The objective of the research presented in this dissertation was to better understand the factors controlling the present and potential future distribution of arctic vegetation. The analysis compares the Circumpolar Arctic Vegetation Map (CAVM) with circumpolar data sets of environmental characteristics. Geographical information system (GIS) software was used to overlay the CAVM with a satellite index of vegetation (normalized difference vegetation index, NDVI) and environmental factors that are most important in controlling the distribution of arctic vegetation, including summer temperature, landscape age, precipitation, snow cover, substrate chemistry (pH and salinity), landscape type, elevation, permafrost characteristics, and distance to sea. Boosted regression tree analysis was used to determine the relative importance of different environmental characteristics for different vegetation types and for different regions. Results of this research include maps, charts and tables that summarize and display the spatial characteristics of arctic vegetation. The data for arctic land surface temperature and landscape age are especially important new resources for researchers. These results are available electronically, not only as summary data, but also as GIS data layers with a spatial context (www.arcticatlas.org). The results emphasize the value and reliability of NDVI for studying arctic vegetation. The relationship between NDVI and summer temperatures across the circumpolar arctic was similar to the correlated increases in NDVI and temperature seen over the time period of satellite records. Summaries of arctic biomass based on NDVI match those based on extrapolation from ground samples. The boosted regression tree analysis described ecological niches of arctic vegetation types, demonstrating the importance of summer temperatures and landscape age in controlling the distribution of arctic vegetation. As the world continues to focus on the Arctic as an area undergoing accelerated warming due to global climate change, results presented here from spatially explicit analysis of existing arctic vegetation and environmental characteristics can be used to better understand plant distribution patterns, evaluate change in the vegetation, and calibrate models of arctic vegetation and animal habitat.

Harvesting wild berries, firewood, and other non-timber forest products (NTFPs) from the boreal forest in Interior Alaska is a common activity amongst local residents. NTFPs are harvested for personal use, subsistence, and commercial purposes. While these activities contribute to informal household economies and livelihoods, harvest of NTFPs are not well documented in Alaska. Availability of these ecosystem services may be altered under changing management and climate regimes. This interdisciplinary dissertation takes a look at the activities and impacts of current NTFP harvesting practices. Survey results from a forest use survey provide insight into harvest activity in the Tanana Valley. Wild blueberries (38.5% of households with mean harvested amount of 7.7 quarts) and firewood (25.0% of households with a mean harvest amount of 4.7 cords) were reported harvested with greatest frequency, and harvesting activities were mostly concentrated around larger population centers. Interviews were conducted with personal use and subsistence NTFP harvesters from Interior Alaska. Participants enjoy harvesting from the forest, and that the importance of harvesting is a combination of both the intangible benefits from the activity and the tangible harvested items. Harvested NTFPs were seen as high-quality products that were otherwise unavailable or inaccessible. Birch syrup is a commercially available NTFP produced in Alaska by a small number of companies. Similar to maple syrup, producing birch syrup is a labor intensive process with marginal profits. Interviews were conducted with workers in the Alaskan birch syrup industry, who reported that they were seeking an alternative to the traditional employment. The effects from mechanical damage from tapping for spring sap on birch's vigor are of concern to birch syrup producers and natural resource managers. This study compared the annual increment growth of Alaskan birch trees, Betula neoalaskana, between tapped and untapped trees. No significant difference was detected from tapping, but annual variability in growth was strongly significant. A temperature index accounted for nearly two-thirds of the annual variability. Pairing this index with two climate scenarios, birch growth was extended out through the 21st century. As temperatures rise, birch in Interior Alaska are projected to face a critical threshold, which may limit or extinguish their ability to sustain growth and yield a sustainable sap resource. Integrating the survey, interview, and dendroclimatological data provides a richer picture of how NTFP harvesters actively use the forest and about the benefits derived. These findings can assist resource managers in balancing these needs with those of other forest uses on public land.

Yellow-cedar is a very long-lived, commercially important tree species found along the coasts of Southeast Alaska and also in small populations in Prince William Sound. However, this is the first study of the tree's annual ring growth patterns in the region. Tree cores were collected from over 400 trees across a large latitudinal gradient and cross-dated using standard dendrochronological techniques. Radial tree-ring growth was measured and compared to reconstructed weather station data to gain a better understanding of the climatic conditions favoring yellow-cedar growth. We found consistent, significant positive correlations between ring widths and mean monthly temperatures in August, previous January, and previous December, and negative relationships with May and December precipitation. Climate indices we created using these variables explain approximately 25% of growth variability in five distinct yellow-cedar populations. Long-term growth patterns in tree populations going back three centuries were similar across all sites, specifically the sustained below mean growth during the 1800s. Yellow-cedar at the northern limits of its distribution shows a common growth signal which may indicate the influence of larger pressure anomalies, such as EI Nino-Southern Oscillation (ENSO), on the climate factors affecting the trees.

Within the forest management community, diversity is often considered as simply a list of species present at a location. In this study, diversity refers to species richness and evenness and takes into account vegetation structure (i.e. size, density, and complexity) that characterize a given forest ecosystem and can typically be measured using existing forest inventories. Within interior Alaska the largest forest inventories are the Cooperative Alaska Forest Inventory and the Wainwright Forest Inventory. The limited distribution of these inventories constrains the predictions that can be made. In this thesis, I examine forest diversity in three distinct frameworks; Recruitment, Patterns, and Production. In Chapter 1, I explore forest management decisions that may shape forest diversity and its role and impacts in the boreal forest. In Chapter 2, I evaluate and map the relationships between recruitment and species and tree size diversity using a geospatial approach. My results show a consistent positive relationship between recruitment and species diversity and a general negative relationship between recruitment and tree size diversity, indicating a tradeoff between species diversity and tree size diversity in their effects on recruitment. In Chapter 3, I modeled and mapped current and possible future forest diversity patterns within the boreal forest of Alaska using machine learning. The results indicate that the geographic patterns of the two diversity measures differ greatly for both current conditions and future scenarios and that these are more strongly influenced by human impacts than by ecological factors. In Chapter 4, I developed a method for mapping and predicting forest biomass for the boreal forest of interior Alaska using three different machine-learning techniques. I developed first time high resolution prediction maps at a 1 km2 pixel size for aboveground woody biomass. My results indicate that the geographic patterns of biomass are strongly influenced by the tree size class diversity of a given stand. Finally, in Chapter 5, I argue that the methods and results developed for this dissertation can aid in our understanding of forest ecology and forest management decisions within the boreal region.

Twice-monthly AVHRR-derived NDVI were used to estimate growing season length across Alaska, north of the Alaska Range. An algorithm, based on the ratio of NDVI to annual maximum NDVI for each pixel, was used to represent percent of maximum greenness for each composite period. Greenup and senescence commenced when NDVI values rose above and fell below a selected percent of maximum greenness. Six different percent of maximum greenness threshholds, ranging from 25 to 50 percent, were evaluated. This algorithm eliminates complications of landscape-specific NDVI thresholds and year-to-year variability. The algorithm was tested against 1) air temperature data from 23 weather stations located in northern Alaska from 1991 to 1997, 2) observed greenup at two sites in Fairbanks, Alaska, from 1991 to 1997, and 3)phenology observations on the Seward Peninsula during the 1996-1997 growing seasons. Best results were obtained with NDVI values at 30% and 40% of maximum NDVI.

Boreal forests store large quantities of carbon (C) and currently act as atmospheric C sinks; however, predicted increases in temperature and fire frequency may change the boreal forest from a net C sink to a net source. This study evaluates the response of organic soil C and nitrogen (N) mineralization, and the bioavailability of C and N to burning in non-permafrost upland black spruce stands in Interior Alaska. Two years after an experimental wildfire, burned soils were warmer than control soils at all depths measured, and decay of common substrates was greater in the burned than in the control soils. Burned soils had higher concentrations of total C, lignin, N, and mineral N, and lower concentrations of dissolved organic carbon (DOC) and soluble organic matter. However, apparent differences in organic matter quality did not correlate well with respiration metrics. In laboratory incubations, burned soils respired less than control soils, and this difference was entirely due to differences on the first day of the incubation. Mean Q₁₀ values ranged from 2.1 to 2.5 and were greater in the burned soils than in the control soils.

Black spruce, Picea mariana (Mill.) B.S.P., is largely overlooked in Alaska because of its small size and slow growth. Growth and yield information is therefore limited or nonexistent. Presented here are the first polymorphic site index (height-age) curves and height-diameter functions for predicting height and volume for Alaska black spruce. Models are accurate for trees up to 50 feet in height and 8 inches DBH. Predicted stem volumes range from 0.006 ft3 to 21.8 ft3 for trees between 0.5 and 11.5 inches DBH Sampled tree dimensions range from 5.5 to 78.0 feet tall and from 0.4 to 11.0 inches DBH. Sampled breast-height ages range from 49 to 257 years; average age-to-breast-height is 26 years. This research, although limited, also characterizes general stand-level structure and community composition for Alaska black spruce. 60 Permanent Sample Plots (PSPs) representing 20 stands were established throughout the Tanana Valley, with stand inventory conducted according to a consistent protocol. Stand densities range from 137 to 2,907 trees per acre; stand volumes ranged from 8 to 2,507 ft3 per acre. Stand density index values range from 6 to 453. Periodic remeasurement of PSPs will yield valuable information about stand evolution and community type change.

The northern perhumid North American Pacific coastal temperate rainforest (NCTR) extends along the coastal margin of British Columbia and southeast Alaska and has some of the densest carbon stocks in the world. Northern temperate ecosystems such as the NCTR play an important role in the global balance of carbon flows between atmospheric and terrestrial pools. However, there is little information on key components of the forest carbon budget in this region. Specifically, the large pool of soluble carbon that is transferred from soils via streamwater as dissolved organic carbon (DOC) certainly plays a role in the total carbon balance in wet forests such as the NCTR. In order to address this information gap, I applied the concept of hydropedology to define functional landscape units based on soil type to quantify soil carbon fluxes and apply these estimates to a conceptual model for determining the carbon balance in three NCTR watersheds. The strong hydrologic gradient among ecosystems served as a template for constructing a conceptual design and approach for constraining carbon budget estimates in the watersheds. Replicated hydropedologic units were identified in three classes: sloping bogs, forested wetlands, and uplands. Estimates of annual soil respiration and DOC fluxes from the hydropedologic types were obtained through seasonal measurements combined with temperature-dependent models. Soil respiration fluxes varied significantly across the hydrologic gradient where soil respiration was 78, 178, and 235 g CO2 m -2 y-1 in sloping bogs, forested wetlands and uplands respectively. Average DOC flux was 7.7, 30.3, to 33.0 g C m-2 y-1 in sloping bog, forested wetland, and upland sites respectively. Estimates of carbon efflux from the terrestrial ecosystem was combined with values of net primary productivity from remote sensing to determine net ecosystem production (NEP). The average NEP estimated in three NCTR watersheds was 2.0 +/- 0.8 Mg C ha-1. Carbon loss as DOC was 10--30% of the total carbon flux from the watersheds confirming the importance of this vector of carbon loss in the NCTR. The watershed estimates indicate that forests of the NCTR serve as a carbon sinks consistent with the average worldwide rate of carbon sequestration in terrestrial ecosystems.

Communities and ecosystems in the Alaskan boreal forest are undergoing substantial change. People contribute to this change. They are also impacted by the consequences. For example, wildfire and spruce bark beetle (Dendroctonus rufipennis) outbreaks have increased in frequency and severity due to warming trends, affecting the ecosystem and services important to people. I conducted a study to explore the social and ecological implications of changing natural disturbances. I evaluated how the occurrence of spruce bark beetle outbreak has altered the probability of wildfire between 2001 and 2009 on the Kenai Peninsula, Alaska. Modeling the effects of bark beetle outbreak on the probability of large wildfire (> 500 ha) and small wildfires (<500 ha), I found that the influence of the outbreak differed as a function of wildfire size. The occurrence and length of outbreak increased large wildfire probability. Small wildfires were mediated by human influence and less so by bark beetle outbreak. I also used spatial econometric techniques to estimate how wildfires and the bark beetle outbreak affected property values on the Kenai Peninsula in 2001 and 2010. I found that wildfires> 3.3 ha and the bark-beetle outbreak increased property values. Wildfires <3.3 ha decreased property values.

The likely direction of change in soil organic carbon (SOC) in the boreal forest biome, which harbors roughly 22% of the global soil carbon pool, is of marked concern because climate warming is projected to be greatest in high latitudes and temperature is the cardinal determinant of soil C mineralization. Moreover, the majority of boreal forest SOC is harbored in surficial organic horizons which are the most susceptible to consumption in wildfire. This research focuses on mechanisms of soil C accumulation in recently burned (2004) and unburned (~1850-1950) black spruce (Picea mariana [Mill.] BSP) forests along gradients in stand productivity and soil temperature. The primary research questions in these three chapters address: (1) how the interaction between stand production and temperature effect the stabilization of C throughout the soil profile, (2) the quantity and composition of water soluble organic carbon (WSOC) as it is leached from the soil across gradients in productivity and climate, and (3) physiographic controls on organic matter consumption in wildfire and the legacy of wildfire in stable C formation (pyrogenic C, or black carbon). Soil WSOC concentrations increased while SOC stocks decreased with increasing soil temperature and stand production along the gradients studied. Stocks of BC were minuscule in comparison to organic horizon SOC stocks, and therefore the C stabilizing effect of wildfire was small in comparison to SOC accumulation through arrested decomposition. We conclude that C stocks are likely to be more vulnerable to burning as soil C stocks decline relative to C sequestered in aboveground woody tissues in a warmer climate. These findings contribute to refining estimates of potential changes in boreal soil C stocks in the context of a changing climate.