Hydrologic and Biogeochemical Fluxes in Land-Water Mosaics

Report to the A.W. Mellon Foundation
1 December 2005 (citations updated on this website on July 15th, 2008)
PIs:  Steve Carpenter, Jon Foley, and Monica Turner
Postdoctoral Associates:  Jeff Cardille, Mike Coe, Paul Hanson
Graduate Student:  Julie Vano

Project Summary

The goal of our project is to understand how the extent of surface water and wetlands affects ecosystem production, respiration, and spatial flow of organic carbon on complex, heterogeneous landscapes.  The approach centers on development of a new integrated spatial simulation model for hydrology and biogeochemistry of the Northern Highland Lake District (NHLD) of northern Wisconsin and Upper Michigan, a complex landscape of with over 7500 lakes and diverse terrestrial ecosystems.

In the past year we have completed the major development and testing of the LUWI model (Lake-Upland-Wetland Interactions) and are actively using it for simulations of the region. We created a new, parsimonious model of spatial mass transport suitable for a wide variety of ecological flows. The model was designed with sufficient generality to be applied to a variety of other settings in which the effect of lateral fluxes of matter (e.g., nutrients, propagules, energy, organisms) is of interest. With an initial focus on hydrologic processes in the NHLD, we created a parameterization of water and carbon flow through a known flow path where we have gauged stream flows and long-term carbon concentration measurements. The parameterization successfully simulates lake type and streamflow using recent monthly climate data. The same parameterization successfully simulates carbon concentrations of inorganic and organic carbon in these lakes while demonstrating that carbon flux to the atmosphere also matches observed values. Simultaneously, we used the general form of the model to pursue our study of the relative effects of uncertainty on estimates of outputs from networks that transport mass through space.

In four areas-- hydrologic modeling, terrestrial modeling, lake modeling, and integrated modeling-- we have made substantial progress in capitalizing on past field work and new simulation modeling to extrapolate across the entire NHLD. Within a spatially explicit framework, this model integrates ecological processing (on land and in wetlands and lakes) with physical hydrology (of the land, lakes, wetlands, and groundwater pools) to estimate, for each watershed, seasonal water cycle and biogeochemical dynamics.

We developed a terrestrial ecosystem model to investigate how climate and land cover change alter the landscape water balance.  The water balance and carbon flows from the terrestrial ecosystems are necessary for the integrated model.  The terrestrial landscape has been quantified and its water balance simulated.  Specifically, the foundations now in place include:  1) the landscape's soil and vegetation heterogeneity have been defined for all of the ~8,000 watersheds, 2) each vegetation and soil type combination has water balance values (e.g. evapotranspiration, soil moisture, surface runoff, drainage) modeled for every day from 1951-2000, and 3) the model's biophysical parameters have been tuned to the Northern Highland environment and modeled values compare well to a diverse range of field measurements including latent heat fluxes, snow depth, stream gauges, soil temperature and moisture values throughout the region.   

In 2005, we used these foundations to further investigate the biophysical and hydrological land surface processes specific to cold temperate climates.  This required special attention to complexity of the seasonal freezing and thawing of the soil and the snow mass and energy balance.  We found that snow cover and a soil surface organic layer (e.g. decomposing litter fall) play a key role in wintertime soil temperatures and the timing of spring thaw.  We also examine how the land cover impacts the flow and storage of water by comparing both seasonal and interannual variations over the land cover types in the Northern Highlands including grasslands, shrublands, coniferous and deciduous forests, as well as soil types ranging from sand to clay.   We found that both land cover and climate variability affect the regional hydrology, but these mechanisms impart change expressed at different timescales. Climate variability has an obvious signal from one year to the next, but did not persist long across the 50 years.  In contrast, land cover and soil type affects the hydrology in subtler, but persistent ways, which have large cumulative differences in the long term.

In 2004 we have progressed from a published lake model that describes the carbon cycling details of a variety of lakes (Hanson et al. 2004) toward a model suited specifically to a spatially-extensive landscape analysis of carbon flow in thousands of hydrologically-linked aquatic ecosystems. We have distilled the original model into a simpler form that retains processes and variables critical to understanding the roles lakes play in carbon cycling at a much larger scale. This compression serves the pragmatic need for model efficiency in a large, integrated project, but also raises a question. What effect does model complexity have on our ability to describe carbon cycling at large scales? Ideally, we would approach this question by confronting our models with data representing the great variety of lakes in the NHLD; however, before 2004 no data sets existed that represent an unbiased sampling of lakes in the NHLD. During the summer of 2004, we sampled 170 lakes chosen at random from the NHLD. Data from these samples will be used to help refine, calibrate, and evaluate our carbon cycling models, leading to answers for the question posed above. Just as importantly, the data, which represent a snapshot of the NHLD carbon cycling state, will serve a critical role in validating the larger landscape carbon cycling model.

The analysis of data from the 170 lake survey has revealed a number of important limnological characteristics for NHLD lakes. Most lakes are small (mean area = 10 ha) and shallow (mean depth = 4 m) and  have moderate to high dissolved organic carbon (mean DOC = 7.6 mg L-1) and total phosphorus (mean TP = 10 µg L-1) and low acid neutralizing capacity (mean ANC = 134 µEq L-1). However, small lakes account for a small proportion of the total surface area of water in the region. Half of the surface area is in lakes larger than about 420 ha, and these larger lakes tend to have low DOC and TP and high ANC. Thus, when describing regional lake characteristics, it is important to differentiate between number of lakes and surface area of lakes. For example, most lakes have DOC concentrations that are twice the dissolved inorganic carbon (DIC) concentrations, but when adjusting for lake surface area, those numbers are reversed --  most of the lake surface area in the region has DIC concentrations that are three times greater than DOC concentrations. These insights into the distributions of standing stocks reveal the current carbon state of lakes for the region and suggest that lakes of different size play different roles in processing carbon at the landscape level.

Through modeling we estimate hydrologic and carbon fluxes through the landscape that lead to the wide range of carbon states found in NHLD lakes. We found that the area and shape of lakes, as well as their positions in the landscape and their hydrologic regimes determine their carbon loads and hydrologic residence times. Smaller lakes have relatively high organic carbon loads and short residence times, whereas larger lakes have higher inorganic carbon loads. Furthermore, the long residence time of larger lakes leads to more complete mineralization of the organic carbon pool, which explains in part the high DIC:DOC ratio in larger lakes. The results from the 170 lake survey were presented at the 2005 Ecological Society of America meeting and serve as the basis for a manuscript under development (Hanson et al. 2005). We anticipate submitting this manuscript for publication before December 2005.

With progress in hydrologic, terrestrial, and lake modeling having proceeded in parallel, we developed the integrated simulation model of water and carbon fluxes. The integrated model links results from the terrestrial and aquatic modeling components to understand the flow and storage of carbon and water across this region. By integrating carbon measurements and hydrologic measurements across three well-studied watersheds of the NHLD, we calibrated the model for both water and carbon processes across multiple nested scales, and are nearly ready to extrapolate to the NHLD as a whole.

In the past year, we applied the calibrated model in a variety of settings that explore the relationship between water and carbon in the Northern Highland Lake District.  Of particular interest were contrasts between simulations that consider both carbon and hydrologic criteria: in particular, how much do carbon measurements tell us about carbon loadings to wetlands and lakes? We also have developed simulations of historic drought and potential climate change, since the timing and amount of precipitation strongly influence lake type, water flow, and lake processes. We also ask whether a more complex lake carbon processing model better simulates lake carbon dynamics than the current simple model. Our terrestrial work continues to explore the major drivers of terrestrial carbon and water flux: weather, climate, and vegetation.

These simulations and analyses have advanced the ability to simultaneously understand connections among a large, tightly coupled set of lakes. Beginning with a focus on about 100 lakes upstream from a well-understood LTER lake, we hypothesized varied groundwater connections among them and found that the type and number of connections can strongly influence both the hydrologic and biogeochemical state of downstream lakes in today’s climate. In addition, because some of our modeled lakes are isolated and disconnected with little surrounding watershed area, we have used the model to isolate the effect of allochthonous carbon input to lakes, and have made new hypotheses about the timing, amount, and biogeochemical importance of aerial inputs to lakes of the NHLD.

Given a study set of hundreds or thousands of lakes, our large set of coupled hydrologic and carbon outputs demanded a new framework for displaying similarities and differences among individual lakes and among lake sets. To display even our most basic results, we derived new methods for summarizing carbon and water budgets at the subregional and regional scale. We developed a novel representation of the major pathways of carbon and, separately, for water. The triangular budget diagram displays the carbon and budget differences for larger sets of lakes, and is also suitable for describing aggregate differences in budgets of chains or categories of lakes. These methods have allowed us to find new relationships among lakes beyond the typical descriptions of lakes as either seepage or drainage, or carbon source or sink.

Our new abilities to display variation among large sets of lakes allows us to understand the lake region’s possible behavior under regional perturbations (e.g., climate change) or local pressures (e.g., land use change). In simulations of these changes, we can isolate differences in the reactions of different lakes to such perturbations as a function of each lake’s morphometry, watershed characteristics, beginning state, and upstream and downstream connections.  We extended our triangular representation of budgets to display, for a large set of lakes, the changes in lake carbon and/or water budgets to different stressors. This extends our understanding of the way that lakes may be similar, for example, in hydrologic type but quite different in their carbon processing.

With the model well calibrated in a handful of lakes and shown to extrapolate well to a hundred lakes, we have the proper foundation for simulating the entire Northern Highlands region. Given our new data set from 170 randomly selected lakes, we have gathered the data necessary to evaluate a regional extrapolation. With new methods developed to analyze results of a large number of variables from thousands of lakes, we can summarize an entire simulation, extract information for a set of lakes, or interpret more detailed results for a single lake. We are now analyzing the results of our 7000-lake simulation in today’s climate, confronting the model with data, and comparing results in selected lakes and watersheds to a range of inputs. Our explicit incorporation of terrestrial and aquatic processes in surface and subsurface connection networks will aid our understanding of the relative roles of on-land, in-lake, and between-lake processes in this lake-rich region.

Publications

Cardille, J. A., Turner, M. G., Clayton, M., Gergel, S., and S. Price, In press. METALAND: a publicly available tool for characterizing spatial patterns and statistical context of landscape metrics across large areas. Bioscience "Biologists' Toolbox".

Cardille, J. A. and M. Clayton. 2007. A regression tree-based method for integrating land-cover and land-use data collected at multiple scales. Environmental and Ecological Statistics 14(2):161-179.

Cardille, J. A., Coe, M. T., and J. A. Vano. 2004. Impacts of climate variation and catchment area on water balance and lake hydrologic type in groundwater-dominated systems: a generic lake model. Earth Interactions 8(13): 1-24.

Carpenter, S.R., J.J. Cole, M. L. Pace, M. Van de Bogert, D.L. Bade, D. Bastviken, C.M. Gille,  J. R. Hodgson, J. F. Kitchell, and E. S. Kritzberg. 2005.  Ecosystem subsidies: terrestrial support of aquatic food webs from 13C addition to contrasting lakes.  Ecology 86:  2737-2750.

Hanson, P. C., A. Pollard, D. L. Bade, K. Predick, S. R. Carpenter, and J. Foley. 2004. A model of carbon evasion and sedimentation in temperate lakes. Global Change Biology. 10: 1285-1298.

Hanson, P.C., S.R. Carpenter, D.A. Armstrong, E.H. Stanley and T.K. Kratz. 2006. Drivers of lake dissolved inorganic carbon and dissolved oxygen across scales from days to decades.  Ecological Monographs 76(3):343-363.

Turner, M. G. and J. A. Cardille. 2007. Spatial heterogeneity and ecosystem processes. Pages 62-77 in: J. Wu and R. J. Hobbs, editors. Key topics in landscape ecology. Cambridge University Press.

Turner, M. G. 2005. Landscape ecology in North America: past, present and future. Ecology 86:1967-1974.

Turner, M. G. 2005. Landscape ecology: what is the state of the science?  Annual Review of Ecology, Evolution and Systematics 36:319-344

Turner, M. G. and F. S. Chapin, III. 2005. Causes and consequences of spatial heterogeneity in ecosystem function. Pages 9-30 in: G. M. Lovett, C. G. Jones, M. G. Turner, and K. C. Weathers, editors. Ecosystem function in heterogeneous landscapes. Springer-Verlag, New York.

Vano, J.A., Foley, J.A., Kucharik, C.J., and M.T. Coe, Evaluating the seasonal and interannual variations in water balance in northern Wisconsin, USA, using a land surface model.  Journal of Geophysical Research - Biogeosciences, 111(G2): G02025.

Vano, J.A., Foley, J.A., Kucharik, C.J., and M.T. Coe,  Investigating the Controls of Land Cover and Climatic Variability on the Land Surface Hydrology of Northern Wisconsin, USA, In preparation.

Presentations

Cardille, J. A., Hanson, P. C., Vano, J. A., Coe, M. T., and S. P. Cornelius. Interactions of terrestrial and aquatic processes among lakes of the Northern Highland Lake District, USA. American Geophysical Union Fall Meeting, December 2004.

Cardille, J. A., Coe, M. T., Turner, M. G., Carpenter, S. R., Foley, J. A., Hanson, P. C., and J. A. Vano. Effect of surface and subsurface flow networks on spatial variation in ecosystem processes of lakes of the Northern Highlands, WI, USA. Ecological Society of America Conference, Portland Oregon, USA, August 1-6, 2004.

Cardille, J. A., Coe, M. T., Turner, M. G., Carpenter, S. R., Foley, J. A., Hanson, P. C., and J. A. Vano. A framework for analyzing terrestrial-aquatic interactions in a land-lake mosaic. Invited presentation in symposium "Applying Landscape Ecology in Forests of the Northern Great Lakes Region", International Association for Landscape Ecology (US-IALE) Conference, March 31, 2004, Las Vegas, Nevada, USA.

Carpenter, S.R.  Molecules and Trees:  Terrestrial Organic Carbon and Lake Processes.  Presentation to Groupe de Recherche Interuniversitaire en Limnologie, Montreal, Quebec, Canada, 6 March 2004.

Carpenter, S.R.  Understanding Change in Complex Regional Systems.  Presentation to University of Stockholm Department of Systems Ecology, Sweden, 2 September 2004.

Carpenter, S.R.  Carbon Cycling in Lake Districts.  Presentation to Center for Ecological Research, Kyoto University, Japan, 28 October 2004.

Hanson, P. C., A. Pollard, D. L. Bade, K. Predick, S. R. Carpenter, and J. Foley. 2003. A model of carbon flux and sedimentation in linked landscape-lake ecosystems. 2003. Paper presented at the American Geophysical Union conference, San Francisco, CA.

Vano, J.A., Foley, J.A., Coe, M.T., and Kucharik, C.J., 2004, Modeling the effects of heterogeneity on landscape water and energy balances in northern Wisconsin, 89th Annual Meeting of the Ecological Society of America, Portland, Oregon.  

Vano, J.A., Foley, J.A., Kucharik, C.J, and Coe, M.T,, 2004, Evaluating the Influence of Various Vegetation and Soil Types on Water and Energy Balances in Northern Wisconsin Using a Dynamic Biosphere Model, American Geophysical Union Fall Meeting, December 2004, San Francisco, CA.  

Vano, J.A., Foley, J.A., Kucharik, C.J., and M.T. Coe,  Investigating the influences of climatic variability and land cover change on the land surface hydrology of northern Wisconsin, USA, American Geophysical Union Fall Meeting, December 2005, San Francisco, CA.  

Hanson, P.C., S.R. Carpenter, J.A. Cardille, and M.I. Coe. 2005. Limnological, watershed, and carbon cycling characteristics of a random sample of lakes from the Northern Highland Lake District of Wisconsin. Paper presented at the Ecological Society of America conference. Montreal, Canada.

Cardille, J. A., Coe, M. T., Hanson, P. C., Vano, J. A., Carpenter, S. R., Turner, M. G., and J. A. Foley. A unified aquatic-terrestrial carbon and water budget for a lake-rich region: modeling today's state and projecting future changes. Ecological Society of America Conference, August 2005.