A 2.5.... .

.4.‘ ,
.. 5?

.... .3. .ﬁ
‘ an ,a

.‘
. 36.».

v: i

1.}

$53».
:4. 3.,

, nun

‘.
.34... 69:,

9.3.1.3. «in: ..

 

. LLB

£|..3.....xc.

{If 5 Inﬁhiiﬂx
“oi-bl a

 

 

7/ LIBRARY
Qty; Michigan State
University

This is to certify that the
thesis entitled

SOIL HYDROCLIMATE, VEGETATION, AND SUBSTRATE
CONTROLS ON CARBON FLUX IN AN ALASKAN FEN

presented by

MOLLY R. CONLIN

has been accepted towards fulﬁllment
of the requirements for the

MS. degree in Plant Biology

 

 

’ C Major Professor’s Signature
07 ‘ 30 ’ QC- CB

Date

MSU is an afﬁrmative-action, equal-opportunity employer

 

a .—--n-n-n-n-o-a—.---u---------

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KIProi/AcoﬁPres/ClRC/DateDue indd

 

SOIL HYDROCLIMATE, VEGETATION, AND SUBSTRATE CONTROLS ON CARBON
FLUX IN AN ALASKAN FEN
By

Molly R. Conlin

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Plant Biology
Ecology, Evolutionary Biology, and Behavior

2008

ABSTRACT

SOIL HYDROCLIMATE, VEGETATION, AND SUBSTRATE CONTROLS ON
CARBON FLUX IN AN ALASKAN FEN

By
Molly R. Conlin
Peatlands store 30% of the world’s terrestrial soil carbon and are located primarily at
northern latitudes, where they are expected to experience severe climate warming. The
goal of my thesis was to determine the effect of experimental soil climate manipulations
on carbon (C) ﬂuxes in an Alaskan rich fen and to assess the indirect inﬂuence of
substrate quality on C mineralization rates in peat. I monitored growing season C02
ﬂuxes across a factorial design of in situ water table and soil warming treatments. The
lowered water table treatment did not alter ecosystem respiration (ER) of C02, but
lowered gross primary production (GPP), making this plot more of an atmospheric source
relative to the control. Relative to the control, the raised water table treatment had more
positive NEE values in 2005, but was a greater C sink in 2006 due to increased early
season GPP. To investigate the effect of the manipulations on carbon mineralization

through changes in soil organic matter (SOM) quality, I measured C02 and CH4
production from incubations at standard laboratory conditions. While CH4 production
rates were not affected by the manipulations, peat taken from the warmed subplots and
lowered water table plot had the lowest C02 production rates, indicating a decrease in

SOM quality induced by these climate treatments. My results suggest that climate change
will impact peatland C ﬂuxes to reduce ecosystem C storage under drought and to

increase ecosystem C storage with ﬂooding conditions.

ACKNOWLEDGEMENTS

I would like to thank my advisor, Merritt Turetsky, for her enthusiastic support of
this research and valuable guidance and mentoring throughout my graduate work. Thank
you to my committee members David Rothstein and Jay Lennon for their support and
valuable comments to the manuscripts. Thank you to David McGuire, Jennifer Harden,
and Mike Waddington for assistance with the development of research hypotheses and
methodology. Thank you also to Evan Kane, Claire Treat, Colin Philippo, Nicole
McConnell, and Spencer Rubin for help with ﬁeldwork, labwork, and logistics. Thanks

to my ofﬁcemates Jason Martina and Sara Kelzer for their support.

This research was funded by the Environmental Protection Agency Students to

Achieve Results (EPA STAR) graduate research fellowship and the National Science

Foundation (NSF) graduate research fellowship.

iii

TABLE OF CONTENTS

LIST OF TABLES .................................................................................... v

LIST OF FIGURES .................................................................................. vi

CHAPTER 1

INTRODUCTION ..................................................................................... 1
1.1 Climate change and projections for high latitudes ....................................... 1
1.2 Introduction to boreal peatlands ........................................................... 3

1.3 Experimental manipulations to study peatland response to climate change. . . . . ....8

CHAPTER 2
CARBON DIOXIDE (C02) F LUXES IN AN ALASKAN F EN: RESPONSE TO
EXPERIMENTAL MANIPULATION OF WATER TABLE AND SOIL

TEMPERATURE ................................................................................... 19
2.1 Introduction ................................................................................ 19
2.2 Methods ..................................................................................... 25
2.3 Results ...................................................................................... 33
2.4 Discussion .................................................................................. 42

CHAPTER 3

LABORATORY C02 AND CH4 PRODUCTION ACROSS AN EXPERIMENTAL
MANIPULATION OF WATER TABLE AND SOIL TEMPERATURE IN AN

ALASKAN FEN: IMPLICATIONS OF SOIL QUALITY .................................... 84
3.1 Introduction ................................................................................. 84
3.2 Methods ..................................................................................... 88
3.3 Results ....................................................................................... 93
3.4 Discussion ................................................................................... 97

CHAPTER 4

CONCLUSION .................................................................................... 124

iv

LIST OF TABLES

Table 2—1. Dominant vascular and bryophyte plant species at the Alaska Peatland ....... 56

Table 2-2. Repeated measures analysis of variance results analyzing C02 ﬂux ........... 57
Table 2-3. Results of a general linear model using environmental ........................... 58
Table 2-4. Results of a general linear model using environmental parameters. . . . . . . .59
Table 2-5. Non-linear NEE-PAR curve parameter estimates for 2006 ..................... 60
Table 2-6. Results of general linear models using vegetation ................................ 61
Table 3-1. Results of two repeated measures analysis of variance ........................ 106
Table 3-2. Results of two repeated measures analysis of variance ......................... 107
Table 3-3. Bulk density (g cm'3), % carbon, % nitrogen, and C/N for peat ............... 108
Table 3-4. Synthesis of northern peatland laboratory microcosm .......................... 109

Figure 1-1.
Figure 1-2.
Figure 2-1.
Figure 2—2.
Figure 2-3.
Figure 2-4.
Figure 2-5.
Figure 2-6.
Figure 2-7.
Figure 2-8.
Figure 2-9.
Figure 2-10.
Figure 2-11.
Figure 2-12.
Figure 2-13.
Figure 2-14.
Figure 3-1.
Figure 3-2.
Figure 3-3.
Figure 3-4.

Figure 3-5.

LIST OF FIGURES

The net ecosystem exchange of C02 between ecosystems ...................... 11
The peatland carbon cycle. Vegetation ﬁxes atmospheric ..................... 12
Schematic illustrating possible thresholds to ecosystem ........................ 62
The Alaska Peatland Experiment (APEX) experimental ........................ 63
At the Alaska Peatland Experiment (APEX) rich fen, 1 established ........... 64
Water table levels and precipitation at the manipulation plots ................. 65
Mean peat temperatures at 25 cm beneath the moss surface ................... 66
Average vascular leaf area index (m2 m'z) across our three water ............. 67
Growing season ecosystem respiration (ER) within each water ............... 68
Growing season gross primary production (GPP) within each ................. 69
Growing season net ecosystem exchange (NEE) within each .................. 7O

Diurnal net ecosystem exchange (NEE) of C02 within our ................... 71

Results of a repeated measures analysis of variance ............................ 72

Ecosystem respiration (ER) versus gross primary production. . . . . . . . . . . . .....73

Early season (Julian Day 140 — 179) ecosystem respiration .................. 74

Relationship between net ecosystem exchange of C02 (NEE) ............... 75
Mean anaerobic CH4 production (nmol cm'3 day") i one ..................... 110
Mean C02 production (umol cm"3 day") i one standard ...................... l l 1
Results of an interaction between depth increment ............................. l 12
Results of an interaction between depth increment and ........................ 113
Results of an interaction between depth increment and ........................ 1 14

vi

Figure 3-6. Results of an interaction between depth increment and water ................ 115
Figure 3-7. Relationships between C02 and CH4 production rates from .................. 117

Figure 38 N03 (mg L") concentrations from the aerobic and anaerobic ................ 1 18

vii

CHAPTER 1
INTRODUCTION

1.] Climate change and projections for high latitudes
Since the industrial revolution in the 1700’s; human activities, including

deforestation and the burning of oil, coal, and gas, have increased greenhouse gas (carbon
dioxide (CO2), methane (CH4), and nitrous oxide (N20)) concentrations in the

atmosphere by 36%, 150%, and 15%, respectively (National Research Council 2001).
Because greenhouse gases absorb and emit heat, their rising atmospheric concentrations
have a warming effect on the Earth. This warming is expected to be greatest in northern
high latitudes due to feedbacks from snow and sea ice, the stability of the lower
troposphere, and thawing of permafrost (Houghton et al. 1992, Ramaswamy et al. 2001,
Serreze et al. 2000).

Climate models project that the North American boreal forest will experience
more warming than any other terrestrial forest biome, with the greatest warming
occurring in the continental interiors (National Research Council 2001). Northem soils
in boreal and subarctic regions store large amounts of C that has slowly accumulated
since the last deglaciation (Harden et al. 1992). Boreal regions contain approximately
27% of the world’s vegetation C (McGuire et al. 1997) and 25 - 30% of the world’s soil
C (estimated between 397-455 Pg or 10'5 g C; Gorham 1991; Zoltai & Martikainen 1996,
Moore et al. 1998). Multiple impacts of climate change including degrading permafrost

(Romanovsky & Osterkamp 1997), reduced snow cover (Magnuson et al. 2000), and

longer growing seasons (Serreze et al. 2000) are likely to impact plant and soil processes,
which will impact C cycling in boreal regions.

Many regions are classiﬁed as “boreal” and grouped together for research
purposes. However, boreal regions cover a wide area (18.5 million kmz; McGuire et al.
1995) and represent a large range in climate and topography. While many boreal regions
are characterized by large, ﬂat glacial lake plains, interior Alaska is bordered on the north
and south by the Alaska Range and the Brooks Range, which successfully block coastal
air masses resulting in a continental climate with cold winters (extremes of -50°C in
January) and warm, relatively dry summers (highs over +33°C in July). Due to its high
latitude, Alaska is characterized by drastic seasonal ﬂuctuation in day length (more than
21 hours on June 21 and less than 4 hours on December 21; Hinzman et al. 2006).
Additionally, interior Alaska has a short growing season (135 days or less from early
May to mid-September) and minimal precipitation due to a montane rain shadow (the
average annual precipitation is only 269 mm in Fairbanks, 30% of which falls as snow;
Slaughter and Viereck 1986). Because of these differences in climate and topography,
ecosystems in Alaska, likely will respond differently to climate change than ecosystems
in Canada or Russian boreal regions.

Interior Alaskan soils have warmed rapidly over the past 30 years due to near-
surface atmospheric warming of approximately 1° C per decade on average (Osterkamp
and Romanovsky 1999). Global climate models predict temperature increases of 2.5 to
6° C and precipitation changes ranging from -10% to 30% for the Fairbanks, Alaska
region by 2050 (Canadian Centre for Climate Modelling and Analysis 2003, Hinzman et

al. 2005). Remote sensing work in Alaska shows that open water bodies in major

wetland regions in Alaska are losing surface area (Riordan et al. 2006), which could be
associated with increased summer water deﬁcits due to increased evapotranspiration with
climate warming (Hinzman et al. 2005). However, wetlands at the margins of
continents may become wetter (Oquist and Svensson 1995) due to thermokarst and
permafrost melting (Romanovsky & Osterkamp 1997). In some areas of Alaska,
wetlands are becoming more saturated due to permafrost thaw and increased runoff from
surrounding uplands, such as the expansion of open water in the Tanana Flats region of
interior Alaska (Osterkamp et al. 2000). Thus, while some wetlands in Alaska are drying,
others currently are becoming wetter, suggesting that future changes in wetland
hydrology could include either drying or inundation due to increased runoff, permafrost

thaw, changes in precipitation, and increased evapotranspiration.

1.2 Introduction to boreal peatlands
Peatlands and their distribution in northern regions

Peat, or partially carbonized vegetation, accumulates where C ﬁxation through net
primary production (NPP) at the surface exceeds losses from decomposition, leaching,
and/or disturbance. Peatlands are deﬁned as any wetland with 40 cm or more of
accumulated peat. Approximately 80% of the world’s peatlands are in high latitudes.
Peatlands globally cover 24% of the circumboreal land area (Wieder et al. 2006) where
they cover major portions of Alaska, Canada, Russia, the Baltic Republics, and
Fennoscandia (Clymo 1983, Gorham 1991, Vitt 2006). The largest expanse of peatlands
is in the boreal regions of Canada, Russia (Siberia), and Alaska (350 x 106 ha; Gorham

1991, Botch et al. 1995).

Five primary factors affect the function of a peatland, including hydrology,
climate, substrate, chemistry, and vegetation (Vitt 2006). These “state factors” have been
used to classify peatlands into three main types: bogs, rich fens, and poor fens. In bogs,
peat is often built up above regional water tables to such an extent that the living
vegetation is raised above sources of surrounding surface water or underlying
groundwater. These peatlands are ombrotrophic, meaning that.they receive water and
nutrients solely from precipitation or dry fall. As such, bogs are usually the most nutrient
poor and highly acidic type of peatland. Acidity comes from organic acid production
with decay and from Sphagnum cation exchange capacity. Sphagnum, or peat moss, is
typically the dominant vegetation type, and because Sphagnum species are strong
competitors, bogs tend to have low species diversity. Fens are minerotrophic and receive
some water and nutrients from ground- or surface- runoff in addition to precipitation.
Fens vary from nutrient rich, emergent vascular or brown moss dominated ecosystems
with a high species diversity (rich fens), to nutrient poor, Sphagnum dominated
ecosystems with low species diversity (poor fens). The fundamental differences in
hydrology and species composition among these peatland types inﬂuence rates of
decomposition and productivity, and ultimately peat accumulation rates. For example,
studies have found that bogs generally have lower rates decomposition and productivity
relative to fens (Frolking et al. 1998, Thormann and Bayley 1997). Even though rich fens
represent the most common peatland type in western boreal North America, most boreal
peatland research has focused on Sphagnum dominated bogs or poor fens (Vitt 2006).
Thus, while the peatland research community has some understanding of the response of

bog (Alm et al. 1997, Moore et al. 2002, Laﬂeur et al. 2005) and poor fen (Silvola et al.

1996, Bubier et al. 2003) peat to climatic ﬂux, less is known about rich fens and their
vulnerability to climate change. Rich fens will likely differ in their response to climate
change relative to these Sphagnum-dominated systems, given that brown moss and
emergent vegetation respond differently than Sphagnum to warming and altered moisture
(Weltzin et al. 2005) and systems with high nutrient availability respond differently to
climate change than when nitrogen (N) is limited (warming increases net N
mineralization rates; Rustad et al. 2000).

In Alaska, about 20% of the landscape is covered by poorly drained ecosystems
(Harden et al. 2003). Wetland abundance in interior Alaska is largely inﬂuenced by
landscape topography (responding to runoff from surrounding areas and permafrost
degradation), and thus peatlands tend to form in valley bottoms. Interior Alaskan
wetlands also are often found in ﬂoodplains that form large wetland complexes, such as
the Minto and Yukon Flats. These large wetland complexes oﬁen are dominated by
tussock marshes or sparsely treed wetlands underlain by permafrost. Areas that are
protected from river erosion (i.e., old oxbows) or ﬁre often accumulate peat to form

peatlands

Studies of Alaskan peatlands to date largely have focused on southeastern Alaska
(Ugolini and Mann 1979, Klinger et al. 1990, Concannon 1995, Hartshom et al. 2003) or
northern tundra ecosystems (Billings 1987, Oberbauer et al. 1992, Klinger 1996).
Despite the documented changes in wetland hydrology in interior Alaska (Riordan et al.
2006), few studies have focused on interior Alaskan peatlands and their response to

climate change.

Peatland carbon storage and ﬂuxes

Peatlands globally cover only 3-5% of the Earth’s terrestrial land base, but
contain 30% of the world’s terrestrial soil C (between 270 - 370 Pg C; Vasander and
Kettunen 2006). Historically, peatlands have acted as a net sink for global atmospheric
C, sequestering an estimated 29 g C m’2 annually from the atmosphere on a millennial
time scale (Gorham 1991, Bartlett and Harriss 1993, Zoltai and Martikainen 1996). In
general, peatlands began accumulating peat about 8-10 k years ago in Canada and Siberia
and about 4-6 k years ago in Alaska. While estimates are not well constrained, Alaska
peatlands today store approximately 41.7 i 50% Pg of C (Bridgham et al. 2000).

A peat accumulation model developed by Clymo (1992) divides the peat proﬁle
into two sections based on the location of the water table. Fast aerobic decomposition
pathways dominate the surface acrotelm (unsaturated oxic zone), while slower, anaerobic
decomposition occurs in the catotelm (saturated deeper anoxic zone). 80-90% of the C
that passes through the acrotelm is lost through decomposition. Therefore, long-term
peat accumulation depends on the rate of C transfer to the slowly decomposing catotelm.
The input of C to this anoxic zone is typically 10-20% of total vegetation litter
production.

In general, peat accumulation occurs because cool, wet conditions in peatlands
limit decomposition. Because of their dependence on soil climate, peatlands and their
extensive C reservoirs are likely to be altered by climate change (Gorham 1994).

Peatlands are strongly controlled by climate, and recent work shows that individual
peatlands can switch from net C02 sinks to sources between wet and dry years (Shurpali

et al. 1995, Alm et al. 1999, Bubier et al. 2003).

Peatlands play a key role in global C cycling by both sequestering and emitting
atmospheric C02 and by emitting CH4 to the atmosphere (Moore 1996). Peatlands
currently act as a net sink for atmospheric C02, sequestering approximately 150 Gt yr"I
of C from the atmosphere (Gorham 1991, Gorham 1994). The net ecosystem exchange
(NEE) of C is a direct measure of the net C02 exchange between ecosystems and the

atmosphere. Net ecosystem exchange is the balance between gross primary production
(GPP), which is plant C uptake, and ecosystem respiration (ER), which is the sum of

heterotrophic and plant respiration. Rates of NEE provide an indication of whether the
ecosystem is serving as a net sink or source of atmospheric C02. Rates of GPP are zero
in the dark, so in dark conditions NEE is also a measure of ER. Root respiration and
decomposition are the primary mechanisms contributing to CO2 emissions from peat
(Moore and Knowles 1987). Air and soil temperature, water table level, plant and
microbial activity, and the quality of organic substrates are the main controls on CO2
production in peatlands (Figure 1-1; Moore et al. 1998, Updegraff et al. 1995). Plant
uptake of C02 depends principally on photosynthetically active radiation (PAR), air

temperature, and plant community structure and composition (Figure 1-1; Moore et al.

1998).

While serving as a net sink for atmospheric C02, peatlands Simultaneously serve
as a net source of methane (CH4; Gorham 1991, Gorham 1994), releasing an estimated 30
to 50 Tg CH4 yr‘l (Chen and Prinn 2006). Over a 100-year time span, a sustained
emission of CH4 has approximately 25 times more global temperature change potential

than C02 (Shine et al. 2005), so small emissions can contribute signiﬁcantly to the total

budget of radiatively active gases (Whalen 2005). Both C02 and CH4 ﬂuxes are
mediated by temperature and moisture in peatlands (Figure 2-1), but CH4 is produced in

the absence of 02. Therefore, the primary controls on CH4 emissions in peatlands are

water table level, soil temperature, and substrate quality (Moore and Knowles 1989,

Moore et al. 1998, Bellisario 1999).

A major uncertainty in the face of climate change is whether peatlands will
continue to act as a net sink for atmospheric C02, or whether changes in climate will

release peatland C pools back to the atmosphere. Atmospheric C concentrations will

likely be inﬂuenced by peatland responses to climate change and will be determined by
either positive feedbacks that occur with enhanced C02 and/or CH4 emissions from

peatlands or negative feedbacks with increased GPP and enhanced C Sinks (McGuire and

Hobbie 1997, McGuire et al. 2000, Chapin et al. 2000, Matthews and Keith 2007).

1.3 Experimental manipulations to study peatland response to climate change

Predicting ecosystem responses to climate change requires a detailed and
mechanistic understanding of climate-ecosystem interactions over long time scales.
Many studies have used natural temporal or spatial gradients to investigate climatic
controls on peatland C cycling (see Updegraff et al. 1995, Silvola et al. 1996, Alm et al.
1999, Bubier et al. 1998, Laﬂeur et al. 2005). Natural gradients are useful for acquiring
large scale understanding of ecosystem responses to climate change. The value of
predictions from gradient analyses depends on the assumption that ecosystems will track

changing climate over time in the same way that ecosystems now vary with climate over

space (Dunne et al. 1996). Long-term adaptation to local climate conditions, ﬁne—scale
environmental heterogeneity, co-varying abiotic factors, and differences in time constants
may confound the use of gradients to predict responses to global warming (Vitousek

1994, Root and Schneider 1995).

Compared with gradient studies, experiments provide a more controlled,
mechanistic approach to predicting ecosystem responses to climate change, and can
identify the most important factors that inﬂuence those responses. Given that soil
environments in northern ecosystems are predicted to change beyond their normal range
of variability, model predictions of ecosystem function under future climate change often
must rely on extrapolations beyond current data or on results from experiments that
simulate climatic regimes outside of contemporary variability. Researchers who utilize
the experimental approach must simulate the desired change in climate while minimizing
confounding changes in other variables. Results from manipulation experiments are
limited by a variety of issues including the difﬁculty in establishing good controls and the
expense and time needed for large scale manipulations and replication (Marion et al.
1997). The short duration of most experiments also lead to dangers of false
understanding of the response and/or predictions; as initial ecosystem responses to
experimental change may differ from responses observed when the manipulation is

sustained over longer periods (Rosswall et al. 1988, Dunne et al. 1996).

In this thesis, I investigate direct and indirect soil hydroclimate (water table

position, soil temperature) controls on C ﬂuxes in an Alaskan fen. Water table level and
soil temperature have been identiﬁed as primary controls on CO2 production; however I

investigated the indirect effects of soil hydroclimate on C ﬂuxes through vegetation and

C quality. Chapter 2 of my thesis describes the response of ﬁeld C02 ﬂuxes (NEE, ER,

GPP) to a water table and soil temperature manipulation. These manipulations were
guided by future climate predictions for interior Alaska, and thus will provide
information beyond the scope of contemporary soil climate variation that will be useful
for modeling the future C balance of poorly drained ecosystems in this region. This
design gave me the opportunity to investigate water table position (3 treatments: control,

soil drying, soil wetting), surface soil warming, and the interactions between water table
and warming on vegetation and C02 dynamics.

Water table level and soil temperature have the potential to indirectly affect
decomposition through changes in soil organic matter quality. Chapter 3 describes a
laboratory experiment designed to isolate soil organic matter quality differences among

peat collected across the experimental ﬁeld treatments (Chapter 2). I measured aerobic
and anaerobic C02 and anaerobic CH4 production rates under constant moisture and

temperature to understand whether the ﬁeld manipulations affected soil organic matter
quality and C mineralization rates. Detailed hypotheses and predictions for these two
studies are described in sections 2.1 and 3.1 of this thesis. This study represents some of
the most detailed C ﬂux dataset from a boreal rich fen as well as from an interior Alaskan
wetland. Given that the response. of peatland C ﬂuxes to a changing climate remains

uncertain, this study will improve estimates of future C emissions from this landscape.

10

Production

Photosynthetic
active radiation

 

 

 

Plant community
structure and
function

 

 

 

 

 

Soil climate

 
 

................................
-

Nitrogen
mineralization

Decomposition

 

    

Soll nucroblal Soil organic
community matter quality

Figure 1-1. The net ecosystem exchange of C 02 between ecosystems and the
atmosphere is the balance between production (plant uptake of CO2) and ecosystem
respiration of CO2 through autotrophic and heterotrophic respiration (decomposition).
Interactions between soil organisms, soil microclimate, and organic matter quality
inﬂuence decomposition processes in peatlands. Air and soil microclimate,
photosynthetic active radiation, and organic matter quality control vegetation community
structure and function. In peatlands, vegetation can have a strong inﬂuence on both soil
climate and quality by altering inputs to the soil.

 

A

 

Atmospheric C02

 

 

 

 

 
  

 

 

 

  
 

  

 

 

Atmospheric CH4 k

 

 

 

 

 

 

Ph"""y""'“,"‘ Aboveground plant
Plant Respiration biomass
,—
Allocation and burial .
Aerobic
decomposition Oxidation
. . CO2 <-——-‘
WATER TABLE LlVlng and dead A
belowground biomass
Aquatic DOC ‘ .
systems Export Anaerobic
decomposition
CO2 and CH4

 

 

Figure 1-2. The peatland carbon cycle. Vegetation ﬁxes atmospheric CO2 through
primary production and a portion of this carbon is transferred to belowground biomass
(peat). Aerobic and anaerobic decomposition transforms solid and dissolved organic

carbon into CO2 and CH4 that diffuse to the atmosphere.

References

Alm J, Talanov A, Saarnio S, Silvola J, Ikkonen E, Aaltonen H, Nykanen H, Martikainen

PJ. 1997. Reconstruction of the carbon balance for microsites in a boreal oligotrophic
pine fen, Finland. Oecologia 110: 423-431.

Alm J, Schulman L, Silvola J, Walden J, Nykanen H, Martikainen PJ. 1999. Carbon
balance of a boreal bog during a year with an exceptionally dry summer. Ecology 80:
161-174.

Bartlett, KB, Harriss RC. 1993. Review and assessment of methane emissions from
wetlands. Chemosphere 26: 261-320.

Bellisario, LM, Bubier JL, Moore TR, Chanton JP. 1999. Controls on CH4 emissions
from a northern peatland. Global Biogeochemical Cycles 13: 81—91.

Billings WD. 1987. Carbon balance of Alaskan tundra and taiga ecosystems: Past,
present and future. Quaternary Science Reviews 6: 165-177.

Bridgham SD, Megonigal JP, Keller JK, Bliss NB, Trettin C. 2006. The carbon balance
of North American wetlands. Wetlands 26: 889-916.

Bubier JL, Frolking S, Crill PM, Moore TR, Savage K, Vamer RK. 1998. Seasonal,

patterns and controls on net ecosystem C02 exchange in a boreal peatland complex.
Global Biogeochemical Cycles 12: 703-714.

Bubier JL, Bhatia G, Moore TR, Roulet NT, Laﬂeur PM. 2003. Spatial and Temporal
Variability in Growing-Season Net Ecosystem Carbon Dioxide Exchange at a Large
Peatland in Ontario, Canada. Ecosystems 6: 353-367.

Canadian Centre for Climate Modelling and Analysis. Accessed April 2008.
http://www.cccma.bc.ec.gc.ca/

Chapin F S, McGuire AD, Randerson J, Pielke R, Baldocchi D, Hobbie SE, Roulet M,
Eugster W, Kasischke E, Rastetter EB, Zimov SA, Running SW. 2000. Arctic and boreal
ecosystems of western North America as components of the climate system. Global
Change Biology 6: 211-223.

Chen Y, Prinn RG. 2006. Estimation of atmospheric methane emissions between 1996
and 2001 using a three-dimensional global chemical transport model. Journal of
Geophysical Research 1 1 1: 1029-1040.

Clymo RS. 1983. Peat. Ecosystems of the World, Corstanje and Reddy.

Clymo RS. 1992. Models of peat growth. Suo 43: 127-136.

13

Concannon, JA. 1995. Characterizing structure, microclimate and decomposition of
peatland, beachfront, and newly logged forest edges in Southeastern Alaska.
University of Washington, Thesis (Ph.D.) 56: 6486-6594.

Dunne JA, Saleska SR, Fischer ML, Harte J. 2004. Integrating experimental and gradient
methods in ecological climate change research. Ecology 85: 904-916.

Frolking S, Bubier JL, Moore TR, Ball T, Bellisario LM, Bhardwaj A, Carroll P, Crill
PM, Laﬂeur PM, McCaughey JH, Roulet NT, Suyker AE, Verma SB, Waddington JM,
Whiting GJ. 1998. Relationship between ecosystem productivity and photosynthetically
active radiation for northern peatlands. Global Biogeochemical Cycles 12: 115-126.

Gorham E. 1991. Northern peatlands: role in the carbon cycle and probable responses to
climatic warming. Ecological Applications 1: 182-195.

Gorham E. 1994. The future of research in Canadian peatlands: a brief survey with
particular reference to global change. Wetlands 14: 206-215.

Halsey L, Vitt D, Zoltai S. 1997. Climatic and physiographic controls on wetland type
and distribution in Manitoba, Canada. Wetlands 17: 243—262.

Harden JW, Sunquist ET, Stallard RF, Mark RK. 1992. Dynamics of soil carbon during
deglaciation of the Laurentide Ice Sheet. Science 258: 1921-1924.

Harden JW, Meier R, Damel D, Swanson DK, McGuire AD. 2003. Soil drainage and its
potential for inﬂuencing wildﬁre in Alaska in Studies in Alaska by the US Geological
Survey edited by Galloway JP. US. Geological Survey Professional Paper 1678,

Hartshom AS, Southard RJ, Bledsoe CS. 2003. Structure and function of peatland-forest
ecotones in southeastern Alaska. Soil Science Society of America Journal. 67: 1572-1581.

Hinzman, LD, Bettez ND, Bolton WR, Chapin F S, Dyurgerov MB, F astie CL, Grifﬁth B,
Hollister RD, Hope A, Huntington HP, Jensen AM, Jia GJ, Jorgenson T, Kane DL, Klein
DR, Koﬁnas G, Lynch AH, Lloyd AH, McGuire AD, Nelson FE, Nolan M, Oechel WC,
Osterkamp TE, Racine CH, Romanovsky VE, Stone RS, Stow DA, Sturm M, Tweedie
CE, Vourlitis GL, Walker MD, Walker DA, Webber PJ, Welker J M, Winker KS,
Yoshikawa K. 2005. Evidence and implications of recent climate change in northern
Alaska and other Arctic regions. Climatic Change 72: 251-298.

Hinzman, LD, Viereck LA, Adams PC, Romanovsky VE, Yoshikawa K. 2006. Climate
and permafrost dynamics of the Alaskan boreal forest, in Alaska’s Changing Boreal
Forest edited by Chapin, F S III, Oswood MW, Van Cleve K, Viereck LA, Verbyla DL,
editors. Oxford University Press, New York.

Houghton, JT, Ding Y, Griggs DJ, Noguer M, van der Linden PJ, Xiaosu D. 2001.
Climate Change 2001: The Scientiﬁc Basis 944 pp, Intergovernmental Panel for

14

Climate Change, Cambridge University Press, UK.

Klinger LF, Elias SA, Behan-Pelletier VM, Williams NE. 1990. The bog climax
hypothesis: Fossil arthropod and stratigraphic evidence in peat sections from Southeast
Alaska, USA.

Holarctic Ecology 13: 72-80.

Klinger LF. 1996. Coupling of soils and vegetation in peatland succession. Arctic and
Alpine Research. 28: 380-387.

Laﬂeur PM, Moore TR, Roulet NT, Frolking S. 2005. Ecosystem respiration in a cool
temperate bog depends on peat temperature but not water table. Ecosystems 8:619-629.

Magnuson JJ, Robertson DM, Benson BJ, Wynne RH, Livingstone DM, Arai T, Assel
RA, Barry RG, Card V, Kuusisto E, Granin NG, Prowse TD, Stewart KM, Vuglinski VS.

2000. Historical trends in lake and river ice cover in the Northern Hemisphere. Science
289: 1743-1746.

Marion GM, Henry GHR, Freckman DW, Johnstone J, Jones G, Jones MH, Le'vesque E,
Molau U, Molgaard P, Parsons AN, Svoboda J, Virginia RA. 1997. Open top designs for

manipulating ﬁeld temperature in high-latitude ecosystems. Global Change Biology 3:
20—32.

Matthews HD, Keith DW. 2007. Carbon-cycle feedbacks increase the likelihood of a
warmer future. Geophysical Research Letters 34: 1-5.

McGuire AD, Melillo JW, Kicklighter DW, Joyce LA 1995. Equilibrium responses of
soil carbon to climate change: empirical and process-based estimates. Journal of
Biogeography 22: 785-796.

McGuire AD, Hobbie JE. 1997. Global climate change and the equilibrium responses of
carbon storage in arctic and subarctic regions. pp. 47-48. Fairbanks, Alaska.

McGuire AD, Melillo JM, Kicklighter DW, Pan YD, Xiao XM, Helfrich J, Moore B,
Vorosmarty CJ, Schloss AL. 1997. Equilibrium responses of global net primary
production and carbon storage to doubled atmospheric carbon dioxide: Sensitivity to

changes in vegetation nitrogen concentration. Global Biogeochemical Cycles 11: 173-
189.

McGuire AD, Clein JS, Melillo JM, Kicklighter DW, Meier RA, Vorosmarty CJ, Serreze
MC. 2000. Modeling carbon responses of tundra ecosystems to historical and projected
climate: Sensitivity of pan-arctic carbon storage to temporal and spatial variation in
climate. Global Change Biology 6: 141-159.

Moore TR, Knowles R. 1987. Methane and carbon dioxide evolution from subarctic fens.
Canadian Journal of Soil Science 67: 77-81.

15

Moore TR, Knowles R. 1989. Inﬂuence of water table levels on methane and carbon
dioxide emissions from peatland soils. Canadian Journal of Soil Science 69: 33-38.

Moore TR. 1996. Trace gas ﬂuxes from northern peatlands. Pp. 115-121 In Proc.
Workshop on northern peatlands in global climate change, Academy of Finland, Helsinki.

Moore, TR, Roulet NT, Waddington JM. 1998. Uncertainties in predicting the effect of

climatic change on the carbon cycling of Canadian peatlands. Climatic Change 40: 229—
245.

Moore TR, Bubier JL, Frolking SE, Laﬂeur PM, Roulet NT. 2002. Plant biomass and

production and CO2 exchange in an ombrotrophic bog. Journal of Ecology 90: 25—36.
National Research Council, Committee on the Science of Climate Change, Division on
Earth and Life Studies. 2001. Climate Change Science: An analysis of Some Key
Questions. National Academy Press. Washington, DC.

Oberbauer SF , Gillespie CT, Cheng W, Gebauer R, Sala Serra A, Tenhunen JD. 1992.

Environmental effect on CO2 efﬂux from riparian tundra in the northern foothills of the
Brooks Range, Alaska, USA. Oecologia 92: 568-577.

Oquist MG, Svensson BH. 1995. Impacts, adaptations, and mitigations of climate change:
Scientiﬁc-technical analyses, in Climate Change edited by Watson RT, Zinyowera MC,
Moss RH, Dokken DJ. pp. 215-240. Cambridge University Press, Cambridge, UK.

Osterkamp TE, Romanovsky VB. 1999. Evidence for warming and thawing of
discontinuous permafrost in Alaska. Permafrost and Periglacial Processes 10: 17-37.

Osterkamp TE, Viereck L, Shur Y, Jorgenson MT, Racine C, Doyle A, Boone RD. 2000.
Observations of thermokarst and its impact on boreal forests in Alaska, USA. Arctic
Alpine Research 32: 303-315.

Ramaswamy V. 2001. Radiative forcing of climate change. pp. 349-416. Cambridge
University Press.

Riordan B, Verbyla D, McGuire AD. 2006. Shrinking ponds in subarctic Alaska based on
1950-2002 remotely sensed images. Journal of Geophysical Research - Biogeosciences:
004002, doi:10:1029/2005J0000150.

Romanovsky VE, Osterkamp TE. 1997. Thawing of the active layer on the coastal plain
of the Alaskan arctic. Permafrost and Periglacial Processes 8: 1-22.

Root TL, Schneider SH. 1995. Ecology and climate: research strategies and implications.
Science 269: 334—341.

16

Rosswall T, Woodmansee RG, Risser PG. 1988. Scales and global change: spatial and
temporal variability in biospheric processes. John Wiley and Sons, New York, New
York, USA.

Serreze MC, Walsh JE, Chapin F S, Osterkamp TE, Dyurgerov M, Romanovsky VE,
Oechel WC, Morison J, Zhang T, Barry RG. 2000. Observational evidence of recent
change in the northern high-latitude environment. Climatic Change 46: 159—207.

Shine KP, Fuglestvedt JS, Hailemariam K, Stuber N. 2005. Alternatives to the global
warming potential for comparing climate impacts of emissions of greenhousegases.
Climatic Change 68: 281-302.

Shurpali NJ, Verma SB, Kim J. 1995. Carbon dioxide exchange in a peatland ecosystem.
Journal of Geophysical Research 100: 14319-14326.

Silvola J, Alm J, Ahlholm U, Nykanen H, Martikainen PJ. 1996. C02 ﬂuxes from peat in
boreal mires under varying temperature and moisture conditions. The Journal of Ecology
84: 219—28.

Slaughter CW, Viereck LA. 1986. Climatic characteristics of the taiga in interior Alaska
in Forest Ecosystems in the Alaskan Taiga pp 9-21 edited by K. Van Cleve, F. S.
Chapin, III, P. W. Flanagan, L. A. Viereck, and C. T. Dymess. Springer-Verlag, New
York.

Thorrnann MN, Bayley SE. 1997. Aboveground plant production and nutrient content of
the vegetation in six peatlands in Alberta, Canada. Plant Ecology 131: 1-16.

Ugolini FC, Mann DH. 1979. Biopedological origin of peatlands in South East Alaska.
Nature 281: 366-368.

Updegraff K, Pastor J, Bridgham SD, Johnston CA. 1995. Environmental and substrate
controls over carbon and nitrogen mineralization in northern wetlands. Ecological
Applications: 5 151-163.

Vasander H, Kettunen A. 2006. Carbon. in boreal peatlands. Ecological Studies 188: 165-
194.

Vitousek PM. 1994. Factors controlling ecosystem structure and function in Factors of
Soil Formation: A 50‘h Anniversary Symposium edited by Amundson R, Harden J, Singer
M, pp 87-97 Soil Science Society of America, Madison, Wisconsin, USA.

Vitt DH. 2006. Functional characteristics and indicators of boreal peatlands, in Boreal

Peatland Ecosystems, edited by Wieder RK, Vitt DK, pp. 1-8, Springer-Verlag Berlin
Heidelberg, New York.

17

Waddington JM, Grifﬁs TJ, Rouse WR. 1998. Northern canadian wetlands: Net
ecosystem CO2 exchange and climatic change. Climatic Change 40: 267-275.

Whalen SC. 2005. Biogeochemistry of methane exchange between natural wetlands and
the atmosphere. Environmental Engineering Science 22: 73-94.

Wieder RK, Vitt DK, Benscoter BW. 2006. Peatlands and the boreal forest, in Boreal
Peatland Ecosystems, edited by Wieder RK, Vitt DH, pp. 1-8, Springer-Verlag Berlin
Heidelberg, New York.

Zoltai SC, Martikainen PJ. 1996. The role of forested peatlands in the global carbon
cycle. Nato ASI Series I 40: 47-58.

18

CHAPTER 2

Carbon dioxide (CO2) Fluxes in an Alaskan Fen: Response to Experimental
Manipulation of Water Table and Soil Temperature

2.1 INTRODUCTION

2.1.] Peatlands as long term carbon sinks

Peatlands cover 15% of the boreal region (McGuire et al. 1995) and store up to
30% of the world’s terrestrial soil carbon (C) (estimated between 397-455 Pg C; Gorham
1991, Zoltai & Martikainen 1996, Moore et al. 1998). The majority (98%) of North
American soil C is stored in wetland soils and 83% of this C is contained in peatlands
(Bridgham et al. 2006). Currently, peatlands are thought to function globally as a net
sink for atmospheric carbon dioxide (CO2), sequestering approximately 76 Tg (10'2 g) C
yr'l (Zoltai & Martikainen 1996). However, recent work has shown that individual
peatlands can switch from net CO2 sinks to sources between wet and dry years (Shurpali
et al. 1995, Alm et al. 1999). Such temporal trends indicate strong climatic controls on
wetland C balance in northern regions.

Peat accumulates where C ﬁxation through net primary production (NPP) at the
surface exceeds losses from decomposition, leaching, and/or disturbance. Long term
rates of C accumulation in peatlands (29 g C m'2 yr'l averaged over millennia; Gorham
1991) are approximately 10 times larger than long term soil C accumulation rates for

upland ecosystem soils (Schlesinger 1990). For long term C accumulation to occur in

19

boreal peatlands, some proportion of plant production must be transferred to the catotelm
(Clymo 1983). Generally, 80-90% of the C that passes through the acrotelm is lost to
decomposition. Therefore, 10-20% of total vegetation production is transferred to the
catotelm, where it generally is protected from ﬂuctuating atmospheric conditions (Clymo

1992).

2.1.2 Climate change in Alaska and potential consequences for wetland C02 fluxes

Alaska’s ecosystems are expected to experience earlier and more drastic climate
changes compared to lower latitude systems (Keyser et al. 2000). A large amount of the
global peatland C pool is stored in Alaskan peatlands (41.7 i 50% Pg C; Bridgham et al.
2006), but the vulnerability of this C stock to C cycling in unknown. In upland
ecosystems, Alaskan soils have undergone rapid warming over the past 30 years due to
near-surface atmospheric warming of approximately 10 C per decade on average
(Osterkamp and Romanovsky 1999, Serreze et al. 2000). These large changes in climate
will not only warm peatland soils, but also will alter the hydrologic cycle in Alaska.
Multiple impacts of climate change, including the degradation of permafrost
(Romanovsky & Osterkamp 1997), reduced snow cover (Magnuson et a1. 2000), and
longer growing seasons (Serreze et al. 2000), are likely to impact plant and soil
processes, which will have large implications for CO2 dynamics and C storage in Alaskan
peatlands.

Recent remote sensing work in Alaska shows changing hydrologic conditions in

wetland complexes (Riordan et al. 2006). In several regions across interior Alaska, open

water bodies are shrinking, often accompanied by the encroachment of drier terrestrial

20

vegetation (Riordan et al. 2006). Wetland contraction is associated with increased
summer water deﬁcits due to increased evapotranspiration and/or decreased precipitation
(Oechel et al. 2000, Hinzman et a1. 2005). In other areas, however, wetlands are
becoming more saturated due to permafrost thaw and increased upwelling of melt water.
Few studies have explored such hydroclimate changes to wetland soil environments and
how they are likely to inﬂuence C storage and emissions in Alaska.

Air and soil temperature, water table level, photosynthetically active radiation

(PAR), plant community structure and function, microbial community, and the quality of
organic substrates have been identiﬁed as the main controls on CO2 net ecosystem

exchange (NEE) from peatlands (Figure 1-1; Updegraff et a1. 1995, Waddington and
Roulet 1996, Alm et al. 1997, Laﬂeur et al. 1997). Climate change has the potential to
directly and indirectly inﬂuence CO2 ﬂuxes by modifying these primary controls.
Warmer air and soil temperatures generally stimulate microbial activity resulting in high
ecosystem respiration (ER) of CO2 (Billings et al. 1982, Crill et al. 1988, Moore and
Dalva 1993, Frolking and Crill 1994, Silvola et al. 1996, Updegraff et al. 1998, Bergman
et al. 1999, Christensen et a1. 1999) and increased gross primary production (GPP) due to
greater nutrient availability (Arft et al. 1999). For example, ﬁeld warming experiments
conducted in arctic regions of Alaska at Toolik Lake (Hobbie and Chapin 1998, Grogan
and Chapin 2000, Van Wijk et al. 2004), stimulated decomposition and ER (Grogan and
Chapin 2000), increased GPP, and altered plant community structure (Hobbie and Chapin

1998)

Many studies have explored water table controls on peatland C02 ﬂuxes from a

variety of spatial scales, from controlled laboratory microcosm experiments, to larger

21.

mesocosms, and ﬁeld settings. Small-scale laboratory microcosm incubations have

examined soil moisture controls on peat decomposition rates (CO2 production).

Generally, CO2 production rates increased with lower soil moisture due to greater aerobic
mineralization (Hogg et al. 1992, Moore and Dalva 1997), but studies also have found a

positive relationship between CO2 production and soil moisture (Orchard et al. 1992,

Waddington et al. 2001), likely due to an optimum soil moisture content for CO2

production (Silvola and Ahlholm 1989). While these microcosm experiments provide
valuable insight into general controls on decomposition, they do not consider the role of
vegetation in impacting NEE (either through GPP or ER). The use of larger, mesocosm
incubations of peat columns allows for a controlled experiment to manipulate water table
level, while maintaining vegetation to understand controls on NEE. Generally, lowering
the water table in these studies resulted in higher ER rates (Moore and Knowles 1989,
Moore and Dalva 1993, Blodau et al. 2004) and decreased GPP (Williams et al. 1999,
Blodau et al. 2004), while raising the water table resulted in decreased ER likely due to
limited oxygen difﬁision for aerobic mineralization (Moore and Knowles 1989, Moore
and Dalva 1993, Aerts and Ludwig 1997, Chimner and Cooper 2003, Blodau et al. 2004,
Corstanje and Reddy 2004). Mesocosms provide an ideal setting to manipulate water
table while isolating and quantifying C ﬂux responses, but do not accurately portray the
true response of peatlands to climate change because of the lack of in-site variability in

climate, vegetation community, geology, and hydrology. Field water table drawdown
experiments have documented large responses of peatland CO2 ﬂuxes post drawdown,

including increased ER from aerobic decomposition as the water table drops (Nykanen et

al. 1995, Laiho 2006). Complex responses of vegetation community and productivity to

22

water table drawdown experiments also have important consequences for NEE, including
increased vascular vegetation (Strack and Waddington 2007) and decreased cover of
Sphagnum mosses (Weltzin et a1. 2001), which will alter GPP rates. Few studies have
explored interactions between water table and warming in the ﬁeld (but see mesocosm

studies by Laﬂeur et al. 2005 and Updegraff et al. 1995) and how these abiotic factors
might impact biotic controls on CO2 ﬂuxes (vegetation community and productivity)

within peatlands. Experiments designed to understand these interactions found that
temperature was a greater control on ER than water table level in both a temperate bog
(Laﬂeur et al. 2005) and a patterned sedge fen (Updegraff et al. 1995) and also
documented signiﬁcant shifts in plant C allocation, plant community, and biomass
production in response to water table and warming treatments, with few interactions

between the two (Weltzin et al. 2000).

2.1.3 Study hypotheses and objectives

The objective of this study is to use an in situ factorial manipulation of water table

and soil temperature to investigate controls on Alaskan peatland CO2 ﬂux and to

understand the impact soil climate has on ecological controls of CO2 ﬂuxes over time.

The experiment included a factorial design of water table position (three treatments
including a control, a lowered or drying treatment, and a raised or ﬂooded treatment) and
surface soil temperature (two treatments including a control or no warming treatment,

and surface soil warming via open top chambers) manipulations.

23

I hypothesize and that CO2 ﬂuxes will respond to manipulation of both water

table position and soil temperature at the Alaskan fen; such that:

H: Early responses to both water table and soil warming manipulations (ﬁrst
2.5 years) will be driven primarily by changes in decomposition rates.
Changes in vegetation will become more substantial over time, especially

changes in community structure.

I I predict that lowering water table position will increase acrotelm thickness,
stimulating the aerobic mineralization (ER) of newly exposed labile C

substrates followed by the turnover of older soil organic matter, which will

results in greater ER of C02.

' Drier soils will decrease GPP due to vegetative drought stress, which will
cause a shift from more negative (atmospheric sink) to more positive

(atmospheric source) NEE values.

. Because inundation by higher water table levels will favor vascular and
moss productivity, I predict that the raised water table treatment will

decrease aerobic mineralization (ER), and increase GPP.

H2: Warming will stimulate mineralization and ER at the fen. Rates of GPP
will also increase in response to greater nitrogen (N) mineralization and

availability.

H3: Modiﬁed water table position and temperature will alter the balance

between ER and GPP, thereby impacting NEE at the Alaskan fen.

24

This chapter describes the ﬁrst two years of the large-scale manipulation of soil
hydroclimate variables and the response of CO2 ﬂux components NEE, ER, and GPP and
vegetation (% cover, vascular leaf area index (LAI)). Besides testing the hypotheses
listed above, using these ﬁrst two years of data, I also investigated whether the
experimental manipulations of soil climate affected CO2 ﬂuxes while maintaining the
same fundamental relationship between ambient water table position and temperature and
CO2 ﬂux, or whether the experimental manipulations ‘pressed’ the system across a
threshold yielding new relationships between ambient water table position, temperature
and CO2 ﬂuxes (Figure 2-1). Signiﬁcant interactions between experimental treatments

and soil hydroclimate variables would provide evidence of such threshold changes.

2.2 METHODS

2.2.1 Study site

This study was conducted at the Alaska Peatland Experiment (APEX) site, which
is just outside the Bonanza Creek Experimental Forest and within the Tanana River
ﬂoodplain, approximately 35 km southeast of Fairbanks, Alaska, USA (64.82°N, l47.87°
W). The APEX site is a moderately rich fen (surface water pH 5.2-5.4), which is one of
the most common peatland types in western North America (Vitt et al. 2000). The mean
annual temperature of the area is -2.9° C with mean annual precipitation of 269 mm
(Hinzman et al. 2006). This site lacks trees and is dominated by brown moss, Sphagnum,
and emergent vascular species (Equisetum, Carex, Potentilla). There is no obvious

microtopography across the site and the peat depth is approximately 1 meter at the center

25

of the site. Seasonal ice is present until late August and there is no surface permafrost
(i.e., in the top 40 cm of organic soil) at the APEX fen. During early spring of 2005 we
established a factorial design of water table position (control, raised water table, and
lowered water table) and soil warming (control, or no warming, and passive soil

warming) treatments (Figure 2-2), which are explained in more detail below.

2.2.2 Experimental manipulation of water table

We established three 120 m2 plots and assigned each to one of three water table
treatments (raised or ﬂooded, lowered or drought, and control or no change) based on
water ﬂow. In March 2005, while soils were still frozen, we used a small excavator to
dig drainage channels to divert water from the lowered water table plot to a small holding
trench down slope. The goal of this manipulation was to lower water table position inside

the lowered water table plot by about 15-20 cm relative to the control plot, reﬂecting the
level of drying predicted for many boreal wetlands under a double CO2 concentration

atmosphere (Roulet et al. 1992). Boardwalks were constructed around each water table
plot during trench construction. In May and June 2005, solar powered bilge pumps were
installed to pump water into the raised water table plot from a surface well located about
20 m down Slope of the plot (Figure 2-2). Water was added to the raised plot at a rate of
approximately 10 cm of water/day, resulting in a 9-11 cm increase on average in the
raised water table plot. The chemistry of water additions was similar to ambient pore
water in the raised water table plot (no signiﬁcant differences in pH, electrical
conductivity, DOC, anion/cation or organic acid concentrations; data not shown). While

the raised water table treatment does not involve a dilution of pore water DOC

26

concentrations as would be expected with increased precipitation, this treatment does not
lead to major changes in pore water chemistry in the raised plot and is probably a
reasonable simulation of ﬂooding involved in wetland thermokarst formation in this
region.

The goal of the experiment was to maintain both a lowered (drought) and raised
(ﬂooded) water table treatment, without minimizing the considerable ambient variability
in water table position that occurs naturally at this site. A Campbell datalogger
communication system facilitated pumping and drainage across the raised and lowered
water table plot based on natural ﬂuctuations in water table levels in the control plot (by
turning pumps on and off). In general, water levels in the raised and lowered water table
plots tracked control plot water table changes in response to precipitation and seasonal

drying trends.

2.2.3 Experimental manipulation of temperature

Within each water table plot, we established Six 3 m2 subplots, which were
randomly assigned to one of two warming treatments; including no warming (control)
and seasonal warming (Figure 2-3). Warming treatments were thus replicated in
triplicate within each water table plot. We manipulated air and surface soil temperatures
within the warming treatment subplots using open top chambers (OTCS) during the snow-
free period following Walker et al. (2006). The OTCS were constructed out of 0.16 cm
thick Lexan, with base dimensions of 0.8 m2. Throughout the growing season of 2005

and 2006, OTCS passively warmed surface soil (2 cm beneath moss) and air temperatures

27

by about 1 °C (0.5 °C and 0.7 °C, respectively). Relative humidity and temperature was

recorded in each gas ﬂux collar during the CO2 ﬂux campaigns (data not shown).

2.2.4 Environmental variables

Beginning in June 2005, mean hourly water table level, photosynthetically active
radiation (PAR), and air and soil temperature were logged continuously with Campbell
Scientiﬁc dataloggers in each subplot over two growing seasons. Air temperature and a
depth array of soil temperatures (above moss surface and 2, 10, 25, and 50 cm beneath
moss surface) were measured within each subplot using thermistors (Y oshikawa et al.
2004) (6 arrays per water table plot, 18 total). Water table levels were measured using
one transducer (Campbell Scientiﬁc, Logan, UT) installed at the bottom of 5 cm
diameter, 1 m long PVC wells in each water table plot (1 per water table plot, 3 total),
and were calibrated against manual measurements from the same well. The spatial
variability in water table position inside each water table plot was determined with
weekly manual measurements of water table position collected from shallow wells within

each subplot during the 2005 and 2006 growing seasons.

In 2006 I experienced datalogger malfunction, which impacted the lowered and
control plot water table data. Weekly manual water table measurements, the continuous
raised water table plot data, and calculations of peat storativity were used to model
continuous water table position in the lowered (July 23 — September 2006) and control

(July 14 — September 20, 2006) water table plots.

Photosynthetically active radiation (PAR) was measured at each gas ﬂux collar

using a Li-COR (Lincoln, Nebraska) quantum light sensor (umol m2 5"). These variables

28

were logged every 5 seconds and hourly averages were recorded at each water table plot
using Campbell Scientiﬁc CRIOX dataloggers (Logan, Utah). Hourly averaged tipping
bucket precipitation was measured at a meteorological station located within the Tanana

River Floodplain.

2.2.5 Carbon dioxide exchange measurements

Net ecosystem CO2 exchange (NEE) was measured using conventional chamber

techniques following Carroll and Crill (1997). Immediately following snowmelt in 2005,
I installed permanent replicate clear Lexan chamber bases, or gas ﬂux collars, in all 18
subplots. The gas ﬂux collars were inserted approximately 10 cm into the soil, taking
care not to damage roots, and allowed to equilibrate for one week before taking the ﬁrst
ﬂux measurement. A clear plexiglass chamber constructed out of 0.3 cm thick Lexan
(area = 0.362 m2 and volume = 0.227 m3) was placed and sealed over the collars using
foam tape around the chamber base during each ﬂux campaign. An internal fan system

maintained ambient temperature and humidity conditions within the chamber.

Carbon dioxide (CO2) exchange measurements were conducted weekly

throughout the growing season from late May - early October during 2005 and 2006.
Carbon dioxide concentration inside the chamber was determined every 1.6 seconds for
2-3 minutes using a PP Systems EGM-4 portable infrared gas analyzer (IRGA;
Amesbury, Massachusetts). The IRGA was calibrated before each gas ﬂux campaign. In
2006, temperature, relative humidity, and PAR were logged continuously within the
chamber during each ﬂux measurement with a PP Systems TRP-l sensor attached to the

inside of the chamber (Amesbury, Massachusetts). I attempted to randomize time of day

29

and weather conditions among all measurements to capture full variations of light and

temperature for each collar. The CO2 ﬂux rate (umol C02 m.2 sec'l) was calculated as

the slope of the linear relationship between headspace CO2 concentration and time with r2

> 0.8. By using this goodness of ﬁt criterion 1 excluded all ﬂuxes that did not exhibit
linear change from the ﬂux data set (<3% of the data). Immediately following the NEE
measurements, the area above my gas ﬂux collars was vented by opening the chamber for
~ one minute. Immediately after this venting, I measured total ecosystem respiration
(ER) using a dark shroud that blocked all PAR from entering the gas ﬂux chambers. I
also used a series of opaque shrouds placed over the chamber to quantify NEE as a
function of light intensity (Bubier et al. 1998). This included a shroud made of 1.2 mm2
polyester mesh netting that blocked ~ 50% of PAR, as well as a shroud made of 0.5 mm2
polyester mesh netting that blocked ~ 75% of incoming PAR. Gross primary production
was determined as the difference between NEE and ER values from each gas ﬂux collar

during the same sampling period. Negative ﬂux values represent net C02 uptake from

the atmosphere.

2.2.6 Vegetation surveys

The percent cover of vascular and bryophyte species was visually estimated
within each subplot in July and August 2006. Dominant vascular species included Carex
atherodes, Carex canescens, Potentilla palustris, and Equisetum ﬂuviatile. Dominant
bryophyte species at the site include Sphagnum (Sphagnum obtusum, Sphagnum
platyphyllum) and brown moss (Hamatocaulis vernicosus, Drepanocladus aduncus)

species.

30

In 2007, vascular LAI was measured by counting all vascular plant leaves within

ﬁve 8 x 8 cm subplots systematically distributed throughout each gas ﬂux collar each
month (Wilson et al. 2007). Total leaf numbers within the CO2 gas ﬂux collars were

extrapolated from these subplots. Individuals of each species were selected outside of the
ﬂux collars at each site and their leaf area was measured biweekly using a calyper. An
average biweekly surface area of leaves was computed and multiplied by leaf numbers to

determine vascular LAI.

2.2.7 Data analysis

Treatment and h ydroclimate controls on C 0 2 ﬂuxes

The main goal of my project was to understand controls on CO2 ﬂuxes (NEE, ER,

GPP) within the water table and soil warming manipulation experiment in interior,
Alaska. To investigate whether the experimental treatments impacted ﬂuxes, I used a
repeated measures analysis of variance (ANOVA) and Tukey post hoc analysis of means
tests (Proc Mixed) in SAS 8.1 (SAS Institute Inc., Cary, NC, USA) to determine the

effects of water table treatment, soil warming treatment, year, and all interactions among
these ﬁxed effects on all three CO2 ﬂux components (NEE, ER, and GPP).

The results from the repeated measures ANOVA models provide insight into how
the treatments impacted ﬂuxes, but do not provide insight into principal controls on CO2

ﬂuxes such as water table position and soil temperature. To understand basic soil

hydroclimate controls on ER and GPP as well as whether the experiment induced

threshold changes between hydroclimate and C02 ﬂux, I used two general linear models

31

(Proc GLM) in SAS 8.1 to predict ﬂuxes: (1) a continuous soil climate variables model,
which included year, season, and the continuous soil climate variables (water table
position, air temperature, and peat temperature at 2 cm and 25 cm depth) and (2) a
treatment + continuous soil climate variables model, which included year, season,
experimental treatments (control, lowered, raised water table treatments; warmed and un-
warmed treatments), and continuous soil climate variables (water table position, air
temperature, and peat temperature at 2 cm and 25 cm depth). These two models were
compared using AIC to estimate goodness of ﬁt and the model with the smaller AIC
value was selected. The selection of Model 2 and the presence of signiﬁcant interactions
between treatment (water table or soil warming treatments) and continuous soil climate
variables was used as evidence that the hydroclimate manipulations ‘pressed’ the system

in a way that created new relationships between soil hydroclimate variables (water table

and temperature) and CO2 ﬂuxes (Figure 2-1).

Light and vegetation controls on C0; ﬂuxes

Simultaneous measurements of NEE and PAR were used to generate light-
response curves. The relationship between NEE and PAR is often represented by a
rectangular hyperbola (Thomley and Johnson 1990), where there is a near-linear increase
in productivity at low light levels and an asymptotic approach at high light levels.
Parameter estimation for the rectangular hyperbola to characterize NEE as a function of

PAR was accomplished using nonlinear regression (PROC NLIN in SAS 8.1). I ﬁt a

rectangular hyperbola to the CO2 ﬂux data, estimating three parameters: the maximum

gross photosynthetic CO2 capture at high PAR, >1000 umol m.2 sec'l (GPmax, umol CO2

32

m'2 sec-l), the photosynthetic quantum efﬁciency (a, umol C02 in.2 sec'1 per umol PAR

m'2 sec-l), and dark respiration at PAR = 0 (R, umol CO2 m"2 sec’l).

NEE = GPmax * a * PAR)
(PAR * a) + GPmax - R

This equation has been used successfully to model the relationship between NEE and
PAR in peatlands and other ecosystems (Frolking et a1. 1998, Ruimy et al. 1996,
Waddington and Roulet 1996, Whiting et al. 1992). I used a general linear model (Proc

Mixed) in SAS 8.1 to describe light response curve model parameters (GPmax, or) using a

combination of experimental treatments (control, lowered, raised water table plots;
warmed and un-warmed subplots), cumulative vascular LAI (forbs, grass functional
groups), and % vegetative cover of mosses (Sphagnum, brown moss) for each gas ﬂux

collar.

2.3 RESULTS

2.3.1 Soil climate

Interannual variation in temperature and precipitation at the Bonanza Creek LTER
Tanana River ﬂoodplain site, which is close to the APEX fen, showed warmer and wetter
conditions in 2005 than in 2006. Mean daily air temperatures (May 1 — September 30) on
average were warmer in 2005 (13.4 d: 0.1 °C) than in 2006 (12.3 :t 0.1 0C; F1, 7337 = 53.06,
p < 0.0001). The site was also wetter in 2005 than in 2006, likely due to more

precipitation received as snowfall (snow water equivalent = 120 mm in 2005; 73 mm in

33

2006). Mean annual growing season precipitation, however, did not vary between years
(F.,304= 1.61, p > 0.10).

The water table and soil temperature manipulations caused both the experimental
water table and soil temperature treatments to differ from the controls. Manipulations in
the raised water table plot raised water table levels by 9 cm and 11 cm on average relative
to the control water table plot in 2005 and 2006, respectively. The lowered water table
manipulation lowered water table levels by 5 cm and 8 cm on average relative to the
control water table plot in 2005 and 2006, respectively (Figure 2-4).

During the growing seasons of 2005 and 2006, OTCS passively warmed surface
soil (2 cm beneath moss) at the site by an average of 0.7, 0.9, and 0.6 °C in the control,
lowered, and raised plots, respectively. However, surface peat temperatures varied more
across water table treatments than between warming treatments or across sampling years.
Most notably, both surface peat (2 cm beneath the moss surface) and deeper peat (25 cm
beneath the moss surface) temperatures were consistently higher in the raised water table
plot than in the lowered or control plots (Figure 2—5). In 2005, growing season peat
temperatures (from July 14 — September 30) at 2 cm depth averaged 12.5 i 0.1 0C, 12.5 i
0.1 °C, 16.9 i 0.0 °C in the control, lowered, and raised water table plots, respectively.

These peat temperatures were Slightly lower in 2006 (data not shown).

2.3.2 Species composition and leaf area index
Percent species cover of the dominant vascular and bryophyte plant species within
each gas ﬂux collar did not vary across the experimental treatments in 2005 (Turetsky

unpublished data), and showed no signiﬁcant shifts in species composition between 2005

34

and 2006 (canonical discriminate analysis, comparison of species composition in 2005
vs. 2006 across treatments; Table 2-1). However, upon visual observation, the abundance
of brown mosses declined in the lowered water table plot and Sphagnum cover increased
from 19% to 51% in the raised water table plot from 2005 to 2006 (Table 2-1). Vascular
LAI measurements in 2007 Showed the greatest cumulative vascular LAI in the control
water table plot and the lowest in the lowered water table plot (Figure 2—6) throughout the
growing season. Average vascular LAI was greatest in the middle of the growing season

around Julian day (JD) 205 in 2007 (Figure 2—6).

2.3.3 Carbon dioxide ﬂuxes

Net ecosystem exchange of CO2 across the experimental treatments ranged from
-5.02 to 6.91 umol CO2 m-2 sec.l (control water table plot), -3.64 to 4.578 umol CO2 m'2
sec—l (lowered water table plot), -8.05 to 4.65 umol CO2 m'2 sec.1 (raised water table

plot) (negative values represent net CO2 uptake from the atmosphere, positive values

represent net emission to the atmosphere). There were no pronounced seasonal trends in
NEE in any of the three water table plots (Figure 2-9). The control water table plot in
2005 and the raised water table plot in 2006 showed the greatest variability in seasonal
NEE. The lowered water table plot showed the smallest variability in NEE across the
season in both 2005 and 2006.

Repeated measures ANOVA models show that NEE varied by a water table
treatment (control, lowered, raised) x year (2005, 2006) interaction (Table 2-2, Figure 2-

11) with no other higher level interactions. The control water table plot showed a slight

tendency toward more positive NEE values in 2006, but remained an overall sink of CO2

35

in both years during ﬂux measurement periods. The lowered water table plot switched
from negative NEE values in 2005 to positive NEE values in 2006. The raised water
table plot in 2006 generally had more negative NEE values. There was no signiﬁcant

warming treatment effect or higher level interactions involving warming on NEE.

Ecosystem respiration of CO2 (ER) ranged from 0.23 to 8.8 umol CO2 m'2 sec-l
(control water table plot), 0.15 to 10.79 umol CO2 m"2 sec'l (lowered water table plot),

0.35 to 12.32 umol C02 111'2 sec'I (raised water table plot). In 2005 in the control water

table plot, ER was generally greatest at the beginning of the growing season with
declining values over time (Figure 2-7). In 2006, ER in the control water table plot was
lowest at the beginning and end of the season, with the greatest ﬂux in the middle of the
season (~JD 163). The lowered water table plot responded similarly to the control water
table plot in both years, with ER gradually declining throughout the season in 2005 and
showing a general parabolic trend in 2006. In both 2005 and 2006, ER ﬂuxes in the
raised water table plot were lowest at the beginning and end of the season, with the
greatest ﬂux in the middle of the season (~JD 200-208).

Similar to the trends in NEE, repeated measures ANOVA models show that
ecosystem respiration (ER) varied by a water table treatment x year interaction (Table 2-
2, Figure 2-11). In 2005, BR did not vary among water table treatments but was highest
in the control water table plot. In 2006, the raised water table plot averaged higher ER
ﬂuxes than either the control or lowered water table treatment plots.

The warming treatment signiﬁcantly increased ER (F l, 5:14.18; p = 0.0131) with

no higher level interactions involving warming. The mean ER ﬂux was 3.99 d: 0.13 and

3.46 :1: 0.12 umol CO2 rn'2 sec'l within the warmed and un-warmed subplots, respectively

36

(data not shown). Gross primary production across the treatments ranged from -13.44

to -0.46 umol CO2 in.2 sec'I (control water table plot), -1 1.29 to -0.44 umol C02 m'2 sec-
I (lowered water table plot), -17.35 to -0.17 umol CO2 m-2 sec'] (raised water table plot;

negative values represent net CO2 uptake from the atmosphere). In 2005, GPP ﬂuxes in

the control water table plot were generally constant until JD 208 after which they steadily
declined to the end of the growing season (Figure 2-8). In 2006, GPP ﬂuxes in the
control plot were greatest in the early to middle of the growing season (JD 164-196) and
declined late in the season. Rates of GPP in the lowered water table plot were similar
between years, although ﬂuxes were lower at the beginning of the season during 2006.
Rates of GPP were highest in the middle of the growing season (JD 174-208) in the
lowered water table plot. In the raised water table plot, there was greater overall
variability in GPP in 2006 than in 2005. Fluxes in the raised water table plot were
greatest in the middle of the season (JD 200-207), with the lowest ﬂuxes occurring during
the early and late seasons.

Similar to NEE and ER, the repeated measures ANOVA model showed that GPP

varied by a water table x year interaction (Table 2-2, Figure 2-11). The control and
lowered water table plots both had lower GPP ﬂuxes (less CO2 taken up from the
atmosphere) in 2006 than in 2005. The raised water table plot showed higher GPP (more
CO2 taken up) in 2006 than 2005. Gross primary production showed more interannual

variation in the raised water table plot than in the control and lowered water table plots.
Also similar to ER, GPP ﬂuxes were signiﬁcantly higher in the warmed subplots

than the un-warmed subplots (Fl, 5:27.77; p = 0.0033) with no interactions among

37

warming, year, and/or water table plot. Mean GPP ﬂuxes were -4.31 :t 0.1.3 and -5.08 i
0.13 umol CO2 m"2 sec.l within the warmed and un-warmed subplots, respectively.

To understand how the balance between ER and GPP is contributing to changes in

NEE, I plotted ER versus GPP (uptake of CO2) against a 1:1 line (Figure 2-12). In 2005,

points from each water table plot cluster together near the 1:1 line, with little variability

among water table treatments. However, the control water table plot had the most points
falling below the 1:1 line, indicating an overall net uptake of CO2 from the atmosphere.

In 2006, most data collected from the lowered water table plot tends to fall above the 1:1

line, while data points from the raised water table plot tend to fall below the 1:1 line. The
lowered water table plot had positive NEE of CO2 on average in 2006 and the 1:1 plot

shows that this is likely due to low GPP (plant uptake) in the lowered plot (Figure 2-11,
Figure 2-11). In the raised water table plot, there was an increase in GPP and a slight
increase in ER, which resulted in more positive NEE (Figure 2-11). The 1:1 plot

suggests an increase in GPP, not a decrease in ER, is responsible for this shift because
there are more moles of CO2 taken up by plants for each mole respired (Figure 2-12).

From these 1:1 ER versus GPP relationships and from the repeated measures ANOVA
model results, GPP appears to be driving the differences in NEE during 2006.

To explore the effects of seasonality on these trends, I also divided these data into
early (JD 140-179), mid (JD 180-219), and late (JD 220-277) season 1:1 plots (Figure 2-
13). While the whole season 1:1 plots (Figure 2-12) suggest that GPP largely causes
differences in NEE during 2006, this trend appears to be driven by responses occurring in
the early season (Figure 2-13). The raised water table plot was much more productive

(points fall below 1:1 line) during the early season of 2006, than during the early season

38

of 2005, while mid and late seasons do not show marked differences between years (data
not shown).

Measurements of NEE at the site (across the treatments during one diurnal

experiment) ranged from -5.908 to 11.797 pmol CO2 in2 sec'1 (Figure 2-10). Night time

ER (measured between 2100 and 0200) ranged from 1.67 -— 8.83 umol CO2 in2 sec}.

This falls within the seasonal range of ER. Because the diurnal range of both NEE and
ER fell within (or very close to) the seasonal ranges, I believe I captured a full range in

the environmental parameters (light, soil temperature, water table position) at the APEX

fen during my CO2 ﬂux measurement campaigns.

2.3.4 Soil hydroclimate controls on CO2 fluxes

For ecosystem respiration (ER), the model containing year, season, and
continuous soil climate variables (Model 1) was a better-ﬁt model using AIC than the
model also containing treatment effects (Model 2). This model explained almost 50% of
the variability in log-transformed daily ER ﬂuxes (Table 2-3). In general, temperature
variables (air and peat temperature at 2 cm and 25 cm) were important predictors of ER
and all temperature variables were positively related to ER. Peat temperature (2 cm
depth) was the most important predictor, explaining 25% of the variation in ER. The
model also contained signiﬁcant season x year, season x air temperature, and year x peat
temperature (25 cm depth) interactions. Air temperature (explained 21% of the
variability in ER) exhibited a stronger positive correlation with ER in the early season
than in the mid and late seasons. Peat temperature (25 cm depth) overall was not a very

important predictor of ER, but was a stronger predictor in 2006 (explained 14% of the

39

variance) than in 2005 (explained 10% of the variance). Surprisingly, mean daily water
table level was not a signiﬁcant predictor of ER.

Similar to ER, the model including year, season, and continuous environmental
variables (Model 1), was the best model explaining log-transformed daily GPP using
AIC. This model explained almost 50% of the variability in GPP (Table 2-4). Peat
temperature (2 cm depth) explained 16% of the variation in GPP. Season x year, season
x air temperature, year x peat temperature (25 cm depth), and year x mean daily water
table position were signiﬁcant predictors of GPP. Rates of GPP did not vary
substantially across seasonal periods in 2005, but GPP ﬂuxes were highest in the mid-
season and lowest in the late-season in 2006. Air temperature, which overall was the
most important predictor of GPP explaining 21% of the variation, was a stronger
predictor of GPP in the early season than in the mid and late seasons. Increased air
temperature generally corresponded to increased GPP. Peat temperature at 25 cm depth
was a stronger predictor of GPP in 2006 (explained 19% of the variance) than in 2005
(explained 5% of the variation). Mean daily water table was a stronger predictor of GPP
in 2005 (explaining 17% of the variation) than in 2006 (explaining 5% of the variation).

Increased water table in general corresponded to increased GPP, with more plant uptake

Of C02.

2.3.5 Light and vegetation controls on CO2 ﬂuxes

Light response curves Show that there was little variability in rates of CO2
exchange among individual gas ﬂux collars within the water table plots (Table 2-5).

GPmx ranged from 3.1 in the lowered water table plot to 11.7 in the raised water table

40

plot. a ranged from 0.009 to 0.058. I used a general linear model to describe light
response curve model parameters (GPmax, or) using a combination of experimental

treatments (control, lowered, raised water table plots; warmed and un-warmed subplots),
cumulative vascular LAI (forbs, grass functional groups), and % vegetative cover of

mosses (Sphagnum, brown moss) for each gas ﬂux collar. A model containing water

table treatment and % brown moss cover explained 77% of the variability in GPmax
values (Table 2-6). The raised water table plot had the greatest GPmax, while the lowered

water table plot had the lowest values. GPmx values increased with % brown moss cover

across all water table plots. Generally, drier gas ﬂux collars tend to have lower % brown
moss cover than wetter gas ﬂux collars. The lowered water table plot had the lowest %
brown moss cover while the control water table plot had the highest cover of these
mosses (Table 2-1).

A model containing water table treatment (control, raised, lowered water table
plots), warming treatment (warmed and un-warmed subplots), cumulative vascular LAI
(forbs), and water table treatment x forb interaction explained 65% of the variability in a
values (the maximum rate of increase of NEE versus PAR; Table 2-6). Warrned subplots
had signiﬁcantly higher a values than un-warmed subplots. F orb leaf area was negatively
related to a values in the lowered water table plot (accounted for 44% of the variability in
a), but Showed no relationship with a in the control and raised water table plots. The
lowered water table plot (31.9 m2 m‘z) had lower cumulative forb LAI than either the

2

control or raised water table plots (66.7, 50.7 m2 m' , respectively).

When averaged across gas ﬂux collars within a water table plot, the light response

curves showed that the raised water table plot had larger parameter estimates (GPmax, a,

41

R) than the control and lowered water table plots (Figure 2-13). The raised water table

plot had a higher GPmam (9.83 d: 0.93) than the lowered plot (4.51 i 0.57).

2.4 DISCUSSION

2.4.1 Characterizing CO2 ﬂuxes in an Alaskan fen
The ranges in CO2 ﬂuxes (NEE, ER, GPP) measured at the control water table

plot agree well with other studies that have used static chambers to quantify CO2 ﬂux
from northern peatlands. Across studies, NEE generally has ranged from -3.95 to 7.1
umol CO2 m.2 sec'l, while GPP has ranged from -12 to -5 umol CO2 m'2 sec.l (Silvola et
al. 1996, Alm et al. 1997, Alm et al. 1999, Moore et al. 2002, Strack et al. 2006,

Wickland et al. 2006). Thus, NEE (-5.02 to 6.91 umol C02 m‘2 sec") and GPP (43.44

to -0.46 umol CO2 m'2 sec") ranges from the APEX rich fen control water table plot

were within the range of other published work, although GPP rates in this Study were
slightly higher, which led to slightly more negative NEE relative to other studies. This is
not surprising given that few studies have investigated CO2 ﬂuxes from rich fens, which
could be more productive than other peatland types due to minerotrophic conditions and
more neutral pH than Sphagnum dominated bogs and poor fens. Also, the high latitude
and the open nature (i.e., no trees) of my site contributes to high growing season light

levels, which likely leads to high growing season GPP. Rates of ER in the control plot
ranged from 0.12 to 12 umol CO2 m.2 sec'l (Figure 9). This generally agrees with

published values. For example, ER in an Alaskan forested peatland ranged from 0.28 to

42

4.72 umol CO2 m.2 sec'l (Wickland et al. 2006), in a sub-boreal poor fen ER ranged
from 5 to 12 umol CO2 m'2 sec'1 (Bubier et al. 2003), in a Finnish poor fen ER ranged
from 0.5 to 5.0 pmol CO2 in2 sec'1 (Silvola et al. 1996), and in an ombrotrophic
temperate bog ER ranged from 0.46 to 4.05 umol CO2 m'2 sec'l (Blodau et al. 2007).
While CO2 ﬂux values in the control plot generally seem to agree with

measurements collected from other northern peatlands, the range in CO2 ﬂuxes measured

in the raised water table plot showed greater ER and GPP than other studies. Despite
being located further north than most previously published work, these results suggest
that Alaskan rich fens, when provided ample water, are very productive ecosystem types.
Similarly, peak photosynthesis in an interior Alaskan permafrost collapse bog (Wickland
et al. 2006) was greater than peak photosynthesis in a sub-boreal collapse bog (Bubier et
al. 1998).

Generally, the NEE-PAR relationships developed for the APEX fen have similar

attributes to those in other peatlands (Frolking et al. 1998, Bellisario et al. 1998). The
model components of the light response curves (at, GPmax, and R) in this study fall within
a Similar range as those found in other studies. For example, studies spanning bogs and
fens have found that a ranged from 0.001 to 0.025, GPmax ranged from 2.8 to 15.1, and R
ranged from -6.1 to -1.98 (Frolking et al. 1998, Bellisario et al. 1998, Bubier et al.
2003, McNeil and Waddington 2003; Strack et al. 2006). While the a and GPmax
components observed in this study were slightly lower than those from a poor fen in New

Hampshire (where a: 0.012 to 0.019, GPmax: 10.1 to 13.8, and R: -6.1 to -2.1; Bubier et

al. 2003), they agreed well with values from a poor fen in Ontario measured by static

43

chambers (where a: 0.01 to 0.023, GPmax: 4.0 to 15.1, and R: -2.39 to -l.98; Frolking

et al. 1998) and eddy covariance (where a: 0.0252t0.002, GPmax: 11.5:t0.29, and R: -

2.39i0.1; Frolking et al. 1998).

2.4.2 Drought effects on C02 ﬂuxes

Changes in precipitation, evaporation, and drainage already have caused water
bodies in some wetland regions in interior Alaska to dry (Riordan et al. 2006) and many
northern peatlands are predicted to become drier under future climate change scenarios
(Roulet et al. 1992). Water table level is thought to serve as one of the most important
controls on CO2 ﬂuxes from peatlands (Updegraff et al. 1995), as it determines plant
community structure and function as well as the transition between oxic acrotelm and
anoxic catotelm peat. Generally, the lowering of water table levels increases ER in
peatlands due to increased oxygen diffusion into soils, which stimulates aerobic
decomposition (Moore and Knowles 1989, Moore and Dalva 1993, Silvola et al. 1996,
Nykanen et al. 1997). However, drought, or lowered water table position, has had
differential impacts on GPP depending on the plant community. Several studies have
investigated differences in NEE between wet and dry years to predict the response of

peatland C cycling to potential drought. In most cases, these studies have concluded that
under drier conditions, peatland GPP (CO2 uptake) will be reduced due to reduced rates
of photosynthesis (Alm et al. 1999, Grifﬁs et al. 2000) and that peatlands will become
sources of atmospheric CO2 because of enhanced ER (Moore and Dalva 1993, Bellisario

et al. 1998, Christensen et al. 1998, Bubier et al. 2003, Strack et al. 2006). Similar to my

44

results, experimental manipulations of water table using both bog and fen mesocosms
also found no signiﬁcant difference in ER due to water table (Updegraff et a1. 2001), but
GPP did vary with water table level (Figure 2-11; Weltzin et al. 2000), likely due to
drought stress (Alm et al. 1999, Bubier et al. 2003, Laﬂeur et al. 2003).

In addition to the experimental drought treatment, my sampling characterized
CO2 ﬂuxes across two years representing very different climates in interior Alaska
(Figure 2-4). 2006 was a drier year than 2005 due to a shallow snow pack and lower

snow water equivalent in the winter of 2005-2006, which likely led to less runoff and

lower water table positions at the fen (Figure 2-4). Differences in soil hydroclimate
between sampling years had large consequences for daily CO2 ﬂuxes, resulting in
signiﬁcant water table treatment by year interactions (Figure 2-11, Table 2-2) for NEE,

ER, and GPP. The large interannual variability observed in this study is typical of other
wetland complexes (Bubier et al. 1998, Bubier et al. 2003, Laﬂeur et al. 2003, Myers-
Smith 2007) where differences in CO2 ﬂuxes typically are driven by large ﬂuctuations in
water table level between years, which lead to dramatic changes in both respiration and
photosynthesis. In general, the trends observed in the control plot between 2005 and
2006 (lower ER and GPP in the drier year) agree with the trends between the control and
lowered water table plots (lower ER and GPP in the lowered water table plot; Figure 2-6).

The lowered water table plot switched from average negative NEE to average
positive NEE in 2006 (Figure 2-11). Similarly, Bubier et al. (2003) observed signiﬁcant

differences in NEE in bogs and fens between wet and dry years, with smaller uptake of
CO2 during the dry summer and some sites switching from a net sink to a source of CO2

between wet and dry years. In my study, mean daily NEE ﬂuxes were 89% and 200%

45

more positive (source of CO2) in the lowered water table plot than the control water table

plot in 2005 and 2006, respectively (Figure 2-1 1). The lowered water table treatment did
not signiﬁcantly impact ER ﬂuxes compared to the control water table plot. Averaged
across the warming treatments, mean daily ER ﬂuxes were 10% lower and 3% higher in
the lowered water table plot from the control water table plot in 2005 and 2006,
respectively (Figure 2-11). This is surprising given that Silvola et al. (1996) predict a 50-
100% increase in respiration with a 14 - 22 cm drop in water table. The lack of ER
response to drought could be because lowering the water table at this site exposed peat
layers that were not readily decomposable. Alternatively, microbial populations in the
surface peat layers of the lowered water table plot may have been drought stressed.
Because ER did not differ between the lowered and control water table plots (Figure 2-
11), the observed changes in NEE were due primarily to decreases in GPP in the lowered
water table plot (Figure 2-12).

In both years, GPP was signiﬁcantly lower (reduced plant CO2 uptake) in the
lowered water table treatment than the control water table treatment. Averaged across the
warming treatments, mean daily GPP ﬂuxes were 24% and 21% lower in the lowered

water table plot than the control water table plot in 2005 and 2006, respectively (Figure
2-1 1). Low rates of GPP (Figure 2-11), low vascular LAI (Figure 2-6), low GPmax values

(Figure 2-13), and low % brown moss cover (Table 2-1) in the lowered water table plot
relative to the control water table plot indicate that dry conditions reduced photosynthesis
and productivity, likely due to drought stress. This result contrasts with studies predicting
that lowered water table levels tend to favor vascular woody vegetation (Gorham 1991,

Thorrnann and Bayley 1997, Weltzin et al. 2000). Like the vascular vegetation, it is

46

likely that the photosynthetic capacity of bryophytes was reduced because of mosses’
strong dependence on tissue water content (Titus and Wagner 1984, Silvola 1990).
Strack et al. (2006) and Weltzin et al. (2000) found a decrease in bryophytes (including
Sphagnum and brown mosses) within drained sites other studies have documented a
decrease in photosynthetic capacity of Sphagnum exposed to drought, where
photosynthesis did not recover until 20 days of saturation (Moore 1989, Tuba et al. 1996,
Alm et al. 1999, McNeil and Waddington 2003). It is likely that my data is dominated by
early vegetation responses to these changes in water table level. Over time, I would
expect to ﬁnd increased success of vascular species in the lowered water table plot, which

will have an important inﬂuence of GPP, ER, and substrate quality.

2.4.3 Flooding effects on CO2 ﬂuxes

In many areas of Alaska, peatlands are becoming more saturated due to
permafrost thaw and increased runoff from surrounding uplands (Osterkamp et al. 2000).
Flooding saturates surface soils, which limits the diffusion of oxygen into the peat,
thereby limiting microbial activity and decomposition rates, and usually decreasing CO2
emissions to the atmosphere (Clymo 1983). Few studies have examined the effects of
raised water table positions or ﬂooding on CO2 ﬂuxes in the ﬁeld, but mesocosm studies
that experimentally manipulate water table levels have Shown variable responses of NEE,
ER, and GPP. Chimner and Cooper (2003) raised water table position by approximately
5 cm and measured a 42% decrease in ER in a subalpine fen, while Updegraff et a1.
(2001) found no affect of ﬂooding on ER in bog or fen mesocosms. Research on

permafrost thaw in Canadian peatlands show that ﬂooding due to thermokarst was

47

associated with a 1.6 fold increase in ER (Turetsky et al. 2002), yet organic matter
accumulation increased by 60% following permafrost thaw (Turetsky et al. 2000),
probably due to increased GPP of Sphagnum and Carex in these newly disturbed and
saturated ecosystems. Thus, ﬂooding can reduce ER by minimizing acrotelm thickness,
or can increase GPP and possibly ER through changes in plant community composition
and substrate quality.

In my study, immediately following the onset of water additions in 2005, the raised
water table treatment had 89% more positive NEE values compared to the control water

table plot (Figure 2-11). In 2005, the raised water table plot had lower ER and GPP
ﬂuxes than the control water table plot, causing the site to be less of a C02 sink (more

positive NEE values). This trend is driven by a 31% reduction in GPP (Figure 2-11),
which is likely due to initial vascular vegetation displacement or stress following the start
of the ﬂooding treatment (raised water table plot vascular LAI was lower than vascular

LAI in the control water table plot; Figure 2-6). However, the raised water table plot in
2006 had 200% more negative NEE values (i.e., more of a net CO2 sink) compared to the

control water table plot. In 2006, the raised water table plot had greater ER and GPP
ﬂuxes than the control water table plot (by 28% and 62%, respectively; Figure 2-11),

which resulted in more negative NEE values (i.e., greater sink capacity). This trend, as
well as the high GPmax values observed in the raised water table plot (Figure 2-13, Table

2-5), suggests that the vegetation community in the raised water table plot responded

positively to higher water table levels, with increased productivity and plant uptake in

2006. This is similar to other studies where the maximum GPmax at wet sites was 2.5

times greater than the maximum GPmax at drier sites (McNeil and Waddington 2003).

48

Given that GPmax was correlated with % brown moss, increased GPP in the raised water

table plot likely was dominated by changes in bryophyte cover (Table 2-1). Also, the
interannual difference in GPP at the raised plot largely was driven by early season
responses, as GPP in mid- and late-seasons did not vary between 2005 and 2006 (Figure
2-13). The fact that the change among years is driven by early season differences also
highlights the importance of moss productivity to GPP in the raised water table plot,
given that mosses begin to photosynthesize much earlier than vascular plants in these
wetlands. This increase in GPP (in 2006) occurred despite a thinner snow pack, less
runoff, and drier soil conditions than in 2005 (Figure 2-4), which likely led to reduced
soil insulation and lower soil temperatures during early spring 2006. While reduced soil
temperatures would likely inhibit early season productivity across plots, the greater %
moss cover in the raised water table plot (Table 2—1) likely buffered the plot from climate
by insulating the soils in the raised plots over the winter and early spring.

Increased ER in the raised water table plot could be due to a variety of processes,
including 1) an increase in autotrophic respiration (ER often scales positively with GPP),
2) increased microbial activity in the warm conditions present in the raised water table
plot (Figure 2-5). In the model predicting ER, peat temperature at 2 cm depth explained
25% of the variation in ER (Table 2-2). Given that temperature was the most important
predictor of ER, high ER in the raised water table plot is probably due primarily to warm

conditions at the raised plot (Bubier et al. 1998, Laﬂeur et a1. 2005), or 3) faster rates of

anaerobic CO2 production (Figure 3-9). For example, Aerts and Ludwig (1997) used

mesocosms to document high anaerobic CO2 production and consequently high ER from

49

peat under a high water table treatment compared to a lowered water table treatment.

More detail on anaerobic C mineralization is provided in Chapter 3.

2.4.4 Warming effects on CO2 ﬂuxes

Temperature is often shown to be a major control on CO2 ﬂuxes in boreal systems

(Bridgham and Richardson 1992, Moore et al. 1998, Laﬂeur et al. 2005). Warmer air and
soil temperatures stimulate microbial activity resulting in increased ER (Crill et al. 1988,
Frolking and Crill 1994, Silvola et a1. 1996) and increased GPP (Hobbie and Chapin
1998, Arﬁ et al. 1999). A meta-analysis of terrestrial warming experiments found that 2-
11 years of warming in the range of 0.3 — 6.00 C signiﬁcantly increased soil respiration
rates by 20%, net N mineralization rates by 46%, and plant productivity by 19% (Rustad
et al. 2001). Indeed, in my study, warming increased ER and GPP consistently across all

water table plots and years by 16% (mean ER ﬂux was 3.99 i 0.13 and 3.46 a: 0.12 pmol

CO2 m'2 sec.l and mean GPP ﬂux was

-4.31 i 0.13 and -5.08 d: 0.13 umol CO2 m.2 sec'l within the warmed and un-warmed

subplots, respectively). While there was no warming x water table interaction (Figure 2-
3), warming increased ER by 11% in the control water table plot, 18% in the lowered
water table plot, and 17% in the raised water table plot. Therefore, the warming
treatment had a more pronounced effect on ER in the lowered and raised water table plots
than in the control water table plot. In the lowered water table plot, drier surface moss
and soils have less insulation, which could have led to greater increase in ER than the
control water table plot. In the raised water table plot, which inherently had warmer soils

independent of the warming treatment, the warm and wet conditions in the warmed

50

subplots may have created ideal conditions for both plant and microbial activity, thereby
increasing ER through both autotrophic and heterotrophic pathways.

Because the warming treatment increased both ER and GPP, there was likely
increased C mineralization and N mineralization in the warmed subplots. Given that N is
often a key limiting nutrient in terrestrial ecosystems (Vitousek et al. 1997), increased N
availability due to high mineralization rates could have stimulated plant growth. Similar
soil warming experiments in arctic tundra sites also attributed increases in GPP to
increased N availability (Hobbie and Chapin 1998). There was no signiﬁcant warming
treatment effect or higher level interactions involving warming on NEE, likely due to
both increased GPP and ER rates within the OTC treatments, cancelling a net effect in
either direction in NEE. These results are similar to other temperature manipulations
across arctic tundra Sites, the ITEX experiment, and peatland mesocosms (Hobbie and
Chapin 1998, Arﬁ et al. 1999, Updegraff et al. 2001), where there was an increase in both

ER and GPP with warming, thereby with little inﬂuence on NEE. Surprisingly, neither
GPmax values (Table 2-6) nor vascular LAI were affected by the warming treatment, but

warming was positively correlated with a, or photosynthetic quantum efﬁciency (Table
2-6). This indicates that warming may be increasing vegetation production in a different
way than the raised water table is increasing production—for example, possible

differences in community response or C allocation to warming.

2.4.5 Modeling CO2 ﬂuxes and the potential for threshold changes

Climate change is expected to alter soil climate dynamics beyond the scope of

contemporary variability. A key question is whether our current models will be able to

51

accurately represent CO2 ﬂuxes under these changing climatic conditions, or whether
ecosystems will undergo threshold changes that create new controls on and/or new
trajectories of CO2 ﬂuxes that are not represented in current models. This experiment
allowed me to determine whether this ecosystem showed evidence of such threshold
changes to affect C02 ﬂuxes in the ﬁrst few years of experimentation (Figure 2-1). For

ecosystem respiration (ER), the model containing year, season, and continuous soil
climate variables (Model 1, Table 2-3) was a better ﬁt model than the model that also
contained treatment effects (water table or soil warming treatments). Temperature
variables were important predictors of ER, with peat temperature at 2 cm depth
explaining 25% of the variation in ER. Surprisingly, water table position was not a

signiﬁcant predictor of ER. However, water table position could still have indirect
controls on CO2 emissions via heat transfer. For example, the raised water table plot was
signiﬁcantly warmer than the control and lowered water table plots (Figure 2-5), likely a
result of heat transfer from surface water to deeper peat layers. Thus, the strong CO2 ﬂux

response to water table treatment may be more of a response to soil temperature than to
water table position, although hydrological and thermal regimes in these ecosystems
clearly are coupled.

The model containing treatment variables (Model 2) was not the best model
according to AIC and contained no signiﬁcant interactions between experimental
treatment (water table, warming) and continuous variables (soil temperature or water
table position). Model 1, or the model including year, season, and continuous
environmental variables, also was the best model explaining GPP (Table 2-4). In general,

temperature and mean daily water table position were both important controls on GPP in

52

the model, and together explain 32% of the variation. Together, these modeling exercises
suggest that the water table and warming treatments have simply extended the ambient
relationship between temperature, water table position, and CO2 ﬂuxes at this site and did
not create new relationships between soil hydroclimate and ER. However, processes
operating at longer time scales, such as changes in plant community structure and soil
organic matter quality may invoke such thresholds in future years of this experiment.

In this study, models of ER and GPP explained less than half the variation across

the growing season (Table 2-3, Table 2-4). Similar peatland chamber studies attempting
to model CO2 ﬂuxes also have had either low predictability (McNeil and Waddington

2003) or have built models that explained less than 25% of the variability in ﬂuxes
(Myers-Smith 2007). The lack of predictive power likely can be attributed to the
interactive nature of plant responses to changes in temperature, moisture, and PAR,

especially under drought stress. In contrast to ER and GPP models however, general
linear models explained 77% of the variability in GPmax values (using water table

treatment and % brown moss cover) and 65% of the variability in a values (with water

table treatment, warming treatment, and cumulative forb leaf area; Table 2-6). The large
amount of variability explained by the GPmx and (1 models relative to the poorer-ﬁtting
ER and GPP ﬂux models indicates that while I decently characterized vegetation controls
on CO2 ﬂux, 1 am not characterizing an important control on decomposition at my site.

Further investigations of nutrient levels, reduction oxidation reactions, and microbial
community structure in this ecosystem will improve our ability to effectively model C

exchange in response to variations in climate.

53

2.4.6 Conclusions
Signiﬁcant changes to growing season CO2 ﬂuxes were found following a 2.5
year ecosystem-scale manipulation of soil hydroclimate (water table position and surface

soil temperature) in an Alaskan rich fen. The lowered water table treatment did not alter

ER, but lowered GPP by 21-24%. Differences in GPP were attributed to decreased
photosynthetic uptake of CO2 by vascular vegetation and bryophytes due to drought
stress. This resulted in increased NEE (trend toward increased atmospheric emissions) in
the lowered water table plot relative to the control plot. The raised water table treatment
was less of a C02 sink (more positive NEE values) in 2005, but a greater sink in 2006
compared to the control plot. This trend was driven primarily by interannual changes in
GPP with greater GPmax values and % bryophyte cover in the raised water table plot
relative to the control water table plot in 2006. Soil warming increased both ER and
GPP by 16% resulting in no net effect of warming on NEE. Models of vegetation
components at the APEX fen (GPP, GPmax, 01) indicate that water table position and soil
temperature are both important controls on plant uptake, while temperature was a
stronger control than water table position on ER. There were no signiﬁcant interactions
with water table position and soil warming, indicating that responses to these processes
across other Alaskan peatlands will be similar to my site, despite variations in hydrology

or peat temperature.

Contrary to many hypotheses and model predictions in the literature, drought and
lowered water table levels will not trigger a large increase of CO2 to the atmosphere in

this Site, but will decrease productivity resulting in a loss of C uptake. Flooding and

54

warming are both likely to increase mineralization rates, stimulating vegetation
productivity and causing at least some peatlands to become a greater sink of atmospheric
C. Together, my results Show that studies investigating only climate change effects on
decomposition rates (i.e., most lab incubation and mesocosm studies) will miss important
vegetation controls on ecosystem-level C ﬂuxes in the context of climate change, and will

not capture the direction or magnitude of C ﬂux changes in peatlands.

55

Table 2-1. Dominant vascular and bryophyte plant species at the Alaska Peatland
Experiment (APEX) fen with average percent cover of each species 4: one standard error
in each water table plot (control, raised, lowered) in 2006.

 

 

2006
FAMILY GENUS SPECIES Average % Cover

Control Lowered Raised

. . . . 3% 2% 3%
Eqursetaceae Equzsetumﬂuvzatzle (0.03) (000) (0.03)

5% 5% 5%
Cyperaceae Carex atherodes (0.04) (0.01) (0.04)
. ‘ . 24% 25% 21%
Rosaceae Potentzlla palus Iris (0.19) (0.03) (0.06)

. . . 2% 10% 5%
Rubraceae Galzum trzfzdum (0'02) (004) (0.04)

Poaceae Calamagrostis 0% 0% 0%
canadensis (0.00) (0.00) (0.00)

S ha naceae Sphagnum obtusum 19% 29% 51%
p g Sphagnum platyphyllum (0.19) (0.17) (0.42)

Ambl ste iaceae Hamatocaulis vernicosus 59% 2% 12%
y g Drepanoclaa’us aduncus (0.48) (0.01) (0.11)

 

56

Table 2-2. Results of a repeated measures analysis of variance analyzing net ecosystem
exchange (NEE), ecosystem respiration (ER), and gross primary production (GPP) across
the experimental treatments. Signiﬁcant higher-level predictors are marked in bold (p >

 

 

0.05).
df
(numerator, F p
denominator)
Net E coszstem Exchange (NEE)
Water table treatment 2, 8 44.73 <.0001
Warming treatment 1, 5 0.11 0.76
Year 1, 5 4.07 0.10
Water table treatment x warming treatment 2, 8 1.70 0.24
Water table treatment x year 2, 8 44.77 <.000l
Warming treatment x year 1, 5 3.98 0.10
ill/ﬁg: table treatment x warmrng treatment 2, 8 0.7 6 0. 50
E coszstem Respiration (ER)
Water table treatment 2, 8 0.41 0.68
Warming treatment 1, 5 14.18 0.01
Year 1, 5 0.71 0.44
Water table treatment x warming treatment 2, 8 1.59 0.26
Water table treatment x year 2, 8 8.85 0.01
Warming treatment x year 1, 5 0.67 0.45
r:::: table treatment x warmrng treatment 2, 8 0.71 052
Gross Primagg Production (GPP)
Water table treatment 2, 8 28.64 0.0002
Warming treatment 1, 5 12.65 0.02
Year 1, 5 17.48 0.01
Water table treatment x warming treatment 2, 8 0.53 0.61
Water table treatment x year 2, 8 60.75 <.0001
Warming treatment x year 1, 5 1.54 0.27
Water table treatment x warming treatment 2, 8 2.05 0.19

x year

 

57

Table 2-3. Results of a general linear model using environmental parameters to predict
log transformed ecosystem respiration (ER) of C02. The model explained 47% of the
variation in ER of C02. Peat temperature (2 cm depth) was the most important predictor,

explaining 25% of the variation in ER. Signiﬁcant predictors (p > 0.05) are highlighted
in bold.

 

Type III

 

df SS F p

ER Model 1 15 57.45 18.14 <0.0001
Season 2 8.58 20.3 1 <0.0001
Year 1 1 .99 9.47 0.0023
Air temperature 1 5.28 25.00 <0.0001
Peat temperature (2 cm) 1 0.77 3.63 0.06
Peat temperature (25 cm) 1 1.67 7.88 0.01
Mean daily water table position 1 0.57 2.71 0.10
Season x year 2 6.96 16.50 <0.0001
Season x air temperature 2 4.44 10.52 <0.0001
Year x air temperature 1 0.08 0.35 0.55
Year x peat temperature (2 cm) 1 0.04 0.19 0.67
Year x peat temperature (25 cm) 1 1.44 6.81 0.01
Eggrirgrprean daily water table 1 0.08 0.3 5 055
Error 306 64.60

 

58

Table 2-4. Results of a general linear model using environmental parameters to predict
log transformed gross primary production (GPP) of C02. The model explained 48% of
the variation in GPP of CO2. Peat temperature (2 cm depth) alone was the most

important predictor of GPP explaining 16% of the variation. Signiﬁcant predictors (p >
0.05) are highlighted in bold.

 

 

df Tyg; III F p
GPP Model 1 15 139.72 40.25 <0.0001
Season 2 1 1.20 24.20 <0.0001
Year 1 2.41 10.40 0.0013
Air temperature 1 14.27 61 .64 <0.0001
Peat temperature (2 cm) 1 1.41 6.07 0.01
Peat temperature (25 cm) 1 2.22 9.61 0.0020
Mean daily water table position 1 8.03 34.71 <0.0001
Season x year 2 6.05 26.14 <0.0001
Season x air temperature 2 4.74 10.24 <0.0001
Year x air temperature 1 0.23 0.98 0.32
Year x peat temperature (2 cm) 1 0.08 0.36 0.55
Year x peat temperature (25 cm) 1 2.31 9.99 0.0016
Eggriiglznean darly water table 1 2.10 9.09 0.0027
Error 648 149.97

 

59

Table 2-5. Non-linear NEE-PAR curve parameter estimates for 2006. Coefﬁcients for
the CO2 exchange and light relationships were derived from equation 1 (i 1 standard
error). a is the initial slope of NEE versus PAR (also called the apparent quantum yield),
GPmax is the gross photosynthesis above light saturation, and R is the y (NEE) axis
intercept, or dark respiration.

 

Water Table Warming

 

Treatment Treatment a GP'““ R
CONTROL NON 0.0187 (0.007) 7.115 (1.106) -3.659 (0.288)
WARM 0.026 (0.013) 5.7046 (0.8162) -3.807 (0.335)
LOWERED NON 0.021 (0.014) 4.512 (0.863) -3.792 (0.297)
WARM 0.026 (0.016) 4.529 (0.788) -4.025 (0.348)
RAISED NON 0.028 (0.009) 9.958 (1.262) -4. 129 (0.357)
WARM 0.034 (0.014) 9.942 (1.429) -5.002 (0.433)

 

60

Table 2-6. Results of general linear models using vegetation indices and treatment
parameters to predict GPmax and Alpha ((1) for 2006. The two models explained 77%

and 65% ofthe variation in GPmalx and a, respectively. Signiﬁcant predictors (p > 0.05)
are highlighted in bold.

 

 

df Type 111 SS F p
QM
Model 3 86.851 14.18 0.0002
Water table treatment 2 80.198 19.64 0.0001
% brown moss cover 1 8.798 4.31 0.0583
Error 13 26.542
Alpha (a1
Model 6 0.002 3.06 0.0573
Water table treatment 2 0.001 4.27 0.0455
Warming treatment 1 0.0008 6.34 0.0305
Cumulative leaf area index (forbs) 1 0.001 11.25 0.0073
2:7:qu 3:111: giggent x cumulative leaf 2 0.001 4.61 0.0380
Error 10 0.003

 

61

RAISED RAISED
CH , CH4

LOWERED
LOWE RED

.50 .30 .10 10 30 -50 -30 -10 10 30

MEAN WATER TABLE POSITION (cm)

Figure 2-1. Schematic illustrating possible thresholds to ecosystem function induced by

climate change. This example shows the effect of water table position on methane (CH4)
emissions in an experimental raised and lowered water table plot. Here, -50 represents a
low water table position below the moss surface and 10 represents a high water table

position above the moss surface. Figure A) As expected, an increase in mean water table

position increases CH4 ﬂux to the atmosphere exponentially due to greater anaerobic
mineralization. Thus, there is no interaction between ambient water table position and

water table treatments governing CH4 emissions. This indicates no threshold changes
induced by the climate manipulation. Figure B) In this case, CH4 emissions increased

with water table position in the raised water table plot as expected. However, CH4
emissions decreased with water table drawdown in the lowered water table plot, due to an

increase in sedge production that led to more CH4 transport through the acrotelm peat via
aerenchyma tissue. This illustrates an interaction between water table position and water
table plot, indicating that the experimental manipulation ‘pressed’ the system across a

threshold creating a new relationship between water table position and CH4 production.

62

H 20 ﬂow
path I Inert shallow collar
_ to slow surface ﬂow

 

   

  
 
 
   
 

  
   

Plot 2:
Lowered

I
I/:

I Plot 3:
gRaised

 

 

 

 

— Trench channels
- - — Boardwalk

I Solar pump station
-- Tubing Shallow well

Figure 2-2. The Alaska Peatland Experiment (APEX) experimental design was initiated
in March 2005. Trenches drain water from the lowered water table plot and solar
powered pumps divert water from the shallow well to the raised water table plot.

63

CONTROL LOWERED RAISED

@C .@ @@
.@ @.
<3 <9 <30

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

—-——-
-—-——
__-—-
———
0
O
———-—

 

 

 

- - - - _ _ - - — - -
WATER TABLE PLOTS WARMING SUBPLOTS
CONTROL: No water table manipulation Gas ﬂux collar in the no

C
warming (Control) subplot

 

 

 

LOWERED: Lowered water table treatment

Gas ﬂux collar surrounded
RAISED: Raised water table treatment by an open t09 chamber
(OTC) in the warmed subplot

 

 

 

 

Figure 2-3. At the Alaska Peatland Experiment (APEX) rich fen, I established six 3 m2
subplots within each water table plot (control, lowered, raised). The subplots were
randomly assigned to one of two warming treatments, including warmed subplots
(identiﬁed with a “W”; seasonally warmed using open top chambers (OTCs; octagon
shapes)) and un-warmed subplots (identiﬁed with a “C”; control). Warming treatments
were replicated in triplicate within each water table plot. Gas ﬂuxes were taken at each
subplot on a weekly basis.

64

 

    

 

 

 

 

 

 

 

 

 

   
     

‘ 2005 l
03 .
‘ > T30
-20” . .°oo....°00.00. . .
-20
—— Control Plot *
E '40 Lowered Plot 1
ii -—- Raised Plot ~ 10 .:‘
% £0 — PreCIpItatIon _ '2
2 . i E
.o l "’
N ‘ - I:
t 20: . . [In]: . .---.|;. .J. ,. . 40 .g
g ; 2006 * g
a 0 ' .3
g : r\ - 2
a» -10 : l ‘ — 30 IL
E : \l *
a .
n j .
-30 : — 20

.L
O
._A
:.
a.
3:1...
I.
a"."
'o
I

 

 

.50 l - 10
-60 l :
1 .
-70 - - 0
Jul Aug Sep Oct
Date

Figure 2-4. Water table levels and precipitation at the manipulation plots in 2005 and
2006. Positive values denote water table position above the peat surface (inundated).
Bars represent precipitation events. Precipitation was not signiﬁcantly different between
the two study years (F=1.61, df=l,p=0.21).

65

 

: 2005

 

 

 

Control
Lowered
--- Raised

1 I 1 1 1

 

 

 

l I l
I

l 2006 '

CI!—
Idl-
dr-
cur-
-

 

Peat Temp (°C)
N
0|

N
O
L

15

103

 

 

 

 

I

120 140 160 180 200 220 240 260 280

Day of year

Figure 2-5. Mean peat temperatures at 25 cm beneath the moss surface in 2005 and 2006
across the three water table plots (corresponding to the control, lowered, raised water
table plots).

66

+Contro| - O - Lowered —A- Raised

co

 

\l

m
l

 

Average Leaf Area (m2 m'z)

 

 

 

4

3

2 s l l l l' l T
H H H H H H N N N N N N N N N
\I \l N on LO LO O O H H N N W W h
O U1 G U1 O U1 D U1 O U1 O In O U'l O

Julian Day

Figure 2-6. Average vascular leaf area index (m2 m'z) across our three water table plots
(control, lowered, raised) throughout the 2007 growing season (June 19 — August 28,

2007). Each point represents average leaf area index (averaged across species) i
standard error.

67

Ecosystem Respiration

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Control
:5 12 12
3 2005 2006
'E a . I . 3 - 3
E O : o : . O :: . : O.
o . N O . ..
3; 4 '.f’g‘is:"' 4 ..'.°i: ,1
O . 0 ' . ' O 0 . '0 o
u ‘ 0. 0
85 o o
125 175 225 275 125 175 225 275
Lowered
.17: 12 12
3 2005 . ; 2006
"E s - . a - .
g . . 8 z. E o . ‘ . . 0
a o 0 o o
v 4 O . . ‘ . . . . . .
6 O . . . ! ‘Q ' ' ' . . 4 : . . O
U . . C . . .
E o o o u o 2
125 175 225 275 125 175 225 275
Raised
<7: 12 12
a 2005 . I 2006
'E a - 3 - z :
‘2 'r i I
o O
3. 4- “3.1:: 4- : t 3: .1
N O . ..
8 . 9" I :I . I
E 0 o
125 175 225 275 125 175 225 275
Julian Day Julian Day

Figure 2-7. Growing season ecosystem respiration (ER) within each water table
treatment (control, lowered, raised) and year (2005, 2006). Each point represents a
speciﬁc measure of ER within each subplot.

68

Gross Primary Production

Control

 

2006

 

 

 

 

0..
cl
0 4 s u. a
5
w
'0
II. 2
lo
0'.
I...

 

 

 

.493 NE .25.. N8 Re

275

225

175

125

275

225

175

125

Lowe red

 

2006

"00\

1......

 

 

 

-16

 

I
2005

 

 

 

044nm

{8 NE .25.. 6o RE

275

225

175

125

275

225

175

125

Raised

2006

 

 

 

 

5
o
0 o 0
-0 2
I
a-
..O-o
col.
I!
no i I o
I. o
I
0'-
.'

 

 

 

044nm

fa N.E BE... Ncu Eu

275

225

175

125

275

225

175

125

Julian Day

Julian Day

Figure 2-8. Growing season gross primary production (GPP) within each water table
treatment (control, lowered, raised) and year (2005, 2006). Each point represents a

speciﬁc measure of GPP within each subplot.

69

NEE C0; (nmol m"z sec“)

NEE co; (umol 111‘2 sec")

NEE co. (umol m4 sec“)

Net Ecosystem Exchange

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Control
a . 8
. 2005 . 2006
4 - . ) 4 ‘ z 2
I ' ' o I :
o .- .. '7 '4. ' vii o 2 ' ' I
”Filly“. '13.; I ; I
.: . . . 0'. .4 ‘. . I
-8
125 175 225 275 125 175 225 275
Lowered
8 8
2005 2006
4 . :0 4 7 ; :
l . . I . ' O '3
‘ E I I ' u . a I!
. s'llll lllz' ° «4:1. 1':
-4 ‘i g I ' °
-8
125 175 225 275 125 175 225 275
Raised
8 8
2005 2006
4 i i O o 4 g.
0 . a I . o
0 -40 '3' L 0 ¢.--.14.¥je I3 . 7
'4 l - e 4 . E I. .8 .
-s -s ' '
125 175 225 275 125 175 225 275
Julian Day Julian Day

Figure 2-9. Growing season net ecosystem exchange (NEE) within each water table
treatment (control, lowered, raised) and year (2005, 2006). Each point represents a
speciﬁc measure of NEE within each subplot.

70

0 Control I Lowered A Raised

 

 

 

 

'33 14
9 o
as” 10 — "
E 6 3 l
3
E I 1 I I
3 2 ‘ I i
8 '2 - e F g
a _6 - ‘ ‘—
z o s.» e a; a
8 8 a 5 6
O O O
Hour

Figure 2-10. Diurnal net ecosystem exchange (NEE) of CO2 within our water table
treatment (control, lowered, raised) on August 2, 2007. Each point represents a single
NEE ﬂux measurement at a single gas ﬂux collar.

71

 

 

 

 

 

 

1
q, C
2’: 05 L
,5: ° F 15'
EE -o.5 a ab
mu
“0
32 -1.5 .2005
I."
35 -2— I [12006
2
-2.5 f
5

£3? 4 abi abc
«0 ‘
.53
WW
ea 3
53 2
0,.

O
@E 1
93
[U

 

 

O

 

abc
T

 

c 0
.2
E“? 2
33 l
a” 4.:
e -4‘
$5 a c
-§0 '6‘ d i
is
.1.
$3 '8 e
3
-10

 

 

CONTROL RAISED LOWERED

Figure 2-11. Results of a repeated measures analysis of variance model analyzing C 02
ﬂux components of net ecosystem exchange, ecosystem respiration, and gross primary
production across water table treatments and years. Data are means i one standard error
(not adjusted for model comparisons). Same letter superscripts denote non-signiﬁcant
differences from post hoc comparison of means tests.

72

Uptake (pmol C02 m2 s")

0 Control El Lowered A Raised

 

 

:5

2005
Net Release

 

 

 

 

 

 

‘ u
. . Eng 2006
o 7 ﬁg ° Net Release
in 0
U El
0 o A $3, a EU
‘0 a A n
‘A ‘ i
‘ A A 5A
A 0
Net Uptake A X p
A ‘ e
2 4 6 8 0 12- 14

Respiration (pmol CO2 m'2 s")
Figure 2-12. Ecosystem respiration (ER) versus gross primary production (GPP; uptake)
plotted against a 1:1 line for 2005 and 2006. Each data point represents a speciﬁc
measurement of C02 ER and GPP in a subplot on a particular day. Points falling above
the line represent a net release of C03 to the atmosphere (less moles of C02 taken up by
plants for each mole respired), points falling below the line represent net uptake of C02,
while the 1:1 line represents no net uptake or release of C02.

73

Uptake (umol CO2 rn'2 s")

0 Control E] Lowered A Raised

 

 

 

 

 

 

 

 

‘ Early Season 2005
t g Net Release
El
0
a O
(Epcoq) on u o
o 9 E, o
o u ”a O
o D o 8
O
0
Net Uptake
Early Season 2006
AU 5 . . Eb Net Release
A‘ An ‘ ‘ Ba
00, an
i;:A‘A 9‘ A D
‘ D
A
5‘? ‘n a
A ‘5
‘ 2
Net Uptake A ‘
2 4 6 8 10 12 14 16

Respiration (umol CO2 m'2 s")

Figure 2-13. Early season (Julian Day 140 — 179) ecosystem respiration (ER) versus
gross primary production (GPP) plotted against a 1:1 line for 2005 and 2006. Each data
point represents a speciﬁc measurement of C02 ER and GPP in a subplot on a particular
day. Points falling above the line represent a net release of C02 to the atmosphere falling
below the line represent net uptake of CO2, while the 1:1 line represents no net uptake or
release of C02.

74

 

 

NEE (umol CO2 m'2 s")

I '0- ControI-EJ- Lowered -A- Raised

'15 ' l I I r
0 500 1000 1500 2000

 

 

PAR (pmol m'2 s")

Figure 2-14. Relationship between net ecosystem exchange of C02 (NEE) and
photosynthetic active radiation (PAR) across our three water table treatments (control,
lowered, raised) during 2006. Lines are hyperbolical ﬁts through the data. Model
parameters were ﬁt using PROC NLIN in SAS 8.1.

75

References

Aerts R, Ludwig F. 1997. Water—table changes and nutritional status affect trace gas
emissions from laboratory columns of peatland soils. Soil Biology and Biochemistry 29:
1691-1698.

Alm J, Talanov A, Saarnio S, Silvola J, Ikkonen E, Aaltonen H, Nykanen H, Martikainen
PJ. 1997. Reconstruction of the carbon balance for microsites in a boreal oligotrophic
pine fen, Finland. Oecologia 110: 423-431.

Alm J, Schulman L, Silvola J, Walden J, Nykanen H, Martikainen PJ. 1999. Carbon
balance of a boreal bog during a year with an exceptionally dry summer. Ecology 80:
161-174.

Arft AM, Walker MD, Gurevitch J, Alatalo J M, Bret-Harte MS, Dale M, Diemer M,
Gugerli F, Henry GHR, Jones MH, Hollister RD, Jonsdottir IS, Laine K, Levesque E,
Marion GM, Molau U, Molgaard P, Nordenhall U, Raszhivin V, Robinson CH, Starr G,
Stenstrom A, Stenstrom M, Totland 0, Turner PL, Walker LJ, Webber PJ, Welker JM,
Wookey PA.1999. Responses of Tundra Plants to Experimental Warming: Meta-Analysis
of the International Tundra Experiment.

Ecological Monographs 69: 491-51 1.

Bellisario LM, Moore TR, Bubier JL. 1998. Net ecosystem C02 exchange in a boreal
peatland, northern Manitoba. Ecoscience 5: 534-541.

Bergman I, Lundberg P, Nilssona M. 1999. Microbial carbon mineralisation in an acid
surface peat: Effects of environmental factors in laboratory incubations. Soil Biology and
Biochemistry 31: 1867-1877.

Billings WD, Luken J0, Mortensen DA, Peterson KM. 1982. Arctic tundra: A source or
sink for atmospheric carbon dioxide in a changing environment? Oecologia 53: 7-11.

Blodau C, Basiliko N, Moore TR. 2004. Carbon turnover in peatland mesocosms exposed
to different water table levels. Biogeochemistry 67: 331-351.

Blodau C, Roulet NT, Heitmann T, Stewart H, Beer J, Laﬂeur P, Moore TR. 2007.

Belowground carbon turnover in a temperate ombrotrophic bog. Global Biogeochemical
Cycles 21: GB 1021 , doi:10.1029/2005GBOOZ659.

Bridgham SD, Richardson CJ. 1992. Mechanisms controlling soil respiration (C02 and
CH4 in southern peatlands. Soil Biology and Biochemistry 24: 1089-1099.

Bridgham SD, Megonigal JP, Keller JK, Bliss NB, Trettin C. 2006. The carbon balance
of North American wetlands. Wetlands 26: 889-916.

Bubier JL, F rolking S, Crill PM, Moore TR, Savage K, Vamer RK. 1998. Seasonal

76

patterns and controls on net ecosystem C02 exchange in a boreal peatland complex.
Global Biogeochemical Cycles 12: 703-714.

Bubier JL, Bhatia G, Moore TR, Roulet NT, Laﬂeur PM. 2003. Spatial and Temporal
Variability in Growing-Season Net Ecosystem Carbon Dioxide Exchange at a Large
Peatland in Ontario, Canada. Ecosystems 6: 353-367.

Carroll P, Crill PM. Carbon balance of a temperate poor fen. Global Biogeochemical
Cycles 11: 349-356.

Chimner RA, Cooper DJ. 2003. Inﬂuence of water table levels on C02 emissions in a

Colorado subalpine fen: an in situ microcosm study. Soil Biology and Biochemistry 35:
345-351.

Christensen TR, Jonasson S, Callaghan V, Havstrom M. 1999. On the potential C02
release from tundra soils in a changing climate. Applied Soil Ecology 11: 127-134.

Clymo RS. 1983. Peat. Ecosystems of the World, Corstanje and Reddy.
Clymo RS. 1992. Models of peat growth. Suo 43: 127-136.

Crill PM, Bartlett KB, Harriss RC, Gorham E, Verry ES. 1988. Methane ﬂux from
Minnesota peatlands. Global Biogeochemical Cycles 2: 371—384.

Damman AWH. Geographical changes in the vegetation pattern of raised bogs in the Bay
of Fundy region of Maine and New Brunswick. Plant Ecology 35: 137-151.

F rolking S, Crill PM. 1994. Climate controls on temporal variability of methane flux
from a poor fen in Southeastern New Hampshire: Measurement and modeling. Global

Biogeochemical Cycles 8: 385-397.

Frolking S, Bubier JL, Moore TR, Ball T, Bellisario LM, Bhardwaj A, Carroll P, Crill
PM, Laﬂeur PM, McCaughey JH, Roulet NT, Suyker AE, Verma SB, Waddington JM,
Whiting GJ. 1998. Relationship between ecosystem productivity and photosynthetically
active radiation for northern peatlands. Global Biogeochemical Cycles 12: 115-126.

Frolking S, Roulet NT, Moore TR, Richard PJH, Lavoie M, Muller SD. 2001. Modeling
northern peatland decomposition and peat accumulation. Ecosystems 4: 479-498.

Gignac LD, Halsey LA, Vitt DH. 2000. A bioclimatic model for the distribution of
Sphagnum-dominated peatlands in North America under present climatic conditions.
Journal of Biogeography 27: 1139-1151.

Gorham E. 1991. Northern peatlands: role in the carbon cycle and probable responses to
climatic warming. Ecological Applications 1: 182-195.

77

Grifﬁs TJ, Rouse WR, Waddington JM. 2000. Interannual variability of net ecosystem
C02 exchange at a subarctic sedge fen. Global Biogeochemical Cycles 14: 1109—1 121.

Grogan P, Chapin FS III. Initial effects of experimental warming on above- and

belowground components of net ecosystem C02 exchange in arctic tundra. 2000.
Oecologia 125: 512-520.

Hinzman, LD, Bettez ND, Bolton WR, Chapin FS, Dyurgerov MB, Fastie CL, Grifﬁth B,
Hollister RD, Hope A, Huntington HP, Jensen AM, Jia GJ, Jorgenson T, Kane DL, Klein
DR, Koﬁnas G, Lynch AH, Lloyd AH, McGuire AD, Nelson FE, Nolan M, Oechel WC,
Osterkamp TE, Racine CH, Romanovsky VE, Stone RS, Stow DA, Sturm M, Tweedie
CE, Vourlitis GL, Walker MD, Walker DA, Webber PJ, Welker JM, Winker KS,
Yoshikawa K. 2005. Evidence and implications of recent climate change in northern
Alaska and other Arctic regions. Climatic Change 72: 251-298.

Hinzman L, Viereck L, Adams PC, Romanovsky V, Yoshikawa K. 2006. Climate and
Permafrost Dynamics of the Alaskan Boreal Forest, in Alaska's Changing Boreal Forest,
edited by F. Chapin, et al., pp. 39-61, Oxford University Press, New York.

Hobbie SE, Chapin FS 111. 1998. The response of tundra plant biomass, aboveground
production, nitrogen, and C02 ﬂux to experimental warming. Ecology 79: 1526-1544.

Hogg EH, Lieffers VJ, Wein RW. 1992. Potential carbon losses from peat proﬁles:
Effects of temperature, drought cycles, and ﬁre. Ecological Applications 2: 298-306.

Keyser AR, Kimball J S, Nemani RR, Running SW. 2000. Simulating the effects of
climate change on the carbon balance of North American hi gh-latitude forests. Global
Change Biology 6: 185-195.

Laﬂeur PM, McCaughey JH, Joiner DW, Bartlett PA, Jelinski DE. 1997. Seasonal trends
in energy, water, and carbon dioxide ﬂuxes at a northern boreal wetland. Journal of
Geophysical Research 102: 29009-29020.

Laﬂeur PM, Roulet NT, Bubier JL, Frolking S, Moore TR. 2003. Interannual variability
in the peatland-atmosphere carbon dioxide exchange at an ombrotrophic bog. Global
Biogeochemical Cyles I7: 1036, doi:10.1029/2002GB001983.

Lafleur PM, Moore TR, Roulet NT, Frolking S. 2005. Ecosystem respiration in a cool
temperate bog depends on peat temperature but not water table. Ecosystems 81619-629.

Laiho R. 2006. Decomposition in peatlands: Reconciling seemingly contrasting results on
the impacts of lowered water levels. Soil Biology and Biochemistry 38: 2011-2024.

Magnuson JJ, Robertson DM, Benson BJ, Wynne RH, Livingstone DM, Arai T, Assel
RA, Barry RG, Card V, Kuusisto E, Granin NG, Prowse TD, Stewart KM, Vuglinski VS.

78

2000. Historical trends in lake and river ice cover in the Northern Hemisphere. Science
289: 1743-1746.

McGuire AD, Melillo JW, Kicklighter DW, Joyce LA 1995. Equilibrium responses of
soil carbon to climate change: empirical and process-based estimates. Journal of
Biogeography 22: 785-796.

McNeil P, Waddington JM. 2003. Moisture controls on Sphagnum growth and C02
exchange in a cutover bog. The Journal of Applied Ecology 40: 354-367.

Moore TR. 1989. Growth and net production of Sphagnum at ﬁve fen sites, subarctic
eastern Canada. Canadian Journal of Botany 67: 1203—1207.

Moore TR, Knowles R. 1989. Inﬂuence of water table levels on methane and carbon
dioxide emissions from peatland soils. Canadian Journal of Soil Science 69: 33-38.

Moore TR, Dalva M. 1993. The inﬂuence of temperature and water table position on
carbon dioxide and methane emissions from laboratory columns of peatland soils.
European Journal of Soil Science 44: 651—664.

Moore TR, Dalva M. 1997. Methane and carbon dioxide exchange potentials of peat soils
in aerobic and anaerobic laboratory incubations. Soil Biology and Biochemistry 29:
1157-114.

Moore, TR, Roulet NT, Waddington JM. 1998. Uncertainties in predicting the effect of
climatic change on the carbon cycling of Canadian peatlands. Climatic Change 40: 229—
245.

Moore TR, Bubier JL, Frolking SE, Laﬂeur PM, Roulet NT. 2002. Plant biomass and
production and C02 exchange in an ombrotrophic bog. Journal of Ecology 90: 25—36.

Moosavi SC, Crill PM. 1997. Controls on CH4 and CO2 emissions along two moisture

gradients in the Canadian boreal zone. Journal of Geophysical Research 102: 29261-
29278.

Myers-Smith I, McGuire AD, Harden JW, Chapin F S III. 2007. The inﬂuence of
disturbance on carbon exchange in a permafrost collapse and adjacent burned forest.
Journal of Geophysical Research - Biosciences 112, G04017,
doi:10.1029/2007JG000423.

Nykanen H, Alm J, Lang K, Silvola J, Martikainen PJ. 1995. Emissions of CH4, N20 and

C02 from a virgin fen and a fen drained for grassland in Finland. Journal of
Biogeography: 22: 351-357.

79

Oechel WC, Vourlitis GL, Hastings SJ, Zulueta RC, Hinzman L, Kane D. 2000.

Acclimation of ecosystem CO2 exchange in the Alaskan Arctic in response to decadal
climate warming. Nature 406: 978-981.

Orchard VA, Cook FJ, Corderoy DM. 1992. Field and laboratory studies on the
relationships between respiration and moisture for two soils of contrasting fertility status.
Pedobiologia 36: 21-33.

Osterkamp TE, Romanovsky VE. 1999. Evidence for warming and thawing of
discontinuous permafrost in Alaska. Permafrost and Periglacial Processes 10: 17—37.

Osterkamp TE, Viereck L, Shur Y, Jorgenson MT, Racine C, Doyle A, Boone RD. 2000.
Observations of thermokarst and its impact on boreal forests in Alaska, USA. Arctic
Alpine Research 32: 303—3 15.

Riordan B, Verbyla D, McGuire AD. 2006. Shrinking ponds in subarctic Alaska based on
1950-2002 remotely sensed images. Journal of Geophysical Research — Biogeosciences:
G04002, doi: 10: 1029/2005JG0001 50.

Romanovsky VE, Osterkamp TE. 1997. Thawing of the active layer on the coastal plain
of the Alaskan arctic. Permafrost and Periglacial Processes 8: 1-22.

Roulet N, Moore T, Bubier J, Laﬂeur PM. 1992. Northern fens: methane ﬂux and
climatic change. Tellus B 44: 100-105.

Rustad L, Campbell J, Marion G, Norby R, Mitchell M, Hartley A, Comelissen J,
Gurevitch J. 2001. A meta-analysis of the response of soil respiration, net nitrogen

mineralization, and aboveground plant growth to experimental ecosystem warming.
Oecologia 126: 543-562.

Schelsinger WH. 1990. Evidence from chronosequence studies for a low carbon-storage
potential of soils. Nature 348: 232-234.

Serreze MC, Walsh JE, Chapin FS, Osterkamp TE, Dyurgerov M, Romanovsky VE,
Oechel WC, Morison J, Zhang T, Barry RG. 2000. Observational evidence of recent
change in the northern high-latitude environment. Climatic Change 46: 159.

Shurpali NJ, Verma SB, Kim J, Arkebauer TJ. 1995. Carbon-dioxide exchange in a
peatland ecosystem. Journal of Geophysical Research 100: 14319-14326.

Silvola J. 1990. Combined effects of varying water content and CO2 concentration on
photosynthesis in Spagnum fuscum. Ecography 13: 224—228.

Silvola J, Ahlholm U. 1993. Effects of CO2 concentration and nutrient status on growth,

growth rhythm and biomass partitioning in a Willow, Salix phylicifolia. Oikos 67: 227-
234.

80

Silvola J, Alm J, Ahlholm U, Nykanen H, Martikainen PJ. 1996. C02 ﬂuxes from peat in
boreal mires under varying temperature and moisture conditions. The Journal of Ecology
84: 219—28.

Strack M, Waddington JM, Rochefort L, Tuittila ES. 2006. Response of vegetation and
net ecosystem carbon dioxide exchange at different peatland microforms following water
table drawdown. Journal of Geophysical Research 1 l 1: G02006,
doi:10.1029/2005JG000145.

Strack M, Waddington J M. 2007. Response of peatland carbon dioxide and methane
ﬂuxes to a water table drawdown experiment. Global Biogeochemical Cycles 21:
GB 1007, doi:10.1029/2006GB002715.

Thormann MN, Bayley SE. 1997. Aboveground plant production and nutrient content of
the vegetation in six peatlands in Alberta, Canada. Plant Ecology 131: 1-16.

Thomley JHM, Johnson IR. 1990. Plant and Crop Modelling. Oxford University Press,
Oxford.

Titus JE, Wagner DJ. 1984. Carbon balance for two Sphagnum mosses: Water balance
resolves a physiological paradox. Ecology 65: 1765-1774.

Tuba Z, Lichtenthaler HK, Csintalan Z, Nagy Z, Szente K. 1996. Loss of chlorophylls,

cessation of photosynthetic CO2 assimilation and respiration in the poikilochlorophyllous
plant Xerophyta scabrida during desiccation. Physiologia Plantarum 96: 383-388.

Turetsky MR, Wieder RK, Williams CJ, Vitt DH. 2000. Organic matter accumulation,
peat chemistry, and permafrost melting in peatlands of boreal Alberta. Ecoscience 7: 379-
392.

Turetsky MR, Wieder RK, Vitt DH. 2002. Boreal peatland C ﬂuxes under varying
permafrost regimes. Soil Biology and Biochemistry 34: 907-912.

Turetsky MR, Mack M, Harden J, Manies KL. 2005. Terrestrial carbon storage in boreal
ecosystems: causes and consequences of spatial heterogeneity, in Proceedings from the
10th Cary Conference edited by G. M. Lovett, et al.

Updegraff K, Pastor J, Bridgham SD, Johnston CA. 1995. Environmental and substrate
controls over carbon and nitrogen mineralization in northern wetlands. Ecological
Applications: 5 1 5 1 - 163.

Updegraff K, Bridgham SD, Pastor J, Weishampel P. 1998. Hysteresis in the temperature

response of carbon dioxide and methane production in peat soils. Biogeochemistry 43:
253-272.

81

Updegraff K, Bridgham SD, Pastor J, Weishampel P, Harth C. 2001. Response of CO2

and CH4 emissions from peatlands to warming and water table manipulation. Ecological
Applications 11: 31 1-326.

Van Wijk MT, Clemmensen KE, Shaver GR, Williams M, Callaghan TV, Chapin F S III,
Comelissen JHC, Gough L, Hobbie SE, Jonasson S, Lee JA, Michelsen A, Press MC,
Richardson SJ, Rueth H. 2004. Long-term ecosystem level experiments at Toolik Lake,
Alaska, and at Abisko, Northern Sweden: Generalizations and differences in ecosystem
and plant type responses to global change. Global Change Biology 10: 105—123.

Vitousek PM, Aber JD, Howarth RW, Likens GE, Matson PA, Schindler DW,
Schlesinger WH, Tilman DG. 1997. Technical Report: Human Alteration of the Global
Nitrogen Cycle: Sources and Consequences. Ecological Applications 7: 737—750.

Vitt DH, Halsey LA, Bauer IE, Campbell C. 2000. Spatial and temporal trends in carbon
storage of peatlands of continental western Canada through the Holocene. Canadian
Journal of Earth Science. 37: 683—693.

Waddington JM, Roulet NT. 1996. Atmosphere-wetland carbon exchanges: Scale

dependency of CO2 and CH4 exchange on the developmental topography of a peatland.
Global Biogeochemical Cycles 10: 233-245.

Waddington JM, Rotenberg PA, Warren F J . 2001. Peat CO2 production in a natural and
cutover peatland: Implications for restoration. Biogeochemistry 54: 115-130.

Walker MD, Wahren CH, Hollister RD, Henry GHR, Ahlquist LE, Alatalo JM, Bret-
Harte MS, Calef MP, Callaghan TV, Carroll AB, Epstein HE, Jonsdottir IS, Klein JA,
Magnusson B, Molau U, Oberbauer SF, Rewan SP, Robinson CH, Shaver GR, Suding
KN, Thompson CC, Tolvanen Totland A0, Turner PL, Tweedie CE, Webber PJ, Wookey
PA. 2006. Plant community responses to experimental warming across the tundra biome.
Proceedings of the National Academy of Science 103: 1342-1346.

Weltzin JF, Pastor J, Harth C, Bridgham SD, Updegraff K, Chapin C. 2000. Response of
bog and fen plant communities to warming and water table manipulations. Ecology 81:
3464-3478.

Weltzin JF, Harth C, Bridgham SD, Pastor J, Vonderharr M. 2001. Production and
microtopography of bog bryophytes: response to warming and water-table manipulations.
Oecologia 128: 557-565.

Wickland KP, Striegl RG, Neff JC, Sachs T. 2006. Effects of permafrost melting on CO2

and CH4 exchange of a poorly drained black spruce lowland. Journal of Geophysical
Research 1 l 1: G020] 1, doi:10.1029/2005JG000099.

82

Williams B, Silcock D, Young M. 1999. Seasonal dynamics of N in two Sphagnum moss
species and the underlying peat. Biogeochemistry 24: 285-302.

Wilson D, Alm J, Riutta T, Laine J, Byme KA, Farrell EP, Tuittila E. 2007. A high

resolution green area index for modelling the seasonal dynamics of CO2 exchange in
peatland vascular plant communities. Plant Ecology 190: 37-51.

Yoshikawa K, Hinzman LD. 2003. Shrinking thermokarst ponds and groundwater
dynamics in discontinuous permafrost near Council, Alaska. Permaﬁ'ost and Periglacial
Processes 14: 151-160.

Zoltai SC, Martikainen PJ. 1996. The role of forested peatlands in the global carbon
cycle. Nato ASI Series I 40: 47-58.

83

CHAPTER 3

Effects of an Experimental Manipulation of Water Table and Soil Temperature on
Soil Carbon Mineralization and Organic Matter Quality

3.1 INTRODUCTION

Peatlands store up to 30% of the world’s terrestrial soil carbon (C) and are located
primarily in areas expected to experience dramatic climate warming (Gorham 1991, Vitt
2006). Peat accumulates where C fixation through net primary production (NPP) at the
surface is greater than losses of C from decomposition, leaching, and/or disturbance.
Peat accumulation rates will be altered if the cold, wet conditions that maintain slow
decomposition rates and promote C storage in peatlands are altered with climate change
(Gignac et al. 2000). Given that responses of peatland C cycling to climatic change are
likely to inﬂuence atmospheric C concentrations, how peatlands and their C ﬂuxes will
respond to C cycling remains an important issue.

Air and soil temperature, water table level, and soil quality are the main controls
on carbon dioxide (CO2) and methane (CH4) emissions from peatlands (Figure 1-1;
Updegraff et al. 1995). Many studies have shown that water table level (Nykanen et al.
1995, Moore and Dalva 1993, 1997, Strack and Waddington 2007) and temperature
(Hobbie and Chapin 1998, Updegraff et al. 1995, Christensen et al. 1999, Grogan and
Chapin 2000, Rustad 2001) strongly control C mineralization rates in peatlands. Because
many northern peatlands are expected to experience warmer and drier conditions, future

climate change likely will directly impact C mineralization rates in peatlands.

84

Besides direct soil hydroclimate controls on mineralization rates, indirect controls
on mineralization processes such as substrate quality have also been shown to impact
mineralization even on a short-term basis (Figure 1-1; Keller et al. 2004). Thus,
changing climatic conditions could indirectly impact rates of C mineralization in
peatlands through changes in soil organic matter quality (Moore and Dalva 1993, Keller
et al. 2004) caused by changes in decomposition and/or vegetation structure and function.

Soil quality is a broad term that encompasses many properties of soil including
the sizes of labile C pools, turnover rates of these pools, chemistry, nutrient
concentrations, bulk density, and microbial community. In one study, soil organic matter
quality was shown to be more important than temperature in controlling rates of
mineralization in peatlands (Nadelhoffer et al. 1991). Generally, higher concentrations of
labile organic matter can result in higher C mineralization rates while higher
concentrations of recalcitrant C compounds decrease mineralization (Updegraff et al.
1995).

Shifts in vegetation productivity and community structure will impact the quality
of litter into the soil, with potential positive or negative feedbacks to heterotrophic
respiration (Hobbie and Chapin 1996, Madritch and Hunter 2003). For example, studies
have documented increased bryophyte cover in peatlands exposed to higher water tables
(Strack and Waddington 2007). An increase in bryophytes, especially Sphagnum mosses,
could lead to more recalcitrant organic matter inputs due to the high concentrations of
organochemical compounds such as phenolics and uronic acids in the tissue, which
inhibit microbial activity (Verhoeven and Liefveld 1997, Turetsky 2003). Weltzin et al.

(2000) raised soil temperature and documented an increase in woody vegetation

85

abundance, which would inﬂuence lignin inputs to soil as well as vascular root biomass
and root exudates. Overall, changes in vegetation structure toward emergent vegetation
is likely to increase labile compounds in soils (Van Wijk et al. 2004) while increases in
bryophyte or woody-dominated communities will increase the inputs of recalcitrant
organic matter to soils, with negative feedbacks to heterotrophic respiration.

As fresh litter inputs begin to decompose, soil organic matter quality becomes
more and more dominated by the signature of recalcitrant compounds. Moving down
through a peat proﬁle, concentrations of labile compounds decrease and concentrations of
recalcitrant compounds generally increase (Turetsky et al. 2000). Changing soil
hydroclimate conditions that are conducive to decomposition (warmer soils, water table
drawdown) likely will accelerate the decomposition of labile compounds preferentially to
recalcitrant materials in peat.

The goal of this chapter is to understand whether the ﬁeld manipulation of water
table and soil warming inﬂuenced soil organic matter quality and C mineralization rates.
I measured anaerobic CH4 production and anaerobic and aerobic CO2 production rates
from peat exposed to the soil climate manipulation treatments from Chapter 2. I
incubated all samples at the same moisture content and temperature. Thus, differences in
C mineralization rates (CO2 and CH4 production) among treatments likely are the result
of changes in soil quality due to the experimental manipulations of soil warming and
water table position (Chapter 2). I also quantiﬁed other soil quality properties such as %
C, % nitrogen (N), and bulk density. The C/N ratio, or the ratio of energy to nutrients, is
often used as a proxy for soil quality because it is a good indicator of the chemical

resistance of organic material to break down (quality). Bulk. density (g cm'3) is also an

86

important soil parameter for predicting C mineralization (Bridgham et al. 1998).

Lowered water tables and subsidence as well as high rates of decomposition have both

been shown to increase bulk density (Minkkinen and Laine 1998). I hypothesize and

predict that C mineralization rates will respond to manipulation of both water table

position and soil temperature at the Alaskan fen; such that:

H12

H22

H32

Given that the ﬁeld manipulation inﬂuenced vegetation structure, gross
primary production (GPP), and ecosystem respiration (ER) (Chapter 2),
my overall hypothesis is that relatively short-terrn (2.5 year) changes in

soil temperature and water table level induced by the treatments will effect
C mineralization (as CO2 and CH4 production) in peatlands through

changes in soil quality or microbial community.

Soils exposed to the lowered water table and warming treatments will
have poor soil quality, which will correspond to lower mineralization
rates, because of previous exposure to drier, warmer conditions and

greater decomposition of labile C relative to the control.

Given that higher water tables appeared to increase bryophyte cover, I
hypothesized that the raised water table treatment will lead to more
recalcitrant organic matter inputs to soils which will inhibit microbial

activity and C mineralization rates.

87

3.2 METHODS

3.2.1 Study site and experimental design

The soil used for this incubation experiment was collected from the Alaska
Peatland Experiment (APEX) site; a moderately rich fen located approximately 35 km
southeast of Fairbanks, Alaska (64.82°N, 147.87o W). The APEX site is representative
of other boreal peatlands because rich fens are one of the most common peatland types in
western boreal North America (Vitt et al. 2000). The APEX site is described in detail in

Chapter 2 (2.2.1) and Turetsky et al. (in press).

I examined laboratory C mineralization rates (as CO2 and CH4 production) in

Alaskan fen peat samples that had undergone experimental water table and soil warming
manipulations for approximately two years. Details of these experimental manipulations
are provided in Chapter 2 (2.2.2 and 2.2.3). Brieﬂy, the field experiment consists of a
factorial design of water table position (three treatments including a control, a lowered or
drained treatment, and a raised or ﬂooded treatment) and surface soil temperature (two
treatments including a control, or no warming, treatment and surface soil warming via

open top chambers) that was initiated in the winter of 2005.

3.2.2 Soil sampling and processing

In September 2007 (approximately 2.5 years after initiation of the ﬁeld
manipulations), I removed one soil core (7 cm diameter, 25 cm long) from each subplot
(2 temperature treatments x 3 water table treatments x 3 replicates=18 cores total).

Because the experiment is designed as a long-term study, there is limited space inside the

88

subplots for destructive sampling. However, to increase replication in the sampling of
the water table treatments, I extracted 2 soil extra cores from each water table treatment
(6 extra cores total) away from the established subplots (representing un-warmed soil).
Thus, in total I sampled 24 cores, 8 per water table treatment). In the ﬁeld, the cores
were cut into 5 cm increments at depths of 0-5, 5-10, 10-15, 15-20, and 20-25 cm, where
0 cm was equivalent to the top of the moss surface. Each 5 cm section was sealed in a
ziplock bag to retain moisture and frozen. Once frozen, the cores were placed in ice
chests with ice packs and shipped overnight to Michigan State University where they
were promptly placed in freezers. For this study, we used soil from the 5-10 cm depth
increment (the surface-most increment with no green moss) and the 20-25 cm depth
increment of each core. While frozen, each depth increment was subdivided into
quarters. Bulk density and ﬁeld volumetric moisture content were determined on one
quarter (a sub sample) of peat. The dry mass of the subsample was measured by oven
drying the sample at 55° C for 72 hours; dry mass values were used in bulk density
calculations. After recording the dry weight, each subsample was ground using a cyclone
sample mill (UDY Corporation; Fort Collins, CO) for homogenization, and run on an
elemental combustion system (Costech ECS 4010, Valencia, CA) for % C and N (C/N)
analysis. Peat from another quarter (separate subsample) was thawed at 2° C overnight,
large roots and sticks (greater than 0.5 cm) were removed by hand, and the peat was
homogenized gently by hand. Soils were equilibrated for 9 days at 2° C to allow for the
C mineralization pulse from dieing roots to pass.

I placed approximately 10 g of wet soil (at ﬁeld capacity) from each depth

increment into a 250 mL incubation jar (Chromatographic Specialties, Inc.; Ontario,

89

Canada). To create similar soil moisture conditions across all incubation samples, each
incubation sample was brought to a similar moisture content (40-60% representing ideal
non-limiting moisture) by adding an appropriate volume of distilled water given the bulk

density of that particular sample (Skopp 1990). Jars assigned to the aerobic headspace
treatment were ﬂushed with ambient room air (~400 ppm CO2) and sealed with lids ﬁtted

with septa. Jars assigned to the anaerobic headspace were ﬂushed continuously with N
gas (N2) in an anaerobic glove box for 10 minutes during processing. The jars were then
immediately sealed with lids ﬁtted with septa and were purged with 60 mL of N gas
followed by the removal of 60 mL of headspace; this process was repeated three times to
ensure an anaerobic headspace. All jars (96 total; n = 3 for warming treatment, n = 8 for
water table treatment, n = 3 for warming x water table treatment) were incubated in the
dark at constant room temperature conditions (approximately 22° C) for the 6 week

incubation experiment.

3.2.3 Carbon dioxide and methane production
Rates of CO2 and CH4 production were estimated during ﬁve 24-hour periods

over the 40 day experiment. The ﬁve 24-hour periods began on days 1, 3, 10, 24, and 39
of the incubation. During each period, a 10 mL headspace gas sample was taken from
each jar using syringes equipped with three way stopcocks immediately after jars were
sealed (time=0), and then every 12 hours over the 24 hour period (3 headspace
samples/24 hours). Prior to sampling, each jar was shaken by hand to mix the gases
within peat pore spaces and to release trapped gas bubbles. After headspace samples

were collected, jars were backﬁlled with 10 ml of N gas to ensure constant air

90

volume/pressure. For jars assigned to the aerobic headspace treatment, after each 24 hour
sampling period the lids were removed, samples were gently ﬂushed with a stream of
room air, and the jars were wrapped with plastic wrap to prevent moisture loss. Jars
assigned to the anaerobic headspace treatment were ﬂushed in the anaerobic glove box
with N2 at the end of each 24 hour experiment before being rescaled. Anaerobic jars also
were purged as described above prior to each 24 hour sampling period to ensure

anaerobic conditions throughout the duration of the experiment.

Each headspace sample was analyzed for CH4 and CO2 concentrations. Five mL
of each sample was analyzed for CH4 concentrations using a Varian 3800 gas
chromatograph (GC) equipped with a Haysep Q column (GC; Varian Inc., Palo Alto,
California). The GC was calibrated using four CH4 standard gases (0, 10.2, 100, 1000

ppm) before each sampling day and standardized by running a standard after

approximately every 10 samples. The remaining 5 mL of each sample was analyzed for
CO2 concentrations using a PP-system EGM-4 infrared gas analyzer (IRGA). The IRGA
was standardized using a C02 standard gas after every 10 samples.

Before assessing production rates, the measured concentrations were corrected for
headspace dilution caused by backﬁlling with N gas, corrected for standard pressure and
temperature, and multiplied by the corrected headspace volume. Overall gas production

rates were calculated from the slope of the gas concentrations over incubation time and

were divided by the dry peat weight (g) to report data on a mass basis (nmol CH4 g'l day-
I, umol CO2 g'I day'l) or were divided by sample bulk density (g cm'3) to report data on

a volume basis (nmol CH4 cm.3 dayJ, umol CO2 cm"3 day-l). Slopes with r2 values less

91

than 0.75 were discarded (representing less than 2% of the CO2 data and 5% of the CH4

data).

3.2.4 Statistical analysis

The ﬁrst round of measurements (day 0) was excluded from all analysis due to
possible disturbance. The overall goals of my statistical analyses were 1) to explore
treatment (soil warming and water table treatments) effects on C mineralization rates, and
2) to assess the correlations between C mineralization and indices of organic matter
quality (%C, %N, C/N, bulk density).

To investigate whether the manipulation experiment impacted CO2 and CH4
production rates, I used a repeated measures analysis of variance (ANOVA) and Tukey
post hoc analysis of means tests (Proc Mixed) in SAS 8.1 (SAS Institute Inc., Cary, NC,

USA) to determine the impact of water table treatment, soil warming treatment, depth,

and all interactions among these ﬁxed effects on CH4 and CO2 production rates

To understand basic organic matter quality controls on CH4 and CO2 production

rates, I used a general linear model (Proc GLM) in SAS 8.1 to explain production rates.
This model included %N, %C, C/N, bulk density, depth (10 cm, 25 cm), incubation
treatment (aerobic, anaerobic) and experimental treatment (control, lowered, raised water

table treatments; warmed and un-warmed treatments) and interactions among these

factors as predictors of CO2 or CH4 production.

92

3.3 RESULTS

3.3.] Carbon dioxide and methane production rates

Over the 40 day incubation experiment, there was no aerobic production of CH4.
Anaerobic CH4 production from the incubations ranged from 0 to 47.79 nmol g'I day'l
and 0 to 3.96 nmol cm'3 day-1 on a per mass and per volume basis, respectively. At both
depths and across water table treatments, CH4 production rates were either constant or
increased slightly over the duration of the 40 day incubation experiment (Figure 3-1).

Repeated measures ANOVA models showed that anaerobic CH4 production rates
on a mass basis (nmol g.l day-l) varied by depth (Table 3-1). Rates of CH4 production
from the 5-10 cm depth increment were faster than the 20-25 cm depth increment (2.0 i
0.1 and 1.6 i 0.1 nmol g.l day'1 in the 5-10 and 20-25 cm depth increments,
respectively). On a volume basis (nmol cm'3 day'l), there were no differences in CH4

production rates across depths, warming, or water table treatments (Table 3-1). When

data were analyzed as a cumulative production rate (i.e., nmol g'I 40 day'1 or nmol cm'3

40 day‘l), results were similar to those described above, with a depth effect for CH4

production on a mass basis and no difference among treatments for CH4 production on a
volume basis.

Aerobic CO2 production rates ranged from 0.93 to 26.66 umol g'l day'1 and 0.09 to
2.89 umol cm'3 day'1 on a mass and volume basis, respectively. At both depths and

across water table treatments, aerobic CO2 production rates increased between day 2 and

93

day 10, and then generally decreased over the duration of the 40 day incubation
experiment (Figure 3-2). Anaerobic CO2 production rates ranged from 0 to 71.61 nmol
g'I day'1 and 0 to 5.19 umol cm.3 day'l on a mass and volume basis, respectively.
Anaerobic CO2 production rates increased between day 2 and day 10, and then remained
constant over the duration of the 40 day incubation experiment (Figure 3-2).

A repeated measures ANOVA model showed that CO2 production on a mass basis
(umol g'l day-l) varied by an incubation treatment (aerobic, anaerobic) x depth
interaction (Table 3-2). For both depths, anaerobic CO2 production rates were

signiﬁcantly faster than aerobic production rates (Figure 3-3). The 5-10 cm depth

increment had faster anaerobic CO2 production rates than the 20-25 cm increment, while
there was no difference between depths in aerobic CO2 production rates (Figure 3-3).
A repeated measures ANOVA model showed that CO2 production on a volume basis

(umol cm'3 day-l) varied by treatment. Aerobic CO2 production rates were slower than
anaerobic production rates (means i one standard error; 0.926 i 0.049 and 1.089 at 0.061
umol cm'3 day'1 in the aerobic and anaerobic treatments, respectively; Table 3-2). Rates

of CO2 production on a volume basis also varied by a depth x soil warming treatment

interaction (Table 3-2). For both depth increments, CO2 production rates from peat

collected within the soil warming subplots were signiﬁcantly slower than from peat taken

from the un-warmed subplots (Figure 3-4). In the warmed treatment, peat from the 5-10
cm depth increment had faster C02 production rates than deeper peat (20-25 cm depth

increment), while there was no difference between depths in peat from the un-warmed

94

subplots. Carbon dioxide production rates on a volume basis also varied by a water table

treatment by depth interaction (Table 3-2). At 5-10 cm depth, peat collected from the
lowered water table treatment had the slowest CO2 production rates, while peat from the

same depth interval collected from the control and raised water table treatments had the

fastest. Conversely, for the 20-25 cm depth increment, the lowered water table treatment
peat had the fastest CO2 production rates, while peat taken from the control water table
treatment had the slowest (Figure 3-5). Therefore, peat at the 5-10 cm depth increment
corresponded to faster C02 production rates in the control and raised water table
treatments than deeper peat, but showed the opposite pattern in the lowered water table
treatment, with slower rates of C02 production in the 5-10 cm depth increment than in
the 20-25 cm depth increment.

Across depths and treatments, the ratio of CO2 to CH4 production ranged from 297 to
6620. Ratios of CO2. to CH4 varied by a water table treatment x depth interaction (F 2, 72 =

4.25, p = 0.02). Across water table treatments, ratios were always higher in the 5-10 cm
increment than in the 20-25 cm increment. However, this was particularly true for the

lowered water table plot where the 5-10 cm increment corresponded to very high ratios.

To investigate whether CO2 inhibition caused greater anaerobic CO2 production
than aerobic CO2 production, temporal development of aerobic CO2 production rates was
modeled using linear regression. In the aerobic jars, the C02 concentration at time zero

explained 10% of the variation in relative CO2 production rates, while day of incubation
explained 13% of the variation. The residuals of this linear regression were examined

and C02 concentration at time zero explained less than 13% of the variation in residuals

95

for both aerobic and anaerobic CO2 production rates. This suggests that CO2 inhibition is

not an important factor controlling differences in aerobic and anaerobic CO2 production

rates.

3.3.2 Soil physical and chemical variables

Across the incubation samples, bulk density ranged from 0.049 g cm'3 to 0.199 g
cm'3. Bulk density varied by a warming treatment x depth interaction (Table 3-3).
Generally, peat from the 20-25 cm depth increment had a higher bulk density than peat
from the 5-10 cm depth increment though this was less obvious in the warmed subplots in
the lowered water table treatment than in the other treatments.

Carbon concentrations ranged from 35.1 to 45.7%. Percent (%) C data varied by
a depth effect alone, and peat from the 20-25 cm depth increment had greater %C than
from the 0-10 cm depth (Table 3-3). Nitrogen concentrations ranged from 1.2 to 3.1%
across the ﬁeld treatment and varied by a signiﬁcant water table treatment x warming
treatment x depth interaction (Table 3-3). This interaction is complex, but generally peat
at the 5-10 cm depth increment taken from the raised water table treatment had the
highest %N, while peat within the 20-25 cm depth increment taken from the un-warmed
subplots in the control and lowered water table treatments had the lowest. C/N ratios
ranged from 0.028 to 0.079 across the site and varied by a depth effect alone. C/N was
signiﬁcantly higher in the 20-25 cm depth increment than in the 5-10 cm depth increment

(Table 3-3).

In general, neither CO2 nor CH4 production rates (umol g'l day.1 or nmol cm-3

day'l) showed relationships with chemical variables (Figure 3-7). General linear models

96

investigating controls on CO2 production rates showed that soil chemical parameters

(%C, %N, BD, C/N) and interactions between these soil chemical parameters and the

water table or warming treatments were not signiﬁcant controls on C production rates.

3.4 DISCUSSION

3.4.1 Characterizing carbon mineralization in an Alaskan rich fen

Rates of anaerobic CH4 production measured in my incubation experiment, which

averaged 0 to 47.79 nmol g'] day'I across all treatments, are on the low end but tend to
fall within the range documented in other peatland incubation experiments (0-51,000
nmol g'l day°l; Yavitt et al. 1987, Magnusson 1993, Moore and Dalva 1997, Waddington
et al. 2001, Keller et al. 2004, Glatzel et al. 2004, Turetsky and Ripley 2005; Table 3-3).

Peat incubation studies generally report a wide range of production rates, likely due to

different experimental methods (temperatures, duration), but also due to differences in
microbial abundance, peat characteristics, and substrate quality (Table 3-4). Field CH4
ﬂuxes at the APEX site, where these soils were collected, also were low compared to
other studies (Turetsky et al. in press). There are several possible explanations for why
CH4 production and emissions are low at the APEX site, including 1) high CH4 oxidation
rates in surface peat, 2) acidic conditions that inhibit methanogenesis (Bergman et al.
1998), 3) large water table ﬂuctuations that cause short-duration anaerobic conditions,

which lower methanogen abundance, 4) an inadequate supply of C substrates (Y avitt et

al. 1997), and 5) redox conditions that favor alternate anaerobic mineralization pathways.

97

Since CH4 production rates were low both in the ﬁeld and laboratory incubations; these
results are not likely a direct result of climate, but due to poor substrate quality or low
methanogen abundance. However, it is likely that climate may indirectly impact CH4
production by altering substrate quality and/or methanogen population. Also the fact that
CH4 production rates from the 5-10 cm depth increments (where C quality tends to be

better due to proximity to fresh organic matter inputs) were faster than production rates

from the 20-25 cm depths points to poor C quality as the most plausible cause for the

overall low CH4 production rates.

Generally, rates of aerobic CO2 production from this study (0.93 to 26.66 umol g-l
day'l) coincide with other peat microcosm incubations (including bog and fen peat from
an array of depths; Table 3-4). My anaerobic CO2 production ranged from 0 to 71.61
nmol g'I day], which is substantially higher than most peat incubations (Table 3-4).
Surprisingly, anaerobic CO2 production was greater than aerobic production across all

climate manipulation treatments (Figure 3-3). Aerobic : anaerobic CO2 ratios ranged
from 0.28:1 — 0.84:1, which are extremely low compared to other peatland studies. Most
studies report aerobic: anaerobic CO2 production ratios within the range of 2.5 :1 to 4.3:1

(Bridgham and Richardson 1992, Johnson et al. 1990, Updegraff et al. 1995, Bergman et

al. 1999, Waddington et al. 2001, Glatzel et al. 2004). Post-incubation leachates from
this experiment were analyzed on a Dionex ICS 2000 and show lower N03 (mg L'l)
concentrations in the aerobic treatment compared to the anaerobic treatment (Figure 3-8).

Therefore, the high anaerobic CO2 production is likely a result of denitriﬁcation, or the

98

mineralization of organic N, which reduces nitrate and produces CO2. There was no

difference in sulfate concentrations (804, mg L'l) in post-incubation leachates from the
aerobic and anaerobic treatments; therefore, sulfate reduction is not likely the cause of

high anaerobic CO2 production.

Similar to other peat incubation experiments, CH4 and CO2 production rates generally

decreased with depth (Figure 3-4 and Figure 3-5; Yavitt et al. 1987, Bridgham and

Richardson 1992, Nadelhoffer et al. 1991, McKenzie et al. 1998, Waddington et al. 1998,

Glatzel et al. 2004). Shallow (5-10 cm depth increment) anaerobic CH4 production rates

(nmol g'l day!) and CO2 production (umol g'l day']) were 23% and 28% greater than

production rates from deeper peat (20-25 cm depth increment). This is likely because
older soils found in deeper peat layers contain higher proportions of recalcitrant C than
younger soils at the surface (Yavitt et al. 1987, Nadelhoffer et al. 1991, Hogg et al. 1992,
Christensen et al. 1999). The decrease in production with depth can also be explained by

the accumulation of compounds unfavorable to microbial activity such as lignin,

phenolic, or humic substances (Hogg et al. 1992). Surprisingly, aerobic CO2 production
(umol g'] day'l) did not signiﬁcantly vary with depth (Figure 3-3), suggesting that

microbial groups responsible for anaerobic mineralization are more responsive to changes
in substrate quality and/or changes in soil hydroclimate occurring with depth compare to
aerobic microbes.

The soil physical and chemical variables (bulk density, %C, %N, and C/N) did

vary across the experimental treatments (Table 3-3), but did not exhibit any linear

relationships with CO2 or CH4 production (Figure 3-7). This lack of correlation is

99

surprising because many studies have found strong correlations between these indices

and rates of C mineralization. For example, Bridgham et al. (1998) found a strong
positive correlation between CO2 production and bulk density (bulk density explained
75-80% of production). However, Moore and Dalva (1997) explained only 20-25% of
the variation in CO2 production using botanical origin of peat material, pH, depth, and

water table position.

3.4.2 Soil hydroclimate effects on CH4 and CO2 production rates

All peat was incubated under uniform temperature and moisture conditions;
therefore any differences in CH4 or CO2 production rates were the result of indirect
effects of the manipulations (water table and soil temperature) on soil quality or
microbial populations during the previous 2.5 years. There was no difference in CH4
production in response to the water table or soil warming ﬁeld manipulations. Contrary
to my ﬁndings, Keller et al. (2004) found a 28% increase in CH4 production rates from
fen peat exposed to high water tables than from peat exposed to intermediate water
tables. However, other studies have shown that CH4 production rates were not altered by
exposure to differing water tables in bogs or high soil temperatures in both bogs and fens
(Glatzel et al. 2004, Keller et al. 2004). As discussed previously, CH4 production was

very low across all of my samples. Turetsky et al. (in press) found increased methanogen

concentrations in the APEX fen warming and ﬂooding treatments, with strong
correlations between methanogen abundance and average CH4 emissions. My data

suggest that these trends in ﬁeld emissions are not due to soil substrate quality conditions.

100

However, redox conditions not conducive to methanogens are the likely cause of low

CH4 production rates and lack of signiﬁcant differences across treatments.
Drought and CO2 production: The lowered water table treatment impacted

aerobic and anaerobic CO2 production in the incubations (Figure 3-5). Compared to peat
collected within the control water table treatment, peat taken from the lowered water

table treatment had slower CO2 production rates at shallow depths (5-10 cm increments)

and faster CO2 production rates deeper into the peat (decreased by 34% and increased by
36%, in the shallow and deep increments, respectively). Other studies have found no
difference in CO2 production rates from peat collected from lowered and control water

table treatments (Keller et al. 2004, Glatzel et al. 2004). Given that I found no net effect
of the lowered water table treatment on ER and a decrease in GPP in the ﬁeld (Figure 2-
10), it is possible that lowering the water table position increased heterotrophic
respiration and decreased autotrophic respiration, resulting in no effect on total ER. If
indeed there was an increase in heterotrophic respiration in the lowered water table
treatment, my incubation results likely reﬂect a shift in soil quality with greater C
mineralization towards more recalcitrant compounds. Another explanation for low
surface mineralization rates from the lowered plot could be moisture limitation on the

microbial community, which in other studies have resulted in low microbial abundance in
the lab (Silvola and Ahlholm 1989). Faster CO2 production rates in the 20-25 cm depth

increment in the lowered water table could be due to either 1) increased ﬁne root turnover
due to vegetation shifting resources from above to belowground with drought stress or 2)

more ideal moisture conditions at depth, which could result in greater microbial activity.

101

There were no obvious effects of the lowered water table treatment on %C, %N, C/N, or
bulk density, with the exception of generally lower %N in surface peat samples from the
lowered water table plot relative to the control and raised water table plots (Table 3-3).
This coincides with low C mineralization rates measured from surface peat from the
lowered water table plot and is likely due to reduced vegetation inputs because of drought
stress (Figure 3-5).

Flooding: The raised water table treatment did not have a consistent effect on

CO2 production rates (Figure 3-5). In surface peat (5-10 cm depth), there was no
difference between CO2 production rates from peat taken from the control and raised

water table treatments (Figure 3-5). In deeper peat (20-25 cm depth), CO2 production
rates from the raised water table treatment were 23% greater than C production rates from
control water table peat (Figure 3-5). This is similar to other studies where peat exposed
to high water tables or taken from wetter sites had 10-52% greater CO2 production
compared to peat exposed to lower water tables (Waddington et al. 2001, Glatzel et al.
2004, Keller et al. 2004).

In the ﬁeld, the raised water table plot had higher ER than the control water table
plot (Figure 2-10), likely due to a combination of higher temperatures in the raised water
table treatment increasing mineralization (Figure 2-4), high anaerobic CO2 production,
and increased autotrophic respiration (coinciding with high GPP in this plot). Increased
bryophyte cover in the raised water table treatment (Weltzin et al. 2007) probably led to
more recalcitrant organic matter inputs, which was expected to inhibit microbial activity
at the surface. However, the warm conditions in the raised water table plot likely

stimulated organic matter turnover of these relatively fresh inputs. Thus, the effects of

102

the soil microclimate and vegetation inputs could have led to opposite inﬂuences on

organic matter quality and laboratory C mineralization rates. In deeper peat collected
from the raised water table treatment, CO2 mineralization rates were faster relative to

deep peat in the other treatments (Figure 3-5). This could be due to greater plant inputs
to the deeper soil horizons via belowground NPP or root exudates in the ﬂooding

conditions.
Warming and CO2 production: Temperature has been found to impact many

biogeochemical processes such as soil respiration (Grogan and Chapin 2000, Updegraff

et al. 2001, Chapter 1), litter decomposition (Meentemeyer 1978, Jansson and Burg 1985,

Hobbie 1996), N mineralization (MacDonald et al. 1995), CH4 emission (Crill 1991,

Turetsky et al. in press), ﬁne root dynamics (Gill and Jackson 2000), and plant
productivity (Hobbie and Chapin 1998). Warming has been shown to affect
mineralization rates in the ﬁeld by stimulating microbial activity in soils, which increases
organic matter breakdown and results in lower quality peat (Grogan and Chapin 2000).

This would result in lower amounts of labile C substrates in the warmed subplots
compared to un-warmed subplots resulting in lower CO2 production. However, ﬁeld soil

temperature manipulation treatments similar to Chapter 2 found that warming modiﬁed
soil quality (Keller et a1. 2004), likely due to differences in plant productivity. A change
in vegetation productivity and/or C allocation with warming could enhance the
production of labile C substrates (moss DOC ‘leakage’ or root exudation). While this
would likely stimulate C mineralization, the addition of labile C also can trigger a
priming effect, stimulating the decomposition of more recalcitrant compounds in the ﬁeld

(Weltzin et al. 2000). Differential effects of warming on vegetation types could lead to

103

an increase or decrease in select species (Weltzin et a1. 2000) and potentially impact the
quality of plant inputs into the soil. Hobbie (1996) also found that changes in plant
community structure and composition in response to warming in tundra wetlands had
important consequences on C mineralization, due to differences in litter decomposition

rates.
The warmed subplots were associated with 26% and 45% lower CO2 production

rates compared to peat from the un-warmed subplots in shallow and deep peat,

respectively (Figure 3-4). Therefore, it appears that the soil warming treatment lowered
the substrate quality of peat, which resulted in lower aerobic and anaerobic CO2

mineralization rates in the lab incubations. Despite greater inputs of fresh organic
matter to surface soils with higher GPP with warming, faster organic matter turnover in
the ﬁeld likely reduced the organic matter quality signature in the warmed subplots,
which then affected C mineralization rates. Generally, surface peat collected from the
warmed subplots had higher bulk density than peat collected from un-warmed subplots
(Table 3-3). This is similar to other studies that have found increased bulk density
associated with high rates of mineralization rates and collapse of pore structure in the

peat matrix (Minkinnen and Laine 1998).

3.4.3 Conclusions
To investigate the effects of the ecosystem-scale in-situ climate manipulations on
C mineralization through changes in soil organic matter (SOM) quality, I collected peat

from the experimental ﬁeld treatments and incubated all samples under standard

laboratory conditions. Rates of CH4 production were not affected by the manipulations,

104

suggesting that substrate quality or other aspects of redox reactions are limiting

methanogenesis. Peat taken from the warmed subplots and surface peat from the lowered
water table plot had the lowest CO2 production. rates, indicating a decrease in soil organic

matter quality likely due to increased heterotrophic respiration in these climate
treatments. This suggests that the changes in ER measured in Chapter 2 likely involve
some component of changing heterotrophic respiration.

My results suggest that relatively short-term changes in climate can alter peatland
C ﬂux and result in potential feedbacks to climate change through altered soil organic
matter quality. Water table position has complex affects on soil organic matter quality
and microbial community. Because I found differential effect of the water table
treatments on C mineralization at different depths, is likely that microbial metabolism is
interacting with vegetation inputs across the manipulations to control soil quality at this
site. For example, warming increased both ER and GPP (Chapter 2) while lowering the
soil organic matter quality (Figure 2-11). Future monitoring of this experiment will be
able to clarify whether poor substrate quality driven by faster organic matter turnover in
the ﬁeld eventually will serve as a negative feedback to decomposition and plant

production (through decreased nutrient availability).

105

Table 3-1. Results of two repeated measures analysis of variance (ANOVA) models

. . -l -l -1 -1 .
analyzmg CH4 production as nmol g day and nmol cm day across our experimental
treatments. Signiﬁcant higher-level effects are marked in bold (p > 0.05).

 

 

Model effects (numgi'ator, F p
denominator)
CH4 Production (nmol g'l day-l)
Plot 2, 36 0.76 0.47
Warm 1, 36 0.07 0.79
Depth 1, 36 5.05 0.03
Plot x Warm 2, 36 0.60 0.55
Plot x Depth 2, 36 0.97 0.39
Depth x Warm l, 36 0.00 0.98
Plot x Depth x Warm 2, 36 0.21 0.81
CH4 Production (nmol cm'3 day'l)
Plot 2, 36 0.78 0.47
Warm l, 36 0.52 0.47
Depth 1, 36 0.08 0.78
Plot x Warm 2, 36 2.01 0.15
Plot x Depth 2, 36 0.49 0.62
Depth x Warm l, 36 1.13 0.29
Plot x Depth x Warm 2, 36 0.50 0.61

 

106

Table 3-2. Results of two repeated measures analysis of variance (AN OVA) models

analyzing CO2 production as umol g'1 day'1 and nmol cm" day'l across our experimental
treatments. Signiﬁcant higher-level effects are marked in bold (p > 0.05).

 

df
Model effects ' (numerator, F p
denominator)

 

CO2 Production (nmol g'l day'l)

Plot 2, 72 2.52 0.09
Warm 1, 72 0.68 0.41
Depth 1, 72 188.18 <.000l
Treatment 1, 72 11.22 0.0013
Plot x Warm 2, 72 2.89 0.06
Plot x Depth 2, 72 1.95 0.15
Plot x Treatment 2, 72 1.70 0.19
Depth x Warm l, 72 0.06 0.81
Depth x Treatment 1, 72 7.92 0.01
Treatment x Warm l, 72 0.16 0.69
Plot x Depth x Warm 2, 72 0.90 0.41
Plot x Warm x Treatment 2, 72 0.41 0.67
Plot x Depth x Treatment 2, 72 0.71 0.49
Depth x Warm x Treatment 1, 72 0.01 0.93
Plot x Depth x Warm x Treatment 2, 72 0.27 0.76
CO2 Production (umol cm'3 day")

Plot 2, 72 1.14 0.32
Warm 1, 72 0.01 0.92
Depth 1, 72 43.35 <0.0001
Treatment 1, 72 5.23 0.03
Plot x Warm 2, 72 0.09 0.91
Plot x Depth 2, 72 4.69 0.01
Plot x Treatment 2, 72 0.80 0.45
Depth x Warm 1, 72 4.45 0.04
Depth x Treatment 1, 72 2.76 0.10
Treatment x Warm 1, 72 0.35 0.56
Plot x Depth x Warm 2, 72 2.93 0.06
Plot x Warm x Treatment 2, 72 0.09 0.91
Plot x Depth x Treatment 2, 72 0.16 0.85
Depth x Warm x Treatment 1, 72 0.00 0.99
Plot x Depth x Warm x Treatment 2, 72 0.06 0.94

 

107

 

 

 

 

 

 

 

 

 

N2: .83 A83 32: A33 :2: :03 88.0 82: A33 :2: see 26
.53 .290 .33 .23 .Neod .Neod .390 .33 .486 .890 $.86 .83
so:

69: 96.8 A38 awe No: 858 A89 808 698 84.8 9N9 u .\

. . . . . . . . n . . . o
.-._ z. .N x. .L 9. .8 3. a? 9. 3N :. .5 mm ..o Na. .2 me. am am am 3. .3 9.
9.0.8 3.8 A89 A38 awe $98 8N3 3N9 No.8 2 es 698 898 z .\
a; N .8 N a; N .2. N .8 N .3 N .5 N .NN N .8 N .: N .3 N .3 N

:5 w

82: 2.03 ANNE A83 A33 53 2 Se 8:: :N3 A83 53 :98 Wash
.02 o ..8 o .mN_ o .NS 0 .8o o a: S .2: c .as o .8. o 28 o .5 o .as o ism
Eu mN Eu A: Eu MN Eu c— 50 mN Eu A: :5 mN :5 A: 50 mN Eu A: :5 mN :5 =—

“5853 .5853-.. D 3:555 3853.: D .5553 58.53-: D

«Sam. .533 635% 53.; .5239 553:1— oEgh. 523$ 3.5.50

 

 

 

555553 E e85 Bates...

25 5:3 282: 2a Sea Amodv 3 bag—mama 5&6 Ho: ca 35835 5:2 583 2: .23 3.52 dﬂ Xmm< 2: 8 35.555
w£§3 95 058 553 55 £88 35230 35a 5% 20 95 ADV cognac Ax. AZV 5on: ..\o 5-25 3 bidet ﬁsm .num 535.

108

Tam: N-Eo BE: N 7.3: $80 BE: _

 

 

 

 

 

 

 

 

 

 

. 8.3 . - :5 3.3 :5 on
an S. 8 o 3 mod 5 K 9 o ow 8:568 goo ow \ U can ”5.: not b55532 35:5 £5.
:5 ow :5 .3 .m ma?
Wové 8.3-35 2. Too; NM: 2:558 22.: \ U .2 25.: 552552
coon— 55580: Eaemcaum.
. . . 8:568 5E8: 55a 5%.: 32 «>5Q
mu m3 mm mm cm 2 m 25m \ 0 com 5 2 25m 95 wom 95 5522
5555 \ 0 .2 5505.... :5 on E5 .3 .m .o : com .5
- cod—$20 5% I _N.o N 5:55 mmwon
. . . a 8558.3
\ 0 com 2 w 5505.25. :8on2 55 5535:
8N5: .29.... 8&5 EON
. cwdmLad 3.0 I mmd on 5:55 \ 0 com mm on .
coo G I o 5552 95 5535: 5 5 555—0
. . . 5388 ooﬂBm $2
coo m _ I 2: - mm 2 I S N S 5:55 \ U omamﬁ ”5.: 55580: Ezzmgaw. .5 5 E2;
. . . . 8585.382 .8 25 SON
Nmm Saw 3: Lo HNS o - E. 8:568 22.: \ 0 cm _ 25m .5 5 55M
. . . . 8255.352 .8 m3 EON
Nmo cum: o _mo 52o o - P 8:558 22.: \ 0 on _ “mom .5 5 55M
. . . . - 2:568 5383 85.53 moom 53E
n wmjnm omm cm mom we. oo 22.: \ 9:2 800% 25.: not oEobxm 95 57.53.“.
3.5.55.3. oEe5< oE?55:< 5.33
6?:— vlu .5535...— N00 ecu—2.5 2:55:— \ 2:55:53. 55: E5 2:: 95.55.— «255.5%

 

.385 :5 a £52: 25 mow—:2 5 5:82 55 55% A5538: 4:0 E5 NOD 5w 7.3: _.m BE: 95 7.3: Tm .083 35: 588
55 o: 5:525“. 553 «53 A8558: 4:0 55 N00 @5555 858535 533:2: 880838 95—55: mo mioﬁgm .Ym 535,—.

109

-0- Control -I Lowered "A“ Raised

10 CM DEPTH

 

3
2.5-
;J

1.5 -

 

 

 

CH4 PRODUCTION
(nmol cm 3 day 1)

 

 

 

 

. 25 CM DEPTH

 

 

 

CH4 PRODUCTION
(nmol cm 3 day")

 

 

 

15 20 25 3O 35 40
DAY OF INCUBATION

Figure 3-1. Mean anaerobic CH4 production (nmol cm'3 day'l) i one standard error over
the 40 day incubation experiment. Peat was taken from the 5-10 cm and 20-25 cm depth
increments below the moss surface across three water table treatments (control, raised,
and lowered water table plots, Chapter 2).

110

AN 585:0 .82: 533 553 5532 85 6358 .5888 85858:
538 553 588 $055 35.5% 808 55 3052 358558 5:5: 80 3.8 85 80 on 58 88: :83 553 55: 858595
885858 050525 85 05055 :5: ov 58 5>o 885 85855 580 “178: $80 583 88588: NOD 852 .Num 2:”:—

zOF<m:Oz. “.0 ><D 20.....(m302. “.0 ><o

 

 

um)

(PKep 9w: [0
Mouonooud zoo

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

8

m o
a... w :0
w m. w
:83. If. u P m
32.6.. .0! km
.253 IOI 1:330 .20 mN U_m0¢m< Itwo 2U OH U_m0¢u< 3 .l N

I m

 

 

111

I ANAEROBIC Cl AEROBIC

 

C

MEAN co: Paooucnou
(umo| a" day")

 

 

   

 

 

5-10CM 20-250M

DEPTH BELOW SURFACE

Figure 3-3. Results of an interaction between depth increment and headspace treatment

(aerobic, anaerobic) on C02 production rates on a mass basis. Data are means i one
standard error (not adjusted for model comparisons). Same letter superscripts denote
non-signiﬁcant differences from Tukey post hoc comparison of means tests.

112

I UN-WARMED CI WARMED

 

 

 

 

 

 

   

 

 

 

z 1.6 c
E .. 1.4 —
3 g“ 1.2
g no 1 ‘ L
t E 0.8—
° 3 0.5
g g 04 -
a o
z 0.21

o

5-1OCM 20-25CM
DEPTH BELOW SURFACE

Figure 34. Results of an interaction between depth increment and warming treatment on

C02 production rates on a volume basis. Data are means d: one standard error (not
adjusted for model comparisons). Same letter superscripts denote non-signiﬁcant
differences from Tukey post hoc comparison of means tests.

113

.10 CM 1325 CM

 

 

1.5 d cd
2
g- c
M 1~ .
1:5 a
0'6
250.5
5..
2
o

 

 

 

 

 

 

 

 

 

CONTROL LOWERED RAISED

Figure 3-5. Results of an interaction between depth increment and water table plot on
C02 production rates on a volume basis. Data are means i one standard error (not
adjusted for model comparisons). Same letter superscripts denote non-signiﬁcant
differences from Tukey post hoc comparison of means tests.

114

.10 CM C125 CM

 

3000‘

2000“

C02 - CH4 RATIO

1000’

 

 

   

 

 

 

CONTROL LOWERED RAISED

Figure 3-6. Results of an interaction between depth increment and water table plot on

ratios of C02 — CH4 production rates. Data are means i one standard error (not adjusted
for model comparisons). Same letter superscripts denote non-signiﬁcant differences from
Tukey post hoc comparison of means tests.

115

Figure 3-7. Relationship between C02 and CH4 production rates from incubated peat
and soil quality variables. Closed circles represent data from the 5-10 cm increment,
while open squares represent the 20-25 cm increment. No relationships between gaseous
carbon production rates were signiﬁcant (p>0.05).

116

cu-I4

co2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o n
.u . é... a ..
“a u D an. D. 3 00% .
0 do one 00 0‘00. a T 6 a
0 0 %0 0 0 00 0 2 D 9N0
a a . 2
. D..ﬂ 9§ ” ) C v u . . e m
0 ) /o O O O
M a. a nut .. 01am...“ .
a m 0 as“ g
4 . 1 . . u 1 .
3 m .H W. 5 0 3 m 3 m H m S O m 3 N E m 5 0 4 3 2 1 0
$.60 a.» .085 M .050 vs .085 m $.50 vu .085 m :52. m.80 .085
C
m m . . ..
00H. m m n.
3?. u g . 3 .. 1 am.
1... %ﬁuéa 3 3 n
O O 0&0 n— . O O D D B D D
O O. .9.» O 6 ﬂ
.. n. .0. . 2
000%00 0 0 00 OOOOnmﬂ
50 ﬂ 2 00 . u 8‘?
00 a a 00 0 0 0 00.. 9.0
06 n 0 0 3 0 m o o
. . . 0.0 M. l _ 0 0 1 g 1
mmwmmo mmwwmo manaauuo. aunmumro
.33 H.» .25.. C5. a.» .25.. .33 a.» .25.. .32. 15 .25..

0.2

0.1

0.0

117

02
BULK DENSITY (g cm'3)

0.1

0.0

5-10 cm Depth

I Aerobic
Ij El Anaerobic

Control Lowered Raised

[N03] (mg L")
a

 

20-25 cm Depth

IN03I(mgL')
P P P P
c N 13 OK 00 I-
! i i
1

Control Lowered Raised

Figure 3-8. N03 (mg L'l) concentrations from the aerobic and anaerobic post-incubation
leachates across water table plot and depth.

118

References

Aber JD, Melillo JM. 2001. Terrestrial Ecosystems. 2nd edition Academic Press, p. 556.
San Diego. CA.

Aerts R, Chapin FS 111. 2000. The mineral nutrition of wild plants revisited: a re-
evaluation of processes and patterns. Advances in Ecological Research 30: 1-68.

Bergman I, Lundberg P, Nilssona M. 1999. Microbial carbon mineralisation in an acid
surface peat: Effects of environmental factors in laboratory incubations. Soil Biology and
Biochemistry 31: 1867-1877.

Bridgham SD, Richardson CJ. 1992. Mechanisms controlling soil respiration (C02 and
CH4 in southern peatlands. Soil Biology and Biochemistry 24: 1089-1099.

Bridgham SD, Updegraff K, Pastor J. 1998. Carbon, nitrogen, and phosphorus
mineralization in northern wetlands. Ecology: 79: 1545-1561.

Bubier JL, Moore, TR. 1994. An ecological perspective on methane emissions from
northern wetlands. Trends in Ecology & Evolution 9: 460-464.

Chapin F S III, 1980. The mineral nutrition of wild plants. Annual Review of Ecology and
Systematics 11: 233-260.

Chapin F S III, McGuire AD, Randerson J, Pielke R, Baldocchi D, Hobbie SE, Roulet M,
Eugster W, Kasischke E, Rastetter EB, Zimov SA, Running SW. 2000. Arctic and boreal

ecosystems of western North America as components of the climate system. Global
Change Biology 6: 211-223.

Christensen TR, Jonasson S, Callaghan V, Havstrom M. 1999. On the potential C02
release from tundra soils in a changing climate. Applied Soil Ecology 11: 127-134.

Crill PM. 1991. Seasonal patterns of methane uptake and carbon dioxide release by a
temperate woodland soil. Global Biogeochemical Cyles 5: 319-334.

Gignac LD, Halsey LA, Vitt DH. 2000. A bioclimatic model for the distribution of
Sphagnum-dominated peatlands in North America under present climatic conditions.
Journal of Biogeography 27: 1139-1151.

Gill RA, Jackson RB. 2000. Global patterns of root turnover for terrestrial ecosystems.
New Phytologist 147: 13—31.

Glatzel S, Basiliko N, Moore TR. 2004. Carbon dioxide and methane production

potentials of peats from natural, harvested and restored sites, Eastern Quebec, Canada.
Wetlands 24: 262-267.

119

Gorham E. 1991. Northern peatlands: role in the carbon cycle and probable responses to
climatic warming. Ecological Applications 1: 182-195.

Grogan P, Chapin F S 111. Initial effects of experimental warming on above- and

belowground components of net ecosystem C02 exchange in arctic tundra. 2000.
Oecologia 125: 512-520.

Hobbie SE. 1996. Temperature and plant species control over carbon and nitrogen
cycling through litter and soil in Alaskan tundra. Ecological Monographs 66:503-522.

Hobbie SE, Chapin F S III. 1996. Winter regulation of tundra litter carbon and nitrogen
dynamics. Biogeochemistry 35:327-338.

Hobbie SE, Chapin F S III. 1998. The response of tundra plant biomass, aboveground
production, nitrogen, and C02 ﬂux to experimental warming. Ecology 79: 1526-1544.

Hobbie SE, Schimel JP, Trumbore SE, Randerson JR. 2000. A mechanistic understanding
of carbon storage and turnover in high-latitude soils. Global Change Biology 6:196-210.

Hogg EH, Lieffers VJ, Wein RW. 1992. Potential carbon losses from peat proﬁles:
Effects of temperature, drought cycles, and ﬁre. Ecological Applications 2: 298-306.

Johnson LC, Damman AWH, Malmer N. 1990. Sphagnum macrostructure as an indicator
of decay compaction in peat cores from an ombrotrophic south Swedish peat-bog. Journal
of Ecology 78: 633-647.

Keller J K, White JR, Bridgham SD, Pastor J. 2004. Climate change effects on carbon and
nitrogen mineralization in peatlands through changes in soil quality. Global Change
Biology 10: 1053—1064.

MacDonald NW, Zak DR, Pregitzer KS. 1995. Temperature effects on kinetics of
microbial respiration and net nitrogen and sulfur mineralization. Soil Science Society of
America 59: 233-240.

Madritch MD, Hunter MD. 2003. Intraspeciﬁc litter diversity and nitrogen deposition
affect nutrient dynamics and soil respiration. Oecologia 136: 124-128.

Magnusson T. 1993. Carbon dioxide and methane formation in forest mineral and peat
soils during aerobic and anaerobic incubations. Soil Biology and Biochemistry 25: 877-
883.

McGuire AD, Hobbie JE. 1997. Global climate change and the equilibrium responses of
carbon storage in arctic and subarctic regions. pp. 47-48. Fairbanks, Alaska.

McGuire AD, Clein J S, Melillo JM, Kicklighter DW, Meier RA, Vorosmarty CJ, Serreze
MC. 2000. Modeling carbon responses of tundra ecosystems to historical and projected

120

climate: Sensitivity of pan-arctic carbon storage to temporal and spatial variation in
climate. Global Change Biology 6: 141-159.

McKenzie C, Schiff S, Aravena R, Kelly C, St. Louis V. 1998. Effect of temperature on

production of CH4 and C02 from peat in a natural and ﬂooded boreal forest wetland.
Climatic Change 40: 247-266.

Minkkinen K, Laine J. 1998. Long-term effect of forest drainage on the peat carbon
stores of pine mires in Finland. Canadian Journal of Forest Research 28: 1267-1275.

Moore TR, Dalva M. 1993. The inﬂuence of temperature and water table position on
carbon dioxide and methane emissions from laboratory columns of peatland soils.
European Journal of Soil Science 44: 651—664.

Moore TR, Dalva M. 1997. Methane and carbon dioxide exchange potentials of peat soils
in aerobic and anaerobic laboratory incubations. Soil Biology and Biochemistry 29:
1157-114.

Nadelhoffer KJ, Giblin AE, Shaver GR, Laundre JA. Effects of temperature and substrate
quality on element mineralization in six Arctic soils. Ecology 72: 242-253.

Nykanen H, Alm J, Lang K, Silvola J, Martikainen PJ. 1995. Emissions of CH4, N20 and

C02 from a virgin fen and a fen drained for grassland in Finland. Journal of
Biogeography 22: 351-357.

Rustad L, Campbell J, Marion G, Norby R, Mitchell M, Hartley A, Comelissen J,
Gurevitch J. 2001. A meta-analysis of the response of soil respiration, net nitrogen

mineralization, and aboveground plant growth to experimental ecosystem warming.
Oecologia 126: 543-562.

Silvola J, Ahlholm U. 1993. Effects of CO; concentration and nutrient status on growth,

growth rhythm and biomass partitioning in a Willow, Salix phylicifolia. Oikos 67: 227-
234.

Silvola J, Alm J, Ahlholm U, Nykanen H, Martikainen PJ. 1996. C02 ﬂuxes from peat in
boreal mires under varying temperature and moisture conditions. The Journal of Ecology
84: 219—28.

Strack M, Waddington JM. 2007. Response of peatland carbon dioxide and methane
ﬂuxes to a water table drawdown experiment. Global Biogeochemical Cycles 21:
GB 1007, doi:10.1029/2006GB002715.

Turetsky MR, Wieder RK, Williams C, Jendro J, Vitt DH. 2000. Organic matter

accumulation, peat chemistry, and permafrost melting in peatlands of boreal Alberta.
Ecoscience 7: 379-392.

121

Turetsky MR. 2003. The role of bryophytes in carbon and nitrogen cycling. The
Bryologist 106: 395-409.

Turetsky MR, Ripley S. 2005. Decomposition in extreme-rich fens of Boreal Alberta,
Canada. Soil Science Society of America 69:1856-1860.

Turetsky MR, Treat CC, Waldrop M, Waddington JM, Harden JW, McGuire AD. In
press. Soil climate controls on methane ﬂuxes and methanogen activity in an Alaskan

peatland: Short-term responses to in situ water table and soil warming manipulation.
Journal of Geophysical Research - Biogeosciences. doi:10.1029/2007JG000496, in press.

Updegraff K, Pastor J, Bridgham SD, Johnston CA. 1995. Environmental and substrate
controls over carbon and nitrogen mineralization in northern wetlands. Ecological
Applications: 5 151-163.

Updegraff K, Bridgham SD, Pastor J, Weishampel P, Harth C. 2001. Response of C02

and CH4 emissions from peatlands to warming and water table manipulation. Ecological
Applications 11: 31 1-326.

Van Wijk MT, Clemmensen KE, Shaver GR, Williams M, Callaghan TV, Chapin F S III,
Comelissen JHC, Gough L, Hobbie SE, Jonasson S, Lee JA, Michelsen A, Press MC,
Richardson SJ, Rueth H. 2004. Long-term ecosystem level experiments at Toolik Lake,
Alaska, and at Abisko, Northern Sweden: Generalizations and differences in ecosystem
and plant type responses to global change. Global Change Biology 10: 105—123.

Verhoeven JTA, Liefveld WM. 1997. The ecological signiﬁcance of organochemical
compounds in Sphagnum. Acta Botanica Neerlandica 46: 117-130.

Vitousek PM, Aber JD, Howarth RW, Likens GE, Matson PA, Schindler DW,
Schlesinger WH, Tilman DG. 1997. Technical Report: Human Alteration of the Global
Nitrogen Cycle: Sources and Consequences. Ecological Applications 7: 737-750.

Vitt DH. 2006. Functional characteristics and indicators of boreal peatlands, in Boreal
Peatland Ecosystems, edited by Wieder RK, Vitt DK, pp. 1-8, Springer-Verlag Berlin
Heidelberg, New York.

Waddington JM, Grifﬁs TJ, Rouse WR. 1998. Northern canadian wetlands: Net
ecosystem C02 exchange and climatic change. Climatic Change 40: 267-275.

Waddington JM, Rotenberg PA, Warren FJ. 2001. Peat C02 production in a natural and
cutover peatland: Implications for restoration. Biogeochemistry 54: 115-130.

Weltzin J F, Harth C, Bridgham SD, Pastor J, Vonderharr M. 2001. Production and

microtopography of bog bryophytes: response to warming and water-table manipulations.
Oecologia 128: 557-565.

122

Yavitt JB, Lang GE, Wieder RK. 1987. Control of carbon mineralization to CH4 and C02
in anaerobic, Sphagnum-derived peat from Big Run Bog, West Virginia.
Biogeochemistry 4: 141-157.

Yavitt JB, Williams CJ, Wieder RK. 1997. Production of methane and carbon
dioxide in peatland ecosystems across North America: Effects of temperature, aeration,
and organic chemistry of peat. Geomicrobiology Journal 14: 299-316.

123

CHAPTER 4

CONCLUSION

Northern soils in boreal regions serve as an important reservoir for terrestrial soil
carbon (C) that has accumulated since the last deglaciation (12,000 — 18,000 yr ago;
Harden et al. 1992). Boreal forests contain approximately 27% of the world’s vegetation
C (McGuire et al. 1997) and between 25 — 30% of the world’s soil C (Gorham 1991,
McGuire et al. 1997). The majority of this boreal C stock is stored in peatlands, where
cold soils, saturated conditions, and recalcitrant substrate quality limit decomposition and
promote peat accumulation (Gorham 1991). Due to the controls, soil climate and its
inﬂuence on C ﬂuxes (respiration and net primary production) represents a major control
on boreal C stocks (Harden et al. 1992, Bhatti et al. 2002). However, climate models
generally predict warmer conditions for high latitudes (Raisanen 1997), an area where
peatlands are an important component of the landscape. Thus, how peatland C stocks
will respond to climate change remains an important issue with implications for
atmospheric C concentrations.

The majority of peatland C studies have been conducted in Sphagnum-dominated
poor fens and bogs in Canada and Fennoscandia. Relatively little is known about
Alaskan peatlands, although Bridgham et al. (2006) estimate that Alaska wetlands store
as much as 42 Pg of C. Alaskan peatlands are changing due to near-surface atmospheric
warming of approximately 1° C per decade on average (Osterkamp and Romanovsky

1999, Serreze et al. 2000). These climatic changes have led to regional changes in

124

surface hydrology and vegetation structure throughout interior Alaska (Riordan et al.
2006, Hinzman et al. 2005, Hinzman et al. 2006).
The goal of this thesis was to investigate direct and indirect soil hydroclimate

(water table position, soil temperature) controls on C ﬂuxes in an Alaskan fen. I
investigated the response of ﬁeld C02 ﬂuxes (NEE, ER, GPP) to a water table and soil

temperature manipulation. Water table level and soil temperature have the potential to
indirectly affect decomposition through changes in soil organic matter quality. I
designed a laboratory experiment to isolate soil organic matter quality differences among
peat collected across the experimental ﬁeld treatments. To understand whether the ﬁeld
manipulations affected soil organic matter quality and C mineralization rates, 1 measured
aerobic and anaerobic C02 and anaerobic CH4 production rates under constant moisture
and temperature.

1 had hypothesized that drought would increase decomposition rates, leading to an
increase in ER and an increase in NEE (atmospheric emissions). Two years of ﬁeld
ﬂuxes showed that the lowered water table treatment did not alter ER, but did lower GPP

by 21-24% (Figure 2-11). I attribute this change in GPP to decreased photosynthetic

uptake of C02 by vascular vegetation and bryophytes due to drought stress (Figure 2-6).

The lowered water table plot generally had lower GPmax values than the other water table

treatments, which supports this conclusion (Figure 2-13). Reductions in GPP resulted in
increased NEE (trend toward increased atmospheric emissions) in the lowered water table
plot relative to the control plot. So while drought affected NEE as hypothesized, the

change in NEE was driven primarily by plant productivity rather than microbial activity.

Surface peat from the lowered water table plot also had the slowest C02 production rates

125

during a laboratory incubation experiment (Figure 3-5), indicating a decrease in soil
organic matter quality either from reduced input of fresh organic matter by plants or
accelerated turnover of organic matter in the ﬁeld. Whether this organic matter quality
signature has or will serve as a feedback to ecosystem ﬂuxes could be determined by

additional C ﬂux monitoring at the site.

The raised water table treatment was less of a C02 sink (more positive NEE
values) than the control water table plot in 2005, but was a greater sink in 2006. In 2006,
this trend was driven primarily by interannual changes in GPP, with greater GPmax values

and % bryophyte cover in the raised water table plot relative to the control (Figure 2-13,
Table 2-1). I had hypothesized that increased water table level would decrease
decomposition and lead to decreased ER. However, the raised water table treatment also
had increased ER compared to the control water table plot; likely due to high autotrophic
respiration and/or high rates anaerobic C mineralization. Additionally, temperature was

the strongest control on ER, and warmer soils at the raised plot (Figure 2-5) and likely

stimulated microbial activity. My laboratory experiment showed that in deeper peat, C02

production rates at the raised water table treatment were 23% greater than C02

production rates in control water table peat (Figure 3-5), likely due to high vegetation
productivity under wet conditions that could have led to increased ﬁne root turnover or
rootexudaﬁon.

Soil warming increased both ER and GPP by 16% relative to un-warmed
subplots, resulting in no net effect of warming on NEE. Laboratory C02 production
rates from peat taken from the warmed subplots were 26% and 45% lower than peat from

the un-warmed subplots in shallow and deep peat (Figure 3-4). Despite greater inputs of

126

fresh organic matter to surface soils with higher GPP, faster organic matter turnover in
the ﬁeld in the warming treatment seems to dominate the organic matter quality

signature, thereby controlling potential C mineralization rates. Models of vegetation
components at our site (GPP, GPmax, (1) indicate that water table position and soil

temperature are both important controls on plant C uptake, while temperature was a
stronger control on ER than water table position.

Taken together, these results indicate that 2.5 years of soil warming and water
table manipulations can signiﬁcantly alter C02 ﬂuxes in an Alaskan fen. While much of

the literature has focused on how climate change will inﬂuence decomposition rates (and
thus peat accumulation), my study clearly demonstrates the need to consider vegetation
and substrate quality in addition to microbial activity when investigating ecosystem-level
responses to water table levels.

My data do show that drought is likely to decrease soil C storage in Alaska
peatlands, but primarily through reduced plant productivity. Future work at the APEX
site should monitor GPP with species succession over time in the lowered water table plot
to determine if the trends I observed are likely to change over long time scales (decades).

Peatland ﬂooding has received less attention in the literature compared to drought,
but my results suggest that inundation can have large effects on C ﬂuxes and can
decrease NEE, promoting C storage. The raised water table treatment increased GPP and
ER, but overall caused a net decrease in NEE. This agrees well with C ﬂux studies
focusing on saturated peatlands post-permafrost (Turetsky et al. 2000, Turetsky et al.
2002, Wickland et al. 2006). Thus, as climate change progresses, causing more and more

permafrost degradation, increased soil C storage in peatlands could serve as a negative

127

feedback to climate change. However, Turetsky et al. (2007) showed that the response of
peatlands post-thaw to sequester more C was temporary, lasting only ~lOO years.
Warming might stimulate both GPP and ER in Alaska peatlands, so it seems
likely that the direction of C ﬂux changes will be driven mostly by changing hydrological
conditions. However, this study demonstrates how soil hydrology and thermal regimes in
peatland environments are coupled due to increased buffer capacity and heat transfer of
saturated soil layers compared to dry soils. Peatland and boreal C models need to
account for feedbacks between plant and microbial activity (i.e., through nutrient
availability or substrate quality) as well as the interactions between physical peat
properties, soil hydroclimate, and biological processes that regulate C ﬂuxes and

emissions.

128

References

Bhatti J S, Apps MJ, Tamocai C. 2002. Estimates of soil organic carbon stocks in central
Canada using three different approaches. Canadian Journal of Forest Research 32: 805-
812.

Bridgham SD, Megonigal JP, Keller JK, Bliss NB, Trettin C. 2006. The carbon balance
of North American wetlands. Wetlands 26: 889-916.

Gorham E. 1991. Northern peatlands: role in the carbon cycle and probable responses to
climatic warming. Ecological Applications 1: 182-195.

Harden JW, Sunquist ET, Stallard RF, Mark RK. 1992. Dynamics of soil carbon during
deglaciation of the Laurentide Ice Sheet. Science 258: 1921-1924.

Hinzman, LD, Bettez ND, Bolton WR, Chapin F S, Dyurgerov MB, F astie CL, Grifﬁth B,
Hollister RD, Hope A, Huntington HP, Jensen AM, Jia GJ, Jorgenson T, Kane DL, Klein
DR, Koﬁnas G, Lynch AH, Lloyd AH, McGuire AD, Nelson FE, Nolan M, Oechel WC,
Osterkamp TE, Racine CH, Romanovsky VE, Stone RS, Stow DA, Sturm M, Tweedie
CE, Vourlitis GL, Walker MD, Walker DA, Webber PJ, Welker JM, Winker KS,
Yoshikawa K. 2005. Evidence and implications of recent climate change in northern
Alaska and other Arctic regions. Climatic Change 72: 251-298.

Hinzman, LD, Viereck LA, Adams PC, Romanovsky VE, Yoshikawa K. 2006. Climate
and permafrost dynamics of the Alaskan boreal forest, in Alaska’s Changing Boreal
Forest edited by Chapin, F S III, Oswood MW, Van Cleve K, Viereck LA, Verbyla DL,
editors. Oxford University Press, New York.

McGuire AD, Hobbie JE. 1997. Global climate change and the equilibrium responses of
carbon storage in arctic and subarctic regions. pp. 47-48. Fairbanks, Alaska.

Osterkamp TE, Romanovsky VE. 1999. Evidence for warming and thawing of
discontinuous permafrost in Alaska. Permafrost and Periglacial Processes 10: 17-37.

Raisanen J. 1997. Objective comparison of patterns of C02 induced climate change in
coupled GCM experiments. Climate Dynamics 13:197—21 1.

Riordan B, Verbyla D, McGuire AD. 2006. Shrinking ponds in subarctic Alaska based on
1950-2002 remotely sensed images. Journal of Geophysical Research — Biogeosciences:
G04002, doi: 10: 1029/2005JG000150.

Serreze MC, Walsh JE, Chapin F S, Osterkamp TE, Dyurgerov M, Romanovsky VE,

Oechel WC, Morison J, Zhang T, Barry RG. 2000. Observational evidence of recent
change in the northern high-latitude environment. Climatic Change 46: 159—207.

129

Turetsky MR, Wieder RK, Vitt DH, Evans R, Scott KD. 2007. The disappearance of
relict permafrost in boreal regions: effects on peatland carbon storage and ﬂuxes. Global
Change Biology 13: 1—13.

Turetsky MR, Wieder RK, Williams CJ, Vitt DH. 2000. Organic matter accumulation,
peat chemistry, and permafrost melting in peatlands of boreal Alberta. Ecoscience 7: 379-
392.

Turetsky MR, Wieder RK, Vitt DH. 2002. Boreal peatland C ﬂuxes under varying
permafrost regimes. Soil Biology and Biochemistry 34: 907-912.

Wickland KP, Striegl RG, Neff JC, Sachs T. 2006. Effects of permafrost melting on C02

and CH4 exchange of a poorly drained black spruce lowland. Journal of Geophysical
Research 111: G02011, doi:10.1029/2005JG000099.

130

 

1|1ititljijilﬂijijjiiﬂjijI