B31D-0309
Utilizing remote sensing to supplement ground monitoring of Diorhabda elongata as a control agent for Tamarix ramosissima in Dinosaur National Monument
The plant Tamarix ramosissima has invaded significant riparian habitat along the Green River in Dinosaur National Monument. Commonly known as salt cedar or tamarisk, it was introduced from Eurasia to the Southwestern United States to prevent soil erosion along riverbanks and as an ornamental plant. It has since come to affect water resources, recreation, wildlife, and ecosystem services. Various methods used to control tamarisk's spread have had moderate success but have drained National Park Service of human and monetary resources. In June 2006, the salt cedar leaf beetle (Diorhabda elongata) was released as a biological control agent within the park to defoliate and ultimately eradicate the invasive species. This study examines the efficacy of using Landsat TM imagery to supplement ground monitoring of the beetle's spread and its effects on tamarisk in Dinosaur National Monument, and discusses the development of a GIS model to predict annual change in tamarisk cover and beetle populations. Through fieldwork, we determined four areas of interest with favorable attributes for satellite detection. A change detection model was created by layering 2005-2008 data and quantifying mean NDVI. Results show that intra-year NDVI trends may be more effective for accurate detection than single-image year-to-year comparisons largely because intra-year environmental variability is significantly smaller. Additionally, our GIS model predicted significant growth of beetle population, implying that defoliation will become more apparent in future years. However, challenges to detecting this defoliation include the year-to-year variability of environmental factors, low spatial resolution of Landsat TM data, low visibility into parts of the Green River canyon, and the spectral mixing of tamarisk and native vegetation.
B31D-0310
Statistical Analysis of Extreme Climatic Indices to Determine Environmental Change in Former and Present Karner Blue Butterfly Habitats
The Karner Blue butterfly is a federally endangered species that once was widely distributed throughout 12 states along the northern part of the United States and Ontario, Canada. Now it only exists in seven states. Many factors are considered to have affected the extinction of this species and this study examines the effect of climate change on the persistence of the Karner Blue butterfly. Five sites were selected to study the effect of climate change. Three sites currently have a Karner Blue population (Allegan, MI, Fort McCoy, WI, and Saratoga, NY) and two sites the Karner Blue has disappeared (Oak Openings, OH, and Pinery, Ontario). Daily climate data from the 1950s to 2005 were used for calculating 13 extreme climatic indices related to precipitation and temperature. The data were broken into two time periods (pre-1984 and post-1984) to analyze how those indices have changed since the butterfly disappeared from the two sites. Statistical analyses including t-tests and ANOVA were used to compare these indices within two time periods among five sites. The results showed that different indices have changed differently among the five sites. The number of extreme hot days and number of extreme cold days per year have a statistically significant change in the sites where the Karner Blue butterfly disappeared. The precipitation-related indices do not show a statistically significant different trend among the five sites. Temperature seems to have more of an effect on the existence of the Karner Blue butterfly. Furthermore, butterfly population size and lake effects are also important factors that cannot be neglected. Larger populations seem to have better chances to survive during a dramatic climate change event.
B31D-0311
Disturbance, Climate, and Management Impacts on US West-Coast Forest Carbon Budgets
Forest net uptake of atmospheric CO2 (net ecosystem production, or NEP) is dependent on climate, disturbance history, management practices, forest age, and forest type. Accurate quantification of NEP and forest carbon budgets is necessary for validation of coupled carbon-climate models and monitoring state carbon budgets. To improve understanding of the influence of disturbance, climate, and management on forest carbon stocks and fluxes in the western U.S., federal inventory data and supplemental field measurements were used to estimate several important components of the carbon balance in forests in Oregon, Washington, and California from 2001-2006. Species- and ecoregion-specific allometric equations and ecoregion-specific lookup tables were used to estimate live and dead biomass stores, net primary productivity (NPP), NEP, and mortality for different age classes. Natural and anthropogenic disturbance impacts on forest carbon accumulation and NPP varied by ecoregion, forest type, and ownership. In the semi-arid East Cascades and mesic Coast Range, mean total biomass was 8 and 24 kg C m-2, and mean NPP was 0.30 and 0.78 kg C m-2 yr-1, respectively. Decrease in NPP with age was not general across ecoregions, with no marked decline in old stands (greater than 200 years) in some ecoregions. Within ecoregions, mean live and dead biomass were usually higher on public lands, primarily because of the younger age class distribution on private lands. In the absence of stand-replacing disturbance, total landscape carbon stocks could theoretically double if forests were managed for maximum carbon storage. Although the theoretical limit is probably unattainable given the timber-based economy and fire regimes in some ecoregions, there is still potential to significantly increase terrestrial carbon storage by decreasing anthropogenic disturbance through increased rotation age and reduction in harvest rates.
B31D-0312
Analysis of Forest Fire Disturbance in the Western United States Using Landsat Time Series Images: 1985 to 2005
In this study we used two different disturbance maps (both utilizing 30 m resolution Landsat imagery) to assess disturbance trends in Western US forests. The first are maps developed by the NAFD project (North American Forest Dynamics). Each NAFD data cube contains an annual-biennial record of forest disturbance events from 1984-2005. We complimented the NAFD maps with MTBS maps (Monitoring Trends in Burn Severity). MTBS solely maps fire disturbance, recording historical (1985-2005) and contemporary burn severity and fire perimeter across the United States. We used Landsat time series stacks for four locations: Oregon (Landsat path 45 row 29), California (p43r33), Idaho (p41r29), and Utah (p32r37). In all four stacks, fire was a relatively small percentage of the total forest disturbance (ranging from 8% in Utah to 27% in Oregon for the entire 20 year period). We also found that the years with greatest burned area were years with a high aridity index (lower precipitation and higher temperatures), a condition necessary, but not sufficient for fire activity. To assess post-disturbance vegetation regrowth we used two spectral indices, the Normalized Difference Vegetation Index (NDVI) and the Normalized Burn Ratio (NBR). Both indices are sensitive to well-defined spectral paths that forests follow during and after disturbance. As expected, NDVI and NBR were lowest (highest) for the highest (lowest) severity class burned area. However, NBR and NDVI only appear to respond to vegetative reflectance in the first decade after a burn. Therefore, they give useful information on location, timing, and magnitude of disturbance, but direct measurement of biomass with other sensors would be necessary to obtain additional ecological information.
B31D-0313
Analysis of the spatial and temporal characteristics of the 2004 Alaskan fires
Satellite data were combined with digital elevation data to analyze the spatial and temporal characteristics of the 2004 fires in interior Alaska. Satellite data were used to map pre-fire vegetation cover, burned area, and timing of fire activity during the growing season. While MODIS hot-spot data were used to map the timing of fire activity, both MODIS and Landsat TM/ETM+ information products were used to map vegetation cover and burned area. The results of our study showed that actual burned area was 80 percent of the area within a fire perimeter regardless of fire size. While the estimates of burned area using the two satellite information sources were not dramatically different, the patterns of vegetation cover mapped by the two systems were. In particular, the Landsat vegetation cover maps produced higher estimates of spruce forest cover than did the MODIS land cover map. The study showed the burning of spruce forests was highest during early season fires and decreased during late season fires. The burning of deciduous forests and non-forested lands increased during late season fires.
B31D-0314
A Recent Shift in the Carbon Balance of High-latitude Terrestrial Ecosystems in Response to Changes in Climate and Disturbance Regime
Analyses of the global carbon budget suggest that terrestrial ecosystems have been responsible for slowing the rate of anthropogenic CO2 build-up in the atmosphere through carbon uptake and storage, with northern extratropical regions responsible for most of this land-based CO2 sink. However, recent changes in atmospheric chemistry, climate trends, disturbance regimes, land use and management systems in northern high latitude regions have the potential to alter the terrestrial sink of atmospheric CO2. To determine the recent trends in the carbon balance of the arctic and boreal ecosystems of this region, we performed a retrospective analysis of terrestrial ecosystem dynamics across the pan-arctic (north of 45°N latitude) using a process-based biogeochemistry model. The results of the simulations suggest a shift in direction of the net flux from the terrestrial sink of earlier decades to a net source on the order of 8.5 Tg C per year between 1997 and 2006. The positive carbon balance (sink) estimated for tundra regions is consistent with observations suggesting a "greening" of, or an increase in productivity in, these ecosystems. However, the simulation framework and subsequent analyses presented in this study attribute the overall shift in regional carbon balance primarily to a large loss of carbon as a result of "browning" in boreal forest ecosystems. Model results suggest that primary productivity of the boreal forest declined over this recent time period in response to a decreasing trend in water balance. However, the substantial release of CO2 as a direct result of the large area of boreal forest burned during the past decade was the largest signal in the overall negative carbon balance for the pan-arctic region. Our results, along with those of other recent studies, emphasize the importance of changes in the disturbance regime (e.g., fire events and insect outbreaks) in the weakening and possible disappearance of the terrestrial carbon sink in high latitude ecosystems.
B31D-0315
An integrative approach to quantifying the effects of disturbance on regional forest carbon cycling
Many challenges remain in identifying ecosystem consequences of perturbations, especially in the context of global change. Insect defoliation in particular represents an important process in forest ecosystems, affecting succession, biogeochemical cycling, net primary productivity (NPP), and stand structural characteristics. To address the impacts of defoliation on forest nutrient cycling and productivity, we combined hyperspectral data from the Airborne Visible / Infrared Spectrometer (AVIRIS) and multi-temporal (i.e. Landsat and MODIS) remote sensing data with field measurements to derive landscape scale measures of the variation in light-use efficiency (LUE, ε) based on forest functional properties, canopy architecture, and disturbance regime. This information is used with an improved production efficiency model (PEM), driven by daily MODIS and climatology data, to quantify the short- (i.e. day to month) and long-term (i.e. inter-annual) responses of forest carbon cycling to climate variability and perturbations, specifically insect disturbances of varying intensities. This research furthers our understanding of ecosystem responses to disturbance and environmental change by testing the links between forest ecosystem processes and mechanisms that can be remotely sensed, thereby advancing large-scale modeling abilities. We demonstrate the ability to remotely sense aspects of forest function related to productivity that are directly impacted by perturbations and may be changing in distribution as a result of global change.
B31D-0316
Projecting Carbon Cycling Trajectories in Forests of the Upper Midwest, USA: Has Carbon Storage Peaked?
The mixed deciduous forests of the upper Midwest, USA are approaching an ecological threshold in which dominant early successional aspen and birch trees are reaching maturity and beginning to senesce, giving way to a canopy that is more species diverse and structurally heterogeneous. Widespread ecological changes in maturing forests of the upper Midwest are predicted to reduce terrestrial C storage in the region; however, no empirical evidence exists to support this hypothesis. At the University of Michigan Biological Station in northern Michigan, we are combining long-term C cycling measurements with a large-scale experimental manipulation to forecast how forest C storage will change in response to ongoing succession and disturbance, and to climate variation. At the plot scale, 10-yr trajectories of increasing wood net primary production were accompanied by significant increases in leaf area index (LAI), which were positively correlated with successional advances in canopy species diversity as late-successional species grew into predominately aspen and birch canopies. Surveys of canopy structure indicate that more species diverse canopies support greater LAI by increasing the vertical distribution of leaf area. These results suggest that forests of the upper Midwest may store more C if, as predicted, their canopies become more species diverse and structurally heterogeneous. To examine changes in forest C cycling following successional transition from mature aspen and birch to a young mixed conifer-deciduous ecosystem, we accelerated forest succession by stem girdling all aspen and birch (>6,700 trees, ~35% canopy LAI) within a 39 ha area in Spring 2008. The Forest Accelerated Succession ExperimenT (FASET) will test the hypothesis that forest net ecosystem production will decline temporarily following an initial disturbance that results in partial canopy defoliation and subsequently increase as canopies become more biologically and structurally complex. Our goal is to elucidate biophysical mechanisms that will constrain C cycling in future forests across much of the upper Midwest.
B31D-0317
Controls on the Origin and Cycling of Riverine Dissolved Inorganic Carbon in the Brazos River, Texas
Rivers are generally supersaturated in CO2 with respect to the atmosphere. However, there is little agreement on the sources and turnover times of excess CO2 in river waters. This is likely due to varying dominant controls on carbon sources (e.g. geologic setting, climate, land use, or human activities). In this study, we measured carbon isotopic signatures (δ13C and Δ14C) of riverine dissolved inorganic carbon (DIC), as well as solid state cross polarization/magic angle spinning (CP/MAS) 13C nuclear magnetic resonance (NMR) of particulate organic carbon (POC), to determine carbon sources fuelling respiration of the Brazos River in Texas. We found that sources of riverine CO2 varied significantly along the length of the Brazos. In the middle Brazos (between Graham and Waco), which is partially underlain by limestone, riverine DIC had average Δ14C of 74 ‰ and δ13C of -7.5 ‰, suggesting that riverine CO2 is derived almost entirely from contemporary carbon (less than 5 years old) with little evidence of carbonate input, probably due to the damming upstream of Waco. In the lower Brazos (downstream of Bryan), riverine DIC was highly depleted in 14C (average Δ14C = -148.5 ‰) and enriched in 13C (average δ13C= -9.32 ‰), indicative of the presence of old carbonate. Since there is no carbonate bedrock in contact with the river in this area, the most likely source of old carbonate is the shell used in road and building construction throughout the 19th century. Our results suggest that the effect of human activities superimposes and even surpasses the effect of natural controls (e.g. geologic setting and climate) on C cycling in the Brazos.
B31D-0318
Study of terrestrial carbon cycling as impacted by mountaintop coal mining in the Southern Appalachian forest region using carbon elemental and isotopic data and remote sensing of land cover change
The need to quantify the impact of human disturbance upon carbon flux and storage has been recently highlighted in order to more accurately budget carbon. One understudied but critical area of research is surface coal mining's impact on terrestrial carbon storage and sediment carbon transport processes—which has been identified as potentially important to understanding fluxes in global carbon budgeting. While national attention has focused on U.S. coal production to maintain a vibrant economy, scientists are concerned that increased coal production could have unforeseen environmental implications if the relationship between coal mining practices and the environment is not better understood. This issue is particularly important to the coal mining region of the Southern Appalachian forest region, which has been responsible for 23.3% of the coal produced in the United States over the past twenty years and seen approximately 300,000 ha of forested land disturbed by surface coal mining during that time period. Our presentation provides results that focus upon terrestrial carbon cycling as impacted by mountaintop coal mining in the Southern Appalachian forest region. In order to study carbon redistribution due to the mining disturbance, our methods make use of measurements of total organic carbon, total organic nitrogen, and carbon and nitrogen stable isotopes of soils and eroded sediments collected in the region as well as published data, consultation with experts and remote sensing of land cover change. It was found that disturbed terrestrial carbon, including soil C, non-soil or plant C, and geogenic C, is approximately 10% of the carbon emitted to the atmosphere during coal combusting and transportation and mining of coal. Quantification of the fate of terrestrial carbon in different pools is provided and discussed including the fate of: removed soil C and non-soil C from mountaintop coal mining sites; long-term uptake of carbon from the atmosphere during recovery of the terrestrial system; newly deposited coal fragments within the terrestrial soil reservoir; and carbon that is eroded to streams in mined watersheds with different levels of disturbance.
B31D-0319
Improving Weather Research and Forecast model (WRF) for California Carbon Emission Inventory Study
As part of the effort on cutting Californian carbon emission to 1999 level by 2020, we are building a regional carbon emission monitoring network combining various observations and inventory at both global and regional levels. The core for integrating all data sources is a regional atmosphere model. Since conventional GCMs are designed for emulating large scale event, we choose to improve WRF model with fine-tuned physics options for regional atmosphere dynamics. We show here our initial results on model improvements of capture small scale emission events at community level.
B31D-0320
Aboveground Tree Carbon Stocks and Flux Following Mountain Pine Beetle Outbreaks
Mountain pine beetle outbreaks result in tree mortality across millions of acres in North America, with significant effects on forest ecosystem processes such as carbon cycling. Following outbreak-related mortality, forest stands continue taking up carbon (C) via the growth of 1) surviving trees and/or 2) tree seedlings that establish during and after outbreaks. To date, the degree to which surviving trees can, in the absence of post-outbreak seedling establishment, recover pre-outbreak C stocks and flux is largely unknown. To address this uncertainty we asked: (1) Do aboveground stocks and flux among pure lodgepole pine stands recover to pre-outbreak levels independent of post-outbreak regeneration? 2) What is the long-term effect of the mountain pine beetle outbreaks on modeled aboveground C stocks and flux? We used measurements from several stands affected by a recent mountain pine beetle outbreak as input to the Forest Vegetation Simulator, an individual tree-based growth model, to predict stand-level aboveground C stocks and flux. The simulation time period spanned from just prior to the bark beetle outbreak and for 200 years following outbreak collapse. At five-year intervals, we compared C stocks and fluxes in stands affected by the disturbance to conditions in the same stands immediately preceding the outbreak, as well as identical stands modeled as if the outbreak had not occurred. Crown closure of surviving trees was predicted by the model in all stands following outbreak collapse, but measured outbreak mortality did not significantly reverse increases of growth dominance by relatively large trees. Stand-level growth dominance and increases of stand density following crown closure have been associated with declines in stand-level primary productivity and productivity efficiency. Thus, unlike stand- level C stocks, predicted C flux did not recover relative to pre-outbreak levels, although we observed basic patterns of C flux rise, peak, and long-term declines typical of all forest ecosystems. The structural and biological characteristics of forests surviving mountain pine beetle outbreaks suggest that seedling establishment in the post-disturbance environment likely plays an essential role in the recovery of pre- outbreak annual aboveground C flux, but not C stocks.
B31D-0321
Monitoring a Deciduous Forest Regeneration Following a Severe Ice Storm
Leaf Area Index (LAI) has been used to estimate the carbon budget in several studies and is now mapped routinely from satellite imagery. In this study, overstory forest damage and its regeneration are assessed using in-situ LAI measurements taken from 1997 to 2007 with three optical systems (LAI-2000, TRAC, and different digital hemispherical photography camera systems) following an intense freezing rain that occurred in January 1998. The study site is composed of two deciduous stands in Larose Forest, Ontario that were damaged during an intense freezing rain event in January 1998. Time series of the variables required to estimate LAI, are assessed separately to better understand the complex architectural changes over time resulting from the impact of the ice storm. Results show that the season maximum effective plant area index (PAIe) decreased by almost 50% for both sites the summer following the ice storm (1998), but had a substantial recovery the following year (1999), and did not show any significant increase from 1998 to 2007. The 1997 to 1998 decrease was more notable for site 1, with a change of three PAIe units; while a decrease of only one PAIe unit occurred for site 2. Both site 1 and site 2 regained significant PAIe in 1999. LAI follows a similar decrease from 1997-1998 with an increase in 1999. However, LAI increased yearly by 0.06 and 0.07 units from 1999 to 2007 for site 1 and site 2, respectively. Study results also show that foliage clumping was the main driver of the LAI increase from 1999 to 2007. Site 1—the older of the two sites— regained its pre- storm LAI within six years, while the younger site 2 had not regained its original LAI nine years after the ice storm.
B31D-0322
Satellite data based method for general survey of forest insect disturbance in British Columbia
Regional forest disturbances caused by insects are important to monitor and quantify because of their influence on local ecosystems and the global carbon cycle. Local damage to forest trees disrupts food supplies and shelter for a variety of organisms. Changes in the global carbon budget, its sources and its sinks affect the way the earth functions as a whole, and has an impact on global climate. Furthermore, the ability to detect nascent outbreaks and monitor the spread of regional infestations helps managers mitigate the damage done by catastrophic insect outbreaks. While detection is needed at a fine scale to support local mitigation efforts, detection at a broad regional scale is important for carbon flux modeling on the landscape scale, and needed to direct the local efforts. This paper presents a method for routinely detecting insect damage to coniferous forests using MODIS vegetation indices, thermal anomalies and land cover. The technique is validated using insect outbreak maps and accounts for fire disturbance effects. The range of damage detected may be used to interpret and quantify possible forest damage by insects.
B31D-0323
Climate Change Implications to Vegetation Production in Alaska
Investigation of long-term NOAA series of Advanced Very High Resolution Radiometer normalized difference vegetation index (NDVI) data from 1982 through 2005 revealed statistically significant vegetation response to climate drivers of temperature, precipitation and solar radiation with exclusion of fire disturbance. Abiotic trends were calculated and correlated to satellite remote sensing observations of vegetation productivity to understand biophysical processes that could impact ecosystem carbon storage. Warming throughout Alaska resulted in disparate trajectories for vegetation growth due to precipitation and photosynthetically active radiation variation. Interior spruce forest low lands in late summer had precipitation deficit which resulted in extensive fire disturbance and browning of undisturbed vegetation with reduced post-fire recovery in burned sites; while Northern slope alpine and moist tundra had increased production due to warmer-wetter conditions during the late 1990s and early 2000s. Coupled investigation of vegetation's response to warming climate in Alaska found spatially dynamic processes with and without fire disturbance observed from coarse resolution satellite instruments. Future effort will simulate carbon cycle process with fire disturbance to understand spatially variant source-sink distribution of Alaskan ecosystems.
B31D-0324
A 35-year range of variability in Alaska boreal forest burn severity derived from Landsat
While the historic range of variability concept is often applied to wildfire size and occurrence rates to assess trends and anomalies in ecological disturbance regimes, this concept has not yet been applied to wildfire burn severity due to insufficient data. This research utilized 35 years of Landsat Multispectral Scanner (MSS), Thematic Mapper (TM), and Enhanced Thematic Mapper (ETM+) data to map wildfire severity for over 100 fires in the boreal forest component of the Yukon River Basin in Alaska. Methodology utilized the Relative differenced Normalized Burn Ratio (RdNBR) as an indicator of burn severity and included developing a crosswalk between MSS and TM/ETM+ imagery to make burn severity comparable given the limited spectral range of MSS. Results indicate a contemporary historical range of variability for burn severity and allow for a better understanding of the ecological impacts of the 2004 and 2005 wildfire seasons, both with historically high numbers of fires and area burned.
B31D-0325
Implementation of the North American Forest Dynamics (NAFD) Project: Current Progress and Future Plans
Through the North American Forest Dynamics (NAFD) project we are evaluating forest disturbance and regrowth patterns by combining U.S. Forest Service Forest Inventory and Analysis (FIA) field observations with biennial time series Landsat imagery. Phase I of the NAFD study examined forest dynamics and disturbance history from Landsat time series stacks (LTSS) for 23 sample locations within the United States. These samples were selected with known inclusion probability to allow derivation of unbiased national estimates of disturbance rates with known uncertainty boundaries. FIA field plot data were used to validate the changes mapped in the Landsat data by the vegetation change tracker (VCT) algorithm. Additional accuracy assessment was performed on selected sample LTSS using visual analysis of the time series compared with high resolution imagery. Results of NAFD disturbance mapping to date reveal generally high rates of forest disturbance across the U.S., with these rates varying both spatially and temporally. Ongoing work for the Phase II NAFD study expands on this initial work, in several ways, to meet NACP goals. We are adding additional sample locations in the conterminous U.S. to reduce error in nationwide estimates of disturbance. We are also now partnering with Canada and Mexico to improve our understanding of continent- wide forest dynamics. The NAFD study has also developed collaborative relationships with other NACP- funded scientists who are using our analyses to inform their carbon assessment. In addition, we are currently developing nationwide maps of disturbance rates from model-based estimators combined with the sample site locations to better meet the needs of carbon modelers. The NAFD project is also extending forest dynamics analyses by examining regrowth patterns. We are doing so by converting Landsat data stacks to biomass to better characterize the carbon significance of forest disturbance and regrowth. We are also employing radiative transfer modeling as a means to validate relationships between the FIA and Landsat measurements by employing the FIA field data to parameterize the RT model and compare with the actual Landsat observations.
B31D-0326
North American Forest Dynamics Evaluated Using Landsat Observations and FIA Measurements: Science Goals within NACP
One of the major goals of the North American Carbon Program is to understand the sources and sinks of atmospheric carbon. The first State of the Carbon Cycle Report (SOCCR) with a focus on North America identified forest as one of the largest sinks on this continent. This report attributed much of this carbon uptake to forest regeneration on abandoned agricultural lands based on carbon accounting procedures. Forest disturbance was not fully addressed. While forest disturbance and regrowth are recognized as significant forces modulating North American carbon balance, examining the carbon fluxes arising from these processes is hindered by a lack of understanding of continental-scale forest dynamics. The North American Forest Dynamics (NAFD) project has been developed to address this fundamental question using historical Landsat observations in combination with comprehensive field measurements collected through the USDA Forest Inventory and Analysis (FIA) program. This presentation will provide an overview of the science objectives of the NAFD project and the approaches for achieving these goals.
B31D-0327
The Perfect Fire? Aging Stands in the Alaskan Boreal Forest Encounter Global Warming
The ecological responses of the boreal forest to climate change have global significance because of the large amount of carbon stored in its soils and biomass. Fire, mostly ignited by lightning, is the keystone disturbance agent in this forest. It triggers cycles of forest succession in its wake, and burning is the main avenue for carbon release back to the atmosphere. We studied the interactions between climate, fires, forest succession, and the age distributions of forest stands in a 60-million hectare region of Interior Alaska over the past 150 years. First we developed a statistical model relating climate to area burned over the period of record (1950-2005). Next we combined this model with climate reconstructions to extend the estimates of area burned back to A.D. 1860. We checked the resultant fire history against stand-age data from 5000 living trees sampled in the study region. Then we fed the history of area burned into a computer model that simulates forest succession on real landscapes. Results show striking changes in the means and variances of stand ages over the last 150 years in response to interactions between climate change and the successional dynamics of the boreal forest. Average stand age increased steadily between 1880 and 1940 and has fluctuated at high levels since then, indicating a historically unusual abundance of flammable stands. This accumulation of old stands has created the potential for unusually large fires. Some support for this conclusion comes from the unprecedented large sizes of the areas burned in 2004 and 2005. Further support comes when we add to the analysis the forecasts made by global climate models for Alaska over the next twenty years. Bracketing estimates for climate warming and precipitation change suggest that warmer, drier summers combined with aging forest stands will cause a series of unusually large fires, the like of which have not occurred in the region for >150 years. We infer that the enhanced burning of the Alaska boreal forest over the next 20 years will increase the release of trace gases from this region. We speculate that the forest will be transformed from being conifer dominated to one dominated by deciduous tree species, which could have sweeping effects on the region's other biota, its hydrology, and the role of the boreal forest in the global carbon cycle.
B31D-0328
Plant Community Responses of Alaskan Arctic Tundra After 14 Years of Experimental Warming and Snow Manipulation
Emissions of greenhouse gases are expected to raise global mean temperature over the next century by 1.0- 3.5°C. Global warming trends are amplified at high latitudes because heating converts high-albedo (reflective) ice and snow surfaces to dark absorptive surfaces that absorb more solar energy and transfer it to the atmosphere. Furthermore, scientists have argued that ecological responses to this recent climate change will be complex and varied. For example, the warming of the Alaskan Arctic during the past 150 years has accelerated over the last three decades and is expected to increase vegetation productivity in tundra if shrubs become more abundant. Predicted vegetation changes in arctic tundra, because of climate change, have therefore been based on a warmer climate that is either drier or wetter than at present. In order to investigate how tundra vegetation may respond to increases in temperature and snow cover, we used 1m2 open-topped fiberglass chambers (OTC's) combined with large snow fences to artificially warm and modify winter snow regimes of a series of permanent vegetation plots established in Toolik Lake Field Station, Alaska. Fieldwork consisted in measuring the vegetation growth and height of plant species in these plots using the point-frame method. The snow cover and temperature manipulation was done in two ecosystem types, dry heath tundra and moist tussock tundra. The study sheds light on how the vegetation of these two tundra sites has responded after 14 years into the experiment and focuses in changes is species composition, relative abundance, diversity and canopy height. Preliminary results suggest major changes in vegetation composition in both tundra sites over the 14 year sampling period. Changes were more conspicuous in the moist tussock tundra site, where considerable increases in the abundance of the dominant species Betula nana, Eriophorum vaginatum, Salix pulchra and Carex bigelowii were detected consistently throughout the years. Instead, changes in the dry heath tundra show a recent decrease in the dominance of the species Arctostaphylos alpina and Dryas octopetala and an increase in the abundance of the species Louseleuria procumbens. Changes observed in the moist tundra supports the predictions of shrub/graminoid dominated tundra in a warmer global temperature scenario while changes in the dry heath tundra might suggest successional patterns in vegetation composition.
B31D-0329
The Legacy of Invasive Species on Ecosystem Carbon Dynamics in a Restored Tallgrass Prairie
Restoration of degraded grassland and prairie ecosystems represents a target sink for offsetting rising atmospheric CO2 levels by increasing carbon sequestration in C-depleted soils, as two-thirds of the biomass is allocated belowground. When considering controls on ecosystem C cycling, biodiversity-led productivity has the potential to be a strong biotic influence. However, invasive species can disrupt ecosystem processes by exhibiting functional characteristics which are distinct from their native counterparts. Invasibility has been linked to disturbance history, which might lead to additional vulnerability of managed lands. The restoration of tallgrass prairie at Fermilab, Batavia, IL, is a known C sink, accruing soil organic matter at rates 43 g C m-2 y-1 during the past 20 years. This rate integrates environmental, climatic and vegetation variations that occurred over this period. Typically, the tallgrass prairie is dominated by warm season grasses and forbs with sporadic but recurrent years when invasive species increase productivity. We measured net ecosystem exchange, net ecosystem production (NEP) and soil C at a 19- year-old restored tallgrass prairie in a four year study where plant species dominance varied. In the first year, the prairie restoration was a strong C sink with a NEP 438 g C m-2 despite a pronounced spring drought. During the second year, dominance of the invasive biannual Melilotus alba L., led to a shorter growing season that resulted on a 47% reduction in NEP from the previous year. NEP did not recover in the third year, even when M. alba was present but not dominant and a number of prairie species re-emerged, showing the legacy of the previous year disturbance. At this time soil C for all years and a fourth NEP year is being analyzed. These data suggest that biotic factors can exert large memory effects on NEP and possibly influence the sink capacity of restored ecosystems. Management strategies should aim to control biotic limitations to NEP in order to maximize long term C sequestration of restorations.
B31D-0330
Changes in the energy, water vapour and CO2 fluxes over a semi-arid grassland after fire disturbance
Continuous measurements of energy, water vapour and CO2 fluxes were made over a semi-arid grassland located on the Audubon Research Ranch in south western Arizona (31.5907N, 110.5104W, elevation 1496 m), USA, from first week of June 2002 to 2007 using the eddy covariance technique as a contribution to the global energy and water cycle experiment (GEWEX). The research ranch was established in 1969 as an ecological research preserve and it is now one of the largest ungrazed, privately managed grassland sites in Arizona. Wild fire occurred in April and May 2002 burned all the standing vegetation and litter on in research ranch (~38,000 acres) including 500 acres of grassland. The mean annual temperature and precipitation at this site were ~16 deg C and ~~370 mm, respectively. More than 60% of the annual precipitation was received during the North American monsoon period, which typically extends from July to September. In June of 2002, daily maximum sensible heat fluxes were ~450 Wm-2, highest sensible heat flux in June of the record, latent heat fluxes were < 20 Wm-2 and net radiation was near 550 Wm-2. Upwelling longwave radiation was close to 700 Wm-2. The ecosystem was a carbon source up to mid-July. With the onset of monsoon, sensible heat fluxes, long wave radiation, surface temperatures and albedo dropped dramatically and the ecosystem turned to carbon sink, with daily total net ecosystem exchange (NEE) varying up to <-2 g C m-2, by mid-July to August of 2002. 2002 was followed by driest year of the record (2003) with <50% of normal July-September precipitation. Because of this, the recovery of the ecosystem was delayed. Daily total evapotranspiration during July-August increased from 2 mm d-1 in 2002 to >3 mm d-1 in 2007. On annual basis ecosystem ecosystem turned to a carbon sink in 2006, with daily total NEE reaching up to -8 g C m-2, a year with longest growing season. Our study suggests that energy, water vapour and CO2 fluxes at this site were strongly influenced by precipitation associated with North American monsoon. The details on the recovery of the ecosystem from fire disturbance and its implications will be presented.
B31D-0331
Opportunities And Challenges For Reducing Carbon Emissions From Deforestation And Degradation In The Democratic Republic of Congo
An international regime is under negotiation for compensating tropical nations that succeed in lowering their
greenhouse gas emissions from tropical deforestation and forest degradation, which are responsible for
approximately one fifth of world-wide carbon emissions. One of the barriers to its success is the participation
of countries with current low rates of deforestation. A successful regime must account for both current
emissions baselines as well as future emissions—otherwise leakage into current low deforestation countries
may offset emissions reduction. Here, we provide a baseline for 1990-2000 and potential future rates of
deforestation and forest degradation in the Democratic Republic of Congo, which contains the world's second
largest area of contiguous tropical forest and epitomizes a country with low deforestation rates yet under
rapidly increasing land-use pressures. We mapped above ground biomass (17 billion tons C), modeled
deforestation (1.1 million ha/year), and estimated CO2 emissions (220 million tons/year) for the 1990-2000
period. We then estimated a price range of reducing carbon annual emission from rural household from $ 4
to more than $ 75 per tones of carbon—with 40 million tones of carbon under $ 10 per ton.
http://atlas.whrc.org/DRCEmissions/