INFLUENCE OF SOIL PROFILE PROPERTIES ON RESPIRED
CARBON DIOXIDE (CO2) AND METHANE (CH4) GASES BEFORE,
DURING, AND AFTER DISTURBANCE BY A RETROGRESSIVE
THAW SLUMP ALONG THE SELAWIK RIVER, NW ALASKA.
by
Amy E. Jensen
A thesis submitted in partial fulfillment
of the requirements for the degree of
Masters of Science in Geology
Idaho State University
In presenting this thesis in partial fulfillment of the requirements for an advanced degree at
Idaho State University, I agree that the Library shall make it freely available for inspection. I
further state that permission for copying of my thesis for scholarly purposes may be granted
by the Dean of Graduate Studies, Dean of my academic division, or by the University
Librarian. It is understood that any copying or publication of this thesis for financial gain
shall not be allowed without my written permission.
Signature______________________________________
Date__________________________________________
To the Graduate Faculty:
The members of the committee appointed to examine the thesis of Amy E. Jensen
find it satisfactory and recommend that it be accepted.
_________________________________________
Benjamin T. Crosby, Principal Advisor
_________________________________________
Kathleen A. Lohse, Second Advisor
_________________________________________
Keith Reinhardt, Graduate Faculty Representative
iii
Acknowledgements
Margaret Murnane this is for you! Without your love, devotion, dedication, support and
unwavering ability to hold me accountable, I would have never become the person I am
today. You give more than a ton, and truly you’re the glue that holds our crazy family
together! To the rest of my family, you all are without a doubt both the best and strangest
things in my life. Thank you for all your love and support couldn’t have done this without
all the laughs, distractions, and encouragement from you all, I love you bunches! Ben
Crosby, where do I even begin? Thanks for sharing your person, family, and dance with
me you all are good weird, and I’ve enjoyed myself immensely with you all! Alaska,
what a life experience, thank you so much for allowing me to work with you on such
amazing research in such a gorgeous place! You’ve given a lot of yourself in my pursuit
of these ~100 pages and I hope I’ve managed to take most of it in! Ben Abbott, thank you
for all your insight and willingness to share your expertise. Without your help and
guidance I would have been utterly lost. Kathleen Lohse, you’re the women: an amazing
person and scientist! Thank you for all the time and effort you committed to me. Your
insight and guidance have not only shaped this project into a story with teeth, grit, and
importance, but you’ve also given me great confidence in my ability as a scientist, and
for that I will be forever grateful. It’s all about the SOILS! Claudia Mora, thank you so
much for all your time and influence! You are an incredible person and scientist! Thank
you so much for such an amazing project to work on! You have opened my eyes to how
effective science is conducted. Joel Rowland, thank you for the use of your meterological
data for chapter 2. Sara Godsey, thank you so much for all the statistical help you
provided on this project. Keith Reinhardt, thank you for the willingness to sit on my
iv
committee! To all my fellow students and friends, BOOM, without you all my headspace
would have been an utter mess!
This research received financial and logistic support from the Los Alamos National
Energy Lab (LANL IGPP Minigrant 166259), the Selawik National Wildlife Refuge
(USFWS – CESU 84320-9-J306R) as well as financial support from the NSF’s Arctic
System Science (ARCSS) Program (OPP – 0806399) and the NSF Idaho EPSCoR
Program in Water Resources in a Changing Climate (EPS 0814387).
v
Table of Contents
List of Figures……………………………………………………………….…………. viii
List of Tables……………………………………………………………….………...... ix
Abstract……………………………………………………………………….………… x
1
Introduction
1
1.1
Motivation
1
1.1.1 Thermokarst
2
1.1.2 Soil Organic Carbon
3
1.1.3 Soil Properties and Structure
4
1.1.4 Carbon Dioxide Efflux
5
1.1.5 Controls on CO2 Efflux
7
1.1.6 δ13C of Soil Respired CO2
8
1.2
Thesis Objective
10
1.3
Thesis Structure
10
1.4
References
12
2
Soil properties affect variations in carbon dioxide efflux across a retrogressive
thaw slump chronosequence in northwest Alaska.
15
2.1
Abstract
15
2.2
Introduction
16
2.3
Field Setting
19
2.4
Study Deisgn: a chronosequence enables a space for time substitution 20
2.5
Methods
22
2.5.1 Meteorology
22
2.5.2 Point measurements of CO2 Efflux, Soil Temperature, Soil Moisture
22
2.5.3 Physical Soil Properties
23
2.5.4 Statistical Analysis
24
2.6
Environmental Conditions over Space and Time
25
2.6.1 Precipitation
25
2.6.2 Air Temperature
25
2.6.3 Soil Temperature
25
2.6.4 Soil Moisture
26
2.6.5 Physical Soil Properties
27
2.7
CO2 Efflux over Space and Time
2.7.1 Controls Affecting CO2 Efflux
vi
28
30
2.8
Implications
32
2.9
Conclusion
33
2.10
References
34
2.11
Figures
37
3
Soil matters: Vertical profiles of CO2 and CH4 concentrations and δ13C of CO2
respired from a thaw slump chronosequence reveal that soil properties can control
carbon efflux at the surface.
44
3.1
Abstract
44
3.2
Introduction
45
3.3
Field Setting
49
3.4
Thaw slump chronosequence: a space for time substitution
50
3.5
Methods
53
3.5.1 Sampling Locations
53
3.5.2 Soil Analysis
53
3.5.3 Soil Gas Analysis
55
3.5.4 Statistical Analysis
56
3.6
Results
57
3.6.1 Physical Soil Properties
57
3.6.2 Soil Gas CO2 and CH4 Concentrations
59
3.6.3 Stable Isotopic Composition of Soil and Soil Gases
13
3.6.4 Controls on Gas Concentrations and δ C Isotope Values
3.7
4
5
Discussion
60
61
62
3.7.1 Diffusion Limitation
62
3.7.2 Micro-Soil Environments
64
3.8
Implications
66
3.9
Conclusion
68
3.10
Reference
69
3.11
Figures
75
Thesis Summary
79
4.1
Implications
79
4.2
Conclusion
80
Supplemental Material
82
vii
List of Figures
Chapter 2
Figure 1. Field area location, chronosequence selection………………………….…….36
Figure 2. Day-of-year time series of environmental variables for 2011 and 2012…….. 38
Figure 3. Mean peak growing season soil temperature, moisture and CO2 efflux…….. 39
Figure 4. Soil moisture and temperature controls on CO2 efflux………………....….... 41
Figure 5. Physical soil property controls on CO2 efflux…………...…………………... 42
Chapter 3
Figure 1. Field area location, topographic cross-sections, soil profiles…….………….. 73
Figure 2. Soil profiles physical properties…………………………………..…………. 74
Figure 3. Soil profiles carbon concentration and isotopes…………………………..…. 76
Supplemental Material
Figure 1. Soil Profile Photographs................................................................................... 80
viii
List of Tables
Chapter 2
Table 1. Summary ANOVA value table................................…………….……….…… 37
Table 2. Chronosequence soil and vegetation characteristics…………….……….…… 37
Table 3. Comparisons of CO2 efflux……………………………………………..…….. 40
Chapter 3
Table 1. Comparisons of CO2 efflux and effective diffusivity…….……………….….. 75
Table 2. Controls on CO2, CH4 concentrations, and δ13C of CO2……………………... 75
ix
Abstract
Warming of the arctic landscape results in permafrost thaw, which causes ground
subsidence or thermokarst. On hillslopes, this process is accelerated, altering soil
structure, temperature, moisture, as well as the content and quality of soil organic matter.
In turn, these changes affect both the rate and mechanism of carbon cycling in permafrost
soils. This is a major concern because permafrost soils store nearly half the world’s
global carbon. In order to assess the time-evolution of carbon contributions from
thermokarst features, we use a three-stage chronosequence of thaw slumps to describe
how carbon dioxide (CO2) fluxes from undisturbed, active and stabilized surfaces differ.
We then evaluate the fate of carbon within the above three environments, using a soilprofile chronosequence to determine how thermokarst-driven changes in the soil microenvironment affect soil column CO2 and CH4 concentrations as well as the resulting CO2
efflux. The study site, found along the Selawik River in northwest Alaska, contains one
of the largest active retrogressive thaw slumps known in North America and had a
surface area of ~50,000 m2 in 2012. We designate the undisturbed tundra as the initial
condition and an adjacent, bowl-like depression re-vegetated by multiple generations of
black spruce trees as a stabilized slump.
In this 2-year study, we use peak summer measurements of CO2 efflux across the
chronosequence, as well as, soil profile CO2 and CH4 concentrations, soil profile δ13C of
both soil respired CO2 and soil organic matter. Physical soil parameters such as
temperature, moisture, soil organic carbon %, porosity and water filled pore space at are
measured at depths of 3, 5, 15, 30, and 50cm, to determine how carbon emissions are
affected by changes within the soil environment resulting from the initiation and
x
subsequent stabilization of a rapidly evolving thaw slump. CO2 efflux from the active
slump is less than half that from either the undisturbed tundra or stabilized slump. While
the active slump had the lowest CO2 surface efflux, at depth the CO2 concentrations were
5 orders of magnitude higher than those observed in the other two environments. CH4
concentrations in the active slump were ~3x higher. Although the floor of the active
slump is generally 10 to 15°C warmer than the tundra and stabilized slump, we did not
observe that soil temperature or moisture exerts a strong control on CO2 efflux. Rather,
we found that local physical soil properties such as low porosity, high water filled pore
space, and low calculated effective diffusivity in the active slump inhibit CO2 efflux. In
contrast to our expectation that the active slump would be primarily aerobic due to
landscape position (riverbanks with slopes of 10-25%), observations of soil gleying,
heavier δ13C of CO2 (~-15‰), and the production of methane all point to anaerobic soil
conditions, suggesting a switch in the metabolic pathway toward methanogenesis. This
suggests that though thaw slumps can profoundly alter the biophysical controls on CO2
and CH4 production, the dilution of organic soils by glacial sediment as well as changes
in physical soil profile properties in the active feature reduced CO2 efflux below those
measured in the undisturbed tundra. As the feature matures and is recolonized as a boreal
forest ecosystem, fluxes return toward initial conditions. Catastrophic thermokarst
features are likely to increase frequency and extent in a changing climate. Future studies
will need to assess whether other thermokarst features with higher organic content
substrates respond similarly to these evolving landscape features.
xi
1
Introduction
1.1 Motivation
Arctic ecosystems are presently experiencing unprecedented climate warming,
which dramatically alters topography, soil structure and function, carbon storage, water
balance, and carbon emissions. This introduction first examines past and current work on
geomorphic changes associated with permafrost thaw and carbon emissions as a result of
permafrost degradation. Second, changes in physical soil properties including soil
temperature and moisture as well as substrate quality are examined as controls on carbon
gas emissions and isotopic values. Third, this chapter discusses how changes in soil
properties with depth, such as bulk density, porosity, soil organic matter, soil
temperature, and soil moisture may influence carbon-cycling.
International research has clearly documented a warming climate in Arctic
ecosystems (IPCC, 2007). Since the Little Ice Age, defined as the period between the 16th
and 19th centuries, mean temperatures in the Arctic have risen by 3 -5 °C (ACIA, 2004;
IPCC, 2007). Within Alaska mean temperature has risen by 1 -3 °C in the last three
decades (Hinzman et al., 2005). Results from general circulation models differ somewhat
regarding future trends, but models and scenarios selected for the Arctic Climate Impact
Assessment (2004) agree that, by 2100, average annual temperatures in the Arctic are
expected to increase 3-5°C and winter temperatures may increase by 4-7°C. These
models also suggest that rising temperatures will be accompanied by increased
precipitation, mostly in the form of rain (IPCC, 2007).
1
1.1.1
Thermokarst
A warming climate will have important consequences for arctic ecosystems. One
major impact is permafrost thaw and subsequent formation of thermokarst terrain.
Permafrost is defined as soil or rock that remains at or below 0°C for 2 or more
consecutive years (Jorgenson and Osterkamp, 2005). Found just above the permafrost,
the active layer is defined as the uppermost layer of soils (~0.4 – 2m thick) which thaws
during the summer and freezes again during the autumn (Jorgenson et al., 2008b). As a
consequence of permafrost thaw the land-surface subsides, resulting in surface
subsidence (thermokarst) and/or mass wasting (thermal erosion feature) (Jorgenson et al.,
2010; Rowland et al., 2010).
There are many types of thermokarst features. Formation of these features
depends on the amount and type of ground ice as well as the geomorphic surface
characteristics. This review focuses on a type of thermal erosional features called
retrogressive thaw slumps (Jorgenson et al., 2008b). These features initiate when
permafrost thaw extends beyond the base of the active layer, resulting in a thick,
supersaturated, soil column typically resting on the top of a weakly cohesive layer of ice.
This supersaturated soil detaches and the entire active layer (vegetation, peat, inorganic
sediments, etc.) begins to flow downhill. Once the active layer has been stripped off the
surface, the exposed permafrost thaws rapidly, forming gullies that focus surface
overflow and erosive energy into a small network of channels that thaw and transport
sediments downslope. The gullies erode both vertically and headward, generating a
distinctive arcuate depression, whose edge is defined by a headwall that delineates the
boundary between the thermokarst feature and the stable frozen ground. In a
2
retrogressive thaw slump, new material is supplied by the continued thawing of
permafrost and subsequent soil instability in the headwall of the feature (Jorgenson et al.,
2008a).
Thermokarst features cause fundamental alterations in the physical and ecological
processes within arctic ecosystems. These changes include, but are not limited to: mass
wasting of soils, re-organization of soil profiles, changes in surface albedo, decreases in
soil pore ice, exposure of carbon and nutrient pools contained within frozen soils,
increased carbon release to the atmosphere, temperature profiles, soil moisture changes,
as well as changes in vegetation, hydrology, and topography (Jorgenson and Osterkamp,
2005).
1.1.2
Soil Organic Carbon
Permafrost affected soils and peatlands in the arctic typically function as sinks for
soil organic carbon (SOC), due to the slow decomposition rates of litter input, and
increased retention/storage rates within the ecosystem (Grosse et al., 2011). Stabilization
of soil organic matter is determined by the interaction of cold temperatures, permafrost,
saturation, and substrate quality. However, the importance of these factors in the global
carbon budget and how they will respond to a changing climate are poorly understood
(Hobbie et al., 2000; Tarnocai, 2009). Thawing permafrost and subsequent thermokarst
development has caused arctic SOC pools to become more available for microbial
decomposition, thus leading to an increase in the production and emission of the
greenhouse gases, carbon dioxide (CO2) and methane (CH4) to the atmosphere (Rowland
et al., 2010). Many workers have attributed increased decomposition rates to altered soil
properties, such as increased temperature and soil moisture, as a result of ecosystem
3
changes due to thermokarst development (Lee et al., 2011; Michaelson and Ping, 2003;
Michaelson et al., 2011). Stieglitz et al. (2000) suggested that drier conditions in the
arctic would increase aerobic soil decomposition, increasing CO2 efflux, leading to a net
ecosystem loss of carbon to the atmosphere. In contrast, cooler and moister conditions
result in slower anaerobic decomposition, resulting in overall ecosystem carbon
sequestration (Stieglitz et al., 2000).
1.1.3
Soil Properties and Structure
At a local scale, it is poorly understood how thermokarst-driven changes in soil
properties and structure will affect carbon emissions. Organic soils (peatlands) and
cryoturbated permafrost-affected mineral soils have the highest mean soil organic carbon
contents: 32.2 – 69.6 kg m-2 (Tarnocai, 2009). Thermal erosion features disturb, and
rework arctic soils, increasing the availability of soil organic matter for microbes to
decompose, leading to an increase in soil respiration (Schuur et al., 2008). Carbon
emission from soil can take the form of either carbon dioxide (CO2) or methane (CH4).
Saturated, low-oxygenated environments (anaerobic conditions) are likely to contain
microbes that reduce CO2 to CH4, through the processes of methanogenesis, which has
25 times more potential to warm the atmosphere than CO2 (Schuur and Abbott, 2011).
Conversely in drier, warmer environments (aerobic conditions), soil respiration results in
the release of CO2. Experimental incubations of permafrost soils over a 500 day period at
15°C showed a 3.9-10.0 times greater CO2 release under aerobic rather than anaerobic
soil conditions (Lee et al., 2012). Deep permafrost mineral soils have also been shown to
release similar amounts of carbon as organic soils for some soil types when compared on
4
a per gram carbon basis (Lee et al., 2012). These findings suggest that permafrost carbon
may be very labile, but significant differences exist across soil types.
1.1.4
Carbon Dioxide Efflux
Carbon dioxide efflux is primarily a product of heterotrophic respiration within
the soil environment. CO2 efflux is a product of microbial processes and passes through
soil to the atmosphere following Fick’s law of diffusion. CO2 concentrations vary as a
function of soil depth and are dependent on production and consumption rates in the case
of anaerobic soils and the diffusive ability of the soil profile (Amundson and Davidson,
1990). Controls on the release of CO2 within the soil profile can be contributed to
physical soil properties, soil organic matter content, temperature, and moisture
(Amundson and Davidson, 1990). For example, increased soil organic matter quality
coupled with increased soil temperature results in higher soil CO2 efflux.
Carbon isotope values, which are the ratio of 13C/12C relative to Vienna Pee Dee
Belemnit (VPDB) standard, represented as δ13C, also vary as a function of depth, which
is a function of soil organic matter decomposition and interactions with atmospheric CO2.
The factors controlling δ13C soil profiles are complex but many models suggest that
these profiles, in part, are controlled by decomposing organic matter, the difference in 12C
and 13C diffusion coefficients, and the rate at which CO2 is produced by biological
activity within the soil (Amundson and Davidson, 1990). It has been noted by numerous
studies that CO2 concentrations in the soil increase with depth due to atmospheric mixing
at the surface to values dependent upon the production rate (Amundson and Davidson,
1990; Amundson et al., 1998; Shanhun et al., 2012). While δ13C values decrease from
atmospheric values (-8‰) near the surface to values approaching that of the decomposing
5
organic matter/root respiration with increasing depth (-27‰) (Amundson and Davidson,
1990; Amundson et al., 1998; Shanhun et al., 2012).
Rates of CO2 release from thermal erosional features have been estimated by
numerous studies using both continuous and random measurements of CO2 efflux. (Lee
et al., 2011; Lee et al., 2010; Schuur et al., 2009; Vogel et al., 2008a). Overall, the
terrestrial ecosystems of northern regions are a net source of carbon to the atmosphere at
276 Tg C yr-1 (Tarnocai, 2009). Given current warming trends, it is projected that the
arctic regions will remain a net source of carbon, and could possibly double carbon
released to the atmosphere, to 473 Tg C yr-1 in the next 30 years (Zhuang et al., 2006).
At present, the carbon contribution from thermokarst features to the projected increase in
carbon release is poorly quantified.
Detailed studies of thermokarst carbon cycling and efflux have been conducted
with varying results. Lee et al. (2010) found that for thermokarst features, CO2 release to
the atmosphere averaged between 177 to 270 g CO2- C m-2 per year over a three year
period. These rates were least in the least-disturbed, moist, acidic tundra soils and
greatest where thawing of permafrost and thermokarst formation was most pronounced.
Over half of the variability in CO2 production for this study was explained by surface
subsidence, soil temperature, and site differences. Schuur et al. (2009) examined
thermokarst CO2 respiration along a thaw gradient, and found that on an annual basis the
extensive thaw site had significantly higher CO2 efflux, with greater total CO2 loss than
the minimal thaw site, while the moderate site had intermediate CO2 efflux between the
two. Ground subsidence, permafrost thaw and increased microbial activity accounted for
the above observed trends. Other research conducted by Vogel et al. (2009) provide
6
contrasting results; moderately thawed sites were a sink for CO2, and minimally thawed
sites remained neither a source nor a sink for CO2. The differing results of carbon efflux
from thermokarst terrain have complicated researcher’s ability to spatially scale up to
extract an overall carbon release from these features. Clearly, more studies are needed to
help quantify the overall impact of thermokarst features on total C release from the
Arctic.
1.1.5
Controls on CO2 Efflux
Soil temperature, moisture, and substrate quality (%SOC) exhibit interactive
controls on CO2 flux rates within the arctic ecosystem (Hobbie et al., 2000; Huemmrich
et al., 2010; Michaelson and Ping, 2003; Oberbauer et al., 2007). Many studies have
shown a positive linear correlation between CO2 efflux and soil temperature (Huemmrich
et al., 2010; Oberbauer et al., 2007). Results of an experimental warming study indicated
that at -2°C, the CO2 respiration for both the organic and mineral horizons are about the
same (Michaelson and Ping, 2003). However when warmed to 4°C, average respiration
rates were twice as high in organic soil horizons than those found for mineral horizons
(Michaelson and Ping, 2003). Oberbauer et al. (2007) conducted numerous field warming
studies around Toolik Lake and Barrow, Alaska. They found that, in general warming
increased soil respiration with the largest increase observed at drier sites (Oberbauer et
al., 2007).
The interaction of soil temperature and moisture has been shown to affect the
rates of CO2 efflux. For example Huemmrich et al. (2010) determined that the water table
determined whether the ecosystem was a source or sink for carbon, with temperature
modifying the strength of the source or sink. Increased moisture can shift the soil
7
environment to anaerobic conditions whereby the majority of carbon in the system may
be released as CH4 instead of CO2. Signigicant knowledge gaps exist in the area of how
the physical structure of soils relates to carbon efflux in arctic environments.
1.1.6
δ13C of Soil Respired CO2
δ13C of soil CO2 and CH4 allow for an examination of how thermokarst features
alter soil profiles by either mixing or compacting soil. Variations in carbon isotope
composition (δ13C) between soil organic matter and CO2 efflux enable tracing sources
and sinks of respired CO2. However, this approach has rarely been applied in arctic
ecosystems. In general, δ13C of soil respired CO2 becomes more negative with depth in
soil profiles (Amundson and Davidson, 1990). This trend can be attributed to mixing with
atmospheric CO2 at the surface (less negative), and values approaching that of the
decomposing organic matter/root respiration (more negative) with increasing soil depth
(Flanagan et al., 2005). Variations in the above trend, are a function of ecosystem
productivity, amount of organic carbon burial and vegetation type. It is predicted that
thermokarst features will re-work soil profiles by unburying older sources of carbon
allowing them to be worked on by microbial processes. We hypothesis that typical δ13C
soil profiles will not be seen within thermokarst features. Rather these profiles will show
the influence of older carbon δ13C values (more negative) closer to the surface due to
accelerated mixing by mass wasting.
Research on δ13C vertical soil profiles within the arctic ecosystem is sparse, and
represents a knowledge gap within overall arctic carbon cycling theory. It has been
shown that, in general, SOM at wetter sites has lower δ13C values. This is due to less
drought stress on plants that contribute to δ13C values as litterfall (Flanagan et al., 2005).
8
Research in western Canada along a transect from boreal forest to tundra, shows a
decrease in δ13C of 0.7-0.9‰ in the upper 30cm of soil (Bird et al., 2002). Recent work in
Antarctica and peatland bogs in northern England has investigated the effects of
temperature, soil moisture and substrate quality (% SOC) on the flux and carbon isotope
content of soil-respired CO2 (Hardie et al., 2011; Shanhun et al., 2012). For these studies,
temperature and substrate quality were main drivers of isotopic C compositions within
the soil profile. This research noted that increased temperatures increased the
fractionation of 13C and that an increase in lignin composition down the soil profile lead
to an overall depletion in δ13C.
Within the Arctic, δ13C of soil respired CO2 has been used to determine whether
the soil environment is aerobic or anaerobic. In general, within a thermokarst aerobic soil
conditions occur in well-drained areas, such as slopes and uplands. Conversely, anaerobic
soil conditions occur in poorly-drained areas with water-logged soils, such as lowlands or
areas of excessive ponding (Lee et al., 2012). Within Arctic environments, aerobic soil
conditions tend to produce δ13C of soil respired CO2 values between -23 to -27‰ (Lee et
al., 2012). While anaerobic soil conditions tend to produce values that are heavier,
ranging from -11 to -23‰ (Lee et al., 2012). Less negative δ13C values observed in
anaerobic soils in the Arctic can be contributed to the influence of methanogenesis,
whereby 12C is preferentially used as opposed to 13C. Studies on the effects of carbon
cycling for both anaerobic and aerobic soil conditions within thermokarst systems have
mainly been focused on laboratory incubations of arctic soil, as well as, the forms of
carbon released. It is unclear if the above trends will hold true in a field setting. More
9
studies need to focus on the changes in field δ13C profiles as they relate to both aerobic
and anaerobic conditions.
1.2 Thesis Objective
In this thesis, I use a thaw slump chronosequence composed of active, stabilized
and undisturbed surfaces to examine soil CO2 efflux from a large thermokarst failure. In
2011 and 2012, we quantified soil CO2 efflux during peak summer months. I use these
data to examine the controls of soil moisture, temperature, as well as, soil physical
properties in governing the release of CO2 across the chronosequence. I hypothesize that
soil CO2 efflux will be greater in the active slump and stabilized slump features as
opposed to the tundra due to deep thaw and mixing of soil organic carbon pools, and
increased soil temperature and moisture.
We also use soil profiles from the thaw slump chronosequence to examine how
changes in the soil environment affect CO2 production across the soil column, as well as
the resulting surface CO2 efflux. In 2012 we measured CO2 and CH4 soil concentrations,
δ13C of respired soil CO2 . I use these data to examine the controls of porosity, effective
diffusivity, soil temperature, moisture, as well as other physical soil properties in
governing the production and release of CO2 across the chronosequence.
1.3 Thesis Structure
Chapter one provides a general overview of the relevant literature, including what
is poorly understood about carbon-cycling within artic soils. The second chapter
investigates soil CO2 efflux along a thaw slump chronosequence, as well as the controls
of soil temperature, moisture, bulk density and soil organic matter content on surface CO2
efflux. The third chapter investigates how changes in physical soil properties across thaw
10
slump chronosequence affects soil profile carbon-cycling and the resulting CO2 efflux.
The fourth chapter serves as a brief discussion of the implications and conclusions of
chapters two and three.
11
1.4 References
ACIA, 2004, Impacts of a Warming Climate: Arctic Climate Impact Assessment.
Alewell, C., R. Giesler, J. Klaminder, J. Leifeld, and M. Rollog, 2011, Stable carbon
isotopes as indicators for micro-geomorphic changes in palsa peats:
Biogeosciences Discussions, v. 8, p. 527-548.
Amundson, R., and E. Davidson, 1990, Carbon dioxide and nitrogenous gases in the soil
atmosphere: Journal of Geochemical Exploration, v. 38, p. 13-41.
Amundson, R., L. Stern, T. Baisden, and Y. Wang, 1998, The isotopic composition of
soil and soil-respired CO< sub> 2</sub>: Geoderma, v. 82, p. 83-114.
Bird, M., H. Santruckova, J. Lloyd, and E. Lawson, 2002, The isotopic composition of
soil organic carbon on a north-south transect in western Canada: European
Journal of Soil Science, v. 53, p. 393-403.
Flanagan, L. B., J. R. Ehleringer, and D. E. Pataki, 2005, Stable Isotopes and BiosphereAtmosphere Interactions: Processes and Biological Controls, Elsevier, 310 p.
Grosse, G., J. Harden, M. Turetsky, A. D. McGuire, P. Camill, C. Tarnocai, S. Frolking,
E. A. G. Schuur, T. Jorgenson, S. Marchenko, V. Romanovsky, K. P. Wickland,
N. French, M. Waldrop, L. Bourgeau-Chavez, and R. G. Striegl, 2011,
Vulnerability of high-latitude soil organic carbon in North America to
disturbance: J. Geophys. Res., v. 116, p. G00K06.
Hardie, S. M. L., M. H. Garnett, A. E. Fallick, A. P. Rowland, N. J. Ostle, and T. H.
Flowers, 2011, Abiotic drivers and their interactive effect on the flux and carbon
isotope (14C and δ13C) composition of peat-respired CO2: Soil Biology and
Biochemistry, v. 43, p. 2432-2440.
Hesterberg, R., and U. Siegenthaler, 1991, Production and stable isotopic composition of
CO2 in a soil near Bern, Switzerland: Tellus B, v. 43, p. 197-205.
Hinzman, L., N. Bettez, W. Bolton, F. Chapin, M. Dyurgerov, C. Fastie, B. Griffith, R.
Hollister, A. Hope, H. Huntington, A. Jensen, G. Jia, T. Jorgenson, D. Kane, D.
Klein, G. Kofinas, A. Lynch, A. Lloyd, A. McGuire, F. Nelson, W. Oechel, T.
Osterkamp, C. Racine, V. Romanovsky, R. Stone, D. Stow, M. Sturm, C.
Tweedie, G. Vourlitis, M. Walker, D. Walker, P. Webber, J. Welker, K. Winker,
and K. Yoshikawa, 2005, Evidence and Implications of Recent Climate Change in
Northern Alaska and Other Arctic Regions: Climatic Change, v. 72, p. 251-298.
Hobbie, S. E., J. P. Schimel, S. E. Trumbore, and J. R. Randerson, 2000, Controls over
carbon storage and turnover in high-latitude soils: Global Change Biology, v. 6, p.
196-210.
Huemmrich, K. F., G. Kinoshita, J. A. Gamon, S. Houston, H. Kwon, and W. C. Oechel,
2010, Tundra Carbon Balance under Varying Temperature and Moisture
Regimes: Journal of Geophysical Research, v. 115, p. 154-175.
IPCC, 2007, Climate Change 2007: Synthesis Report. Contribution of Working Groups I,
II and III to the Fourth Assessment Report of the Intergovernmental Panel on
Climate Change, v. 446, IPCC, 104 p.
Jorgenson, M. T., and T. E. Osterkamp, 2005, Response of boreal ecosystems to varying
modes of permafrost degradation: Canadian Journal of Forest Research, v. 35, p.
2100-2111.
12
Jorgenson, M. T., V. Romanovsky, J. Harden, Y. Shur, J. O’Donnell, E. A. G. Schuur, M.
Kanevskiy, and S. Marchenko, 2010, Resilience and vulnerability of permafrost to
climate change: Canadian Journal of Forest Research, v. 40, p. 1219-1236.
Jorgenson, M. T., Y. L. Shur, and T. E. Osterkamp, 2008a, Thermokarst in Alaska, Ninth
International Conference on Permafrost, p. 1-5.
Jorgenson, M. T., Y. L. Shur, and T. E. Osterkamp, 2008b, Thermokarst in Alaska, Ninth
International Conference on Permafrost.
Lee, H., E. Schuur, J. Vogel, and M. Lavoie, 2011, A spatially explicit analysis to
extrapolate carbon fluxes in upland tundra where permafrost is thawing: Global
Change Biology, p. 1379-1393.
Lee, H., E. A. G. Schuur, K. S. Inglett, M. Lavoie, and J. P. Chanton, 2012, The rate of
permafrost carbon release under aerobic and anaerobic conditions and its potential
effects on climate: Global Change Biology, v. 18, p. 515-527.
Lee, H., E. A. G. Schuur, and J. G. Vogel, 2010, Soil CO2 production in upland tundra
where permafrost is thawing: Journal of Geophysical Research: Biogeosciences,
v. 115, p. G01009.
Michaelson, G. J., and C. L. Ping, 2003, Soil Organic Carbon and CO2 Respiration at
Subzero Temperature in Soils of Arctic Alaska: Journal of Geophysical Research,
v. 108, p. 8164.
Michaelson, G. J., C. L. Ping, and M. T. Jorgenson, 2011, Methane and carbon dioxide
content in eroding permafrost soils along the Beaufort Sea coast, Alaska: Journal
of Geophysical Research, v. 116, p. G01022.
Oberbauer, S. F., C. E. Tweedie, J. M. Welker, J. T. Fahnestock, G. H. R. Henry, P. J.
Webber, R. D. Hollister, M. D. Walker, A. Kuchy, E. Elmore, and G. Starr, 2007,
Tundra CO2 fluxes in response to experimental warming across latitudinal and
moisture gradients: Ecological Monographs, v. 77, p. 221-238.
Rowland, J. C., C. E. Jones, G. Altmann, R. Bryan, B. T. Crosby, L. D. Hinzman, D. L.
Kane, D. M. Lawrence, A. Mancino, P. Marsh, J. P. McNamara, V. E.
Romanvosky, H. Toniolo, B. J. Travis, E. Trochim, C. J. Wilson, and G. L.
Geernaert, 2010, Arctic Landscapes in Transition: Responses to Thawing
Permafrost: Eos Trans. AGU, v. 91.
Schuur, E., and B. Abbott, 2011, High risk of permafrost thaw: Nature, v. 480, p. 32-33.
Schuur, E. A. G., J. Bockheim, J. G. Canadell, E. Euskirchen, C. B. Field, S. V.
Goryachkin, S. Hagemann, P. Kuhry, P. M. Lafleur, H. Lee, G. Mazhitova, F. E.
Nelson, A. Rinke, V. E. Romanovsky, N. Shiklomanov, C. Tarnocai, S.
Venevsky, J. G. Vogel, and S. A. Zimov, 2008, Vulnerability of Permafrost
Carbon to Climate Change: Implications for the Global Carbon Cycle:
BioScience, v. 58, p. 701-714.
Schuur, E. A. G., J. G. Vogel, K. G. Crummer, H. Lee, J. O. Sickman, and T. E.
Osterkamp, 2009, The effect of permafrost thaw on old carbon release and net
carbon exchange from tundra: Nature, v. 459, p. 556-559.
Shanhun, F., P. Almond, T. Clough, and C. Smith, 2012, Abiotic processes dominate CO2
fluxes in Antarctic soils: Soil Biology & Biochemistry, v. 53, p. 99-111.
Stieglitz, M., A. Giblin, J. Hobbie, M. Williams, and G. Kling, 2000, Simulating the
effects of climate change and climate variability on carbon dynamics in Arctic
tundra: Global Biogeochem. Cycles, v. 14, p. 1123-1136.
13
Tarnocai, C., 2009, Soil organic carbon pools in the northern circumpolar permafrost
region: Global Biogeochemical Cycles, v. 23, p. 1-11.
Vogel, J., E. A. G. Schuur, C. Trucco, and H. Lee, 2008, Response of CO2 exchange in a
tussock tundra ecosystem to permafrost thaw and thermokarst development:
Journal of Geophysical Research, v. 114, p. G04018.
Zhuang, Q., J. Melillo, M. Sarofim, and D. Kicklighter, 2006, CO2 and CH4 exchanges
between land ecosystems and the atmosphere in northern high latitudes over the
21st century: Geophysical Research Letters, v. 33.
14
2
Soil properties affect variations in carbon dioxide efflux across a
retrogressive thaw slump chronosequence in northwest Alaska.
2.1 Abstract
Warming of the arctic landscape results in permafrost thaw, which causes ground
subsidence or thermokarst. On hillslopes, these features can grow rapidly, altering soil
structure and likely temperature, moisture, as well as the content and quality of soil
organic matter. In turn, these changes likely affect both the rate and mechanism of
carbon cycling in permafrost soils. This is a major concern because permafrost soils
store nearly half the world’s global carbon. In order to assess the time-evolution of
carbon contributions from features like these, we use a three-stage chronosequence of
thaw slumps to describe the differences in carbon dioxide (CO2) fluxes from undisturbed,
active and stabilized features. Our study site, found along the Selawik River in northwest
Alaska, contains one of the largest active retrogressive thaw slumps known in North
America and had a surface area of ~50,000 m2 in 2012. We designate the undisturbed
tundra as the initial condition and an adjacent, bowl-like depression revegetated by
multiple generations of black spruce trees as the stabilized slump.
In this 2-year study, we use peak summer measurements of CO2 efflux across the
chronosequence to determine how belowground carbon emissions compare among
features of different maturity. Using measurements of soil temperature, soil moisture,
soil organic matter, and bulk density, we also evaluate how environmental factors control
CO2 emissions within each of the three stages. CO2 efflux from the active slump is less
than half that observed in either the undisturbed tundra or stabilized slump. Although the
floor of the active slump is generally 10 to 15°C warmer than the tundra and stabilized
slump, we did not observe soil temperature or moisture to exert a strong control on CO2.
15
Rather we found local physical soil properties such as soil organic matter and bulk
density to be strongly and inversely related among these features (r2=0.97) and could
potentially explain ~50% of the variation in CO2 efflux from our system. Because of the
low organic matter content in the active slump (1%-3%), as well as bulk densities close to
particle density (~2.3g-dry/cm3), there is a substrate quality and diffusivity limitation
within the active slump which decreases CO2 efflux. Our findings suggest that though
thaw slumps can profoundly alter the biophysical controls on CO2 production, the
mixture of low-organic content mineral substrates with peaty soils actually reduces CO2
efflux below initial conditions. As the feature matures and is recolonized as a boreal
forest ecosystem, fluxes from this altered stable state return toward initial conditions. In
order to assess the significance of catastrophic thermokarst on global carbon budgets,
future studies must assess whether CO2 fluxes from features in high organic content
substrates are significantly elevated above initial conditions as well as provide higher
spatial resolution maps of carbon content in thawing permafrost soils.
2.2 Introduction
Permafrost degradation and thermokarst features are dramatically altering carbon
cycling across the Artic landscape (Grosse et al., 2011; Rowland et al., 2010; Schuur et
al., 2008; Tarnocai, 2009). Although numerous types of thermokarst exist across the
Arctic landscape (Jorgenson et al., 2008b), carbon cycling studies primarily measure
fluxes from gradual ground subsidence in low relief terrain (Lee et al., 2010; Schuur et
al., 2009). At the other end of the spectrum, fluxes from rapid erosional features such as
retrogressive thaw slumps have received little attention. The increasing recognition of
these catastrophic thermokarst throughout the arctic (Gooseff et al., 2009; Jorgenson et
16
al., 2008b; Lacelle et al., 2010), creates an explicit need to quantify soil carbon dioxide
(CO2) emissions and study the governing controls within these features.
Warming trends in the arctic and associated landscape changes have received
international attention and are projected to increase with future changes in climate
(ACIA, 2004; IPCC, 2007; Winton, 2006). Over the past 30 years, temperatures within
Alaska have risen by 1-3°C (ACIA, 2004). Results from general circulation models differ
somewhat regarding future trends, but models and scenarios selected for the Arctic
Climate Impact Assessment (2004) generally agree that average annual temperatures in
the arctic will increase a further 3-5°C and winter temperatures may increase by 4-7°C by
2100. Model scenarios also suggest that rising temperatures will be accompanied by
increased precipitation, mostly in the form of rain.
This warming trend has accelerated widespread permafrost degradation throughout
Alaska (Jorgenson et al., 2010; Jorgenson et al., 2006). A direct consequence of this
degradation is thermokarst, where thaw leads to ground subsidence (Jorgenson et al.,
2010; Rowland et al., 2010). Thermokarst features directly impact soil and hydrologic
properties, ecosystem energy balance, increase exposure of soil carbon and nutrient pools
due to active layer deepening, as well as changes to vegetation and landscape topography
(Jorgenson and Osterkamp, 2005). Thermokarst can broadly be characterized by either
gradual thaw or catastrophic failure. Gradual thaw occurs in areas with high soil organic
matter content, low relief, and stable topography result in increased water storage
associated with active layer deepening (Jorgenson et al., 2008b). In contrast, catastrophic
failures occur on hillslopes, where vegetation is typically stripped, exposing bare soil
which causes a multitude of cascading effects: increased exposure to incoming solar
17
radiation, decreased available soil organic matter, exposure of buried soil material,
increased erosion by surface overflow and rapid fluctuations in soil moisture (Gooseff et
al., 2009).
Progressive permafrost degradation has the potential to release of vast stores of
previously frozen carbon contained within permafrost soils. Through thaw, this carbon
becomes available for microbial decomposition, causing the release of greenhouse gases
such as CO2 and methane (CH4) to the atmosphere through increased soil respiration
(Amundson and Davidson, 1990; Schuur and Abbott, 2011). This transfer of organic
carbon from permafrost to the atmosphere could induce a positive feedbacks with global
temperatures (Schuur et al., 2008). However, the magnitude and timing of carbon release
is poorly constrained due to changes in soil physical properties associated with
permafrost thaw, uncertainty in how permafrost will thaw across a heterogeneous
landscape, as well as temperature and moisture constraints on the decomposition of
previously frozen carbon (Grosse et al., 2011; Hinzman et al., 2005; Schuur and Abbott,
2011; Schuur et al., 2008).
Most CO2 efflux studies in the Arctic have been small in spatial scale (< 10m2)
and conducted in areas of gradual thaw, where active layer thaw deepens into organic
rich substrates and changes the rate of carbon cycling within the ecosystem (Lee et al.,
2011; Schuur et al., 2009; Vogel et al., 2009). These studies generally show that CO2
production increases as the thermokarst matures. In recent years, catastrophic
thermokarst failures have been increasing in frequency across the Alaskan landscape
(Gooseff et al., 2009), yet few studies have examined CO2 efflux from these features as
well as the physical and biological controls on these fluxes.
18
Studies have shown that soil temperature, moisture, and substrate quality exhibit
interactive controls on CO2 efflux rates within the arctic ecosystem (Huemmrich et al.,
2010; Michaelson and Ping, 2003; Oberbauer et al., 2007). In general, increases in soil
temperature lead to an increase in CO2 soil respiration, with the largest increase observed
under drier conditions (Huemmrich et al., 2010). Increasing soil moisture has the effect of
shifting CO2 production from an aerobic process to an anaerobic process, whereby the
majority of carbon is released in the form of methane (Anthony, 2009). It is unclear how
catastrophic thermokarst will alter soil moisture and temperature controls on soil CO2
efflux and whether these features are sources or sinks of carbon to the atmosphere.
Here we use a thaw slump chronosequence to examine soil CO2 efflux from a
catastrophic thermokarst failure. In 2011 and 2012, we quantify soil CO2 efflux during
peak summer months and examine the influence of soil moisture, temperature, as well as,
soil physical properties on the release of CO2 across the chronosequence. We hypothesize
that soil CO2 efflux should be higher in the active slump and stabilized slump as opposed
to the undisturbed tundra due to deeper thaw and mixing of soil organic carbon pools, and
increased soil temperature and moisture.
2.3 Field Setting
The Selawik slump field area is located along the upper Selawik River
(157°36”43”W, 66°29’54”N) in northwest Alaska (FIGURE 1). This area is composed of
an upland tussock shrub tundra community, an active retrogressive thaw slump that
initiated in 2004, and a stabilized retrogressive thaw slump that is estimated to be >500
years old.
19
Mean annual precipitation in this area is 256 mm per year and mean annual air
temperature is -5.67°C. Soil temperature regime is gelic with a mean soil temperature of 4.67°C and the soil moisture regime is classified as aquic (Schoeneberger et al., 2002).
The parent material in this area is a glacial diamict composed of poorly sorted silt, gravel
and cobble overlain by fine grained loess (Lanan, 2013). The active layer (the upper
portion of the soil column that freezes and thaws each year) varies between the different
sites and extends to depths of 0.3 to 0.8 m.
Because this field site is remote (~108 km east of Selawik Village) and difficult to
access during spring, fall, and winter (helicopter only), no sampling occurred during
these seasons. Sampling focused on summer months when we expected to observe the
highest CO2 effluxes owing to higher temperatures and biologic productivity.
2.4 Study Deisgn: a chronosequence enables a space for time substitution
We used a space for time substitution (thaw slump chronosequence) to examine
the evolution of thaw slumps and their subsequent stabilization. Our field site includes
three stages of feature evolution: undisturbed tundra, active slump, and stabilized slump
(FIGURE 1). Other factors such as climate, slope, parent material, and biota are all held
relatively constant (Chapin et al., 2011; Jenny, 1941). Since there are no biotic barriers
across this chronosequence, differences in vegetation are a function of time since
disturbance. In our chronosequence conceptual model, the final state of our system does
return to initial tundra conditions. We postulate that the ending trajectory of the active
slump will be similar in nature to the forested stabilized slump feature, due to
geomorphic and ecosystem drivers.
20
The tundra environment has not undergone any physical disturbance and
represents our arctic landscape reference condition or control. The tundra ~20m north of
the Selawik slump headwall (FIGURE 1). Primary vegetation consists of sedge, tussocks
(Eriophorum vaginatun, and Carex spp.) with interspersed moss and lichen. Summer
active layer thickness ranges from 30 to 60 centimeters. A Gellic Typic Fibristels, which
is characterized by a wet, poorly decomposed thick (15-20 cm) peat, that transitions to a
mineral soil with permafrost in the upper meter. This soil is also characterized as loamy
that is poor to moderately drained (Jorgenson et al., 2009) (TABLE 2).
The active slump initiated in 2004 is one of largest known active retrogressive
thaw slump in the world. This feature has mobilized more than a half million cubic
meters of ice and sediment, and is growing at a rate of ~20 m/year. The slump floor is
approximately 50,000 m2, with a headwall measuring 15-18 m tall. The headwall is
composed mostly of glacial diamict and loess, with a small (20 cm) upper cap of organic
rich peat and tussock community as previously described for the tundra. Because the
active slump grows quickly and the vast majority of the eroded material is diamict, the
slump floor is largely composed of overlapping mudflows of unsorted silts and gravels
with no detectable soil profiles. Three percent of the slump floor is mantled by vegetated
rafts (0.05-0.5 m2 each), which are composed of intact pieces of tundra that have fallen
off the headwall and been transported with sediment downslope. Frost probing the slump
floor revealed no permafrost in the upper meter (TABLE 2).
The stabilized slump is directly adjacent and west of the active slump (FIGURE
1) and represents a likely trajectory of our active thaw slump. Following catastrophic
failure, this area has since been recolonized by black spruce, alder (Picea mariana and
21
Alnus spp.), and various grasses. Soil profiles have been re-established and are classified
as Gellic Typic Aquorthels, which are wet mineral soils over permafrost, which lack
cryoturbation. This soil type is characterized as loamy and is poorly to moderately
drained (Jorgenson et al., 2009). Summer active layer thicknesses range from 50 to 80 cm
(TABLE 2). The feature is thought to be greater than 500 years old based on the presence
of numerous generations of black spruce logs found decaying on the surface. The active
slump has eroded the eastern flank of the stabilized slump revealing both a texture of the
ancient thaw slump deposit and the new soil profiled developed above.
2.5 Methods
2.5.1
Meteorology
Continuous air temperature and precipitation were collected in thirty minute
intervals from a weather station located approximately 50 m due north of the active
slump headwall in the tundra domain (FIGURE 1). Air temperature was collected using
a shielded sensor 215 cm above ground. Precipitation was collected using a tippingbucket 242 cm above ground. Data were collected from January to September in 2011,
and from June to early August in 2012. Due to the close spatial proximity of our
chronosequence to the weather station, we assume precipitation and air temperature data
are consistent throughout the study area.
2.5.2
Point measurements of CO2 Efflux, Soil Temperature, Soil Moisture
Sampling sites for soil CO2 efflux, soil temperature, and soil moisture were
established along our defined chronosequence: six sampling sites within the tundra,
eleven sites within the active slump, and nine sites within the stabilized slump (26 total,
FIGURE 1). Sites within a given region sample different types of vegetation and
22
disturbance to assess heterogeneity within each category. After sites were established, a
12.7x10.2 cm polyvinyl chloride pipe (soil collar) was driven 8cm into the upper portion
of the soil column. The soil was then allowed to stabilize for twenty-four hours prior to
any measurements (LI-COR, 2007). In July of 2011, each site was sampled
approximately five times in a six-day sampling period. In July and August of 2012, each
site was sampled approximately twelve times over an eighteen-day sampling period.
Sampling times for each collar were distributed throughout the day and night to average
across expected temporal variability in measured parameters.
Soil CO2 efflux, soil temperature, and soil moisture were measured using a LiCor8100A analyzer (Lincoln, NE). CO2 efflux was collected using a LiCor-8100-102, 10cm
diameter survey chamber attachment, placed on top of the soil collar platforms and set at
a 1.5 minute sampling interval. Soil temperature was measured at depths of 5, 10, and 15
cm next to the chamber with a LiCor 8100-203 soil thermistor probe. Volumetric water
content (VWC) (volume water/volume soil), hereafter referred to as soil moisture, was
measured using a Delta-T ML2x probe (Lincoln, NE) at a depth of 10cm. Post-processing
of soil CO2, temperature and soil moisture was conducted using the LI-8100 Data File
Viewer (v2.0). Soil volumetric water content was adjusted for high organic content soils
using the equation provided in the Li-COR manual (2007).
2.5.3
Physical Soil Properties
In 2012, soil bulk density and soil organic matter content were determined for a
subset of collars along the chronosequence. Two rectangles of known volume were
excised from the soil collar after all flux measurements had been completed. The soil
sample was subsequently frozen and transported back to Idaho State University where
23
they were analyzed for soil organic matter (SOM) and bulk density following methods
described by Schumacher (2002) and Blake and Hartge (1986), respectively. In brief,
SOM was determined by loss-on-ignition by combusting 100 g of dry soil for 24 hours at
500°C and reweighing the remaining mass. Bulk density was determined by drying the
soil at 105°C and weighing the dry mass and sieving the soil to retain the fine fraction <2
mm.
2.5.4
Statistical Analysis
Repeated CO2, soil temperature and moisture measurements were averaged for
each ring location each day within the chronosequence environment. CO2 efflux
measurements were log transformed to improve normality and homoscedasticity. Other
parameters such as soil moisture and soil temperature were normally distributed and thus
did not require transformation. ANOVA and Tukey-Kramer HSD statistical tests for soil
CO2 efflux, soil temperature, and soil moisture were used to determine statistical
differences within each environment along the chronosequence (TABLE 1). Collars
within each environment (tundra, active, and stabilized slump) showed no significant
differences within each environment so that data collected within each environment was
pooled for both 2011 and 2012. ANOVA tests and Tukey-Kramer HSD were then
conducted to determine statistical differences between chronosequence environments for
2011 and 2012. Linear and multiple linear regression models were used to determine
environmental controls on CO2 efflux. All statistical analyses were performed in JMP 10
(Cary, NC) software.
24
2.6 Environmental Conditions over Space and Time
2.6.1
Precipitation
Both magnitude and intensity of precipitation during the 2012 campaign were
higher than 2011. Cumulative precipitation during the 2011 sampling period (FIGURE
2a) was relatively low, totaling 9 mm from two small events. In contrast, the 2012
sampling period was much wetter, including three significant events, the largest of which
occurred during the first week of sampling and delivered 20 mm. The second and third
events occurred in the latter half of the 2012 campaign each delivered approximately 5
mm.
2.6.2
Air Temperature
Air temperature was generally higher in 2011 than during the 2012 sampling period
(FIGURE 2b). Average air temperature during the 2011 sampling period was ~16±5.3°C,
and showed an overall cooling trend. In contrast, 2012 average air temperature was
lower, ~14±3.8°C, and showed a slight warming trend in the first half of the sampling
period, followed by a slight cooling trend throughout the latter half of the period. For
both 2011 and 2012, periods of precipitation reduced the magnitude of the diel cycle in
air temperature.
2.6.3
Soil Temperature
Average daily soil temperatures at 15 cm were consistently higher in 2011 for all
three chronosequence environments (FIGURE 2c). This mean behavior corresponded
with the higher air temperatures and lower precipitation in 2011. Within individual years,
similar correlations existed between measurements of air temperature, precipitation and
soil temperature. These correlations were most apparent in the active slump.
25
Soil temperatures were significantly different among the environments for both
2011 and 2012 (2011 F2,21=64.07, p<0.0001; 2012 F2,25=78.4, p=0.0001) (FIGURE
3a)(TABLE 1). Post-hoc Tukey tests showed that although there are no significant
differences between soil temperatures in the tundra and the stabilized slump, soil
temperatures were significantly elevated; 3-4 times higher, in the active slump relative to
these two environments (p<0.05). We attribute these differences to the higher thermal
conductivity and lower albedo of the dark, bare mineral soil in the active slump. The
abundant, multi-story vegetation at both the tundra and stabilized slump insulate the soil
and shade against solar radiation, dampening variability in soil temperature. The tundra
and stabilized slump also have thinner active layers, placing permafrost closer to the
surface.
2.6.4
Soil Moisture
Average daily soil moisture for both the 2011 and 2012 seasons were more
temporally variable than soil temperatures, with generally higher soil moistures observed
in the 2012 season (FIGURE 2d). The tundra had the highest soil moisture, whereas the
mineral soil in the active slump has the lowest soil moisture in both years.
Average soil moisture differed significantly among chronosequence environments
(FIGURE 3b) in both years with the tundra having higher soil moisture than the other two
environments (2011 F2,21=7.52, p<0.0039; 2012 F2,25=5.08, p=0.015)(TABLE 1). Posthoc Tukey tests revealed no significant differences between the active and the stabilized
slump whereas the tundra was significantly elevated relative to these two environments
(p<0.05). We attribute the differences between the tundra and both the active and
stabilized slump to differences in topographic setting. The tundra in our study area
26
consists mostly of mosses, lichen and tussock communities on a relatively flat lying
slope. These characteristics allow water to have a longer residence time in that
environment. In contrast, both the active and stabilized slumps are located on steeper
surfaces, sloping 10-25°toward the Selawik River such that water likely drains faster,
thus reducing soil moisture. Alternatively, differences in soil organic matter, albedo and
bulk density affecting soil water holding capacity could drive differences in moisture
contents across environments.
2.6.5
Physical Soil Properties
Soil organic matter (SOM) and bulk density differed considerably among the
three environments. SOM content ranged from 1-95%. In the active slump, where the
parent material for this active depositional surface is dominated by glacial diamict, we
observed the lowest SOM content while the stabilized slump contained the highest SOM
content (TABLE 2). The tundra’s SOM values ranged between the active and stabilized
slumps.
Bulk density along the chronosequence was also highly variable (0.1-2.5 gdry/cm3) (TABLE 2). However, a strong inverse exponential correlation existed between
soil organic matter and bulk density and corresponded to different chronosequence
environments (r2=0.97) (FIGURE 5a). The active slump floor, for example, has the
lowest organic matter content because it is composed of sediment remobilized from the
thawing headwall, which is dominated by glacial silts and gravels. The thawed materials
flow away from the headwall as a thick, viscous slurry and stabilize on the slump floor.
Because of the high silt content and low porosity, the active slump samples also had the
highest bulk densities. In contrast, the stabilized slump had the highest SOM values and
27
to the lowest bulk density. Though the stabilized slump had the same bulk density
material at depth, this environment has been completely re-vegetated and has a mature
soil profile with a greater proportion of grasses than the mosses and lichens common in
the tundra environment. This results in lower bulk densities and higher SOM values in
the stabilized slump.
2.7 CO2 Efflux over Space and Time
CO2 efflux was generally greater in 2011 than in 2012 in each of the
chronosequence environments. CO2 efflux for both seasons decreased over the sampling
period (Figure 2e) corresponding to increases in soil moisture as well as slight increases
in soil temperature. This correlation is most evident in the tundra and active slump in
2012.
CO2 effluxes were significant different among the chronosequence environments
(figure 3c) and differed among years (2011, F3,22=12.69 p<0.0001, 2012 F3,25=27.48,
p<0.0001)(TABLE 1). In 2011, fluxes from the tundra and stabilized slump were similar.
The fluxes from the active slump were significantly lower (p<0.05) than those from the
other two environments. Conversely, in 2012, the CO2 fluxes from the tundra, active
slump and stabilized slump were all significantly different from each other. The highest
fluxes came from the stabilized slump and lowest from the active slump.
Measurements of CO2 efflux along the chronosequence for both 2011 and 2012
were scaled up to estimate total grams of carbon lost to the atmosphere per day during
peak growing season (TABLE 2). Total carbon released as CO2 in the tundra for 2011
was greater than in 2012. Numerous studies have determined average carbon release
specifically for the tundra environment via CO2 during the growing season (Giblin et al.,
28
1991; Poole and Miller, 1982; Raich and Schlesinger, 1992) and our observed values
were 1.5 to 2 times higher than these studies report. These differences may be due to a
dampening effect of cooler temperatures inhibiting soil CO2 efflux at the more northerly
comparison sites.
Average carbon released from CO2 in the active slump during peak summer in
2011 was 1.75±0.15 g CO2-C m-2 day-1. In 2012, these fluxes were half that (0.86±0.24 g
CO2-C m-2 day-1). These values fall within the lower range of fluxes observed along a
chronosequence of minimally thawed to extensively thawed gradual thermokarst sites in
Alaska (Lee et al., 2010; Schuur et al., 2009)(TABLE 2). Unlike our system, there has
been no significant vegetation stripping, sediment esposure or remobilization. These
changes have not altered the structure of the soil column beyond deepening the frost
table, allowing more carbon to become functionally available. Compared to our site, their
albedo changes were insignificant. Because of these differences, the values that we
observed for CO2 efflux in the active slump were closer to sites that had been minimally
thawed, but are not directly comparable.
Average carbon release from the stabilized slump during peak summer 2011 was
5.14±0.27 g CO2-C m-2 day-1, while in 2012 we observed only 3.24±0.13 g CO2-C m-2
day-1. Again, our estimates (TABLE 3) fell within in the higher range of measured
estimates for black spruce forests within Alaska (which we use as a proxy for the
stabilized slump) (Raich and Schlesinger, 1992; Schlentner and van Cleve, 1985; Vogel
et al., 2008b). Most of the values from other studies were averaged over a full year. Our
values represent peak growing season and therefore tended to be skewed to the upper end
of existing measurements.
29
2.7.1
Controls Affecting CO2 Efflux
2.7.1.1 Soil Temperature and Moisture
Soil temperature and moisture exhibit a surprisingly weak control on soil CO2
efflux across our chronosequence. In 2011, soil moisture exerts a slightly stronger
control while in 2012, soil temperature exerts a slightly greater control on CO2 efflux
(FIGURE 4). We suggest that during 2011, our system was water-limited, and therefore
soil moisture exerted a stronger control on soil CO2 efflux. In 2012, due to the increased
precipitation, the system was no longer water limited and soil temperature exerted
stronger control on CO2 efflux along our chronosequence.
Our weak correlations between either soil temperature or moisture and CO2 efflux
along the chronosequence are likely due to limited environmental variation during our
sampling campaign. Data were collected only during the peak growing season (~18 hours
of sun) when the driving variables did not change much. We believe that if given the
ability to collect data during shoulder seasons and winter we would have observed higher
correlations between soil temperature and moisture with soil CO2 efflux. Also, although
we measured CO2 efflux intermittently during both day and the night, we did not use
automated sensors to collect data at regular, high frequency intervals. Because
measurements were made at irregular times over days when environmental conditions
such as precipitation were also changing, we have less explanatory power. For example,
at one collar, one measurement might have been made on a warm, dry night and then next
on a cool, dry night.
Although we did not find strong correlations between soil CO2 efflux and soil
temperature or moisture, the correlations between variables that we did observe were
consistent with the observations of other workers (Huemmrich et al., 2010; Michaelson
30
and Ping, 2003; Natali et al., 2011; Oberbauer et al., 2007; Peterson and Billings, 1975).
For example, it has been shown that increased soil temperature, increases soil CO2 efflux
with the greatest increases observed in sites in drier ecosystems. In general, soil CO2
efflux increased with increasing soil temperature, and decreased with increasing soil
moisture (Michaelson and Ping, 2003; Oberbauer et al., 2007).
2.7.1.2 Physical Soil Properties
In contrast to our weak correlations for soil moisture and temperature, we found a
strong inverse correlation between bulk density and soil CO2 efflux (R2= 0.50) (FIGURE
5b). Along the chronosequence, the tundra and stabilized slump have some of the lowest
bulk densities and highest average soil CO2 effluxes (TABLE 2&3). In contrast, the
active slump has some of the highest bulk densities and lowest soil CO2 effluxes
(FIGURE 5b). Indeed, soil bulk density values in the active slump approached typical
particle density values of 2.65 g cm-3. These findings suggest that higher bulk densities
may provide a diffusion limitation inherent in the soil profiles along the chronosequence
and possibly inhibiting gas from rising to the surface (Davidson and Trumbore, 1995).
In contrast, we found a strong positive correlation between soil organic matter and
soil CO2 efflux (R2=0.51) (FIGURE 5c). The tundra and stabilized slump had some of the
highest organic matter contents and some of the highest soil CO2 efflux. In contrast, the
active slump had the lowest soil organic matter content and the lowest soil CO2 efflux
possibly suggesting an organic substrate limitation to respiration. These results suggest
that physical soil properties, such as bulk density and soil organic matter, act as stronger
local control on CO2 efflux than environmental factors such as soil temperature and
moisture. Though we cannot differentiate whether SOM or bulk density are more
31
influential in controlling CO2 efflux because of the correlation between SOM and bulk
density (FIGURE 5a), these observations suggest fruitful avenues for future studies.
2.8 Implications
In a warming world, the prevalence of large, rapidly evolving, thermal erosional
features like the Selawik slump are predicted to increase in frequency (Gooseff et al.,
2009; Lacelle et al., 2010; Rowland et al., 2010). However, no published studies to our
knowledge have examined soil CO2 emissions, as well as the governing controls on these
fluxes, within these catastrophic thaw slumps. This study contributes to our knowledge of
how these retrogressive thaw slumps may initially function and give rise to changes in
soil physical properties that either limit or enhance soil CO2 efflux and how these
geomorphic thaw slumps may evolve over time.
Findings from this study indicates that quantifying the initial substrate quantity, used
as a proxy for carbon availability, as well as the material’s diffusivity will likely be of
great importance to understanding CO2 efflux in these rapidly changing environments.
The Arctic is composed of a heterogeneous landscape where substrate soil organic matter
ranges from 1-3% as seen at the active Selawik slump, to 90-100% as seen in yedoma
soils of northern Alaska and Siberia (Kanevskiy et al., 2011) and the stabilized slump in
our study. Initial substrate appears to play a large role in driving differences in CO2 efflux
from a given environment, with lower fluxes from the active slump compared to the
tundra and stabilized slump, suggesting a critical need to quantify soil carbon and other
soil properties in these environments.
Our research also suggests that geomorphic controls governing the initiation and
development of thaw slumps influences overall CO2 efflux. Our findings show that
32
thermokarst features such as retrogressive thaw slumps ultimately result in compaction of
the soil structure increasing soil bulk density and decreasing porosity. These soil
structural changes appear to reduce the release rate of soil CO2 due to diffusivity
limitations and merit further study.
2.9 Conclusion
We examined CO2 efflux along a thaw slump chronosequence from one of the
largest known retrogressive thaw slumps. To our knowledge, this is the first study to
publish CO2 efflux, and their possible controls in a thermokarst feature of this type and
magnitude. In contrast to our initial expectations that the active slump would have higher
soil CO2 effluxes, we found that soil CO2 effluxes in the active thaw slump were
significantly lower, approximately half that measured in the tundra and stabilized slump
for 2011 and 2012. While some studies have shown a strong correlation with soil
temperature and moisture controlling CO2 efflux, our research suggests that physical soil
properties such as bulk density and soil organic matter also exhibit a control on CO2
efflux, and that further study of these local controls would improve our understanding of
carbon-cycling within these features.
33
2.10 References
ACIA, 2004, Impacts of a Warming Climate: Arctic Climate Impact Assessment.
Amundson, R., and E. Davidson, 1990, Carbon dioxide and nitrogenous gases in the soil
atmosphere: Journal of Geochemical Exploration, v. 38, p. 13-41.
Anthony, K. W., 2009, Methane: a menace surfaces: Scientific American, v. 301, p. 6875.
Blake, G. R., and K. H. Hartge, 1986, Bulk Density1: Methods of Soil Analysis: Part 1—
Physical and Mineralogical Methods, v. sssabookseries, p. 363-375.
Chapin, F. S., P. A. Matson, and P. M. Vitousek, 2011, Principles of terrestrial ecosystem
ecology, Springer.
Davidson, E. A., and S. E. Trumbore, 1995, Gas diffusivity and production of CO2 in
deep soils of the eastern Amazon: Tellus B, v. 47, p. 550-565.
Giblin, A. E., K. J. Nadelhoffer, G. R. Shaver, J. A. Laundre, and A. J. McKerrow, 1991,
Biogeochemical Diversity Along a Riverside Toposequence in Arctic Alaska:
Ecological Monographs, v. 61, p. 415-435.
Gooseff, M. N., A. Balser, W. B. Bowden, and J. B. Jones, 2009, Effects of Hillslope
Thermokarst in Northern Alaska: Eos Trans. AGU, v. 90.
Grosse, G., J. Harden, M. Turetsky, A. D. McGuire, P. Camill, C. Tarnocai, S. Frolking,
E. A. G. Schuur, T. Jorgenson, S. Marchenko, V. Romanovsky, K. P. Wickland,
N. French, M. Waldrop, L. Bourgeau-Chavez, and R. G. Striegl, 2011,
Vulnerability of high-latitude soil organic carbon in North America to
disturbance: J. Geophys. Res., v. 116, p. G00K06.
Hinzman, L., N. Bettez, W. Bolton, F. Chapin, M. Dyurgerov, C. Fastie, B. Griffith, R.
Hollister, A. Hope, H. Huntington, A. Jensen, G. Jia, T. Jorgenson, D. Kane, D.
Klein, G. Kofinas, A. Lynch, A. Lloyd, A. McGuire, F. Nelson, W. Oechel, T.
Osterkamp, C. Racine, V. Romanovsky, R. Stone, D. Stow, M. Sturm, C.
Tweedie, G. Vourlitis, M. Walker, D. Walker, P. Webber, J. Welker, K. Winker,
and K. Yoshikawa, 2005, Evidence and Implications of Recent Climate Change in
Northern Alaska and Other Arctic Regions: Climatic Change, v. 72, p. 251-298.
Huemmrich, K. F., G. Kinoshita, J. A. Gamon, S. Houston, H. Kwon, and W. C. Oechel,
2010, Tundra Carbon Balance under Varying Temperature and Moisture
Regimes: Journal of Geophysical Research, v. 115, p. 154-175.
IPCC, 2007, Climate Change 2007: Synthesis Report. Contribution of Working Groups I,
II and III to the Fourth Assessment Report of the Intergovernmental Panel on
Climate Change, v. 446, IPCC, 104 p.
Jenny, H., 1941, Factors of soil formation: New York, McGraw-Hill Book Co., 281 p.
Jorgenson, M. T., and T. E. Osterkamp, 2005, Response of boreal ecosystems to varying
modes of permafrost degradation: Canadian Journal of Forest Research, v. 35, p.
2100-2111.
Jorgenson, M. T., V. Romanovsky, J. Harden, Y. Shur, J. O’Donnell, E. A. G. Schuur, M.
Kanevskiy, and S. Marchenko, 2010, Resilience and vulnerability of permafrost to
climate change: Canadian Journal of Forest Research, v. 40, p. 1219-1236.
Jorgenson, M. T., J. E. Roth, P. F. Miller, M. J. Macander, M. S. Duffy, A. F. Wells, G.
V. Frost, and E. R. Pullman, 2009, An ecological land survey and landcover map
34
of the Arctic Network, Natural Resource Technical Report NPS/ARCN/NRTR2009/270, Fort Collins, Colorado.
Jorgenson, M. T., Y. L. Shur, and T. E. Osterkamp, 2008, Thermokarst in Alaska, Ninth
International Conference on Permafrost, p. 1-5.
Jorgenson, M. T., Y. L. Shur, and E. R. Pullman, 2006, Abrupt increase in permafrost
degradation in Arctic Alaska: Geophysical Research Letters, v. 33, p. L02503.
Kanevskiy, M., Y. Shur, D. Fortier, M. T. Jorgenson, and E. Stephani, 2011,
Cryostratigraphy of late Pleistocene syngenetic permafrost (yedoma) in northern
Alaska, Itkillik River exposure: Quaternary Research, v. 75, p. 584-596.
Lacelle, D., J. Bjornson, and B. Lauriol, 2010, Climatic and geomorphic factors affecting
contemporary (1950–2004) activity of retrogressive thaw slumps on the Aklavik
Plateau, Richardson Mountains, NWT, Canada: Permafrost and Periglacial
Processes, v. 21, p. 1-15.
Lanan, K. M., 2013, Glacial and geomorphic history of the upper Selawik valley,
northwestern Alaska, with implications for thermokarst formation, M.S. Thesis,
Idaho State University.
Lee, H., E. Schuur, J. Vogel, and M. Lavoie, 2011, A spatially explicit analysis to
extrapolate carbon fluxes in upland tundra where permafrost is thawing: Global
Change Biology, p. 1379-1393.
Lee, H., E. A. G. Schuur, and J. G. Vogel, 2010, Soil CO2 production in upland tundra
where permafrost is thawing: Journal of Geophysical Research: Biogeosciences,
v. 115, p. G01009.
LI-COR, 2007, LI-8100 automated soil CO2 flux system and LI-8150 multiplexer
instruction manual, Lincoln, Nebraska, LI-COR Inc.
Michaelson, G. J., and C. L. Ping, 2003, Soil Organic Carbon and CO2 Respiration at
Subzero Temperature in Soils of Arctic Alaska: Journal of Geophysical Research,
v. 108, p. 8164.
Natali, S. M., E. A. G. Schuur, C. Trucco, C. E. Hicks Pries, K. G. Crummer, and A. F.
Baron Lopez, 2011, Effects of experimental warming of air, soil and permafrost
on carbon balance in Alaskan tundra: Global Change Biology, v. 17, p. 13941407.
Oberbauer, S. F., C. E. Tweedie, J. M. Welker, J. T. Fahnestock, G. H. R. Henry, P. J.
Webber, R. D. Hollister, M. D. Walker, A. Kuchy, E. Elmore, and G. Starr, 2007,
Tundra CO2 fluxes in response to experimental warming across latitudinal and
moisture gradients: Ecological Monographs, v. 77, p. 221-238.
Peterson, K. M., and W. D. Billings, 1975, Carbon Dioxide Flux from Tundra Soils and
Vegetation as Related to Temperature at Barrow, Alaska: American Midland
Naturalist, v. 94, p. 88-98.
Poole, D. K., and P. C. Miller, 1982, Carbon dioxide flux from three Arctic tundra types
in North-central Alaska, U.S.A.: Arctic and Alpine Research, v. 14, p. 6.
Raich, J. W., and W. H. Schlesinger, 1992, The global carbon dioxide flux in soil
respiration and its relationship to vegetation and climate: Tellus B, v. 44, p. 81-99.
Rowland, J. C., C. E. Jones, G. Altmann, R. Bryan, B. T. Crosby, L. D. Hinzman, D. L.
Kane, D. M. Lawrence, A. Mancino, P. Marsh, J. P. McNamara, V. E.
Romanvosky, H. Toniolo, B. J. Travis, E. Trochim, C. J. Wilson, and G. L.
35
Geernaert, 2010, Arctic Landscapes in Transition: Responses to Thawing
Permafrost: Eos Trans. AGU, v. 91.
Schlentner, R. E., and K. van Cleve, 1985, Relationships between CO2 evolution from
soil, substrate temperature, and substrate moisture in four mature forest types in
interior Alaska: Canadian Journal of Forest Research, v. 15, p. 10.
Schoeneberger, P. J., D. A. Wysocki, E. C. Benham, and W. D. Broderson, 2002, Field
book for describing and sampling soils,, in N. S. S. C. Natural Resources
Conservation Service, ed., Lincoln, NE.
Schumacher, B. A., 2002, Methods for the determination of total organic carbon (TOC)
in soils and sedimnets, in E. P. Agency, ed., U.S. Government, p. 8.
Schuur, E., and B. Abbott, 2011, High risk of permafrost thaw: Nature, v. 480, p. 32-33.
Schuur, E. A. G., J. Bockheim, J. G. Canadell, E. Euskirchen, C. B. Field, S. V.
Goryachkin, S. Hagemann, P. Kuhry, P. M. Lafleur, H. Lee, G. Mazhitova, F. E.
Nelson, A. Rinke, V. E. Romanovsky, N. Shiklomanov, C. Tarnocai, S.
Venevsky, J. G. Vogel, and S. A. Zimov, 2008, Vulnerability of Permafrost
Carbon to Climate Change: Implications for the Global Carbon Cycle:
BioScience, v. 58, p. 701-714.
Schuur, E. A. G., J. G. Vogel, K. G. Crummer, H. Lee, J. O. Sickman, and T. E.
Osterkamp, 2009, The effect of permafrost thaw on old carbon release and net
carbon exchange from tundra: Nature, v. 459, p. 556-559.
Tarnocai, C., 2009, Soil organic carbon pools in the northern circumpolar permafrost
region: Global Biogeochemical Cycles, v. 23, p. 1-11.
Vogel, J., E. Schuur, C. Trucco, and H. Lee, 2009, Response of CO2 Exchange in a
Tussock Tundra Ecosystem to Permafrost Thaw and Thermokarst
Development Journal of Geophysical Research, v. 114, p. G04018.
Vogel, J. G., B. P. Bond-Lamberty, E. A. G. Schuur, S. T. Gower, M. C. Mack, K. E. B.
O'Connell, D. W. Valentine, and R. W. Ruess, 2008, Carbon allocation in boreal
black spruce forests across regions varying in soil temperature and precipitation:
Global Change Biology, v. 14, p. 1503-1516.
Winton, M., 2006, Amplified Arctic climate change: What does surface albedo feedback
have to do with it?: Geophysical Research Letters, v. 33, p. L03701.
36
2.11 Figures
Figure 1. Map view of the Selawik slump. Each environment is symbolized as follows:
black squares denote tundra, black triangles denote active slump, and black circles denote
stabilized slump. Stippled pattern in the stabilized slump is made by Black Spruce
shadows. The boundary for each chronosequence is outlined. The Selawik River flows
east to west at the base of the image. The inlay map in the bottom right corner shows the
location of the field area in Alaska. Weather station location is denoted by a black circle
with white ring.
37
Table 1. Summary ANOVA value table.
2011 Soil Temperature
2012 Soil Temperature
2011 Soil Moisture
2012 Soil Moisture
2011 CO2 Efflux
2012 CO2 Efflux
N
Value
F
Value
P
Value
Degrees
Freedom
21
25
21
25
22
25
64.07
78.4
7.52
5.08
12.69
27.48
<0.0001
0.0001
<0.0039
0.015
<0.0001
<0.0001
2
2
2
2
3
3
Table 2. Soil and vegetation characteristics for each chronosequence environment. Active slump
has some of the highest bulk densities, lowest soil organic matter %, with no detectable soil
profile.
Vegetation
Tundra
Active Slump
Stabilized
Slump
Shrub, Sedge,
Tussock, Moss,
Lichen
None. Bare
mineral diamict,
3% - vegetated
tundra rafts
Black spruce,
alder, grasses
Soil Type
Soil Organic
Matter (%)
Bulk Density
(cm3/cm3)
July Active
Layer
Thickness
(cm)
Gellic typic
fibristel
25 - 90
0.06 - 0.8
30 - 60
No soil profile,
bare mineral
diamict
1-8
1.3 - 2.2
>100
Gellic typic
aquorthel
85 - 95
0.1 – 0.3
50 - 80
38
Figure 2. Day-of-the-year time series of environmental variables for 2011 and 2012 field
seasons (grey shading). (a) cumulative precipitation (mm), (b) air temperature (°C), (c)
soil temperature at 15cm depth (°C), (d) soil moisture (cm3/cm3), (e) soil CO2 efflux (g
CO2-C/m2/day). For figures a and b, light grey and black lines represent continuous
weather station data from 2011 and 2012, respectively. For figures c-e, daily means are
shown as black squares for tundra, black triangles for the active slump, and white circles
for the stabilized slump.
39
Figure 3. Average mean peak growing season (a) temperature (°C), (b) soil moisture
(cm3/cm3), (c) soil CO2 efflux (g CO2-C/m2/day), in both 2011 (light grey) and 2012
(dark grey) for the chronosequence environments: tundra, active slump, stabilized slump.
Bars represent the mean of all measurements and error bars represent one standard error.
Lettering above each bar represents the results for significant differences within year and
across environments using a Tukey-Kramer HSD. Columns with similar letters are not
significantly different.
40
Table 3. Average daily values of CO2 efflux (g CO2-C/m2/day) in 2011 and 2012 for each
chronosequence environment. Also included for comparison are CO2 efflux from
relatively similar environments.
2011 CO2 efflux
(g-CO2-C/m2/day)
Tundra
Active Slump
Stabilized Slump
2012 CO2 efflux
(g-CO2-C/m2/day)
Comparable CO2 efflux
(g-CO2-C/m2/day)
3.32±0.027
1.89±0.16
0.46 to 1.71,2,3
1.75±0.15
0.86±0.23
0.72 to 2.644,5
5.14±0.27
3.24±0.13
0.264 to 5.896,7,8
1
Giblin et al., 1991, 2 Poole and Miller, 1982, 3 Raich and Schlesinger, 1992, 4 Lee et al., 2010, 5 Schuur et al., 2009, 6
Raich and Schlesinger, 1992, 7 Schlentner and van Cleve, 1985, 8 Vogel et al., 2008.
41
Figure 4. Soil temperature (°C) and soil moisture (cm3/cm3) controls on CO2 efflux (g
CO2-C/m2/day) in 2011 and 2012, along the chronosequence: tundra, active slump,
stabilized slump. For all graphs, regressions of soil moisture with CO2 efflux are
represented by black lines. Regressions of soil temperature with CO2 efflux are
represented by grey lines. r2 and p values for each soil temperature regression (grey) and
soil moisture regression (black) are in the bottom left corner of each graph.
42
Figure 5. (a) Correlation between soil organic matter (%) and bulk density (g-dry/cm3),
exponential fit. (b) Negative correlation between bulk density (g-dry/cm3) and CO2 efflux
(g CO2-C/m2/day). (c) Positive correlation between soil organic matter (%) and CO2
efflux (g CO2-C/m2/day). For all graphs, chronosequence is represented as follows: black
squares are tundra, grey triangles are active slump, white circles are stabilized slump.
43
3
Soil matters: Vertical profiles of CO2 and CH4 concentrations and
δ13C of CO2 respired from a thaw slump chronosequence reveal that
soil properties can control carbon efflux at the surface.
3.1 Abstract
Arctic landscapes are experiencing unprecedented warming resulting in
permafrost thaw, which causes ground subsidence or thermokarst. On hillslopes, these
features result in erosion, altering physical soil structure and properties such as
temperature and moisture. This raises concerns because permafrost soils store nearly half
the world’s global carbon. In order to evaluate the time evolution of carbon released from
thaw slumps, we evaluated a soil-profile chronosequence at the Selawik slump in
northwest Alaska, including measurements from undisturbed tundra, the active slump and
a stabilized, revegetated slump. We use these three environments to assess how
thermokarst-driven changes affect soil column CO2 and CH4 production and the resulting
surface efflux. Our previous measurements at this site determined that CO2 efflux from
the active slump was half that from the tundra or stabilized slump. This current study
aims to resolve whether this difference can be explained by soil diffusivity or the quantity
and quality of the soil organic matter in the soil profile.
Toward these ends, we measured soil profile CO2 and CH4 concentrations, δ13C of
both soil respired CO2 and soil organic matter, as well as physical soil parameters such as
soil organic carbon %, porosity and water filled pore space at depths of 3, 5, 15, 30, and
50 cm. These data constrain how belowground carbon emissions are affected by changes
within the soil environment resulting from the initiation and subsequent stabilization of a
rapidly evolving thaw slump. Though the active slump had the lowest CO2 surface efflux,
we were surprised to find that the CO2 concentrations at depth were 5 orders of
44
magnitude higher than those observed in the other two environments. CH4 concentrations
in the active slump were ~3x higher. The low porosity, high water filled pore space, and
low calculated effective diffusivity in the active slump soil profile all indicate that these
soil properties limit carbon fluxes at the surface. Observations of soil gleying, heavier
δ13C of CO2 (~-15 ‰), and elevated methane concentrations all point to anaerobic soil
conditions in these environments. This suggests that thaw slumps switch the metabolic
pathway toward methanogenesis. Though thermal erosion features like thaw slumps are
predicted to become more prevalent on the landscape, their soil properties may ultimately
inhibit them from becoming hot-spots of carbon efflux. Our observations suggest that
rapidly evolving thaw slumps with high topographic relief profoundly alter the local
environmental conditions and biophysical controls on carbon cycling but do not appear to
produce a rapid ‘flush’ of carbon to the atmosphere.
3.2 Introduction
Permafrost soils in the arctic store approximately 1600 Pg of carbon as soil organic
matter which constitutes ~50 % of global terrestrial carbon (Schuur et al., 2008; Tarnocai,
2009; Zimov et al., 2006). These landscapes are undergoing dramatic geomorphic
transformations due to increased warming associated with climate change that have
important consequences for soil carbon cycling (ACIA, 2004; IPCC, 2007; Serreze et al.,
2000). Over the past 30 years, mean temperatures within Alaska, for example, have risen
by 1-3 °C (ACIA, 2004). By 2050, temperatures are projected to increase by 3-5 °C
during the summer, and 4-7 °C during winter (ACIA, 2004). This warming may also be
accompanied by increased precipitation in the form of rain.
45
Within Alaska, permafrost soils have experienced a 1-4 °C rise in temperature over
the past 30 years, which has caused widespread permafrost degradation (Osterkamp and
Romanovsky, 1999; Osterkamp, 2005). A direct result of permafrost degradation has
been thermokarst formation, whereby thaw leads to ground subsidence (Hinzman et al.,
2005; Jorgenson et al., 2010; Jorgenson et al., 2008b; Rowland et al., 2010). Thermokarst
features directly impact soil and hydrologic properties, physically changing soil profiles
and soil microenvironments, ecosystem energy balance, increase exposure of soil carbon
and nutrient pools due to active layer deepening, as well as changes to vegetation and
landscape topography (Jorgenson and Osterkamp, 2005; Rowland et al., 2010). In
particular, thermokarst features occurring on hillslopes can strip vegetation and expose
frozen soils to incoming solar radiation and warm air temperatures. On sloping surfaces,
thawed material is transported away by mass movements or surface overflow, leading to
continued thaw and feature growth. Rapidly evolving thermokarst currently composes
only 2-4 % of the landscape but are expected to increase in spatial extent, magnitude and
frequency with warming climate (Gooseff et al., 2009). Despite these trends, few studies
have examined how altered physical soil properties in thermal erosion features could
affect the delivery of carbon to the atmosphere.
Permafrost degradation has the potential to expose vast stores of frozen soil carbon to
microbial decomposition. This would result in the release of greenhouse gases to the
atmosphere, such as carbon dioxide (CO2) and methane (CH4) (Amundson and Davidson,
1990; Schuur and Abbott, 2011). Most CO2 efflux studies within the arctic have focused
on slowly evolving thermokarst features on relatively low relief terrain with relatively
small spatial extents. In these studies, CO2 efflux increases with thermokarst maturity
46
(Lee et al., 2010; Schuur et al., 2009; Vogel et al., 2009). % SOC, soil temperature and
moisture have been shown to be the strongest controls on efflux (Lee et al., 2011; Schuur
et al., 2009; Vogel et al., 2009). In general, increases in soil temperature result in
increases in CO2 soil respiration (Hobbie et al., 2000; Huemmrich et al., 2010;
Michaelson and Ping, 2003), while increases in soil moisture can shift CO2 production
from an aerobic process to an anaerobic process, whereby the majority of carbon is
released in the form of methane through methanogenesis (Anthony, 2009; Huemmrich et
al., 2010; Lee et al., 2012; Michaelson and Ping, 2003; Oberbauer et al., 2007).
Rapidly evolving thermokarst create both aerobic and anaerobic soil conditions in
relatively small spatial scales (>1 m), which effects carbon cycling within these systems
(Lee et al., 2012). CO2 and CH4 concentration within the soil profile are governed by
both rate of production, gas conversions, and the soil diffusivity which is a function of
porosity (amount of air-filled pore space within material) and soil moisture (Amundson
and Davidson, 1990).
In general, CO2 concentrations in soil profiles affected by permafrost thaw, are shown
to decrease with depth, while CH4 concentrations increase with depth (Michaelson et al.,
2011) Depth to water-table, landscape position, soil moisture, and temperature as well as
vegetation composition have been shown to have interactive controls on carbon
production within these systems (Lee et al., 2012; Michaelson et al., 2011; Olefeldt et al.,
2013). Furthermore, stable isotope studies suggest that the δ13C of soil respired CO2 in
arctic environments ranges from -28 to -11 ‰ (Alewell et al., 2011; Hesterberg and
Siegenthaler, 1991) with aerobic soil conditions tending to produce δ13C of soil respired
CO2 values between -23 to -27 ‰ (Lee et al., 2012) and anaerobic soil conditions
47
producing values that are heavier, ranging from -11 to -23 ‰ (Lee et al., 2012), due to
the preferential use of 12C over 13C during methanogenesis . Studies on the effects of
carbon cycling for both anaerobic and aerobic soil conditions within thermokarst systems
have mainly focused on the forms of carbon released as well as the production rate under
varying temperature and moisture conditions from laboratory soil incubations
(Huemmrich et al., 2010; Lee et al., 2012).
No studies have focused on gas characteristics in soil profiles on rapidly evolving
thermokarst features. Previous work by Jensen et al., (in review) showed that, when
compared to both the undisturbed tundra and an adjacent stabilized thaw slump, the
lowest soil CO2 efflux was from the surface of the active thaw slump. Soil moisture and
temperature were found to be weak controls on these fluxes. Rather, findings from this
previous study suggested that changes in physical soil properties such as initial substrate
quality and/or the material’s diffusivity likely play a key role in governing CO2 efflux in
these rapidly changing environments (Amundson and Davidson, 1990; Davidson and
Trumbore, 1995).
In this effort, we use soil profiles across a thaw slump chronosequence to examine
how changes in soil characteristics affect the concentration and isotopic character of CO2
and CH4 . We use these data to examine the influence of porosity, effective diffusivity,
soil temperature, moisture, as well as other physical soil properties on the production and
release of carbon across the chronosequence. We hypothesize that lower soil CO2 efflux
in the active slump compared to the tundra and stabilized slump can be explained by
changes in physical soil properties associated with the initiation and geomorphic
evolution of thermal erosion features.
48
3.3 Field Setting
The Selawik slump field area is located in the upper Selawik River basin
(157°36”43”W, 66°29’54”N) in northwest Alaska (FIGURE 1a). This field area is
located ~108 km east of Selawik Village, and is only accessible by helicopter. The field
site contains a thaw slump chronosequence composed of the undisturbed upland tussock
shrub tundra, an active retrogressive thaw slump, and a stabilized retrogressive thaw
slump. The active slump, which initiated in 2004, is the largest known retrogressive thaw
slump in Alaska. The active slump floor is approximately 50,000 m2, with a headwall
measuring 15-18 m tall, which erodes headward into a tundra upland at a rate of ~10
m/yr.
Mean annual air temperature at the Kotzebue airport weather station is -5.67 °C,
averaged over 30 years. Our field site is further inland and is expected to be warmer in
the summer and colder in the winter with a similar annual mean. The soil temperature
regime is gelic, with a mean soil temperature of -4.67 °C. Mean annual precipitation for
this area is 256 mm per year. The soil moisture regime is classified as aquic
(Schoeneberger et al., 2002). The parent material is a glacial diamict composed of poorly
sorted cobble, gravel and silt (Lanan, 2013). Factors such as climate, slope, parent
material, and biota are relatively consistent across the chronosequence (Chapin et al.,
2011; Jenny, 1941; Jenny and Amundson, 1997). Since there are no biotic barriers across
this chronosequence, differences in vegetation are a function of time since disturbance.
Due to the difficulty in accessing this area, sampling was targeted during the summer
months, when we expected to observe the highest CO2 production and efflux, as well as,
the greatest variability in environmental conditions.
49
3.4 Thaw slump chronosequence: a space for time substitution
In order to assess the temporal evolution of a thaw slump, we utilize a space for
time substitution and observe three contemporary surfaces that range in feature maturity
from undisturbed tundra to active slump, and finally stabilized slump (FIGURE 1a). In
this chronosequence, the final state of our system does not represent a return to tundra
conditions but rather in a trajectory toward a re-vegetated, forested surface due to oneway geomorphic and ecosystem drivers.
The tundra environment has not undergone any physical disturbance and
represents our arctic landscape “reference” condition or control. The tundra is ~20m
north of the Selawik slump headwall (FIGURE 1a). Typical soil profiles for this area
(FIGURE 1I) consist of an organic layer (0-18 cm), a cryoturbated A horizon (19-23 cm),
and a mottled B horizon (23-55 cm). Summer active layer depths for this area range from
30-60 cm (FIGURE 1II). The O horizon in the tundra is composed of three distinct units
all in various states of decomposition: Oi, Oe, Oa. The Oi horizon (0-4 cm) is the least
decomposed containing moss, and lichen. The Oe horizon (4-9 cm) shows evidence of
primary decomposition. This horizon is mainly composed of sedge and tussock
(Eriophorum vaginatun, and Carex spp.) root systems. The Oa horizon (9-18 cm) is the
most decomposed, which consists of peat, humus and lignin. Soil color within the O
horizon ranges from brown to light brown (10YR 3/4 dry, 10YR 3/2 moist) (FIGURE
1II) (SPP 1). The A horizon represents a transition from organic to mineral soils. Horizon
A (19-23 cm) is characterized by cryoturbation, consisting of decomposed material from
above being worked into the upper portion of the A horizon due to freezing and thawing
(FIGURE 1I). Soil colors are grey (2.5Y 6/3-7/2 dry, and 5GY 5/1 moist) (FIGURE 1II)
50
(SPP 1). Horizon B (23-55 cm) is composed of silts and clays which, show evidence of
gleying and redoximorphic soil features (FIGURE 1I). Gleying within the B horizon is
characterized by soil colors ranging from grey (2.5Y 6/2-7/2 dry, and 5GY 4/1-5/1 moist)
(FIGURE 1II) (SPP 1). Our observed soil profiles are consistent with those defined by
Jorgenson (2009) for the Selawik area.
Cross section 1-1’ in Figure 1b, shows the transition from the control tundra to the
active slump floor. The headwall is composed mostly of glacial diamict and ice, with a
small (20 cm) upper cap of organic rich peat and tussock community (the tundra
environment). As a consequence, the vast majority of the material eroding into the active
slump is glacial diamict from the headwall, which is transported downslope through
mudflows and rill erosion on the floor of the active slump. ~60 % of the slump floor is
covered with these slurries of water, fine sediment, cobbles and gravels. These materials
are sufficiently saturated to support flow. There are no detectable soil profiles within the
active slump. Instead, the profile (0-50 cm) is composed of unconsolidated pebbles,
gravels, cobbles, silts, and clays with sparse interspersed peat material (FIGURE 1III).
There was no detectable active layer in the upper meter. All active slump soil profiles
showed strong evidence of gleying, with soil colors from dark grey to grey (2.5Y 5/2-6/1
dry, and N 3/ - 5GY 4/1 moist) (FIGURE 1IV) (SPP 1), due to water saturation.
The stabilized slump is directly adjacent to and west of the active slump
(FIGURE 1a). This feature is >500 years old based on the presence of numerous
generations of black spruce logs found decaying on the surface. We use this feature to
represent the likely trajectory of our active thaw slump. Cross section 2-2’ (FIGURE 1c)
shows the landscape transition from tundra to stabilized slump. Unlike the transition from
51
tundra to active slump, this transition is denoted by a gently sloping hillside that
transitions into a slight depression downslope. Typical soil profiles within this area
consist of an O horizon (0-9 cm), an A horizon (9-13 cm), and a B horizon (13-60 cm),
which contains material similar to the active slump at the base (FIGURE 1V). The O
horizon consists of both an Oi and an Oe horizon, and is less decomposed than the O
horizon in the tundra. The Oi horizon consists of grass, moss and lichen, which transitions
into an Oe horizon that is mainly composed black spruce, and alder (Picea mariana and
Alnus spp.) root systems, with primary decomposed material (FIGURE 1V). Soil colors
range from 10YR 3/4-6 dry, and 10YR 3/1 moist (FIGURE 1VI) (SPP 1). The A horizon
(9-13 cm) is denoted by a transition from organic to mineral material within the soil
profile. Within this horizon, cryoturbation is present with material from the Oe horizon
being cycled into the upper portion of the A horizon through freeze thaw action. Soil
color ranges from light brown to brown (2.5Y 5/2-6/2 dry, and 5GY 4/1-N 4/0 moist)
(FIGURE 1VI) (SPP 1). Horizon B (13-60 cm) is composed of silts and clays which,
show evidence of gleying and redoximorphic soil features (FIGURE 1V). Gleying
throughout the B horizon is characterized by soil colors ranging from light grey to grey
(2.5Y 6/2-3 dry, and 5G 4/1-N 4/0 moist) (FIGURE 1VI) (SPP 1). Within the B horizon
there is also a layer of unconsolidated gravels and cobbles (35-50 cm) at the base. We
consider this material to be similar in composition and origin that observed on the floor
of the active slump (FIGURE 1V).
52
3.5 Methods
3.5.1
Sampling Locations
Sampling sites were established along our chronosequence for the collection of
both soil physical properties as well as soil gases. At each of these sites, soil gases from
various depths were collected first, then soil pits were dug to 60cm. Three sites were
established in both the tundra and the stabilized slump. In the active slump, 4 sites where
established, with one site, SL12-07 dominated by organic peat. This profile is anomalous
as it represents only ~4% of the area of the slump floor. Each of these sites represents
slightly different environment, vegetation composition as well as landscape position
within the stabilized slump.
3.5.2
Soil Analysis
Soil temperature and soil moisture were taken at depths of 3, 5, 15, 30, and 50 cm
at each soil pit using a LiCor-8100A analyzer (Lincoln, NE). Soil temperature was
measured using a LiCor 8100-203 soil thermistor probe attachment. Volumetric water
content (VWC, volume water/volume soil), hereafter referred to as soil moisture, was
measured using a Kelta-T ML2x probe (Lincoln, NE). Post-processing of soil
temperature and moisture was conducted using the LI-8100 Data File Viewer (v2.0). Soil
volumetric water content was adjusted for high organic content soils using the equation
provided in the (LI-COR, 2007).
Total soil cores of known volume were taken at each soil pit at depths of 3, 5, 15,
30, and 50 cm. These 50 samples were subsequently frozen and transported back to Idaho
State University and analyzed for bulk density and gravimetric water content (GWC)
following methods described by (Schumacher, 2002), and (Black, 1965), respectively. In
53
brief, bulk density was determined by drying the soil at 105°C and weighing the dry mass
and sieving the soil to retain the fine fraction <2 mm. GWC was determined by drying 20
g of soil for 24 hours at 105 °C and reweighing the remaining mass.
Porosity and water filled pore space (WFPS) are factors thought to be limiting to
diffusivity (Firestone et al., 1989). We calculated the above parameters using measured
bulk density and GWC. Porosity was calculated using the equation, % porosity = (1-(bulk
density/particle density)*100). We assumed a constant particle density of 2.65 g/cm3.
WFPS was calculated using the equation, WFPS (%) = (gravimetric water content x soil
bulk density/1-(soil bulk density/particle density))*100 (Firestone et al., 1989; Hall and
Matson, 2003).
Soil samples were also analyzed for % soil organic carbon (SOC) and δ13C of soil
organic carbon. Soil samples were prepared for analysis by drying 100 g of soil for 24
hours at 60 °C and then grinding it to a fine powder. Samples for carbon elemental
concentrations and stable isotope ratios were analyzed at the Idaho State University’s
Interdisciplinary Laboratory for Elemental and Isotopic Analysis (ILEIA) using a
Costech ECS 4010 elemental analyzer interfaced to a Thermo Delta V Advantage
continuous flow isotope ratio mass spectrometer. All stable isotopic data are reported in
standard delta notation (δ13C) relative to the Vienna PeeDee Belemnite (VPDB) reference
standard. Analytical precision, calculated from analysis of standards distributed
throughout each run, was ≤ ± 0.2‰ for carbon isotopes, and < ± 0.5% of the sample
value for %SOC.
Effective diffusivity (DEFF)(m2 s-1) was calculated for all three chronosequence
environments using the Fick’s first law transformed to calculate DEFF = -Jc x ∂c/∂x)-1
54
(Preuss et al., 2012), where Jc (µmols m-2 s-1) values are our previous measurements of
CO2 efflux across the same chronosequence (Jensen et al., in review). We also
differenced CO2 concentrations (∂c) (mol m-3) from depths of 0.03 and 0.50 (∂x)(m) to
determine effective diffusivity. We assume that there are no net additions or losses of
CO2 along this path.
3.5.3
Soil Gas Analysis
Soil gas sampling equipment “dogbones” were used to collect soil gas samples
from each site location at depths of 3, 5, 15, 30, and 50 cm. A dogbone is composed of a
1/4” stainless steel cylinder cut to lengths of 10, 10, 25, 40, or 60 cm. Inserted into the
1/4” cylinder was a 1/8” stainless steel rod, that was 3 cm longer. The dogbone was then
pushed into the ground so that the tip of the cylinder is at the designated depth. The inner
1/8” stainless steel rod is then removed and a 10 cm section of flexible tubing connects
the hollow steel to a three-way stopcock. The stopcock was then left open for 24 hours
allowing free exchange to the atmosphere, re-establishing equilibrium with the soil
environment. After 24 hours, the stopcock was closed, allowing for soil gases to fill the
dogbone over a period of ~24 hours. Accumulated soil gas samples (5-8 ml) were then
extracted from the dogbones using a syringe fitted with a stainless steel needle, and
evacuated into glass exetainers (Labco Limited, United Kingdom). These samples were
analyzed for concentration of both CO2 and CH4, as well as, δ13C of soil respired CO2, at
Los Alamos National Energy Lab (LANL), Los Alamos, NM, USA.
Delta 13C of soil respired CO2 values were measured using a GV Instruments
Isoprime continuous flow isotope ratio mass spectrometer (GV Instruments, Manchester,
UK) coupled to a TraceGas peripheral instrument. Briefly, 15 to 1500 uL of gas was sent
55
through an injection port and CO2 was cryogenically separated from other gases prior to
introduction to the mass spectrometer. Reported results are the average of 2
duplicates. δ13C of CO2 results are reported relative to VPDB, and were calibrated off
NOAA Climate Monitoring and Diagnositics Laboratory “gold” standards for CO2
concentration and isotope ratios.
CO2 concentrations were measured using a GV Instruments Isoprime continuous
flow isotope ratio mass spectrometer (GV Instruments, Manchester, UK) coupled to a
multiflow peripheral instrument. CO2 concentration was proportionate to mass
spectrometer response and calibrated using CO2/air standards of known
concentration. High concentration samples were diluted 1:12 prior to analysis. Precision
on 8 replicates of air within a sealed glove box was 3 ppm.
CH4 concentrations were measured using an Agilent 6890A GC equipped with a
packed column (Restek RT-Msieve 5A) and flame ionization detector (FID). Carrier
helium flow was 1.5 mL/min, with the GC oven at 50 °C. Reported results are the
average of two 200 uL injections. Concentrations were proportionate to FID response
and calibrated off methane/helium standards prepared in-house.
3.5.4
Statistical Analysis
Multiple linear regression and step-wise linear regression models were used to
determine environmental controls on concentration of both CO2 and CH4 as well as, δ13C
of soil respired CO2. Concentrations of both CO2 and CH4 were log transformed to
improve normality and homoscedasticity. Other parameters such as soil moisture,
temperature, porosity, and % SOC were normally distributed and thus did not require any
transformation. All statistical analyses were performed in JMP 10 (Cary, NC) software.
56
3.6 Results
3.6.1
Physical Soil Properties
Soil moisture in the three different chronosequence environments were relatively
similar (0.3 to 0.6 cm3/cm3) (FIGURE 2a,g,m); the lowest values were in the active slump
where there as little variation down the soil profile. Both the tundra and the stabilized
slump at 3 cm had soil moistures equivalent to the active slump, however when
transitioning through the organic layer, soil moisture doubles compared to the active
slump and stays relatively constant through the mineral horizon down the soil profile
(FIGURE 2a,g,m).
Soil temperatures within the active slump were 3x higher than both the tundra and
stabilized slump (15 to 20 °C) (FIGURE 2h). This trend was consistent down profile with
the active slump displaying the least variation in temperature with depth. In contrast, both
the tundra and stabilized slump decreased in temperature (~3 °C) with depth (FIGURE
2b,n) as you approach the frost table.
Soil organic carbon % (SOC) showed greatest variation in the organic layers of
both the tundra and the stabilized slump (FIGURE 2c,o) where values ranged from 5-40
% SOC. However, % SOC was dramatically reduced to ~3 %, when transitioning into the
mineral layers in the above two environments. The active slump, which had no organic
layer, had the lowest % SOC (~1-2 %) with little variation with depth (FIGURE 2i). In
general, the mineral horizons of all three chronosequence environments, showed similar
low trends in % SOC, as well as, little variation with soil profile depth.
The highest measured bulk densities were found in the active slump (0.9-2.3 gdry/cm3), with the greatest values in the upper 10 cm. Values decreased 1.5x at 15 cm
57
and remainded relatively constant in the lower portion of the profile (FIGURE 2j). In
contrast, both the tundra and stabilized slump had the lowest bulk densities (~0.2-0.4 gdry/cm3) near the surface and increased slightly with depth to values of ~0.6 g-dry/cm3
(FIGURE 2d,p).
Calculated porosity has an opposite patterns compared to bulk density in all
chronosequence environments. The active slump had the lowest porosity at surface (0.10.3, values are dimensionless), and doubled down profile to a depth of 15cm (FIGURE
2k). In contrast both the tundra and stabilized slump had the highest porosity (~0.8-0.9) at
vegetated surface and decreased slightly to a depth of 15cm (FIGURE 2e,q). The tundra,
active slump and stabilized slump showed similar porosities in their mineral horizons
(~0.8) starting at 15 cm and continuing down the soil profile (FIGURE 2e,k,q).
The organic horizon within the tundra showed the most variability in calculated
water filled pore space (WFPS) with values ranging from (1-99 %); however, after 15
cm, the variability decreased, with most values ~25 % WFPS throughout the mineral
horizon to depth (15-50 cm)( FIGURE 2f). In contrast, the stabilized slump had low
variability in the upper organic layer with the lowest calculated WFPS (~15-25 %) in the
upper 15 cm (FIGURE 2r). However, below 15 cm the mineral horizon within the
stabilized slump showed the highest variability between sites (~1-50 %). The active
slump showed no site variability and had the highest WFPS near the surface (~97 %) and
drastically decreased to 10 % from 15 to 50 cm (FIGURE 2l).
58
3.6.1.1
Calculated Effective Diffusivity
All environments had different calculated effective diffusivities (DEFF). Higher
values indicate more connectivity between soil gases and the surface. The stabilized
slump had the highest DEFF, 138 m2 day-1. In contrast, the active slump had the lowest
DEFF, and was approximately four orders of magnitude lower than the stabilized slump
(TABLE 1). The DEFF in the tundra was both two orders of magnitude less than the active
slump and two orders more than the stabilized slump (TABLE 1).
3.6.2
Soil Gas CO2 and CH4 Concentrations
We observed similar trends in CO2 and CH4 concentrations across the
chronosequence soil profiles. CO2 concentrations were 2 orders of magnitude more than
CH4 concentrations. CO2 concentrations in the active slump were the highest, showed the
most variability throughout the soil profile among sites, and showed a general trend of
increasing CO2 concentrations with depth: ~600 ppm at surface and ~10,000 ppm at
depth (FIGURE 3e). In contrast, both the tundra and stabilized slump had less severe
gradients in CO2 concentration (~600 ppm at surface to ~3000 ppm at depth, FIGURE
3a,i). Atmospheric CO2 concentrations for the study area were determined to be ~400
ppm.
CH4 concentrations were surprisingly high within the active slump, 3x higher than
either the tundra or stabilized slump, and showed the most variability with depth
(FIGURE 3f). These concentrations increased with depth: ~6 ppm at surface and ~10 to
1200 ppm at depth (FIGURE 3f). In contrast, both the tundra and stabilized slump had
similar CH4 concentrations (~6-30 ppm), and were consistent throughout the soil profile
59
(FIGURE 3b,j). Atmospheric CH4 concentrations for the study area were determined to
be ~3 ppm.
3.6.3
Stable Isotopic Composition of Soil and Soil Gases
The δ13C of soil organic matter (SOM) was similar across all soil profile
chronosequence environments (-29 to -25 ‰), with the active slump being slightly
heavier in composition. There was little variation in the values with depth across the
chronosequence, except in the stabilized slump, which showed a slight increase in δ13C of
SOM with depth.
The δ13C of soil CO2 gas was variable across the chronosequence during both
2011 and 2012. In 2011, δ13C in the upper 15 cm, for all environments, ranged from -20
to -15 ‰. The stabilized slump was the only environment to become lighter (more
negative) with similar value to the δ13C of soil organic matter at depth. In contrast, both
the tundra and active slump δ13C of soil CO2 was extremely variable throughout the soil
profile between sites. Among sites for both the tundra and active slump, we observed a
shift from becoming both lighter and heavier with depth. For both the tundra and the
active slump, the difference between δ13C of SOM and δ13C of soil CO2 gas was ~-10 ‰.
In 2012, we observed no typical δ13C of soil CO2 gas profiles. Instead all
environments were heavier than -22 ‰, with average values ~18 ‰ (FIGURE 3d,h,l).
We observed the most variability both with depth and between sites in the active slump
(FIGURE 3h). In general, the tundra and the active slump showed a slight trend of heavy
to light δ13C within the upper 30cm (FIGURE 3d,h) whereas the stabilized slump was
similar through all depths (FIGURE 3l).
60
3.6.4
Controls on Gas Concentrations and δ 13C Isotope Values
Soil temperature and moisture were strong controls on CO2 concentration within
the tundra and stabilized slump (r2=0.45, r2=0.49, respectively) (TABLE 2). CO2
concentrations within the tundra were controlled by soil temperature (p=0.04), whereas
concentrations within the stabilized slump showed an interactive effect between soil
temperature and moisture (p=0.04). In contrast, porosity (p=0.02) was the strongest
control on CO2 concentrations within the active slump, with soil moisture (p=0.03) and
temperature (p=0.05) exhibiting a weaker control (r2=0.61) (TABLE 2).
Soil temperature and the δ13C of soil organic matter (SOM) exhibited strong
controls on CH4 concentration across the soil profile chronosequence (table 2). Within
the tundra environment, δ13C of SOM was the only control (r2=0.49) (p=0.03). In
contrast, within the active slump, the control switched to soil temperature (r2=0.42)
(p=0.003). Within the stabilized slump both soil temperature (p=0.03) and δ13C of SOM
(p=0.008) had a strong control on CH4 concentrations within the soil profile (r2=0.50)
(TABLE 2).
Both CO2 and CH4 concentrations, as well as, the environmental variables of soil
moisture and temperature exerted control on δ13C of soil CO2 gas. The only variable that
exerted control in the tundra environment (r2=0.51) was CO2 concentration (p=0.006).
Within the stabilized slump, the strongest control was also CO2 concentration (r2=0.87,
p<0.0001), although soil temperature (p=0.03) was also controlling δ13C of CO2 gas
(TABLE 2). In contrast, CH4 concentration (p=0.002) and soil moisture (p=0.004)
(TABLE 2) were the main controls on the δ13C values of soil CO2 gas in the active slump
(r2=0.72) (TABLE 2).
61
3.7 Discussion
3.7.1
Diffusion Limitation
Findings from our study showed the highest CO2 and CH4 concentrations occur at
depth in the active slump. This was seemingly in contrast with our previous work where
we measured the lowest soil CO2 efflux from the active slump. Following analysis of soil
properties, we found that low porosity, high water filled pore space (WFPS), and low
calculated effective diffusivity all suggest that our low efflux measurements were a
consequence of gas trapping rather than low production rates. Though we did not make
measurements of methane flux, low diffusivity would limit the rate of these fluxes as
well.
In contrast to the relatively depth-invariant porosity in the tundra and stabilized
slump, porosities in the upper 20 cm of active slump profiles decreased from ~0.8 to 0.4
(FIGURE 2), which we suggest limits gas diffusion and elevates both CO2 and CH4
concentrations at depth (FIGURE 3). Multiple linear regression also confirmed porosity
as the strongest control on CO2 concentrations within the active slump. The active slump
floor is composed of sediment remobilized from the thawing headwall, which is
dominated by glacial silts and gravels. The thawed materials flow away from the
headwall as a thick, viscous slurry and stabilize upon reaching lower slopes. Our
findings suggest that the fine, silty sediment that rises to the top of these flows acts to
reduce air-filled pore space within the upper 20 cm, decreasing porosity and “capping”
the slump floor resulting in a build-up of CO2 and CH4 gases at depth.
Gas concentration profiles in thawing or eroding permafrost are limited, but have
been shown to both increase and decrease with depth depending on location (Michaelson
et al., 2011). While there is no published work relating greenhouse gas concentrations
62
within active thaw slumps to physical soil properties, work conducted in temperate and
tropical environments have shown increased CO2 concentrations at depth corresponding
to either higher production rates, which correspond to organic horizons within soil
profiles, or decreases in a material’s effective diffusivity (Davidson and Trumbore, 1995;
Fierer et al., 2005; Scanlon et al., 2000). The active slump contains no organic horizon,
evident by extremely low soil organic carbon (FIGURE 2), suggesting that rather than
observing increased gas production rates; we are observing the effects of lower
diffusivity in the active slump.
Drastic increases (~95 %) in water filled pore space (WFPS) were also observed
in the upper 20 cm of the active slump (FIGURE 2). Soils are typically considered
aeration-limited when soil WFPS starts to exceed 60 % (Linn and Doran, 1984). These
high values (FIGURE 2) suggest anaerobic conditions are present in the slump. This
conclusion reinforced by the fact that most of the area of the slump floor (~65 %) is so
saturated that it impassable (quicksand-like conditions). The combination of reduced pore
volume within the top 20 cm of the active slump and the shift from air filled to water
filled pore space increases gas entrapment due to slower diffusions rates of CO2 in water
as opposed to air. Other studies have shown that increases in water filled pore volume
slow gas diffusivity in a material, as well as, increasing gas trapping at depth (Davidson
and Trumbore, 1995; Millington and Shearer, 1971; Nazaroff, 1992; Risk et al., 2008).
Finally, calculated effective diffusivity across the chronosequence supported the
patterns of lower CO2 efflux from the active slump compared to the other two
environments. We found that the lowest CO2 efflux occurred within the active slump for
both 2011 and 2012, corresponding to the lowest calculated effective diffusivity (TABLE
63
1). In contrast, the stabilized slump’s effective diffusivity was five orders of magnitude
greater than the active slump, corresponding to the highest mean CO2 efflux in both 2011
and 2012. Although the stabilized slump has the same material as the active slump at
depth, this environment has been completely re-vegetated with black spruce and a more
diffusive soil profile has been re-established.
Research relating soil moisture to CO2 and CH4 production within retrogressive
thaw slumps is limited, but collectively, our findings suggest that the retrogressive thaw
slump dramatically alters the soil physical environment and strongly influences the
production and fate of the CO2 and CH4 through the interactive controls of diffusivity and
moisture. Consistent with our findings, the few other studies that have evaluated
increasing soil moisture conditions on CO2 indicate decreases in CO2 efflux and shifts
carbon production to methanogenesis (Alewell et al., 2011; Lee et al., 2012; Michaelson
et al., 2011; Oberbauer et al., 2007; Walter et al., 2008). We expand on this idea below.
3.7.2
Micro-Soil Environments
In contrast to our initial expectation that soil profiles within the active slump would
have more aerobic activity due to increased availability carbon after thaw and slumping,
we found that the majority of soil profiles across the chronosequence to be anaerobic in
nature as indicated by (1) the abundance of methane, (2) gleying and redoximorphic
features, and (3) significant fractionation of carbon, leading to lighter δ13C values in CO2
gas. Collectively, we interpret the results to suggest that conditions in the active thaw
slumps may shift carbon cycling, at least in the short term, towards anaerobic metabolic
pathways such as methanogenesis owing to the drastically altered soil environment
conditions.
64
Methane concentrations were 3-5 orders of magnitude higher in the active slump and
increased with depth, when compared to the tundra or stabilized slump. Indeed, methane
concentrations within the active slump were consistent (mean 97 ppm) with those
reported from Alaskan lakes undergoing permafrost degradation, and other waterlogged
tundra environments (5-1000 ppm) (Michaelson et al., 2011; Thebrath et al., 1993; Torn
and Chapin, 1993; von Fischer et al., 2010; Walter et al., 2008). Saturation existed within
the active slump in 2012, as indicated by the ~95 % water filled pore space in the upper
part of the soil column in the more well drained areas and the fact that 60 % of the slump
was too saturated to walk on.
Anaerobic soil conditions, as evidenced by gleying and redoximorphic features, were
also observed throughout the soil profile in the active slump and in the lower portions of
the soil profile for both the tundra and stabilized slump. This assertion is supported by
dark grey to grey gleyed soil colors (5GY 4/1-5/2, N 4/0-3/0). Gleying was observed in
all the soil profiles across the chronosequence. Gleying appeared to be the strongest in
the active slump, where the entire profile was dominated by dark grey soil colors of 2.5Y
6/1 dry, and 5GY 4/1 – N 4/0 moist (FIGURE SP1). In contrast the mineral horizons in
the tundra and stabilized slump where extremely mottled with patchy dark grey colors
2.5Y 6/2 dry, and 5GY 4/1-3 moist (FIGURE SP1)driven by fluctuations in water-table
position (Simonson and Boersma, 1972). Within the arctic, it has been shown that the
gleying and redoximorphic features we observed across our chronosequence are primarily
due to anaerobic soil conditions (Jorgenson et al., 2009; Melke and Chodorowski, 2006;
Sjögersten et al., 2006).
65
Finally, the δ13C values of soil CO2 was, across the sites, on average -15 ‰, which is
2-3x heavier than our theoretically predicted value of ~-22.6 ‰ (based on -4.4‰
fractionation) if CO2 was derived from SOC (Amundson et al., 1998; Cerling et al.,
1991), suggesting that metabolic pathways were shifting from aerobic to anaerobic ones.
Our expectations were that the δ13C of soil respired CO2 would increase toward the
surface, starting with values close to the δ13C of soil organic matter (-27 ‰) and
equilibrating with the atmosphere (-8 ‰) (Amundson et al., 1998; Cerling et al., 1991).
However, this pattern was only observed in the warmer and dryer 2011 and only in the
stabilized slump (FIGURE 3). Rather than observing a fractionation effect of the SOM
pool, our values of δ13C of CO2 appear to reflect fractionation associated with anaerobic
soil conditions and the processes of methanogenesis over a shorter time period thereby
resulting in heavier δ13C-CO2 values. Arctic soil lab incubations under both aerobic and
anaerobic soil conditions, as well as, field work on littoral lakes in Alaska, support this
conclusion showing that heavy δ13C values of CO2 (-15 ‰) are formed under anaerobic
soil conditions and can be influenced by methanogenesis (Bates et al., 2011; Lee et al.,
2012; Thebrath et al., 1993). Similar patterns of δ13C of CO2 becoming heavier with
depth as seen in our soil profiles (FIGURE 3) have been observed across the arctic in
waterlogged soils, and these patterns have also been correlated to increases in methane
production within the soil profile (Alewell et al., 2011; Krull and Retallack, 2000;
Zaccone et al., 2008).
3.8 Implications
No studies to our knowledge have examined changes in physical soil characteristics
and micro-environment shifts that occur as a result of large, rapidly evolving, thermal-
66
erosional feature development. Although thermokarst features are predicted to increase in
frequency, (Gooseff et al., 2009; Lacelle et al., 2010; Rowland et al., 2010) our results
from a low % SOC substrate slump indicate that changes to the physical soil structure
may prevent features like ours from functioning as hotspots of CO2 efflux to the
atmosphere during their active phase as previously thought.
Our study determined that quantifying a material’s diffusivity limitations are key to
controlling the efflux of CO2 and possibly CH4 to the atmosphere from retrogressive thaw
slumps. Although CO2 efflux from active features is below the initial tundra efflux, with
subsequent feature stabilization and rebuilding of the soil profile, our research indicates
that these features will become greater source of CO2 efflux to the environment.
Geomorphic controls governing the initiation and development of thaw slumps
influence overall CO2 efflux. Our findings show that thermokarst features on hillslopes,
such as thaw slumps, may be expected to have well-drained aerobic soil conditions but
material properties may dictate largely anaerobic conditions. This ultimately affects
carbon cycling by shifting decomposition pathways to methanogensis, resulting in
possible greater CH4 emissions than previously considered. These shifts in soil conditions
could lead to underestimation of these features carbon fluxes and warrant further study.
Thermokarst CO2 efflux studies primarily rely on surface efflux measurements, which
may not represent production occurring at depth. Quantifying physical soil profile
properties, as well as, soil micro-environment shifts over small spatial scales, will
become important predictors of the effects of rapidly evolving thermo-erosional features
on the global carbon cycle. Until we focus on not only the surface but also soil profile
67
changes, our ability to accurately spatially upscale these features effects on the global
carbon cycle will be limited.
3.9 Conclusion
We use a soil profile thaw slump chronosequence to examine how changes in the soil
environment due to rapidly evolving thermokarst features, affect soil column CO2
production as well as the resulting soil CO2 efflux. We determined that carbon fluxes at
the surface of the active slump are limited by the low permeability of the consolidated
slump materials. We also determined that the majority of our chronosequence exists
under anaerobic soil conditions. Though remobilized carbon is made available for
decomposition during the active phase, lower soil diffusivity inhibits gas efflux to the
surface.
68
3.10 Reference
ACIA, 2004, Impacts of a Warming Climate: Arctic Climate Impact Assessment.
Alewell, C., R. Giesler, J. Klaminder, J. Leifeld, and M. Rollog, 2011, Stable carbon
isotopes as indicators for micro-geomorphic changes in palsa peats:
Biogeosciences Discussions, v. 8, p. 527-548.
Amundson, R., and E. Davidson, 1990, Carbon dioxide and nitrogenous gases in the soil
atmosphere: Journal of Geochemical Exploration, v. 38, p. 13-41.
Amundson, R., L. Stern, T. Baisden, and Y. Wang, 1998, The isotopic composition of
soil and soil-respired CO< sub> 2</sub>: Geoderma, v. 82, p. 83-114.
Anthony, K. W., 2009, Methane: a menace surfaces: Scientific American, v. 301, p. 6875.
Bates, B. L., J. C. McIntosh, K. A. Lohse, and P. D. Brooks, 2011, Influence of
groundwater flowpaths, residence times and nutrients on the extent of microbial
methanogenesis in coal beds: Powder River Basin, USA: Chemical Geology, v.
284, p. 45-61.
Bird, M., H. Santruckova, J. Lloyd, and E. Lawson, 2002, The isotopic composition of
soil organic carbon on a north-south transect in western Canada: European
Journal of Soil Science, v. 53, p. 393-403.
Black, C. A., 1965, Methods of Soil Analysis: Part I Physical and mineralogical
properties: Madison, Wisconsin, USA, American Society of Agronomy.
Blake, G. R., and K. H. Hartge, 1986, Bulk Density1: Methods of Soil Analysis: Part 1—
Physical and Mineralogical Methods, v. sssabookseries, p. 363-375.
Cerling, T. E., D. K. Solomon, J. Quade, and J. R. Bowman, 1991, On the isotopic
composition of carbon in soil carbon dioxide: Geochimica et Cosmochimica Acta,
v. 55, p. 3403-3405.
Chapin, F. S., P. A. Matson, and P. M. Vitousek, 2011, Principles of terrestrial ecosystem
ecology, Springer.
Davidson, E. A., and S. E. Trumbore, 1995, Gas diffusivity and production of CO2 in
deep soils of the eastern Amazon: Tellus B, v. 47, p. 550-565.
Fierer, N., O. A. Chadwick, and S. E. Trumbore, 2005, Production of CO2 in soil profiles
of a California annual grassland: Ecosystems, v. 8, p. 412-429.
Firestone, M., E. Davidson, M. Andreae, and D. Schimel, 1989, Microbiological basis of
NO and N2O production and consumption in soil: Exchange of trace gases
between terrestrial ecosystems and the atmosphere., p. 7-21.
Flanagan, L. B., J. R. Ehleringer, and D. E. Pataki, 2005, Stable Isotopes and BiosphereAtmosphere Interactions: Processes and Biological Controls, Elsevier, 310 p.
Giblin, A. E., K. J. Nadelhoffer, G. R. Shaver, J. A. Laundre, and A. J. McKerrow, 1991,
Biogeochemical Diversity Along a Riverside Toposequence in Arctic Alaska:
Ecological Monographs, v. 61, p. 415-435.
Gooseff, M. N., A. Balser, W. B. Bowden, and J. B. Jones, 2009, Effects of Hillslope
Thermokarst in Northern Alaska: Eos Trans. AGU, v. 90.
Grosse, G., J. Harden, M. Turetsky, A. D. McGuire, P. Camill, C. Tarnocai, S. Frolking,
E. A. G. Schuur, T. Jorgenson, S. Marchenko, V. Romanovsky, K. P. Wickland,
N. French, M. Waldrop, L. Bourgeau-Chavez, and R. G. Striegl, 2011,
69
Vulnerability of high-latitude soil organic carbon in North America to
disturbance: J. Geophys. Res., v. 116, p. G00K06.
Hall, S. J., and P. A. Matson, 2003, Nutrient status of tropical rain forests influences soil
N dynamics after N additions: Ecological Monographs, v. 73, p. 107-129.
Hardie, S. M. L., M. H. Garnett, A. E. Fallick, A. P. Rowland, N. J. Ostle, and T. H.
Flowers, 2011, Abiotic drivers and their interactive effect on the flux and carbon
isotope (14C and δ13C) composition of peat-respired CO2: Soil Biology and
Biochemistry, v. 43, p. 2432-2440.
Hesterberg, R., and U. Siegenthaler, 1991, Production and stable isotopic composition of
CO2 in a soil near Bern, Switzerland: Tellus B, v. 43, p. 197-205.
Hinzman, L., N. Bettez, W. Bolton, F. Chapin, M. Dyurgerov, C. Fastie, B. Griffith, R.
Hollister, A. Hope, H. Huntington, A. Jensen, G. Jia, T. Jorgenson, D. Kane, D.
Klein, G. Kofinas, A. Lynch, A. Lloyd, A. McGuire, F. Nelson, W. Oechel, T.
Osterkamp, C. Racine, V. Romanovsky, R. Stone, D. Stow, M. Sturm, C.
Tweedie, G. Vourlitis, M. Walker, D. Walker, P. Webber, J. Welker, K. Winker,
and K. Yoshikawa, 2005, Evidence and Implications of Recent Climate Change in
Northern Alaska and Other Arctic Regions: Climatic Change, v. 72, p. 251-298.
Hobbie, S. E., J. P. Schimel, S. E. Trumbore, and J. R. Randerson, 2000, Controls over
carbon storage and turnover in high-latitude soils: Global Change Biology, v. 6, p.
196-210.
Huemmrich, K. F., G. Kinoshita, J. A. Gamon, S. Houston, H. Kwon, and W. C. Oechel,
2010, Tundra Carbon Balance under Varying Temperature and Moisture
Regimes: Journal of Geophysical Research, v. 115, p. 154-175.
IPCC, 2007, Climate Change 2007: Synthesis Report. Contribution of Working Groups I,
II and III to the Fourth Assessment Report of the Intergovernmental Panel on
Climate Change, v. 446, IPCC, 104 p.
Jenny, H., 1941, Factors of soil formation: New York, McGraw-Hill Book Co., 281 p.
Jenny, H., and R. Amundson, 1997, On a State Factor Model of Ecosystems: BioScience,
v. 47, p. 536-543.
Jorgenson, M. T., and T. E. Osterkamp, 2005, Response of boreal ecosystems to varying
modes of permafrost degradation: Canadian Journal of Forest Research, v. 35, p.
2100-2111.
Jorgenson, M. T., V. Romanovsky, J. Harden, Y. Shur, J. O’Donnell, E. A. G. Schuur, M.
Kanevskiy, and S. Marchenko, 2010, Resilience and vulnerability of permafrost to
climate change: Canadian Journal of Forest Research, v. 40, p. 1219-1236.
Jorgenson, M. T., J. E. Roth, P. F. Miller, M. J. Macander, M. S. Duffy, A. F. Wells, G.
V. Frost, and E. R. Pullman, 2009, An ecological land survey and landcover map
of the Arctic Network, Natural Resource Technical Report NPS/ARCN/NRTR2009/270, Fort Collins, Colorado.
Jorgenson, M. T., Y. L. Shur, and T. E. Osterkamp, 2008a, Thermokarst in Alaska, Ninth
International Conference on Permafrost.
Jorgenson, M. T., Y. L. Shur, and T. E. Osterkamp, 2008b, Thermokarst in Alaska, Ninth
International Conference on Permafrost, p. 1-5.
Jorgenson, M. T., Y. L. Shur, and E. R. Pullman, 2006, Abrupt increase in permafrost
degradation in Arctic Alaska: Geophysical Research Letters, v. 33, p. L02503.
70
Kanevskiy, M., Y. Shur, D. Fortier, M. T. Jorgenson, and E. Stephani, 2011,
Cryostratigraphy of late Pleistocene syngenetic permafrost (yedoma) in northern
Alaska, Itkillik River exposure: Quaternary Research, v. 75, p. 584-596.
Krull, E. S., and G. J. Retallack, 2000, δ13C depth profiles from paleosols across the
Permian-Triassic boundary: Evidence for methane release: Geological Society of
America Bulletin, v. 112, p. 1459-1472.
Lacelle, D., J. Bjornson, and B. Lauriol, 2010, Climatic and geomorphic factors affecting
contemporary (1950–2004) activity of retrogressive thaw slumps on the Aklavik
Plateau, Richardson Mountains, NWT, Canada: Permafrost and Periglacial
Processes, v. 21, p. 1-15.
Lanan, K. M., 2013, Glacial and geomorphic history of the upper Selawik valley,
northwestern Alaska, with implications for thermokarst formation, M.S. Thesis,
Idaho State University.
Lee, H., E. Schuur, J. Vogel, and M. Lavoie, 2011, A spatially explicit analysis to
extrapolate carbon fluxes in upland tundra where permafrost is thawing: Global
Change Biology, p. 1379-1393.
Lee, H., E. A. G. Schuur, K. S. Inglett, M. Lavoie, and J. P. Chanton, 2012, The rate of
permafrost carbon release under aerobic and anaerobic conditions and its potential
effects on climate: Global Change Biology, v. 18, p. 515-527.
Lee, H., E. A. G. Schuur, and J. G. Vogel, 2010, Soil CO2 production in upland tundra
where permafrost is thawing: Journal of Geophysical Research: Biogeosciences,
v. 115, p. G01009.
LI-COR, 2007, LI-8100 automated soil CO2 flux system and LI-8150 multiplexer
instruction manual, Lincoln, Nebraska, LI-COR Inc.
Linn, D. M., and J. W. Doran, 1984, Effect of Water-Filled Pore Space on Carbon
Dioxide and Nitrous Oxide Production in Tilled and Nontilled Soils1: Soil Sci.
Soc. Am. J., v. 48, p. 1267-1272.
Melke, J., and J. Chodorowski, 2006, Formation of arctic soils in Chamberlindalen,
Bellsund, Spitsbergen: Polish Polar Research, v. 27, p. 119-132.
Michaelson, G. J., and C. L. Ping, 2003, Soil Organic Carbon and CO2 Respiration at
Subzero Temperature in Soils of Arctic Alaska: Journal of Geophysical Research,
v. 108, p. 8164.
Michaelson, G. J., C. L. Ping, and M. T. Jorgenson, 2011, Methane and Carbon Dioxide
content in Eroding Permafrost Soils along the Beaufort Sea Coast, Alaska: Journal
of Geophysical Research, v. 116, p. G01022.
Millington, R., and R. Shearer, 1971, Diffusion in aggregated porous media: Soil
Science, v. 111, p. 372-378.
Natali, S. M., E. A. G. Schuur, C. Trucco, C. E. Hicks Pries, K. G. Crummer, and A. F.
Baron Lopez, 2011, Effects of experimental warming of air, soil and permafrost
on carbon balance in Alaskan tundra: Global Change Biology, v. 17, p. 13941407.
Nazaroff, W. W., 1992, Radon transport from soil to air: Reviews of Geophysics, v. 30,
p. 137-160.
Oberbauer, S. F., C. E. Tweedie, J. M. Welker, J. T. Fahnestock, G. H. R. Henry, P. J.
Webber, R. D. Hollister, M. D. Walker, A. Kuchy, E. Elmore, and G. Starr, 2007,
71
Tundra CO2 fluxes in response to experimental warming across latitudinal and
moisture gradients: Ecological Monographs, v. 77, p. 221-238.
Olefeldt, D., M. R. Turetsky, P. M. Crill, and A. D. McGuire, 2013, Environmental and
physical controls on northern terrestrial methane emissions across permafrost
zones: Global Change Biology, v. 19, p. 589-603.
Osterkamp, T., and V. Romanovsky, 1999, Evidence for warming and thawing of
discontinuous permafrost in Alaska: Permafrost and Periglacial Processes, v. 10,
p. 17-37.
Osterkamp, T. E., 2005, The recent warming of permafrost in Alaska: Global and
Planetary Change, v. 49, p. 187-202.
Peterson, K. M., and W. D. Billings, 1975, Carbon Dioxide Flux from Tundra Soils and
Vegetation as Related to Temperature at Barrow, Alaska: American Midland
Naturalist, v. 94, p. 88-98.
Poole, D. K., and P. C. Miller, 1982, Carbon dioxide flux from three Arctic tundra types
in North-central Alaska, U.S.A.: Arctic and Alpine Research, v. 14, p. 6.
Preuss, I., C. Knoblauch, J. Gebert, and E. Pfeiffer, 2012, Improved quantification of
microbial CH4 oxidation efficiency in Arctic wetland soils using carbon isotope
fractionation: Biogeosciences Discussions, v. 9, p. 16999.
Raich, J. W., and W. H. Schlesinger, 1992, The global carbon dioxide flux in soil
respiration and its relationship to vegetation and climate: Tellus B, v. 44, p. 81-99.
Risk, D., L. Kellman, and H. Beltrami, 2008, A new method for in situ soil gas diffusivity
measurement and applications in the monitoring of subsurface CO2 production:
Journal of Geophysical Research: Biogeosciences (2005–2012), v. 113.
Rowland, J. C., C. E. Jones, G. Altmann, R. Bryan, B. T. Crosby, L. D. Hinzman, D. L.
Kane, D. M. Lawrence, A. Mancino, P. Marsh, J. P. McNamara, V. E.
Romanvosky, H. Toniolo, B. J. Travis, E. Trochim, C. J. Wilson, and G. L.
Geernaert, 2010, Arctic Landscapes in Transition: Responses to Thawing
Permafrost: Eos Trans. AGU, v. 91.
Scanlon, B. R., J. P. Nicot, and J. W. Massmann, 2000, Soil gas movement in unsaturated
systems: Handbook of soil science, p. A277-A319.
Schlentner, R. E., and K. van Cleve, 1985, Relationships between CO2 evolution from
soil, substrate temperature, and substrate moisture in four mature forest types in
interior Alaska: Canadian Journal of Forest Research, v. 15, p. 10.
Schoeneberger, P. J., D. A. Wysocki, E. C. Benham, and W. D. Broderson, 2002, Field
book for describing and sampling soils,, in N. S. S. C. Natural Resources
Conservation Service, ed., Lincoln, NE.
Schumacher, B. A., 2002, Methods for the determination of total organic carbon (TOC)
in soils and sedimnets, in E. P. Agency, ed., U.S. Government, p. 8.
Schuur, E., and B. Abbott, 2011, High risk of permafrost thaw: Nature, v. 480, p. 32-33.
Schuur, E. A. G., J. Bockheim, J. G. Canadell, E. Euskirchen, C. B. Field, S. V.
Goryachkin, S. Hagemann, P. Kuhry, P. M. Lafleur, H. Lee, G. Mazhitova, F. E.
Nelson, A. Rinke, V. E. Romanovsky, N. Shiklomanov, C. Tarnocai, S.
Venevsky, J. G. Vogel, and S. A. Zimov, 2008, Vulnerability of Permafrost
Carbon to Climate Change: Implications for the Global Carbon Cycle:
BioScience, v. 58, p. 701-714.
72
Schuur, E. A. G., J. G. Vogel, K. G. Crummer, H. Lee, J. O. Sickman, and T. E.
Osterkamp, 2009, The effect of permafrost thaw on old carbon release and net
carbon exchange from tundra: Nature, v. 459, p. 556-559.
Serreze, M. C., J. E. Walsh, F. S. Chapin, III, T. Osterkamp, M. Dyurgerov, V.
Romanovsky, W. C. Oechel, J. Morison, T. Zhang, and R. G. Barry, 2000,
Observational Evidence of Recent Change in the Northern High-Latitude
Environment: Climatic Change, v. 46, p. 159-207.
Shanhun, F., P. Almond, T. Clough, and C. Smith, 2012, Abiotic processes dominate CO2
fluxes in Antarctic soils: Soil Biology & Biochemistry, v. 53, p. 99-111.
Simonson, G. H., and L. Boersma, 1972, Soil Morphology and Water Table
Relations: II. Correlation between Annual Water Table Fluctuations and Profile
Features.: Technical paper 3273. Oregon Agriculture Experimental Station.
Sjögersten, S., R. van der Wal, and S. J. Woodin, 2006, Small-scale hydrological
variation determines landscape CO2 fluxes in the high Arctic: Biogeochemistry,
v. 80, p. 205-216.
Stieglitz, M., A. Giblin, J. Hobbie, M. Williams, and G. Kling, 2000, Simulating the
effects of climate change and climate variability on carbon dynamics in Arctic
tundra: Global Biogeochem. Cycles, v. 14, p. 1123-1136.
Tarnocai, C., 2009, Soil organic carbon pools in the northern circumpolar permafrost
region: Global Biogeochemical Cycles, v. 23, p. 1-11.
Thebrath, B., F. Rothfuss, M. Whiticar, and R. Conrad, 1993, Methane production in
littoral sediment of Lake Constance: FEMS Microbiology Letters, v. 102, p. 279289.
Torn, M. S., and F. S. Chapin, 1993, Environmental and biotic controls over methane flux
from arctic tundra: Chemosphere, v. 26, p. 357-368.
Vogel, J., E. Schuur, C. Trucco, and H. Lee, 2009, Response of CO2 Exchange in a
Tussock Tundra Ecosystem to Permafrost Thaw and Thermokarst
Development Journal of Geophysical Research, v. 114, p. G04018.
Vogel, J., E. A. G. Schuur, C. Trucco, and H. Lee, 2008a, Response of CO2 exchange in a
tussock tundra ecosystem to permafrost thaw and thermokarst development:
Journal of Geophysical Research, v. 114, p. G04018.
Vogel, J. G., B. P. Bond-Lamberty, E. A. G. Schuur, S. T. Gower, M. C. Mack, K. E. B.
O'Connell, D. W. Valentine, and R. W. Ruess, 2008b, Carbon allocation in boreal
black spruce forests across regions varying in soil temperature and precipitation:
Global Change Biology, v. 14, p. 1503-1516.
von Fischer, J. C., R. C. Rhew, G. M. Ames, B. K. Fosdick, and P. E. von Fischer, 2010,
Vegetation height and other controls of spatial variability in methane emissions
from the Arctic coastal tundra at Barrow, Alaska: Journal of Geophysical
Research: Biogeosciences (2005–2012), v. 115.
Walter, K., J. Chanton, F. Chapin, E. Schuur, and S. Zimov, 2008, Methane production
and bubble emissions from arctic lakes: Isotopic implications for source pathways
and ages: Journal of Geophysical Research: Biogeosciences (2005–2012), v. 113.
Winton, M., 2006, Amplified Arctic climate change: What does surface albedo feedback
have to do with it?: Geophysical Research Letters, v. 33, p. L03701.
73
Zaccone, C., D. Said-Pullicino, G. Gigliotti, and T. Miano, 2008, Diagenetic trends in the
phenolic constituents of< i> Sphagnum</i>-dominated peat and its corresponding
humic acid fraction: Organic Geochemistry, v. 39, p. 830-838.
Zhuang, Q., J. Melillo, M. Sarofim, and D. Kicklighter, 2006, CO2 and CH4 exchanges
between land ecosystems and the atmosphere in northern high latitudes over the
21st century: Geophysical Research Letters, v. 33.
Zimov, S. A., E. A. Schuur, and F. S. Chapin III, 2006, Permafrost and the global carbon
budget: Science(Washington), v. 312, p. 1612-1613.
74
3.11 Figures
Figure 1. Map view of the Selawik slump (a). Each environment is symbolized as
follows: black squares denote tundra, black triangles denote active slump, and black
circles denote stabilized slump. The boundary for each chronosequence is outlined. The
Selawik River flows east to west at the base of the image. The inlay map shows the
location of the field area in Alaska. The line 1 to 1’ on map corresponds to the cross
section (b), which represents the topographic transition from tundra to active slump. The
line 2 to 2’ on map corresponds to cross section (c), which represents the topographic
transition from tundra to stabilized slump. In b and c, boxes labeled A, B, C represent the
rough positions of representative soil profiles shown in I, III, and V respectively. These
schematic soil profiles from the tundra, active slump, and stabilized slump are paired
with photographs (II, IV, and VI) of soil profiles from the Selawik slump.
75
Figure 2. Soil depth profiles of physical soil properties across the chronosequence: tundra
(a-f), active slump (g-l), and stabilized slump (m-r). Representative soil profiles for each
chronosequence are shown on the far left and correspond to representative profiles in
figure 1-I, III, V. The horizontal grey line in the tundra and stabilized slump plots, marks
the approximate transition from organic to mineral horizons.
76
Table 1. Calculated effective diffusivity and average daily CO2 efflux (g CO2-C/m2/day)
in 2011 and 2012 for each chronosequence environment: tundra, active slump, stabilized
slump. Active slump has the lowest effective diffusivity corresponding to the lowest CO2
efflux for both 2011 and 2012.
Environment
Effective Diffusivity
(m2/day)
Tundra
Active Slump
Stabilized Slump
2011 CO2 efflux
(g-CO2-C/m2/day)
5.24
0.012
3.32±0.0271
1.75±0.151
2012 CO2 efflux
(g-CO2-C/m2/day)
138
5.14±0.271
3.24±0.131
1.89±0.161
0.86±0.231
1
Jensen et al., (in review)
Table 2. Controls on CO2 and CH4 concentrations and δ13C of soil CO2 gas. Porosity
exhibited the strongest control on CO2 concentrations in the active slump.
CO2 Concentration
(ppm)
CH4 Concentration
(ppm)
δ 13C of soil CO2 gas
(‰)
Tundra
R2 = 0.45
Soil Temperature (p=0.039)
R2 = 0.49
δ 13C of SOM (p=0.029)
R2 = 0.51
CO2 Concentration (p=0.0056)
Active Slump
R2 = 0.61
Porosity (p=0.017)
Soil Moisture (p=0.025)
Soil Temperature x
Porosity (p=0.048)
R2 = 0.49
Soil Temperature x
Soil Moisture (p=0.041)
R2 = 0.42
Soil Temperature (p=0.0028)
R2 = 0.72
CH4 Concentration (p=0.0015)
Soil Moisture (p=0.0037)
R2 = 0.50
Soil Temperature (p=0.026)
δ 13C of SOM (p=0.0076)
R2 = 0.87
CO2 concentration (p<0.0001)
Soil Temperature (p=0.029)
Stabilized Slump
77
Figure 3. Soil depth profiles of CO2 and CH4 concentrations, as well as, δ13C of soil
organic matter (SOM) and δ13C-CO2 across the chronosequence: tundra (a-d), active
slump (e-h), stabilized slump (i-l). Representative soil profiles for each chronosequence
are shown in the first column. Note that 2012 CO2 and CH4 gas concentrations are on log
axes. Comparison of both δ13C of SOM (solid lines) and δ13C of CO2 (dashed lines) in
2011 and 2012 show that soil gases are considerably lighter than the soil material that
they are derived from. For all plots within the tundra and stabilized slump, the grey
horizontal line represents the transition from organic to mineral horizons.
78
4
Thesis Summary
4.1 Implications
This study is the first to examine how retrogressive thaw slumps give rise to changes
in soil physical properties that either limit or enhance soil CO2 efflux and how these
characteristics evolve over time. In a warming world, rapidly evolving, thermal erosional
features like the Selawik slump are predicted to increase in frequency (Gooseff et al.,
2009; Lacelle et al., 2010; Rowland et al., 2010), and therefore warrant further study.
Our findings indicate that quantifying the initial substrate quantity, used as a proxy
for carbon availability, as well as the displaced material’s diffusivity will be of great
importance to understanding CO2 efflux in these rapidly changing environments. The
arctic is a heterogeneous landscape where soil organic matter (SOM) ranges from 1-3%
as seen at the active Selawik slump, to 90-100% as seen in the stabilized slump and
yedoma soils of northern Alaska and Siberia (Kanevskiy et al., 2011). Higher carbon
fluxes are observed in environments with higher SOM (our tundra and stabilized slump),
suggesting a critical need to quantify soil carbon and other soil properties in these
environments.
On the floor of the active slump, materiel diffusivity also appears to play an
important role in controlling the efflux of CO2 and possibly CH4 to the atmosphere.
Although we observed CO2 concentrations at depth in the active slump to be 3x greater
than that of the tundra or stabilized slump, the surface efflux was ~2x lower than in the
tundra or stabilized slump in both 2011 and 2012. Low porosity, high water filled pore
space and calculations of effective diffusivity across the chronosequence all suggest that
the soil profile changes occurring due to the initiation of rapidly evolving thermokarst
79
dampen the release of CO2 gas to the atmosphere. However, with subsequent stabilization
and development of a soil profile, the active Selawik slump may become a greater source
of CO2 efflux to the environment than the initial tundra environment. Quantifying the
evolution of physical soil properties of thaw slumps will become important in
determining their long-term effect on the global carbon cycle.
Our research also suggests that geomorphic controls governing the initiation and
development of thaw slumps influences overall CO2 efflux. Thermokarst features such as
retrogressive thaw slumps which occur on hillslopes might be expected to be welldrained with aerobic soil conditions but we found largely anaerobic soil environments.
This ultimately affects carbon cycling by decreasing decomposition rate, which may
ultimately lead to slower soil C emissions. Due to the heterogeneous soil microenvironments that occur within these features at relatively small spatial scales, our ability
to project the effect of thaw slumps on the global carbon cycle remains limited.
4.2 Conclusion
We examined CO2 efflux along a thaw slump chronosequence from one of the
largest known retrogressive thaw slumps. To our knowledge, this is the only study to
measure CO2 efflux, and possible controls in a thermokarst feature of this type and
magnitude. In contrast to our initial expectations that the active slump would have higher
soil CO2 effluxes, we found that soil CO2 effluxes in the active thaw slump were
approximately half that measured in the tundra and stabilized slump for 2011 and 2012.
While some studies have shown a strong correlation with soil temperature and moisture
controlling CO2 efflux, our research suggests that physical soil properties such as bulk
80
density and soil organic matter also influence CO2 efflux. Further study of these local
controls will improve our understanding of carbon-cycling within these features.
We use a soil profile thaw slump chronosequence to examine how changes in the
soil environment due to rapidly evolving thermokarst features, affect soil column CO2
production as well as the resulting soil CO2 efflux. To our knowledge, this is the only
study to examine changes in physical soil characteristics and the resulting effect on CO2
efflux. Within the active slump, consolidated slump floor material limit gas diffusion,
inhibiting CO2 surface efflux and impounding CO2 at depth. We also determined that the
majority of our chronosequence exists under anaerobic soil conditions, despite high
surface slopes.
81
5
Supplemental Material
Figure 1. Photographs of each soil profile within the thaw slump chronosequence. Note
the absence of a dark, upper organic horizon in the active slump. The wooden pegs
delineate the depths where soil and gas were sampled: 3, 5, 15, 30, and 50cm. Because
photographs were taken looking down into the pits, the length scale becomes compressed
with depth.
82