TECHNICAL SECTION

SERDP RELEVANCE This research is designed to understand the mechanistic connections among vegetation, the organic soil layer, and permafrost ground stability in Alaskan boreal ecosystems. Permafrost is a major control of structure and function of boreal ecosystems, and the soil organic layer mediates the effects of a changing climate on the ground thermal regime and permafrost stability. Understanding the links among vegetation, organic soil, and permafrost is critical for projecting the impact of climate change on permafrost in ecosystems that are subject to abrupt anthropogenic and natural disturbances (fire) to the organic layer. This research responds to several elements within the SERDP Statement of Need by using Department of Defense (DoD) and surrounding lands in Interior Alaska to explore and test the conceptual and mechanistic basis for threshold change and regime shift in the Alaskan boreal forest (SON 4). Based on the understanding of these mechanisms developed from field measurements, this research will use spatially explicit numerical modeling to predict the response of permafrost ground and disturbance regimes to projected changes in climate (SON 2). These modeling approaches will provide a dynamic mapping tool to help land managers identify those DoD lands that are resistant and those that are vulnerable to permafrost degradation under scenarios of disturbance and climate change (SON 1,3). In addition, these approaches will allow land managers to explore the consequences of interactive changes in climate and management for vegetation composition, establishment of invasive species, fire dynamics, and ecosystem structure and function (SON 3).

TECHNICAL OBJECTIVES This research will combine field measurements (Objective 1) with models (Objective 2) to detect and predict state changes in boreal ecosystems of Interior Alaska in response to changing climate and land management. OBJECTIVE 1) Determine mechanistic links among fire, soils, permafrost, and vegetation succession in order to develop and test field-based ecosystem indicators that can be used to directly predict ecosystem vulnerability to state change. We propose three activities to develop and assess indicators of ecosystems at risk of state change in response to changes in climate, fire regime, or fire management:

a) Monitoring vegetation recolonization, soils, and permafrost on a previously existing network of sites located in recent, severe wildfires adjacent to, and on, DoD lands in Interior Alaska.

b) Extending this network to include parallel measurements at sites located in recent prescribed fires and fuel treatments on DoD lands.

c) Conducting studies of vegetation stand history and organic layer re-accumulation on an established network of mid-successional boreal ecosystems adjacent to, and on, DoD lands.

Questions that will be addressed: How does fire severity (i.e. deep burning versus shallow burning of the soil organic layer) and fire management affect rates of permafrost degradation? Do sites with unusually deep burning of the soil organic layer have higher rates of permafrost degradation? How does fire severity affect re-establishment of future deciduous and conifer forests? How does soil organic layer re-accumulation differ between deciduous and conifer trajectories? What effects do natural wildfires and human disturbances associated with fire management (prescribed fire, fuel treatments) have on invasion by exotic plant species? OBJECTIVE 2) Forecast landscape change in response to projected changes in climate, fire regime, and fire management. We propose four activities to be able to accurately forecast how fire regime and fire management will interact with climate change to shape the future structure, function, and distribution of Alaskan boreal ecosystems on DoD and surrounding lands. These activities include:

a) Incorporating field data sets on vegetation, soils, and permafrost (developed as part of Objective 1) into a model of landscape fire dynamics and into a model of ecosystem structure and function.

b) Coupling these two stand-alone models so that the influence of a changing climate on permafrost and vegetation can be assessed together with natural and managed changes in the fire regime.

c) Evaluating the performance of the coupled model using retrospective statistical datasets of past fire regime and forest structure in Interior Alaska.

d) Projecting future landscape distribution of vegetation and permafrost using the coupled model in combination with different scenarios of climate change, fire regime, and fire management.

Questions that will be addressed: What boreal ecosystems on DoD and surrounding lands are most at risk of losing permafrost from the direct effects of climate change? What boreal ecosystems are most at risk of wildfire under future scenarios of climate change? Which of these are most at risk of losing permafrost following fire? How will fire- or climate-induced changes in the distribution of deciduous and conifer forest types feedback to influence patterns of fire and permafrost on the landscape? Can fire management affect the probability of severe wildfires and ecosystem state change in a future climate?

TECHNICAL APPROACH 1) Background The boreal fire cycle: The boreal region of Interior Alaska comprises a mosaic of evergreen, deciduous, and mixed forest ecosystems interspersed with herbaceous or shrubby wetlands. These ecosystems are often underlain by permanently frozen—permafrost—soils, and range across gradients of soil moisture that vary with topography, from very poorly drained to well drained soils. Evergreen stands dominated by black spruce (Picea mariana) are the most abundant forest type in Interior Alaska and are frequently underlain by permafrost (Van Cleve et al. 1991). Black spruce forests are highly flammable and typically burn during stand-replacing fires every 70-130 years (Johnstone et al. in press). Fire offers an opportunity for plant community reorganization that can be strongly influenced by fire characteristics (Chapin et al. 2006). After high severity fires, mono-dominant spruce stands can give way to fully deciduous stands where permafrost degrades or is lost completely. After less-severe fires, spruce replaces itself and permafrost can stabilize and recover over succession. Stable cycles of fire disturbance and spruce self-replacement have persisted for over 8kya, since black spruce came to dominate the evergreen forests of Interior Alaska (Chapin et al. 2006). Forecasted changes in future climate, however, could affect the stability of boreal

ecosystems directly, by warming permafrost in undisturbed ecosystems, and indirectly, through an increase in fire size and severity (Kasischke and Turetsky 2006)(Figure 1). These direct and indirect effects of climate warming could, together, drive a regional fire regime shift and alter the distribution of boreal ecosystems, changing the landscape of Interior Alaska and the goods and services it provides to humans (Chapin et al. 2008).

Figure 1. Climate change can affect boreal ecosystems (vegetation/soils/permafrost) directly (solid lines), and perhaps more importantly, indirectly (dashed lines) via the fire cycle.

Permafrost control over ecosystem dynamics: Permafrost is central for understanding potential threshold change faced by boreal ecosystems (Schuur et al. 2008). Because permafrost ground includes large volumes of ice, permafrost thaw can trigger major changes in surface topography—known as thermokarst—as ice melts and the ground surface subsides (Shur and Jorgenson 2007). As the freezing point threshold is passed, permafrost thaw and ground subsidence can have profound biological and physical effects on natural ecosystems. Vegetation is dependent on the surface water table created by permafrost, so much so that ground subsidence and redistribution of soil moisture can have much larger effects on ecological functioning than changes in ground temperature alone (Schuur et al. 2009). In addition, changes in the ground structure due to loss of ground ice have catastrophic effects on facilities, infrastructure, and military testing and training. Permafrost temperature, thickness, and geographic continuity are controlled to a large extent by the surface energy balance and thus vary strongly with latitude (Brown et al. 1998). Interior Alaska lies within the discontinuous permafrost zone, where regional temperature is not low enough to sustain permafrost everywhere and existing permafrost is especially susceptible to observed and anticipated climate warming (Hinzman et al. 2005). Here, permafrost stability and distribution is a function of local factors including the vegetation structure and, in particular, by the characteristics of the thick organic soil layer present in many boreal ecosystems. Organic soil layers buffer the response of permafrost to climate variability and delay heat propagation into the soil (Osterkamp et al. 1994). Thus, ecosystem structure, particularly the thickness of the soil organic layer as well as rapid changes in that layer via fire or human activities, can have profound effects on permafrost stability and its degradation in a changing climate.

Vegetation composition and ecosystem dynamics: Plant-soil-microbial (PSM) feedbacks between vascular plants, mosses, and microbial decomposition maintain deep organic soils in the black spruce forests and wetlands of Interior Alaska (Figure 2). This internal feedback has been a key source of ecosystem resilience under the historical fire regime; moist, cold soils, poorly drained due to permafrost, burn at low severity and create a seedbed that favors the re-establishment of black spruce (Johnstone and Chapin 2006) and the recovery of the organic soil layer. In extreme fires, however, these soils can burn deeply (Boby et al. in press). When less than ~5 cm of organic soil remains after fire, deciduous tree species such as aspen and birch establish at high densities (Johnstone et al. 2010) and catalyze a switch to alternate plant successional trajectories that are dominated or co-dominated by deciduous trees. Here, a new PSM feedback domain emerges where shallow organic soils are maintained by rapidly decomposing litter from highly productive deciduous species. Degradation and loss of permafrost is likely once this threshold is reached (Yoshikawa et al. 2002), leading to a state change that permanently alters ecosystem structure and function. Shifts between domains of spruce vs. deciduous dominance and the resulting effects on permafrost have large implications for ecosystem productivity and carbon storage (Mack et al. 2008), feedbacks to regional climate (Randerson et al. 2006), the goods and services that boreal ecosystems provide to humans (Chapin et al. 2008), including ecosystem resilience to anthropogenic activities such as those

Figure 2. Conifer and deciduous forest stands maintain distinct ecosystem dynamics through internal cycles mediated by the organic soil layer. Fire severity is an important determinant of successional trajectory.

related to military operations. In addition, fire and shifts in PSM feedbacks may also create new opportunities for the spread of invasive plant species that capitalize on soil disturbance and associated warming and reduced acidity of the soil (Conn et al. 2008). Because the organic soil layer plays this dual role, mediating climate effects on the ground thermal regime and controlling vegetation successional dynamics, it is a key indicator for understanding resilience and state change of boreal ecosystems on permafrost soils.

Managing interactions among fire, vegetation, and permafrost: Fire management has the potential to influence the natural fire regime by determining the spatial patterning and timing of fire occurrence, and thus the successional state of the ecosystems within the management area (Figure 1). Controlled burning of selected evergreen stands can increase deciduous stand frequency on the landscape with the potential for altering spread of natural wildfires because of the differential flammability of boreal ecosystems (Chapin et al. 2006). White spruce is generally less flammable than black spruce, as illustrated by a long history during the Holocene of white spruce dominance (8-10 kya) with low fire frequency (Higuera et al. 2009). However, the juxtaposition of black and white spruce forest stands on the landscape means that white spruce often burns in tandem within the fire regime of black spruce, as do shrubby or herbaceous wetlands where surface organic soils can serve as a ground fuel to carry fire during dry months (Turetsky et al. 2002). In contrast, deciduous or early successional stands have little ground fuel and while they can burn, they often reduce the spread of fire relative to other ecosystem types (Cumming 2001). Human manipulation of the fire cycle via lower intensity controlled burns has the potential to reduce the risk of high intensity large fires –the type most likely to burn the organic soil deeply and cause threshold changes in permafrost and vegetation– by creating a patchwork of early successional vegetation that could reduce fire spread under some conditions. However, fire management also includes the installation of fire containment lines (firebreaks). Although these affect a smaller area than controlled burns, containment lines can be a much more intense disturbance of the vegetation and the organic soil layer, causing threshold changes in permafrost that could alter landscape patterns of drainage (L. A. Viereck, unpublished data). In sum, fire management has the potential to reduce the risk of the severe fires that are most likely to cause future state changes in boreal ecosystems, but has associated localized disturbances that are potentially novel to these ecosystems.

Fire management on military lands: The DoD manages approximately 7,200 km2 of land in the State of Alaska. Over 95% of Alaskan military land is located in the boreal forest of Interior Alaska, associated with Fort Wainwright and Eielson Air Force Base near the city of Fairbanks, and with Fort Greeley near the city of Delta Junction (Appendix Figure 1). These lands cross two ecologically, economically, and culturally important boreal eco-regions: the Tanana-Kuskokwim Lowlands, which covers about 52,000 km2 of Interior Alaska, and the Yukon-Tanana Uplands, which covers about 102,000 km2. Fire is the most widespread natural disturbance in these regions. The Yukon-Tanana Uplands have the highest incidence of lightning strikes in Alaska and the Yukon Territory (Dissing and Verbyla 2003). In addition to high natural sources of ignition, these military lands also experience high human ignition pressures due to their proximity to the road system and urban areas, and the frequency of military testing and training activities. Thus, these military lands are designated in a distinct fire management zone by the Alaska Interagency Coordination Center (Appendix Figure 2) so that local fire management officers (FMO) can address the unique needs of military land management. Because of this fire management designation, DoD lands offer the potential to understand both the threshold changes associated with severe fires, and the potential for fire management to either contribute to or mitigate vulnerability of DoD lands to threshold change in ecosystem state.

Project goals: Detecting and predicting state change in the boreal landscape in response to changes in climate and fire management against a backdrop of substantial variation caused by the natural disturbance regime is a technical and scientific challenge. This research is designed to answer this challenge by focusing on the mechanistic connections among vegetation, the organic soil layer, and permafrost ground stability in Alaskan boreal ecosystems. Permafrost is

a major control of the structure and function of boreal ecosystem, and the soil organic layer mediates effects of changing climate on the ground thermal regime and permafrost stability. Understanding this linkage, and the control of vegetation over soil organic layer dynamics, is critical for projecting the impact of climate change on permafrost in ecosystems that are subject to abrupt anthropogenic and natural disturbances (fire) to the organic layer. Within this conceptual framework, we focus on aspects that are central to understanding threshold change in the boreal region. First, we seek to develop field-based measurements that can indicate where and when threshold changes are likely. We hypothesize that major threshold change is more likely occur in ecosystems that are already at the margins—forests that, historically, are already stressed—and in fires that are at the extremes in terms of size or severity. Second, the coupling of field data and models will be used to identify where and when threshold changes might be expected in the future. Lastly, we hypothesize that fire management can be used as a tool to mitigate risk of large and severe fires most likely to cause threshold changes. We will use this newly developed model framework to test the influence that management can have on fire patterns in a future climate. 2) Methods OBJECTIVE 1) Determine mechanistic links among fire, soils, permafrost, and vegetation succession in order to develop and test field-based ecosystem indicators that can be used to directly predict ecosystem vulnerability to state change. Our approach towards this objective is to use experiments and field surveys of soils and vegetation in disturbed and undisturbed boreal ecosystems to identify thresholds in key ecosystem responses to natural and human-manipulated fire treatments. We will focus our studies on three time scales of ecosystem recovery: 1) early post-fire responses that occur within a decade of disturbance and represent the initiation of vegetation successional trajectories, 2) mid-successional responses that demonstrate the longer-term consequences and stabilizing processes associated with different successional trajectories, and 3) mature ecosystem responses that indicate the structure and function of successional endpoints in response to current climatic conditions. We will explicitly compare ecosystem responses to natural disturbance caused by wildfire with responses to human fire management activities on DoD lands such as prescribed fire, fuel treatments, and containment lines. Field measurements will provide key information for both the initial parameterization of ecosystem models and data to be used later in model validation. In addition, our field measurements are aimed at determining indicators that forecast current or future ecosystem change, such as organic layer consumption during fire, tree ring growth, or moss abundance. These indicators will allow us, and land managers, to predict the conditions and ecosystems most vulnerable to state change under altered climate and fire disturbance regimes. The key elements of our field research methods relating to each step in completing the objective are as follows: a) Monitor vegetation recolonization, soils, and permafrost in recent, severe wildfires in Interior Alaska. We will measure ecosystem responses to recent wildfires across landscape gradients in site moisture (well drained vs. poorly drained), permafrost (temperature, active layer thickness), and fire severity (high to low). These are the key ecosystem aspects that control variation in ecosystem vulnerability and resilience to fire across different landscape units and vegetation types.

Study sites: Our study sites for this activity will build off an existing network of 90 black spruce sites that burned in three large fires in 2004 in Interior Alaska (Appendix Figure 3) (Johnstone et al. 2010, Boby et al. in press)). These plots are 30x30 m in size and each is located within a relatively homogenous stand. We will increase the spatial coverage of this network to include pre-existing sites located within recent fires on DoD lands in Interior Alaska, such as the 2001 Survey Line fire within the Fort Wainwright bombing range. Our sampling network will provide a comprehensive representation of black spruce stands, which are a dominant forest type in Interior Alaska and are expected to be sensitive to changes in the fire

and climate regime (Chapin et al. 2006). In addition, we will expand our network to include a broader set of ecosystem classes that include burned upland white spruce and deciduous stand types, shrub communities, and tundra. Selection of new study sites will be based on existing Landsat-derived maps of pre-fire vegetation and fire severity for recently burned areas (Verbyla and Lord 2008). In total, our sampling network will provide a suite of >100 sites that are broadly representative of ecosystems in Interior Alaska, are found within DoD-managed lands, and encompass the full range of fire, soil moisture/drainage, and permafrost characteristics. Post-fire responses at a representative sub-set of these sites will be compared to measured unburned conditions from a network of paired unburned sites (expanding from the network of 28 sites in Boby et al. in press (Appendix Figure 3)). Unburned sites are selected to have similar conditions of stand age, composition, and structure as the paired burned sites, and provide a measure of the variability in pre-fire ecosystem dynamics.

Soils and permafrost: The potential for fire to act as a mechanism to destabilize permafrost depends on the influence of fire severity on organic layer thickness. Three important factors include: 1) the amount of remaining organic matter following fire, 2) the proportional consumption of the organic layer, and 3) the age of the organic layer that is consumed by fire. We hypothesize that severe fires that consume a larger proportion of the soil organic layer have the greatest potential for permafrost destabilization, but that this effect will be evident only in forest stands where such deep burning has not occurred as part of the recent fire cycle. To identify the links between deep soil combustion and permafrost degradation, we will measure the remaining organic matter, the proportion of soil organic matter combusted, and the radiocarbon age of the remaining soil organic layer in 25 burned sites within the network described above. Particular sites will be selected that span a range from shallow to deep burning based on previous measurements (Boby et al. in press). The first two measurements, made with the adventitious root method (Boby et al. in press) at 10 random points per site, can tell us whether a particular site has experienced deep burning in the most recent fire, and what the original soil organic layer thickness was pre-fire. The radiocarbon age of the remaining soil organic layer can identify the amount of time required to accumulate the combusted organic soil (Trumbore 2000). Five soil cores will be sampled from each of the sites and moss macrofossils will be removed under a dissecting scope from the top layer exposed after the fire. Cellulose is extracted from the macrofossils (Gaudinski et al. 2005) and the sample is converted to graphite by reduction with an iron catalyst in a hydrogen atmosphere (Vogel 1992). The 14C content of the graphite prepared at the University of Florida (UF) is then measured at the University of California, Irvine accelerator mass spectrometer facility to determine the age of the organic matter (Southon et al. 2004). A subsample will be analyzed for δ13C at UF on an isotope ratio mass spectrometer to correct the final Δ14C value. In this way we can differentiate forest stands that burned only the surface organic matter that accumulated since the last fire from stands that burned deeply into organic soils accumulated over hundreds to thousands of years. This can be a used as an indicator for identifying forest stands that may have been pushed past a threshold by a recent fire. To address the mechanistic links between deep burning and permafrost degradation, we will use the same sites to monitor the seasonally thawed active layer thickness. In these sites we will install fixed benchmarks into the permafrost and measure rates of ground subsidence of the soil surface with differential GPS, and the active layer thickness over time (Nelson et al. 2004). These measurements, relative to natural variability measured in the paired unburned sites, will link changes in soil organic layer thickness due to fire to rates of permafrost degradation, and will test the hypothesis that permafrost will degrade most substantially in sites with unprecedented fire severity. This dataset will also be used for model validation as part of Objective 2, described in more detail below.

Vegetation: We will use post-fire measurements of vegetation recruitment and growth across different stand types and fire severities to develop a mechanistic understanding of how an altered fire regime affects the initiation of alternate successional trajectories. We will relate measurements of soil organic layer combustion described above to the recruitment success of different dominant tree species. Because tree canopies in mid-successional or mature forests

are dominated by trees that establish within 10 years after fire (Gutsell and Johnson 2002), measures of early tree recruitment provide a strong predictor of future overstory composition and structure (Johnstone et al. 2004, Peters et al. 2005). We will compare post-fire successional trajectories with previous stand composition by describing pre-fire tree composition and species basal area through surveys of standing dead wood (Johnstone and Chapin 2003). These data will be analyzed using flexible statistical models such as regression trees or multivariate regression splines in order to identify thresholds in site conditions, pre-fire vegetation, or fire characteristics that lead to the initiation of alternate successional trajectories via tree seedling establishment (Johnstone et al. 2010). From these statistical models we will be able to predict which parts of the landscape are most likely to undergo a shift in successional trajectory and a state change in the ecosystem feedbacks that promote organic layer re-accumulation after fire. Relationships identified in these statistical analyses can be directly incorporated as probabilistic rules into our numerical models (described in more detail in Objective 2) to provide realistic simulation of tree seedling establishment after fire. This study will also include surveys of understory vegetation establishment in burned and unburned sites and adjacent road corridors to monitor for the establishment of any non-native plant species. These data will provide an initial assessment of factors that may be influencing the spread of invasive plant species in disturbed and undisturbed forest stands in Interior Alaska (Conn et al. 2008). b) Extend this network to include parallel measurements from sites located in recent prescribed fires and fuel treatments. The wildfire site network described in the previous section covers a wide range of ecosystem and fire conditions on the landscape. This is critical for understanding where and when boreal ecosystems may undergo threshold change in the future. Within this framework, we can then assess how human land management activities such as prescribed fire and fuel treatments fall into (or outside of) the spectrum of natural ecological responses to wildfire. This in turn provides answers to where and when human intervention either contributes to or mitigates vulnerability of the boreal landscape to state change.

Study sites: To understand the effects of fire management on the soil organic layer, vegetation, and permafrost, we will study prescribed burns, fuel treatment areas, and containment lines on DoD land. In consultation with the Alaska Fire Service military zone FMO, we will first select approximately 10 stands that burned in separate, prescribed fires during the last 10 years and that did not receive pre-fire fuel treatments such as thinning, hydro-axe mastication of trees and/or shear blading (Defries, Rees pers. comm.). Second, we will select 10 stands that received these types of fuel treatments but are otherwise unburned. Finally, we will select 10 stands where shear blading or hydro-axing were employed to create fire containment lines, but were otherwise unburned. Fire containment lines represent fuel treatments that often expose mineral soil and thus have the most severe impact on the soil organic layer. For both the thinning and the containment line treatments we will pair nearby unmanipulated stands as controls. Overall, our study design will consist of 10 prescribed fire plots, 10 fuel treatment plots that are unburned, 10 fire containment line plots that are unburned, and 20 unburned, untreated control plots paired with the treatment plots. The unburned sites in the larger landscape network will serve as the control for the prescribed fire plots. Occasionally, fuel treatment stands will burn in a subsequent wildfire, but because these are rare cases and the management intention of the fuel treatment is to prevent fire and remain unburned, we do not include this special case in our study design because site availability is likely too low. Plots in prescribed fire and fuel treatment areas will be 30x30 m in size, while fire line plots will be 5x5 m due to the typically smaller spatial scale of this treatment. We will use multivariate statistical analyses to determine how the effects of fire management on organic soils and vegetation compare to the effects of wildfire, and to determine where and how management techniques influence the outcome of fire-permafrost interactions.

Soils and permafrost: We will conduct similar measurements of soil organic layer consumption, the age of the remaining organic matter, and active layer thickness using the same methods as were used in the naturally burned and unburned control stands (see section above) so that responses to human fire management interventions can be directly compared to ecosystem responses to wildfire. Here we will use the sites that vary in fuel treatment and in prescribed fire to determine the impact of these fire management techniques on the soil organic layer and the underlying permafrost.

Vegetation: We will conduct similar measurements of post-disturbance vegetation recovery to those used in naturally burned stands (see section above) so that responses to human fire management interventions can be directly compared to ecosystem responses to wildfire. In addition, we will use root and shoot architecture indicators to estimate modes of plant regeneration (recruitment from seed, resprouting, or residual surviving vegetation) and their representation in the plant community. This information will allow us to link the disturbance attributes of different management interventions with mechanisms of plant regeneration to predict subsequent patterns of vegetation recovery. Vegetation surveys will also be used to explicitly determine whether fire management interventions influence the establishment and spread of non-native species on DoD lands. c) Conduct studies of organic layer re-accumulation and vegetation stand history on an established network of mid-successional boreal ecosystems in Interior Alaska. We will use detailed studies in ecosystems recovering from disturbance to improve our ability to predict how interactions among vegetation composition, ecophysiological traits of plant species, and environmental conditions lead to different trajectories of vegetation recovery following disturbance. Our approach focuses on two aspects of ecosystem resilience: a) climate effects on tree growth that may affect ecosystem vulnerability to state change, and b) feedbacks among plants, microbes, and soil conditions that determine trajectories of organic soil re-accumulation and ecosystem resilience to fire.

Study sites: For this portion of our field research, we will expand our study sites to include an established network of 45 mid-successional stands that includes conifer-dominated, deciduous-dominated, and mixed conifer-deciduous stands recovering from fire in the past 20 to 60 years (Appendix Figure 4). This will allow us to expand our range of investigation to processes that are important not only to the initiation but also to the maintenance of alternate successional trajectories.

Tree ring reconstructions: We will compare pre- and post-fire composition of forest stands to assess shifts in successional trajectory and analyze the climate sensitivity of tree rings from the pre-fire stand to test whether stress signals in tree growth can be used to predict ecosystem vulnerability to a state change in successional trajectory. We will test the hypothesis that drought-stressed forest stands, identified by negative correlations between tree growth and summer temperature (Lloyd and Bunn 2007), are more likely to shift to an alternate successional trajectory after fire than stands that exhibit positive or neutral ring-width responses to temperature. Recent analyses of tree-ring records in Alaska provide evidence of increasing drought stress for boreal trees such as white spruce in both upland and treeline habitats (Barber et al. 2000, Wilmking et al. 2004), a pattern that appears to be widespread in the boreal forest (Lloyd and Bunn 2007). We propose to use tree-ring records from pre-fire trees, regularly obtained for the aging of pre-fire stands, to assess the climate sensitivity of tree radial growth in the pre-fire stand using standard techniques in dendroclimatology (Cook and Kairiukstis 1989). Negative correlations of tree ring widths with growing season temperature and positive correlations with precipitation are indicators of drought stress that may influence the potential for seedling regeneration in the post-fire stand (Johnstone, unpublished manuscript). Because regeneration feedbacks generally act to ensure the self-replacement of boreal forest stands after fire, large changes in the relative species composition of post-fire versus pre-fire stands can be used to identify a loss of stand-scale resilience leading to an alternate successional trajectory (Johnstone and Chapin 2003). Slow rates of wood

decomposition in Alaskan boreal forests permits the identification and sampling of wood from the pre-fire stand well into mid-succession. We will use tree-ring analyses and pre- vs. post-fire stand comparisons in both early- and mid-successional stands to evaluate whether there is evidence for increased climate stress in recent decades leading to a greater frequency of stand type conversions after fire in Alaskan boreal forests. These data will provide a test of whether the climate sensitivity of tree-rings can be used as an ecological indicator of stand vulnerability to a state change in ecosystem type.

Organic soil re-accumulation: The balance between production and decomposition of moss litter is a key control over re-accumulation of the soil organic layer after fire (Figure 2). We will test the hypothesis that moss percent cover and organic soil re-accumulation are negatively related to the percent cover of deciduous canopy tree species by surveying moss abundance and soil organic layer re-accumulation across the 45 mid-successional forest stands. We predict that there will likely be a threshold effect of deciduous tree cover on mosses that will have consequences for soil organic layer re-accumulation: stands that fall below a critical deciduous tree cover value will have abundant moss cover and re-accumulate thick soil organic layers after fire, while stands that fall above will have little moss cover and thin soil organic layers. Surveys of moss composition and percent cover will be compared to previous surveys of forest composition and structure (Mack et al. unpublished data). In a subset of 18 sites selected to represent landscape variation in drainage, we will estimate annual moss productivity by species and by unit area (Mack et al. 2008). In these sites, we will also examine the relationship between deciduous tree cover and decomposition. We will use a common substrate decomposition experiment to test the hypothesis that litter decomposition rate is positively related to deciduous tree cover (Schuur 2001). Again, we predict that there will be a threshold effect of deciduous tree cover relating to environmental controls over decomposition rate, and this threshold will be similar to the moss threshold above; stands that fall below a critical deciduous tree cover value will have abundant mosses and slow litter decomposition rates due to cool, moist soils and low soil nutrient availability, while stands that fall above will have few mosses and warm, dry soils with high nutrient availability. At each site we will also monitor soil moisture with ECH2O probes, soil temperature with thermocouple probes, and soil nutrients with mixed bed ion exchange resin bags (Binkley and Vitousek 1989). In order to better understand the mechanisms that underlie the hypothesized negative relationship between deciduous trees and mosses, we will use a moss and litter transplant experiment to manipulate two key controls over moss production: site drainage and deciduous litterfall. This experiment will be carried out within the Granite Mountain 1954 fire scar located on Fort Greeley and described in Mack et al. (2008). In this fire scar, we will locate experimental plots in three spatially separate pairs of deciduous and black spruce stands. We will harvest replicate shallow soil cores with live mosses in the black spruce stands, and transplant these either back into their site of origin, or into the paired deciduous site. Moss cores in deciduous stands will receive two treatments: ambient litterfall or deciduous litter exclusion. Moss cores in black spruce stands will receive two treatments: ambient litter, or deciduous litter addition. Thus, this experimental design will allow us to determine the effects of stand microclimate on moss growth, the effects of deciduous litter on moss growth, and the interactions between deciduous litter and stand microclimate. In each block of each vegetation type, we will replicate the experiment five times. Response variables include moss species composition and production per species, as well as ambient litterfall and environmental conditions. OBJECTIVE 2) Provide maps of landscape change in response to projected changes in climate, fire regime, and fire management. Our approach towards this objective is to couple together two stand-alone models previously developed by our group in order to be able to accurately forecast how fire regime and fire management will interact with climate change to shape the future structure, function, and distribution of Alaskan boreal ecosystems (Figure 3).

a) Incorporate field data sets on vegetation, soils, and permafrost (in part developed as part of Objective 1) into a model of landscape fire dynamics and into a model of ecosystem structure and function. In this study we will use two models that have focused on describing various aspects of linkages among climate, fire, and ecosystem structure and function. These models will incorporate information about 1) combustion of soil organic layer and permafrost degradation following fire and 2) vegetation succession and the re-accumulation of the soil organic layer following fire being generated as parts of Objective 1.

Fire modeling: The ALaska FRame Based EcoSystem COde (ALFRESCO) was originally developed to simulate the response of subarctic vegetation to a changing climate and disturbance regime (Rupp et al. 2000a, Rupp et al. 2000b). Additional research, focused on boreal forest vegetation dynamics, has emphasized that fire frequency changes–both direct (climate-driven or anthropogenic) and indirect (as a result of vegetation succession and species composition)–strongly influence landscape-level vegetation patterns and associated feedbacks to future fire regime (Rupp et al. 2002, Chapin et al. 2003, Turner et al. 2003). The boreal forest version of ALFRESCO was developed to explore the interactions and feedbacks between fire, climate, and vegetation in Interior Alaska (Rupp et al. 2002, Duffy et al. 2005, Duffy et al. 2007, Rupp et al. 2007) and associated impacts to natural resources (Rupp et al. 2006, Butler et al. 2007). ALFRESCO models the relationship between monthly climate variables and flammability of a given grid cell. Fire-climate rules come from a statistical model (Duffy et al. 2005) explaining 79% of the variability in the logarithm of interannual area burned in Interior Alaska as a function of monthly climate parameters over the period 1950-2005. Fires are stochastically ignited and then burned recursively. The fire routine in ALFRESCO generates patterns of burning on the landscape that are consistent with the frequency, size distribution, and footprint (burned area) that have been observed in the historical record since 1950 (Rupp et al. 2007). ALFRESCO also models the changes in vegetation flammability that occur during succession through a flammability coefficient that varies with vegetation type and stand age (Chapin et al. 2003). The Boreal ALFRESCO version utilizes probabilistic rules for recovery of spruce forests after fire to relate patterns of seedling establishment and alternative trajectories of succession to variations in fire severity.

Ecosystem modeling: The parent version (Balshi et al. 2007) of the dynamic organic soil – dynamic vegetation model version of the Terrestrial Ecosystem Model (DOS-DVM-TEM) was originally developed to simulate the dynamics of the major carbon and nitrogen fluxes and pools in boreal forest ecosystems, and now has been expanded to explicitly model how fire disturbance and post-fire vegetation recovery affects the structure of the soil (DOS), and how different plant functional types compete for light, water, and nitrogen (DVM). TEM has been used to estimate fire emissions and the carbon dynamics of northern high latitude ecosystems across the pan-boreal region (Balshi et al. 2007) and the North American Boreal Region (Balshi et al. 2009). Fire disturbance implemented in these model versions reduced the amount of soil carbon without affecting organic soil thickness and associated changes in the thermal and hydrological

Figure 3. An ecosystem model and a fire model will be coupled to determine the effects of changing climate and fire management on ecosystem dynamics of Alaskan boreal ecosystems.

properties of soil. To help overcome this problem, Yi et al. (2009a) explicitly coupled soil thermal and hydrological processes of TEM so that the model was capable of simulating soil environmental changes, including permafrost, in the context of changing soil organic thickness. There are four components in dynamic organic soil version of TEM: (1) the environmental module (Yi et al. 2009a), the ecological module (Yi et al. submitted), the fire effects module (Yi et al. submitted) and the dynamic organic soil module (Yi et al. 2009b). In DOS-DVM-TEM, the representation of carbon and nitrogen biogeochemical processes and pools in the ecological module has been replaced by the representation in DVM-TEM. The DVM-TEM includes interactions among soil thermal dynamics, multiple vegetation pools (leaf, wood, and roots), and a dynamic vegetation component (TEM-DVM) that includes competition for light and nitrogen among the plant functional types (PFTs) in an ecosystem (Euskirchen et al. 2009). The coupling of DOS-TEM with DVM-TEM now allows the model to consider competition for water, light, and nitrogen in the context of a dynamic organic soil that can be altered by fire disturbance, post-fire vegetation recovery, and changes in permafrost. b) Couple these two stand-alone models so that the influence of a changing climate on permafrost and vegetation can be assessed together with natural and managed changes in the fire regime. While both ALFRESCO and DOS-DVM-TEM models can run in isolation, they need to be merged in a coupled framework to take advantage of the established ability of ALFRESCO to model characteristics of the fire regime and tree seedling establishment, and the ability of DOS-DVM-TEM to model how characteristics of the fire regime and tree seedling establishment influence ecosystem structure and function. The modeling framework we will use in this study consists of a synchronous coupling of the ALFRESCO and DOS-DVM-TEM models (Figure 3) in which each component will exchange spatially explicit 1 km2 maps at an annual time step. For a given year, the spatially explicit fire perimeters will be simulated by ALFRESCO as a function of both climate and vegetation type. The flammability in ALFRESCO will be modified by information provided by DOS-DVM-TEM regarding vegetation biomass and organic matter from the previous year. In addition to simulating area burned, ALFRESCO will simulate burn severity based on information from Objective 1. Using the mechanistic links between burn severity and succession also developed for Objective 1, the successional trajectory for the post-fire vegetation will then be classified. This linkage between burn severity and post-fire successional pathway will be quantitatively characterized and translated into probabilistic rules using statistical models. For each year, once ALFRESCO has simulated the fire perimeters and burn severity patterns within the fires, a spatially explicit map of tree seedling establishment within fire perimeters will be passed to DOS-DVM-TEM, which will simulate biogeochemical processes for that year. At the end of the year, information on vegetation biomass and organic matter thickness will be passed back to ALFRESCO in the form of a spatially explicit map. Information from DOS-DVM-TEM on vegetation biomass in different PFT categories and organic matter thickness will influence patterns of flammability and burn severity, respectively, modeled in ALFRESCO for the subsequent year. c) Evaluate the performance of the coupled model using retrospective statistical datasets of past fire regime and forest structure in Interior Alaska. We will apply the modeling framework to estimate aspects of the historical fire regime and subsequent vegetation recovery from 1860-2010 in Interior Alaska (essentially the Yukon River Drainage Basin west of the Alaska-Canada border) at 1 km2 resolution. This area includes the military land associated with Fort Wainwright and Eielson Air Force Base near the city of Fairbanks, and with Fort Greeley near the city of Delta Junction. It is important to consider the larger domain Interior Alaska in our analysis so that the spatial aspects of fire regime are properly represented with respect to the effects of future fire on military lands. Aspects of the model simulation to be evaluated include area burned, the composition and distribution of vegetation types, stand-age distribution, and the spatial distribution or soil organic matter in Interior Alaska. Area burned in the model simulation will be compared with two data sets, one which spans 1950-2009 from

the Alaska Large Fire Database (Kasischke et al. 2002) and second spanning 1860-1949 from a modified version of a statistical model presented in Duffy et al. (2005). The simulation of the composition and distribution of vegetation types will be compared with the 2001 landcover in Interior Alaska obtained from the National Land Cover Database (NLCD: http://www.mrlc .gov/index.php) and with other landcover datasets after 2005 if they become available to us during the course of the project. A key analysis in this comparison is to evaluate the ratio of coniferous to deciduous vegetation across the landscape, which is an integrated metric of the ability to simulate both fire disturbance and succession dynamics through time. Stand-age distribution will be evaluated with data from available forest inventory surveys in Interior Alaska, with which both Rupp’s and McGuire’s labs have been working. Finally, we will compare the distribution of organic matter near the end of the simulation to that estimated by a spatially explicit empirical model currently being developed by a postdoctoral researcher in McGuire’s lab as part of the USGS project “Assessing the role of deep soil carbon in Interior Alaska: Data, models, and spatial/temporal dynamics”, which is synthesizing all of the available information on soil carbon in Interior Alaska into an accessible data base as a pilot project for the National Soil Carbon Network. d) Project future landscape distribution of vegetation and permafrost using the coupled model in combination with different scenarios of climate change, fire regime, and fire management.

Climate Scenarios: The coupled modeling framework will be applied to downscaled scenarios of future climate from 2011 to 2100 developed by University of Alaska’s Scenarios Network for Alaska Planning (SNAP) program (http://www.snap.uaf.edu/gis-maps). The IPCC AR4 provides output from a suite of GCMs that can be driven by trace gas emission scenarios (Nakicenovic 2000) to produce spatially-explicit representations of climate (IPCC 2007). Individual model scenario performance was evaluated based on model performance for the historical period 1958-2000 (Walsh et al. 2008) and the five best performing models were selected. SNAP scaled down these coarse GCM outputs with a statistical approach using PRISM (www.prism.oregonstate .edu) data, which accounts for variations in slope, aspect, elevation, and coastal proximity coupled with data from nearby weather stations (http://www.snap.uaf.edu/downloads /validating-snap-climate-models). SNAP data are currently available at 2km resolution; for this project we will further downscale these data to 1 km resolution. Using these downscaled GCM scenarios, SNAP now has spatially explicit data of mean monthly temperature and precipitation projections for three different emissions scenarios, including the A2 scenario, which predicts rapid and unchecked increases in anthropogenic greenhouse gas emissions, the B1 scenario, which predicts swift leveling followed by significant decline of emissions, and the A1B scenario, which falls between the other two. In addition, we will use cloud cover data from the Climate Research Unit (CRU; www.ipcc-data.org) for the historical period and develop future data sets of cloud cover for use in the project using appropriate aspects of the SNAP downscaling methodology. These spatially explicit data layers will be used to drive the coupled model framework, which will generate spatially explicit maps of fire, fire severity, vegetation dynamics, ecosystem function, and permafrost for Interior Alaska.

Fire Management Applications: Alaska holds an unprecedented level of public responsibility for its land-based resources—70% of the state is under the jurisdiction of federal and state resource managers. Fire management in Alaska is multifaceted as federal and state land managers seek ways to manage fuels and fire to both improve community safety and meet multiple natural resource objectives. Distribution of information to fire and land managers poses a major challenge to effective fire science delivery. We will investigate the potential influence of specific management scenarios applied on top of the simulations outlined above. One of the mitigation options available to land managers is to modify fire suppression or prescribed fire policies. However, development of effective policies depends upon how the vulnerability of different landscape units (e.g., topography/vegetation cover) to burning changes under different fire regimes. We will use a suite of our prognostic climate change scenarios to

identify influences of fire management strategies (e.g., suppression, prescribed fire) at a landscape-level. We will meet with the collaborating fire and natural resource management groups (Alaska Fire Service military zone and Center for Environmental Management of Military Lands) on this project to identify relevant management scenarios for the purpose of exploring the effects of these scenarios (including potential feedbacks) on future fire regimes and associated permafrost and vegetation responses. These simulations will help inform managers about the potential future impacts of climate change on fire regimes at landscape scales. Specifically, the simulations will provide estimates (including measures of uncertainty) of the likely success or failure of different landscape-level mitigation strategies under scenarios of climate change. Success/failure evaluations will be assessed through paired comparisons between no-management simulations and simulations with different management scenarios. 3) Milestones We propose a 5-year timeline to develop a conceptual and mechanistic model of resilience and vulnerability in Alaskan boreal forest and to integrate findings into a decision support framework for management of DoD lands.

Project timeline and milestones

Obj. Milestone 2011 2012 2013 2014 2015

1 Expand and monitor wildfire network X X X

1 Develop and monitor fire management network X X X

1 Monitor and experiment in mid-successional network X X

2 Parameterize individual models with data X X

2 Couple DOS-DVM-TEM and ALFRESCO models X

2 Evaluate and test coupled models with data X X

2 Forecast future landscape distribution with coupled models X X

2 Develop dynamic modeling tools for land managers X

1,2 Transfer technology to land managers via workshops X X

1,2 Submit technical reports on project progress X X X X

1,2 Submit scientific manuscripts X X X X

1,2 Archive datasets on BNZ LTER website X X

RESEARCH TEAM

These researchers have previously worked together under the auspices of the Bonanza Creek Long Term Ecological Research (BNZ-LTER) program funded by the NSF, where each of the participants on this project is either a Principle Investigator or Senior Scientist. Many of the overarching questions of the BNZ-LTER address related issues of change in boreal ecological communities in response to a changing climate. This multi-disciplinary team consists of:

Dr. Edward Schuur. Associate Professor, Department of Biology, University of Florida. Dr. Schuur is an ecosystem ecologist with research focused on carbon cycling in high latitude ecosystems. Schuur is budgeted for 1 summer month per year for supervising the UF postdoctoral researcher and conducting fieldwork, and is also committed to an additional month of the academic year (non-budgeted) for data analysis and project management.

Dr. Jill Johnstone. Assistant Professor, Department of Biology, University of Saskatchewan, Canada. Dr Johnstone is a plant community ecologist with a particular interest in understanding multiple successional trajectories in response to disturbance in North American boreal forests. Dr. Johnstone commits 1.5 months/year to work on this project, including 2-3 weeks of field research/year in Years 1-3 and time for data analysis, integration, and project management. She will also supervise three graduate students (1 Ph.D and 2 M.Sc.) who will work full-time on the project.

Dr. Michelle Mack. Associate Professor, Department of Biology, University of Florida. Dr. Mack is an ecosystem ecologist who is interested in the interface between plant community

composition and ecosystem function. Mack is budgeted for 1 summer month per year for interfacing with UF postdoctoral researcher and research crew conducting fieldwork. She is also committed to an additional month of the academic year (non-budgeted) for data analysis and project management.

Dr. A. David McGuire. Professor, Institute of Arctic Biology, University of Alaska, Fairbanks. Dr. McGuire is an ecologist with an expertise in biogeochemical modeling of high latitude ecosystems. He is the director of the Spatial Ecology lab in the Institute of Arctic Biology, and is an employee of the USGS. Dr. McGuire will commit 2 months in each year of the project (non-budgeted) to work on model development, model testing, model application, model coupling and technical outreach.

Dr. Scott Rupp. Associate Professor, School of Natural Resource and Agricultural Sciences, University of Alaska, Fairbanks. Dr. Rupp is an ecologist with an expertise in numerical modeling of ecological systems. He is also the director of the Scenarios Network for Alaska & Arctic Planning (SNAP), which provides timely access to management-relevant scenarios of future conditions in Alaska. Dr. Rupp will commit 1 month in each year of the project to work on model development, model testing, model application, model coupling and technical outreach.

Dr. Eugenie Euskirchen. Research Assistant Professor. Dr. Euskirchen is an ecologist with expertise in modeling vegetation dynamics of high latitude ecosystems. She is budgeted for 1 month of research work in Years 1-5 of the project; she will assist the postdoctoral researchers with model development and participate in the information transfer workshops.

Other Project Members: At UF, a full-time postdoctoral researcher (Years 1-3) will be responsible for the bulk of the fieldwork including management of the seasonal field crew (3 individuals for 3 months each summer for Years 1-3). This researcher will interface with the project PIs to ensure coordination of all fieldwork, and will also work closely with the UF laboratory technician to ensure that samples collected during fieldwork are processed in the laboratory. This technician is budgeted for 6 months/year in Years 1-3 to assist the postdoctoral researcher, and for 3 months/year in the final two project years to coordinate the project as the postdoctoral researcher (and fieldwork) is finished. These project members will coordinate closely with the personnel at the University of Saskatchewan conducting fieldwork including 2 Master’s level students and one PhD level student. The PhD student at U of S will focus on field research integrating vegetation and soil responses to fire. Each of the U of S Master’s students will take on one element of the field research for their graduate theses, such as tree-ring analyses, surveys of invasive species, or assessment of successional trajectories. U of S will also contribute 6 months/year of undergraduate research assistants to support field and laboratory work on the study in Years 1-3, plus an additional 3 months of support for tree-ring analyses in Year 2 and for data integration with models in Year 4. At the University of Alaska Fairbanks, one of the postdoctoral researchers will be responsible for model development involving DOS-DVM-TEM (supervised by McGuire) and the other will be responsible for model development involving ALFRESCO (supervised by Rupp). The first postdoctoral researcher will also interface closely with Dr. Euskirchen who has been extending the capabilities of the DOS-DVM-TEM model. Both postdoctoral researchers will be involved in complementary aspects of coupling, testing, and applying the two models to make projections for future climate and land management scenarios.

COOPERATIVE DEVELOPMENT We have established collaborative interest in this research with the Center for Environmental Management of Military Lands (contact point – D. Rees) that has ongoing studies monitoring post-disturbance vegetation establishment patterns in several prescribed fire and fuel treatment sites within the Ft. Wainwright DoD lands complex. In addition, Co-PI Rupp has ongoing research collaborations with the Alaska Fire Service (AFS) specifically addressing fuel treatment effects on fire behavior (a Joint Fire Science Program funded project) and has identified

common interests in this proposed research with the AFS military zone FMO (contact point – T. Defries). The Scenarios Network for Alaska & Arctic Planning located at the University of Alaska Fairbanks (directed by Co-PI Rupp) will directly support this project through technical expertise associated with the SNAP climate scenarios. Technical expertise will include computer programming, data management, GIS and remote sensing, and statistical analysis. SNAP will also provide infrastructure required for project data management, analysis, and technological transfer (estimated in-kind value of expertise and service is $40K per year). The proposed project is also aligned in parallel with the Bonanza Creek Long Term Ecological Research (BNZ LTER) Program located at the University of Alaska Fairbanks. The BNZ LTER Program is funded by the National Science Foundation and was established in 1987 to examine the interactions between climate and disturbance, and their effects on ecosystem processes in the boreal forests of Interior Alaska (http://www.lter.uaf.edu/). The current phase (2007-2010) of the BNZ LTER program is studying the dynamics of change in Alaska’s Boreal Forests with a focus on the resilience and vulnerability in response to climate warming. McGuire, Schuur, Mack, Rupp, and Johnstone are investigators in the current phase of the BNZ LTER Program, which has collectively provided $311,527 ($77,881 per year) to support their research activities in the program between 2007 and 2010. All senior personnel on this proposal are involved in the BNZ LTER renewal proposal currently under review that would run from 2011 – 2016 and would collectively provide approximately $100,000 per year to support the ongoing research programs of these investigators.

TRANSITION PLAN

The outcome of this research will be transferred as the following deliverables: >10 peer-reviewed scientific publications containing the detailed findings of the research, an interactive workbook and mapping tool to support management decision making on Interior Alaska lands, and two information transfer workshops designed to make results accessible to Alaska land managers. Our rubric for the number of scientific publications that this project will produce is based on our collective past record of productivity and consists of the following estimates: one first authored paper per PI for the project (5 publications), one first authored paper for each post doctoral researcher (5 publications), one first authored publication per masters student (2 publications) and two first authored papers for the Ph.D. student (2 publications). In addition, we will organize two technical transfer workshops at the end of Years 3 and 5 in order to interact with fire managers in Interior Alaska, particularly those involved with fire management on Department of Defense lands in Interior Alaska. In the first workshop, which will be held at the end of Year 3, we will communicate the results of the project to date and present plans for the no-management and management model simulations we envision conducting. The workshop will be primarily focused on helping us to modify the details of the planned management simulations so that the results of the simulations will be most useful to the fire managers. The second workshop will occur late in the fifth year of the project to present and provide an interactive workbook and mapping tool designed make the results of the no-management and management simulations easily accessible and useful to the fire managers in Interior Alaska. These workshops and the related technological transfer of proposed deliverables will be coordinated through the Alaska Fire Science Consortium. This consortium is funded by the Joint Fire Science Program (Rupp is Consortium Co-PI) with a mission to: 1) coordinate current science delivery efforts; 2) create a formal outreach mechanism for two-way communication between fire scientists and diverse fire and land managers; 3) provide an organized, centralized arena for effectively delivering fire science information to managers; and 4) work to ensure that the science delivery and outreach mechanisms are both practical and readily implemented in the field. The consortium includes representatives from all sectors of the fire research and management communities and technological service providers (those providing web, mapping and teleconferencing/webinar capabilities). Integrating the delivery of results from this project with the Fire Science Consortium will ensure that our results reach the broader fire management community in Alaska as well as DoD managers.

REFERENCES

Balshi, M. S., A. D. McGuire, P. Duffy, M. Flannigan, D. W. Kicklighter, and J. Melillo. 2009. Vulnerability of carbon storage in North American boreal forests to wildfires during the 21st century. Global Change Biology 15:1491-1510.

Balshi, M. S., A. D. McGuire, Q. Zhuang, J. Melillo, D. W. Kicklighter, E. Kasischke, C. Wirth, M. Flannigan, J. Harden, J. S. Clein, T. J. Burnside, J. McAllister, W. A. Kurz, M. Apps, and A. Shvidenko. 2007. The role of historical fire disturbance in the carbon dynamics of the pan-boreal region: A process-based analysis. Journal of Geophysical Research-Biogeosciences 112:-.

Barber, V. A., G. P. Juday, and B. P. Finney. 2000. Reduced growth of Alaskan white spruce in the twentieth century from temperature-induced drought stress. Nature 405:668-673.

Binkley, D., and P. M. Vitousek. 1989. Soil nutrient availability. Page (in press) in R. W. Pearcy, J. R. Ehleringer, H. A. Mooney, and P. Rundel, editors. Physiological Plant Ecology: Field Methods and Instrumentation. Chapman and Hall, London.

Boby, L., E. A. G. Schuur, M. C. Mack, D. Verbyla, and J. F. Johnstone. in press. Quantifying fire severity, carbon, and nitrogen emissions in Alaska's boreal forest: the adventitious root method. Ecological Applications.

Brown, J., O. J. F. Jr., J. A. Heginbottom, and E. S. Melnikov. 1998. Circum-Arctic map of permafrost and ground-ice conditions. National Snow and Ice Data Center/World Data Center for Glaciology, Boulder, CO.

Butler, L. G., K. Kielland, T. S. Rupp, and T. A. Hanley. 2007. Interactive controls of herbivory and fluvial dynamics on landscape vegetation patterns on the Tanana River floodplain, interior Alaska. Journal of Biogeography 34:1622-1631.

Chapin, F. S., A. D. McGuire, R. W. Ruess, D. A. Walker, R. D. Boone, M. E. Edwards, B. P. Finney, L. D. Hinzman, J. B. Jones, G. P. Juday, E. S. Kasischke, K. Kielland, A. H. Lloyd, M. W. Oswood, C. L. Ping, E. Rexstad, V. E. Romanovsky, J. P. Schimel, E. B. Sparrow, B. Sveinbjornsson, D. W. Valentine, K. VanCleve, D. L. Verbyla, L. A. Viereck, R. A. Werner, T. L. Wurtz, and J. Yarie. 2006. Summary and Synthesis:Past and future changes in the Alaskan boreal forest. Oxford University Press, New York.

Chapin, F. S., S. F. Trainor, O. Huntington, A. L. Lovecraft, E. Zavaleta, D. C. Natcher, A. D. McGuire, J. L. Nelson, L. Ray, M. Calef, N. Fresco, H. Huntington, T. S. Rupp, L. Dewilde, and R. L. Naylor. 2008. Increasing wildfire in Alaska's boreal forest: Pathways to potential solutions of a wicked problem. Bioscience 58:531-540.

Chapin, F. S. I., T. S. Rupp, A. Starfield, L. DeWilde, E. Zavaleta, N. Fresco, J. Henkelman, and A. McGuire. 2003. Planning for resilience: Modeling change in human-fire interactions in the Alaskan boreal forest. Frontiers in Ecology and the Environment 1:255-261.

Conn, J. S., K. L. Beattie, M. A. Shephard, M. L. Carlson, I. Lapina, M. Hebert, R. Gronquist, R. Densmore, and M. Rasy. 2008. Alaska Melilotus invasions: Distribution, origin, and susceptibility of plant communities. Arctic Antarctic and Alpine Research 40:298-308.

Cook, E., and L. Kairiukstis. 1989. Methods of Dendrochronology: Applications in the Environmental Sciences. Kluwer Academic, Dordrecht.

Cumming, S. G. 2001. Forest type and wildfire in the alberta boreal mixedwood: What do fires burn? Ecological Applications 11:97-110.

Dissing, D., and D. L. Verbyla. 2003. Spatial patterns of lightning strikes in interior Alaska and their relations to elevation and vegetation. Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere 33:770-782.

Duffy, P. A., J. Epting, J. M. Graham, T. S. Rupp, and A. D. McGuire. 2007. Analysis of Alaskan burn severity patterns using remotely sensed data. International Journal of Wildland Fire 16:277-284.

Duffy, P. A., J. E. Walsh, J. M. Graham, D. H. Mann, and T. S. Rupp. 2005. Impacts of large-scale atmospheric-ocean variability on Alaskan fire season severity. Ecological Applications 15:1317-1330.

Euskirchen, E. S., A. D. McGuire, F. S. Chapin, S. Yi, and C. C. Thompson. 2009. Changes in vegetation in northern Alaska under scenarios of climate change, 2003-2100: implications for climate feedbacks. Ecological Applications 19:1022-1043.

Gaudinski, J. B., T. E. Dawson, S. Quideau, E. A. G. Schuur, J. S. Roden, S. E. Trumbore, D. R. Sandquist, S. W. Oh, and R. E. Wasylishen. 2005. Comparative analysis of cellulose preparation techniques for use with C-13, C-14, and O-18 isotopic measurements. Analytical Chemistry 77:7212-7224.

Gutsell, S. L., and E. A. Johnson. 2002. Accurately ageing trees and examining their height-growth rates: implications for interpreting forest dynamics. Journal of Ecology 90:153-166.

Higuera, P. E., L. B. Brubaker, P. M. Anderson, F. S. Hu, and T. A. Brown. 2009. Vegetation mediated the impacts of postglacial climate change on fire regimes in the south-central Brooks Range, Alaska. Ecological Monographs 79:201-219.

Hinzman, L. D., N. D. Bettez, W. R. Bolton, F. S. Chapin, M. B. Dyurgerov, C. L. Fastie, B. Griffith, R. D. Hollister, A. Hope, H. P. Huntington, A. M. Jensen, G. J. Jia, T. Jorgenson, D. L. Kane, D. R. Klein, G. Kofinas, A. H. Lynch, A. H. Lloyd, A. D. McGuire, F. E. Nelson, W. C. Oechel, T. E. Osterkamp, C. H. Racine, V. E. Romanovsky, R. S. Stone, D. A. Stow, M. Sturm, C. E. Tweedie, G. L. Vourlitis, M. D. Walker, D. A. Walker, P. J. Webber, J. M. Welker, K. Winker, and K. Yoshikawa. 2005. Evidence and implications of recent climate change in northern Alaska and other arctic regions. Climatic Change 72:251-298.

IPCC. 2007. Climate Change 2007: The Physical Science Basis. The Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Johnstone, J., and F. Chapin. 2006. Effects of soil burn severity on post-fire tree recruitment in boreal forest. Ecosystems 9:14-31.

Johnstone, J. F., F. Chapin, T. Hollingsworth, M. Mack, V. Romanovsky, and M. Turetsky. in press. Fire, climate change, and forest resilience in interior Alaska. Canadian Journal of Forest Research.

Johnstone, J. F., and F. S. Chapin. 2003. Non-equilibrium succession dynamics indicate continued northern migration of lodgepole pine. Global Change Biology 9:1401-1409.

Johnstone, J. F., F. S. Chapin, J. Foote, S. Kemmett, K. Price, and L. Viereck. 2004. Decadal observations of tree regeneration following fire in boreal forests. Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere 34:267-273.

Johnstone, J. F., T. Hollingsworth, F. Chapin III, and M. Mack. 2010. Changes in fire regime break the legacy lock on successional trajectories in Alaskan boreal forest. Global Change Biology 16:1281-1295.

Kasischke, E. S., and M. R. Turetsky. 2006. Recent changes in the fire regime across the North American boreal region-Spatial and temporal patterns of burning across Canada and Alaska (vol 33, art no L09703, 2006). Geophysical Research Letters 33:-.

Kasischke, E. S., D. Williams, and D. Barry. 2002. Analysis of the patterns of large fires in the boreal forest region of Alaska. International Journal of Wildland Fire 11:131-144.

Lloyd, A. H., and A. Bunn. 2007. Responses of the circumpolar boreal forest to 20th century climate variability. Environmental Research Letters.

Mack, M. C., K. K. Treseder, K. L. Manies, J. W. Harden, E. A. G. Schuur, J. G. Vogel, J. T. Randerson, and F. S. Chapin. 2008. Recovery of aboveground plant biomass and

productivity after fire in mesic and dry black spruce forests of interior alaska. Ecosystems 11:209-225.

Nakicenovic, N. 2000. Greenhouse gas emissions scenarios. Technological Forecasting and Social Change 65:149-166.

Nelson, F. E., N. I. Shiklomanov, H. H. Christiansen, and K. M. Hinkel. 2004. The circumpolar-active-layer-monitoring (CALM) Workshop: Introduction. Permafrost and Periglacial Processes 15:99-101.

Osterkamp, T. E., T. Zhang, and V. E. Romanovsky. 1994. Thermal regime of permafrost in Alaska and predicted global warming. Journal of Cold Regions Engineering 4:38-42.

Peters, V. S., S. E. MacDonald, and M. R. T. Dale. 2005. The interaction between masting and fire is key to white spruce regeneration. Ecology 86:1744-1750.

Randerson, J. T., H. Liu, M. G. Flanner, S. D. Chambers, Y. Jin, P. G. Hess, G. Pfister, M. C. Mack, K. K. Treseder, L. R. Welp, F. S. Chapin, J. W. Harden, M. L. Goulden, E. Lyons, J. C. Neff, E. A. G. Schuur, and C. S. Zender. 2006. The impact of boreal forest fire on climate warming. Science 314:1130-1132.

Rupp, T. S., F. S. Chapin, and A. M. Starfield. 2000a. Response of subarctic vegetation to transient climatic change on the Seward Peninsula in north-west Alaska. Global Change Biology 6:541-555.

Rupp, T. S., X. Chen, M. Olson, and A. D. McGuire. 2007. Sensitivity of simulated boreal fire dynamics to uncertainties in climate drivers. Earth Interactions 11:-.

Rupp, T. S., M. Olson, L. G. Adams, B. W. Dale, K. Joly, J. Henkelman, W. B. Collins, and A. M. Starfield. 2006. Simulating the influences of various fire regimes on caribou winter habitat. Ecological Applications 16:1730-1743.

Rupp, T. S., A. M. Starfield, and F. S. Chapin. 2000b. A frame-based spatially explicit model of subarctic vegetation response to climatic change: comparison with a point model. Landscape Ecology 15:383-400.

Rupp, T. S., A. M. Starfield, F. S. Chapin, and P. Duffy. 2002. Modeling the impact of black spruce on the fire regime of Alaskan boreal forest. Climatic Change 55:213-233.

Schuur, E. A. G. 2001. The effect of water on decomposition dynamics in mesic to wet Hawaiian montane forests. Ecosystems 4:259-273.

Schuur, E. A. G., J. Bockheim, J. G. Canadell, E. Euskirchen, C. B. Field, S. V. Goryachkin, S. Hagemann, P. Kuhry, P. M. Lafleur, H. Lee, G. Mazhitova, F. E. Nelson, A. Rinke, V. E. Romanovsky, N. Shiklomanov, C. Tarnocai, S. Venevsky, J. G. Vogel, and S. A. Zimov. 2008. Vulnerability of permafrost carbon to climate change: Implications for the global carbon cycle. Bioscience 58:701-714.

Schuur, E. A. G., J. G. Vogel, K. G. Crummer, H. Lee, J. O. Sickman, and T. E. Osterkamp. 2009. The effect of permafrost thaw on old carbon release and net carbon exchange from tundra. Nature 459:556-559.

Shur, Y. L., and M. T. Jorgenson. 2007. Patterns of permafrost formation and degradation in relation to climate and ecosystems. Permafrost and Periglacial Processes 18:7-19.

Southon, J., G. Santos, K. Druffel-Rodriguez, E. Druffel, S. Trumbore, X. M. Xu, S. Griffin, S. Ali, and M. Mazon. 2004. The Keck Carbon Cycle AMS laboratory, University of California, Irvine: Initial operation and a background surprise. Radiocarbon 46:41-49.

Trumbore, S. 2000. Age of soil organic matter and soil respiration: radiocarbon constraints on belowground C dynamics. Ecological Applications 10:399-411.

Turetsky, M., K. Wieder, L. Halsey, and D. Vitt. 2002. Current disturbance and the diminishing peatland carbon sink. Geophysical Research Letters 29:-.

Turner, M. G., S. L. Collins, A. E. Lugo, J. J. Magnuson, T. S. Rupp, and F. J. Swanson. 2003. Disturbance dynamics and ecological response: The contribution of long-term ecological research. Bioscience 53:46-56.

Van Cleve, K., F. S. Chapin, III, C. T. Dryness, and L. A. Viereck. 1991. Element cycling in taiga forest: State-factor control. Bioscience 41:78-88.

Verbyla, D., and R. Lord. 2008. Estimating post-fire organic soil depth in the Alaskan boreal forest using the Normalized Burn Ratio. International Journal of Remote Sensing 29:3845-3853.

Vogel, J. S. 1992. A rapid method for preparation of biomedical targets for AMS. Radiocarbon 34:344-350.

Walsh, J. E., W. L. Chapman, V. Romanovsky, J. H. Christensen, and M. Stendel. 2008. Global climate model performance over Alaska and Greenland. . Journal of Climate 21:6156-6174.

Wilmking, M., G. P. Juday, V. A. Barber, and H. S. J. Zald. 2004. Recent climate warming forces contrasting growth responses of white spruce at treeline in Alaska through temperature thresholds. Global Change Biology 10:1724-1736.

Yi, S., A. McGuire, E. Kasischke, J. Harden, K. Manies, M. Mack, and M. Turetsky. submitted. A dynamic organic soil biogeochemical model for analyzing carbon responses of black spruce forests in Interior Alaska. Journal of Geophysical Research-Biogeosciences.

Yi, S., A. D. McGuire, J. Harden, E. Kasischke, K. Manies, L. Hinzman, A. Liljedahl, J. Randerson, H. Liu, V. Romanovsky, S. Marchenko, and Y. Kim. 2009a. Interactions between soil thermal and hydrological dynamics in the response of Alaska ecosystems to fire disturbance. Journal of Geophysical Research – Biogeosciences 114:G02015.

Yi, S. H., K. Manies, J. Harden, and A. D. McGuire. 2009b. Characteristics of organic soil in black spruce forests: Implications for the application of land surface and ecosystem models in cold regions. Geophysical Research Letters 36:-.

Yoshikawa, K., W. R. Bolton, V. E. Romanovsky, M. Fukuda, and L. D. Hinzman. 2002. Impacts of wildfire on the permafrost in the boreal forests of Interior Alaska. Journal of Geophysical Research-Atmospheres 108:-.