Introduction

Although they are small features on Earth’s surface, lakes process a tremendous amount of carbon.  While surface waters are a source of carbon to the atmosphere (0.14*10¹⁵ g C/yr; Cole et al 1994), the lake sediment sink stores a globally significant amount of carbon annually (0.3*10¹⁵ g C/yr in lakes, reservoirs, and peatlands; Dean and Gorham 1998).  Sedimentary records of the past 10,000 years show rapid shifts in the magnitude and nature of organic and inorganic (mineral) carbon storage, indicating that burial rate and style are sensitive to relatively minor changes in the state of the water body.  Under changing climatic, atmospheric, and land-use conditions, such threshold behavior may lead to abrupt changes in the direction and magnitude of the carbon flux through lakes.  Because the processes which govern production and preservation of particulate carbon in lakes are incompletely understood, however, it is unclear whether such a shift would act to buffer atmospheric carbon dioxide or exacerbate its rising levels.  I hope to determine which parameters directly control whether inorganic or organic carbon is preserved in lakes, and the rates at which burial occurs.  Because style and rate of carbon burial change rapidly, my research takes the form of a test of system sensitivity across four key natural and anthropogenic transitions in the sedimentary record.
 

Context and Significance

I have chosen to study lakes for my thesis research because they encompass, on a manageable scale, two of my major interests:  the global carbon cycle, and threshold system behavior.  Lakes are elegant microcosms of terrestrial carbon cycling processes, acting as crucibles for the transformation of carbon from one form to another.  Phytoplankton productivity, precipitation and dissolution of carbonate minerals, bacterial respiration of organic matter, and air-water gas transfer rapidly shift the size and dominance of particulate and dissolved organic and inorganic carbon pools over the course of a year.  Because of their rapid response time, lakes are ideal (though complex) natural laboratories for studying these processes.  Their sediments provide a record of the variability of fluxes over longer time scales and under environmental conditions different from those of today.

We have learned from paleorecords that Earth systems exhibit rapid switches rather than the gradual transitions we would imagine on the basis of looking at change on a human timescale.  Snapshots of the Earth do not show thresholds, nor do models adequately simulate threshold behavior.  While mine is by no means a paleoclimate or paleoecology study, it does take advantage of a well-defined and -timed Holocene and Recent sequence of climate and landscape scenarios in Minnesota as boundary conditions.  We know what experiments Nature has run, so regional and watershed changes, such as increased aridity, reforestation, and changes in nutrient flux can be tested as potential causes for the shifts in carbon storage observed in the sedimentary record.

Based on initial results and a conceptual model of the lake carbon cycle, my starting hypothesis is that residence time of the system is the main control on organic and inorganic carbon storage.  A lake with a relatively long water residence time (e.g. a closed basin, which has no surface outflow) will have a longer nutrient residence time, which increases the biologic productivity of the system.  Sinking organic matter is oxidized, producing organic acids and carbon dioxide, which lowers the pH of bottom waters and dissolves carbonate minerals (inorganic carbon).  Similarly, a lake with a high nutrient influx has a high nutrient residence time, which will likewise increase productivity and lead to dissolution of carbonates.  Either of these situations may have an abrupt initiation:  in the first case, corresponding to the natural transition, a lowering of the water table (such as might occur under increased aridity) cuts off the surface outflow; in the second, corresponding to the anthropogenic transition, human land clearance and shoreline development lead to increased nutrient flux to the lake.
 
 
 

Modeling

Lakes may be treated as box models for carbon (see lakecmodel.pdf) having a number of measurable parameters with known geochemical characteristics and effects.  To measure or monitor every component of even a small lake system would be beyond the scope of a Ph.D. thesis project, so I propose testing of the most important and relevant components, along with the use of results of prior studies (e.g. Winter, ed., 1997) with the assumption that this approach sufficiently characterizes the system (see Methods below).  For my purposes, the significant reservoirs are dissolved inorganic carbon (DIC; comprising CO2(aq), H2CO3, HCO3^-, and CO3^2-), particulate organic carbon (POC; algal, bacterial, and inwashed terrestrial organic matter), and particulate inorganic carbon (PIC; carbonate minerals precipitated within the lake surface waters).  The carbon stable isotopic signal (d¹³C) is of great importance:  DIC imparts its isotopic signature to both organic and inorganic carbon produced within the lake, and these components are used to reconstruct productivity and lake state through time.

During the course of my research, I will develop a lake carbon model using Stella™ systems modeling software, and iteratively refine the model as new data come in.  While some models of lake carbon exist, mine would be among the first to deal with changing carbon storage.  Because all lakes share common aspects of the carbon system, this model would be broadly applicable to lakes around the world.  An added value of the development of such a model is its application to teaching:  students find the carbon cycle a more easily-understandable beast when it is presented in terms of a familiar and manageable system such as a lake.
 

Study Site and Plan of Attack

For my thesis research I have chosen as an experimental site a pair of small lakes which are nearly identical in setting and chemistry, but which represent contrasting systems with respect to carbon storage.  In lakes only 1 km apart, with the same geologic substrate and similar water chemistry, one would intuitively expect similar sediment composition; however, one lake has surface sediments high in organic carbon (algal and terrestrial plant matter), while the other is rich in inorganic carbon (calcite precipitated in the lake surface waters).  In a sediment core of the organic-rich lake, I have found that a few thousand years ago the lake stored a high percentage of inorganic carbon, and that the switch from inorganic to organic dominance occurred abruptly, over a period of tens to hundreds of years.  Similarly, the inorganic-rich lake has a dramatic increase in organic carbon , and a drop in inorganic carbon, in the late 19th century, as a consequence of tree clearance in the immediate region and residential development of the lake shore.  In sediments dated to the 1970s and later, there is a decrease in organic, and increase in inorganic carbon, probably related to a decrease in nutrient inputs to the lake.

My goal is to compare the rate and nature of these shifts:  is an increased nutrient input also the cause for the pre-cultural organic increase?  Or is the shift instead the result of climate and vegetation change?  How quickly did the natural transition occur?  Obviously, water chemistry and biologic productivity in the lake and its watershed will have a direct influence on the particulate matter sedimented in the lake; however, we have few studies which address either the interrelationship of organic and inorganic production and preservation, or the effects on chemistry and sediment of different types of productivity (e.g. terrestrial vs. algal vs. bacterial).  My high-resolution analysis of sediments across threshold transitions will employ geochemical equilibrium modeling to determine the conditions for precipitation and preservation of mineral forms of carbon, and organic geochemical analysis to characterize the relative importance and effects of different groups of primary producers.

Lakes respond rapidly to internal and external forcing because of their small reservoir size.  This sensitivity, however, does not fully explain why they exhibit threshold behavior, nor does it illuminate the nature of the forcing functions responsible.  Why would carbon within two similar systems respond in different ways to the same climate, vegetation, and land-use conditions?  Are observed threshold switchovers caused by threshold forcing or some sudden event (pulse), or is the input function a gradual change which ultimately causes the system to respond abruptly?  I have chosen this pair of lakes to represent two parallel natural experiments with known historical forcings:  climate and vegetation change (gradual); basin closure (a step function); land clearance (a pulse); and anthropogenic nutrient loading (which exhibits gradual, step, and pulse characteristics).  My study proposes a detailed comparative analysis of the sediments across four transitions:  Late Glacial to Holocene, mid-Holocene, late 19th century settlement, and late 20th century remediation.  The inverse relation of inorganic and organic carbon holds true for many Minnesota lakes (Dean, in press), so this detailed study and model provide a basis for a regional comparative analysis of lower analytical complexity, and generalization of results to terrestrial lake carbon storage at large.
 

Methods

I am currently studying five cores from Green and Spectacle Lakes.  One more from each, spanning the Late Glacial-Holocene boundary, is planned for 1999, as well as a transect of 3 to 4 cores from the shore to the center of the lake to determine the distribution and thickness of sediments.

To build a model of the lake I must have calibration values for the modern system.  To this end, I will sample for a suite of water chemical analyses at three depths in the lake, three to four times during the summer.  Ca²⁺, Mg²⁺, alkalinity, and pH will determine carbonate saturation using geochemical modeling software; dissolved O2 and CO2 will be tested to clarify conditions to which sinking particulates are subject.  The degree of CO2 saturation of lake surface waters is an essential measurement which may have bearing on carbon stable isotopic values; CO2 measurements in surface waters and the overlying atmosphere will be conducted by the method of Raymond et al (1997).  To distinguish production from preservation signals in the sediment, I need isotopic and geochemical information about particulates produced in the water column during the summer.  Lake water filtrate (organic and inorganic particulates) obtained with an N2-pressurized system (Hollander, pers. comm) from three depths will be characterized, and analyzed for stable isotopes.  Combined with dissolved inorganic carbon (DIC) stable isotopes taken at the same time, these show how particulate inorganic and organic d¹³C tracks DIC d¹³C over the course of the summer (Hollander and MacKenzie 1991).

Characterization of the relative contributions of terrestrial and aquatic organic carbon, and detrital and authigenic inorganic carbon in the sediments is the key to understanding the significance of carbon stable isotopic measurements.  Terrestrial, aquatic and bacterial organic matter can be distinguished with relatively simple techniques by a combination of their d¹³C and C:N ratios (Meyers and Ishiwatari 1993), by plant pigments (Leavitt 1993), and by hydrogen-index values (Dean 1999b).  If more in-depth analysis is necessary, organic geochemical separations and compound-specific isotopic techniques will be employed at high-resolution across the chosen transitions.  A novel but simple technique for assessment and counting of sediment components is smear slide analysis, which uses a small amount of sediment mounted on a slide and viewed under a petrographic microscope with polarized light (http://lrc.geo.umn.edu/smears/smear.html).

Samples for bulk density enable calculation of mass accumulation rates, necessary for meaningful characterization of quantitative changes in storage.  Total carbon and total inorganic carbon measurements by coulometric titration, determine weight percentages of inorganic and organic carbon, and provide the initial framework for sample selection and evaluation of the switchovers in dominance from one form to the other. Across the transitions, cm-scale samples for organic and inorganic stable isotopic analysis will characterize signals of productivity, redox, hydrology, and temperature. Dating of the cores is done with ²¹⁰Pb for the cultural section and AMS radiocarbon on terrestrial macrofossils or aeolian charcoal below.
 
 

Conclusion

The global carbon cycle is a fascinating, esoteric, and truly significant subject, and one on which lifetimes of research are spent.  A Ph.D. thesis project is necessarily limited in scope, but can make a meaningful contribution if the global context and questions are held firmly and continually in mind.  Basic carbon cycle research is vital:  my project has a local focus and a practicable experimental sweep, but wrapped into it are greater implications for the understanding of system response.  I have been inspired by my participation in the USRA Earth System Science Education community, and the three components of my study - thresholds, modeling, and education - together provide a basis for a career in research and teaching on questions of biogeochemical dynamics central to our knowledge of the planet and our place within the system.
 
 
 
 
 

REFERENCES
 
 

Cole, J.J., N.F. Caraco, G.W. Kling, and T.K. Kratz 1994.  Carbon dioxide supersaturation in the surface waters of lakes.  Science 265 (9 Sept 94):  1568-1570.

Dean, W.E. 1999b.  Characterization of organic matter in lake sediments from Minnesota and Yellowstone National Park.  Submitted.

Dean, W.E., and E. Gorham 1998.  Magnitude and significance of carbon burial in lakes, reservoirs, and peatlands.  Geology 26:6, 535-538.

Hollander, D.J., and J.A. MacKenzie 1991.  CO2 control on carbon-isotope fractionation during aqueous photosynthesis: a paleo-CO2 barometer.  Geology 19, 929-932.

Leavitt, P.R. 1993. A review of factors that regulate carotenoid and chlorophyll deposition and fossil pigment abundance. J. Paleolimnol. 9: 109-127.

Meyers, P.A., and R. Ishiwatari 1993.  Lacustrine Organic Geochemistry – an overview of indicators of organic matter sources and diagenesis in lake sediments.  Org. Geochem. 20:7, 867-900.

Raymond, P.A., N.F. Caraco, and J.J. Cole 1997.  CO2 concentration and atmospheric flux in the Hudson River. Estuaries 20, 381-390.

Winter, T.C., ed., 1997.  Hydrological and Biogeochemical Research in the Shingobee River Headwaters Area, North-Central Minnesota.  U.S. Geological Survey Water-Resources Investigations Special Report 96-4215.
 
 

This document, in a slightly different form, was submitted as a proposal (ultimately unsuccessful) to the NASA-Earth System Science Graduate Student Fellowship program in March 1999.