<%BANNER%>

Effects of Thermokarst Slumps on Ecosystem Carbon and Nitrogen in Upland Arctic Tundra

  • STANDARD VIEW
  • MARC VIEW
MISSING IMAGE

Material Information

Title:
Effects of Thermokarst Slumps on Ecosystem Carbon and Nitrogen in Upland Arctic Tundra a Chronosequence Approach
Physical Description:
1 online resource (111 p.)
Language:
english
Creator:
Baron, Andres F
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Botany, Biology
Committee Chair:
Schuur, Edward A
Committee Co-Chair:
Mack, Michelle C
Committee Members:
O'connor, George A

Subjects

Subjects / Keywords:
arctic -- carbon -- chronosequence -- nitrogen
Biology -- Dissertations, Academic -- UF
Genre:
Botany thesis, M.S.
Electronic Thesis or Dissertation
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )

Notes

Abstract:
In the arctic, warming climate is leading to increased permafrost degradation. Thermo-erosional disturbances have the potential to alter permafrost characteristics and vegetation composition, ranging from gentle ground subsidence to abrupt hill slope features. Four retrogressive thaw slumps (RTS) chronosequences were located in upland arctic tundra in the vicinity of the Toolik Field Station and the Noatak National Preserve, Alaska, USA. The chronosequence approach assumed that the constant state factors of the sites studied were climate, potential pool of organisms that could colonize new niches open by RTS disturbance, and parental material. Relief and time were changing factors. We tested the hypotheses that C and N pools in surface soils will decline immediately after RTS due to erosional loss of the organic horizon; this initial decline will be followed by a net increase of surface C and N pools as a result of rapid re-accumulation of the soil organic layer as vegetation composition shifts from graminoid to shrub tundra, and several decades after disturbance, RTS surface soils will re-accumulate C and N pools similar to those found in the stabilized RTS and will support a plant community composition similar to the one found on the undisturbed tundra due to the stabilization of the permafrost and reduced organic matter inputs from plant communities as graminoid tundra once more dominates.  The results suggest that the undisturbed tundra soils contain large pools of buried organic matter that has been protected from decomposition by cold temperatures that inhibit microbial activity. RTS, as an abrupt spatial and temporal disturbance, have the capacity to rearrange massive quantities of surface soil and the C and N it contains, modifying the ecosystem soil and vegetation composition dynamics. RTS acts as a generator of new niches for plant colonization, exposes rich-nutrient mineral soil layers, mobilizes sediment to receiving waters, and modifies the local topography. RTS dynamics through time have the capacity to substantially alter the form and function in upland arctic tundra. Understanding the effects of RTS on ecosystem C and N dynamics offers a unique opportunity to establish links between climate change and the ecological impacts of changing disturbance regimes
Statement of Responsibility:
by Andres F Baron.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
General Note:
Description based on online resource; title from PDF title page.
General Note:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2012.
General Note:
Adviser: Schuur, Edward A.
General Note:
Co-adviser: Mack, Michelle C.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044679:00001

  • STANDARD VIEW
  • MARC VIEW
MISSING IMAGE

Material Information

Title:
Effects of Thermokarst Slumps on Ecosystem Carbon and Nitrogen in Upland Arctic Tundra a Chronosequence Approach
Physical Description:
1 online resource (111 p.)
Language:
english
Creator:
Baron, Andres F
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Botany, Biology
Committee Chair:
Schuur, Edward A
Committee Co-Chair:
Mack, Michelle C
Committee Members:
O'connor, George A

Subjects

Subjects / Keywords:
arctic -- carbon -- chronosequence -- nitrogen
Biology -- Dissertations, Academic -- UF
Genre:
Botany thesis, M.S.
Electronic Thesis or Dissertation
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )

Notes

Abstract:
In the arctic, warming climate is leading to increased permafrost degradation. Thermo-erosional disturbances have the potential to alter permafrost characteristics and vegetation composition, ranging from gentle ground subsidence to abrupt hill slope features. Four retrogressive thaw slumps (RTS) chronosequences were located in upland arctic tundra in the vicinity of the Toolik Field Station and the Noatak National Preserve, Alaska, USA. The chronosequence approach assumed that the constant state factors of the sites studied were climate, potential pool of organisms that could colonize new niches open by RTS disturbance, and parental material. Relief and time were changing factors. We tested the hypotheses that C and N pools in surface soils will decline immediately after RTS due to erosional loss of the organic horizon; this initial decline will be followed by a net increase of surface C and N pools as a result of rapid re-accumulation of the soil organic layer as vegetation composition shifts from graminoid to shrub tundra, and several decades after disturbance, RTS surface soils will re-accumulate C and N pools similar to those found in the stabilized RTS and will support a plant community composition similar to the one found on the undisturbed tundra due to the stabilization of the permafrost and reduced organic matter inputs from plant communities as graminoid tundra once more dominates.  The results suggest that the undisturbed tundra soils contain large pools of buried organic matter that has been protected from decomposition by cold temperatures that inhibit microbial activity. RTS, as an abrupt spatial and temporal disturbance, have the capacity to rearrange massive quantities of surface soil and the C and N it contains, modifying the ecosystem soil and vegetation composition dynamics. RTS acts as a generator of new niches for plant colonization, exposes rich-nutrient mineral soil layers, mobilizes sediment to receiving waters, and modifies the local topography. RTS dynamics through time have the capacity to substantially alter the form and function in upland arctic tundra. Understanding the effects of RTS on ecosystem C and N dynamics offers a unique opportunity to establish links between climate change and the ecological impacts of changing disturbance regimes
Statement of Responsibility:
by Andres F Baron.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
General Note:
Description based on online resource; title from PDF title page.
General Note:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (M.S.)--University of Florida, 2012.
General Note:
Adviser: Schuur, Edward A.
General Note:
Co-adviser: Mack, Michelle C.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044679:00001


This item has the following downloads:


Full Text

PAGE 1

1 EFFECTS OF THERMOKARST SLUMPS ON ECOSYST EM CARBON AND NITROGEN IN UPLAND ARCTIC TUNDRA: A CHRONOSEQUENCE APPROACH By ANDRES FELIPE BARON LOPEZ A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR TH E DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Andres Felipe Baron Lopez

PAGE 3

3 To my famil y: Maria Amelia, Cesar Esteban and Juan Camilo for being that constant reminder that I was raised to always give the best of me You did a good job mom. To my extended family and friends in Alaska and Florida for gave me the tools to build a better version of myself Th is is for all of you

PAGE 4

4 ACKNOWLEDGMENTS I thank my advisors Edward Schuur and Michelle Mack for sharing their knowledge and inspired me to do more than I ever tough I was capable of. It was an honor to work at your side and be part of your laboratories I thank the Arctic System Science Thermokarst Project (ARCSS/TK) members and the Toolik Field Station (TFS) Staff for all their good advice, training and friendship since 2009. I am very grateful with my Trucco, Sue Natali and her family Grace Crummer, Fay Belshe, Catilin Pries Hicks, Jordan Mayor, Rosvel Bracho and his family Christina Schadel, Verity Salmon, Garrett Arnold and Camilo Mojica Leda Kobziar, Kira Taylor Hoar, Camila Piz ano, Kamala Earl, Julia Reiskind, Heather Alexander April Melvin, Jennie De Marco, Silvia Alavarez, Undergrad Army F ou were there in the good and hard times and I will never forget that I thank the University o f Florida Master in Science program for extending my professional formation to higher levels and expose to me the great feeling of being part of the gator nation. Salvador Gezan, Larry Winte r, Willard Harrison, Rebecca Darner, and Kent Vliet for teaching me new concepts, and made me a more complete biologist. I thank Dr. Juan Posada for introducing me to the world of ecology and allow me the opportunity to have one of the best y ears in my lif e back in the i sland of Gorgona I thank my mentors and friends at the National University of Colombia Luis Carlos Monten e gro and Luz Marina Melgare jo for believing in me and share with me skills that were key in my success.

PAGE 5

5 I thank my friends Mike and Cor in n e for reminding me that the simple things in life are the ones t hat bring the most satisfaction, John Wood for show ing me that there is a good ending for guys like us and my parents and my grandparents for making me feel that I make this world a better place I am writing these pages right now just because you guys were holding my back without any conditions, just love.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ ........ 10 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 14 2 METHODS ................................ ................................ ................................ .............. 23 Study Site ................................ ................................ ................................ ............... 23 Chronosequence Aging ................................ ................................ .......................... 27 Soils Characterization ................................ ................................ ............................. 32 Rock Volum e Correction for Mineral Soil C and N Pool Estimates ......................... 35 Organic Soil C/N, Top Ten Mineral Soil C/N, and Organic Layer Loss/Gain Esti mation ................................ ................................ ................................ ............ 36 Surface Soil Organic C Pool Re Accumulation ................................ ....................... 37 Plant Community Composition ................................ ................................ ................ 37 Perc ent Tall Shrub Estimation ................................ ................................ ................ 38 Statistical Analysis ................................ ................................ ................................ .. 38 3 RESULTS ................................ ................................ ................................ ............... 46 Study Sites Aging ................................ ................................ ................................ ... 46 Surface Soils Characterization ................................ ................................ ................ 47 Rock Volume Correction for Mineral Soil C and N Pool ................................ .......... 54 Organic Soil C/N, Top Ten Mineral Soil C/N, and Organic Layer ............................ 55 Surface Soil Organic C Pool Re Accumulation ................................ ....................... 56 Percent Tall Shrub Estimation ................................ ................................ ................ 57 4 DISCUSSION ................................ ................................ ................................ ......... 80 5 CONCLUSIONS ................................ ................................ ................................ ..... 86 APPENDIX A SYSTEMATIC SHRUB SURVEYS ON FOUR RTS CHRONOSEQUENCES ACROSS UPLAND ARCTIC TUNDRA, ALASKA. AGE DISTRIBUTION OF SITES (LOBES) WHERE SHRUBS ROOTED IN MI NERAL SOIL WERE SAMPLED. ................................ ................................ ................................ .............. 88

PAGE 7

7 B DISSOLVE INORGANIC NITROGEN AND MELICH 1 EXTRACTABLE PHOSPHORUS IN SURFACE ORGANIC AND MINERAL SOILS ......................... 92 C PERCENT LOSS AND REACCUMULATION OF ORGANIC LAYER DEPTH AND SURFACE SOILS CARBON AND NITROGEN THRU TIME IN FOUR RTS CHRONOSEQUENCES IN UPLAND ARCTIC TUNDRA, AK ................................ 98 LIST OF REFERENCES ................................ ................................ ............................. 106 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 111

PAGE 8

8 LIST OF TABLES Table page 2 1 Study sites locations across upland arctic tundra in the North Slope of AK, USA. 1 GPS coordinates in decimal degrees ................................ ....................... 40 2 2 Plant species list for four RTS chronosequences in upland arctic tundra, North Slope, Alaska, USA ................................ ................................ .................. 41 2 3 Age of plant succesional processes started by RTS disturbance in four chronosequences in upland arctic tundra, North Slope, Alaska, USA. P. glauca rings we re count in this site instead of Salix sp. rings ............................. 43 2 4 Surface soil characteristics and tall shrub presence in four RTS chronose quences across upland arctic tundra, North Slope, AK, USA. *From the top 10cm of the mineral soil ................................ ................................ .......... 44 2 5 Fitted models used to predict carbon re accumulation rates in four RTS chronosequences in upland arctic tundra, AK, USA ................................ ........... 45 3 1 NE 14 RTS chronosequence s hallow soil characterization ................................ 58 3 2 Loon Lake RTS chronosequence shallow soil characterization .......................... 59 3 3 Itkillik RTS chronosequence shallow soil characterization ................................ .. 61 3 4 I minus 1 RTS chronosequence shallow soil characterization ............................ 63 3 5 Proportion of large rocks (>5cm in diameter) present on 50cm 3 soil pits dug on three RTS chronosequences around Toolik Field Station, AK, USA .......... 65 3 6 C pool change on the organic layer, top 10cm of the mineral soil and total pool change due to initial thermokarst disturbance in four RTS chronosequences across arctic tundra, AK, USA ................................ ............... 66 3 7 N pool change on the organic later, top 10cm of the mineral soil and total pool change due to initial thermokarst disturbance in four RTS chronosequences across arctic tundra, AK, USA ................................ ............... 67 3 8 Bulk density, %C and %N change at the top 10cm of the mineral soil due to initial thermokarst disturbance in four RTS chronosequences across arctic tu ndra, AK, USA ................................ ................................ ................................ 68 B 1 Ammonia, nitrate and phosphorus concentration in surface organic and mineral soil at NE 14 ................................ ................................ .......................... 93

PAGE 9

9 B 2 Ammonia, nitrate and phosphorus concentration in surface organic and mineral soil at Loon Lake ................................ ................................ .................... 94 B 3 Ammonia, nitrate and phosphorus concentration in surface organic and mineral soil at Itkillik ................................ ................................ ............................ 95 B 4 Ammonia, nitrate and phosphorus concentration in surface organic and mineral soil at I minus 1 ................................ ................................ ...................... 96

PAGE 10

10 LIST OF FIGURES Figure page 1 1 Helicopter view of a retrogressive thaw slump associated with the I minus 1 Lake, in the vicinity of Toolik Field Station, North Slope, Alaska, USA ............... 20 1 2 Distribution of study sites across upland arctic tundra, north and south slope of the Brooks Range, Alaska, USA ................................ ................................ ..... 21 1 3 Retrogressive thaw slump at NE 14 Lake in the vicinity of Toolik Field Station, North Slope, Alaska, USA ................................ ................................ ..... 22 3 1 Organic layer dynamics at four RTS chronosequences in upland arctic tundra, AK, USA ................................ ................................ ................................ 69 3 2 Surface soil organic C pool re accumulation dynamic after disturbance in four RTS chronosequences in upland arctic tundra, AK, USA ................................ ... 70 3 3 Surface soil organic N pool re accumulation dynamic after disturbance in four RTS chronosequences in upland arctic tundra, AK, USA ................................ ... 71 3 4 NE 14 RTS total C and N pools re accumulation dynamic through time. 72 3 5 Loon Lake RTS total C and N pools re accumulation dynamic through time ...... 73 3 6 Itkillik 1 RTS total C and N pools re accumulation dynamic through time ........... 74 3 7 Itkillik 2 RTS total C and N pools re accumulation dynamic through time ........... 75 3 8 Itkillik 3 RTS total C and N pools re accumulation d ynamic through time ........... 76 3 9 I minus 1 RTS total C and N pools re accumulation dynamic through time ........ 77 3 10 Surface soil organic C pool re accumulation predicted by ln(time), after initial disturbance, in four RTS chronosequences in upland arctic tundra, AK, USA ... 78 3 11 Estimation of the presence of tall shrub populations (i.e. Salix alaxensis, S. glauca, S. pulchra, S, hastata and Betula nana higher than 1.3m ) ..................... 79 A 1 Age distribution of shrubs rooted in mineral soil at Loon Lake RTS chronosequence, Noatak Nati onal Preserve, AK, USA ................................ ...... 88 A 2 Age distribution of shrubs rooted in mineral soil at NE 14 RTS chronosequence, Toolik Field Station, AK, USA ................................ ................. 89 A 3 Age distribution of shrubs rooted in mineral soil at Itkillik RTS chronosequence, Toolik Field Station, AK, USA ................................ ................. 90

PAGE 11

11 A 4 Age distribution of shrubs rooted in mineral soil at I minus 1 RTS chronosequence, Toolik Field Station, AK, USA ................................ ................. 91 C 1 Organic layer depth range of loss and re accumulation due to RTS disturbance in four chronosequences in upland arctic tundra, AK, USA ............. 99 C 2 Surface soil organic C pool range of loss and re accumulation due to RTS disturbance in four chronosequences in upland arcti c t undra, AK, USA ........... 100 C 3 Surface soil mineral C pool range of loss and re accumulation due to RTS disturbance in four chronosequenc es in upland arctic tundra, AK, USA ........... 101 C 4 Surface soil total C pool range of loss and re accumulation due to RTS disturbance in fo ur chronosequences in upland arctic tundra, AK, USA ........... 102 C 5 Surface soil organic N pool range of loss and re accumulation due to RTS disturbance in four chronosequences in upland arctic tundra, AK, USA ........... 103 C 6 Surface soil mineral N pool range of loss and re accum ulation due to RTS disturbance in four chronosequences in upland arctic tundra, AK, USA ........... 104 C 7 Surface soil total N pool range of l oss and re accumulation due to RTS disturbance in four chronosequences in upland arctic tundra, AK, USA ........... 105

PAGE 12

12 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for th e Degree of Master of Science EFFECTS OF THERMOKARST SLUMPS ON ECOSYST EM CARBON AND NITROGEN IN UPLAND ARCTIC TUNDRA: A CHRONOSEQUENCE APPROACH By Andres Felipe Baron Lopez August 2012 Chair: Edward Schuur Cochair: Michelle Mack Major: Botany In the arctic, warming climate is leading to increased permafrost degradation T hermo erosional disturbances have the potential to alter permafrost characteristics and ve getation composition ranging from gentle ground subsidence to abrupt hillslope features. Four retrogressive thaw slumps (RTS) chronosequences were located in upland arctic tundra in the vicinity of the Toolik Field Station and the Noatak N ational Preserve Alaska USA. The chronosequence approach assumed that the constant state factors of the sites studied were climate, potential pool of organisms that could colonize new niches open by RTS disturbance, and parental material. Relief and time were changing f actors. We tested the hypotheses that C and N pools in surface soils will decline immediately after RTS due to erosional loss of the organic horizon; this initial decline will be followed by a net increase of surface C and N pools as a result of rapid re accumulation of the soil organic layer as vegetation composition shifts from graminoid to shrub tundra, and several decad es after disturbance, RTS surface soils will re accumulate C and N pools similar to those found in the stabilized RTS and will support a plant community composition similar to the one found on the undisturbed

PAGE 13

13 tundra due to the stabilization of the permafro st and reduced organic matter inputs from plant communities as graminoid tundra once more dominates. The results suggest that the undisturbed tundra soils contain large pools of buried organic matter that has been protected from decomposition by cold tempe ratures that inhibit microbial activity. RTS as an abrupt spatial and temporal disturbance, have the capacity to rearrange massive quantities of surface soil and the C and N it contains, modifying the ecosystem soil and vegetation composition dynamics. RT S acts as a generator of new niches for p lant colonization, exposes rich nutrient mineral soil layers, mobilize s sediment to receiving waters, and modifies the local topography. RTS dynamics through time have the capacity to substantially alter the form an d function in upland arctic tundra U nderstanding the effects of RTS on ecosystem C and N dynamics offers a unique opportunity to establish links between climate change and the ecological impacts of changing disturbance regimes

PAGE 14

14 CHAPTER 1 INTRODUCTION Permafrost has been warming in many regions of the Arctic (Zhang et al. 1997, Osterkamp and Romanovsky 1999, Pollack et al. 2003, Frauenfeld et al. 2004, Oelke and Zhang 2004). Recent summaries report that in the decades between 1954 and 2003 annual average temperatures in the Arctic rose 1 C and that average winter temperatures increased 2 4 C Annual temperatures in the Arctic are expected to increase by 3 5 C and winter temperatures may increase by 4 7 C (e.g., Chapin et al. 2000, IPCC 2001, US/ARC 2003, ACIA 2004). Thawing of ice rich permafrost is leading to increased formation of thermo erosional features (TEF) collectively known as thermokarst (van Everdi ngen, 1998). TEF formation is highly variable, and its topographic and ecological consequences depend on interactions among slope position, soil texture, hydrology, and ice content (Jorgenson and Osterkamp, 2005). The nature and magnitude of disturbance as sociated with TEF is directly related to the thermal stability of the upper part of permafrost including the depth of the active layer and ground ice content (Lantuit and Pollard, 2008). This results in TEF ranging from gentle subsidence due to thawing of ice rich ground layers that result in altered drainage and slow shifts in plant species composition, to catastrophic hillslope failure and exposure of mineral substrate that initiates primary succession of plant communities. TEF formation, triggered by in creasing permafrost temperature, has the potential to abruptly alter ecosystem soil carbon (C) and nitrogen (N) pools and fluxes. Additionally, new niches for colonization by plant species such as tall deciduous shrubs can result from TEF formation that ma y further alter the dynamics of these elements as

PAGE 15

15 warming increases frequency of disturbance, which is important to understand the rate of element loss and re accumulation to predict effects on arctic C balance. There is growing evidence that tundra vegeta tion is responding to increasing temperatures (Chapin et al., 1995, Stow et al., 2004; Tape et al., 2006). In the arctic, a widespread shift from graminoid tundra to shrub dominated vegetation appears to be underway (Sturm et al. 2001, Lloyd et al. 2003, T ape et al. 2006). TEF formation may be an important driver of shrub expansion because exposure of mineral soil favors recruitment of deciduous shrub species (Callaghan et al., 2004, Lloyd et al. 2003) and surface subsidence appears to increase deciduous sh rub dominance (Schuur et al., 2007). Plant nutrients in Arctic soils, particularly nitrogen (N), are available to plants at low rates because microbial decomposition and mineralization rates of organic matter are constrained by low temperature (Russell, 1 940). By influencing biophysical and biogeochemical processes, tall deciduous shrubs can significantly alter ecosystem structure and function, and feedbacks to climate warming (Shaver and Chapin, 1980; Chapin et al., 1995; Epstein et al., 2004; Schimel et al., 2004; Sturm et al., 2005; Rhoades et al., 2008). Specifically for N, shrubs could increase cycling rates through either increased N concentrations in litter (Chapin and Shaver, 1988) or via winter soil warming resulting from greater snow accumulation beneath shrubs (Sturm et al. 2001; Sturm et al. 2005). Feedbacks between shrub abundance and TEF may lead to further changes in permafrost conditions, wildlife habitat, and ecosystem function (Forbes et al., 2001, McGuire et al., 2006).

PAGE 16

16 Research on the rev egetation of tundra after TEF disturbance has largely focused on describing the initial stages of plant succession and/or the trajectory of this process over one or two decades (Lambert, 1972; Lambert, 1976; Ovenden, 1986; Bartleman et al, 2001; Jorgenson et al., 2001; Mackay and Burn, 2002). There are few studies, however, of the effects of TEF on plant species and community level responses over multiple decades, based on the specific developmental and stabilization processes of a TEF. Studies of TEF in th e Canadian sub arctic tundra (Lantz et al. 2009) and the boreal ecotone (Burn and Friele, 1989) stand as one of the most complete attempts to understand successional processes established by TEF disturbance on a multi decadal time scale, ranging from the o pening of new niches due to exposure of mineral substrate to the established plant communities after TEF stabilization. The extent of the alteration of the organic horizon is one of the key determinants of the ecological consequences of TEF disturbance in the Arctic (Kershaw, 1983a; Ebersole, 1985). The initial formation of TEFs mobilizes, redistributes and/or erodes the soil organic horizon. In addition to reducing pools of C and N stored in the organic layer, disturbance may expose deeper layers of organ ic or mineral soil, thus altering substrates for plan colonization. Disturbance of the organic layer may also compact organic soils or mix mineral soil into the organic layer, decreasing its insulating capacity and increasing heat penetration into the grou nd (Haag and Bliss, 1974). The extent of the disturbance to the organic layer and its properties plays a significant role in the alteration of the ecosystem stock of C and N and the re accumulation rate post disturbance, especially in a system where most o f the C and N accumulation occurs in the soil. As for the aboveground component, plant productivity

PAGE 17

17 increases post TEF disturbance (Lantz et al 2009), but it is not clear whether the increased plant growth, including shifts in vegetation composition (i.e. a more shrub dominated tundra) is sufficient to compensate for losses of soil organic matter. The study of TEF, as an abrupt spatial/temporal disturbance associated with thawing permafrost, offers a unique opportunity to establish links between climate ch ange and the ecological impacts of changing disturbance regimes. The study of changes in ecosystem dynamics over long periods of time can be successfully accessed by the adequate implementation of the chronosequence co ncept (Walker et al. 2010; Figure 1 1 ) In order to track TEF effects on ecosystem C and N stocks and plant composition through several decades, four TEF chronosequences of time since disturbance, were characterized in Arctic tundra on the North and South slope of the B r ooks Range, Alaska, USA ( Figure 1 2 ). The mode of permafrost degradation associated with these chronosequences fit the description of a retrogressive thaw slump (RTS) as defined by Jorgenson and Osterkamp (2005). RTS have three main elements ( Figure 1 3) that can be easily inden tified on active features: a vertical or sub vertical headwall, which contains most of the active layer and ice poor organic or mineral materials, a headscarp whose angle varies between 20 and 50 and which retreats by the thawing of ice rich expose perma frost due to direct effect of solar radiation, and the slump floor, which contains the mix of materials eroded from the headwall and acts as a transitional deposit zone (Lewkowicz, 1987; de Krom, 1990). Three of the RTS chronosequences (NE 14, Loon Lake an d Itkillik) are surrounded by non acidic, non tussock sedge, dwarf shrub, moss tundra (CAVM Team, 2003). The

PAGE 18

18 fourth RTS chronosequence (I minus1) is surrounded by acidic, tussock sedge, dwarf shrub, moss tundra on top of wet sedge peat (CAVM Team, 2003; M. C. Mack personal observation). An important step towards understanding how increased TEF and erosive disturbances will affect landscape scale C and N balance and biogeochemical feedbacks to climate is characterizing the effects of the disturbance on tempo ral dynamics of plant communities and ecosystem C and N stocks contained in surface soils across multiple decades characterizing losses of C and N stocks during disturbance and re accumulation after disturbance. The goal of our study was to address three k ey questions about the effect of TEF disturbance in upland arctic ecosystems: (1) How much C and N is lost from surface soil due to RTS formation? (2) What is the rate of surface soil C and N re accumulation after disturbance? and (3) After several decades do surface soil C and N pools return to pre disturbance levels? We tested the following hypothesis: (1) C and N pools in surface soils will decline immediately after TEF formation because of erosional loss of the organic horizon; (2) this initial decline in surface soil element pools will be followed by a net increase of surface C and N pools as a result of rapid re accumulation of the soil organic layer because vegetation composition shifts from graminoid to more productive shrub tundra and (3) several d ecades after disturbance, RTS surface soils will re accumulate C and N pools similar to those found in the stabilized RTS and will support a plant community composition similar to the one found on the adjacent, undisturbed tundra; this is due to the stabil ization of the permafrost (i.e. exhaustion of the ground ice and/or insulation of

PAGE 19

19 the headwall by organic/mineral material) and reduced organic matter inputs from plant communities as graminoid tundra once more dominates.

PAGE 20

20 Figure 1 1 Helicopter view o f a retrogressive thaw slump associated with the I minus 1 Lake, in the vicinity of Toolik Field Sta tion, North Slope, Alaska, USA Undisturbed tundra, early, mid and late succession sites were identified

PAGE 21

21 Figure 1 2. Distribution of st udy sites across upland arctic tundra, north and south slope of the Brooks Range, Alaska, USA. Imagery provided by Google Earth/Digital Globe. Figure created by Toolik Field Station GIS Office, 2012

PAGE 22

22 Figure 1 3 Retrogressive thaw slump at NE 14 La ke in the vicinity of Toolik Field Station North Slope, Alaska, USA

PAGE 23

23 CHAPTER 2 METHODS Study S ite Three of the four RTS chronosequences (NE 14, I minus 1 and the Itkillik) are located in the vicinity of the Toolik Field Station (TFS, N68 38 0 W149 36 0 1 ), which is approximately 255 km north of the Arctic Circle and at an elevation of 720 m above sea level in the foothills province of the Brooks Range, AK. These sites were accessed via helicopter from TFS. The major vegetation types found in the study regio n near TFS include graminoids (mainly the tussock forming sedge Eriophorum vaginatum and Carex bigelowii ), deciduous shrubs (mainly Betula nana and several species of Salix ), evergreen shrubs mosses and lichens (Shaver & Chapin 1991). The dominant vegetati on community is moist acidic tussock tundra, defined by the presence of E. vaginatum. TFS is located in the warmest of the Arctic subzones ( Subzone E, CAVM, 2003). The fourth RTS chronosequence (Loon Lake) is located in the vicinity of the Noatak National Preserve (NNP, N 67 802 68 839, W155 850 162 855 1 ), which is on the south slope of the Brooks Range in northwestern Alaska. This site was accessed via bush plane from Kotzebue, and then via helicopter from a central field camp. The general vegetation types present in northwestern Alaska are moist acidic tundra, dominated by E. vaginatum Dryas fell field, ericaceous shrub tundra, Eriophorum Carex wet meadow, solifluction slopes, and boreal forest (Viereck et al. 1992). Low shrub or tussock tundra dominates much of the NNP. Forests, dominated by Picea 1 Coordinates in decimal degree s

PAGE 24

24 glauca occur almost strictly in the south western corner of the Preserve. All RTS chronosequences were located with helicopter over flights, aerial photographs (ARCSS/TK and TFS GIS office aerial image files), satellite imagery (Google Earth, 2009, 2010 images) and ground exploration of the terrain (Garmin GPSMap, 60CSx; accuracy of le ss than 10m 95% of the times ; datum WGS 84) Within each chronosequence all lobes were located in the same parental material (Hamilton, 2003, Hamilton 2008): NE 14 and I minus 1 were located on the drift of Itkillik phase II (till and ice contact deposits) Itkillik series on undifferentiated lacustrine deposits, and Loon Lake on Holocene floodplain deposit (alluvium). Lobes within a chronosequence were close enough that the climate and the biotic potential capable of colonizing substrate exposed by the di sturbance were likely to be similar. Relief (topography and aspect), however, varied somewhat among lobes within chronosequences: NE 14 lobes (three distinct features) were located in the north portion of the NE 14 Lake. All of them had a southern aspect a nd changes in topography were due to differences in plant and organic matter cover of the headwalls, direct exposure of the slump floor to radiation and differences in snow accumulation capacity during the winter and spring seasons. Itkillik lobes (three d istinct features) had southern and eastern aspects according to their origin along the Itkillik River. Slope steepness was considerably less pronounced on the lobe with southern aspect and brakes in slope were frequently observed inside the two lobes with eastern aspect. I minus 1 lobes (seven features) were surrounding the coastal area of the I minus 1 Lake. Three of them (lobes 1, 2 and 3) had a northern aspect and four of them (lobes 4, 5, 6

PAGE 25

25 and 7) had southern aspects according to their locations along the coast of the lake. Lobes with a northern aspect had a greater snow accumulation and also displayed well defined headwalls with steeper slopes in comparison to the southern aspect lobes. Loon Lake lobes (five distinct features) had northern aspect and w ere located on the south portion of the Loon Lake (Noatak, AK). Headwalls were well defined and differences in topography responded to changes in exposed slump floor and headwall activity. New lobes were much steeper than older lobes, and in the longer chr onosequences, there was variation in aspect and inclination of the slope due to stabilization of the headwall by re accumulation of organic and mineral material on its active face, which left a poorly defined headwall scar. Finally, each site within each c hronosequence varied in putative time since disturbance. Sites were first selected based on visual characteristics that indicated time: younger sites had large expanses of mineral substrate exposed and the main three elements of an RTS (headwall, headscar p and slump floor) were well defined. Intermediate aged sites were re vegetation was evident and had less pronounced RTS elements. The oldest sites were only detectable using light detection and ranging imagery ( AK DOT LiDAR imagery prepared by Krieger, 20 12 ), which allowed us to detect the three main elements of RTS despite them being difficult to identify from the ground The Itkillik chronosequence was the only one where each of its features (lobes) were not located adjacent to each other, but separated i n space depending on the zone along the Itkillik River where the disturbance initially took place. In addition, organic layer depths of adjacent, undisturbed tundra were substantially different across the

PAGE 26

26 sites, thus we treated this chronosequence as three distinct pairs: each lobe paired with its own undisturbed, adjacent tundra. Once the study sites were selected and characterized in terms of TFS type, boundaries and the three main elements of each RTS were mapped (i.e. headwall, headscarp and slump floor ). Areas inside each feature and adjacent undisturbed tundra were surveyed to find appropriate locations for study transects. Within each feature, a 50m by 4m belt transect was located in vegetation representative of the slump floor of each feature. Our pr imary goal within each feature was to sample a zone that was not being affected by the headwall dynamics, drastic changes in slope, the associated body of water (i.e. lake or river) or the borders of the feature (potential erosion zones). In the adjacent, undisturbed tundra, transects were located that were relatively close to the RTS but were not obviously affected by the disturbance. GPS coordinates were re co rded for each transect (Table 2 1). For NE 14 chronosequence, we located one transect inside each of the three lobes and located two pre disturbance (control) transects on the adjacent, un disturbed tundra. One pre transects, and the second pre disturbance transect was paired with the third lobe transect. The Itkillik chronosequence consisted of three paired disturbed and control transects. I minus 1 had the highest number of lobes and control transects: seven transects inside well defined lobes, and three control transects. The spatial dist ribution of the lobes along the coast of the I minus 1 Lake was represented by two clusters, matching the slope aspect groups defined above: three lobes (lobes 1, 2 and 3) were located on the south west coast of the lake, and four of them (lobes 4, 5, 6 an d 7) on

PAGE 27

27 the north east coast of the lake. Loon Lake chronosequence consisted of five lobes with one transect within each, and one control transect. For NE 14, Itkillik and Loon Lake chronosequences, the vegetation of the surrounding undisturbed tundra was classified as non tussock sedge, dwarf shrub, moss tundra, with peaty non acidic soils (Walker et al. 2002) Frost boils (barren patches of cryoturbated soil) were common. This vegetation type spans from Fennoscandia to Russia and contributes 11.2% of circ umpolar arctic tundra cover. Plant heights in this vegetation type are generally 10 20cm. Salix species and other dwarf shrubs (i.e. Rhododendron lapponicum ) are common, but once in the open tundra, tend to display an hemiprostrate growth form, reducing t heir potential to be erect, which decreases their heights. Well developed moss layers (5 20cm thick) are common. A list of the plant species present in these three chronosequences was recorded (Table 2 2 ). For the I minus 1 chronosequence, the surrounding vegetation was classified as moist tussock tundra dominated by tussock cottongrass ( Eriophorum vaginatum ) and dwarf shrubs <40cm tall (Walker et al. 2002). This vegetation type is characteristic of landscapes with ice rich permafrost and shallow active layers. It contributes about 6.6% of circumpolar tundra cover (Walker et al. 2002) Plant cover is nearly continuous (80 100%). The heigh t of the plant canopy varies from about 20 to 40cm. A list of the plant species present in this chronosequence was recorded (Table 2 2 ). Chronosequence Aging Across all chronosequence sites, we used dendrochronological methods to determine the age of the o ldest shrub rooted in the mineral soil layer. We assumed that this age represents when RTS soils stabilized and plant colonization began. It is likely a minimum age for the feature. In a subset of features (i.e. 2 features at NE 14 and 4

PAGE 28

2 8 features at I minu s 1; Table 2 3), we also used radiocarbon dating of dead mosses at the base of the recovering soil organic layer as a second constraint on site age. Like the shrubs rooted on mineral soil, we assumed that these moss bodies at the organic mineral soil inter face would have come from colonization after RTS soil stabilization. In tundra vegetation, shrub rings can be used to date stands in much the same way as they are used in forest stands (Johnstone and Henry 1997, Rayback and Henry 2005) Salix spp. (mainly S. alaxensis, S. glauca and S. pulchra ) stems were analyzed from adult individuals rooted in mineral soil inside each of the different chronosequence sites. We assumed that adult shrubs rooted in mineral soil were representat ive of individuals that colonized the TEF as a seed, once the landscape was disturbed but the surface had stabilized. Systematic surveys of adult shrubs were conducted at most of the sites of each chronosequence. The survey area was limited to the zones i nside the features that were not subjected to any further disturbance due to the headwall processes, drastic changes in slope or the effect of the body of water associated with the RTS. Two additional belt transects were established parallel to each side o f the established transect in each site (see above), with an extended belt from 4 to 8m in order to delimit the survey area. The distance between belt transects depended on the area of each one of the sites sampled. These three belt transect delimited a sa mpling area of 500m 2 where adult shrubs rooted in mineral soil were harvested. Soil pits were dug at each harvest point to expose root system and visually confirm the soil type as mineral. When the visual characterization of the soil type was difficult, s oil samples were collected and sent to the

PAGE 29

29 University of Florida for bulk soil %C analysis ( ECS 4010 elemental analyzer; Costech Analytical, Valencia, California, USA). There were few adult shrubs rooted in mineral soil of young RTSs compared to revegetate d features in most of the chronosequences. In young sites, most of the adult individuals were rooted in buried organic material, presumably coming from residual did not include any individuals rooted in organic debris in our survey. Revegetated sites exhibited a higher density of individuals rooted in mineral soil. Similarly to the young RTSs, there were few adult shrubs rooted in mineral soil at the stabilized sites. It is important to mention that on the stabilized feature found at the Loon Lake chronosequence, Picea glauca sections were used to determine the age of the plant succesional process established by this feature. These systematic surveys took place in three d ifferent summers. In 2009, 9 individuals from the revegetated lobes in NE 14 were harvested and aged, and 19 individuals from one of the revegetated lobes in I minus 1. In 2010, 2 individual from the recently from lobe in NE 14 were harvested and aged, and 34 individuals from the revegetated lobes. From the Itkillik chronosequence, 24 individuals were harvested and aged, and 19 individuals (plus 3 P. glauca individuals) from the Loon Lake chronosequence 7 individuals were harvested and aged from recently form lobe in I minus 1 and 16 individuals from the revegetated lobes. Finally, in 2011, 5 individuals from one of the revegetated lobes in I minus 1 were harvested and aged (Table 2 3). The portion of t he shrub where the root system transitions into the main stem (i.e.

PAGE 30

30 cleaned and sliced into three horizontal sections. Only the bottom most section was used for aging. Sections were dried at 60 C for 48 hours and sanded using progressively finer grades of sandpaper (180, 220, 320, 400 and 600 grit) then scanned at 2400 dpi to determine the number of growth rings using an image analyzing system ( Windendro TM Basic 2009, Regent Instruments Canada Inc.). Where ring boundaries were difficult to discern, a stereoscope was used to confirm the software accuracy. For the P. glauca encountered in the stabilized lobe at Loon Lake, stems sections of three different individuals (ba sal diameters of 11.8cm, 12.75cm and 8.25cm) were used instead of Salix processing of the tree stem sections was The ring count of all the samples processed generated a shrub age dist ribution for each lobe sampled (Table 2 3; Appendix A ). Revegetated sites across all the chronosequences displayed a wide distribution of ages (Table 2 3; Appendix A ). Estimates of RTS age based on ring counts were compared to moss radiocarbon dates in the two revegetated sites at the NE 14 chronosequence and one revegetated site at the I minus 1 chronosequence (Table 2 3). Estimates of RTS ages based on Salix sp. dendrochronology and 14 C from dead mosses at the base of the recovering soil organic layer wer e used together at sites wherever possible to cross check dates For selection of radiocarbon sampling sites, well developed moss colonies were selected from the sampling area. Monoliths were selected where a clear distinction between the mineral and the organic layers was visible, and there was no evidence of mixing of the two layers, as might happen during the erosion event. Our goal was to

PAGE 31

31 select patches were the organic layer had developed after the RTS first impacted the area. In most of the sites sampled on each chronosequence, a set of five moss cores was collect ed; only a subset was analyzed. Moss core samples for radiocarbon dating consisted of soil monoliths where the interface between the organic and the mineral layers was well defined. For the younger sites at each chronosequence, the search for adequate samp ling sites was difficult due to the sparse moss cover. This limited the collection of samples from these sites. For the revegetated and stabilized site on each one of the four RTS chronosequences, a systematic survey was conducted using the same area as th e shrub surveys. The use of radiocarbon as an age estimator takes advantage of the increased levels of 14 C in atmospheric CO 2 as a consequence of thermonuclear weapons testing in the early 1960s (Gaudinski et al. 2000). This global 14 be used to trace the time elapsed since C in plant tissues was fixed from the atmosphere by photosynthesis, and to estimate C cycling rates in an ecosystem. According to Gaudinski et al. (2001), after the nuclear test ban treaty in 1963, the amount of 14 C in atmospheric CO 2 has decreased due to exchange with the ocean and terrestrial biosphere, and dilution by burning of 14 C free fossil fuels. Moss cores were frozen and shipped to the University of Florida for processing, where the mineral and organic layer were separated. From the organic soil, we sampled the bottom most layer (A) from 0 (interface with the mineral soil) to 1.75cm above the min eral soil and a second layer (B) from 1.75 to 3.5cm above the mineral soil. Moss stems were assumed to be the first colonists and thus their age should indicate the time of TEF soil surface stabilization.

PAGE 32

32 Moss samples were sequentially extracted to remove carbon compounds, leaving behind purified holocellulose (Gaudinski et al. 20 05) Cellulose samples were combusted, purified, and converted to graphite, and 14 C content was analyzed on the University of California Irvine Keck Carbon Cycle Accelerator Mass Spectrometer ( NEC 0.5MV 1.5SDH 2 AMS system. National Electrostatics Corp., Middleton, WI ) Because 14 C values above zero have equivocal dates due to the shape of the bomb c urve, we analyzed both A and B increments. Samples where the B section was more enriched in 14 C than the A section were assigned to the ascending slope of the bomb peak ( 1966), while samples where the B section was more depleted than A section were assign ed to the descending slope ( 1966) (Hicks et al. 2011). Soils Characterization Surface s oils were sampled to determine pools of C and N in organic matter and mineral soils along a 50m by 4m belt transect inside each RTS and in undisturbed, adjacent tundra. We also sampled Melich 1 extractable P and KCl extractable d issolved inorganic N (Appendix B ), which may be indicative of plant available nutrients. Transects inside the chrono sequences sites were located parallel to the headwall, but in a zone far enough from it that was not being affected by its dynamics, drastic changes in slope, the associated body of water (i.e. lake or river) or the borders of the feature (erosion zones). This transect area was large enough to include all growth forms ( i.e. shrubs, graminoids, forbs, and non vascular plants) present in each feature. Organic and mineral soils were sampled volumetrically to the depth at which we encountered rock or ice was en countered at six sampling points per transects.

PAGE 33

33 For organic layer samples, a small pit was dug to the surface of the mineral soil, and then an intact organic profile was removed from the side of the pit with a serrated knife. Depth, length and width of the profile were recorded. Mineral samples were collected, when possible, using a 7cm internal diameter by 15cm in length soil core, as a sequential soil profile based on the area where the organic sample was extracted. Mineral soil cores were taken at the ba se of the pit where the organic profile was removed. Samples were placed on ice and transported in coolers back to the laboratory facilities at TFS for preliminary processing including bulk density, gravimetric water content, pH, bulk soil percent C and N, and total surface soil C and N pools. Organic and mineral soil samples were sectioned into depth increments (organic layer: 0 5cm increment and 10cm increments thereafter; mineral layer was 10cm increments). For comparison among sites, all soil increments were standardized (i.e. binned) by depth using discrete soil depth ranges for the organic and mineral layer. This process allowed us to account for the variability in organic and mineral layer depths along transects on each site on each chronosequence. O rganic layer depth was taken every 5m on each transect to detect the loss and re accumulation of organic matter thru time on each chronosequence. Small pits were dug until the surface of the mineral soil, making sure that the interface in between the organ ic and mineral layers was evident in the pit profile. Depth was recorded from the top of the surface moss to the surface of the mineral soil. The different organic layer depths were also standardize d by depth in the same fashion as the organic and mineral soil increment depths.

PAGE 34

34 Soil bulk density (Db, g/cm3), the bulk soil percent of C and N (% on a mass:mass basis), and the depth of each soil increment (cm) were used to calculate the total stock of C and N for all the surface soil profiles sampled per trans ect as follows: Soil increment pool = increment Db (g/cm3) % C or N increment depth (cm) 100 = g/m2 Total C and N pools per site were expressed in g/m2, and calculated as an average of C and N pool values per surface soil sampling point per transect. Prior to analysis, organic and mineral soil samples were weighed, sectioned into increments, and homogenized by hand to remove the larger than 2mm diameter fraction (e.g. coarse woody debris, roots, rhizomes, fiber and rock). Gravimetric water content was calculated by subtracting the dry (organic soils: 60 C for 48 hours; mineral soils: 110 C for 48 hours) weight of the soil from the wet weight of the soil and then dividing by the dry weight of the soil. Soil bulk density was calculated for each organic and mineral soil increments as the mass (g) per unit volume (cm3) of dry soil. For both organic and mineral soil increments the dry weight was calculated by multiplying the homogenized wet weight of the increment by the dry:wet ratio. Volume (cm3) of soil was calculated as total increment volume minus the volume displaced by rocks. Soil pH was estimated using a ratio of 1:1 between an air dry subsample of the homogenized soil fraction and DI water. The mixture was allowed to settle for 30 minutes before sub merging a calibrated pH electrode ( Model 250Aplus, Thermo Orion; Beverly, MA) Values of pH were transformed to hydrogen concentration values and means per site per chronosequence were obtain. To determine bulk soil percent C and

PAGE 35

35 N, a subsample of the homo genized soil fraction was dried at 60 C for 48 hours, ground to a fine powder on a Wiley mill ( T4276 Wiley Intermediate Mill, 115V, 60HZ; Thomas Scientific, Swedesboro, NJ) with a #40 mesh screen, and then analyzed using an ECS 4010 elemental analyzer ( Costech Analytical, Valencia, California, USA). For Melich 1 extractable P, an air dry subsample of the homogenized soil fraction was extracted using a double acid solution (0.05N HCl: 0.025N H2SO4). The mixture was shaken for 2 hours, and then centrifuge and filter through a Whatman #5 paper filter. The filtrate was analyzed using a BioTek PowerWave XS micro plate reader ( BioTek Instruments, Inc., Winooski, VT) after being processed following the ascorbic acid, molybdenum blue method (Murphey and Riley, 19 62) amended for small volumes. Rock Volume Correction for Mineral Soil C and N Pool Estimates In order to quantify the soil volume display by large rocks (i.e. 5cm in diameter) that would have be avoided in the above soil sampling protocol, three 50cm3 pi ts were dug inside the youngest lobes of the three chronosequences located at the vicinity of TFS (i.e NE 14, I minus 1 and the Itkillik series). These pits were located parallel to the 50m by 4m belt transects. The limit of 50cm on depth was arbitrarily s et assuming that, below this depth, the constitution of the soil profile is highly variable, often frozen and difficult to access. Pits were dug during the summer of 2011, in a period when the active layer depth was deeper than 50cm for all the sites (i.e. late in July). For each pit, soil was shoveled onto a large plastic tarp. Once the dimensions were reached (50cm in depth by 50cm in length by 50 cm in width) the total amount of soil was homogenized and separated into three categories: buried peat, soil and rocks. Rocks were weighed in the field on a digital scale. A random sub sample of rocks was

PAGE 36

36 collected and returned to TFS for estimation of the relationship between mass and volume. Buried peat and soil was weighed in the field and sub samples were ret urned to the lab for estimation of the relationships between volume and mass. To estimate total pit volume, we lined the pit with a plastic bag and filled it with Styrofoam packing peanuts. Bag was returned to TFS, where we measured the volume of the peanu ts The proportion of the total pit volume that was displaced by large rocks was calculated as an average from the three pits dug for each site, and used as a multiplier factor for each one of the mineral C and N surface soil pools in order to correct for t he effect of the large rocks present in the study sites located in the vicinity of TFS. Or ganic Soil C/N, Top Ten Mineral Soil C/N, and Organic L ayer Loss/Gain Estimation After characterizing important surface soil characteristics (i.e. organic C and N poo ls, organic and mineral C and N pools, and re accumulation of organic layer thru time) altered by RTS disturbance, a quantification of the actual mass (Kg) per unit area (m 2 ) of organic C and N loss or gain, and organic layer depth re accumulation on each site on each chronosequence, was conducted. The values of organic C and N pool, organic and mineral C and N pool, and organic layer depth estimated for the undisturbed, adjacent tundra used as control on each chronosequence, were used as reference values ( t 0 ). The trajectory of these surface soil characteristics was established once the reference values were compared to the values reported for each one of the aged lobes for each chronosequence. The propagation of uncertainty for each comparison was calcula ted following the rule of addition/subtraction of errors as follows:

PAGE 37

37 U f = (U a ) 2 + (U b ) 2 b Where U f equals the error term of resulting comparison, U a equals the standard error of the parameter at t0, and U b equals to the standard error of the % of C and N lost from the organic layer, the amount lost or gained from the top 10 cm of the mineral layer, changes in bulk density, % of C and N from th e top 10 cm of the mineral layer, and the net loss of C and N from the surface soils at each chronosequence were established comparing the predisturbance values ve rsus the values detected right after a young RTS altered the surrounding tundra. Surface Soil Organic C Pool Re A ccumulation Recovery curves of C pool thru time were established using logarithmic models for all the four chronosequences (plus a lineal model for Loon Lake). Projection were based on the percent undisturbed C predicted from ln(time). I minus 1 and Loon Lake projected beyond 50 years (i.e. these chronosequences had sites older than 50 years), but NE 14 and Itkillik did not project beyond this point in time. Correlation coefficient ( r 2 ) had a range from 0.51 to 0.93 (Table 2 5), expressing a reliable estimation of C pool re accumulation and showing differences in the rates per study sites. Plant Community Composition The same transects established for the surface soil sampling were used for visual estimates of percent cover, talle st shrub, point intercept, aboveground biomass estimates, shrub density, surface soil temperature, thaw depth measurements, and depth of organic layer for all the four RTS chronosequences and the adjacent, undisturbed tundra.

PAGE 38

38 A 1m 2 frame with a grid of a hundred 10 X 10cm 2 was used to estimate percent cover at locations every 10m along the sampling transect (50m). This method allowed for a relatively fast estimation of the most frequent species and the abundance of species per area (although rare species c an be missed using this technique). A 1m point intercept scope was also used to identify vegetation composition via point intercept every 0.5m along the sampling transect. Species frequency and vertical distribution of species, including the substrate, wer e recorded. Percent Tall Shrub Estimation A variation of the point intercept technique was applied in order to estimate the presence / absence as well as the difference in tall shrub populations (i.e. Salix alax ensis, S. glauca, S. pulchra, S. hastata and Betula nana ) in between the different lobes that compose each of the four RTS chronosequences. A 1m point intercept scope was used, every 0.5m, as a reference to detect shrubs taller than the scope height. The same transects established for the surface soi l sampling were also used for this procedure. Complementary to this index, measurements of the tallest shrubs associated to the percent cover frame were recorded: these two pieces of information allowed us to obtain a better estimation of changes in tall s hrub populations in time, for each chronosequence. Statistical Analysis We used mixed linear analysis of variance (ANOVA, JMP 8.0) to test for RTS effects on surface soil variables and tall deciduous shrubs through time. Site (chronosequence), transect lo cation (inside or outside RTS) and soil type (organic or mineral soil) were fixed treatment effects. Surface soil variables for the organic and

PAGE 39

39 mineral soil included: organic layer depth, bulk density, % C and % N C:N ratio and C and N pools Lobe (each one of the features in a particular chronosequence) was considered as a random effect, nested within chronosequence and transects location. Sample points (random effect) along the 50m by 4 m belt transects, were experimental units for transect location and soi l type. The F statistic, degrees of freedom, and P va lues are reported (Table 2 4 ). To explore the relationship between lobe age and percent shrubbiness on each of the four RTS chronosequences, linear regression analysis (JMP, version 8.0) were used. When significance was obtained from ANOVAs, we used LSMeans Tukey HSD test (JMP, version 8.0) to control for family wise error rates. For NE 14, Itkillik and I minus 1, logarithmic regression models were applied using the changes in carbon pool detect ed thru ti me, in order to establish carbon re accumulation rates For Loon Lake, both logarithmic and linear regression s were applied (Table 2 5) All the data set used were transformed using the Box Cox transformation method and tested for normality using the Shapiro Wilk W Goodness of fit (JMP, version 8.0). All the tests were cond ucted at an alpha level of 0.05

PAGE 40

40 Table 2 1. Study sites locations across upland arctic tundra in the North Slope of AK USA 1 GPS coordi nates in decimal degrees Chronosequence Site Coordinates (lat, long) 1 Loon Lake control 67.927667 N, 161.963717 W 1 67.928467 N, 161.961333 W 2 67.927817 N, 161.96455 W 3 67.92905 N, 161.959283 W 4 67.928378 N, 161.962161 W 5 67.929883 N, 161.95645 W Itkillik 1 control 68.633222 N, 149.7985 W 1 68.632611 N, 149.798639 W Itkillik 2 control 68.666256 N, 149.817472 W 1 68.666 75 N, 149.817472 W Itkillik 3 control 68.671592 N, 149.844794 W 1 68.673667 N, 149.844333 W NE 14 control 1 68.679292 N, 149.621064 W control 2 68.679244 N, 149.627725 W 1 68.678833 N, 149.623278 W 2 68.679417 N, 149.625361 W 3 68.6785 N, 149.628139 W I minus 1 control 1 68.553 N, 149.574444 W control 2 68.562267 N, 149.571017 W 1 68.553417 N, 149.573639 W 2 68.553778 N, 149.57475 W 3 68.554483 N, 149.57545 W 4 68.559117 N, 149.573283 W 5 68.560483 N, 149.5742 W 6 68.56055 N, 149.563533 W 7 68.56065 N, 149.569767 W

PAGE 41

41 Table 2 2. Plant species list for four RTS chronosequences in upland arctic tundra, North Slope, Alaska, US A Chronosequence Plant species Loon Lake Undisturbed tundra type description: non tussock sedge, dwarf shrub, moss tundra, with peaty non acidic soils Andromeda polifolia, Alnus viridis spp. crispa, Astragalus umbellatus, Aulacomnium palustre, A. turgidum, Bartramia ithyphylla, Betula nana, Bryum sp., Carex biglowii, C. misandra, Cassiope mertensiana, C. tetragona, Cetraria sp., Cladina sp., Dactylina arctica, Dicranum sp., Equisetum arvense, E. scirp oides, Eriophorum angustifolium, E. vaginatum, Flavocetraria sp., Hylocomium splendens, Ledum palustre, Pedicularis capitata, P. kanei, P. landsdorfii, Peltigera aphthosa, Pleurozium schreberi, Picea glauca Pogonatum urnigerum, Pohlia nutans, Polygonum bistorta, Polytrichum commune, P. strictum, Ptilium crista castrensis, Potentilla stipularis, Rhododendron lapponicum, Rhytidium rugosum, Salix alaxensis, S. glauca, S. phlebophyla, S. reticulata, Saxifraga oppositifolia, Senecio congestus Sphagnum fuscum, S. angustifolium, S. warnstorfii., Stellaria longipes, Tomentypnum nitens, Vaccinium uliginosum, V. vitis ideae Itkillik 1,2 and 3 Undisturbed tundra type description: non tussock sedge, dwarf shrub, moss tundra, with peaty non acidic soils Andromeda polifolia, Andrea rupestris, Arctostaphylos alpina, Asahinea chrysantha, Aulacomnium palustre, A. turgidum, Bartramia ithyphylla, Betula nana, Bryum sp., Calamagrostis sp., Carex biglowii, C. misandra, C. scirpoides, Cassiope mertensiana, C. tetragona, Cet raria sp., Cladina sp., Dactylina arctica, Dicranum sp., Dicranella schreberiana, Drepanocladus sp., Dryas integrifolia, D. octopetala, Equisetum arvense, E. scirpoides, Eriophorum angustifolium, E. vaginatum, Flavocetraria sp., Hylocomium splendens, Ledum palustre, Pedicularis capitata, P. kanei, P. landsdorfii, Peltigera aphthosa, Poa sp., Pleurozium schreberi, Pogonatum urnigerum, Pohlia nutans, Polygonum bistorta, Polytrichum commune, P. strictum, Potentilla stipularis, Rhododendron lapponicum, Rhytidiu m rugosum, Salix alaxensis, S. chamissonis, S. phlebophyla, S. reticulata, Saxifraga oppositifolia, Sphagnum angustifolium, S. warnstorfii., Stellaria longipes, Tomentypnum nitens, Vaccinium uliginosum, V. vitis ideae

PAGE 42

42 Table 2 2. C ontinued Chronose quence Plant species NE 14 Undisturbed tundra type description: non tussock sedge, dwarf shrub, moss tundra, with peaty non acidic soils Andromeda polifolia, Andrea rupestris, Arctostaphylos alpina, Asahinea chrysantha, Aulacomnium palustre, A. turgidum, Bartramia ithyphylla, Betula nana, Bryum sp., Carex biglowii, C. misandra, C. scirpoides, Cassiope mertensiana, C. tetragona, Cetraria sp., Cinclidium arcticum, Cladina sp., Dactylina arctica, Dicranum sp., Dicranella schreberiana, Drepanocladus sp., Dryas integrifolia, D. octopetala, Equisetum arvense, E. scirpoides, Empetrum nigrum, Eriophorum angustifolium, E. vaginatum, Flavocetraria sp., Hylocomium splendens, Ledum palustre, Pedicularis capitata, P. kanei, P. landsdorfii, Pleurozium schreberi Pogonatum urnigerum, Pohlia nutans, Polygonum bistorta, Polytrichum commune, P. strictum, Ptilium crista castrensis, Potentilla stipularis, Rhododendron lapponicum, Rhytidium rugosum, Salix alaxensis, S. chamissonis, S. hastata, S. glauca, S. phlebophyla S. reticulata, Saxifraga oppositifolia, Sphagnum fuscum, S. angustifolium, S. warnstorfii., Stellaria longipes, Thalictrum alpinum, Tomentypnum nitens, Thuidium sp., Vaccinium uliginosum, V. vitis ideae I minus 1 Undisturbed tundra type description: acidic, tussock sedge, dwarf shrub, moss tundra on top of wet sedge peat Andromeda polifolia Arctagrostis latifolia, Arctostaphylos alpina, Aulacomnium turgidum, Betula nana, Boykinia richardsonii, Brachythecium turgidum, Carex biglowii, C. misandra, C. scirpoides, Eriophorum vaginatum, Ledum decumbens, Ceratodon purpueus, Dicranum sp., Drepanocladus sp., Dryas integrifolia, Epilobium angustifolium, Equisetum arvense, E. scirpoides, Eriophorum angustifolium, E. vaginatum, Flavocetraria sp., Gentiana glauc a, Geum glaciale, Hedysarum alpinum, Hookeria sp., Hylocomium splendens, Ledum palustre, Mnium thomsonii, Oxytropis maydelliana Pedicularis capitata, P. kanei, Pedicularis oederi, Poa sp., Pleurozium schreberi, Pogonatum urnigerum, Pohlia nutans, Polygonum bistorta, Polytrichum commune, P. juniperinum, P. strictum, Polygonum viviparum, Pyrola grandiflora, Racomitrium canescens, R. lanuginosum, Rhytidium rugosum, Rubus chamaemorus Salix alaxensis, S. chamissonis, S. glauca, S. phlebophyla, S. pulchra S. reti culata, Saussurea angustifolia, Saxifraga nelsoniana, Sphagnum fuscum, S. angustifolium, S. magellanicum, S. warnstorfii., Stellaria longipes, Tomentypnum nitens, Vaccinium uliginosum, V. vitis ideae

PAGE 43

43 Table 2 3. Age of plant succesional processes s tarted by RTS disturbance in four chronosequences in upland arctic tundra, North Slope, Alaska, USA. P. glauca rings were count in this site instead of Salix sp. rings Chronosequence Site Mean age of shrubs (years) n SE Oldest shrub age (years) 14 +/ 14 C years of growth 3 n SE Loon Lake 1 1.67 4 0.49 2 NA NA NA NA NA 2 11.22 5 7.38 22 NA NA NA NA NA 3 29.89 5 3.43 33 4 41.35 5 16.82 64 5* 220.58 3 64.52 266 Itkillik 1 1 6.75 6 1.03 8 NA NA NA NA NA Itkillik 2 1 6.32 10 1.36 8 NA NA NA NA NA Itkillik 3 1 34.36 7 12.21 54 NE 14 1 4.75 2 0.89 5 NA NA NA NA NA 2 25.24 19 7.24 42 121.3 2.3 18.32 2 0.38 3 28.94 24 7.25 46 131.3 2.3 22.71 2 0.54 I minus 1 1 5.79 7 1.45 8 NA NA NA NA NA 2 10.81 28 2.91 16 NA NA NA NA NA 3 19.35 7 4.22 25 4 28.76 5 7.67 37 98.4 2.2 54.14 1 5 NA NA NA NA 93.9 2.2 53.46 1 6 NA NA NA NA 46 2.2 324 2 112 7 NA NA NA NA 34.4 2.3 451.5 2 2.6

PAGE 44

44 Table 2 4. Surface soil characteristics and tall shrub presence in four RTS chronosequences across upland arctic tundra, North Slope, AK, USA. *F rom the top 10cm of the mineral soil. All surface soil properties and % tall shrub presence were analyzed usin g a mix linear ANOVA (JMP, 8.0) with site (chronosequence), transect location (inside or outside RTS) and soil type (organic or mineral soil) as fixed treatment effects NE 14 Itkillik I minus 1 Loon Lake Layer Characteristic compared df F ratio P value df F ratio P value df F ratio P value df F ratio P value Vegetation % Tall shrub presence 4 12.58 <0.0001 5 14.5 <0.0001 8 15.02 <0.0001 5 12.71 <0.0001 Organic Organic layer depth (cm) 4 46.32 <0.0001 5 27.91 <0.0001 8 15.64 <0.0001 5 5.09 0.03 Organic soil bulk density (g per cm3) 4 0.56 0.5933 5 0.81 0.5584 8 2.8 0.0801 5 0.45 0.7299 Organic soil %C 4 0.06 0.9384 5 14.37 0.0018 8 3.49 0.0426 5 2.86 0.1436 Organic soil %N 4 0.76 0.5046 5 15.05 0.0015 8 1.38 0.3181 5 3.61 0.1006 Organic C:N ratio 4 0.32 0.8125 5 4.84 0.0312 8 4.29 0.0234 5 3.64 0.0944 Organic pools Organic soil C pool (g per m2) 4 32.53 <0.0001 5 25.74 <0.0001 8 9.03 <0.0001 5 3.84 0.004 Organic soil N pool (g per m2) 4 28.71 <0.0001 5 23.99 <0.0001 8 9.68 <0.0001 5 1.72 0.016 Mineral Mineral soil bulk density (g per cm3) 4 11.41 0.0776 5 0.06 0.9921 8 12.74 0.0032 5 2.37 0.183 Mineral soil %C 4 0.001 0.983 5 0.13 0.9588 8 21.02 0.0008 5 10.59 0.0108 Mineral soil %N 4 0.2 0.6999 5 0.15 0.949 8 3.64 0.0684 5 28.54 0.0011 Mineral C:N ratio 4 12.24 0.0729 5 22.76 0.1578 8 9.62 0.0068 5 3.2 0.1137 Mineral pools Mineral soil C pool (g per m2)* 4 1.15 0.36 5 2.52 0.056 8 4.45 0.002 5 0.24 0.92 Mineral soil N pool (g per m2)* 4 0.82 0.5 5 2.57 0.052 8 1.24 0.314 5 0.13 0.98

PAGE 45

45 Table 2 5. Fitted models used to predict carbon re accumulation rates in four RTS chronosequences in upland arctic tundra, AK USA Time scale was projected up to 300 years. NE 14 and Itkillik do not project beyond 50 years y=C po ol Site X Slope Intercept n r 2 F ratio p value NE 14 ln(time) 1485.0 513.3 3 0.6 1.516 0.4342 I minus 1 ln(time) 1658.3 1324.4 7 0.93 65.032 0.0005 Loon Lake (ln) ln(time) 756.6 1398.8 5 0.51 3.0928 0.1769 Itkillik ln(time) 1955.9 3297.3 3 0.89 7.843 0.2183 Loon lake (linear) Time 17.4 171.8 5 0.98 144.903 0.0012

PAGE 46

46 CHAPTER 3 RESULTS Study Sites Aging Ring counts of systematic surveyed adult shrubs rooted in mineral soil (Table 2 3; Appendix A ), allowed us to estimate the age of the plant succesional processes initiated by RTS disturbance. Shrub aging and radiocarbon dating of moss macrofossils at the b ase of the organic layer produced conflicting results. In revegetated lobes of NE 14, maximum shrub ages were ~24 years older and mean shrub ages were 7 years older than the radiocarbon ages of mosses (Table 2 3). In one revegetated lobe of I minus, the p attern was reversed: radiocarbon age of mosses was 17 years older than the oldest shrub and 25 years older than the mean shrub age (Table 2 3). Because of the larger sample size for shrubs, we chose to use the oldest shrub age as the time at which successi on was initiated by RTS disturbance. For the older sites in the I Minus 1 sequence where there were no shrubs, we used the radiocarbon dates with the caveat that their age may be over or underestimated by two plus decades relative to the younger sites Af ter grouping all the RTSs ages per chronosequence, and arranging them from younger to older, three time categories were created: 1) recently formed RTS (i.e. younger, well defined formations not older than ten years since disturbance), 2) revegetated RTS a fter disturbance (i.e. established formations in between ten and sixty years) and 3) stabilized RTS (i.e. old scars of hundreds of years old, where the headwall slopes are not pronounced and the active faces are stabilized with organic and mineral material covering and insulating the once exposed permafrost).

PAGE 47

47 Surface Soils Characterization Most of the pre disturbance surface soil profiles shown a well developed top 15cm of organic l ayer (Tables 3 1, 3 2, 3 3, and 3 4). Loon Lake (Table 3 2) and Itkillik 3 ( Table 3 3 ) had the least developed organic layer in depth, condition that becomes particularly intriguing when the soil profile is analyzed in terms of number of samples detected at each soil increment established: Loon Lake stands as the site with the sha llower organic layer depth when compare with the rest of the chronosequences, exhibiting a well represented t o p 5cm of organic layer (Table 3 2 ). The top 10cm of mineral layer was detected and well represented for all the pre disturbance soil profiles with the ex ception of I minus 1 (Table 3 4 ). The mineral layer was present in this site, but the particular feature of the type of tundra that constitutes this area of being on top of wet sedge peat, made the mineral layer harder to reach. For NE 14 and I minu s 1, sites where multiple control transects were set in the undisturbed tundra, the results reported for the pre disturbance conditions, and used for comparisons with the RTS post disturbance conditions, are averages of each set of surface soil characteris tics measured at each control transect. Recently formed RTSs diminished the organic matter from the pre disturbance conditions in one of two ways: reduction in depth (i.e. partial or complete loss of organic matter) and/or compaction of the organic layers (i.e increased bulk density when comparing an organic layer profile from the undisturbed tundra with the one sampled inside the recently formed RTS). In the case of the recently formed RTS at NE 14, there was a partial loss of organic matter, which reduced the soil increments detected in the pre disturban ce organic soil profile (Table 3 1 ), and a considerably increase in the bulk density ( D b ) of the remaining organic matter inside the feature: the top 15cm once well

PAGE 48

48 represented and developed in the pre dist urbance organic soil profile was partially reduced, and its D b increased almost twice when compared with the pre disturbance value. For Loon Lake, the recently formed RTS had a drastic effect on the organic matter, reducing the top 5cm from the pre disturb ance organic soil pr ofile almost completely (Table 3 2 ). This abrupt reduction in organic matter, and its extremely slow re accumulation rate through time, was characteristic of the Loon Lake chronosequence. Itkillik 1 stands as the site where the disturba nce to the surface soil organic matter was catastrophic. The recently formed RTS that affected this area completely reduced the organic soil profile characterized for the undisturbed tundra: the well represented and defined top 15cm of organic matter was c ompletely loss once the RTS impacted the landscape (Table 3 3 ), exposing the mineral layer in large extension inside the feature. In a very contrasting way, the recently form RTS in Itkillik 2 just partially reduced the top 15cm of organic layer characteri zed for the undisturbed tundra, and no effect was evident on the D b of the remaining organic matter inside the feature with compare with the pre disturbance values (Table 3 3 ). I minus 1 recently formed RTS had an interesting effect on the landscape: the o rganic layer was certainly reduced in depth, and the remaining organic matter had an increase on its D b (Table 3 4 ). The reduction in organic layer depth exposed the mineral layer, once completely covered by the characteristic wet sedge peat of the area. Recently form RTSs had a site specific effect on the surface soi l mineral lay er (Table 2 4, 3 1, 3 2, 3 3, 3 4 ). Overall, an increased exposure of the top 10cm of the mineral layer was detected in all sites, but the changes in the mineral soil characterist ics thru time could be attributed to the possible mixing with organic material during initial

PAGE 49

49 stages of the disturbance, which could explain the reduction in D b with an increased concentration of C and N observed in NE 14 (Table 3 1) and Loon Lake (Table 3 2 ), or by a complete incorporation to surface soil profile once the RTS affected the area, a s observed in I minus 1 (Table 3 4 ). Summarizing, recently formed RTS in all chronosequences exhibited a significant reduction in the organic layer depth when comp ared to the adjacent, undisturbed tundra pr e disturbance conditions ( Figure 3 1 Figure C 1 Table 2 4 ). In these young and active features, it is common to find large areas of expose mineral soil and/or compaction of the organic layer as well as defined e rosional and depositional zones. Revegetated RTS after disturbance showed a gradual re accumulation and development of organic soil on top of the newly exposed mineral layers ( Figure 3 1 Figure C 1 Table 2 4 ), which is a typical trend for soil developmen t in arctic tundra ecosystems (Harper and Kershaw, 1997). The organic layer depth re accumulation patterns, when analyze for each chronosequence in particular, displayed differences in the rates of organic matter re accumulation comparable to the ones dete cted on the undisturbed tundra. NE 14 revegetated RTSs re accumulated an organic layer depth close to 84 and 86% of the initi al pre disturbance value ( Figure C 1 ), respectively, in less than 60 years after disturbance. In contrast, the comparable revegetat ed sites in age at Loon Lake re accumulated an organic layer depth close to 25 and 51% of the initia l pre disturbance value (Fig ure C 1 ) after 33 and 64 years, respectively, after disturbance. After 54 years since disturbance, the revegetated site at Itkil lik 3 re accumulated an organic layer depth value that is 25% above the one reported fo r t he undisturbed tundra ( Figure C 1 ), but no significant difference was detected between

PAGE 50

50 these two organic layer depths ( Figure C 1 ). Similarly to Loon Lake, I minus 1 revegetated RTSs had a slow organic layer depth re accumulation, reaching a peak around 85% of the initial pre disturbance value after 55 years since disturbanc e ( Figure C 1 ). Although, this process had a notorious increase on its rate after 16 years since disturbance, reaching a plateau in between 55 and 324 years since disturban ce ( Figure 3 1 ). A possible driver for this trend exhibited by all the revegetated RTSs across the four chronosequences, could be the presence of tall and highly productive deciduo us shrub populations dominating the canopy of those areas and acting as a considerably increase in C inputs (i.e. leaf litter accumulation contributing to a thicker surface organic layer) when compared with the graminoid dominated undisturbed tundra the sy stem. Stabilized RTS sites found in Loon Lake and I minus 1, when compared to the pre disturbance values of organic layer depth, presented similar levels implying that the re accumulation process that took place on each of those chronosequences lasted cent uries before the initial organic la yer depths were achieved ( Figure 3 1 ). For Loon Lake, 87% of the pre disturbance depth was attained after 266 year since disturbance ( Figure 9). For I minus 1, 1% above the initial value of organic layer depth was re accu mulated after 451 years since disturbance ( Figure C 1 ). At the pool level, organic C and N and total C and N pools had significant differences between sites on eac h chronosequence ( Figure 3 2, 3 3, 3 4, 3 5, 3 6, 3 7, 3 8, 3 9, Table 2 4 ). For the mineral soil pools, significant differences were detected only am ong the I minus 1 sites (Table 2 4 ). The extent of the initial disturbance to the surface soil organic layer and its re accumulation rates through time, the pre disturbance surface soil characteristi cs such as organic layer depth, bulk density, %C

PAGE 51

51 and %N (Table 3 1, 3 2, 3 3 and 3 4 ), and the nature of C and N inputs once a plant succesional process was established due to RTS disturbance, could determine the C and N pool dynamics observed at each of the four chronosequences. Surface soil C and N pools dynamics through time at NE 14 were driven by mass loss/gain, changes in the nature of post disturbance C and N inputs, and changes in the concentration of those nutrients in the soil as follows: 1) inc reased organic matter D b due to compaction, and consequent reduction in organic layer depth, which resulted in a significant reduction in bot h C and N organic pools ( Figure 3 2A and 3 3 A) as direct consequence of mass erosion caused by a recently formed RT S, 2) increased C and N concentration in the top 10cm of the mineral layer during revegetation processes s tarted post disturbance (Table 3 1 ), and 3) organic layer dept re accumulation that spanned around 40 years before reaching values similar to the ones detected at the pr e disturbance conditions ( Figure 3 1 A). These three mechanisms also directed the surface soil C and N dynamics through time at Loon Lake. But, when compared to NE 14, the enormous difference in organic layer depth re accumulation rate that spanned 266 years without reaching similar values as the ones detected at the pr e disturbance conditions ( Figure 3 1 B), stands as a factor that could reflect key differences in the nature of C and N inputs established for those two RTS chronosequences The Itkillik recently formed RTSs total C and N pool dynamics were driven by the mineral layer contributions after disturbance in very contrasting ways. A complete organic matter mass loss, that reflected a total reduction of organic C and N pools, was t he case in Itkillik 1 (Table 3 3, Figure 3 1D, 3 2 D, 3 3D 3 6, C 1 ). An increased D b and

PAGE 52

52 a slight increased concentration of C in the top 10cm of mineral layer after disturbance was observed after the RTS affected the landscape (Table 3 3 ). N concentratio n in the top 10cm if mineral layer exhibited a reduction, that could be associated to catastrophic organic matter mass loss which exposed this layer to the top of the soil profile inside the feature, higher temperatures and possibly increased N consumption by plant colonizers and microbe communities. The slight C concentration increase could be associated to possible compaction of this layer with deeper and C richer layers in the soil profile. The fact that D b increased rather than decreased, could be an in dicative of compaction with deeper layers rather than mixing with organic matter removed after disturbance. In the case of Itkillik 2, a reduction in organic layer depth was detected after disturbance, but not to the extent observed at Itkillik 1 (Table 3 3 Figure 3 1D, 3 2D, 3 3D, 3 7 ) The remaining organic layer inside the feature still had a contribution to the total C and N pools once the area was affected by RTS keeping an organic C and N contribution to the total pool size, but was the notorious incr ease in C and N concentration at top 10cm of mineral layer which increased the total C and N pools when compared with the pre disturbance conditions ( Figure 3 7 ). This increased in concentration could be associated to a possible mixing with the disturbed o rganic matter once this material was eroded by the RTS. Looking at the D b pre and post disturbance, a slight reduction was observed on the top 10cm of mineral layer inside the feature. Total surface soil C and N pools dynamic at the Itkillik 3 revegetated RTS were directed by the increase D b of the re accumulated organic matter after disturbance rather than increase in the concentration of these elements in the organic layer (Table 3 3, Fig ure 3 8 ). Similarly to NE 14, the re accumulation rate spanned arou nd 54 years

PAGE 53

53 before reaching similar values that the ones observed at the undisturbed tundra conditions. The top 10cm of mineral layer reminded similar in terms of C and N concentrations and D b when compared to the pre distu rbance condition values (Table 3 3 ). I minus 1, site where we detected the most components for any chronosequence, shown surface soil total C and N pool dynamics that were driven by: 1) the organic matter re accumulation rate that spanned around 451 year before reaching similar values to the one observed at the undisturbed tundra, 2) changes in D b of the expose mineral layer after disturbance, and 3) changes in C and N concentration in the total mineral soil layer. As the main effect of the recently formed RTS, a well developed 20cm of min eral layer was exposed, which was not enough contribution to keep the total C and N pools to similar levels as the ones observed at the undisturbed tundra ( Figure 3 9 ). The net effect of the organic mass loss was the key factor in the significant reduction of the total pools aft er a recently RTS formed ( Figure 3 1C, 3 2 C 3 3C ). After 16 years of revegetation started by RTS, the organic matter re accumulated exhibited a notorious increased D b and N concentration. These, in addition to the contributions of the mineral layer, are the factors that boosted the total C and N pools at this poin t in the chronosequence ( Figure 3 9 ). After 25 years of revegetation, the exposed 20cm of mineral layer shown signs of well mixing with the organic matter re accumulated: 1 ) mineral D b was notoriously reduced when compared with the value inside the recently formed RTS and the 16 y ear old revegetated RTS (Table 3 4 ), and 2) the observed concentration of C in the mineral soil was the highest for a ny mineral soil sampled (Table 3 4 ). Although the N concentration in the mineral soil was the highest for any

PAGE 54

54 mineral soil sampled at any site in the chronosequence, the reduction in D b resulted on a reduction in the total N pool when compared with the younger revegetated RTS ( Figure 3 9 ). The differences detected at the two 55 year old revegetated RTS are driven by decreased D b in the mineral layer and increased concentra tion of organic C and N (Table 3 4 ), being the site with the higher %C and %N the one where tall deciduo us shrubs we re detected ( Figure 3 17 ). This could be potential evidence of the effect of shifts in nature of C and N inputs in the C and N dynamics at this site. Stabilized RTS shown a reduced organic C and N concentration when compared with yo unger revegetated sites (Table 3 4 ), and after 451 years of organic matter re accumulation the organic layer depth was significantly recovered to the undisturbed tundra values. The driver of the difference between pre disturbance and stabilized RTS total C and N pools are the red uced organic %C and %N detected at the stabilized sites, and the contribution in C and N by the mineral layer exposed by t he recently formed RTSs ( Figure 3 9 ). The mineral layer at the oldest site could potentially being mixing with the re accumulated orga nic matter due to its reduction in D b when compared to some of the revegetated sites, and its incr ease in C concentration (Table 3 4 ). Rock Volume Correction for Mineral Soil C and N Pool Soil volume displaced by large rocks (i.e. >5cm in diameter) acc ount ed for at least 30% (Table 3 5 ) of the total pit volume for all of the three chronosequences sampled. The overall proportion of large rock volume was 35% across sites sampled at NE 14, Itkillik and I minus 1, confirming the necessity of this correction fac tor for accurate pool size estimation in the mineral surface soils

PAGE 55

55 Organic Soil C/N, Top Ten Mineral Soil C/N, and Organic L ayer Organic layer depth and surface soil org anic C and N ( Figure C 1, C 2, C 4 ) had a considerable reduction in percent remaining f rom the initial pre disturbance values across all the recently form RTSs. Surface soil mine ral N ( Figure C 5 ), with the exception of Loon Lake, had a gradual increase from the initial pre disturbance values. This is particularly evident for I minus 1, where the whole mineral layer was exposed once the ar ea was disturbed by RTS (Table 3 4 ). Initial RTS disturbance in all the study sites had a major effect on C pool in both organic and mineral surface soils (Table 3 6). The percent change in organic C po ol after initial RTS disturbance across sites, showed the catastrophic impact of RTS on the organic horizon when compared with pre di s turbance condition values. The top 10cm of mineral soil C pool had a more intricate behavior, which depended on specific p re disturbance conditions of each of the study sites. Overall, a reduction on the C pool was detected, but in the cases of the younger Itkillik RTSs and the younger Loon Lake site, an increased C pool on the top 10cm of mineral layer was the case due to th e exposure of deeper mineral layers as consequence of thermokarst disturbance (Table 3 6) In terms of N pool change due to initial RTS activity in all of the study sites, a similar trend as the C pool changes was detected, displaying a relevant reduction in the surface organic soil and an overall reduction on the top 10cm of mineral layer (Table 3 7). The effects of RTS disturbance on the top 10cm of mineral soil are clearly obse rved when values of D b %C and %N were compared with the pre disturbance condition values (Table 3 8) The mineral layer C and N pool changes, at each st udy site, can be associated to : 1) increments on the bulk density of t he top 10cm due to the physical erosion of organic soil and compaction of material coming from the undistu rbed

PAGE 56

56 tundra inside the slump floor or 2) well mixing of the eroded organic layer with exposed surface mineral soil, which can be observed when %C values increase from pre disturbance conditions (Table 3 8) Surface Soil Organic C Pool Re A ccumulation Recov ery curves showed differences in the C pool re accumulation rates thru time, placing NE 14 as the chronosequence with the fastest re accumulation of soil organic layer C pool 50 years after initial disturbance ( Figure 3 10). Itkillik recovery curve follows a very similar trend as NE 14, but with a lower re accumulation rate. By the other extreme, Loon Lake appears to be the chronosequence with the lowest re accumulation rate, expressed by both logarithmic and lineal models ( Figure 3 10). I minus 1 chronoseq uence is the study site with the most temporal components (i.e. lobes), which accounts for the well established recovery curve from 10 to 300 years, corresponding to the most complete reconstruction in time of soil organic layer C pool dynamic after RTS di sturbance ( Figure 3 10). Time projection for the recovery curves extends until 300 years after initial RTS disturbance, but when the soil organic layer C pool re accumulation rate for all the four RTS chronosequences were compared in a 10 40 years period o f recovery, Itkillik appears as the site with the highest re accumulation (78 g C*m 2 ) followed by I minus 1 (66.7 g C *m 2 ), NE 14 (59.8 g C *m 2 ) and Loon Lake (30.7 g C m 2 using the logarithmic model, and 17.4 g C m 2 using the lineal model) These tendencies, and the relevant differences detected between sites, in C pool re accumulation could be determine not only by RTS activity itself and the initial magnitude of the disturbance, but also by particular characteristics of each site like: 1) v egetation type associated to the pre disturbance conditions, 2) microclimate

PAGE 57

57 associated to each of the RTS features, 3) relieve (i.e. aspect and slope steepness), 4) differences in the percent of tall shrubs present inside each RTS feature, and 5) parental material. Percent Tall Shrub Estimation Populations of tall deciduous shrubs increased in abundance following disturbance for all revegetated sites when compared against recently formed and stabilized sites The presence of ta ll deciduous shrubs across s ites in any particular chronosequence had a ver y distinct distribution ( Figure 3 11 ): low percents on both recently formed and stabilized RTSs with presence peaks on revegetated RTSs (sites in between 10 and 60 years old). Some of the tall deciduous shrub species detected inside the features were encountered at the undisturbed tundra (i.e. Salix alaxensis, S pulchra, S. glauca, S. hastata and Betula nana ) but any of them were taller than 1m. S. alexensis was only found within RTF features or in other types of disturbances, such as river banks or road cuts. The increased in abundance at revegetated RTSs, could be consider as a shift in the nature of C and N inputs to the system, once dominated by less productive lower litter quality graminoid vegetation.

PAGE 58

58 Table 3 1. NE 14 RTS chronosequence sh allow soil characterization. TK=thermokarst, Tr=transect, T0= control transect on the adjacent, undisturbed tundra, T1 =transect inside the feature, SE= standard error TK Tr Soil Type Section depth (cm) n D b (g/cm3) SE pH SE %C SE %N SE NE 14TK1 T0 Org 0 5 6 0.081 0.005 4.71 0.33 37.387 2.926 1.434 0.235 Org 0 15 4 0.165 0.023 5.12 0.324 40.952 1.461 1.896 0.354 Org 15 25 1 0.258 6.17 22.889 1.525 Org 25 35 1 0.304 6.42 31.898 2.035 Min 0 10 4 1.141 0.344 4.92 0.435 5.249 2.615 0.275 0.107 NE 14TK1 T1 Org 0 5 2 0.234 0.011 7.61 0.41 30.316 5.781 1.717 0.595 Org 0 15 2 0.27 0.012 6.79 0.562 34.523 4.579 2.253 0.02 Min 0 10 6 0.791 0.145 7.7 0.12 3.326 1.554 0.213 0.091 NE 14TK2 T1 Org 0 5 5 0.193 0.107 5.45 0.348 36.121 2.087 1.709 0.18 Org 0 15 4 0.239 0.047 6.18 0.277 31.35 4.623 1.859 0.326 Org 15 25 3 0.305 0.018 6.56 0.057 25.9 3.38 1.74 0.167 Min 0 10 6 0.706 0.179 6.23 0.221 6.061 2.006 0.362 0.112 NE 14TK3 T1 Org 0 5 6 0.225 0.084 6.04 0.17 32.359 3.113 1.69 0.12 Org 0 15 2 0.136 0.066 6.06 0.456 28.192 7.501 1.721 0.195 M in 0 10 4 0.884 0.175 6.79 0.076 6.224 2.086 0.408 0.145

PAGE 59

59 Table 3 2. Loon Lake RTS chronosequence sh allow soil characterization. TK=thermokarst, Tr=transect, T0= control transect on the ad jacent, undisturbed tundra, T1= transect inside t he feature, SE= standard error TK Tr Soil Type Section depth (cm) n D b (g/cm3) SE pH SE %C SE %N SE Loon Lake T0 Org 0 5 6 0.052 0.028 5.23 0.085 41.218 3.188 1.503 0.249 Org 0 15 1 0.153 5.54 38.66 1.801 Org 15 25 1 0.231 6.68 39.631 1.998 Min 0 10 4 0.969 0.079 5.55 0.18 2.004 0.022 0.153 0.001 Min 10 20 2 0.958 0.202 6.37 0.155 1.997 0.172 0.166 0.023 Loon Lake TK1 T1 Org 0 5 1 0.024 5.46 42.443 1.008 Org 0 15 1 0.026 4.93 40.47 1.35 Min 0 10 6 0.869 0.041 7.09 0.55 3.028 0.763 0.228 0.07 Min 10 20 1 0.643 7.71 4.501 0.303 Loon Lake TK2 T1 Min 0 10 6 0.809 0.086 7.36 0.535 2.908 0.799 0.186 0.07 Min 10 20 2 0.9 0.078 7.32 0.59 2.184 0.562 0.174 0.026 Loon Lake TK3 T1 Org 0 5 1 0.063 7.92 35.364 1.845 Min 0 10 6 0.785 0.02 8.24 0.104 3.198 1.232 0.207 0.103 Min 10 20 2 0.947 0.021 8.39 0.045 2.714 0.439 0.171 0.031 Loon Lake TK4 T1 Org 0 5 6 0.088 0.03 7.56 0.227 30.569 0.793 1.721 0.164 Min 0 10 1 0.802 7.49 4.726 0.411 Min 10 20 1 0.586 7.53 4.58 0.396

PAGE 60

60 Table 3 2. C ontinued TK Tr Soil Type Section depth (cm) n D b (g/cm3) SE pH SE %C SE %N SE Loon Lake TK5 T1 Org 0 5 6 0.032 0.007 5.51 0.45 44.599 0.68 1.753 0.353 Org 5 15 2 0.085 0.006 5.64 0.638 42.553 1.668 1.615 0.207 Org 15 25 2 0.21 0.027 6.69 0.075 34.673 2.183 1.604 0.238 Min 0 10 1 0.549 7.05 5.943 0.413

PAGE 61

61 Table 3 3 Itkillik RTS chronosequence sh allow soil characterization. TK=thermokarst, Tr=transect, T0= control transect on the adjacent, undisturbed tundra, T1 = transect inside th e feature, SE= standard error TK Tr Soil Type Section depth (cm) n D b (g/cm3) SE pH SE %C SE %N SE IKTK1 T0 Org 0 5 6 0.079 0.017 5.29 0.319 32.514 3.797 1.51 0.23 Org 5 15 5 0.164 0.026 6.19 0.191 29.597 2.553 2.11 0.229 Org 15 25 2 0.199 0.037 6.62 0.075 27.349 0.102 1.83 0.053 Org 25 35 2 0.298 0.067 6.57 0.035 25.187 1.866 1.84 0.055 Org 35 45 1 0.459 6.41 29.482 2.02 Min 0 10 4 0.831 0.068 6.17 0.182 1.958 0.627 0.15 0.056 T1 Min 0 10 6 1.021 0.049 6.2 0.184 2.29 0 0.786 0.1 3 0.051 IKTK2 T0 Org 0 5 6 0.082 0.002 6.77 0.054 31.334 2.24 1.87 0.083 Org 5 15 4 0.163 0.006 6.64 0.076 27.323 1.521 2.00 0.151 Min 0 10 6 1.193 0.032 6.3 0.075 2.386 0.651 0.19 0.045 Min 10 20 1 0.373 6.14 18.161 1.41 T1 Org 0 5 4 0.08 0.004 6.68 0.079 34.517 0.817 1.69 0.409 Org 5 15 4 0.156 0.009 6.62 0.039 30.628 1.566 1.94 0.093 Min 0 10 6 0.997 0.058 6.35 0.218 4.066 1.558 0.30 0.131

PAGE 62

62 T able 3 3. C ontinued TK Tr Soil Type Section depth (cm) n D b (g/cm3) SE pH SE %C SE %N SE IKTK3 T0 Org 0 5 3 0.075 0.011 5.69 0.601 36.549 4.122 1.5 9 0.108 Org 5 15 2 0.172 0.012 6.28 0.25 0 26.263 3.234 1.7 6 0.173 Min 0 10 6 0.713 0.166 5.66 0.198 9.153 2.595 0.6 3 0.176 T1 Org 0 5 4 0.156 0.086 7.05 0.157 32.478 2.897 1.36 0.189 Min 0 10 6 0.73 0.189 7.04 0.191 8.43 2.607 0.7 6 0.279

PAGE 63

63 Table 3 4. I minus 1 RTS chronosequence sh allow soil characterization. TK=thermokarst, Tr=transect, T0= control transect on the ad jacent, undisturbed tundra, T1= transect inside t he feature, SE= standard erro r TK Tr Soil Type Section depth (cm) n D b (g/cm3) SE pH SE %C SE %N SE Iminus1TK1 T0 Org 0 5 6 0.076 0. 019 4.66 0.173 36.498 1.647 1.553 0.257 Org 5 15 4 0.132 0.017 4.8 0.109 37.336 0.707 2.375 0.271 Org 15 25 2 0.202 0.007 4.87 0.04 0 35.155 1.229 2.524 0.056 Iminus1TK1 T1 Org 0 5 1 0.086 4.85 22.536 1.36 0 Org 5 15 1 0.194 4.44 33.026 2.21 0 Min 0 10 6 1.319 0.049 6.82 0.132 1.245 0.292 0.094 0.018 Min 10 20 2 1.33 0 0.048 6.86 0.035 1.147 0.102 0.09 0 0.005 Iminus1TK2 T1 Org 0 5 1 0.249 5.93 33.73 0 2.343 Org 5 15 1 0.38 0 6.93 29.147 1.911 Min 0 10 6 1.06 0 0.055 5.87 0.159 1.208 0.203 0.092 0.011 Min 10 20 3 1.116 0.094 5.84 0.38 0 0.567 0.072 0.059 0.003 Iminus1TK3 T1 Org 0 5 3 0.062 0.015 5.06 0.024 36.357 5.624 1.082 0.272 Org 5 15 3 0.15 0 0.017 5.2 0 0.305 36.857 3.284 1.425 0.203 Min 0 10 6 0.354 0.142 5.04 0.045 13.106 2.273 0.701 0.122 Min 10 20 1 0.346 5.4 0 14.91 0 0.555

PAGE 64

64 Table 3 4. C ontinued TK Tr Soil Type Section depth (cm) n D b (g/cm3) SE pH SE %C SE %N SE Iminus1TK4 T1 Org 0 5 6 0.064 0.009 5.5 0.352 31.123 1.697 1.618 0.121 Org 5 15 1 0.14 6.39 31.207 2.011 Min 0 10 6 0.656 0.026 6.67 0.186 4.543 1.994 0.367 0.156 Min 10 20 5 0.901 0.035 6.94 0.128 1.354 0.151 0.115 0.011 Iminus1TK5 T1 Org 0 5 6 0.071 0.006 5.59 0.402 27.225 1.381 1.381 0.188 Org 5 15 4 0.146 0.024 6.2 0.237 22.906 0.613 1.144 0.133 Min 0 10 6 0.488 0.068 6.55 0.162 7.442 1.211 0.459 0.073 Min 10 20 6 0.674 0.069 6.66 0.179 8.427 1.749 0.583 0.115 Iminus1TK6 T1 Org 0 5 6 0.08 0.008 6.9 0.253 24.558 0.746 1.43 0.143 Org 5 15 5 0.15 0.01 6.74 0.221 30.819 3.279 1.868 0.213 Min 0 10 4 0.436 0.127 6.29 0.109 0.993 0.059 0.076 0.007 Min 10 20 4 0.826 0.122 5.59 0.201 4.444 1.962 0.506 0.318 Iminus1TK7 T1 Org 0 5 6 0.077 0.004 6.15 0.221 29.17 1.931 1.848 0.169 Org 5 15 5 0.156 0.013 6.13 0.127 30.243 3.183 1.653 0.107 Min 0 10 6 0.797 0.024 5.7 0.206 11.446 0.661 0.406 0.197

PAGE 65

65 Table 3 5 Proportion of large rocks (>5cm in diameter) present on 50cm 3 soil pits dug on three RTS chronosequences around Toolik Field Station, AK USA These values were used to correct for mineral C and N pool sizes. Number s in parenthesis represent the standar d error Chronosequence Soil vol (ml)/Total Pit vol (ml) Itkillik series 0.48 0.27 Mean 0.39 (0.105) I minus 1 0.13 0.35 0.42 Mean 0.30 (0.088) NE 14 0.43 0.55 0.08 Mean 0.37 (0.138)

PAGE 66

66 Table 3 6. C pool change on the orga nic lay er, top 10cm of the mineral soil and total pool change due to initial thermokarst disturbance in four RTS chronosequences across arctic tundra, AK USA C pool (g*m 2) Site t0 t1 Change % change NE 14 site 1 Organic 6959 2840 4119 59 Mineral (0 10 cm) 5989 2631 3358 56 Total 12948 5471 7477 58 Itkillik 1 Organic 9388 0 9388 100 Mineral (0 10 cm) 1627 2338 +711 +43 Total 11015 2338 8677 79 Itkillik 2 Organic 5408 1540 3868 72 Mineral (0 10 cm) 2846 4054 +1207 +42 Total 8254 5594 2661 32 Loon Lake site 1 Organic 2694 260 2434 90 Mineral (0 10 cm) 1942 2631 +689 +35 Total 4636 2891 1745 38 I minus 1 site 1 Organic 8107 1027 7080 87 Mineral (0 10 cm) 8291 1642 6649 80 Total 16398 2669 13729 84 Average Organic 6511.2 1133 5378 82 Mineral (0 10 cm) 4139 2659 5003 68 Total 10650 3793 6858 64

PAGE 67

67 Table 3 7. N pool change on the organic later, top 10cm of the mineral soil and total pool change due to initial thermokarst disturbance in four RTS chronosequences across arctic tundra, AK USA N pool (g*m 2 ) Site Category t0 t1 pool change % change NE 14 site 1 Organic layer 356 162 194 54 Mineral layer (0 10 cm) 314 168 145 46 Total 670 330 339 51 Itkillik 1 Organic layer 633 0 633 100 Mineral layer (0 10 cm) 126 132 +6 +4 Total 759 132 627 83 Itkillik 2 Organic layer 378 92 286 76 Mineral layer (0 10 cm) 234 300 +66 +28 Total 612 392 220 36 Loon Lake site 1 Organic layer 130 8 122 94 Mineral layer (0 10 cm) 148 198 +50 +33 Total 278 206 72 26 I minus 1 site 1 Organic layer 518 67 451 87 Mineral layer (0 10 cm) 497 124 374 75 Total 1015 191 825 81 Average Organic 403 66 337 60 Mineral (0 10 cm) 264 184 259 61 Total 667 250 417 55

PAGE 68

68 Table 3 8 Bulk density, %C and %N change at the top 10cm of the mi neral soil due to initial thermokarst disturbance in four RTS chronosequences across arctic tundra, AK USA Bulk density (g/cm3) %C %N Site t0 t1 %c hange t0 t1 %c hange t0 t1 %c hange NE 14 site 1 1.141 (0.344) 0.791 (0.145) 35.0 5.249 (2.615) 3.326 (1.554) 36.2 0.275 (0.107) 0.213 (0.091) 22.54 Itkillik 1 0.831 (0.068) 1.021 (0.049) 19 1.958 (0.627) 2.29 (0.786) 16.95 0.152 (0.056) 0.129 (0.051) 15.13 Itkillik 2 1.193 (0.032) 0.997 (0.058) 0.196 2.386 (0.651) 4.066 (1.558) 70.41 0.196 (0.045) 0.301 (0.131) 53.57 Loon Lake site 1 0.969 (0.079) 0.869 (0.041) 10 2.004 (0.022) 3.028 (0.763) 51.097 0.153 (0.001) 0.228 (0.07) 49.019 I minus 1 site 1 0.944 (0.025) 1.319 (0.049) 37.5 8.783 (4.074) 1.245 (0.292) 85.82 0.527 (0.219) 0.094 (0.018) 82.16

PAGE 69

69 Figure 3 1. Organic layer dynamics at four RTS chronosequences in upland arctic tundra, AK, USA. (A) NE 14, (B) Loon Lake, (C) I minus 1 and (D) Itkillik. Grey bands represent the pre disturbance conditions assuming no abrupt alteration thru time. Red marker represents the undisturbed tundra value. Age of the undisturbed tundra was not determined. Error bars represent standard errors. Lower case letters indicate a significant difference between sites. X axis at Loon Lake and I minus 1 were log transformed to better display the changes in organic layer depth though time. Large standard errors at (B) are result of the small sample size for those sites

PAGE 70

70 Figure 3 2 Surface soil organic C pool re accumulation dynamic after disturbance in four RTS chronosequences in upland arctic tundra, AK USA (A) NE 14, (B) Loon Lake, (C) I minus 1 and (D) Itkillik. Grey bands represent the pre disturbance conditions assuming no ab rupt alteration thru time. Red marker represents the undisturbed tundra value. Age of the undisturbed tundra was not determined. Error bars represent standard errors. Lower case letters indicate a significant difference between sites. X axis at Loon Lake a nd I minus 1 were log transformed to better display the changes in organic layer depth though time. Large standard errors at (B) are result of the small sample size for those sites

PAGE 71

71 Figure 3 3 Surface soil organic N pool re accumulation dynamic after disturbance in four RTS chronosequences in upland arctic tundra, AK USA (A) NE 14, (B) Loon Lake, (C) I minus 1 and (D) Itkillik. Grey bands represent the pre disturbance conditions assuming no abrupt alteration thru time. Red marker represents the undi sturbed tundra value. Age of the undisturbed tundra was not determined. Error bars represent standard errors. Lower case letters indicate a significant difference between sites. X axis at Loon Lake and I minus 1 were log transformed to better display the c hanges in organic layer depth though time. Large standard errors at (B) are result of the small sample size for those sites

PAGE 72

72 Figure 3 4 NE 14 RTS total C and N pools re accumulation dynamic through time. These pools where calculated from the whole organic layer profile plus the top 10cm of mineral layer sampled at each site. Grey bands repre sent the pre disturbance conditions assuming no abrupt alteration thru time. Red marker represents the undisturbed tundra value. Age of the undisturbed tundra was not determined. Error bars represent standard errors. Lower case letters indicate a significa nt difference between sites a b a b a a a b a b

PAGE 73

73 Figure 3 5 Loon Lake RTS total C and N pools re accumulation dynamic through time. These pools where calculated from the whole organic layer profile plus the top 10cm of mineral layer sampled at each site. Grey bands r epresent the pre disturbance conditions assuming no abrupt alteration thru time. Red marker represents the undisturbed tundra value. Age of the undisturbed tundra was not determined Error bars represent standard errors. Lower case letters indicate a signi f icant difference between sites a b c b b ab c b b ab b a

PAGE 74

74 Figure 3 6 Itkillik 1 RTS total C and N pools re accumulation dynamic through time. These pools where calculated from the whole organic layer profile plus the top 10cm of mineral layer sampled at each site. Grey bands represent the pre disturbance conditions assuming no abrupt alteration thru time. Red marker represents the undisturbed tundra value. Age of the undisturbed tundra was not determined Error bars represent standard errors. Lower case letters indicate a signi ficant difference between sites

PAGE 75

75 Figure 3 7 Itkillik 2 RTS total C and N pools re accumulation dynamic through time. These pools where calculated from the whole organic layer profile plus the top 10cm of mineral layer sampled at each site. Grey bands represent the pre disturbance conditions assuming no abrupt alteration thru time. Red marker represents the undisturbed tundra value. Age of the undisturbed tundra was not determined Error bars represent standard errors. Lower case letters indicate a signi ficant difference between site s b a b a

PAGE 76

76 Figure 3 8 Itkillik 3 RTS total C and N pools re accumulation dynamic through time. These pools where calculated from the whole organic layer profile plus the top 10cm of mineral layer sampled at each site. Grey bands represent the pre disturbance conditions assuming no abrupt alteration thru time. Red marker represents the undisturbed tundra value. Age of the undisturbed tundra was not determined Error bars represent standard errors. Lower case letters indicate a signi ficant difference between sites b a b a

PAGE 77

77 Figure 3 9 I minus 1 RTS total C and N pools re accumulation dynamic through time. These pools where calculated from the whole organic layer profile plus the whole mineral layer sampled at each site. Grey bands represe nt the pre disturbance conditions assuming no abrupt alteration thru time. Red marker represents the undisturbed tundra value. Age of the undisturbed tundra was not determined Error bars represent standard errors. Lower case letters indicate a significant difference between sites a c ab ab b ab ab ab a d abc c bc ab ab abc

PAGE 78

78 Figure 3 10 Surface soil organic C pool re accumulation predicted by ln(time), after initial disturbance, in four RTS chronosequences in upland arctic tundra, AK USA

PAGE 79

79 Figure 3 11. Estimation of the presence of tall shrub populations (i.e. Salix alaxensis, S. glauca, S. pulchra, S, hastata and Betula nana higher than 1.3m ) inside the different lobes at each RTS chronosequence in upland arctic tundra, AK, USA. (A) = NE 14, (B) = Itkillik, (C ) = Loon Lake an d (D) = I minus 1 Error bars represent the standard error. Lower case letters indicate a significant difference between sites A B C D a b b a b b a a a b a a b b c a a

PAGE 80

80 CHAPTER 4 DISCUSSION NE 14 RTS chronosequence consisted of three distinct lobes: one rec ently formed RTS and two revegetat ed RTS after disturbance (Table 2 3) Surface soil characteristics, for both the organi c and mineral component (Table 3 1 ), revealed a loss in the organic depth belo w 15cm as a result of the impacted area by the recently formed TEF ( Figure 3 1 ); which also exposed big extensions of mineral soil and mobilized the organic material eroded towards the body of water present at the end of the slope (Lake NE 14). Mineral soi ls exposed in RTSs are likely to provide substantially different soil structure and belowground resources for plant recruitment and growth, but little is known about how soil characteristics are altered by RTS, and how they change over time after disturban ce. The re accumula tion of organic matter occurred in a period of 40 years after the effect of the recently formed RTS on the initial organic layer depth at the pre disturbance conditions ( Figure 3 10) This rapid re accumulation of organi c material was dr iven by the shifted vegetation inputs from the established post disturbance vegetation, where the abundance of highly productive, tall deciduous shrubs was significantly higher than the values observed for the recently form RTS or even when visually compar ed with the undisturbed tundra gramino id dominated vegetation ( Figure 3 11 ). The presence of tall deciduous shrubs showed a drastically change through time, incrementing more than 60% from the status m easured on the recently formed RTS to the oldest revege tate d RTS after disturbance ( Figure 3 11 ). For the total surface C and N pools, the impact of a recently fo rmed TEF caused a reduction on close to10Kg of C and 0.6Kg of N when compared with the pr e disturbance

PAGE 81

81 pool sizes ( Figure 3 4 ). Both C and N pool si zes were recovered and surpassed pre disturbance level at one instance, d uring the revegetation processes g enerated by RTS activity ( Figure 3 4 ). Loon Lake chronosequence consisted of five distinct lobes: one recently fo rmed RTS, three r evegetated RTS aft er disturbance one stabilized RTS The organic matter reduction that took place in this chronosequence was ver y dramatic, showing a significant reduction in the already shallow organic layer depth observed in the pre disturbance conditions. This mass loss was due to the erosion and mobilization of organic material once RTS affected the landscape (Table 3 2, Figure 3 1 B ). The partial re accumulati on of organic matter, when compared with the values of organic layer depth observed in the undisturbed tundra, s panned 266 years This extremely slow re accumulation rate could reflect particularities on the microclimate of each site conforming the chronosequence, an overall higher soil temperature when compared with the sites established in the vicinity of the Tool ik Field Station, or smaller inputs from the vegetation established post disturbance. Being one of the two chronosequences where stabilized RTS s were found, Loon Lake stands as an example of alternate states in terms of tall deciduous shrub abundance Reac hing a percent shrubbiness peak around 40 % during revegetated stages and exhibiting individuals taller than 1m, a re duction in percentage was dete cte d at the stable stage ( Figure 3 11 ). For the total surface C and N pools, the impact of a recently forme d TEF caused a reduction on close to15Kg of C and 0.7Kg of N when compared with the pr e disturbance pool sizes ( Figure 3 5 ). Both C and N pool sizes were significantly recovered during the revegetation process es started by RTS disturbance in the area ( Figu re 3 5 ), without

PAGE 82

82 reaching levels similar to the pool sizes measured on the pre disturbance conditions. The re accumulation of material exhibited a lineal progressi on after the recently formed RTS induce a significant mass loss Post disturbance surface so il total C and N pool sizes are not recovered even after 266 years of disturbance. The Itkillik chronosequence consisted of three distinct f eatures paired with their own adjacent, undisturbed tundra: two recently formed RTSs, very close in age, but showing a very contrasting development and impact on the la ndscape, and one revegetated RTS a fter disturbance (Table 2 3 ). The most drastic impact in terms of the changes detected on the surface soil organic matter depth was found i n the recently formed RTS at I tkillik 1 The complete organic soil profile was not detected inside the feature suggesting that the impact and mobilizat ion of materials inside this RTS were critically influenced by physical properties of the feature itself and/or its location on the la ndscape. This behavior is contrasting with the surface soil characteristics found on Itkillik 2 The overall organic matter depth suffered a reduction as a result of the impact, as expected ( Figure 3 1 D ), but the re distribution of the eroded material from the undisturbed tundra was so that the organic fraction of the soil profile was still representa tive inside the feature (Table 3 3 ). This pattern could be playing an important role as the driving force behind the increment of the surface soil total C and N pools when compared with th e pre disturbance values ( Figure 3 7 ). The presence of tall deciduous shrubs for these two sites also contributed to differentiate the magnitude of the disturbance experience d at each TEF. No tall decidous shrubs wer e found ins ide Itkillik 1, fact that can be associated with the complete reduction on organic matter in thi s site. Contrastingly, Itkillik 2 exhibited a

PAGE 83

83 presence of tall deciduous shrubs closer to 3 0%, including in dividuals talle r than 1m ( Figure 3 11 ). At Itkillik 3 revegetated RTS after disturbance, an increment in organic matter was observed when compared with the pos t disturb ance conditions ( Figure 3 1D, Table 3 3 ), and similarly to the revegetated RTSs at NE 14 this increment could be due to the shift in vegeta tion co mposition established during post d isturbance conditions (i.e. tall deciduous shrubs ) and its inputs towards the accumulation of organic material In a similar fashion than the revegetated sites at NE 14, the organic matter re accumulation rate also showed a fast rate when compared with revegetated stages at Loon Lake and I minus 1. Also, the abundance of tall deciduous shrubs were very comparable between Itkillik 3 and the oldest revegetat ed RTS at NE14, both sites showing more than 60% of presence of this functional gro up inside the feature ( Figure 3 11 ). I minus 1 chronosequence contains the older sites determine by the radiocarbon dating of the moss cores sampled on the area. By compari son with the rest of the chronosequences studied, I minus 1 is the only representative of acidic, tussock sedge, dwarf shrub, moss tundra type on top of wet sedge peat. The other chronosequences are group into the non acidic, non tussock sedge, dwarf shrub moss tundra type Seven lobes constitute this chronosequence, five of them very conspicuous, and two of them completely blended with the undisturbed, adjacent tundra: one recently formed RTS four revegetated RTSs with contrasting ag es, and two old and s tabilized RTS. The pattern of re accumulation of organic matter depth through time is very distinct in this chronosequence, having its higher re accumulation after 25 years in comparison

PAGE 84

84 to pre disturbance conditions; after this, the increment is gradual a nd more paused until reaching similar levels with the organic matter depth and organic C and N pools of the undisturbed, adj acent tundra. The presence of tall deciduous shrubs reached its peak after 25 years, starting from a l ack of tall shrubs right after the recently form RTS impacted the la ndscape ( Figure 3 11 ). Considering the type of tundra that defines the pre disturbance conditions, the reason behind the lack of mineral soil, for the control soil profile (Table 3 4 ), could be attributed to the actua l depth of this layer (i.e. the layer must exist, but underneath the wet sedge peat where the acidic tundra lies on). The organic sec tion below 15cm, for all the RTSs constituting this chronosequence, is not present, but the inclusion of the top 10cm of mi neral layer can be the result of the disturbance and alteration of the permafro st cause by RTS disturbance in the area (Table 3 4 ). The two stabilized RTS are more similar, in terms of plant composition, to the pre disturbance condition than to any of the other five features included in the chronosequence. The lack of tall shrubs after more than 400 years after disturbance could be proof of permafrost stabilization and no additional disturbance during the re vegetation processes ins ide this stable scars ( F igure 3 11 ). Forbes et al. (2001) have stable states following perturbation can exert a long term influence on a range of biotic and abiotic processes in the disturba nce and on the surrounding terrain. It is clear that understanding the effects of TEF on arctic landscapes is a growing field and a challenging ecosystem ecology question. Improved prediction of a dynamic arctic

PAGE 85

85 landscape cannot be achieved without signif icant advances in observations and process based studies. Burn and Friele (1989) documented the succesional process over a 43 years period after RTS disturbance in the Canadian boreal ecotone noticing distinct vegetation units established inside the RTS w hen compared with the surrounding undisturbed forest during their survey in 1987, but detecting that all the important constituents of the undisturbed surrounding forest had become established in the stabilized slump. In Low Arctic Subarctic Transition in the Mackenzie delta region, Lantz et al (2009) found altered plant composition and increasing alder ( Alnus viridis ) density in RTSs compared to undisturbed tundra over a 56.9 years (mean age of their inactive, stable slumps) period. The enhanced alder growth and reproduction inside RTSs was related, in part, to the increased opportunity for colonization on exposed non acidic substrates (Forbes et al., 2001). Our study fou nd similar results in upland arctic tundra, using a similar chronosequence approach that included several decades, and for some sites even centuries, of changes in the vegetation composition and ecosystem C and N dynamics in areas disturbed by RTS.

PAGE 86

86 CHAPTER 5 CONCLUSIONS The appearance of a recently formed RTS on the landscape mobilizes carbon contained on the organic horizon, exposing the mineral soil and the permafrost to higher temperatures, affecting the normal rate of permafrost soil processes. This mobilization decreases the vegetation cover inside the RTS, opening new spaces for colonizing species and altering the plant composition as one the initial consequences of this disturbance. Revegetat ed RTSs after disturbance, showed an inc reased organic layer depth, car bon and nitrogen pools, and presence of tall deciduous shrubs (as a key contributor of C inputs inside the features). This increased C input reflects not only a re accumulation of the vegetation cover itself, but also a notor ious change in plant community composition lead by the tall deciduous shrubs populations. The increment of these populations throughout the re vegetation process that took place on the different chronosequences, could promote higher rates of litter decompo sition and an increase in nutrient availability as the snow fence capacity, proper of the species that dominate the canopy, could be raised in comparison to adjacent areas with few or no shrubs (Wahren et al 2005). As a localized physical disturbance, RTS acts as a nutrient re distributor and, according to Callaghan (2004), the effect of these results in greater soil warming and permafrost thawing, which tend to increase soil organic matter and nutrient turnover. Shrubs are able to utilize these available n utrients more efficiently than other tundra plants (Tape, 2006). When RTS stabilizes, due to the accumulation of organic and mineral material on top of the headwall, reducing its slope angle and covering the once exposed permafrost

PAGE 87

87 front, surface soil char acteristics inside the features such as active layer depth and C and N pools are not suitable to maintain the tall deciduous shrubs populations established on previous stages of RTS development. A reduction in the canopy dominant shrub component was observ ed for the stabilized RTS at Loon Lake and I minus 1. These features exhibited a lower % tall shrub presence, values which resembled the adjacent, undisturbed tundra canopy dominance more than any revegetated feature. Further studies are necessary in order to understand what is driving the differences in the re accumulation rate of the surface soil C and N pools across the different chronosequences. Complementary analysis of physical aspect inherent to each RTS (i.e. slope, aspect, shape, ice rich permafros t versus ice depleted permafrost, snow accumulation, etc.), vegetation effects on microclimate, trace of inputs from surrounding undisturbed tundra (i.e. nutrients, water, litter) and characterization of the mobility of materials inside the features are ne cessary in order to provide a more complete answer for the question arise by the differences observed on the re accumulation rates of C and N surface soil pools for each of the study sites characterized. This study presents evidence of how RTS dist urbance affect ecosystem C and N pools contained in surface soil, and explains relevant surface soils and vegetation shifts through several decades, using a chronosequence approach in four altered a reas across upland arctic tundr a

PAGE 88

88 APPENDIX A SYSTEMATIC SHRUB SURVEYS ON FOUR RTS CHRONOSEQUENCES ACROSS UPLAND ARCTIC TUNDRA, ALASKA. AGE DISTRIBUTION OF SITES (LOBES) WHERE SHRUBS ROOTED IN MINERAL SOIL WERE SAMPLED Figure A 1 Age distribution of shrubs rooted in mineral soil at Loon Lake RTS chronoseque nce, Noatak National Preserve, AK USA In lobe 5, P. glauca tree rings were count instead of shrub species rings

PAGE 89

89 Figure A 2. Age distribution of shrubs rooted in mineral soil at NE 14 RTS chronosequence, Toolik Field Station, AK USA

PAGE 90

90 Figure A 3. Age distribution of shrubs rooted in mineral soil at Itkillik RTS chronosequence, Toolik Field Station, AK USA

PAGE 91

91 Figure A 4 Age distribution of shrubs rooted in mineral soil at I minus 1 RTS chronosequence, Toolik Field Station, AK USA

PAGE 92

92 APPENDIX B DISSOLVE INORGANIC NITROGEN AND MELICH 1 EXTRACTABLE PHOSPHORUS IN SURFACE ORGANIC AND MINERAL SOILS

PAGE 93

93 Table B 1 Ammonia, nitrate and phosphorus concentration in surface organic and mineral soil at NE 14. Ammonia and nitrate were extracted from air dried soils (ADS) and field moist soils (FMS). Phosphorus was extracted from ADS soils ODE=oven dried equivalent TK Tr Soil Type Section depth (cm) n %N SE NO 3 ( g per g of ODE soil) ADS SE NH 4 + ( g per g of O DE soil) ADS SE NO 3 ( g per g of ODE soil) FM S SE NH 4 + ( g per g of ODE soil) FM S SE P ( g of P per g of ODE soil) SE NE 14TK1 T0 Org 0 5 6 1.434 0.235 1.245 0.798 15.012 2.844 3.637 2.334 48.746 10.453 3.85 1.53 Org 5 15 4 1.896 0.354 0.344 0.344 13.593 1.051 2.255 2.255 64.708 11.859 5.16 2.61 Org 15 25 1 1.525 0.502 0.000 1.384 0.000 0.18 Org 25 35 1 2.035 2.512 10.608 6.701 28.295 0.24 Min 0 10 4 0.275 0.107 1.996 1.171 9.148 4.168 2.477 1.548 21.046 12.510 13.80 4.34 NE 14TK1 T1 Org 0 5 2 1.717 0.595 2.290 0.014 0.277 0.098 8.160 3.235 0.986 0.077 0.22 0.19 Org 5 15 2 2.253 0.020 0.257 0.257 6.092 5.944 1.073 1.073 25.426 24.803 1.02 0.76 Min 0 10 6 0.213 0.091 0.646 0.633 8.878 4.399 1.349 1.334 13.001 6.357 0.62 0.55 NE 14TK2 T1 Org 0 5 5 1.709 0.180 1.323 0.565 11.122 2.337 8.182 5.295 68.524 14.129 9.44 6.87 Org 5 15 4 1.859 0.326 0.853 0.853 4.290 1.715 6.273 6.273 12.725 3.346 3.66 2.59 Org 15 25 3 1.740 0.167 0.204 0.121 8.165 2.095 0.637 0.417 30.663 12.842 0.85 0.49 Min 0 10 6 0.362 0.112 0.478 0.458 9.237 2.906 1.077 1.046 15.294 5.791 6.38 2.91 NE 14TK3 T1 Org 0 5 6 1.690 0.120 1.318 0.357 1.085 0.629 5.254 2.092 7.349 5.634 3.41 3.25 Org 5 15 2 1.721 0.195 1.252 0.938 5.582 5.582 4.270 2.014 40.147 40.147 2.16 1.10 Min 0 10 4 0.408 0.145 0.585 0.585 10.524 4.632 0.584 0.584 11.544 5.072 13.74 3.33

PAGE 94

94 Table B 2. Ammonia, nitrate and phosphorus concentration in surface organic and mineral soil at Loon Lake Ammonia and nitrate were extracted from air dried soils (ADS) and field moist soils (FMS). Phosphorus was extracted from ADS soils ODE=oven dried equivalent TK Tr Soil Type Section depth (cm) n %N SE NO 3 ( g per g of ODE soil) ADS SE NH 4 + ( g per g of ODE soil) ADS SE NO 3 ( g per g of ODE soil) FM SE NH 4 + ( g per g of ODE soil) FM SE P ( g of P per g of ODE soil) SE Loon Lake T0 Org 0 5 6 1.503 0.249 0.180 0.095 30.279 6.855 0.738 0.358 146.260 31.770 7.07 4.78 Org 5 15 1 1.801 0.253 4.748 1.116 17.309 3.41 Org 15 25 1 1.998 0.106 8.832 0.221 18.425 26.49 Min 0 10 4 0.153 0.001 0.051 0.003 2.252 0.960 0.069 0.002 3.243 1.532 5.26 2.43 Min 10 20 2 0.166 0.023 0.754 0.629 1.572 0.337 1.120 0.961 2.144 0.275 12.14 7.07 Loon Lake TK1 T1 Org 0 5 1 1.008 0.062 30.098 0.829 1.967 28.61 Org 5 15 1 1.350 0.237 68.746 1.399 2.335 0.14 Min 0 10 6 0.228 0.070 0.460 0.190 6.437 4.893 1.257 0.849 314.318 297.179 0.33 0.13 Min 10 20 1 0.303 0.917 1.530 0.545 266.374 0.16 Loon Lake TK2 T1 Min 0 10 6 0.186 0.070 1.329 0.306 0.560 0.047 2.393 0.668 0.981 0.154 0.14 0.04 Min 10 20 2 0.174 0.026 1.210 0.591 3.411 2.792 2.022 1.037 5.761 4.776 11.00 10.88 Loon Lake TK3 T1 Org 0 5 1 1.845 0.414 89.365 3.316 715.603 1.94 Min 0 10 6 0.207 0.103 0.147 0.054 4.883 1.000 0.212 0.047 7.347 0.562 0.43 0.28 Min 10 20 2 0.171 0.031 0.123 0.005 3.873 0.542 0.176 0.011 5.500 0.637 0.18 0.02 Loon Lake TK4 T1 Org 0 5 6 1.721 0.164 3.221 1.293 19.282 11.786 16.367 5.875 104.941 57.529 5.85 4.66 Min 0 10 1 0.411 2.589 16.199 3.703 23.173 2.19 Min 10 20 1 0.396 0.924 5.546 1.799 10.793 0.22 Loon Lake TK5 T1 Org 0 5 6 1.753 0.353 1.139 0.960 54.103 35.304 10.654 10.103 458.711 393.156 10.06 8.20 Org 5 15 2 1.615 0.207 0.633 0.459 13.119 8.335 4.764 3.985 55.896 41.585 6.02 5.80 Org 15 25 2 1.604 0.238 0.807 0.472 18.220 16.740 2.054 1.461 34.223 30.154 10.08 9.68 Min 0 10 1 0.413 5.423 2.787 5.429 9.112 16.23

PAGE 95

95 Table B 3. Ammonia, nitrate and phosphorus concentration in surface organic and mineral soil at Itkillik. Ammonia and nitrate were extra cted from air dried soils (ADS) and field moist soils (FM S). Phosphorus was extracted from ADS soils. ODE=oven dried equivalent TK Tr Soil Type Section depth (cm) n %N SE NO 3 ( g per g of ODE soil) ADS SE NH 4 + ( g per g of ODE soil) ADS SE NO 3 ( g per g of ODE soil) FM SE NH 4 + ( g per g of ODE soil) FM SE P ( g of P per g of ODE soil) SE IKTK1 T0 Org 0 5 6 1.510 0.230 0.753 0.753 11.856 1.232 1.533 1.533 43.515 9.444 3.85 1.10 Org 5 15 5 2.110 0.229 0.235 0.181 5.554 3.257 1.011 0.835 27.007 15.626 2.69 1.44 Org 15 25 2 1.827 0.053 0.048 0.048 2.346 2.346 0.137 0.137 7.650 7.650 0.18 0.09 Org 25 35 2 1.843 0.055 1.765 1.765 3.755 2.235 5.521 5.521 13.319 5.419 0.23 0.13 Org 35 45 1 2.023 0.000 0.000 0.000 0.000 0.07 Min 0 10 4 0.152 0.056 0.561 0.561 5.050 1.328 0.592 0.592 5.678 1.571 8.80 2.33 T1 Min 0 10 6 0.129 0.051 0.566 0.383 8.054 2.904 0.631 0.420 9.004 3.324 1.62 0.73 IKTK2 T0 Org 0 5 6 1.874 0.083 2.417 2.417 9.724 2.324 9.785 9.785 37.243 9.629 7.00 2.69 Org 5 15 4 2.001 0.151 1.296 0.787 12.794 2.005 4.765 2.834 45.733 7.614 0.61 0.14 Min 0 10 6 0.196 0.045 0.948 0.689 6.650 1.697 1.116 0.812 10.755 4.437 35.54 7.59 Min 10 20 1 1.412 0.219 9.665 0.521 23.028 0.15 T1 Org 0 5 4 1.691 0.409 0.761 0.469 11.118 3.364 1.824 1.130 24.833 6.780 2.26 1.19 Org 5 15 4 1.943 0.093 0.016 0.016 7.232 3.930 0.044 0.044 18.385 9.490 1.67 0.48 Min 0 10 6 0.301 0.131 0.904 0.835 6.984 1.583 1.030 0.942 8.683 1.421 4.95 2.34 IKTK3 T0 Org 0 5 3 1.586 0.108 1.020 1.020 10.718 0.389 5.360 5.360 46.555 6.350 3.84 1.63 Org 5 15 2 1.756 0.173 0.000 0.000 11.944 0.457 0.000 0.000 47.812 9.852 1.74 1.14 Min 0 10 6 0.628 0.176 1.086 0.629 11.773 4.469 2.206 1.417 23.918 10.698 16.65 7.40 T1 Org 0 5 4 1.365 0.189 0.000 0.000 14.163 1.255 0.000 0.000 48.271 5.621 2.58 1.55 Min 0 10 6 0.759 0.279 0.173 0.138 6.812 2.534 0.276 0.188 12.810 5.001 12.24 5.33

PAGE 96

96 Table B 4. Ammonia, nitrate and phosphorus concentration in surface organic and mineral soil at I minus 1. Ammonia and nitrate were extracted from air dried soils (ADS) and field moist soils (F MS). Phosphorus was extracted from ADS soils. ODE=oven dried equivalent TK Tr Soil Type Section depth (cm) n %N SE NO 3 ( g per g of ODE soil) ADS SE NH 4 + ( g per g of ODE soil) ADS SE NO 3 ( g per g of ODE soil) FM SE NH 4 + ( g per g of ODE soil) FM SE P ( g of P per g of ODE soil) SE Iminus1TK1 T0 Org 0 5 6 1.55 0.26 0.285 0.25 19.480 5.62 1.058 0.74 71.691 18.773 4.81 1.81 Org 5 15 4 2.38 0.27 0.000 0.00 18.296 2.38 0.000 0.00 76.751 25.660 1.09 0.44 Org 15 25 2 2.52 0.06 0.212 0.21 15.751 0.90 0.877 0.88 59.634 9.328 0.18 0.03 Iminus1TK1 T1 Org 0 5 1 1.36 0.000 0.362 0.000 1.105 1.02 Org 5 15 1 2.21 0.000 18.670 0.000 91.331 1.00 Min 0 10 6 0.09 0.02 0.772 0.77 0.266 0.13 1.125 1.12 0.368 0.195 31.19 5.86 Min 10 20 2 0.09 0.01 0.279 0.28 0.368 0.37 0.369 0.37 0.580 0.580 35.23 5.60 Iminus1TK2 T1 Org 0 5 1 2.34 0.010 7.989 0.030 22.949 39.12 Org 5 15 1 1.91 0.000 0.000 0.000 0.000 31.69 Min 0 10 6 0.09 0.01 1.186 0.62 11.486 4.86 2.020 1.02 17.553 6.359 23.89 4.21 Min 10 20 3 0.06 0.01 3.217 1.36 7.494 4.38 5.407 2.42 9.696 5.050 31.00 3.13 Iminus1TK3 T1 Org 0 5 3 1.08 0.27 0.068 0.02 14.454 6.15 0.477 0.06 120.705 66.697 1.50 0.59 Org 5 15 3 1.42 0.20 0.163 0.03 23.999 6.09 0.656 0.08 94.002 13.408 4.05 3.70 Min 0 10 6 0. 70 0.12 0.695 0.39 8.280 0.74 1.785 0.18 9.263 4.398 1.31 0.39 Min 10 20 1 0.56 1.615 3.750 3.218 7.472 0.53 0.22 Iminus1TK4 T1 Org 0 5 6 1.62 0.12 0.047 0.03 57.691 16.4 0.284 0.18 397.439 117.20 2 1.40 0.35 Org 5 15 1 2.01 0.025 19.088 0.139 104.431 7.87 Min 0 10 6 0.37 0.16 0.001 0.00 7.413 1.02 0.001 0.00 8.332 1.230 27.43 1.07 Min 10 20 5 0.12 0.01 0.004 0.00 5.035 0.84 0.005 0.00 5.573 0.973 30.54 0.74 Iminus1TK5 T1 Org 0 5 6 1.38 0.19 0.045 0.03 25.726 6.26 0.358 0.30 171.623 43.650 4.69 0.29 Org 5 15 4 1.14 0.13 0.000 0.00 16.999 8.29 0.000 0.00 107.144 52.547 8.82 0.22 Min 0 10 6 0.46 0.07 1.932 1.92 6.465 1.11 3.896 3.88 15.333 2.799 31.47 1.12 Min 10 20 6 0.58 0.12 0.075 0.04 5.212 0.35 0.104 0.06 6.590 0.746 38.15 0.83

PAGE 97

97 Table B 4. C ontinued TK Tr Soil Type Section depth (cm) n %N SE NO 3 ( g per g of ODE soil) ADS SE NH 4 + ( g per g of ODE soil) ADS SE NO 3 ( g per g of ODE soil) FM SE NH 4 + ( g per g of ODE soil) FM SE P ( g of P per g of ODE soil) SE Iminus1TK6 T1 Org 0 5 6 1.43 0.14 0.158 0.07 32.787 4.14 1.945 1.01 317.841 41.594 2.31 0.16 Org 5 15 5 1.87 0.21 0.292 0.12 11.677 0.94 3.061 1.08 125.072 31.216 1.32 0.1 0 Min 0 10 4 0.08 0.01 0.013 0.01 3.741 0.48 0.016 0.01 4.766 0.442 36.02 1.83 Min 10 20 4 0.51 0.32 0.038 0.02 5.036 0.03 0.044 0.01 5.784 0.044 32.97 0.92 Iminus1TK7 T1 Org 0 5 6 1.85 0.17 0.066 0.04 34.450 7.48 0.523 0.37 260.321 46.382 1.82 0.06 Org 5 15 5 1.65 0.11 0.092 0. 04 15.318 5.09 0.597 0.26 81.333 12.118 4.21 0.21 Min 0 10 6 0.41 0.20 0.025 0.02 6.016 0.97 0.045 0.03 20.113 4.666 30.46 1.70

PAGE 98

98 APPENDIX C PERCENT LOSS AND REA CCUMULATION OF ORGAN IC LAYER DEPTH AND SURFACE SOILS CARBON AND NITROGEN THRU TI ME IN FOUR RTS CHRONOSEQUENCES IN U PLAND ARCTIC TUNDRA, AK

PAGE 99

99 Figure C 1. Organic layer depth range of loss and re accumulation due to RTS disturbance in four chronosequences in upland arctic tundra, AK USA Grey band represent the undisturbed tundra stage. Reference lines in the X axis (starting from 0) represent pre disturbance values of organic layer depth per chronosequence

PAGE 100

100 Figure C 2. Surface soil organic C pool range of loss and re accumulation due to RTS disturbance in four chronosequen ces in upland arctic tundra, AK, USA. Grey band represent the undisturbed tundra stage. Reference line s in the X axis (starting from 0) represent pre disturbance values of % organic C pool per chronosequence. Organic C pools were calculated from the whole organic layer per site per chronosequence

PAGE 101

101 Figure C 3. Surface soil mineral C pool range of loss and re accumulation due to RTS disturbance in four chronosequences in upland arctic tundra, AK USA Grey band represent the undisturbed tundra stage. Reference line s in the X axis (starting from 0) represent pre disturbance values of % mineral C pool per chronosequence. NE 14, Itkillik and Loon Lake mineral C pools were calculated from the top 10cm of the mineral layer sampled on each site. I minus 1 mineral C pools were calculated from the whole mineral layer sampled at each site. For Itkillik 3, t1=0.0026%, for Loon Lake, t1=0.1037%, t2= 0.00664% and t3=0.0249% *

PAGE 102

102 Figure C 4. Surface soil total C pool range of loss and re accumulation due to RTS disturbance in four chronosequences in upland arctic tundra, AK USA Grey band represent the undisturbed tundra stage. Reference lines in the X axis (starting from 0) represent pre disturbance values of % total C pool per chronosequence. NE 14, Itkillik and Loon Lake total C pools were calculated from the whole organic layer plus the top 10cm of the mineral layer sampled on each site. I m inus 1 total C pools were calculated from the whole organic layer plus the whole mineral layer sampled at each site. For I minus 1, t5=0.012%

PAGE 103

103 Figure C 5. Surface soi l organic N pool range of loss and re accumulation due to RTS disturbance in four chronosequences in upland arctic tundra, AK USA Grey band represent the undisturbed tundra stage. Reference lines in the X axis (starting from 0) represent pre disturbance values of % organic N pool per chronosequence. Organic N pools were calculated from the whole organic layer per site per chronosequ ence. For I minus 1, t7= 0.0753%, for NE 14, t2=0.0286% and t3= 0.0786%

PAGE 104

104 Figure C 6. Surface soil mineral N pool range of loss and re accumulation due to RTS disturbance in four chronosequences in upland arctic tundra, AK USA Grey band represent the undisturbed tundra stage. Reference line s in the X axis (starting from 0) represent pre disturbance values of % mineral N pool per chronosequence. NE 14, Itkillik and Loon Lake mineral N pools were calculated from the top 10cm of the mineral layer sampled on each site. I minus 1 mineral N pools were calculated from the whole mineral layer sampled at each site. For I minus 1, t2=0.10%, for Itkillik 1, t1=0.1329%, and for Itkillik 3, t1=0.1258%

PAGE 105

105 Figure C 7. Surface soil total N pool range of loss and re accumulation due to RTS disturbance in four chronosequences in upland arctic tundra, AK USA Grey band represent the undisturbed tundra stage. Reference lines in the X axis (starting from 0) represent pre disturbance values of % total N pool per chronosequence. NE 14, Itkillik and Loon Lake total C pools were calculated from the whole organic layer plus the top 10cm of the mineral layer sampled on each site. I m inus 1 total N pools were calculated from the whole organic layer plus the whole mineral layer sampled at each site. For I minus 1, t5= 0.0786% and t7= 0.0225%, for Itkillik 2, t1=0.1154%

PAGE 106

106 LIST OF REFERENCES Bartleman AP, Miyanishi K, Burn CR, Cote MM ( 2001 ) Development of vegetation communities in a retrogressive thaw slump near Mayo, Yukon Territory: a 10 year assessment. Arctic 54 149 156 Bliss LC, Cantlon JE ( 1957 ) Succession on river alluvium in northern Alaska. American Midland Naturalist 58 452 46 9 Burn CR, Friele PA ( 1989 ) Geomorphology, vegetation succession, soil characteristics and permafrost in retrogressive thaw slumps near Mayo, Yukon Territory. Arctic 42 31 40 Callaghan T, Bjorn LO, Chernov Y, Chapin FS, Christensen T, Huntley B, Im s R, Johansson M ( 2004 ) Effects on the function of Arctic ecosystems in the short and long term perspectives. Ambio 33 448 458 Chapin FS, Shaver GR, Giblin AE, Nadelhoffer KG, Laundre JA ( 1995 ) Response of arctic tundra to experimental and observed changes in climate. Ecology 76 694 711 CAVM Team ( 2003 ) Circumpolar Arctic Vegetation Map. (1:7,500,000 scale), Conservation of Arctic Flora and Fauna (CAFF) Map No. 1. U.S. Fish and Wildlife Servic e, Anchorage, Alaska. ISBN: 0 9767525 0 6, ISBN 13: 978 0 9767525 0 9 de Krom V (1990) Retrogressive thaw slumps and active layer slides on Herschel Island, Yukon. M.Sc. Thesis, McGill University, Montral, Qubec. Epstein HE, Beringer J, Gould WA (2004) The nature of spatial transitions in the Arctic. Journal of Biogeography 31 1917 1933 Forbes BC, Ebersole JJ, Strandberg B (2001) Anthropogenic disturbance and patch dynamics in circumpolar Arctic ecosystems. Conservation Biology, 15 954 969 Frauenfeld OW Zhang TJ, Barry RG, Gilichinsky D (2004) Interdecadal changes in seasonal freeze and thaw depths in Russia. Journal of Geophysical Research 109 D05101 Gaudinski JB, Trumbore SE ( 2000 ) Soil carbon cycling in a temperate forest: radiocarbon based estimates of residence times, sequestration rates and partitioning of fluxes. Biogeochemistry 51 (1) 33 69 Gaudinski J, Trumbore SE, Davidson EA, Cook AC, Markewitz D, Richter DD ( 2001 ) The age of fine root carbon in three forests of the United States measured by radiocarbon. Oecologia 129 420 429

PAGE 107

107 Gaudinski JB, Dawson TE, Quideau S (2005) Comparative analysis of cellulose preparation techniques for use with C 13, C 14, and O 18 isotopic measure ments. Analytical chemistry 77 (22) 7212 7224 Haag RW, Bliss LC ( 1974 ) Energy budget changes following surface disturbance to upland tundra. Journal of Applied Ecology 11 355 374 Harper KA, Kershaw GP (1996) Natural revegetation on borrow pits and vehicle tracks in shrub tundra, 48 years following construction of the CANOL no. 1 pipeline, NWT, Canada. Arctic and Alpine Research 28 163 17 Hicks Pries CE, Schuur EAG, Crummer KG (2011) Holocene Carbon Stocks and Carbon Accumulation Rates Altered in Soils Undergoing Permafrost Thaw. Ecosystems DOI 10.1007/s10021 011 9500 4 Holt EA, McCune B, Neitlich P ( 2009 ) Macrolichen communities in relation to soils and vegetation in the Noatak Nati onal Preserve, Alaska Botany 87 241 252 Johnson LC, Shaver GR, Giblin AE, Nadelhoffer KJ R astetter ER, Laundre JA, Murray GL (1996) Effects of drainage and temperature on carbon balance of tussock tundra microcosms. Oecologia 108 737 748 Johnstone JF, Henry GHR (1997) Retrospective analysis of growth and reproduction in Cassiope tetragona and relations to climate in the Canadian High Arctic. Arctic and Alpine Research 29 459 469 Jorgenson MT, Racine CH, Walters JC, Osterkamp TE (2001) Permafrost degradation and ecological changes associated with a wa rming climate in central Alaska. Climate Change 48 551 579 Jorgenson MT, Osterkamp TE (2005) Response of boreal ecosystems to varying modes of permafrost degradation. Canadian Journal of Forest Research 35 2100 2111 Kreiger K (2012) The Topographic Form and Evolution of Thermal Erosion Features: A First Analysis Using Airborne and Ground Based LiDAR in Arctic Alaska, Idaho State University, M.S. Thesis in Geology, 115p Kokelj SV, Lantz TC, Kanigan J, Smith SS, Coutts R (2009) On the origin and polycyclic behavior of tundra thaw slumps, Mackenzie Delta region, Northwest Territories, Canada. Permafrost and Periglacial Processes 20 Landhausser SM, Wein RW (1993) Postfire Vegetation Re accumulation and Tree Establishment at the Arctic Treeline: Climate Change Vegetation Response Hypotheses. J ournal of Ecology 81 665 672 Lantz TC, Kokel SV, Gergel SG, Henry GHR (2009) Relative impacts of disturbance and temperature: persistent changes in microenvironment and vegetation in retrogressive thaw slumps. Global Change Biology 15 1664 1675

PAGE 108

108 Lewkowicz AG (1987) Headwall retreat of ground ice slumps, Banks Island, Northwest Ter ritories. Canadian Journal of Earth Sciences 24 1077 1085 Lloyd AH, Yoshikawa K, Fastie CL Hinzman I, Fraver M (2003) Effects of permafrost degradation on woody vegetation at arctic treeline on the Seward Peninsula, Alaska. Permafrost and Periglacial Processes 14 93 101 Mack MC, Schuur EAG, Bret Harte SM, Shaver G R, Chapin FS (2004) Ecosystem carbon storage in arctic tundra reduced by long term nutrient fertilization. Nature 431 440 443 Mackay JR, Burn CR (2002) The first 20 years (1978 1979 to 19 98 1999) of active layer development, Illisarvik experimental drained lake site, western Arctic coast, Canada. Canadian Journal of Earth Sciences 39 1657 1674 McGuire DM, Chapin FS, Walsh JE, Wirth C (2006) Integrated Regional Changes in Arctic Climate Feedbacks: Implications for the Global Climate System. A nnual Review of Environment and Resources 31 61 91 Murphy J, Riley JP (1962) A modified single solution method for the determination of phosphate in nat ur al waters. Anal ytica Chim ica Acta 27 31 36 Nadelhoffer KJ, Giblin AE, Shaver GR, Linkins A E (1992) Microbial processes and plant Pages 281 300 in E S. Chapin III, R. L. Jefferies J,. F Reynolds, G.R. Shaver, and J. Svoboda, editors. Arctic ecosystems in a changing climate: an ecophysiological perspective. Academic Press, San Diego, California, USA Oelke C, Zhang TJ (2004) A model study of circum arctic soil temperatures. Permafrost Per iglacial Process 15 103 121 Osterkamp TE, Ro manovsky VE (1999) Evidence for warming and thawing of discontinuous permafrost in Alaska. Permafrost and Periglacial Processes 10 17 37 Ovenden L (1986) Vegetation colonizing the bed of a recently drained thermokarst lake (Illisarvik), Northwest Territories Canadian Journal of Botany 64 2688 2692 Pollack HN, Demezhko DY, Duchkov AD, Golovanova IV, Huang SP, Shchapov VA, Smerdon JE (2003) Surface temperature trends in Russia over the past five centuries r econstructed from bo rehole temperatures. Journal of Geophysical Research, 108(B4) 2180 Rayback SA, Henry GHR (2005) Dendrochronological potential of the arctic dwarf shrub, Cassiope tetragona Tree Ring Research 61 43 53

PAGE 109

109 Rhoades C, Binkley D Oska rsson H, Stottlemyer R (2008) Soil nitrogen accretion along a floodplain terrace chronosequence in northwest Alaska: influence of the nitrogen fixing shrub Shepherdia canadensis. Ecoscience 15 223 230 Russell RS (1940) Physiological and ecological studies on Arctic vegetation. II. The development of vegetation in relation to nitrogen supply and soil micro organisms on Jan Mayen Island. J ournal of Ecology 28 269 288 Shaver GR, Chapin FS (1980) Response to fertilization by various plant growth forms in an Alaskan tundra: nutrient accumulation and growth. Ecology 61 662 675 Schimel JP, Chapin FS (1996) Tundra Plant Uptake of Amino Acid and NH 4 + Nitrogen in Situ: Plants Complete Well for Amino A cid N. Ecology 7 7 2242 2147 Schimel JP, Bilbrough C, Welker JA (2004) Increased snow depth affects microbial activity and nitrogen mineralization in two Arctic tu ndra communities. Soil Biology and Biochemistry 36 217 227 Schuur EAG, Crummer KG (2007) Plant species composition and productivity following permafrost thaw and thermokarst in alaskan tundra. Ecosystems 10 280 292 Stow DA, Hope A, McGuire D (2004) Remote sensing of vegetation and land cover change in Arctic Tundra Ecosystems. Remote Sens in g of Environment 89 281 308 Stuiver M, Polach HA (1977) Discussion: reporting of 14 C data. Radiocarbon 19 355 363 Sturm M, Schimel J P, Michaelson G, Welker JM, Oberbauer S, Liston GE, Fahnestok J, Romanovsky VE (2005) Winter biological processes could help convert arctic t undra to shrub land. Bioscience 55 17 26 Sturm M R acine C, Tape K (2001) Climate change: Increasing shrub abundance in the Arctic. Nature 411 546 547 Svoboda J, Henry GHR (1987) Succession in mar ginal Arctic environments. Arctic and Alpine Research 19 373 384 Tape K, Sturm M, Raci ne C (2006) The evidence for shrub expansion in Northern Alaska and the Pan Arctic. Global Change Biology 12 686 702 Van Everdingen RO ( 1998 ) Multi language Glossary of Permafrost and Related Ground ice Terms. International Permafrost Ass ociation. University of Calgary Viereck LA, Dyrness CT, Batten AR, Wenzlick KJ (1992) The Alaska Vegetation Classification. USDA Forest Servic e Pacific Northwest Research Station, General Technical Report 286 Portland, Oregon

PAGE 110

110 Wahren CHA, Walker MD (2005) Vegetation responses in Alaskan arctic tundra after 8 years of a summer warming and winter snow manipulation expe riment. Global Change Biology 1 1 537 552 Walker DA, Gould WA, Maier HA, Raynolds MK (2002) The circumpolar Arctic vegetation map: AVHRR derived base maps, environmental controls, an d integrated mapping procedures. International Journal of Remote Sensing 23 4551 4570 Walker LR, Wardle DA, Bardgett RD, Clarkson BD (2010) The use of chronosequences in studies of ecological succession and soil development. Journal of Ecol ogy, 98 725 736 Zhang T, Osterkamp TE, Stamnes K (1997) Effects of climate on the active layer and permafrost on the north slope of Alaska, USA. Permafrost Periglacial Process, 8 45 67

PAGE 111

111 BIOGRAPHICAL SKETCH iology from the National University of Colombia. During the course of his undergraduate studies, he was influen ced by the vast world of plant ecop hysiology and the exciting challenge that field work in the Tropics implies. After graduating from college, he was introduced to the field of ecosystem e cology research In other words: he left his home in B ogota, Colombia, and went to explore the factors that influence the productivity of a super wet forest in the island of Gorgona, Colombia Then he went to the boreal f ores t in Q uebec and New Brunswick to learn about moss es and lichens and their vast contr ibution to the ecosystem function. F inally, he accepted to be part of an adventure to reveal the secrets of the soils and plan ts in the arctic tundra of the great s tate of Alaska and detect how they are altered by geo morphological disturbances Thro ugh t hese research experiences, he obtained numerous field and laboratory skills as a research assistant. He also made excellent friends and partners The early introduction to the particular rhythm o f work inside research group s motivated him to apply for a master in s cience degree at the University of Florida. As a result of his experi ence as a young graduate student, Andres discovered a natural passion for fiel d and laboratory work, which enhanced his disposition to learn new concepts and techniques in e col ogy. During his graduate studies, Andres realize d that being a productive, constructive and reliable unit in side a team is a role that gives him a great satisfact ion. He also was a teaching assistant for the courses Introduction to Biology I and General Ec ology While finishing his master in s cience, Andres was hired as the new lead ecological field technician at the Joseph W. Jones Ecological Research Center at Ichauway, Georgia