Carbon Cycling in Northern Peatlands E-Book
133,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Serie: Geophysical Monograph Series
- Sprache: Englisch
Published by the American Geophysical Union as part of the Geophysical Monograph Series, Volume 184.
Carbon Cycling in Northern Peatlands examines the role that northern peatlands play in regulating the atmospheric carbon budget. It summarizes current research in four interconnected areas: large-scale peatland dynamics and carbon cycling; plant and microbial dynamics and their effect on carbon fluxes to the atmosphere; methane accumulation in, and loss from, peatlands; and water and dissolved carbon fluxes through peatlands. The volume highlights include
- A thorough assessment of the challenges involved in incorporating carbon cycling in northern peatlands into global climate models;
- A conceptual model to examine the partitioning of terminal carbon mineralization into production of CO2 and CH4;
- A comprehensive review of the evidence for the accumulation of methane in deep and shallow peat; and
- A description of the hydrologic changes induced by peat harvesting and associated challenges in restoring altered peatlands to their natural hydrologic regime.
Carbon Cycling in Northern Peatlands will be of interest to research scientists and graduate and undergraduate students, particularly those who wish to know more about the role of peatlands in the global carbon cycle and their role as modifiers of climate.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 1025
Veröffentlichungsjahr: 2013
Ähnliche
Contents
PREFACE
Understanding Carbon Cycling in Northern Peatlands: Recent Developments and Future Prospects
Section I: Large-Scale Peatland Dynamics and Carbon Cycling
Nonlinear Dynamics of Peatlands and Potential Feedbacks on the Climate System
1. INTRODUCTION
2. EVIDENCE OF NONLINEAR DYNAMICS IN NORTHERN PEATLANDS
3. STABILIZING AND DESTABILIZING FORCES
4. POTENTIAL FEEDBACKS ON THE CLIMATE SYSTEM
5. RESEARCH NEEDS
Issues Related to Incorporating Northern Peatlands Into Global Climate Models
1. INTRODUCTION
2. DEVELOPMENTS OF COUPLED CARBONCLIMATE MODELS, WITH REPRESENTATION OF ECOSYSTEMS
3. NORTHERN PEATLANDS IN THE COUPLED CLIMATE-CARBON SYSTEM
4. NORTHERN PEATLAND ECOSYSTEM PROPERTIES THAT WILL REQUIRE NEW CLIMATE MODEL DEVELOPMENTS
5. CONCLUDING COMMENTS
Upscaling of Peatland-Atmosphere Fluxes of Methane: Small-Scale Heterogeneity in Process Rates and the Pitfalls of “Bucket-and-Slab” Models
1. THE NEED FOR MODELS OF PEATLAND ATMOSPHERE CARBON EXCHANGES
2. VARIABILITY OF CARBON BALANCE PROCESSES WITHIN PEATLANDS
3. UPSCALING CARBON BALANCE PROCESSES IN PEATLANDS: PROBLEMS WITH THE BUCKET AND SLAB
4. REPLACING THE BUCKET AND SLAB: DEALING WITH HETEROGENEITY
5. AN AGENDA FOR FUTURE RESEARCH
Sensitivity of Northern Peatland Carbon Dynamics to Holocene Climate Change
1. INTRODUCTION
2. CLIMATE CONTROLS OVER DISTRIBUTION, INITIATION, AND EXPANSION OF NORTHERN PEATLANDS
3. SPATIAL AND TEMPORAL PATTERNS OF CARBON ACCUMULATION DURING THE HOLOCENE
4. CONCEPTUAL MODEL OF LONG-TERM CARBON ACCUMULATION IN NORTHERN PEATLANDS
5. PEATLAND RESPONSE TO CLIMATE CHANGE AND IMPLICATIONS FOR THE GLOBAL CARBON CYCLE
6. CONCLUDING REMARKS: OUTSTANDING ISSUES AND FUTURE DIRECTIONS
Direct Human Impacts on the Peatland Carbon Sink
1. INTRODUCTION
2. DIRECT HUMAN INFLUENCE: LAND USE CHANGE
3. LAND USE IMPACTS ON THE CARBON SINK FUNCTIONS
4. How To Reduce Uncertainties in Emission Estimates
5. CLIMATIC IMPACTS OF THE LAND USE CHANGE IN PEATLANDS
6. SUMMARY AND PERSPECTIVES
Section II: Near-Surface Processes of Peatland Carbon Cycling
Northern Peatland Vegetation and the Carbon Cycle: A Remote Sensing Approach
1. INTRODUCTION
2. REMOTE SENSING APPROACH
3. MAPPING NORTHERN PEATLAND LAND COVER WITH REMOTELY SENSED DATA
4. RETRIEVAL OF VEGETATION PHYSIOLOGICAL AND BIOPHYSICAL PROPERTIES FROM REMOTELY SENSED DATA
5. SUMMARY
Plant Litter Decomposition and Nutrient Release in Peatlands
1. INTRODUCTION
2. MASS LOSS
3. NUTRIENT RELEASE
4. CONCLUSIONS
Microbial Community Structure and Carbon Substrate Use in Northern Peatlands
1. INTRODUCTION
2. DIRECT DRIVERS OF MICROBIAL COMMUNITY STRUCTURE AND SUBSTRATE USE
3. FEEDBACKS AND THEIR EFFECTS ON THE FATE OF PEATLAND C STOCKS
4. SYNTHESIS AND RECOMMENDATIONS FOR FURTHER RESEARCH
Partitioning Litter Mass Loss Into Carbon Dioxide and Methane in Peatland Ecosystems
1. INTRODUCTION
2. DEGRADATION OF ORGANIC MATERIAL
3. TERMINAL CARBON MINERALIZATION AND ITS PARTITIONING INTO CO2 AND CH4
4. EVALUATION OF LITERATURE DATA ON THE PARTITIONING OF MINERALIZED CARBON FROM PEAT TO CO2 AND CH4
5. RESEARCH NEEDS
SECTION III: METHANE ACCUMULATION IN, AND LOSS FROM, PEATLANDS
Methane Accumulation and Release From Deep Peat: Measurements, Conceptual Models, and Biogeochemical Significance
1. INTRODUCTION
2. INITIAL MODELS OF METHANE PRODUCTION AND TRANSPORT IN PEATLANDS
3. DETECTION OF GAS BUBBLES IN PEAT
4. MODELS OF METHANE PRODUCTION IN DEEP PEAT
5. MODELS OF GAS TRANSPORT, STORAGE, AND EBULLITION IN PEATLANDS
6. REGIONAL COMPARISONS
7. ROLE OF EBULLITION IN THE GLOBAL METHANE CYCLE
Noninvasive Field-Scale Characterization of Gaseous-Phase Methane Dynamics in Peatlands Using the Ground-Penetrating Radar Method
1. INTRODUCTION
2. GPR METHOD
3. APPLICATIONS AND EXAMPLES
4. DISCUSSION
Methane Dynamics in Peat: Importance of Shallow Peats and a Novel Reduced-Complexity Approach for Modeling Ebullition
1. METHANE LOSS FROM PEATLANDS
2. A NEW APPROACH TO MODELING CH4 BUBBLE BUILDUP AND EBULLITION: UPSIDE DOWN AVALANCHES
3. RESEARCH NEEDS
THE STABLE CARBON ISOTOPE COMPOSITION OF METHANE PRODUCED AND EMITTED FROM NORTHERN PEATLANDS
1. INTRODUCTION
2. METHODS
3. RESULTS AND DISCUSSION
4. SUMMARY AND FUTURE WORK
Laboratory Investigations of Methane Buildup in, and Release From, Shallow Peats
1. INTRODUCTION
2. SHALLOW PEAT MONOLITHS
3. GAS EXCHANGE DYNAMICS
4. PEAT GAS PROFILE
5. MANIPULATION EXPERIMENTS
6. THE 14C IMPULSE LABELING
7. CONCLUSION
Physical Controls on Ebullition Losses of Methane From Peatlands
1. INTRODUCTION
2. IMPORTANCE OF GAS-PHASE CH4 STORAGE IN RELATION TO PHYSICAL CONTROL ON EBULLITION
3. AN APPROACH FOR ASSESSING EFFECTS OF CHANGING TEMPERATURE, WATER TABLE, AND ATMOSPHERIC PRESSURE ON THE GAS VOLUME AND BUOYANCY
4. PHYSICAL FACTORS CONTROLLING EBULLITION
5. FUTURE RESEARCH NEEDS
Section IV: Water and Dissolved Carbon Transfers Within and From Peatlands
DISSOLVED ORGANIC CARBON PRODUCTION AND TRANSPORT IN CANADIAN PEATLANDS
1. INTRODUCTION
2. CONTROLS ON DOC PRODUCTION
3. DOC TRANSPORT WITHIN AND EXPORT FROM CATCHMENTS
4. EFFECT OF DISTURBANCE ON DOC
5. SIGNIFICANCE OF DOC
Hydrological Controls on Dissolved Organic Carbon Production and Release From UK Peatlands
1. INTRODUCTION
2. DOC EXPORT
3. HYDROLOGICAL CONTROLS UPON DOC PRODUCTION
4. TREND GENERATION MECHANISMS
5. MANAGEMENT OPTIONS TO REDUCE DOC LOSSES FROM PEATLANDS
6. CONCLUSIONS
The Role of Natural Soil Pipes in Water and Carbon Transfer in and From Peatlands
1. INTRODUCTION
2. PEAT PIPE FORMATION AND DISTRIBUTION
3. PIPE FLOW
4. PIPE CARBON EXPORT
5. GEOCHEMICAL INTERACTIONS WITHIN PIPE NETWORKS
6. NEXT STEPS
Improving Conceptual Models of Water and Carbon Transfer Through Peat
1. INTRODUCTION
2. GLACIAL LAKE AGASSIZ PEATLANDS
3. CARBON CYCLING
4. CHALLENGES
5. SUMMARY AND CONCLUSIONS
Water Relations in Cutover Peatlands
1. INTRODUCTION
2. ECOHYDROLOGICAL FUNCTIONS IN UNDISTURBED PEATLANDS
3. NATURE AND EXTENT OF HYDROLOGICAL CHANGE EXPERIENCED BY CUTOVER PEATLANDS
4. POSTHARVEST CHANGES
5. HYDROLOGICAL PROCESSES RELATED TO SYSTEM RESTORATION
6. SUMMARY AND CONCLUSION
The Influence of Permeable Mineral Lenses on Peatland Hydrology
1. INTRODUCTION
2. METHODOLOGY
3. RESULTS AND DISCUSSION
4. CONCLUSION
Index
Geophysical Monograph Series
Published under the aegis of the AGU Books Board
Kenneth R. Minschwaner, Chair; Gray E. Bebout, Joseph E. Borovsky, Kenneth H. Brink, Ralf R. Haese, Robert B. Jackson, W. Berry Lyons, Thomas Nicholson, Andrew Nyblade, Nancy N. Rabalais, A. Surjalal Sharma, Darrell Strobel, Chunzai Wang, and Paul David Williams, members.
Library of Congress Cataloging-in-Publication Data
Carbon cycling in northern peatlands / Andrew J. Baird ... [et al.].
p. cm. — (Geophysical monograph ; 184)
Includes bibliographical references and index.
ISBN 978-0-87590-449-8
1. Carbon cycle (Biogeochemistry)—Northern Hemisphere. 2. Peatlands—Environmental aspects—Northern Hemisphere.
3. Carbon sequestration—Northern Hemisphere. 4. Greenhouse gases—Northern Hemisphere. I. Baird, Andrew J., 1969-
QH344.C385 2009
577’.144—dc22
2009024462
ISBN: 978-0-87590-449-8
ISSN: 0065-8448
Cover Photo: Open pools in the central unit of Caribou Bog, Orono, Maine. Modified from R. B. Davis original (taken June 2006).
Copyright 2009 by the American Geophysical Union
2000 Florida Avenue, N.W.
Washington, DC 20009
Figures, tables and short excerpts may be reprinted in scientific books and journals if the source is properly cited.
Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by the American Geophysical Union for libraries and other users registered with the Copyright Clearance Center (CCC) Transactional Reporting Service, provided that the base fee of $1.50 per copy plus $0.35 per page is paid directly to CCC, 222 Rosewood Dr., Danvers, MA 01923. 0065-8448/09/$01.50+0.35.
This consent does not extend to other kinds of copying, such as copying for creating new collective works or for resale. The reproduction of multiple copies and the use of full articles or the use of extracts, including figures and tables, for commercial purposes requires permission from the American Geophysical Union.
PREFACE
Even though they cover only between 2 and 3% of its land mass, peatlands are a major component of the Earth’s carbon cycle, containing about one third of the carbon in the pedosphere. These large carbon stores remove carbon from, and release it to, adjacent systems (most significantly the atmosphere) in a complex cycle. Although large peatlands are found in the tropics, this monograph focuses on recent developments in our understanding of carbon dynamics in northern peatlands, that is, those peatlands that occur at latitudes higher than 45°N. We focus exclusively on northern peatlands because of their significance in terms of surface areal extent and their importance as carbon stores. Northern peatlands are also most likely to be affected by climate change and warming.
Peatlands science cuts across disciplines, and it can be difficult for peatland researchers in one discipline to find advances made by peatland scientists in other disciplines. Recognition of this problem was partly behind our decision to propose the monograph; in other words, we wanted to produce a collection of papers that brought together the state of knowledge on peatland science. Obviously, a monograph that considered all areas of peatland science would have been a huge, if not impossible, undertaking. We chose to focus on carbon cycling and climate, and we did so for two reasons. We felt that much previous work has looked at peatlands primarily as archives of climate change, with less emphasis on the processes that control how a peatland responds to variations in climate and how it may itself influence climate. We were also aware of the need to include peatlands in climate models and the need to communicate current understanding of the role of peatlands in the global carbon cycle to a larger audience, especially, but not exclusively, climate modelers.
Many individuals helped with the production of this monograph. Each chapter was independently reviewed, and we are indebted to those academics who undertook reviews, and in some cases rereviews, on very tight deadlines. Finally, this effort was inspired by a National Science Foundation funded “Peatlands Geophysics Workshop” held at the University of Maine in June 2007. The purpose of the workshop was to bring together peatland scientists from a range of disciplines to consider novel ways of mapping the subsurface structure of peatlands. We hope some of that ambition and novelty is reflected in the current collection.
Lee D. SlaterRutgers, State University of New Jersey
Andrew J. BairdUniversity of Leeds
Lisa R. BelyeaQueen Mary University of London
Xavier ComasFlorida Atlantic University
A. S. ReeveUniversity of Maine
Understanding Carbon Cycling in Northern Peatlands: Recent Developments and Future Prospects
Andrew J. Baird,1 Xavier Comas,2 Lee D. Slater,3 Lisa R. Belyea,4 and A. S. Reeve5
1School of Geography, University of Leeds, Leeds, UK.
2Department of Geosciences, Florida Atlantic University, Boca Raton, Florida, USA.
3Department of Earth and Environmental Sciences, Rutgers, State University of New Jersey, Newark, New Jersey, USA.
4Department of Geography, Queen Mary University of London, London, UK.
5Department of Earth Sciences, University of Maine, Orono, Maine, USA.
Although one of the earliest recorded investigations of peatlands is attributed to King [1685] more than 3 centuries ago [Gorham, 1953], Weber’s [1902] treatise on the Aukštumala Raised Bog in Lithuania is still considered the first comprehensive ecohydrological study of peatlands and the foundation for modern peatland science. Weber’s monograph was pioneering for different reasons: (1) it integrated disciplines such as stratigraphy, hydrology, chemistry, and ecology to describe, classify, and model peatlands forms and their development (e.g., internal processes), and (2) it investigated potential interactions between peatland (mostly hydrology) and changes in climate and sea level (e.g., external forcing). Since then, most peatlands science has focused on the peatland archive (e.g., pollen and macrofossils) for environmental and climate reconstruction over the Holocene, whereas processes controlling the response of peatlands to climate change have tended to be overlooked.
The effect of peatlands on global climate is currently unclear. Peatlands influence climate by sequestering CO2 from the atmosphere and storing it in living and dead biomass. They return some of this CO2 via the decay of plant litter and peat and are the largest natural terrestrial source of atmospheric methane (CH4), which is produced during anaerobic decomposition. Through the Holocene, peatlands have been a persistent sink for atmospheric CO2 and a persistent source of atmospheric CH4. Although CH4 is a much more potent greenhouse gas than CO2, modeling work by Frolking et al. [2006] suggests that peatlands have had a negative radiative forcing (cooling) effect on climate through the Holocene. However, that may change in the coming decades as peatlands respond to climate change. For example, existing peatlands may become net emitters of CO2 as peat warms and rates of decomposition increase, while in areas of permafrost the formation of thaw lakes may lead to higher rates of CH4 loss. On the other hand, new peatlands may develop in areas that are currently tundra and become large sinks of atmospheric CO2 in the next 100–200 years, thus offsetting, at least in part, greater losses of CO2 and CH4 from northern peatlands at lower latitudes. To understand how peatlands affect global climate, we need to represent them as land surface schemes coupled to global climate models (GCMs). Before we can do so, there is a need to understand better how peatlands function as ecohydrological entities. Part of this understanding can come from the paleorecord, but part must come from process and modeling studies at a range of scales. Where possible, peatland models should be process-based and applicable to a range of climatic, geomorphological, and geological settings; that is, they should be transportable.
Scale is an important factor to consider in studies of carbon cycling in northern peatlands for at least two reasons. First, when considering processes operating across the whole peatland or at the regional scale, it is uncertain whether small-scale processes (those at plan scales of about 1 m) can be ignored or simplified adequately. Second, it is becoming clear that little is known about the off-site components of the carbon budget of peatlands; carbon exchanges are not just between peatlands and the atmosphere; dissolved carbon can also be exported from peatlands, where its fate remains uncertain. As part of this second reason, other off-site factors such as topographic setting are also important controls on carbon cycling in individual peatlands. Although this monograph focuses on peatlands occurring across a limited latitudinal range, study sites are included across geographical longitudes spanning three different continents and a total of seven countries.
The chapters consider some of these scale and geographical issues and also how we can improve our understanding of key ecohydrological processes in peatlands and how they affect carbon cycling. The monograph is divided into four sections, and the content of each is discussed briefly below.
Section 1 considers the role of peatlands in the global carbon cycle. It does so from a variety of perspectives. The first chapter considers how peatlands respond to climate change. Many workers still consider peatlands to respond in simple linear ways to climate. However, it is becoming clear that peatlands are complex adaptive systems that do not show simple responses to climate. Sometimes peatlands undergo large ecohydrological changes in the absence of a climate driver, and sometimes they show extreme sensitivity to apparently small climatic changes. If we are to model peatlands adequately, then we need to know the reasons behind such nonlinear behavior. Climate modelers are only now starting to recognize the importance of the peatland carbon store and the need to include peatlands in GCMs. How peatlands should be represented in climate models is the focus of the second and third chapters. The second looks at how peatlands differ from other land surface types and the challenges these differences present when trying to incorporate peatlands into land surface schemes. The third looks at the problem of what it means to ignore small-scale variability when modeling CH4 losses from peatlands and the importance of such small-scale variability when trying to represent peatlands in GCMs. The fourth chapter provides a broad temporal and spatial perspective and uses meta-analyses of data from previous studies to investigate the factors that influenced peatland initiation and carbon accumulation during the Holocene. Finally, there is a chapter on direct human impacts on the peatland carbon store. Much of the current concern is with the indirect human impacts (climate change) on peatlands; therefore, it is useful also to consider how direct impacts such as changes in land use to forestry and agriculture affect carbon cycling.
Section 2 focuses on processes operating at and near the peatland surface, where climate change is likely to have the greatest impact through disturbance (e.g., fire and permafrost thaw) and changes in water table and temperature regimes. One of the unique characteristics of peatland land surfaces is the prevalence of mosses, which have distinct biophysical and biochemical properties compared with vascular plants. The relative abundance of these plant types has a profound effect on carbon cycling because the biochemistry of plant-derived substrates is a key control on rates and pathways of decomposition. The other major control on carbon cycling relates to vertical and horizontal heterogeneity in water table regime, which, in turn, controls oxygenation, the distribution of plants and microbes, and microbial metabolic pathways. The first chapter in this section reviews remote sensing approaches to obtaining land surface data relevant to the carbon cycle, both for generating land surface classifications and for retrieving biophysical properties that can be used to parameterize process-based models. The second chapter considers mass loss and nutrient release from fresh plant litter, noting that changes that affect the relative production of Sphagnum versus vascular plant litter are likely to have feedbacks on carbon and nutrient cycling. The third chapter examines carbon flow from a microbial perspective, identifying the main microbial players, metabolic pathways, and factors controlling substrate use in the oxic, periodically oxic/anoxic, and permanently anoxic zones. The final chapter considers the relative amounts of CO2 and CH4 produced during terminal mineralization of carbon under anoxic conditions, noting the importance of both substrate characteristics and physical factors and making a plea for further studies using consistent methodologies.
Section 3 describes the state of knowledge on CH4 accumulation in, and release from, peatlands by considering both deep and shallow sources of the gas at a wide range of scales. Methane is lost to the atmosphere through three main mechanisms: diffusion through the peat matrix, transport through vascular plants, and ebullition (as bubbles). Until recently, most studies considered the first two mechanisms almost exclusively. However, there is burgeoning interest in the significance and causes of ebullition. Nonsteady or episodic ebullition events have generated particular interest because of the potentially large amounts of CH4 involved. Ebullition fluxes in northern peatlands typically exceed average diffusive fluxes on a per-event basis and often on a seasonal basis as well. Current ebullition estimates are unclear for several reasons: (1) our poor understanding of gas spatial variability related to the heterogeneous nature of the peat matrix; (2) uncertainties related to contrasting models of gas accumulation (e.g., shallow entrapment in poorly decomposed peat versus deep entrapment below confining layers of woody peat); and (3) factors affecting ebullition dynamics, often related to environmental parameters such as soil chemistry, substrate quality, or plant community structure. To further complicate estimates, biogenic gas emissions from wetlands are often related to changes in temperature, atmospheric pressure, and/or water table elevation. The first and second chapters in this section emphasize the importance of CH4 accumulation in deep peats (i.e., >2 m) and describe the use of minimally invasive techniques such as global positioning systems (GPS) and ground-penetrating radar to investigate deep free-phase gas accumulations in peatlands. The third and fourth chapters examine the role of shallow peat soils (i.e., <1 m) as both zones of CH4 production and zones from which CH4 is lost to the atmosphere. The third chapter proposes a new conceptual model for bubble buildup and release in shallow peat soils, while the fourth identifies key zones of enhanced methanogenesis at shallow depths based on carbon isotope composition. The fifth chapter presents an overview of an experimental design that can be used for investigating the accumulation and release of CH4 from shallow peats under controlled laboratory conditions. Finally, the last chapter looks at some of the controls that may induce losses of CH4 gas from peatlands such as atmospheric pressure, peat temperature, and water table position for both deep and shallow peats.
All of the topics presented in this monograph reveal the importance of the physical and chemical processes related to water supply to and movement within peatlands. Section 4 focuses on peatland hydrology and its role in carbon dynamics. Efforts to understand peatland hydrological processes typically focus on (1) saturation state and water table position and (2) rates and directions of water movement. Water in peatlands isolates organic matter from the atmosphere, altering the redox state and slowing the oxidation of organic matter while creating an environment favorable for methane production. This mixture of organic matter and water indirectly results in high concentrations of dissolved organic carbon (DOC) in peat pore waters. The production of DOC within and export from peatlands is discussed in the first two chapters of section 4. The rate and direction of surface and groundwater flow within a peatland regulate the export of DOC from the peatland and influence the supply of nutrients to it. The third chapter describes the hydrological and hydrochemical importance of natural soil pipes in the peatlands in which they occur. Chapter five focuses on the hydrodynamics of the unsaturated zone in harvested peatlands. Chapter six discusses the role of subsurface heterogeneity on groundwater flow patterns. There are a variety of feedback mechanisms between the hydrology of a peatland and associated carbon dynamics that complicate this relationship. The relationship between hydrology and biogenic gas dynamics is one of these feedback systems and is discussed in chapter four. While there are many similarities among the peatland systems discussed in this section, it is important to note the differences between individual systems and to use caution when generalizing processes observed in one peatland to another.
REFERENCES
Frolking, S., N. Roulet, and J. Fuglestvedt (2006), How northern peatlands influence the Earth’s radiative budget: Sustained methane emission versus sustained carbon sequestration, J. Geophys. Res., 111, G01008, doi:10.1029/2005JG000091.
Gorham, E. (1953), Some early ideas concerning the nature, origin, and development of peat lands, J. Ecol.,41(2), 257–274.
King, W. (1685), On the bogs and loughs of Ireland, Philos. Trans. R. Soc. London, 15, 948–960.
Weber, C. A. (1902), Über die Vegetation und Entstehung des Hochmoors von Augstumal im Memeldelta mit vergleichenden Ausblicken auf andere Hochmoore der Erde, Paul Parey, Berlin.
A. J. Baird, School of Geography, University of Leeds, Leeds LS2 9JT, UK.
L. R. Belyea, Department of Geography, Queen Mary University of London, Mile End Road, London E1 4NS, UK.
X. Comas, Department of Geosciences, Florida Atlantic University, Boca Raton, FL 33431, USA.
A. S. Reeve, Department of Earth Sciences, University of Maine, Orono, ME 04469, USA.
L. D. Slater, Department of Earth and Environmental Sciences, Rutgers, State University of New Jersey, Newark, NJ 07102, USA. ([email protected])
Section I: Large-Scale Peatland Dynamics and Carbon Cycling
Nonlinear Dynamics of Peatlands and Potential Feedbacks on the Climate System
Lisa R. Belyea
Issues Related to Incorporating Northern Peatlands Into Global Climate Models
Steve Frolking, Nigel Roulet, and David Lawrence
Upscaling of Peatland-Atmosphere Fluxes of Methane: Small-Scale Heterogeneity in Process Rates and the Pitfalls of “Bucket-and-Slab” Models
A. J. Baird, L. R. Belyea, and P. J. Morris
Sensitivity of Northern Peatland Carbon Dynamics to Holocene Climate Change
Zicheng Yu, David W. Beilman, and Miriam C. Jones
Direct Human Impacts on the Peatland Carbon Sink
Jukka Laine, Kari Minkkinen, and Carl Trettin
Nonlinear Dynamics of Peatlands and Potential Feedbacks on the Climate System
Lisa R. Belyea
Department of Geography, Queen Mary University of London, London, UK
Peatlands have potential for strong feedback on the global climate system, but their response to future climate change is highly uncertain. In this chapter, I review a range of evidence demonstrating that peatland dynamics are nonlinear. Rather than gradual change that converges on a single dominant pathway and matches the frequency of external forcing, peatlands show (1) sensitivity to initial conditions and divergence onto multiple pathways of development, (2) long periods of little change, punctuated by abrupt transitions of state even under weak or steady environmental forcing, and (3) responses to external forcing at unexpected frequencies. Nonlinear systems exhibit persistence when stabilizing forces (i.e., negative feedback mechanisms) dominate and undergo rapid transformation when destabilizing forces (i.e., positive feedback mechanisms) dominate. In peatlands, stabilizing and destabilizing forces result from interactions among hydrological processes, organic matter dynamics, and energy exchanges. The depth dependence of peat hydraulic conductivity tends to stabilize hydrological conditions, whereas local flow networks may amplify water losses when vascular plant transpiration is high. Peat formation rate is generally constrained by water storage change but occasionally can trigger a rapid increase or decrease in thickness of the unsaturated zone. Regionally, increases in evapotranspiration may be counteracted by recycling and precipitation of evaporated water over peatlands, whereas contrasts in albedo and energy partitioning across peatlands and surrounding forests may promote rapid spring thaw. In order to predict feedbacks on the climate system, it will be essential to reduce the complexity of peatlands by identifying the key variables and interactions that control nonlinear behavior.
1. INTRODUCTION
Throughout the Holocene, the radiative forcing function of peatlands has shifted from a net warming to a net cooling [Smithet al., 2004; Frolkinget al., 2006], but the future impact of peatlands on a changing climate system is highly uncertain. Of central concern is the vulnerability of the large peatland carbon pool to processes that might release CO2 and CH4 to the atmosphere, thereby amplifying human-induced changes to atmospheric chemistry. Peatlands are complex systems [Belyea and Baird, 2006], and as I will show in this chapter, their dynamics are often nonlinear. Abrupt, step-like changes in peatland structure (e.g., the distribution of vegetated and nonvegetated land surface types) and function (e.g., hydrological processes, organic matter dynamics, and energy exchanges) may be linked only weakly to climate forcing [e.g., Belyea and Malmer, 2004; Yu, 2006a]. If these nonlinear changes involve the peatland carbon pool, they may have profound effects on the climate system.
A key question, then, is “What controls nonlinear dynamics in northern peatlands and what are the potential feedbacks on the climate system?” I will approach this question by reviewing some of the evidence for nonlinear behavior in peatland ecosystems, identifying some of the stabilizing (negative feedback) and destabilizing (positive feedback) forces that operate within them, and discussing possible biogeochemical and biogeophysical feedbacks on the global climate system for two “what if?” scenarios of future environmental change.
2. EVIDENCE OF NONLINEAR DYNAMICS IN NORTHERN PEATLANDS
The nonlinear behavior of some complex systems is characterized by long periods of stasis punctuated by occasional, abrupt shifts to alternative regimes, which differ in fundamental structure and processes from the previous regime. Such threshold (step-like) dynamics occur in many physical and biological systems and can be classified into three basic types [Andersen et al., 2009]: (1) “driver threshold,” the system state responds linearly to an environmental driver, which itself undergoes a step-like change; (2) “state threshold,” the system state responds in a step-like way after a slowly varying environmental driver exceeds a threshold value; (3) “driver-state hysteresis,” the threshold value for step-like response of the system state differs, depending on whether the environmental driver is increasing or decreasing. The last two types are of particular concern in the context of climate change and the carbon cycle, because abrupt shifts in fundamental structure and function can occur unexpectedly under weak external forcing.
Paleorecords and long-term instrumental records can provide evidence of three types of behavior that are characteristic of nonlinear systems [Rial et al., 2004]: (1) Nonlinear systems are highly sensitive to initial conditions and show divergence onto multiple pathways rather than convergence onto a single dominant pathway of evolution. (2) Even under weak or steady external forcing, nonlinear systems show rapid, abrupt transitions of state rather than slow, gradual changes proportional to external forcing. (3) Nonlinear systems respond to oscillations at unexpected frequencies rather than matching the frequency of the external forcing. All of the following examples from peatlands suggest at least one of these distinctively nonlinear behaviors.
2.1. Peatland Initiation and Expansion
Peatlands initiate by three mechanisms [Rydin and Jeglum, 2006]: terrestrialization by infilling of a lake or pond, primary peat formation on newly exposed mineral soil, and paludification by “swamping” of mineral soil that was previously covered by forest, grassland, heathland, or tundra. In previously glaciated landscapes, peat formation often begins by terrestrialization of small depressions, and these initially small peatlands subsequently expand across the landscape by primary peat formation or paludification [Korhola et al., 1996; Anderson et al., 2003]. Very few studies have examined the processes directly involved in lateral expansion by primary peat formation and paludification, but it seems clear that local positive feedback must be involved. Specifically, the groundwater mound must grow with the peat deposit, with the result that surrounding mineral soils become “swamped.” A range of pedogenic processes that decrease soil permeability may also be involved [Rydin and Jeglum, 2006]. The increase in wetness of the mineral soil allows establishment of peat-forming plants such as Sphagnum, triggering a further network of positive feedbacks, involving water retention, changes in soil water chemistry, and organic matter accumulation [van Breemen, 1995]. The switch from mineral soil to peat soil can occur within decades [Hulme, 1994], and the rate of lateral expansion can reach 8 m a−1 [Rydin and Jeglum, 2006]. The rapidity and abruptness with which peatland initiation and expansion can occur suggests a threshold response from one regime (mineral soil) to another (peat). At a continental scale, postglacial peatland expansion is described very well by a simple sigmoidal model [Gorham et al., 2007]. The phase of “explosive” peatland expansion (i.e., the near-vertical part of the sigmoidal curve) occurred much earlier in Siberia than in North America [Smith et al., 2004] and was accompanied by similarly abrupt increases in storage of organic carbon and emission of CH4 to the atmosphere [Gajewski et al., 2001; Smith et al., 2004]. The difference in timing of the explosive phase may indicate a disproportionate response to changes in climate, although this hypothesis should be tested explicitly.
2.2. Successional Dynamics
Once initiated, peatlands undergo a wide variety of vegetation changes at scales ranging from individual microforms (“microsuccession,” occurring at length scales of 100–101 m, e.g., hollow to hummock transition) to the whole peatland (“macrosuccession,” occurring at length scales of 102–103 m, e.g., fen to bog transition). Regional analyses, which compile paleoecological records of sediment cores from many sites, show a complicated network of developmental pathways inferred from transitions in plant macrofossil composition [e.g., Walker, 1970; Aaviksoo et al., 1993; Bunting and Warner, 1998]. In these records, self-replacement of vegetation types is very common; that is, the peatland tends to persist in one vegetation state over many sampling intervals. When transitions do occur, they can follow one of a number of alternative pathways rather than a single dominant pathway. More detailed analyses of single cores suggest that long periods of little or no change are punctuated by occasional, brief episodes of abrupt change, both in terms of vegetation and carbon storage [Belyea and Malmer, 2004; Yu, 2006a]. These step-like transitions can occur even through periods where environmental forcing is gradual or weak [Belyea and Malmer, 2004; Yu, 2006a]. These results support the idea that successional dynamics are nonlinear: a peatland will remain static in one regime for a long period of time despite external forcing and then occasionally undergo rapid transition to one of a number of alternative regimes, with the change in state disproportionately large compared with any change in the strength of environmental forcing.
2.3. Ombrotrophication (Fen-Bog Transition)
Ombrotrophication is a particular type of macrosuccession that occurs when the surface of a fen becomes isolated from minerotrophic water and undergoes transition to poor fen or bog, triggering a vegetation switch to dominance by Sphagnum mosses and marked changes in soil water chemistry. In large part, isolation from minerotrophic water occurs because of height growth of the peat deposit. The timing of fen-bog transition, therefore, is highly dependent on local factors such as time of peat initiation, local rate of peat growth, and local topography of the mineral substrate [Bunting and Warner, 1998; Anderson et al., 2003]. The process is driven largely by Sphagnum “engineering” the physical and chemical environment [van Breemen, 1995], and the new, ombrotrophic state (i.e., bog or poor fen) will persist if Sphagnum achieves a threshold abundance. Detailed paleoecological analyses of peat cores show that ombrotrophication occurs rapidly on a time scale of decades [Kuhry et al., 1993]. In some cases, the records show one or two previous, failed attempts before the successful transition occurs (P. Kuhry, personal communication, 2005). Once the peatland has reached a critical stage of development, a period of drier climate may help to trigger the transition [Hughes, 2000; Anderson et al., 2003]. These observations suggest a steplike shift from one state (fen) to another (bog or poor fen). Although regional climate change may help push the system up and over the step, the transition is driven mainly by internal, positive feedbacks involving the effects of Sphagnum on the physicochemical environment. Given that fens and bogs differ markedly in rates of CH4 emission [Bubier et al., 2005] and carbon storage [Yu, 2006a], ombrotrophication is almost certain to result in a step-like change in carbon cycling.
2.4. Permafrost Formation and Thaw
In the discontinuous permafrost zone, peat plateau landforms undergo a cyclical succession, switching between permafrost and nonpermafrost regimes [Zoltai, 1993; Camill and Clark, 2000]. In the nonpermafrost state, peat accumulates differentially, leading to formation of hummocks, which allow colonization by trees [Camill and Clark, 2000]. In winter, the (evergreen) canopy reduces snow depth [Camill and Clark, 2000], and the frozen peat conducts heat rapidly to the atmosphere [Zoltai, 1993]; in summer, the tree canopy shades the peat surface [Camill and Clark, 2000] and the dry surface peat acts to insulate deeper layers [Williams, 1968; Zoltai, 1993]. As a result, temperature within the hummocks is reduced and permafrost begins to form. The permafrost features gradually expand and coalesce, forming a densely forested permafrost plateau. Disturbance of forest cover on the permafrost plateau by fire or windthrow initiates permafrost thaw in isolated “collapse scars” [Camill and Clark, 2000]. Without the sheltering effect of the trees, the peat within the scar thaws, and thermokarst features such as wet lawns or small lakes form as the surface collapses. Over time, differential peat accumulation can lead to the formation of hummocks within the collapse scar. The unfrozen area expands in size until the thawing edge of the collapse scar is stabilized, either by shade of fringing trees or by the insulating effects of Sphagnum hummocks [Camill and Clark, 2000]. The dominance of local factors (tree size and density and differential peat microtopography) in this cycle suggests that responses to climate change will be nonlinear. In continental Canada, the southern limit of the discontinuous permafrost zone has shifted northward since the Little Ice Age, but this shift has lagged increases in temperature, leaving behind relict permafrost in regions where it could not form today [Turetsky et al., 2007]. Over the past 50 years, the rate of permafrost thaw has increased, with the rate increasing more quickly in southern regions than in those farther north, possibly linked to greater increases in winter and spring temperatures [Camill, 2005]. In Siberia over the past 30 years, thaw lake area has increased by about 12% in the continuous permafrost zone and decreased by about the same percentage in the discontinuous permafrost zone [Smith et al., 2005], suggesting that the link of climate change with permafrost degradation and lake drainage is mediated by regional factors. These temporal lags and differential responses in permafrost distribution to climate forcing are suggestive of nonlinear dynamics. Carbon dynamics are also likely to be nonlinear and transient, because collapse scars have higher rates of CH4 emission but also higher rates of peat accumulation than do permafrost landforms [Camill et al., 2001; Turetsky et al., 2007].
2.5. Water Table and CO2 Flux Dynamics
Long-term instrumental records of climate variables, water table height, and atmospheric carbon flux have been collected at a number of peatlands, and these high-resolution records can be tested for concordance of climate and ecosystem variables. Yu [2006b] analyzed frequency variation in such records for three continental fens and one maritime bog, using power density spectra. His analyses showed that water table behavior is highly dependent on history and past events, suggesting a nonlinear response to climate and dominance of rare (low frequency) events. Precipitation records were not included in the analysis, but temperature (and relative humidity for the maritime bog site) showed a spectrum very different from that of water table height. These results support the idea that the peatland is self-regulating with respect to water table height [Ivanov, 1981; Ingram, 1983] and that it responds to external forcing at unexpected frequencies. Yu’s [2006b] analyses of CO2 flux (for the three fen sites only) showed a more complicated spectrum, showing concordance with water table height at time scales greater than 1 month and less than 1 day but concordance with temperature at intermediate time scales. In its role as a control on CO2 flux, water table height is likely to be an indirect indicator of soil moisture at subdaily time scales and of microform type at time scales greater than 1 day, whereas temperature may be a direct control as well as an indirect indicator of vascular leaf area at intermediate times scales [cf. Riutta et al., 2007; Laine et al., 2007]. In any case, the analyses suggest that peatland CO2 flux behaves in a nonlinear way, with dominant controls switching across time scales of observation.
3. STABILIZING AND DESTABILIZING FORCES
The empirical examples presented in section 2 provide qualitative evidence suggestive of nonlinear behavior: peatland dynamics seem to be characterized by long periods of stasis punctuated by occasional, abrupt shifts to alternative regimes. Further investigation of time series data using appropriate quantitative techniques is required to verify the existence of regime shifts in peatlands and to explore how they are related to possible drivers [see Andersen et al., 2009].
What forces and mechanisms might underlie threshold dynamics and regime shifts in peatlands? As in other complex systems, negative feedback mechanisms that act to damp fluctuations in the system compete with positive feedback mechanisms that act to amplify initially small fluctuations [Holling and Gunderson, 2002; Rial et al., 2004]. Stabilizing forces dominate most of the time, but occasionally, the balance will shift in favor of destabilizing forces. The system then undergoes abrupt change until stabilizing forces once again dominate. External forcing may sometimes tip the balance, but perturbations are either damped or amplified primarily by feedback mechanisms internal to the system.
As in other complex systems [Werner, 1999; Holling and Gunderson, 2002; Rial et al., 2004], forces that act to damp or amplify fluctuations in peatlands are linked to processes interacting across spatial and temporal scales [Belyea and Baird, 2006]. A fundamental concept borrowed from physics is the “enslaving principle” [Haken, 2004]: when stabilizing forces dominate, the dynamics of components that have the potential to change rapidly (“slaves”) are entrained by the dynamics of a few components that always change more slowly (“masters”). In a dune system, for example, the long-term movement of sand grains (slaves) is determined by migration of the sand dune (master), which itself is a dynamic pattern emerging from the collective behavior of many individual sand grains [Werner, 1999]. The number of degrees of freedom of the system (i.e., its entropy) is greatly reduced as the fast variables (slaves) almost instantaneously come into a slowly varying “quasi steady state” dictated by the slow variables (masters) [Rinaldi and Scheffer, 2000]. In some systems, destabilizing forces occasionally dominate when the fast components “revolt” [Holling and Gunderson, 2002] and temporarily escape the constraints imposed by the slow components. At certain values of the slow variable (“bifurcation points”), the system can undergo a qualitative change of behavior in which the fast variables leave the quasi steady state and begin to vary much more rapidly [Rinaldi and Scheffer, 2000]. Depending on the system, the resulting bifurcation can take one of several different forms, including catastrophic shift to a new state. The classic ecological example is the “driver-state hysteresis” type of regime shift between clear and turbid conditions in shallow lakes, which occurs as external nutrient loading slowly increases or decreases across two distinct thresholds [Scheffer et al., 1993].
In studies of ecosystems, theoretical investigation of threshold dynamics and regime shifts focuses on concepts of resilience and alternative stable states [Holling and Gunderson, 2002]. “Ecosystem resilience” refers to the capacity of an ecosystem to absorb perturbation through changes in function rather than fundamental structure: the more resilient the ecosystem, the larger the disturbance it can absorb without change in fundamental structure or loss of key processes [Holling, 1973]. When the resilience of an ecosystem is exceeded, the system undergoes rapid transformation to another (alternative) state. In the case of peatlands, these transformations may occur across biomes (e.g., switch from forest to peatland or from permafrost to nonpermafrost landform) or within the ecosystem (e.g., switch from fen to bog or from homogeneous vegetation to a two-phase mosaic of hummocks and hollows). Holling and Gunderson [2002] point out that resilience is not fixed in time or space but operates at a range of scales and may change as the ecosystem develops. For example, a switch from homogeneous vegetation to a two-phase mosaic of hummocks and hollows (transformation at within-ecosystem scale) may allow the peatland to persist under a changing climate (persistence at biome scale).
The most pressing challenges to understanding nonlinear dynamics in ecosystems are to identify, first, the slow (master) variables, which stabilize the system, and, second, the subset of fast (slave) variables, which have the potential to destabilize the system at critical points. In section 3.1, I examine some of the stabilizing and destabilizing forces that arise in peatland ecosystems through hydrological processes, organic matter dynamics, and surface energy exchanges and attempt to identify some of the key slow and fast variables involved. This analysis is purely qualitative, somewhat speculative, and certainly biased by my experience and understanding of northern peatlands.
3.1. Hydrological Processes
Depth-integrated rates of water flow through peat are highly dependent on fluctuations in the water table, and this dependence provides a potential mechanism for stabilization of peatland hydrology on daily and seasonal time scales. Saturated hydraulic conductivity (Ksat) and specific yield generally decrease nonlinearly with depth, because pore spaces become smaller as peat decays and compresses [Boelter, 1969; Ingram, 1983]. As the water table rises closer to the peatland surface, transmissivity (i.e., Ksat integrated through the entire thickness of saturated peat) increases and water is discharged more rapidly. At the same time, more water can be stored per unit rise in water table because the pore volume is greater near the surface and because the total peat volume can expand because of the elastic nature of near-surface peat [Kellner and Halldin, 2002]. Consequently, excess water from a rainfall event is either stored within the peat or rapidly discharged through near-surface peat, minimizing the occurrence of saturation-excess overland flow. Conversely, as the water table falls during a period without rain, transmissivity decreases and water is discharged more slowly. Total peat volume may also contract as water is lost [Kellnerand Halldin, 2002]. The net effect is that the peatland is self-regulating with respect to its water table, which tends to be maintained most of the time within a narrow range of elevations [Ingram, 1983].
The influence of evaporation and transpiration on peatland dynamics is complicated, because the two processes have different controls and their rates depend on a multitude of factors, including the relative abundances of mosses and vascular plants, water table depth, atmospheric conditions, and local advection [Kellner, 2001]. Rates of evaporation from mosses are high when the water table is close to the peat surface [Nichols and Brown, 1980] but may (or may not) decline as water table depth increases [Ingram, 1983; Lafleur et al., 2005]. During prolonged drought, the mosses may lose their pigments and turn white [Gerdol et al., 1996], reducing net radiation by the increase in albedo. A surface crust may also form as the moss capitula dry out, “sealing” the surface and increasing resistance to evaporation. In contrast to moss evaporation, transpiration by vascular plants is unaffected by drought as long as the water table remains within the rooting zone [Lafleur et al., 2005]. When the water table is close to the surface, however, transpiration may decline [Lafleur et al., 2005], presumably because of plant stress caused by oxygen deficiency. Other than during extreme drought when the water table drops below the rooting zone causing decreases in both evaporation and transpiration, losses of water to the atmosphere may be largely decoupled from peatland hydrological conditions and therefore offer no direct mechanism for stabilization or destabilization of water table dynamics. Spatial differences in rate, however, may contribute to an indirect mechanism that involves local redistribution of water.
Spatial heterogeneity in hydraulic properties and water flux rates can promote local flow networks that redistribute water (and possibly carbon and nutrients) among microforms. Limited data suggest that transmissivity and storativity are higher in hollows than in hummocks because the water table in a hollow is situated in more porous, less decomposed peat [Ivanov, 1981; Kellner and Halldin, 2002]. After a rainfall event, the absolute elevation of the water table may be slightly higher in hummocks than in hollows, such that hummocks or ridges can act as local watersheds, with a tendency for water to flow down a hydraulic gradient to adjacent hollows or pools [Price and Maloney, 1994]. During wet periods, flooded pools may expand laterally onto adjacent microforms and coalesce, storing excess water and rapidly discharging it by overland flow [Quinton and Roulet, 1998]. During drier periods, the pools may shrink and become isolated from one another, so that no overland flow occurs and water is lost by evaporation over a smaller area of open water [Quinton and Roulet, 1998]. These mechanisms are stabilizing, because pool expansion and contraction modulate water losses in response to fluctuating inputs. In some situations, however, densely vegetated ridges may act as water pumps, with high rates of transpiration drawing water (and nutrients) from surrounding hollows [Eppinga et al., 2008]. The flux of nutrients may promote further growth of vascular plants on the ridges, which, in turn, may promote higher rates of advection of water and nutrients from surrounding hollows [Rietkerk et al., 2004; Eppinga et al., 2008]. This mechanism is destabilizing in the long term, because it amplifies initially small differences in water loss. Whether the net redistribution of water is from ridge to pool (i.e., precipitation-driven flow, stabilizing) or from pool to ridge (i.e., transpiration-driven flow, destabilizing) may change seasonally with weather conditions [Eppinga et al., 2008] and may also vary regionally with climate wetness (M. B. Eppinga et al., Resource contrast in patterned peatlands increases along an evapotranspiration gradient, manuscript in preparation, 2009).
Local flow networks provide a mechanism for shifting water losses between evapotranspiration and runoff, and the way in which these networks function may depend at least partially on regional climate, specifically the excess of precipitation, P, over evapotranspiration, E. In wetter climates (high P − E), the amount of water lost by runoff must be greater than in drier climates (low P −E). Since hydraulic gradient is controlled mainly by topography of the peatland surface, the increased discharge in wetter climates must be accommodated by an area-averaged increase in transmissivity. This increase could be accomplished through either a general decrease in thickness of the unsaturated zone (i.e., water table positioned within peat of high Ksat) or an expansion of high-transmissivity microforms, such as hollows and pools [Ivanov, 1981]. Evidence from stratigraphic analyses of large peat exposures provide support for the idea that peatlands respond to decadal changes in climate wetness through expansion and contraction of microforms of contrasting transmissivity [Barber, 1981]. In most cases, these responses are likely to damp fluctuations in climate wetness: when the climate is wetter, hollows expand and allow greater runoff losses; when the climate is drier, hummocks expand and reduce runoff losses. This stabilizing effect may be weakened or reversed in continental climates, because the expansion of hummocks would tend to increase losses by the mechanism of “transpiration-driven flow” mentioned above (Eppinga et al., manuscript in preparation, 2009).
The formation of surface water bodies on peatlands (ponds, pools, and lakes) may become destabilizing in the long term. Once pool initiation begins, it seems to proceed in one direction: individual pools expand laterally and coalesce to cover a greater proportion of the peatland, and the pool complex as a whole extends outward, with new pools initiating on the periphery of the complex [Foster et al., 1988; Belyea and Lancaster, 2002]. Catastrophic drainage may occur once the pools have expanded to such an extent that the intervening ridges are structurally compromised [Foster et al., 1988]. Pool initiation and expansion, therefore, may be stabilizing at the within-ecosystem scale (by regulating water losses) but destabilizing at the biome scale (by eventually leading to peatland degradation). Pool formation is also likely to have a strong effect of decreasing CO2 sequestration and increasing CH4 emission [Hamilton et al., 1994].
3.2. Organic Matter Dynamics
At the scale of individual microforms (1–10 m), height of the peat surface relative to the upper boundary of the saturated zone (hereinafter referred to as “thickness of the unsaturated zone”) is a key variable controlling plant and microbial community structure, peat formation, and carbon flux [Sjörs, 1990; Waddington and Roulet, 1996; Alm et al., 1997; Belyea and Clymo, 2001; Laine et al., 2007]. Mechanistically, this control is related to differences in production of litter by vegetation (i.e., species composition, especially the relative abundance of vascular plants) and differences in the mass of litter/peat exposed to aerobic decomposition (i.e., thickness of the “acrotelm”) [Belyea and Clymo, 2001]. Although the upper boundary of the saturated zone is relatively homogeneous across a scale of meters, the peat surface is highly heterogeneous across this scale, with thickness of the unsaturated zone varying tremendously, from high and dry hummocks to low and wet hollows or open water pools. In defining the upper boundary of the saturated zone, daily and seasonal fluctuations in the water table (and the fact that some of the pore space in the “saturated” zone may be filled by biogenic gas bubbles rather than water [e.g., Strack et al., 2005]) are ignored. Water storage change, therefore, refers to long-term rise or fall of the lowest water table over years or decades due, for example, to growth of the groundwater mound. As such, storage change is a slow variable, which, most of the time, controls the fast variables of peat formation and carbon flux (i.e., CH4 emission and CO2 exchange).
Changes in thickness of the unsaturated zone depend on the relative rates of two processes: formation of new litter/peat and water storage change in the saturated zone [Belyea and Clymo, 2001]. If new litter/peat is added at the peatland surface at the same rate as water storage increases, then thickness of the unsaturated zone remains in a steady state (Figure 1a). If litter/peat forms more quickly or if storage increases more slowly, then the microform increases in height, e.g., from lawn to hummock (Figure 1b). Conversely, if litter/peat forms more slowly or if storage increases more quickly, then the microform decreases in height (Figure 1c), e.g., from lawn to hollow. Surface wetness, as indicated by thickness of the unsaturated zone, is not simply a function of the climatic water balance, as is implicitly assumed in many climate reconstructions based on proxies such as plant macrofossils, testate amoebae, or peat humification [e.g., Blackford, 2000]. If storage change is held constant, then change in thickness of the unsaturated zone depends entirely on the rate of peat formation. Nonlinear responses arising from negative and positive feedback mechanisms must be considered.
Figure 1. Changes in thickness of the unsaturated zone, resulting from the relative rates of new litter/peat formation and water storage change. (a) Steady state. Peat formation equals storage change. (b) Thickening of the unsaturated zone. Peat formation increases relative to storage change (time 2a), or storage change decreases relative to peat formation (time 2b). The alternative scenarios result in the same thickness of the unsaturated zone but different absolute elevations of the peat surface. (c) Thinning of the unsaturated zone. Peat formation decreases relative to storage change (time 2a), or storage change increases relative to peat formation (time 2b). The alternative scenarios result in the same thickness of the unsaturated zone but different absolute elevations of the peat surface.
The negative and positive feedbacks become evident if we impose small perturbations on an otherwise constant slow variable, storage change, and examine the dynamics of the system when the fast variable, peat formation, is allowed to vary in response to these small perturbations [Rinaldi and Scheffer, 2000]. Empirical studies show that there is a nonlinear relationship between rate of peat formation and thickness of the unsaturated zone, with peat forming more quickly in intermediate microforms (lawns and low hummocks) than in either wetter (hollows and pools) or drier (high hummocks) microforms [Alm et al., 1997; Belyea and Clymo, 2001; Laine et al., 2007]. Since the relationship between rate of peat formation and thickness of the unsaturated zone is humpbacked, there are potentially two points where peat formation equals storage change, one on the rising limb and one on the falling limb of the “hump” [Belyea and Clymo, 2001]. The point on the rising limb is unstable, because small changes in thickness of the unsaturated zone (e.g., due to a run of dry or wet years) will be amplified by positive feedback. The point on the falling limb is stable, because small changes in thickness of the unsaturated zone will be damped by negative feedback. Hence, stabilizing forces operate on the falling limb of the hump, and destabilizing forces operate on the rising limb [Belyea and Clymo, 2001].
The dynamics of the system can be illustrated by a thought experiment (Figure 2). In a run of wet years, storage change will increase slightly, leading to a slight decrease in thickness of the unsaturated zone as the upper boundary of the saturated zone rises. If the system is on the rising limb, the rate of peat formation will decrease, and, in turn, thickness of the unsaturated zone will decrease even further, amplifying the initial perturbation (Figure 2a). Conversely, if the system is on the falling limb, the rate of peat formation will increase, and, in turn, thickness of the unsaturated zone will increase, counteracting the initial perturbation (Figure 2b). Complementary responses would be observed with the opposite sort of perturbation. In a run of dry years, storage change will decrease slightly, leading to a slight increase in thickness of the unsaturated zone as the upper boundary of the saturated layer falls. If the system is on the rising limb, the rate of peat formation will increase, and, in turn, thickness of the unsaturated zone will increase even further, amplifying the initial perturbation (Figure 2a). Conversely, if the system is on the falling limb, the rate of peat formation will decrease, and, in turn, thickness of the unsaturated zone will decrease, counteracting the initial perturbation (Figure 2b). Hence, a high hummock (falling limb, stable point) is resistant to perturbations and will tend to persist, whereas a hollow (rising limb, unstable point) is sensitive to perturbations and may undergo rapid transition to either a hummock or a pool. At the unstable point, the trajectory of change may vary stochastically across space; the system may undergo a bifurcation or “divergent succession” [Sjörs, 1990], such that some hollows become hummocks, whereas others become pools, depending on small differences in initial conditions.
Figure 2. Stabilization and destabilization of organic matter dynamics. The solid curve is rate of litter/peat formation as a function of thickness of the unsaturated zone (TUZ), whereas the dashed line is storage change (constant). TUZ is in a steady state when peat formation equals storage change (intersection of curve and line). The equation for the curve is from Belyea and Clymo [2001]. (a) Destabilization (positive feedback) at the unstable point, where rate of peat formation increases with increasing TUZ. In a dry year, the unsaturated zone thickens (see Figure 1b). In a wet year, the unsaturated zone thins (see Figure 1c). (b) Stabilization (negative feedback) at the stable point, where rate of peat formation decreases with increasing TUZ. In both dry and wet years, TUZ is maintained close to a steady state (see Figure 1a).
