1997) (Fig. 1). Although lengths (fetches) of the two basins are roughly equal, the southern basin is wider, with a maximum width of ~25 km as opposed to 16.5 km for the northern basin. In addition, the northern basin has a greater maximum depth (14 m, as opposed to 12 m) and a greater mean depth (Arnal, 1961; Ferrari & Weghorst, 1997). Because the southern basin is shallower and has a greater surface area, lower wind speeds may result in more complete mixing of the southern basin during spring and summer. This would produce differences between the two basins in their thermal, dissolved oxygen and sulfide dynamics and in the impacts on biota during spring and summer months. Differences between basins would disappear due to horizontal advection during periods high wind speeds. The greater extent of well-mixed water in the southern basin probably makes it an important refuge from hypoxia and anoxia for fish and benthic macroinvertebrates from May to September. Arnal's (1961) analysis of wind patterns suggested that prevailing west to northwest winds in the northern basin and west to southwest winds in the southern basin drove circulation patterns of the Sea. During

our study, winds in the southern basin (CIMIS 128) usually originated from the northwest and rarely originated from the west or southwest, and when they did, they were of short duration and low velocity. The discrepancy between this finding and that of Arnal is most likely due to Arnal's wind data coming from a weather station in the city of El Centro, ~40 km to the south of the Sea.

Currents and salinity gradients During windless and to some extent even windy periods, the fresher water (2.5-5 g 1-1) entering the Sea from the New and Alamo rivers flows over the surface of the southernmost portion of the Salton Sea creating a surface layer of less saline, less dense water. We detected salinity gradients only at one station (SA). These scant data combined with salinity and current data from other studies (Arnal, 1961; Parsons 1986) suggest that salinity gradients may inhibit mixing over portions of the southern basin. The phenomenon does not appear significant in the northern basin. Salinity gradients of 0.2-2 g 1-1 extending 11 km from the Whitewater River in the north were found by Arnal (1961), however. Such gradients were not observed at the northern stations monitored by Parsons (1986) or ourselves. Distribution of salinity gradients in the southern basin is determined primarily by wind and currents. Salinity at the surface of station S-4 was often 2 g 1-1 less than mean Sea salinity and occasionally substantially lower, as on 23 June 1999 (Fig. 6). Under windy conditions, the current pattern in the southern basin is a counter clockwise gyre that causes freshwater to flow northeast along the southeast shoreline and then northwest along the eastern shoreline (Arnal, 1961; Cook & Orlab, 1997). Vertical salinity gradients of > \-2 g I-I are then created along a 2-8 km wide strip along these shorelines. Strong winds would accelerate mixing and minimize the spatial extent of areas with lowered surface water salinity. Current strength will be at a minimum during low wind periods when strong vertical salinity gradients are likely to extend well out from the delta areas of the New and Alamo rivers. Wherever they occur, salinity gradients arising from these inflows will inhibit mixing of surface and bottom waters and the movement of heat and oxygen from surface to bottom waters. The water column stability caused by these salinity gradients is usually much greater than that caused by thermal stratification. It may counter the tendency for the shallower southern

171 basin to mix more readily at lower wind speeds than does the deeper northern basin. The impact of inflows on mixing is increased by the fact that the majority of water entering the Sea enters from these two rivers and their mouths are only 12.5 km from each other on a ~ 150 km shoreline.

may still be 0.4 mg I-I after 3 days. Such levels can be toxic for fish and invertebrates (Bagarinao & Vetter, 1989; Hagerman & Vismann, 1995). To persist, mid-lake fish and metazoan populations must exist in hypoxic, sulfide-rich conditions for a period of days.

Sulfide and deoxygenation of the mid-lake

Sulfide concentrations seem to have increased in the Salton Sea over the last half century. Carpelan (1958) found maximum sulfide concentrations of only 0.085 mg I-I (and average concentrations of approximately 0.02 mg I-I) in anoxic bottom waters 7 km to the westsouthwest of S-I during biweekly monitoring in 1956. This compares with sulfide concentrations measured in this study of >5 mg I-I, and we measured sulfide only on two dates. The increased sulfide concentrations could be a consequence of an increase in sulfate concentrations from 6.8 in 1956 to 9.1 g I-I in 1999 and an increase in biomass in the Sea (Carpelan, 1961; Hurlbert & Detwiler, unpubi. data; Riedel et aI., 2002; Tiffany et aI., 2002; unpubi. data). The increased organic matter in the Sea may be in the form of both phytoplankton and tilapia (Oreochromis mossambicus Peters) biomass and fecal matter. Tilapia, an exotic fish that invaded the Sea in the 1960s, is now the most abundant fish present, and periodically suffers massive moralities, along with other fish, during some mixing events. Carpelan (1958, 1961) noted a single deoxygenation of the mid-lake water column in 1955. It followed a windstorm that lowered surface dissolved oxygen concentration to 0.8 mg I-Ion 16 September 1955. Although sulfide oxidation and simple dilution by anoxic bottom waters most likely the caused this deoxygenation event, sulfide was not monitored in the water column until 1956. The deoxygenation event only lasted I day (Carpelan, 1958), perhaps because sulfide concentrations were lower in 1955 than in 1997-1999. Later studies also noted the odor of sulfide in the air following wind driven summer mixing events, but did not measure sulfide concentrations in the lake (Bain et aI., 1970).

The deoxygenation events consistently occurred during the time of maximum water temperatures (JulySeptember), most notably in September of all three years. These were a result of bottom waters low in oxygen and rich in sulfide, and presumably organic matter (dissolved and particulate) and microbial heterotrophs, mixing with surface waters. The chemistry and kinetics of sulfide production and oxidation are complex and important in understanding development of these deoxygenation events. In the absence of oxygen, bacteria may use nitrate, metal oxides, carbon dioxide, or sulfate as a terminal electron acceptor to decompose organic matter (J0rgenson, 1982). In the Salton Sea, high concentrations of sulfate (9.1 g 1-1) make it likely to be the predominant electron receptor used by bacteria for anaerobic decomposition (Goldhaber & Kaplan, 1974). The reduction of sulfate coupled with organic matter decomposition generates sulfide in the anoxic bottom waters. In some systems, photosynthetic bacteria oxidize any sulfide produced by decomposition (Huxtable, 1986). However, in the Salton Sea, anoxic bottom waters usually are found below the euphotic zone (top 4-5 m) so sulfide is not oxidized by this mechanism. Therefore, sulfide accumulates in the bottom waters, as it does in most aquatic systems with high organic matter loading to bottom waters during thermal stratification, e.g. Baltic Sea (Ehrhardt & Wenck, 1983) and Lake Kinneret (Eckert & Hambright, 1996). In the Sea, enough sulfide is present in bottom waters that mixing events may leave surface waters hypoxic and laden with sulfide. The length of time that sulfide is present may be as important for biota as the absolute concentrations present after a mixing event (Theede et aI., 1969). Sulfide oxidation is not instantaneous and although the reaction half-life has not been directly measured at the Salton Sea, it has been estimated at 10 - 50 h in the laboratory (Cline & Richards, 1969; Almgren & Hagstrom, 1974). If the reaction half-life is conservatively assumed to be 24 h and initial concentrations are 2.8 mg HS I-I (as on 31 July 1999), sulfide concentrations

Changes in sulfide levels since 1950s

Changes in oxygen levels since 1950s Since 1954-1956, when Carpelan investigated the oxygen regime of the Sea, notable changes have occurred in the system that led us to suspect that dissolved oxygen concentrations would be generally lower during this study. Salinity increased from ~ 33 to ~42 g I-I, lowering oxygen solubility. Fish biomass

172

.-40

-0

0

... Q)

L..

:::J

CO

30

Q)

25

E

20

L..

c. Q)

A

35

I- 15 10 20

B

Surface (0-1 m)

16 ,....I

--

12

0)

E 8 cQ)

0)

4

~ 0 0 "'0 Q)

>

12

(/)

en

C

Bottom (12m)

0

8

0

4 0

' "

.

Ii

J

F

M

A

M

----1997 ---.-1998 ~1999

J

J

A

S

0

N

0

"'0···1954 "'0.. ·1955 .. ·t:. .. ·1956

Figure 7. Salton Sea temperature and dissolved oxygen comparison of conditions between 1954-1956 and 1997-1999. (A) Surface temperature (0-1 m). (B) Dissolved oxygen in surface waters. (e) Dissolved oxygen in bottom waters. Measurements for 1954-1956 were taken biweekly at a sampling station 7 km west-southwest of S-1 from August 1954 to July 1956. Those for 1997-1999 were taken at S-1 (Fig. 1).

has increased, mainly due to the introduction of til apia to the Sea, thereby increasing respiration and, presumably, decomposition rates in the Sea. Temperature affects dissolved oxygen, but the thermal regime has

not changed notably since 1954-1956 (Fig. 7A). But to our surprise, a pattern of generally lower oxygen levels in 1997-1999 was not observed.

173 In surface waters, mid-day dissolved oxygen concentrations were, however, more variable during 1997-1999 when compared to 1954-1956 (Fig. 7B). Little inter-annual variation and only small seasonal differences were found in 1954-1956 whereas large variation was found among years and seasons in 19971999. In 1954-1956, surface waters were well oxygenated at all seasons (5-11 mg 02 1-1 ) except for one notable date in September 1955 when surface waters were hypoxic (0.8 mg 02 1-1). In contrast, surface waters in 1997-1999 were supersaturated at times during january-july (10-20 mg 02 1-1) and then hypoxic or anoxic at times in August and September. During the cooling period when the Sea was mixing daily or nearly daily, oxygen concentrations of surface waters were similar among all six years. Bottom waters in 1954-1956 were usually oxic, with periods of hypoxia or anoxia developing during June-August for only a few days at a time (Fig. 7C). During 1997-1999, hypoxic or anoxic conditions began occurring as early as February and persisted for longer periods than in 1954-1956. These changes in oxygen dynamics are most likely a result of increases since 1954-1956 in rates of primary production, fish production, decomposition, and sulfide production. Higher phytoplankton densities (M.A. Tiffany, unpubl. data) lead to increased photosynthetic rates and, therefore, a greater tendency for surface waters to be supersaturated during spring and summer when the Sea is stratified. Increased primary production supports higher densities of other organisms in the Sea and, therefore, higher rates of decomposition and sulfide production in bottom waters throughout the year. Sulfide production may also have been enhanced by the 34% increase in sulfate concentrations in the Sea since the 1954-1956. These higher sulfide concentrations can now reduce more dissolved oxygen in the water column during a mixing event. During fall and winter, when the Sea is well-mixed, processes favoring supersaturation and deoxygenation are inhibited and dissolved oxygen concentrations in 1997-1999 are comparable to 1954-1956.

ive mixing of oxygen poor bottom waters into surface waters nearshore. The potential biological significance of the thickened oxic stratum nearshore region is great, as the oxic nearshore acts as a refuge from hypoxic or anoxic conditions for fish and benthic macroinvertebrates, even during major deoxygenation events. Breaking waves at the shoreline, friction at watersediment interface generated by wind driven surface currents, and convection driven by greater nocturnal cooling of shallow waters all probably contribute to the greater oxygenation and mixing of the nearshore waters. These mechanisms were not directly investigated in this study, but would be expected to result in a thicker oxic stratum and less opportunity for build-up of sulfide concentrations close to shore. Strong wind events that mix the Sea occur sporadically throughout spring and summer at mid-lake. These reduce oxygen concentrations in surface waters by mixing into them, to a greater or lesser degree, oxygen poor, sulfide-rich bottom waters. These midlake surface waters then gradually reoxygenate. In shallower nearshore areas, however, these low quality bottom waters are either absent or occupy a smaller percentage of the water column. Thus a mixing event that reduces oxygen levels in mid-lake surface waters by 50, or 90, or 99% may have negligible effects on oxygen levels in nearshore surface waters. Under the right conditions, however, currents can carry oxygen poor mid-lake surface waters toward shore. The thickening of the oxic stratum nearshore affects estimates of lake bottom exposed to various dissolved oxygen concentrations (Fig. 3C). Plotted values slightly overestimate the areal extent of low dissolved oxygen conditions. Only oxygen values at the mid-lake stations were used for calculating these percentages. Overestimates only occur during periods of stratification and are on the order of ~5% when conditions mimic those found on 17 June 1999 (Fig. 5A), somewhat greater under conditions such as those found on 29 August 1999 (Fig. 5B).

Consequences for biota Thickening of oxic stratum nearshore The greater thickness of the oxic stratum in nearshore waters seems likely to be a general phenomenon during the warmer part of the year. Though our own data are limited, we suspect it is attributable to two principal factors: greater turbulence of surface waters nearshore than offshore, and less frequent and intens-

Sulfide and oxygen regimes, themselves driven largely by wind and temperature regimes, have an influence on plankton, benthos and fish populations in the lake. Anoxia and sulfide have well documented adverse effects on aquatic organisms, especially metazoans (Bagarinao, 1992; Hagerman & Vismann, 1995). As the two factors usually coincide in the Salton Sea, their

174 separate effects were not distinguishable in this study, and, for convenience, are sometimes both subsumed under phrases such as 'deoxygenation effects'. Major direct effects of deoxygenation are abandonment of the lake center by fish and benthic macro invertebrates during May-September and abrupt plankton and fish die-offs in August and September. A recent gillnet survey of fish distribution found that during summer few fish are found in the midlake (Riedel et aI., 2002); and observations with a depth/fish sonar unit mounted on the sampling vessel show that when fish are found mid-lake, they are only found in the top 6 m of the water column. In spring, fish migrate to nearshore waters. This movement may be in response to the first injections of hypoxic and sulfide-bearing bottom waters into surface waters during wind events after early stratification periods in April and May. Reduction of mid-lake fish abundance likely affects mid-lake plankton populations (Tiffany et aI., 2002, unpubi. data). Movement of fish to nearshore occurs when tilapia is reproductively active (Riedel et aI., 2002). Tilapia are mouth brooders that build nests on the lake bottom in the form of depressions a few decimeters in diameter. High densities of their nests have been documented by sonar imaging in offshore areas at depths down to 12 m. It is not known at what time of year these nests were created. Clearly it would have been prior to strong thermal stratification, during which nesting would be restricted to shallow ( 5 mg HS-I- 1) are in the lethal range for many aquatic organisms (Smith et aI., 1976; Bagarinao, 1992). Nearshore areas provide refugia for metazoan benthic macroinvertebrates during summer. Their densities are low even here in summer, however, which represents a further reduction in fish food supplies. The beginning of strong and deep convectional circulation, usually in September, signals when effective recolonization of deep sediments by benthic macroinvertebrates can begin (Detwiler et aI., 2002).

The deoxygenation events in August and September cause mortality of fish and plankton. These events in 1997, 1998 and 1999 were accompanied by crashes in popUlations of the dominant metazoan zooplankters (Apocyclops dengizicus MUller and Brachionus rotundiformis Lepeschkin), reduction in phytoplankton densities (dominated by diatoms, dinoflagellates and a raphidophyte), and increases in anoxia-tolerant ciliates (Tiffany et aI., 2002, unpubi. data). Moribund and freshly dead fish were observed on the lake surface in large numbers following deoxygenation events. Large fish mortalities events are common at the Sea. One fish die-off in August 1999 was estimated to involve 7.6 million tilapia. A deoxygenation event in September 1955 resulted in mortality of bairdiella (Bairdiella icistia Jordan and Gilbert), the most abundant fish in the Sea at the time (Carpelan, 1958). Not all mortality events are associated with deoxygenation events. Mortalities of tilapia that sometimes occur during winter months have been attributed to low water temperatures ( 10 mS had 64 (±0.6) species compared to 70.1 (± 1.8) taxa in habitats with conductance values < 10 mS. There were no significant correlations between N03-N1- and either of the two biotic indices. Members of the Achnanthaceae (39.2% ±2.3), Eunotiaceae (16.3% ±1.4), Naviculaceae (11.3% ±1.7) and Fragilariaceae (10.7% ±1.1) were the dominant diatom groups in upland streams. Numerically dominant taxa in the upland systems included Achnanthes minutissima (31.5% ±5.7), Eunotia pectinatus (16.1 % ±2.9), Fragilaria pinnata (15.9% ±5.4), Navicula radiosa (14.7% ±0.5) and Tabellariaflocculosa (4.9% ±2). Rhizosolenia eriensis was collected only in the upland streams at low densities where pH was -

00.6

z

~0.4

SO.2 II:

u.

~ o

°

DEGREES

°

DEGREES

~0.3

0°·2 a. 00.1 II:

a.

c. >-

90

CONTROL.

~0.4 U.I

8°·2 U.I

Funeral Mountains Springs

II:

u.

60

~

o

TREATMENT

~0.8

0°·6 a. 0.4

~0.2

a.

Badwater is a small spring province covering approximately 0.5 ha of the eastern-most extent of the Badwater Basin, which is a barren salt flat composed of sodium, calcium, and magnesium salts (Hunt, 1975). Vegetation occurs peripherally in the basin and consists of widely dispersed patches of pickleweed, wire grass (Juncus spp.) and honey mesquite (Prosopis glandulosa). The basin is only wet during winters with high precipitation. Permanent aquatic habitat is limited to three small spring clusters at Badwater (referred to as South Pool, Middle Pool and North Pool). All of these springs are 0.5. Negative electivity indicated by for values -0.1 to -0.5 and by - for values 0.23, T-test), spatial differences between pools ns (p > 0.24, T-test).

Population studies Shell height distributions at Badwater were similar to those observed at Cottonball Marsh, where each population also included several cohorts during spring and autumn. Contrary to Cottonball Marsh populations, there were no significant spatial or temporal differences (p > 0.05, T-test) in height distributions at Badwater (Table IB). Abundance of snails at Middle Pool and North Pool ranged from 0 to 19000 m- 2 (0-40 snails/21 cm- 2 at both sites) (Fig. 4). Mean snail density in quadrats I, 2 and 3 was always lower at Badwater than at Cottonball Marsh. Spatial and temporal variations in densities were small. Approximately 75% of the population consistently occupied quadrat 1 (located nearest water) and densities decreased sequentially in quadrats 2 and 3 as distance from water increased. Funeral Mountains Springs A. infima were limited to habitat within approximately IS cm of water, and they were most common in decaying cattail (Typha sp.), sedges (Scirpus sp.), and fallen fan palm (Washingtonia felifera) leaves. They

263 were rare in decaying salt grass (Distich lis spictata) and wire grass (funcus spp.). At Nevares Springs, 11 were found in 83 grab samples taken along a total of 370 m of the largest spring in the province. It was absent in 37 grabs taken from all other springs and seeps in the area. Absence from seeps can be attributed zeric conditions caused by small discharge and periodic drying. Reasons they did not occur at flowing springs are not clear, however, evidence of dredging at these springs suggests that existing habitat conditions have naturalized from past disturbance that may have extirpated populations. At Travertine Springs, 28 were observed in 162 grabs along approximately 1 km of springbrook. It was absent from 56 grabs taken from springbrooks without vegetation or debris, from Texas Spring, and all seeps in the area. Absence from Texas Spring can be attributed to dry conditions caused by total diversion of its outflow for many years. Threloff & Koenig (1999) estimated that diversion has eliminated approximately 85 % of historical springbrooks at Travertine and Texas springs, which suggests that A. infima habitat has been reduced by a similar amount at these springs.

Discussion Even though desert wetlands are comparatively small they provide the only persistent water over expansive landscapes of most continents. In the Death Valley region of North America, persistent water is additionally distinguished by providing habitat for at least 60 endemic plants and animals (Sada et al., 1995). A number of studies have examined fish ecology and fish and invertebrate taxonomy in these systems, but invertebrate ecology has received little attention. Characteristics of the semi-aquatic habitat occupied by A. infima suggests that its demography may be highly influenced by seasonal conditions affecting soil temperature and water availability. Its demography and habitat use indicate, however, that seasonal variation in abundance, size class structure, and habitat use is small. Size class structure of all Badwater and Cottonball Marsh populations were similar during spring and autumn, and there was little spatial and temporal variation in population size. Snail abundance at Badwater was consistently less than at Cottonball Marsh. It is difficult to determine if this is attributable to impacts of human disturbance, the small size of Badwater springs, or differences in chemical composition of salts at the two areas. Results of trampling exper-

iments indicate, however, that snail abundance and habitat may be detrimentally impacted by visitors. Preferred habitat for A. infima at Badwater and Cottonball Marsh consisted of high banks, high bank angles and bank overhangs. There was little seasonal variation in habitat use, but the amount of available habitat during autumn at Cottonball Marsh appeared to be greater than during spring. This was indicated by densities in quadrats 2 and 3 that were highest in all three sites during autumn and lowest in spring. Increased abundance in habitats more distant from water may result from seasonal changes in water availability and temperature that expand habitat availability during periods when air temperatures are low and the amount of water greatest. Snails may be excluded from these habitats during spring and summer when air temperatures are highest and soil moisture is lowest. On the Death Valley floor, A. infima relies on persistent water and high-relief salt crust that forms when evaporation draws moisture and salts from soils deposited during pluvial periods. Its preferred habitat occurs along springbrooks where salt crust banks are high, steeply angled and deeply undercut. Experimental manipulations indicated that trampling decreased bank heights, angles, undercuts, and snail abundance. The comparatively large declines that were caused by only 60 s of trampling suggest that populations can be easily extirpated with similar, but more extended, activities. The specificity of A. infima habitat use also suggests that it is poorly suited to occupy other habitat types, and that populations can be eliminated when habitats are altered. Reliance on persistent water and their absence from habitats without undercuts, and with low bank angles and height also indicates that populations may be extirpated when spring brooks are physically modified. These impacts are similar to effects to riparian and aquatic habitats that are caused intensive livestock grazing (Fleischner, 1994). Sada & Vinyard (2002) concluded that these types of impacts have resulted in extinctions and population extirpations of many aquatic taxa endemic to the Great Basin region of North America. Impacts of human trampling on salt crust relief, examination of recent and historical photographs, and the influence of trampling on A. infima habitat and abundance all indicate that public use has altered physical and biological characteristics of Badwater. Death Valley National Park archival photographs indicate that Badwater remained in comparatively natural condition until the first road was constructed into the area during 1933 (Lingenfelter, 1986). Photographs

264 taken over the past 50 years show the extent of highrelief salt crust dwindling over much of Badwater. It is difficult to compare extant and historical A. infima populations at Badwater without knowledge of its abundance and distribution prior to human disturbance. However, historical changes in the extent of flattened salt crust indicates that little habitat has been lost at North Pool and that approximately 30 m has been lost at Middle Pool. This represents a reduction of approximately 30% of historical habitat at Middle Pool. Abundance and distribution of A. infima has also been affected by diversions that dried aquatic and riparian habitats supported by Texas, Travertine and Nevares Springs. Influence of these diversions on A. infima is also difficult to determine because there were no studies documenting its historical abundance and distribution. Its present occurrence in all springfed riparian areas with persistent water unaffected by frequent flood events suggests, however, that diversions eliminated several kilometers of habitat that may have been occupied by millions of individuals. Other aquatic macroinvertebrates endemic to Funeral Mountain springs have also been affected by these diversions. It is likely that continuation of historical use patterns at Badwater and future springbrook diversions will cause additional declines in abundance and distribution of many aquatic macro invertebrates endemic to Death Valley. Fortunately, A. infima populations at Cotton ball Marsh appear to occupy its historical habitat and to be comparatively unaffected by cultural uses. Although information is needed to design recovery programs for Badwater populations of A. infima, observations at Cottonball Marsh indicated that trampled salt crust can regain its relief when protected from trampling. Growth rates were comparatively slow, approximately 2 cm yr- I in trampled areas, and considerable time may be required to allow the crust to mature and provide A. infima habitat. Water rights granting diversions from Funeral Mountain Springs may preclude restoration of these systems, but diversions should be curtailed as much as possible to prevent additional losses and restore as much A. infima habitat as possible. Public use of aquatic and riparian resources in Death Valley National Park should consider biological integrity of these spring resources. These uses should be allowed only when management programs result in functional conditions at these springs that are more similar to those at Cottonball

Marsh than to those currently existing at Badwater and in the Funeral Mountains.

Acknowledgements Funding and assistance with this work was provided by the U.S. National Park Service, Death Valley National Park. D. Threloff, B. Davenport and B. Thornburg assisted with field work. S. Smith, D. Herbst and J. Wehausen, D. Blinn and an anonymous reviewer provided helpful comments to early manuscripts.

References Baldinger. A. 1., W. D. Shepard & D. L. Threloff. 2000. Two new species of Hyalella (Crustacea: Amphipoda: Hyalellidae) from Death Valley National Park, California, U.S.A. Proc. BioI. Soc. Wash. 1l3: 443-457. Berry, S. S., 1947. A surprising molluscan discovery in Death Valley. Leaflets in Malacology I: 5-8. Fleischner. T. L., 1994. Ecological costs of livestock grazing in western North America. Cons. Bio. 8: 629-644. Fowler, B. H., 1980. Reproductive biology of Assiminea californica (Tyron, 1865) (MesogaSlropoda: Rissoacea). Veliger 23: 163166. Hershler, R., 1987. Redescription of Assiminea infima Berry, 1947, from Death Valley, California. Veliger 29: 274-288. Hershler, R., 1989. Springsnails (Gastropoda: Hydrobiidae) of Owens and Amargosa River (exclusive of Ash Meadows) drainages, Death Valley system, California-Nevada. Proc. BioI. Soc. Wash. 102: 176-248. Hershler, R., 1998. A systematic review of the hydrobiid snails (Gastropoda: Rissooidea) of the Great Basin, western United States. Part I. Genus Pyrgulopsis. Veliger 41: 1-132. Hershler. R., 1999. A systematic review of the hydrobiid snails (Gastropoda: Rissooidea) of the Great Basin, western United States. Part II. Genera Colligyrus, Eremopygrus, Fluminicola, Pristinicola and Tryonia. Veliger 42: 306-337. Hershler, R. & D. W. Sada, 1987. Springs nails (Gastropoda: Hydrobiidae) of Ash Meadows, California-Nevada. Proc. BioI. Soc. Wash. 100: 776-843. Hunt, C. B., 1975. Death Valley Geology, Ecology, and Archaeology. U. California Press, Berkeley, CA: U.S.A., 234 pp. Hunt, C. B., T. W. Robinson, W. A. Bowles & A. L. Washburn, 1966. Hydrologic basin, Death Valley, California. U.S.G.S. Prof. Pap., 494-B: BI-B138. Jacobs, 1., 1974. Quantitative measurement of food selection: a modification of the forage ratio and Ivlev's electivity index. Oecologia (Berlin) 14: 413-417. La Bounty, 1. F. & J. E. Deacon, 1972. Cyprinodon miller;, a new species of pup fish (Family Cyprinodontidae) from Death Valley, California. Copeia 1972: 769-780. La Rivers, I., 1948. A new species of Ambrysus from Death Valley with notes on the genus in the United States (Hemiptera: Naucoridae). Bull. S. Calif. Acad. Sci. 47: 103-110. Levins, R., 1968. Evolution in Changing Environments. Princeton U. Press, Princeton, NJ, U.S.A.: 120 pp. Lingenfelter, R. E., 1986. Death Valley and the Amargosa. A Land of Illusion. U. California Press, Berkeley, CA, U.S.A.: 664 pp.

265 Miller, R. R., 1981. Coevolution of deserts and pupfishes (genus Cyprinodon) in the American southwest. In Naiman, R. 1. & D. L. Soltz (eds), Fishes in North American Deserts. John Wiley & Sons, NY, U.S.A.: 39-94. Sada, D. W. & G. L. Vinyard, 2002. Anthropogenic changes in biogeography of Great Basin aquatic biota. Smiths. Contr. Earth Sci.: in press. Sada, D. w., H. B. Britten & P. F. Brussard, 1995. Desert aquatic ecosystems and the genetic and morphological diversity of Death Valley system speckled dace (Rhil1ichthys osculus).ln Nielsen, J. L. (ed.), Evolution and the Aquatic Ecosystem. Defining unique units in population conservation. Am. Fish. Soc. Sym. 17: 350359. Shepard, W. D., 1993. Desert springs - both rare and endangered. Aq. Conser.: Mar. and Freshwat. Eco!. 3: 351-359. Shepard, W. D., 1992. Riffle beetles (Coleoptera: Elmidae) of Death Valley National Monument, California. Great Basin Natur. 52: 378-381.

Smith, G. R., 1978. Biogeography of intermountain fishes. Great Basin Natur. Mem. 2: 17-42. Taylor, D. w., 1985. Evolution offreshwater drainages and molluscs in western North America. In Smiley, C. J. (ed.), Late Cenozoic History of the Pacific Northwest. Amer. Ass. Adv. Sci. San Francisco, CA, U.S.A.: 265-321. Threloff. D. L. & S. Koenig, 1999. Effects of water diversion activities on stream lengths in the Travertine-Texas Spring complex, Death Valley National Park, summer 1999. Unpublished U.S. National Park Service report, Death Valley National Park, CA, U.S.A.: 7 pp. Williams, 1. E., D. B. Bowman, 1. E. Brooks, A. A. Echelle, R. J. Edwards, D. A. Hendrickson & 1. 1. Landye, 1985. Endangered aquatic ecosystems of North American deserts, with a list of vanishing fishes of the region. J. Arizona-Nevada Acad. Sci. 20: 1-62.

Hydrobiologia 466: 267-277, 2001. J.M. Melack. R. Jellison & D.E. Herbst (eds), Saline Lakes. © 200 I Kluwer Academic Publishers.

267

Holocene hydrological and climatic changes in the southern Bolivian Altiplano according to diatom assemblages in paleo wetlands S, Servant-Vildaryl, M, Servant2 & O. Jimenez l I Antenne [RD, ESA 7073, Laboratoire de Geologie, MNHN, 43 Rue Buffon, 75005 Paris, France 2[RD, Centre de Recherche Ile de France, 32 Rue Henri Varagnat, 93143 Bondy, France

Key words: Bolivia, Altiplano, river deposits, diatoms, paleoclimate, Holocene

Abstract

This paper presents the first Holocene continuous record from the southern Bolivian Altiplano. In this area, the climate is now characterized by weak summer monsoon rains. The record is located north of Salar de Uyuni in a non-glacial valley (Rio Baja). Between ~ 11600 and ~221 0 cal year BP, the rivers accumulated fine deposits, while under the present climatic conditions, the fine particles are carried downstream by strong water floods. These deposits contain a rich diatom flora showing that the valley floor was occupied by paleowetlands. Water input needed to be more or less continuous to explain that the paleowetlands survived over a long period of time. We show that diatoms can be used to reconstruct the relative variations in the water level and the salinity throughout time, despite of the spatial complexity of this type of environment. During the Holocene, the water level was low except during some periods, dated ~ II 600-9800, ~6330-5300, and ~ 3110-2210 cal year BP. Saline and freshwater microhabitats were simultaneously present in the valley floor as indicated by a mixed diatom flora evidenced throughout the record. We propose a paleoclimatic scenario based on the assumption that the NE wet atmospheric flow of the monsoon was replaced by the westerlies of the southern hemisphere at the latitude of the study site.

Introduction The tropical southern Andes are characterized between 19° and 24°S by the largest salt flats and salt lakes of the world. Lake Poop6 is a 3000-km2 salt lake which episodically dries up, Uyuni and Coipasa salt flats cover 10 000 km2, and numerous shallow saline lakes and salt pans are observed in the southern subtropical Andes in Bolivia and Chile (Fig. I). These salt environments reflect dry climatic conditions that arose at the beginning of the Holocene. This dry period, marked by a 10-m deep salt bed in the Salar de Uyuni (Risacher & Fritz, 1992), followed a relatively wet phase (the Coipasa event) marked by the presence of a shallow saline lake (Servant et al., 1995; Sylvestre et aL, 1999). At the same latitudes in the northern Chilean Altiplano, dryness set in at ~ 1a900 cal year BP, water level decreased between ~ 10900 and 8000 cal year BP and the lakes were desiccated between ~8000 and 4000 cal year BP (Geyh et aL, 1999).

Despite of these dry conditions, Holocene paleowetlands were well-developed and widespread in the valley floors, both in northern Chile (Betancourt et aL, 2000) and in Bolivia, particularly in the nonglacial valleys (Servant & Fontes, 1984). The paleowetland sediments which deposited in a continuous fashion through time in the southern Bolivian Andes, provide excellent material for detailed paleoclimatic reconstruction. They contain a rich diatom flora (Frohlich & Servant-Vildary, 1989, Betancourt et al., 2000) which can be used, as currently done in lake deposits, to reconstruct changes in water level and salinity in a reliable river record (e.g., see reviews by Smol & Cumming, 2000). In this paper, we present a diatom study from a paleowetland located north of the Salar de Uyuni. Our objectives were (I) to reconstruct changes in water level and salinity in a river system, (2) to construct a paleoclimatic scenario in order to explain the presence of evaporites in the Salar de Uyuni while wetlands

Figure 1. The Bolivian Altiplano (modified from Cross et at.. 2000). Detailed map of the northern border of Salar de Uyuni shows the location of the Rio Baja profile.

N

0\ 00

269 were developed in the catchment and finally (3) to discuss this scenario according to the data available from the different areas of the Southern tropical Andes, both in the northern Altiplano (Lake Titicaca area) and the southern Altiplano (the Uyuni-Coipasa and northern Chile areas),

Climatic and paleoclimatic setting At the present time, the climate is characterized by low precipitation and by two contrasted seasons: a short rainy season between December and March (austral summer) and a long dry season between April and November. The seasonal distribution of precipitation reflects changes in the atmospheric circulation throughout the year at the latitudes where Bolivia is located. During the austral winter, the southern westerlies spread toward the equator and reach the Lake Titicaca area at the mid and upper level of the troposphere (Vuille, 1999) and the Bolivian Andes are submitted to dry and subsident air from the Pacific region. During the austral summer, the westerlies shift southwards, the NE humid air flow (hereafter referred to as 'South American monsoon') from the Atlantic reaches the southern tropical Andes. Summer precipitation consistently decreases from the NE to the SW, 800 mm year- 1 in the Lake Titicaca area, 300 mm year- 1 in the Uyuni area, and -100 mm year- 1 in the northern Chile. These seasonal changes are modulated on the synoptical time scale by atmospheric processes related to interactions between high and low latitudes. West of the Andes, along the Pacific coast of northern Chile and southern Peru, cold air masses from the southern polar region episodically reach the sub-tropical latitudes at the mid levels of the troposphere (Vuille & Caspar, 1997). East of the Andes the low troposphere is characterized all year round by south-north cold air incursions, that can reach the central Amazon Basin (Garreaud, 1999). Occasional non-stormy rainfall and snowfall occur in the southern tropical Andes when the tropical atmosphere is destabilized by cold air incursions, particularly during the winter dry season. Under these climatic conditions, the rivers in the non-glacial valleys are submitted to strong floods during the wet season (December to March), the valley floors are intensively eroded, and fine particles are carried downstream. During the rest of the year (April to November) the non-glacial valley floors are desiccated, except in some areas where water input from

extended groundwater reservoirs buffers the effect of the seasonal dryness. Most data concerning the Holocene climatic changes in the Bolivian Andes were obtained from the northern Altiplano. Lake Titicaca water-level variations indicate that the precipitation minus evaporation (P-E) balance in the watershed changed fundamentally throughout the Holocene. The water level was lower and the climate drier than at present between ~ 11 500 and ~4500 cal year BP (Ybert, 1992; Mourguiart et aI., 1998; Cross et aI., 2000). However, increased lake level phases occurred during this general dry trend, they are dated ~ 10 000-8500 and 7000-6000 cal year BP (Baker et aI., 2001). Between ~4500 and 3500 cal year BP, the Lake Titicaca water level increased substantially and reached the present level after 1000 cal year BP (Mourguiart et aI., 1998) showing a wet climatic trend. This wet trend has been interrupted by short dry phases, recorded by strong drops in the water level, at ~3000, 2800, 1800, and 700 cal year BP (Abbott et aI., 1997a). These climatic changes have been interpreted to be the result of modifications in the intensity of the South American monsoon.

Material and Methods The outcrop is exposed on the bank of the Rio Baja, located near the village of Alianza (67°45'W, 19°50'S) at about 3700 m above sea level, ~56 m above the Salar de Uyuni (Fig. 1). The 8-m deep profile has been continuously sampled using 30-cm deep boxes. Twenty-six boxes, representing the totality of the outcrop, were collected (courtesy of D. Wirrmann). Between 850 and 725 cm, sands and gravels are predominant, diatoms are lacking, except in a silt clay layer intercalated at 806 cm. Between 725 and 529 cm, silts and clays are dominant. From 529 to 280 cm, the fine deposits are calcareous (a calcareous crust is interbedded at 500 cm). From 280 to 180 cm, fine sands are abundant in silts, and from 180 cm to the top, silt and clays are followed by a calcareous fine deposit. Dark organic layers are intercalated in the sequence, particularly in the lower and the upper parts. Time control was based on six total organic matter samples that were 14C dated. Ages were calibrated according to the method outlined for CALIB 3.0. by Stuiver & Reimer (1993). Uncalibrated radiocarbon ages and calibrated ages (cal year BP) versus depth are reported in Figure 2. The sedimentation rate was low: 0.04 cm year- 1 before 6610 yr BP, with an increase

270

Ages 0

2500

5000

7500

10000

12500

0

100

200

300

i

--

~ 400

:r n

3

500

600

700

6OOL---------------------~

Figure 2. Chronological control of Rio Baja profile. 14C year BP dates (squares) and calibrated year BP (circles) versus profile depth (cm).

between 6610 and 3940 year BP (0.19 cm year-I) and another decrease (0.07 cm year-I) after 3940 year BP. The increased sedimentation rate coincided with the appearance of carbonates in the sediments. Thirty-nine sub-samples were collected for the diatom study. Valves were counted until no new species appeared. Thus, the number of valves counted in each sample largely depended on the abundance and the specific diversity of the diatom assemblages. Only 200 valves were counted in samples 137 and 276, 5000 valves in samples 184 and 190 and 600-1000 valves in the 35 other samples. The fossil diatom flora is composed of 182 species. Hierarchical cluster analysis was used to group the samples and define the species that are characteristic of each group of samples. This method reveals the latent structure of the 'objects' (here the diatom composition of the samples) in terms of groups of 'similar elements' (here groups of similar samples) (Roux, 1991 a). The agglomerative hierarchical procedures include two steps repeated alternately. The

first step involves searching for the distance matrix for the two closest objects as a single individual and then computing the distances between this new element and the rest of the objects. The first step is again activated on the reduced distance matrix, and so on. Cluster analysis deals with a rectangular data table where the rows are the organisms while the columns are variables. To prevent a lack of continuity and lack of stability in the data table, Roux ( 1991 b) recommends processing the data set using both a clustering method and a correspondence analysis, also called reciprocal averaging (Hill, 1973). We then take the coordinates of the most significant axes of the cluster analysis (here the first six axes) as objects which explain 55% of the variance. The usual euclidean distance formula was chosen as the best one. The dendrogram obtained from the data showed seven cut groups. In order to know which species were characteristic of clusters, we used the coefficient of correlation (,1) which is the ratio of the 'between groups' sum of squares (BSS) over the total sum of squares (TSS) (see legend of Table I). A species can be considered as highly related to a cluster when ,1 is >0.50, the species is then the best ecological indicator for the cluster. These species are generally dominant or sub-dominant in the cluster. When a species is abundant in several clusters, ,1 decreases under 0.50, this means that the species is not strongly allied to one cluster. However, this species is able to give additional ecological information on the clusters where it is abundant (e.g., Navicula joubaudii is abundant in clusters 3 and 1, Navicula cari in clusters 4 and 5). The paleoecological interpretation was based on a large range of modern environments in Bolivia. The modern data set is composed of 58 samples collected in several shallow salt lakes of the southernmost part of the Bolivian Altiplano in the Lipez area (ServantVildary, 1984, Servant-Vildary and Roux, 1990), Lake Poop6 (Servant-Vildary, 1978), in the border of Salar de Uyuni (Sylvestre, 1997), and 90 samples in lakes and wetlands of a glacial valley (Hichu Kkota) located north-east of Lake Titicaca (Servant-Vildary, 1982, 1986; Miskane, 1997). Most of the fossil species were found in the modern assemblages, even if the modern analogues cannot be found in the non-glacial valleys which are now desiccated for most of the time. On the metric to hectometric scale, the wetlands are characterized by a mosaic of aquatic environments including wet meadows, flowing streams and more or less extended shallow water bodies. According to a detailed study of the diatom flora in the wetlands of the gla-

271 Table 1. Results of cluster analysis of 182 species in 39 samples. Mean abundance of 30 selected species in seven clusters Genera

Species/Samples:

Cluster 2 Cluster 3 Cluster 4 Cluster 7 Cluster 5 Cluster 1 Cluster 6 Mean Frequency Correl84,101, 97,178, 113,119, 681, 346 22,70 656 ation 149, 152, 221,423, 437,443, 464,529, 571

ACHNANTHES AMPHORA ANOMOEONEIS CALONEIS COCCONE~

CYCLOTELLA CYMBELLA CYMBELLA DENTICULA EPITHEMIA FRAGILARIA FRAGILARIA FRAGILARIA FRAGILARIA HANTZSCHIA NAVICULA NAVICULA NAVICULA NAVICULA NAVICULA NAVICULA NITZSCHIA NITZSCHIA NITZSCHIA NITZSCHIA NITZSCHIA NITZSCHIA NITZSCHIA NITZSCHIA RHOPALODIA

LANCEOLATA 23 VENETA 2 SPHAEROPHORA 0 BACILLUM 26 PLACENTULA 8 KUElZINGIANA 15 FALAISENSIS 5 MESIANA 2 VALIDA S9 ARGUS 34 CAPUCINA 0 FASCICULATA 0 LAPPONICA 14 ULNA 0 AMPHIOXYS 8 CARl 58 ClNCTA 4 COHNII 8 GALLICA 438 JOUBAUDII 51 SEMINULUM 19 AMPHIBIA 8 COMMUTATA 4 FRUSTULUM 3 HANTZSCHIANA 10 LlNEARIS 1 PALEA 5 PSEUDOAMPHIBIA 3 UMBONATA 2 GIBBERULA 5

188,190, 128, 137, 692, 257,402, 158,608, 806 521,542, 633 714

20 15

4

o

o

23

26 15 37 2

6

4

18 25 10 63

9

3

42 1

13 2 4

3 5

194

6

277 39 8

o o

3

o

21

o o

70

19 72

137 20 33

45 6

o 3

39 9 9

o

15 12

55 16

73

6

18 6

40

7 11 15 5 2 24

4 4

o 6

14

11

122 62 12

o 4

14

184,238, 276,310, 325

o 2 264

o o o o

44 2

I 7

4

4

20 28 15 22 26 26 58

32 23 19 23 26 30 30

3

3

o o

88 98 17

o

18

o o 2

o 80 174

222

397

4

5 93

8

4

o

12

104

o o 4

38 94

4

o

7

10

o o

2 7

14 27

168

4

15

14 4

o o o

o o o

o

3 87

o o o 258

o

o

1 72 7 1 96

174 77 26 1

o

6

o 12 8 14 14 16

o

12

o o o

47

4

7 11 13

20

9

o o o

5 74 7 II 197 84 38 12 3

5 4

13 18 4

12 9

26 19

0.06 0.16 1 0.09 0.97 OJ OJ1 0.17 OJ5 0.15 0.97

2 27 11 25 39 18 19 39 38 31 25 11 10 17 25 28 14 17 21

0.74 0.58 OJ5 0.21 OJ OJ5 0.67 0.21 0.2 0.06 0.27 0.15 0.23 0.15 0.16 0.16 0.91 0.04

Thirty species were selected: 27 were present in more than nine samples, three present only in four, three and two samples were also included because they were very abundant. The mean abundance (Xk) of each species (X) in samples (i) which belong to a cluster (k) is used for ecological reconstruction of a cluster when Xk > i (i is the mean abundance of the species in the data set). The two columns on the right give respectively the 'Frequency' which is the number of samples where each species is present and the correlation coefficient 'Correlation' ,2, such as: ,2 =BSSITSS.

cial valley of Ichu Kkota, two main groups of samples were recognized. The first one, collected in aerophilic habitats, is composed of more than 50% of aerophilous species. The second one, collected in permanent shallow waters (water-depth 5mllont • 1-6 million t • 0.1-1 mliiont C PotentIally IIgnIftcant depoeIt, but ___ not quantltled



a

oI

Scale 50 km I

Figure 1. Distribution and size of sodium sulfate deposits of the Northern Great Plains (after Last & Slezak. 1987).

recently, Last and other workers have studied saline lake sediments as sensitive proxies for past climate variations (Last, 1992; Vance et aI., 1992, 1993; Lemmen et aI., 1997). The economic geology of the sodium sulfate deposits of the region has been welldocumented (Cole, 1926; Witkind, 1952; Tompkins, 1954; Grossman, 1968; Rueffel, 1970; Broughton, 1984; Last & Slezak, 1987; Murphy, 1996). Most authors have noted that the deposits are located in areas of active groundwater discharge, as manifested by springs and seeps peripheral to the lakes and in the lake beds, and agree that groundwater is a causative agent for the formation of the deposits. Previous workers disagree, however, on the ultimate source of ions and the scale of groundwater flow systems that deliver those ions to discharge at the deposits. McIlveen & Cheek (1994) summarized the hypotheses proposed by

previous workers on the sources of dissolved Na+ and S04= ions as follows: (1) till (derived primarily from underlying Cretaceous marine shales) containing abundant smectite with exchangeable sodium, (2) Cretaceous or older marine rocks containing bentonite with exchangeable sodium, (3) connate water from marine rocks, and (4) dissolution of deeply buried (> 1000 m) Paleozoic evaporites. Flow systems that previous workers have cited as potential agents for the transport of ions to the surface include: (I) Meteoric waters flowing over the land surface (runoff)

281 (2) Shallow flow systems, involving groundwater of Recent (meteoric) origin andlor Pleistocene (glacial) origin that circulates through fractured till andlor intertill aquifers (3) Flow systems of intermediate depth that involve Recent meteoric water, Pleistocene water, or Cretaceous connate water (4) Deeply-circulating flow systems involving Paleozoic connate water or water that dissolved salt from the Devonian Prairie Evaporite or other evaporites in the Western Canada Basin.

nian Prairie Evaporite, have disrupted the horizontal continuity of the underlying stratigraphy over much of southern Saskatchewan (Christiansen, 1967a; Broughton, 1988). Furthermore, the bedrock surface was strongly modified by riverine erosion that took place prior to the onset of Pleistocene glaciation (Christiansen, 1967b). Witkind (1952), Grossman (1968) and Rueffel (1970) noted that alkali lakes often occupy elongate surface depressions that overlie pre-glacial drainage Valleys.

None of the previous workers tested their hypotheses. Our initial focus has been on testing the hypotheses presented by previous workers regarding groundwater flow path(s) and solute source(s). Therefore, we have undertaken a reconnaissance study of the H, 0 and Sr isotope compositions of groundwater associated with six Saskatchewan lakes that contain documented sodium sulfate resources, in an attempt to evaluate the tracer potential of H, 0 and Sr isotopes as tools for isotopically fingerprinting specific aquifer inputs to a lake.

Economic geology

Geologic setting Regional geology

Southern Saskatchewan and adjacent parts of Montana, North Dakota and Alberta are underlain by a thick (in excess of 1000 m) sequence of nearly-horizontal sedimentary rock. The Paleozoic section consists predominantly of carbonates and evaporites, whereas the Mesozoic rocks are dominantly marine clastic sedimentary rocks. The Cenozoic section is composed of nearshore and terrestrial clastic sedimentary rocks. The region was subjected to multiple episodes of glaciation during the Pleistocene. The unconsolidated glacial, glaciofluvial and glaciolacustrine sediment (drift) that mantles the bedrock is over 300 m thick in places, and averages about 100 m thick in southern Saskatchewan (Simpson, 1997). The drift is derived primarily from the Cretaceous marine shale (principally the Bearpaw Formation) that underlies the southern part of the province, with varying proportions of the Paleozoic carbonates and Precambrian crystalline rocks that crop out in the northern part of the province. During deglaciation, meltwater carved numerous channels and spillways in the glacial sediment. Collapse structures caused by the dissolution of Paleozoic evaporites, most notably from the Devo-

Three types of sodium sulfate accumulations are recognized in saline lakes of the northern Great Plains. (1) As lake brine. (2) As beds of intermittent crystal, deposited from the brine in autumn as the ambient temperature cools. Intermittent crystal re-dissolves the following spring with dilution of the brine by runoff by the Spring snow-melt. (3) As permanent beds of mirabilite mixed with other salts, clastics, and organic sediment. The permanent crystal beds are typically 1-5 m thick, but exceed 30 m in a few deposits. Most Saskatchewan companies that mine Na2S04 pump brine into crystallization ponds, where it is concentrated by evaporation over the summer. In the fall, as the ambient temperature cools, Glauber's salt (the commercial name for mirabilite) crystallizes and accumulates on the pond liner. The pond is then drained of any remaining liquid, and the Glauber's salt is harvested from the pond. At one operation, intermittent and permanent crystal is mined from the lake bed using a dredge-mounted excavator. Solution mining has been tried in the past, but is not currently being used. Figure 2 illustrates schematically the three types of sodium sulfate accumulations and mining methods.

Hydrogen, oxygen and strontium isotope systematics in groundwater systems Water isotopes

Elements of low atomic mass, such as hydrogen or oxygen, can undergo mass-dependent isotope fractionation as a result of chemical reactions, diffusion or phase changes. Multiple increments of isotopic fractionation are recorded by water as it navigates the

282

Brine Lake

+ ~ Intermittent

~"-+"---;,~~

Crystal Bed

+----Permanent Crystal Bed

NOT TO SCALE

Drying

AnhYdrols sodium sulfate

2. Schematic cross-section of idealized Saskatchewan sodium sulfate deposit. Mining methods illustrated are: (I) Evaporative concentration of lake brine in crystallization pond, followed by precipitation of crystals of Glauber's salt (mirabilite) as the brine cools in autumn; (2) Solution mining of thick crystal beds. The brine return is treated as in I; and (3) Dredging of lake bottom crystal beds. Figure

Earth's meteoric water cycle. Therefore, continental surface waters and groundwaters exhibit isotope compositions indicative of water source and history of movements in the global water cycle. The isotopic composition of water changes with evaporation, condensation, freezing, thawing, sublimation, reaction with hydrous minerals and mixing with other waters. Oxygen and hydrogen isotopes are measured as isotope abundance ratios (80/ 16 0, DIH), and reported as 818 0 and 8D, defined as: 8I 80sample= {[(80/ 16 0)sample/ cI 80/ 16 0)SMow]-I}.1000%0 The 818 0 value represents the permil (%0) enrichment or depletion of 18 0 in the sample relative to Standard Mean Ocean Water (SMOW). For example, a sample with 818 0 = -25%0 contains 25 parts-per-thousand (or 2.5%) less 18 0 than standard mean ocean water. The convention for hydrogen is identical, with D eH) taking the place of 18 0 and H (I H) taking the place of 16 0 (for reviews, see Kendall & Caldwell, 1998 and Clark & Fritz, 1997).

Strontium isotopes Although the relative mass differences between the four isotopes of Sr are too small to exhibit natural mass dependent isotope variation, one isotope, 87 Sr, varies in abundance relative to the other isotopes of Sr because it is the natural radioactive decay product of 87Rb. By convention, differences in 87 Sr abundance are reported as the ratio 87 Sr/86 Sr. Old crustal rocks have typically high 87Srl86Sr ratios, whereas younger crustal rocks, and marine carbonates, have lower 87Sr/86Sr ratios. Strontium readily dissolves in aqueous solutions, following calcium in geochemical behavior. Strontium isotopes (87Sr/86Sr) are generally conservative in mixing of waters in lakes or estuaries. In surface waters, strontium concentration may be changed by precipitation or dissolution of calcite, gypsum or other Sr-containing minerals, but the 87 Sr/86 Sr ratio is not affected by these reactions. This is in contrast to Hand 0 isotopes, which are often non-conservative in mixing of surface waters due to evaporation effects causing enrichment of the residual water in D and 18 0.

283 In groundwater systems dissolved Sr is less conservative than Hand 0 isotopes, due to Sr-exchange between rocks and water along the flow path. The degree of non-conservative behavior is a function of the groundwater velocity and rock type. In general, faster moving groundwater is more conservative for Sr isotopes than slower moving groundwater, the latter undergoing more exchange of strontium with minerals in the rocks. Therefore, the isotopic composition of dissolved Sr in a slow moving groundwater will evolve towards a 87Sr/86Sr ratio that is similar to the rock (Bullen & Kendall, 1998; Bullen et aI., 1996; Johnson & DePaolo, 1997a, b).

Methods Field methods

During 1998, we sampled groundwater from seeps, discrete spring orifices, and shallow wells in the immediate vicinity of four saline lakes (Vincent, Corral, Boot and Chain) that host documented sodium sulfate resources. In 1999, we sampled groundwater and lake brine at three ofthe lakes sampled during 1998 (Vincent, Corral, Chain) and two additional lakes (Grandora N., Whiteshore). The lakes that were sampled are shown in Figure 3. Water samples were filtered in the field using 0.45 11m filters and stored in acid-cleaned 250 ml highdensity polyethylene (HDPE) bottles. The samples were split, and one aliquot was acidified to pH=2. Alkalinity, major anions, 8D and 818 0 were determined on the non-acidified sample, whereas 87Sr/86Sr and major cations were determined on the acidified sample. Specific gravity, temperature, and pH were measured at each sampling event. The well samples are generally from depths of 5-30 m in domestic or stock (mostly flowing) wells within a few hundred meters of the respective lake. Exceptions include a municipal well that is "-'50 m deep at Chain Lake, which samples an intertill aquifer; and two wells that are approximately 150 m deep at Whiteshore Lake. Both White shore wells sample the Judith River Formation, an upper Cretaceous aquifer with good water quality. The spring samples are from springs or seeps that discharge near the lakeshore, generally at or above lake level. The discharge water immediately mixes with the lake brine. Lake brines were sampled in locations remote from springs and seeps. Where more than one sample

was collected from a lake (Vincent, Grandora, and Whiteshore), the sampling locations are up to several kilometers apart, in order to capture any compositional heterogeneity. Laboratory methods

The 0 isotope composition of water was analyzed using the C02-H20 equilibration method of Epstein & Mayeda (1953) using an equilibrium isotope fractionation factor of 1.0412. Hydrogen isotope analyses were performed by the V-reduction methods of Bigeleisen et al. (1952). Hand 0 isotope analyses were performed on a Finnigan MAT Delta E, gas isotope-ratio mass spectrometer. External precision is based on the reproducibility of our internal water standard over the course of this work, and is better than ±2%0 (10-) for 80, and 0.2%0 (10-) for 818 0. Strontium was purified from the solution using standard cation exchange chromatograpy in a clean lab environment. Purified Sr was loaded onto a single Re filament with Ta-gel and measured using a multidynamic peak-hopping routine on a Finnigan MAT 261 thermal ionization mass spectrometer. The 87 Sr/86 Sr ratios were normalized to 88Sr/86 Sr = 8.375209. External precision is ±0.00002 (20-), based on multiple analysis of the SRM 987 SrC03 standard which yielded 0.71026 during the course of this study. Major element concentrations were performed by ICP-AES at the Saskatchewan Research Council with external precision typically ± 10% (20-). Evaluation of tracer potential Water isotopes

Field and analytical results for all samples collected in this study are summarized in Table 1. Hydrogen and oxygen isotope results for groundwater sampled from springs, seeps and shallow wells near sodium sulfate deposits ranged from approximately -145%0 to -170%0 for 80 and from -15%0 to -20%0 for 818 0 (Fig. 4). These data are similar to the range reported for shallow Saskatchewan groundwater by McMonagle (1987) and very depleted of D and 18 0 compared to the data from deeper saline Paleozoic aquifers reported by Rostron et al. (1998). The data fall between the averages for Saskatchewan meteoric water (8D ~ -138%0 and 818 0 ~ -17%0 (McMonagle, 1987)) and a component of porewater in the glacial till that was last recharged during the

G2-I G2-2 G2-3 G2-4 G2-5

Grandora North

VL-4 VL-5 VL-6 VL-7 Vincent-4 Vincent -4 Vincent -I Vincent -I Vincent-2 Vincent -2 Vincent -3 VL-9

CL-3 Corral-I Corral-I CI-I CI-5

ChL-2 Chain-I Chain-I Chain-2 Chain-2 Chain-3 Chain-3

Boot-I Boot-2

Vincent Lake

Corral Lake

Chain Lake

Boot Lake

Whiteshore Lake WS-I WS-3 WS-2 WS-5 WS-6

Station

Lake name

Sample type

lake brine municipal well municipal well spring spring domestic well domestic well

lake brine spring spring spring-fed pond spring-fed pond

lake brine lake brine lake brine lake brine stock wen stock well stock well stock well stock well stock well stock well spring

lake brine lake brine bedrock water well spring-fed pond bedrock water well

29-May-98 spring 29-May-98 spring

28-May-99 21-Jul-98 28-May-99 21-JuI-98 28-May-99 21-JuI-98 28-May-99

27-May-99 20-JuI-98 27-May-99 27-May-99 27-May-99

26-May-99 26-May-99 26-May-99 26-May-99 26-May-99 23-Jul-98 26-May-99 23-Jul-98 26-May-99 23-Jul-98 23-Jul-98 29-Jul-99

28-Apr-99 29-Apr-99 30-Apr-99 ll-Mar-99 ll-Aug-99

26-Apr-99 lake brine 27-Apr-99 lake brine 28-Apr-99 spring sand point well water well

Date

Table 1. Summary of field and analytical results

8.55 7.4 7.7 7.4 7.62 7.6 7.59

21.1 7.5 8.3 9.0 4.6 7.0 6.5 6.5 7.3 10 7.3

8.2 7.5 7.74 8.71 8.63

8.70 8.82 nd 8.54 7.54 7.3 7.50 7.6 7.76 7.7 7.4 nd

8.28 9.45 nd 7.20 8.49

23.4 15.0 15.2 19.7 24.0

18.8 15.4 nd 15.4 7.4 8.0 8.3 8.0 9.5 14.0 12.0 nd

16.4 17.9 nd 22.5 5.8

Na+ 665 614 50 21 31

0.998 0.998

113 819

1.150 49000 0.992 200 1.000 218 0.992 331 1.000 345 1.000 86 1.000 90

1.135 46500 0.998 16 nd 32 1.008 1720 1.000 245

1.140 51 100 1.020 7000 1.134 38400 1.135 55700 nd 1330 1.000 1410 nd 1180 1.000 1050 nd 1060 1.002 1140 1.000 118 nd 51

108 441

4920 104 90 87 93 128 III

2290 27 36 902 65

2040 302 2600 2440 49 46 23 18 21 20 56 25

165 146 3 3 5

97 191

664 54 56 115 135 87 128

395 94 112 49 46

210 47 270 176 113 68 63 35 64 50 85 37

6 13

3 7 7 5 6

3

454

208 1 3 81 19

191 39 262 202 11 9 5 6 5 7 8 14

488 3110

6.3 126000 0.7 422 0.7 423 903 0.9 0.8 912 0.7 520 0.7 567

3.7 107000 0.2 55 0.6 150 1.1 5070 0.8 616

1.0 2.0

4580 4310 42 56 139

51 168

72 9 9 17 16 15 16

25

1380

90 4 35

7050 1300 8850 6700 201 200 263 262 280 283 16 8

860 nd 550 nd 526 nd 452

80

nd nd

13

600 800

460 nd 338 643 291

;.



__

._-

1000





o. ..._c~ .........

10

........ fIIww ......

-'

0

o

01

00

00

01

Log[SO:IHCO,l .0

10,

Figure 5. Na+; Ca2+ vs. S042-IHC03- for samples from this study compared with data from various shallow «200 m) aquifers reported by McMonagle (1987).

between the Judith River and intertill compositions. The highly permeable Empress Group occupies paleovalleys that cut through various other, generally less permeable, aquifers. Van der Kamp (1986) characterized the valley-fill aquifers of the Empress Group as regional drains. The intermediate composition of Empress Group waters may reflect the homogenization of multiple drained aquifers with different water chemistry. Strontium isotopes

Results for strontium isotopes are summarized in Figure 6. Each of the three lake brines (Vincent, Grandora, Whiteshore) that were sampled in multiple locations showed little spatial variation in 87 Sr/86 Sr. This indicates that the lake brines are isotopically wellmixed, despite being exceedingly shallow (commonly 10-50 cm) and dense (SG > 1.15). Strontium isotopic compositions are generally within the range derived for Phanerozoic seawater (Burke et al., 1982), perhaps, in part, reflecting the presence of authigenic marine precipitates in the underlying Cretaceous marine shale (Bearpaw Formation) and derived glacial till. Alternatively, the provenance of the clastic component in the shales is known to be the Western Cordillera, which consists of relatively juvenile crust with low 87 Sr/86 Sr. The average of 12 rivers draining the Western Cordillera, today, yields a mean 87Sr/86Sr of 0.7077 (Holmden et aI., 1997). Therefore, cation-exchange between reworked Cretaceous sediments and Recent groundwater can also explain the relatively low 87 Sr/86 Sr ratios for the ground waters measured in this study. The groundwater samples from the White shore Lake wells screened

in the Judith River Formation are a case in point. The sediment that was deposited to form the Judith River Formation was eroded from the Cordillera in the late Cretaceous, in a continental setting, with material transported across the Great Plains by streams discharging from the Cordillera. The very low 87 Sr/86 Sr of 0.7065 is too low to be derived from authigenic seawater precipitates, but consistent with Sr-exchange between groundwater and aquifer materials derived from juvenile igneous rocks of the Cordillera. In other cases, samples with much higher 87Sr/86Sr (e.g. the domestic well at Chain Lake, with 87Sr/86Sr = 0.7099) most likely reflect greater proportions of distallysourced, glacially-transported Precambrian detritus from the Canadian Shield, mixed with the reworked Cretaceous shales. The 87 Sr/86 Sr ratio for groundwater discharging from springs and seeps near the shores of lakes is generally similar in isotopic composition to the lake brine. For example, 87Sr/86Sr in Grandora Lake brine (mean 87Sr/86Sr of 0.70824) is virtually identical to that of discharge spring waters (0.70830). The 87Sr/86Sr for Vincent Lake brine (0.70855) is bracketed by two discharge springs (0.70836 and 0.70938). One notable exception is Chain Lake, where groundwater discharging into the lake from one site had lower (less radiogenic) 87Sr/86Sr ratios than the lake brine. In general, groundwater sampled from shallow wells near the lakes had lower 87Sr/86Sr ratios than either lake brine or groundwater discharging from springs. Again, Chain Lake is the exception. The 87Sr/86Sr of groundwater sampled from a spring that discharges directly into Chain Lake is lower than the lake water, and very close to that of groundwater sampled from a nearby municipal well that is 50 m deep. The 87Sr/86Sr ratio of water sampled from a domestic well, in which the water level is approximately equal to the elevation of the lake, was intermediate between the compositions of the lake brine and the spring discharge. Although the Sr isotope systematics of Chain Lake are at first glance more complicated than the other lakes, the utility of using Sr isotopes as a sensitive tracer of the relative contributions of water and salts from specific aquifers is amply demonstrated. Figure 6 also shows the effect of Ca-mineral precipitation on lake water Sr/Ca ratios. Calcite and gypsum discriminate against Sr during precipitation (Kushnir, 1984) causing the lake brine to increase in Sr/Ca ratio. For example, the 1000 (Sr/Ca) of Vincent Lake brine samples are all in excess of nine while the

287 07110

0.7105 0.7100

07095

..

07090

~

07085

(/) (/)

.... eo

• •

0.7080

0.7075

0.7070

0.7065

07060 0

2

4

6

8

10

12

14

1000{Sr/Ca molar ratio) Figure 6. Plot of 87Sr/86Sr vs. lOOO(Sr/Ca molar ratio). The number in the lower right of each box in the legend indicates the number of samples analyzed from that class. Wells near Whiteshore Lake are screened in the Judith River Fm. All others are screened in intertill aquifers.

1000 (Sr/Ca) in the discharge spring waters (that have 87 Sr/86 Sr ratios similar to the lake brine) are about four. This is indicative of in situ precipitation of calcite and/or gypsum within Vincent Lake. Therefore, lake brines with 87Sr/86Sr ratios similar to those of aquifer inputs, but higher Sr/Ca ratios, suggest Ca-mineral precipitation. Strontium isotopic ratios for the lake brines were mostly similar to those of groundwater sampled from springs and seeps that discharged directly into the lakes. However, these discharging spring waters had different 87Sr/86Sr ratios than groundwater sampled from intertill aquifers in wells screened at depths of 5-30 m. We offer two possible explanations for this observation. First, there may be no connection between the groundwater discharging through springs and the shallow groundwater sampled in wells near sodium sulfate deposits. The spring discharge may be the result of very shallow local flow cells that are recharged by precipitation and snowmelt from the highlands around the lake basin. We consider this hypothesis unlikely because the springs flow year-

around, and some have flowed at the same location for generations, based on the recollections of landowners and deposit descriptions of Cole (1926) and Tompkins (1954). Shallow local flow cells seem unlikely to be the source of such persistent discharge over such long periods of time. An alternative explanation is that the discharge springs are connected to the larger and deeper flow systems that are sampled by the wells, but that 87 Sr/86 Sr undergoes considerable evolution due to rock (till)-water interaction as the water moves toward the discharge springs. In three of the four lakes studied, inferred evolution of 87Sr/86Sr is towards higher isotopic compositions. However, the specific direction of Sr isotope evolution depends on the isotopic composition of the initial fluid and the isotopic composition of the aquifer component that is most susceptible to Sr-exchange, and the homogeneity of the aquifer (Johnson & DePaolo, 1997a). Johnson & DePaolo (1997b) described Sr isotope evolution along a flow path through a homogenous low-permeability siltstone aquifer and were able to relate changes in

288

87 Sr/86 Sr to fluid flow rates, and identify preferential fluid flow paths. The large variations seen thus far in Sr isotope compositions of saline lakes, and associated springs and groundwater indicates that there is sufficient isotopic sensitivity to be of use in constructing water balances in these lakes. For example, we know that the Sr isotope balance of Chain Lake is incomplete in that the lake has a higher Sr isotope composition (0.7104) than any of the groundwaters (0.7088-0.7099) or discharge springs (0.7090) that we have measured, thus far. Clearly, there is another aquifer input of Sr to this lake that we have not identified. Furthermore, Sr isotopes appear to be a powerful tool for elucidation of the mechanisms controlling till-water Sr isotope interactions, which in tum may improve our understanding of how groundwaters evolve towards the Na-S04 compositions needed to form a Na2S04-dominated evaporite deposit in a c1osed-basinlake.

Conclusions Sodium sulfate deposits of the northern Great Plains continue to be supplied with dissolved ions via groundwater discharge. The net rate of salt accumulation in these modem ore-forming systems is related to fluid and solute mass balances. The geochemical tracers investigated in this reconnaissance study will be tested in more detailed hydrogeological investigations of individual deposits. The water isotope data suggests that there are subtle isotopic differences in H and 0 isotopes that could be useful for discriminating aquifers with a substantial component of seasonally recharged water from aquifers that carry older, more evolved waters, or those that have a component of Pleistocene glacial melt water. Comparison of ion ratios for data from this study with data of McMonagle (1987) suggests that groundwater sampled from springs and seeps near sodium sulfate deposits may be at least partly from a bedrock source, perhaps mixed with shallower groundwater during ascent. Strontium isotopes hold the greatest promise for fingerprinting aquifer inputs to the saline lake basins that host sodium sulfate deposits. 87 Sr/86 Sr ratios must be examined in the context of flow-path evolution, constrained by physical hydrogeology, to be ultimately useful in identifying solute source(s), elucidating till-water evolution of the major cations (Na,

Ca), and quantifying aquifer input(s) to the lake basins that host sodium sulfate deposits.

References Bigeleisen, J., M. L. Perlman & H. C. Prosser, 1952. Conversion of hydrogenic materials to hydrogen for isotopic analysis. Anal. Chern. 24: 1356-1357. Broughton, P. L., 1984. Sodium sulphate deposits of western Canada. In Guillet, G. R. & W. Martin (eds), The Geology ofIndustrial Minerals in Canada. Can. Inst. Min. Met. Special Volume 29: 195-200. Broughton, P. L., 1988. Formation of Tertiary coal basins in southern Saskatchewan. Sask. Energy Mines, Open File Rep. 88-1: 53 pp. Bullen, 1. D. & C. Kendall, 1998. Tracing of weathering reactions and water flowpaths: a mUlti-isotope approach. In Kendall, C. & J. J. McDonnell (eds), Isotope Tracers in Catchment Hydrology. Elsevier Science B.V, Amsterdam: 611-646. Bullen, 1. D., D. P. Krabbenhoft & C. Kendall, 1996. Kinetic and mineralogic controls on the evolution of groundwater chemistry and 87Sr/86Sr in a sandy silicate aquifer, northern Wisconsin, U.S.A. Geochim. Cosmochim. Acta. 10: 1807-1821. Burke, W. H., E. A. Denison, R. B. Hetherington, R. B. Koepnick, H. F. Nelson & J. B. Otto. 1982. Variation of seawater 87 Sr/ 86 Sr ratio throughout Phanerozoic time. Geology 10: 516-519. Christiansen, E. A., 1967a. Collapse structures near Saskatoon, Saskatchewan, Canada. Can. J. Earth Sci. 4: 757-766. Christiansen, E. A., 1967b. Preglacial valleys in southern Saskatchewan;. Sask. Resear. Counc., Map 3, Scale I: 1500000. Clark, I. & P. Fritz, 1997. Environmental Isotopes in Hydrogeology. Lewis Publishers, Boca Raton: 328 pp. Cole, L. H., 1926. Sodium sulphate of western Canada: Occurrence, uses, and technology. Canada Department of Mines, Report 646, Ottawa: 160 pp. Epstein, S. & 1. K. Mayeda, 1953. Variations of the 18 0/ 16 0 ratio in natural waters. Geochim. Cosmochim. Acta. 4: 213. Grossman, l. G., 1968. Origin of the sodium sulfate deposits of the northern Great Plains of Canada and the United States. Prof. Pap. 600-B: B104-B 109. Hammer, U. 1., 1978a. The saline lakes of Saskatchewan. I. Background and rationale for saline lakes research. Int. Rev. ges. Hydrobiol. 63(2): 173-177. Hammer, U. 1., I 978b. The saline lakes ofSasktachewan III. Chemical characterization. Int. Rev. ges. Hydrobiol. 63: 311-335. Hammer, U.T., 1986. Saline Ecosystems of the World. Monographiae Biologicae, 59. Dr W. Junk Publishers, Dordrecht: 616 pp. Hammer, U. 1. & R. C. Haynes, 1978. The saline lakes of Saskatchewan II. Locale, hydrography and other physical aspects. Int. Rev. ges. Hydrobiol. 63: 179-203. Hendry, M. J. & L. I. Wassenar, 1999. Implications for transport of 8D in porewaters for groundwater flow and the timing of geologic events in a thick aquitard system. Wat. Resour. Res. 35: 17511760. Holmden, C., K. Muehlenbachs & R. A. Creaser, 1997. Depositional environment of the early Cretaceous Ostracode Zone: Paleohydrologic constraints from 0, C and Sr isotopes. In Pemberton, S. G. & D. P. James (eds), Petroleum Geology of the Cretaceous Mannville Group, Western Canada. Memoir. Canadian Society of Petroleum Geologists, Calgary: 77-92.

289 Johnson, T M. & D. 1. DePaolo. 1997a. Rapid exchange effects on isotope ratios in groundwater systems I. Development of a transport-dissolution-exchange model. Wat. Resour. Res. 33(1): 187-195. Johnson, T M. & D. J. DePaolo, I 997b. Rapid exchange effects on isotope ratios in groundwater systems 2. flow investigation using Sr isotope ratios. Wat. Resour. Res. 33(1): 197-209. Kendall, C & E. A. Caldwell, 1998. Fundamentals of isotope geochemistry. In Kendall, C & J. 1. McDonnell (eds), Isotope Tracers in Catchment Hydrology. Elsevier Science, Amsterdam: 51-86. Kushnir, K., 1984. The co-precipitation of strontium, magnesium, sodium, potassium and chloride ions with gypsum: an experimental study. Geochim. Cosmochim. Acta. 44: 1471-1482. Last, W. M., 1984. Sedimentology of playa lakes of the northern Great Plains. Can. J. Earth Sci. 21: 107-125. Last, W. M., 1988. Salt lakes of western Canada: a spatial and temporal geochemical perspective. In Nicholaichuk, W. & H. Steppuhn (eds), Proceeding of the Symposium on Water Management Affecting the Wet-to-Dry Transition: Planning at the Margins. Water Studies Institute. University of Regina, Regina: 99-113. Last, W. M., 1989a. Continental brines and evaporites of the northern Great Plains of Canada. Sedimentary Geology 64: 207-221. Last, W. M., 1989b. Sedimentology of a saline playa in the northern Great Plains, Canada. Sedimentology 36: 109-123. Last, W. M., 1990. Paleochemistry and paleohydrology of Ceylon Lake, a salt-dominated playa basin in the northern Great Plains, Canada. 1. Paleolimnol. 4: 219-238. Last, W. M., 1992. Salt lake paleolimnology in the northern Great Plains: the facts, the fears, the future. In Robarts, R. D. & M. L. Bothwell (eds), Aquatic Ecosystems in Semi-arid Regions: Implications for Resource Management. National Hydrology Research Institute Symposium. Environment Canada. Saskatoon: 51-62. Last, W. M., 1993. Geolimnology of Freefight Lake: an unusual hypersaline lake in the northern Great Plains of western Canada. Sedimentology 40: 431-448. Last, W. M. & T H. Schweyen, 1983. Sedimentology and geochemistry of saline lakes of the Great Plains. Hydrobiologia 105: 245-263. Last. W. M. & L. A. Slezak, 1986. Paleohydrology, sedimentology and geochemistry of two meromictic lakes in southern Saskatchewan. Geographie physique et Quaternaire. 40(1): 5-15.

Last. W. M. & L. A. Slezak, 1987. Sodium sulphate deposits of western Canada: Geology, mineralogy and origin. In Gilboy, C. F. & L. W. Vigrass (eds), Economic Minerals of Saskatchewan. Saskatchewan Geological Society, Special Publication 8. Regina: 197-205. Lemmen, D. S., R. E. Vance, S. A. Wolfe & W. M. Last, 1997. Impacts offuture climate change on the southern Canadian prairies: a paleoenvironmental perspective. Geoscience Canada 24(3): 121-133. McIlveen, S. & R. L. Cheek, 1994. Sodium sulfate resources. In Carr, D. D. (ed.), Industrial Minerals and Rocks, 6th edn. Soc. Mining. Metal. Explor: 1129-1158. McMonagle, A. L.. 1987. Stable isotope and chemical compositions of surface and subsurface waters in Saskatchewan. M. Sc. Thesis. University of Saskatchewan, Saskatoon: 108 pp. Murphy, E. C, 1996. The sodium sulfate deposits of northwestern North Dakota. North Dakota Geological Survey. Report 99: 73 pp. Rostron, B. 1., C Holmden & L. K. Kreis. 1998. Hydrogen and oxygen isotope compositions of Cambrian to Devonian formation waters, Midale area, Saskatchewan, Proceedings, 8th International Williston Basin Symposium. Sask. Saskatchewan Geo!. Soc., Spec. Pub!. 13: 267-273. Rueffel, P. G.. 1970. Natural sodium sulfate in North America, 3rd Symposium on Salt. Northern Ohio Geological Society. Columbus: 429-451. Simpson. M. A.. 1997. Surficial Geology Map of Saskatchewan. Sask. Energy Mines/Sask. Resear. Coune., scale 1:1 000000. Tompkins, R. V, 1954. Natural sodium sulphate in Saskatchewan. Sask. Dep. Miner. Resour.. Report 6: 71 pp. Van der Kamp. G.. 1986. The groundwater resources of Saskatchewan's deep buried-valley aquifers, Canadian Hydrology Symposium No 16, Regina: 529-543. Vance, R. E., 1. J. Clague & R. W. Matthewes. 1993. Holocene paleohydrology of a hypersaline lake in southeastern Alberta. 1. Paleolimnol. 8: 103-120. Vance, R. E.. R. W. Matthewes & 1. 1. Clague. 1992. 7000 year record oflake-Ievel change on the northern Great Plains: A highresolution proxy of past climate. Geology 20: 879-882. Witkind, I., 1952. The localization of sodium sulfate deposits in northeastern Montana and northwestern North Dakota. Am. 1. Sci. 250: 667-676.

Hydrobiologia 466: 291-297,2001. J.M. Melack, R. Jellison & D,E, Herbst (eds), Saline Lakes, © 2001 KhlWer Academic Publishers,

291

Benthos of a seasonally-astatic, saline, soda lake in Mexico Javier Alcocer!, Elva G, Escobar2, Alfonso Lugo 1, L Maritza Lozano2 & Luis A, Oseguera2 1Limnology Lab, Environmental Conservation & Improvement Project, UIICSE, FES Iztacala, UNAM, Av. de los Barrios sin, Los Reyes /ztacala, 54090 Tlalnepantla, Edo. de Mexico, Mexico Fax: +52 5277.1829. E-mail: [email protected] 2Instituto de Ciencias del Mar y Limnologia, UNAM, Apdo. Postal 70-305, Coyoacan 04510, Mexico, D.F., Mexico

Key words: benthic macroinvertebrates, physical and chemical variables, crater-lake, tropical lake, shallow lake, temporary pond

Abstract The benthic macroinvertebrate community (BMC) of Lake Tecuitlapa Sur, central Mexico, was monitored to determine the structure of the community (i.e. species composition, richness, abundance and biomass), throughout an annual cycle. Tecuitlapa Sur is shallow, seasonally-astatic, warm, meso saline, and soda-alkaline. The physical, chemical and biological variables were determined monthly for a yearly cycle. Tecuitlapa Sur displayed a seasonal patterns of dilution (June-August) and concentration (September-November) phases. Salinity and pH were the most important parameters explaining environmental variance. The BMC consisted of two species: Culicoides occidentalis sonorensis (Diptera: Ceratopogonidae) and Tanypus Apelopia sp. (Diptera: Chironomidae). C. occidentalis was the most important species both numerically and in biomass (2':95%). Annual density (mean ± sd) of C. occidentalis (1141 082 ± 2765879 indo m- 2 , n = 120) was notably higher than other reported for other saline water bodies. However, the mean annual density of T. Apelopia (6782 ± 8310 indo m -2, n = 120) was similar to other saline lakes. Seasonal abundance and biomass dynamics of the BMC showed an increasing trend until October (T. Apelopia) and November (c. occidentalis), when massive emergence occurred, just before the lake dried out. Contrary to most temporal waters, Tecuitlapa Sur did not show taxonomic or trophic succession. C. occidentalis, a transient detritivore, dominated over T. Apelopia, a resident predator during the wet period.

Introduction Alkaline-saline or soda lakes, are inhabited by specialized biota that can often tolerate extremes of high temperature, high salinity, high pH and low dissolved oxygen concentration. These lakes may also desiccate during the dry months and rehydrate during the wet season, requiring adaptations for recolonization. Temporary waters, in their cycle of wetting and drying, represent excellent examples of the 'extreme' in water regimes for most aquatic biota. The high concentrations and substantial fluctuations of salinity during the wetting-drying cycle in saline, temporary waters pose an even greater challenge for aquatic biota (Boulton & Brock, 1999). If the temporary saline waters are soda-type, high pH may cause further environmental constraints. It has been suggested that habitat duration may mediate a shift in the relative importance of abiotic

and biotic processes in determining the distribution and abundance of species in temporary ponds (Wiggins et aI., 1980; Wilbur, 1987). In naturally stressful habitats, physical stress and the adaptation of organisms to the environment would be expected to exert a dominant influence over community composition. As the time period between disturbances increases, biotic interactions among the species would gain importance (Wilbur, 1987). Temporary or intermittent aquatic systems are ideally suited for examining patterns of community succession, especially benthic macroinvertebrates (Moorhead et aI., 1998). Schneider & Frost (1996) have found that abiotic factors are the most important determinants of community structure in temporary ponds, but biotic interactions (i.e. competition and predation) become increasingly important as the period of inundation increases.

292 Studies of the temporal dynamics of taxa or trophic groups for temporary, saline, soda lakes are few. We present the results of a study of benthic macroinvertebrate community (BMC) development in Tecuitlapa Sur, a temporary astatic, saline, soda lake in central Mexico. It displays an extreme and stressful environment in salinity (high and fluctuating), pH (high and stable) and a temporary water-regime. Our objective was to determine the structure of the benthic macroinvertebrates community (i.e. species composition, richness, abundance and biomass), as well as its changes along an annual cycle. Since the aim of this study was exploratory, it is descriptive in nature so it cannot explicitly test any hypothesis.

Study site Tecuitlapa Sur is one of three water-bodies in the Pleistocene crater of an extinct strato-volcano in the Oriental basin (19° 08'-19° 30' N, 97° 20'-97° 51' W; 2300 m.a.s.\.; 4982 km 2) (Fig. I). This endorheic basin is located in the extreme south-eastern portion of the Mexican Plateau at the conjunction of the States of Puebla, Tlaxcala and Veracruz. The area is characterized by a subhumid, temperate climate: mean annual temperature and precipitation are 13°C and 706 mm, respectively. The warm rainy season extends from May to November, and the cold-dry season from December to April (Garda, 1988). Lake Tecuitlapa Sur is shallow (maximum depth ~ 0.5 m), and small (100 x 20 m when fully inundated). Its waters are saline, soda type (Na2C03), alkaline, and deep sepia or dark brown in color ('true color', it remained after 0.22 (tm filtration) due to the presence of dissolved organic matter and/or colloidal silica (Alcocer & Hammer, 1998). In addition to color, large quantities of suspended sediments keep water turbid. Aquatic macrophytes are absent, and cyanobacterial blooms, common in the other two neighboring water bodies (Spiruiina in Tecuitlapa Norte and Microcystis in Tecuitlapa crater lake) (Fig. I), are lacking. The sediment consists of silty sand with a high organic matter content (8%). There are no fish or other large aquatic vertebrate predators. Tecuitlapa Sur resembles (Authors' personal observations) in size, depth, color and probably water chemistry the pond close to the northern shore of Big Soda Lake, Nevada (not to be confused with Small Soda Lake, few meters away from Big Soda Lake).

Methods A monthly sampling program was carried out from December 1993 to December 1994. Midday measurements of temperature, pH, conductivity (K2S), dissolved oxygen and redox potential (referenced to H electrode) were measured with a previously calibrated Hydrolab DS3/SVR3. We used conductivity as a proxy for salinity. Dissolved oxygen percent saturation values were temperature and salinity corrected. On each sampling date, faunal samples were obtained from 20 (6 cm in diameter) sediment cores from the top 15 cm of the sediment. The cores were randomly located within the lake area. The samples were sieved in the laboratory through 0.25 mm mesh screen. The fauna was sorted, preserved in 70% ethanol, identified, quantified, weighed wet and transformed to mg C m- 2 (10% of wet-weight, Margalef, 1983). Multivariate analysis (Principal Components Analysis - PCA) was used to examine the temporal variation of the environmental factors in Tecuitlapa Sur. Transformed environmental (log n+ 1, except pH) data were used in the statistical analysis.

Results and discussion Environmental patterns Lake Tecuitlapa Sur holds water (wet period) during the warm-rainy season, from June to November and remains dry from December to May, in the cold-dry season. This pattern is comparable to other tropical temporary pools containing water immediately after the monsoons. However, the exact length of the aquatic phase varies according to both geographic location and local hydrological conditions (Williams, 1987). Rain and groundwater are the main water sources to Tecuitlapa Sur (Alcocer & Hammer, 1998). Tecuitlapa Sur experienced two phases: dilution and concentration (Fig. 2). The dilution phase occurred from June to August and was characterised by a trend of decreasing conductivity (13.6-10.5 mS cm- 1) and nearly constant temperature (24-25 0C). The first runoff from the catchment containing dissolved salts caused an initial peak in conductivity, as in Boulton & Brock (1999), and Williams (1987). The pH remained almost constant (9.8-10), while the redox potential values increased (430-457 mV). The concentration phase took place from September to November and was characterized by a trend

293

"

.........

106'

91'

~-

29'

Gulf of

Ocean

puebl~ 15'

99"00'

97"00'

... Tecuitlapa Norte :::::::::::::::: _................. .................. ................. ................ . . .....

·fl .....

..

0~m~. . .1~0~om!!!!~2~00m

Tecuitlapa Sur

Figure 1. Geographic location of Tecuitlapa Sur lake, Mexico.

Table 1. Principal Components Analysis of environmental characteristics of Tecuitlapa Sur (K 25 = conductivity standarized at 25°C, D.O. =dissolved oxygen, PC =principal component) Parameter

Min-Max

PC I (62.8%)

PC 2 (20.2%)

(Factor loading) (Factor loading) Temperature (0C) 24.1-27

-0.737

-0.562

9.8-10

0.470

0.671

K25 (mScm- l )

10.5-62.6

-0.976

0.076

D.O. (mgL -I)

2.4-3.6

0.894

-0.291

pH

D.O. (% Sat.)

5\.7-71.7

0.561

-0.592

Eh(mV)

355-457

0.967

-0.065

of increasing conductivity (11.4-62.6 mS cm -I) and temperature (25.3-27 0c). While pH (~ 9.8) remained constant, redox potential (448-355 mY) decreased. Dissolved oxygen percent saturation showed a bimodal pattern during the inundation period (Fig. 2) with two maxima, the first one occurring in July and the second in September (68.9 and 71.7%, respectively) separated by a minimum in August (51.7%). The PCA showed that salinity (PC 1) and pH (PC 2) were the most important parameters in explaining environmental variance in Tecuitlapa Sur (Table 1).

294

Biotic structure "Filling"

"Drying"

"Filled" Temperature (OC)

-

~I : :::::= JUN

JUL

SEP

AUG

OCT

NOV

D.O.I%S.,)

80 70

,.

60 50

,,-, .... _"

~-"'-

........

.....

40 30 20 10 0 JUN

JUL

SEP

AUG

OCT

NOV

Conductivity emS/em) 70



60 50

,

40

I

30

,•

-- -. -- -- - - -- ....."••

20 10

o+---~----~----P----.----.---~ JUL SEP OCT NOV JUN AUG pH

11 10.5 10 9.5

- -- - - - - - - - JUN

JUL

AUG

SEP

OCT

NOV

- -- _.- -_ .......... .. .. EhlmV)

500 400 300 200 100 0

JUN

JUL

AUG

SEP

OCT

NOV

Figure 2. Changes in the physical and chemical variables over time during the phases of dilution (filling) and concentration (drying) in Tecuitlapa Sur.

Two dipteran species (Insecta, Diptera) composed the simple BMC in Tecuitlapa Sur. In order of dominance, the species were Culicoides occidentalis sonorensis J0rgensen (Ceratopogonidae), and Tanypus Apelopia sp. Meigen (Chironomidae). C. occidentalis is a transient detritivore species, while T. Apelopia is a predator ocurrying most of the time. Such a low species richness (1-2 species) has been previously reported in saline lakes around the world (Timms, 1983; Hammer et aI., 1990; Dejoux, 1993). In other saline lakes worldwide that have a similar salinity range as Tecuitlapa Sur, the species richness is higher than 2 but still low (Timms, 1982, 1983; Colburn, 1988; Herbst, 1988; Williams & Kokkinn, 1988; Dejoux, 1993; Alcocer et al., 1997, 1999). Low species richness is associated but not restricted to saline lakes that lack aquatic macrophytes (e.g. Tecuitlapa Sur in Mexico, this study; Little Manitou and Aroma Lakes in Canada, Hammer et aI., 1990; Werowrap, Missen and Gnarlinegurk Lakes in Australia, Timms, 1983). Duration of inundation and the pattern of loss of the water, whether predictable or unpredictable, are two of the most important factors that influence the invertebrate assemblages in temporary aquatic habitats (Wiggins et aI., 1980; Williams, 1987, 1996; Boulton & Brock, 1999). In comparison to other temporary waters, Tecuitlapa Sur could be considered of intermediate-duration (water present six months), and predictable in both its time of onset and duration (personal observations for a lO-year period before this study took place). Both species inhabiting Tecuitlapa Sur belong to families (Ceratopogonidae and Chironomidae) well represented through a gradient of habitat permanence from ephemeral to permanent lentic habitats (Williams, 1996). The abundance and biomass of C. occidentalis (99 and 95% of the total, respectively) overshadowed that of the T. Apelopia (1 and 5% ofthe total, respectively). Mean annual benthic macroinvertebrate density and biomass (mean ± sd) were 1147865 ± 2 765 660 indo m- 2 and 1290 ± 2960 mg C m- 2 (n =120), respectively. Higher total densities (6 793 095 ± 339 654 indo m- 2, n =20) and biomasses (7 319.29 ± 365.97 mg C m- 2 , 20) were detected in November, while the lowest were observed in June (1706 ± 205 indo m- 2 and 5.8 ± 0.7 mg C m- 2 , n =20) (Fig. 3). Mean annual density of C. occidentalis in Tecuitlapa Sur (1141082 ± 2765879 indo m- 2, n = 120)

295 m-2_ _ _ _ _ _ _ _ _ _ _ _--, 10000000 Ind. _ ~

1000000 100000 10000 1000 100

-I

••

_-I

t"l-

f

,"=c:-=-..,.=-=-==-=~-,----___--=,-------'

JUN JUL AUG SEP OCT NOV

C. occidentaJis

JUN JUL AUG SEP OCT NOV

T. Apelopia

10000 mg C m-2

1000

t

100 10

It JUN JUL AUG SEP OCTNOV

C. occidentalis



JUN JUL AUG SEP OCT NOV

T. Apelopia

Figure 3. Abundance (ind. m- 2) (top) and biomass (mg C m- 2) (bottom) of Culicaides occidentalis and Tanypus Apelapia in Tecuitlapa Sur. (Mean ± I sd).

is notably higher (10-800 times) than the other values reported for other saline water bodies (Hammer et aI., 1990; Hayford et aI., 1995; Alcocer et aI., 1999). Mean annual density of T. Apelopia in Tecuitlapa Sur (6782 ± 8310 indo m- 2, n = 120) is similar to values recorded in other saline lakes worldwide (see Stahl, 1986). Seasonal biotic changes The seasonal abundance changes of both species are similar and show a one-month emergence temporal lag between them. Both species showed an increasing trend of abundance and biomass along the time (Fig. 3). C. occidentalis reached its maximum density values (6786857 ± 475080 indo m- 2 , n = 20) and biomass (7305 ± 876.6 mg C m- 2, n = 20) in November, right before the lake dried out completely. T. Apelopia displayed its maximum density values (22222 ± 2667 indo m- 2, n = 20) and biomass (381.55 ± 45.79 mg C m- 2, n =20) in October, and decreased in November (6238 ± 312 indo m- 2, and 14.29 ± 11.43 mg C m- 2, n =20).

Many previous studies have failed to find a consistent relationship between abiotic factors and species richness, diversity, or abundance of particular taxa (Williams, 1996). In spite of this, three main groups of constraints, physical, chemical and biological factors, seem to strongly influence aquatic biota in some temporary waters (Williams, 1987). Biotic factors, such as life-history characteristics, competition and predation, control the invertebrate assemblages in other temporary aquatic habitats (Moorhead et aI., 1998). We did not find a statistically significant (p > 0.05) correlation between abiotic factors and abundance or biomass of the BMC of Tecuitlapa Sur. Our data suggest that changes in abundance of the BMC (Fig. 3) follow a similar trend to the one observed with the environmental factors (e.g. temperature, salinity) (Fig. 2), although some environmental fluctuations are rather small (e.g. pH = 9.75-10.04, Eh = 355-457 mY). Changes in abundance of the benthic macroinvertebrate fauna were related to key periods in the reproductive cycles of the species in Tecuitlapa Sur. The massive emergence period of C. occidentalis was indicated by the presence of multiple pupae in November, 'clouds' of adults flying at the lake shore, and numerous exuvia found on the dry bed of the lake occurred in December. At the time when C. occidentalis emerged (September-November), an increase in water temperature from 25.3 °C to 27°C and a decrease in the water level from 0.5 m to 0.05 m, took place. Linley et aI. (1970), Linley & Adams (1972) and Alcocer et aI. (1999) observed similar events in C. jurens, C. melleus and C. occidentalis, respectively. However, since there is no definitive evidence to confirm that temperature or water level is the main causal factor, the massive emergence of C. occidentalis could be related to the seasonal emergence pattern or simple part of the annual development cycle. The constant high temperature (above 23°C) probably favors continuous emergence of C. occidentalis, which could explain why C. occidentalis adults were seen flying near Tecuitlapa Sur throughout the entire sampling period. The large number of pupae found in October explained the large decrease in numbers of T. Apelopia at the end of the drying period of the lake. The development of both families (Ceratopogonidae and Chironomidae) displaying rapid growth and a short lifespan is strongly linked to temperature (Williams, 1996).

296 Seasonal succession

Fauna of temporary ponds generally display characteristic taxonomic and trophic successional patterns (Bayly & Williams, 1973). Taxa lacking active dispersal mechanisms increase densities early in the wet season. Taxa with dispersal mechanisms (flight), especially insects, tend to reach higher numbers later in the season. In contrast, trophic structure shows a progressive shift from numerical dominance of filterfeeders and detritivores to predators (Moorhead et aI., 1998). Schneidder & Frost (1996) suggest a model in the community development of temporary ponds in which the first phase is a rapid increase in species richness as a result of colonization. A second phase includes a period of constant species richness but continuous turnover of taxa, probably as a result of increasing competition and predation. High levels of predation usually dominate the final phase. None of these successional patterns - taxonomic and trophic - were observed in Tecuitlapa Sur and may be attributed to the extreme environmental constraints encountered. The two species inhabiting Tecuitlapa Sur have active dispersal mechanisms, and colonized the lake as soon as it held water and remained there until it dried. Species richness remained constant and there was no turnover of taxa. Moreover, at the later stages, abundance of predators was diminished. Recruitment

To colonize Tecuitlapa Sur successfully, any species must be able to tolerate high and fluctuating salinity and high pH (saline, alkaline, soda water species) as well as cope with a temporary water regime. Biological interactions may act to further restrict species colonization in Tecuitlapa Sur. There are two additional potential colonization loci within the same crater that Tecuitlapa Sur is located: Tecuitlapa crater-lake and Tecuitlapa Norte (Fig. 1). The former is an alkaline (pH = 9.8 ± 0.1) and eutrophic (chlorophyll a = 200 mg m- 3 ) freshwater (K25 = 1.65 ± 0.05 mS cm- I ) crater-lake, located less than 100 m from Tecuitlapa Sur. Tecuitlapa Norte is an alkaline (pH =10.4-11.6) and hypertrophic (chlorophyll a = 1500 mg m- 3), mesosaline (K25 = 21.2-55.7 mS cm-I) lake, located about 200 m from Tecuitlapa Sur. All three lakes have a similar alkaline carbonate (soda) ionic composition (Vilaclara et al., 1993; Alcocer et aI., 1999).

Twenty-nine species of benthic macroinvertebrates (Ephemeroptera, Odonata, Hemiptera, Trichoptera, Coleoptera, Chironomidae, Oligochaeta, Hirudinea, Gastropoda) have been recorded in Tecuitlapa craterlake (Alcocer, 1995) and none of these occur in Tecuitlapa Sur. Thus, there must be limitations impeding the colonization of species from Tecuitlapa crater-lake into Tecuitlapa Sur. We consider that salinity is of primary importance in this respect since Tecuitlapa crater-lake has a similar alkaline carbonate composition to Tecuitlapa Sur but is freshwater and not saline. The dominant species of Tecuitlapa Sur (c. occidentalis), characteristic of saline, alkaline carbonate waters, is not found in Tecuitlapa crater-lake, most likely due to competition and/or predation. Saline, alkaline lakes, in which predators and competitors are lacking, or greatly diminished due to salinity stress offer an excellent habitat for ceratopogonids. The flourishing population of C. occidentalis in Tecuitlapa Norte (Alcocer et aI., 1999) is consistent with the reduced role of biotic interactions at elevated salinity. Two dominant species characterize Tecuitlapa Norte: C. occidentalis and Ephydra hians Say (Alcocer et aI., 1999). Why is E. hians not present in Tecuitlapa Sur if saline alkaline lakes offer an excellent habitat for ephydrids to live in and multiply in large numbers? Adult ephydrids have been seen flying near Tecuitlapa Sur throughout the wet period. T. Apelopia inhabits the soft bottom of Tecuitlapa Sur and could be feeding on the eggs and early larval stages of E. hians, thus preventing its development and successful colonization of Tecuitlapa Sur. McCafferty (1981) has suggested that ceratopogonids feed on aquatic insect (e.g. alkali fly) eggs. While capable of affecting taxon abundance, predation and competition seem not to determine the distribution of taxa (species composition) (Schneider & Frost, 1996). If habitat duration is the major factor controlling the community structure, communities should be increasingly structured by biotic interactions as the period between disturbances (dry periods) increases. Duration acts by mediating the relative importance of life histories and biotic interactions, particularly predation, in determining the distribution and abundance of taxa in temporary ponds (Williams, 1996). One interesting question is why T. Apelopia does not control the popUlation of C. occidentalis which is dominant in Tecuitlapa Sur? We consider they occupy different spatial positions in this rather shallow water column. C. occidentalis lives near or at the

297 water surface; especially pupae that maintain contact with air at the water surface by means of thoracic respiratory horns. T. Apelopia is a bottom dweller (bottom browser). In addition, ceratopogonids are highly mobile swimmers and may elude T. Apelopia. Summarizing, salinity and pH were the most important parameters explaining environmental variance of Tecuitlapa Sur. Two species of BMC occurred in Tecuitlapa Sur: C. occidentalis and T. Apelopia. Both species coexisted throughout the year. C. occidentalis was most important both numerically and in biomass (2:95%). While mean annual density of C. occidentalis was notably higher than the values reported for other saline water bodies, T. Apelopia density was in the same range of other studies. Seasonal dynamics of the BMC showed an increasing trend until October (T. Apelopia) and November (c. occidentalis), when massive emergence occurred, just before the lake dried out completely. Tecuitlapa Sur did not show taxonomic or trophic succession. C. occidentalis, a transient detritivore species, dominated over T. Ape/opia, a resident predator during the wet period.

Acknowledgements We gratefully acknowledge the help of M. R. Sanchez, M. M. Chavez, L. Peralta and M. 1. Montoya in the field and the laboratory work. Financial support was partially given by CONACYT project T-25430, and DGAPA project IN204597. The authors thank Dr D. W. Webb and Dr A. Borkent (Ceratopogonidae) and Dr A. Contreras (Chironomidae) for taxonomic identification.

References Alcocer, J., 1995. Amilisis hoHstico de la comunidad de macroinvertebrados bent6nicos litorales de seis lagos-crater con un gradiente de salinidad. Ph.D. Thesis. Universidad Nacional Aut6noma de Mexico. Facultad de Ciencias. Mexico: 106 pp. Alcocer, J. & U. T. Hammer, 1998. Saline lake ecosystems of Mexico. Aquat. Ecosystems Health and Management I: 291-315. Alcocer. J., A. Lugo, E. Escobar & M. Sanchez, 1997. The macrobenthic fauna of a former perennial and now episodically filled Mexican saline lake. Int. J. Salt Lake Res. 5: 1-14. Alcocer, J., E. Escobar, A. Lugo & L. A. Oseguera, 1999. Benthos of a perennially-astatic, saline, soda lake in Mexico. Int. J. Salt Lake Res. 8: 113-126. Bayly, 1. A. E. & W. D. Williams, 1973. Inland Waters and their Ecology. Longman, Melbourne: 314 pp.

Boulton, A. J. & M. A. Brock, 1999. Australian Freshwater Ecology. Processes and Management. Cooperative Research Center for Freshwater Ecology, Australia: 300 pp. Colburn, E. A., 1988. Factors influencing species diversity in saline waters of Death Valley, U.S.A. Hydrobiologia 158: 215-226. Dejoux, C., 1993. Benthic macroinvertebrates of some saline lakes of the Sud Lipez region, Bolivia. Hydrobiologia 267: 257-267. Garda. E., 1988. Modificaciones al Sistema de Clasificaci6n CIimatica de Koppen. E. Garcia, Mexico: 217 pp. Hammer, U. T.. J. S. Sheard & J. Kranabetter, 1990. Distribution and abundance oflittoral benthic fauna in Canadian prairie saline lakes. Hydrobiologia 197: 173-192. Hayford, B. L., J. E. Sublette & S. 1. Hermann, 1995. Distribution of chironomids (Diptera: Chironomidae) and ceratopogonids (Diptera: Ceratopogonidae) along a Colorado thermal spring etfluent. J. Kansas Entomo!' Soc. 68: 77-92. Herbst, D. B., 1988. Comparative population ecology of Ephydra hians Say (Diptera: Ephydridae) at Mono Lake (California) and Abert Lake (Oregon). Hydrobiologia 158: 145-166. Linley, J. R. & G. M. Adams, 1972. Ecology and behavior of immature Culicoides melleus (Coq.) (Dipt., Ceratopogonidae). Bull. Entomo!. Res. 62: 113-127. Linley. J. R., F. D. S. Evans & H. T. Evans, 1970. Seasonal emergence of Culicoides furens (Diptera: Ceratopogonidae) at Vero Beach, Florida. Ann. Entomo!. Soc. Am. 5: 1332-1339. Margalef, R., 1983. Limnologfa. Omega, Barcelona: 1010 pp. McCafferty, W. P., 1981. Aquatic Entomology. Science Books International, Boston: 448 pp. Moorhead, D. 1.. D. 1. Hall & M. R. Willig, 1998. Succession of macroinvertebrates in playas ofthe Southern High Plains, U.S.A. 1. n. am. Benthol. Soc. 17: 430-442. Schneider, D. W. & T. M. Frost, 1996. Habitat duration and community structure in temporary ponds. 1. n. am. Benthol. Soc. 15: 64-86. Stahl, 1. B., 1986. A six-year study of abundance and voltinism of Chironomidae (Diptera) in an Illinois cooling reservoir. Hydrobiologia 134: 67-79. Timms, B. v., 1982. A study of the benthic communities of twenty lakes in South Island. New Zealand. Freshwat. BioI. 12: 123138. Timms, B. v., 1983. A study of benthic communities in some shallow saline lakes of western Victoria, Australia. Hydrobiologia 105: 165-177. Vilaclara, G., M. Chavez, A. Lugo, H. Gonzalez & M. Gaytan, 1993. Comparative description of crater-lakes basic chemistry in Puebla state, Mexico. Verh. int. Ver. Limno!. 25: 435-440. Wiggins, G. B., R. 1. Mackay & I. M. Smith, 1980. Evolutionary and ecological strategies of animals in annual temporary pools. Arch. Hydrobio!., Supple. 58: 97-206. Wilbur, H. M., 1987. Regulation of structure in complex systems: experimental temporary pond communities. Ecology 68: 14371452. Williams, D. D., 1987. The Ecology of Temporary Waters. Croom Helm, London: 205 pp. Williams, D. D., 1996. Environmental constraints in temporary fresh waters and their consequences for the insect fauna. J. n. am. Benthol. Soc. 15: 634-650. Williams, W. D. & M. 1. Kokkinn, 1988. The biogeographical at1inities of the fauna in episodically filled salt lakes: a study of Lake Eyre South, Australia. Hydrobiologia 158: 227-236.

Hydrobiologia 466: 299-306,2001. J.M. Me/ack, R. Jellison & D.E. Herbst (eds), Saline Lakes. © 2001 Kluwer Academic Publishers.

299

Phytoplankton dynamics in a deep, tropical, hyposaline lake Ma. Guadalupe Oliva), Alfonso Lugo2, Javier Alcocer2 , Laura Peralta2 & Ma. del Rosario Sanchez2 1Botany Laboratory, UMF, FES Iztacala, UNAM, Av. de los Barrios sin, Los Reyes [ztacala, Tlalnepantla 54090, Edo. de Mexico, Mexico 2Limnology Laboratory, Environmental Conservation & Improvement Project, UIICSE, FES Iztacala, UNAM

Key words: crater lake, monomictic lake, Nodularia spumigena, sodium chloride lake, Central Mexico

Abstract The annual variation of the phytoplankton assemblage of deep (64.6 m), hyposaline (~8.5 g 1-1) Lake Alchichica, central Mexico (19 0 N, 97 0 W), was analyzed in relation to thermal regime, and nutrients concentrations. Lake Alchichica is warm monomictic with a 3-month circulation period during the dry, cold season. During the stratified period in the warm, wet season, the hypolimnion became anoxic. N-NH3 ranged between non detectable (n.d.) and 0.98 mg 1-1, N-N02 between n.d. and 0.007 mg 1-1, N-N03 from 0.1 to 1.0 mg I-I and P-P04 from n.d. to 0.54 mg I-I. Highest nutrient concentrations were found in the circulation period. Chlorophyll a varied from < I to 19.8 J.Lg 1-1 but most values were I % PAR) usually comprised the top 15-20 m. Nineteen algae species were identified, most of them are typical inhabitants of salt lakes. Diatoms showed the highest species number (l0) but the small chlorophyte Monoraphidium minutum, the single-cell cyanobacteria, Synechocystis aquatilis, and the colonial chlorophyte, Oocystis parva, were the numerical dominant species over the annual cycle. Chlorophytes, small cyanobacteria and diatoms dominated in the circulation period producing a bloom comparable to the spring bloom in temperate lakes. At the end of the circulation and at the beginning of stratification periods, the presence of a bloom of the nitrogen-fixing cyanobacteria, N. spumigena, indicated nitrogen-deficit conditions. The well-stratified season was characterized by low epilimnetic nutrients levels and the dominance of small single-cell cyanobacteria and colonial chlorophytes. Phytoplankton dynamics in tropical Lake Alchichica is similar to the pattern observed in some deep, hyposaline, North American temperate lakes.

Introduction Studies of phytoplankton dynamics are accumulating from an increasing, although still limited, number of tropical sites located in Africa, Asia, Australia and Central and South America (Talling & Lemoalle, 1998). Using mainly data from Africa, Melack (1979) demonstrated the lack of latitudinal trend in temporal fluctuation within the tropics and identified several patterns of seasonal and interannual variation. In most tropical lakes, pronounced seasonal fluctuations usually correspond to variations in rainfall, runoff or vertical mixing within the lake (Melack, 1996). Studies of phytoplankton dynamics from deep, warm monomictic tropical lakes are scant (e.g. Tailing, 1966; Lewis, 1978, 1986; Hecky & Kling,

1981; Bootsma, 1993; Kebele & Belay, 1994; Matsumura-Tundisi & Tundisi, 1995). Deep tropical lakes show one or two peaks of algal abundance that seem to be primarily determined by changes of hydrographic structure related to the radiation, temperature and wind (Payne, 1986; TaIling & Lemoalle, 1998). Saline lakes (S > 3 g 1- 1) occur in many parts of the tropics but most of them are shallow, fluctuate in depth and salinity and vary widely in chemical composition (Melack, 1996). Deep (lm> 15m) tropical saline lakes are much less common, and the phytoplankton succession of these lakes have been investigated in few, most of them located in Africa (Melack, 1981). In the temperate region of North America, there are several deep saline lakes (e.g. Lake Mono and

300 Big Soda Lake, California; Lakes Pyramid and Walker, Nevada). Pyramid and Walker lakes are warmmonomictic, hypos aline (3-20 g 1-1), terminal lakes located in arid or semi-arid regions (Galat et aI., 1981; Cooper & Koch, 1984). Dominant ions are Na+ and Cl- but high concentrations of HC03 - and C03-2 also are present. A characteristic trait of phytoplankton in both lakes is the presence of blooms of the nitrogenfixing cyanobacteria, Nodularia spumigena, during the stratified period, indicating N-deficient conditions (Galatetal.,1981). Lake Alchichica is a Mexican tropical lake (19° N, 97° W) sharing several characteristics with Pyramid and Walker lakes. All are deep, warm-monomictic, hyposaline lakes. Water chemistry is comparable. Furthermore, the most striking resemblance is the presence of N. spumigena blooms. The aim of the present study is to provide the first report of phytoplankton dynamics of a deep warm-monomictic, North American tropical lake. The objectives were to determinee the phytoplankton taxonomic composition, to delineate the general successional pattern of the phytoplankton and to elucidate the factors governing phytoplankton fluctuations. Our hypothesis is that, in spite of the latitudinal differences, the similar stratification pattern, salinity concentration and ionic composition, results in the phytoplankton dynamics of Lake Alchichica resembling those of Pyramid and Walker Lakes.

Description of site Six maars lakes occur in the endorheic Basin of Oriental (4982 km2, mean altitude 2300 m) in the central portion of Mexico (Alcocer et al., 1998). Lake Alchichica (19° 24' 22" Nand 97° 23' 52" W, 2345 masl), the largest and deepest of the six lakes, shows a circular form (diameter 1733 m). The lake is filled by ground water mainly and secondarily by rainwater. Lake Alchichica is 64.6 deep (mean depth= 38.55 m), with a surface area of 1.81 km2, a volume of 69.92 x 106 m3 and a shoreline lenght of 5.06 km (Arredondo et aI., 1983). The climate is dry temperate, with a mean annual temperature of 12.9 °C and the mean annual precipitation less than 400 mm (Garda, 1988). Evaporation is high (Arredondo et aI., 1984). Rain occurs during the summer, along with the highest temperatures (14.5-15.4 °C mean monthly temperature). In winter, precipitation is rather low and temperature is the coldest (9.2-13 0c) (Garda, 1988). Lake Alchichica is hyposaline (~8.5 g I-I) and basic (pH

around 9); sodium and chloride are the dominant ions but bicarbonates and carbonates are also important with a total alkalinity of 37 meq I-I (Vilaclara et ai., 1993).

Materials and methods Samples were colected monthly from January to December, 1998, with the exception of May when two field trips were done. Due to logistic problems, it was not possible to sample in September. Five samples were obtained with a Niskin (61) water sampler at 2,5, 10,20 and 50 m. Subsamples (500 ml) for phytoplankton counts were preserved with acid Lugol's solution (1 % final concentration). Phytoplankton counts were made using 1O-100-ml settling chambers with an Invertoscope D of Carl Zeiss by the Utermohl (1958) method. At the same depths, another set of 5 I samples was obtained for chlorophyll a and nutrients analysis. These samples were mainteined cold (4°C) and in darkness until their analysis. In the laboratory, 3 I of each sample were filtered through a 0.45 f.,Lm poresize Millipore membrane filter. After 24 h cold (4°C) 100% methanol extraction (Marker et aI., 1980), absorbance was measured in a Hewlett-Packard 8450A UVNIS spectrophotometer. Chlorophyll a concentration was calculated using the equation proposed by Talling & Driver (in Vollenweider, 1969) without phaeophytins correction. A portion of the filtered samples were used to measure phosphorous as orthophosphate (P04-P, ascobic acid method, detection limit 0.01 mg I-I.) and nitrogen as nitrate (N03N, cadmiun reduction method, detection limit 0.1 mg I-I), nitrite (N02-N, diazotization method, detection limit 0.001 mg 1-1) and ammonia (NH3-N, Nessler method modified considering salinity, detection limit 0.01 mg 1-1) using a HACH DRELl2000 laboratory. Vertical profiles of temperature, dissolved oxygen, specific conductivity standarized to 25°C (K2S) and pH were measured at each sample date employing a Hydrolab Datasonde 3/Surveyor 3 multiparameter water quality logging system (Hydrolab Co., Tx, U.S.A.).The photic zone (> 1% of surface incident radiation) was measured using an scalar PAR irradiance collector mounted on a PFN-300 (Biospherical Instruments, Ca., U.S.A.) vertical profiler. The zooplankton information included along the results and discussion of this paper, was taken from Lugo et al. (1999).

301 Table 1. Nutrients concentrations in Lake Alchichica. 1998. n =60

Maximum Mimimum Mean Stand. dev.

P-P04 mgl- I

N-NH3 mgl- I

N-N02 mgl- I

N-N03 mgl- 1

DIN mgl- 1

0.56 nd 0.04 0.08

0.98 nd 0.16 0.19

0.007 nd 0.002 0.001

1.0 0.1 0.2 0.1

1.85 0.12 0.38 0.27

DIN = Dissolved inorganic nitrogen nd = non detectable

?rJ(O~~(J

il ~Airf~~3t :.

~I .. , , MONTHS

Jr

Figure 2. Depth-time diagram of dissolved oxygen (mg I-I) concentrations. 1998.

Results Environmental conditions Data obtained in 1998 confirmed the alkaline and hyposaline conditions of Lake Alchichica water. pH values ranged from 8.8 to 10.0. Highest values were recorded in May associated with the highest densities of N. spumigena. Lowest values were measured from August to October, when lake was well-stratified. Specific conductivity (K2S) varied between 13.27 and 14.14 mS cm- I . Temperature profiles in Lake Alchichica during 1998 confirmed the lake is warm monomictic. Mixing occurred from January to March, in the dry and cold season when the water was about 15 cC. Starting in April, the surface layer warmed. From June to October, the lake was well-stratified (Fig. 1). In November and December, the surface layer cooled (Fig. 1). During the circulation period, dissolved oxygen (DO) was high (6-8.1 mg I-I) throughout the water column (Fig. 2). Gradually, anoxic conditions appeared near the bottom in April and increased to

J(~~~_ (. 20 phosphorous is the limiting nutrient and when DIN/SRP < 10 nitrogen is the limiting factor (Danielidis et aI., 1996). DIN/SRP varied in Lake Alchichica. P limitation was found in January while along February and March it happened to be N. In April and May,

302

Phytoplankton composition and abundance

Figure 4. Depth-time diagram of dissolved inorganic nitrogen (DIN mg 1-1) concentrations. 1998.

N. spumigena appeared, DIN concentrations increased and P became the limiting factor. During stratification DIN levels decreased but P continued as the limiting nutrient.

PAR The euphotic zone ranged between 15 and 35 m. Lower values (15-21 m) were measured in the circulation period. During stratification the euphotic zone deepened (28-35 m).

Chlorophyll a Chlorophyll a concentrations were generally low (mean column value 3.5%, while that for saline lakes sensu lato 2:0.30%1 or in excess of 3 gil (Willians, 1996). Saline lakes can contain important raw materials for industry, agriculture, and medicine, e.g. halite, mirabilite, lithium, magnesium, boron, gypsum calcium chloride, tungsten, cesium, rubidium, strontium, hydromagnesite and zeolite. Considerable amounts of biological resources, such as halophilic algae, Artemia, Spirulina, of economic and scientific value, occur in saline lakes. Moreover, saline lakes are important for tourism. The heat-storing features of lake brine solar evaporation ponds have also been used in electricity. Finally, saline lakes are sensitive indicators of the past and important for reconstructing paleoclimatic, paleoenvironmental and tectonic events. 1 Saline lakes sensu stricto refer to those lakes which use the limit of saline lake salinity used commonly in geological communities and this limit of salinity is higher than the average salinity of the world ocean water; saline lakes sensu lato refer to those lakes in which some organic groups show significant change at the limit of saline lake salinity used commonly in biological communities.

Trends of scientific and technological development If we consider that research on saline lakes star-

ted from an analysis of the brines of the Karabugaz Lagoon in the mid-nineteenth century, saline lake research has experienced three stages. The first stage lasted from the middle of the nineteenth to the beginning of the twentieth century (the Van't-Hoff school and N.S. Kumakov school). Research during this period was dominated by physical-chemical analysis. The second stage, from the early-twentieth century to the 1960s, can be referred to as the stage of traditional geological, biological and chemical disciplinary research. Since the 1970s, research on saline lakes has developed into a stage with multidisciplinary research (Matter and Tucker, 1978; Kushner, 1978; Eugster and Hardie, 1978; Smith, J 979; Brock, 1979; Dietor, 1979; Gwynn, 1980; Nissenbaum, 1980; Borowitzka et aI., 1981; Hammer, 1981; Williams, 1981; Yuan, 1982; Javor, 1989; Hurlbert, 1993; Zheng, 1995, 1996; Oren, 1999). Development and expansion of the scope of research on saline lakes Saline lakes are valuable natural resources. With continued reconnaissance, exploration and development of their resources, research on them has deepened and its scope expanded, with progress in experimental studies of their values and uses. The concept of sa-

340 line lakes as mainly solid mineral resources, has been outdated. Large amounts of important raw materials for the chemical industry, agriculture, metallurgical industry and medicine are now obtained from saline lakes. In some brines, biological resources are present in large amounts. In addition, the heat-storing features of saline lake brines have been used in 'solar energy salt pond' electricity and playas have been used to build highways, railways and even airfields.

Value and use of saline lake brines Mineral resources of saline lakes and their applications are shown in Table 1. In saline lakes, there are usually complex ores with solid and liquid states. Significant advances in the investigation, valuation and uses have been made with considerable economic benefits. For example, in Great Salt Lake, long-term fundamental and applied research has focused on mineral and biological resources, hydrochemistry, and engineering and includes meteorological and hydrological observations over 140 years (Gwynn, 1980). The annual gross value of the potash, magnesium, sodium sulfate and halite reaches about one billion dollars. Another example is the Atacama Salt Lake in Chile, where the Cyprus-Foot Company and SQM Company invested 60 million dollars for feasibility studies of development of lithium production and then established a lithium plant that is now the largest in size and produces lowest cost lithium. Simultaneous production of lithium salt, potash salt, borate and sodium sulfate occur with an annual gross product of more than one billion dollars. Since 1997, SQM has greatly reduced the cost of Li2C03 and eliminated many hard rock lithium ventures (McCracken, 1998). Potassium-containing saline lakes as an important source of the world's potash Since the 1970s, Quaternary potassium-bearing saline lakes have become an important source of the world's potash resources. In some countries (e.g. Jordan, Israel and China), potassium-bearing saline lakes have become the main source of potash. As other saline minerals can also be extracted as by-products and the potash resources are located at shallow depths and are easy to mine (Zheng Mianping, 1989), significant economic benefits have been obtained. There are more than 25 large potassium-bearing saline lakes in the world, of which 10 have been mined, with an annual production of about 5 million tons, accounting for 15% of the world's potash production.

Development of halophilic organisms in saline lakes and saltfields As early as 1913, the existence of microorganisms specific to hypersaline environments (Artari, 1913) were reported. In the 1970s, it was found that the rhodopsin membranes of halophiles convertz sunlight into electric energy (Dietor, 1979). In order to produce large amounts of halophilic algae, at the end of the 1970s and at the beginning of the 1980s, Australia, the United States and Israel invested considerable funds in research on the purification and culture of algae and to determine the optimum conditions for the development of ,B-carotene or glycerine. For example, the Westfarmers Company of Australia invested more than 3.5 million dollars in research (1984-1986). The development of halophilic algae have now become a newly-emerging industry, and in Australia, the United States, Israel and China, it has reached the practical stage, i.e. factory-scale artificial culture of halophilic algae to a natural ,B-carotene has been widely used in making artificial butter and soft drinks. It is considered that natural ,B-carotene has an anti-cancer effect. As the use of Artemia eggs developed in some saline lakes and coasted marshes to high-quality fish and shrimps yielded good economic effects, the practice has developed rapidly in recent years. Many saline lakes and coasted marshes and beaches that are not suited for growing crops can be used to breed halophilic organisms. This is an important supplement to agriculture. Saline lake resources and related environmental protection Arid and semi-arid areas cover almost one third of the world's land area. Here, saline lakes are extensive. Saline lakes are distributed on all continents except the Arctic. Over a half the saline lakes lie in Asia, Africa and South America. The ecological environment of these regions are attracting increasing attention from scientist (Zheng, 1995) as well as commercial groups in many countries. The Jartai saline lake, Inner Mongolia, is an example. Before 1980, the exploitation of halite took place without regard to the protection of the surrounding vegetation, so that the erosion of drift sand became severe. By 1983, the sand was as thick as 0.5 mover 29% of the saline lake area of 87 km2. The sand dunes expanded towards the mining area at a rate of 33 m per year. This attracted great attention and a special fund was allocated for sand control and saline lake protection. A program was instituted to prohibit the

341 Table 1. Summary of types, magnitudes and uses of saline lake resources Type Classic material

Solidstate minerals

Dominant evaporites

Carbonates

Magnitude and occurrence

Uses

Clay, sand, gravel, volcanic ash

Large quantities available; economic-geographic conditions appropriate; exploitable

I. Calcite, aragonite, dolomite, magnesite Hydromagnesite

Commonly dispersed; low purity; but calcite occurs as large accumulations A few thousand tons discovered in Tibet, China Vary in magnitude from several hundreds tons and several hundred million tons (in U.S.A., Russia, Inner Mongolia of China, Kenya) Commonly contain a few per mil to a few hundred ppm Li2C03 (Zubuye Lake and Bangkog Lake, Tibet)

Used in building and road paving, and especially clay used in brick-making; volcanic ash used as absorbents for automobiles, ship docks and factories Calcite can be used as lime

2. Trona, soda and other alkalibearing materials

3. Li-bearing magnesite, Zabuyaite (Li2C03)

Sulfates

I. Gypsum and anhydrite

2. Thenardite and mirabilite

3. Epsomite and kieserite 4. Bloedite

5. Aphthitalite, picromerite, kainite, polyhalite, syngenite, and hanksite I. Halite

Mostly dispersed and locally occur as a few million tons of gypsum accumulations Mirabilite deposits and part of thenardite deposits have large magnitude and high purity, larger ones attain a few hundred million to a few billion tons (e.g. in Karabujiaz Gulf, Russia) Range from a few hundred to a few hundred thousand tons Generally dispersed, and in some areas reserves attain a few tens of million tons (e.g. in Russia) Dispersed and locally picromerite occurs as patash layers (Dalang P. and Da Qaidam L., China); can be mined for mUlti-purpose uses Ranges from a few hundreds of thousands of tons to a few hundreds of billions oftons (e.g. Qarhan P., China); widely distributed in world

Can be used for refractory magnesium bricks Widely applied in glass, textile, paper-making, dye, rubber, synthetic rubber, medicine, battery, detergent and chemical industries Mainly used in special glass, ceramic, aluminium smelting, chemical products, synthetic rubber, medicine, batteries, lubricant, air purifier Used in light building materials, agriculture, gypsum products and other daily life products Used in glass, paper-making, detergent, dynamite, medicine and metallurgical industries

Medicine and chemical industries Used in manufacturing magnesium glass for laboratory use yielding good effects Mainly used in agriculture and chemical industry

Mainly used in edible salt and basic chemical industry; those used for edible salt and fodder make up to ca. 65-70%

Continued on p. 342

342 Table 1. Continued Type Solidstate minerals

(Dominant) evaporites

Halides

2. Potash

3. Carnallite

4. Bischofite

Borates

Borax

Mgborate

Ulexite

Colemanite

Nitrates

Soda niter

Niter

10dates

Uranates

Calciurn chloride Silicates Others, Iron minerals

Lautarite, Brueggenite

Pitchblende, Ursilite, Carnotite, Autunite Antarcticite and other CaCl2 minerals Hydroglass, magadiite Goethite, limonite, pyrite, marcasite

Magnitude and occurrence

Uses

Commonly dispersed; in places reserves up from nx 104 to nx 107 tons (Potash Lake of Qaidam and Potash Valley of Red Sea) nx 104 to nx 107 tons KCl, and nx 104 to 107 tons MgCl; low grade; generally difficutlt to mine as an independent mineral (e.g. Qarhan P., Qaidam, China) Locally forms separate accumulations, up to a few million tons (Dalangtan P., China) B203 ranges from a few tens to a few millions tons; high-grade boron ores (e.g. W. U.S.A., plateau of S. America, Tibet of China)

About 90% used in agriculture, and the rest in the chemical and military industries (1) Use of KCl same as above; (2) magnesium salt used in producing metal magnesium and its alloy and medicine Same as above

B203 ranges from n x 104 to n x 106 tons; acid-soluble boron ore (e.g. Zacang Caka and Da Qaidam L. of China, Argentina) B203, a few tens to a few tens of thousand tons; can be partly decomposed in hot water (e.g. N. Chile and W. U.S.A.) B203, a few tens to a few tens of thousand tons; acid-soluble boron ores About a few tens to a few millions tons (e.g. desert area in N. Chile) Smaller in magnitude than soda niter (e.g. N. Chile and Uzong Bulak of Xinjiang, China Dispersed in niter ore; can be mined as a by-product (e.g. N. Chile) Can form accumulations of cornmercial value (e.g. W. Australia, W. U.S.A.) Can form commercial mineral deposits; known ones dispersed (e.g. N. Antarctic and Qaidam); can form commercial ore beds (e.g. Bristol Lake, W. U.S.A.) Mostly dispersed; locally occur as thin beds (Magadi Lake) Occur in a few acid saline lakes; goethite and limonite can make up modem deposits with over a few millions tons (e.g. Tyrrell L., Australia); high grade

Widely applied in more than 100 industries, such as glass, glass fiber, porcelain glaze, detergent and steel smelting Same as above

Same as above

Same as above

Used in dynamite, chemical industry, agriculture Used in chemical industry, dynamite and firewords Mainly used in medical industry, chemical industry, agriculture and environmental hygiene Used in nuclear power station and military aspeets Used in driers, chemical reagents

Not yet used Possible use in iron smelting

Continued on p. 343

343 Table J. Continued Magnitude and occurrence

Uses

Zeolite

Can form industrial ore beds (e.g. salt lakes in W. U.S.A.)

Bedded chert. opal Organic ooze

Bedded cherts may have larger magnitude (e.g. Magadi Lake, Kenya) Locally contains large amount of organic ooze and may form large bedded accumulations

Environmental protection, metal recovery and agriculture Not yet used

Type

Solidstate ore

Liquidstate ore

Sedi mentary

Silicates

Integrated brines

K-MgBr-Ca brine K-Mg brine

K-MgLi-B brine

B-Li-KW Brine B-Li-KRb-Cs brine Alkali brine Na sulfatehalite brine Cl- and alkalibearing brine BrBearing brine

Other components (e.g. Sr etc.)

Liquid fuel can be extracted Through heat treatment

Death Sea contains 143 km 3 brine of commercial value; has been developed Mainly Mg salt mined (e.g. Qarhan P. and Da Langtan P., China); magnitude, nx 107 to nx 108 tons K and Mg in modern saline lakes mainly mined from this kind of brine; K and Mg salts, a few million to a few tens of million tons (e.g. Yiliping P., China) For total uses; potash, nx 106 to nx 107 tons; Band Li grades high (e.g. Searles Lake, U.S.A.) B-Li-K-Rb-Cs mined for total uses; grade high; B-Li brine has large magnitude (e.g. Zabuya Lake, China) Mainly trona extracted (e.g. Kulongda alkali Lake, Russia) Mainly sodium sulfate extracted e.g. Kalabugaz, Russia)

K and Mg as well as CaCl2 can be extracted; used in driers and chemical reagents Same as above

Mainly trona extracted; F remains to be developed (e.g. Magadi Lake, Kenya)

Same as above

Belongs to a major type of Br deposit; magnitude of a few deposits up to IlX 107 tons (e.g. Kuqiuk Salt Lake, Russia); in a few saline lakes in Inner Mongolia, China, brine may attain commercial grade Used in manufacturing firefew thousand mg/l (Great Salt Lake, U.S.A.)

Widely used in petroleum, chemical, medical, dye, photographic, synthethic fiber, military and extinguisher industries as well as agriculture

Same as above

Same as above

Same as above

Same as above Same as above

works, sugar, glass etc.

Continued on p. 344

344 Table 1. Continued Type

Gas state

Natural gas Natural gas

Magnitude and occurrence

Uses

Occurs in salt beds or pore spaces of clastic beds, and its magnitude not clear (e.g. Searles Lake); sedimentary

Fuel and chemical industries

rock series of early Quarternary saline lakes contains natural gas accumulations of commercial value (e.g. structure of Qaidam Salt Lake, China) Halophilic Organisms

Halophilic Algae

Artemia

Spirulina

Halophiles and alkaliphiles

Physical properties and uses of solid-

Brine

liquidgaseous phases of saline lakes

Brine lakes at surface and intermediate and great depth (or man-made saline

Used to produce ,6-carotene, glycerin, fodder protein, tetratevebene oil, chlorophyll etc. (Australia, U.S.A., Israel etc.)

Rich in protein, several kinds of amino acid and unsaturated fatty acid, EPA etc. (U.S.A., Russia, China) Contains high-quality protein, several kinds of amino acid, chlorophyll, flax acid, carotene etc. (Mexico, U.S.A., China) Possess stable structure, function and genetic factors of salt and alkali resistance

/i-carotene used in food and medical industries; glycerin in dynamite and light industry; fodder protein in animal busbandry; tetratevebene oil in painting industry Used in culturing fish, shrimps and crabs

Mainly used in medicine and nutrients

Have important theoretical and practical significance; e,g, alkali-resistant micro-organisms used in treating black specks of alkali pulp

Possess heat -storing property, temperature close to boiling point (water depth) required to be > 3m)

Can be used in 'solar energy saline lake' electricity generation; good potential

lakes Solid-state

Transparent

Larger transparent halite

Used in ultra-red micro-

phase

halite crystal

crystals (e.g. early Quarternary salt lake pit in Dafeng-

scopes

shan, China); small magnitude, a few tens to a few hundred tons Solidgaseous phase

Medical sludge

Belongs to alga-bearing H2S sludges (e.g. Saji Lake and Death Sea, Russia)

Treats chronic diseases such as arthritis and skin diseases

345 felling of trees, strengthen afforestation and combine scientific research with production. Over a period of 10 years, Picea asperata Mast, Pinus sylvestris var. mongolica Litvin., Populus alba var. L.c. Pyramidalis Bunge, Ulmus pumila L., Elaeagnus angustifolia L., Tamarix chinensis Lour., Caragana microphylla Lam., etc. and medical plants such as Glycyrrhiza uralensis Fisch, Ephedra sinica Stapf and Licium chinese were planted. Now, shelter-forests of 1400 ha., and barricades of 70 ha., have been established and 23 km2 of trees and grasses recovered. The ecological environment of the saline lake region has been notably improved and fruit trees and fish ponds spread throughout the region which has now become a place for local recreation. Multidisciplinary research promoting the saline lake studies

Saline lake sediments have become an important subject of research on global change. Such study has brought about the integration of disciplines such as sedimentology and Quaternary geology and thus given impetus to the study of saline lakes and lacustrine paleo-environment and paleoclimatology (Smith, 1979; Gwynn, 1980; Zheng et aI., 1998). By using thermodynamic theory and computer simulation, the formation and evolutionary processes of brines have been deduced quantitatively (Eugster & Hardie, 1978). On the basis of Pitzer fluid theory, the law of salt removal of brines with various components in the process of evaporation and concentration has been simulated (Pitzer, 1991). Studies of saline lake organisms, ecological environments and biogeochemistry have aroused great interest among researchers in various fields. In all international symposia or conferences on salt lakes since 1979, large numbers of papers about saline lake biology, ecology and biogeochemistry have been presented, and new research directions have appeared, e.g. ecology of saline lakes (Brock, 1979; Williams, 1981), ecology of hypersaline environments, geoecology of saline lakes (Zheng et aI., 1985), environmental microbiology (Nissenbaun, 1980; Javor, 1989), biogeochemistry of hypersaline environments, agriculture of saline lakes (Zheng Mianping, 1995), and saline lake conservation and management. Recently, research on saline lake organisms has developed in scope and intensity. Besides the large number of workers who study halophilic algae and bacteria, Artemia, Spirulina and Rotifera, ad-

vances have been made in studies of some halophilic vertebrates and their ecological regimes. SaJinology The trends in research and utilization of saline lakes indicate that traditional research methods concerning saline lakes do not suit the needs of social and commercial development in saline lake science and technology. According to traditional classifications, saline lake research is a part of limnology. However, limnology is the study of the physical, chemical and biological properties and features of lakes (Moore, 1975), with emphasis on physico-chemical properties and biological resources. Thus, limonology does not include research on mineral deposits or resource engineering. Geological studies of saline lakes do not usually include consideration of biological resources and ecology. Therefore, traditional limnology or saline lake geology cannot cover all research and development of saline lakes. Hence, I suggest there is a need to establish a new term that covers the study of saline lake resources, 'salinology'. It is an applied branch of science dealing with the study of the chemistry, physics and biology of saline lake systems as well as the properties and features of their environment and resources and the promotion of their engineering development (Fig. 1). Basic task

The basic task of salinology is to study and explore the features of saline lakes by systematically using multidisciplinary principles. In this way, it will provide a scientific and technological basis for coordination and integration between man and saline lakes, and the comprehensive development and rational utilization of saline lakes and their sedimentary systems and make contributions to the development of 'saline lake agriculture, mining industry and tourism and their sustainable development. Salinology largely comprises the following disciplinary: (1) geology (including mineral deposits and mineralogy of salts, geochemistry, tectonic geology, hydrogeology and engineering geology); (2) biology (including biological taxonomy, ecology, geoecology, molecular biology and bioengineering); (3) chemistry (including salt solution chemistry, crystalline dynamics, inorganic chemistry of the salt-formation, organic chemistry and biochemistry in the saline environments); (4) engineering (including salt chemical

346

I

Salinology

A applied branch of science dealing with the study of the chemistry, physics and biology of saline lake systems, of the properties and features of their environments and resources, and promotion of their engineering development

I

I

I

Study, exploration, longterm observation, experiment and comprehensive evaluation of mineral resources of saline lakes

I

-

I

Saline Lake Chemistry

Saline Lake Geology

Saline Lake Engineering

I

I

Research on Phase separaseparation tion engineering of solar and extracpond, multition theory and technolopurpose use tgy of salts engineering, technical plant testing and economic assessment

I

-

Saline Lake Biology

I

Investigation and evaluation of organisms and ecology of saline lakes

I

/

Fishing technique, biological technique and medical sludge technique

Integrated studies of saline lakes. salt-field technology, mining and dressing techniques and high-value output

I

I

/

Extraction of information of paleoclimatic and paleoenvironmental changes

-

I

"""-

Studies of protection of ecological environment of saline lades and coordination between man and saline lakes

Prediction of the future climate and environment in the saline lake area Study of ecological engineering r--- measures, making-up and combination of ecological measures and engineering development

Reveal scientific laws

I

I

Saline Lake Environments

Serve the desisn-pilot test-scale production (inorganic salt industry and "agriculture" of saline lakes) and tourism industry

L ~

I Provide advice t policy-deciding Departments

Figure 1. Conceptual diagram of salinology.

engineering, solar pond technology and engineering of solar pond and beneficial saIt engineering); (5) environmental studies (Quaternary geology, geomorphology, hydrology, meteorology and conservation of saline environments).

Main research directions (1) Research, exploration and evaluation of mineral resources of saline lake sedimentary systems should be carried out according to the principles of the geology of mineral deposits, geochemistry, tectonic geology, hydrogeology and engineering geology by using conventional and new techniques. (2) Biological and ecological investigations and evaluation of biological resources of saline lakes

should be conducted according to the principles of biology, ecology and geoecology by integrating conventional and modem techniques. The following subjects are included: a. Ecological systems of saline lake and their biological resources: species composition, community structure, bio-productivity, food chains, models of energy flow, dynamics of dominant species-groups and environmental dynamics, rational use of saline lake resources and management for sustainable development. b. Geoecology of saline lakes: geological and geographical conditions of saline lake organisms, interactions between biological activity and the environment and deposition and mineralization.

347 c. Biological engineering of saline lakes: principles, means and feasibility of engineering development of biological resources of saline lakes. (3) Saline lake chemical study focuses on the processes of salt dissolution in solvent or crystallization from solution, relations among composition, structure and characters of solution, thermal chemistry, thermodynamics, phase balance thermodynamics of solution and crystallization kinetic mechanism of the water-salt system, principle and application of heat storage and electrochemical energy of 18 salt-forming elements. (4) On the basis ofthe knowledge obtained by studies of (I), (2) and (3) above, we should carry out investigation of well-integrated engineering techniques with high-value output and comprehensive development as the target, including division of saline lake areas, saltfield technology, mining and dressing techniques, fisheries and biological techniques, medical sludges and other uses. (5) In light of principles of sedimentology and Quaternary geology, and through a study of such indicators as the high resolution of sediments and geomorphological features of saline lakes, we should extract information on past global changes and elucidate paleoclimatic changes and the response of lakes to the interaction of man and nature. In key saline lakes, we should establish long-term meteorologicalhydrological and biological observation stations to obtain systematic values of changes of ecological environments and integrate the observations with geographic information system (GIS) techniques to monitor the dynamic change in saline lake areas. On that basis, we will study the engineering development of protective measures for the ecological environments of saline lakes.

Acknowledgements This research was supported by the National Natural Science Foundation of China Grant 49833010.

References Bogoslovskii, B. B. & C. D. Muraveiskii, 1955. Introduction to Limnology. Moscow University Press, Moscow (in Russian). Borowitzka, L. J. et aI., 1981. The microflora-adaptations to life in extremely saline lakes. Hydrobiologia 81: 33-40. Brock, 1. D., 1979. Ecology of Saline Lakes. Strategies of Microbial Life in Extreme Environments, M. Shilo: 29-47.

Dietor, w.. 1979. Reviews on bioenergetics. Biochem. Biophys. Acta 505: 279-353. Dzens-Litovskii, 1957. Comprehensive Research and Exploration Methods of Saline Lake Mineral Deposits. Leningrad Publishing House (in Russian). Eugster, P. & A. Hardie, 1978. Saline Lakes - Lakes Chemistry Geology Physics (Chapter 8), Springer Verlag, New York INC: 237-283. Guillcrino, C. D., 1988. The Cenozoic saline deposits of the Chilean Andes between 18° 00 and 27° 00' south latitude. Lecuture Notes in Earth Sciences, Vol. 17. Springer-Verlag. Berlin Heiidelbeg. Gwynn, J. w., 1980. Great Salt Lake - a Scientific, Historical and Econonmic Overview. In Utah Geological and Mineral Survey: 1-400. Hammer. U. T., 1981. Primary production in saline lakes: a review. Hydrobiologia 81: 47-57. Hurlbert, S. H. (ed.), 1993. Salt Lakes V-Developments in Hydrobiology. Kluwer Academic Publishers, Dordrecht: 87 pp. Javor, B.. 1989. Hypersaline Environments. In: Microbiology and Biogeochemistry, Springer-Verlag, Berlin. Kudhner, D. J., 1978. Microbial Life in Extreme Environments, Academic Press INC. London: 319-352. Matter, A. & M. E. Tucker, 1978. Modern and Ancient Lake Sediments. Blackwell Schentilic Publications, Oxford, London. McCracken, D. & Mike, H., 1998. Lithium, Minerals, Supplement to Mining Journal, Aug. 7, II. Nisserbaum, A. (ed.), 1980. Hypersaline brines evaporite environments. Dev. Sedimentol. 28: 23-29. Oren, A., 1999. Bioenergetic aspects of Halophilism. Microbial. Molec. BioI. Rev., 63: 334-348. Pitzer, K. S., 1991. Activity Coefficients in Electrolytes Solutions, 2nd edn. CRC, Florida: 75-153. Smith, G. I.. 1979. Supsurface Stratigraphy and Geochemistry of Lake Quaternary Evaporite, Scarles Lake California, U.S. Geol. Survey Prof. 1043: 98-100. Usiglio, I., 1849. L.' Analyse de Ipeau de la Mediterannee sur les coles de France. Etutes sur la composition leau dela Mediterannee et sur Ipexploitation des sels quele contient. Ann. de Chim. et de phys. XXVII. Williams, W. D., 1981. Inland salt lakes: Introduction. Hydrobiologia 81: 1-13. Williams, W. D., 1996. How important is salinity in structuring biological communitities in salt lakes" Resources environment and global change of saline lakes. In Selected papers of the 6th International Conference on Salt Lakes. Beijing: Geological Publishing House: 103-114. Yuan Jianqi, 1982, New development of the genetic theory of saline mineral deposits and its position in the geology of mineral deposits. Geology of Mineral Deposits, 1(1): 16-33 (in Chinese with English abstract). Zheng Mianping & Liu Wengao, 1985. The discovery of halophilic algae and halobacteria at Zabuye Salt Lake Tibet and preliminary study on the geoecology. Acta Geologica Sinica, 2: 162-171. Zheng Mianping, 1995. On 'agriculture of saline lakes'. J. Geosci. 4: 404 (in Chinese with English abstract). Zheng Mianping (ed.), 1996. Resources Environment and Global Change of Saline Lakes. In Selected Papers of the 6th International Conference on Salt Lakes. Geological Publish House, Beijing: 1-20: 145-157 (in Chinese). Zheng Mianping, Zhao Yuanyi & Liu Junying, 1998. Sedimnetation and paleoclimate of Quarternary saline lakes. Quaternary Research 4: 297-307 (in Chinese).

Loading...

Saline Lakes

Saline Lakes Developments in Hydrobiology 162 Series editor H. J. Dumont Saline Lakes Publications from the 7th International Conference on Salt ...

50MB Sizes 1 Downloads 0 Views

Recommend Documents

No documents