B33E-01 INVITED
An overview of the role of disturbance in the terrestrial carbon budget
Disturbance occurs throughout all terrestrial ecosystems to some degree, and affects the cycling of carbon to varying degrees. In this introductory paper to this special session, we will review the concept of the disturbance continuum as a paradigm for studying the short and long-term impacts of disturbance on terrestrial carbon cycling. We will present an overview of the different types of disturbance that are important to the carbon cycling and outline the research challenges associated with reducing uncertainties associated with disturbance.
B33E-02 INVITED
Quantifying the Impacts of Disturbance on the Canadian Managed Forest Carbon Budget
Several studies have demonstrated that natural disturbances have a very strong impact on the carbon (C) budget of Canada's boreal forests, but how important are these impacts within the managed forest, where both natural and anthropogenic disturbance factors are at play? Is it possible to quantify the C impacts of forest management (e.g harvesting, silviculture, fire suppression) in a landscape so heavily impacted by natural disturbances? We used national disturbance monitoring datasets, including a National Burn Area Composite product compiled using remote sensing data fusion techniques, forest health survey data, and forest harvest and silviculture statistics as input to an empirically-driven simulation model (the Carbon Budget Model of the Canadian Forest Sector, CBM-CFS3) to analyze the C budget of Canada's managed forest since 1990 and to quantify the direct and indirect impacts of disturbances. We found both strong direct and indirect impacts, including large direct transfers of CO2 and non-CO2 greenhouse gases (CH4, N2O) to the atmosphere, very large transfers of C from living biomass to dead organic matter pools, and high inter- annual variability in both. Changes in future disturbance regimes will determine the future contribution of Canada's managed forest to the global carbon cycle.
B33E-03
Hurricane impacts on tree mortality and carbon cycling in coastal forest ecosystems
Forests recovering from land-use, the encroachment of woody vegetation, and other ecological processes, cause terrestrial ecosystems to act as a net sinks for atmospheric carbon dioxide. Changes in the strength and sign of this sink over the coming decades are difficult to predict. One process that can act to diminish the terrestrial carbon sink is an increase in disturbance frequency and intensity, which transfers greater amounts of biomass from live to dead respiring pools, and shifts the forest size distribution toward smaller average tree size and lower biomass stocks. A number of studies predict an increase in the frequency of extreme weather events and the intensity of tropical cyclones under a warming climate, which may ultimately result in elevated forest disturbance regimes. Here we present a novel synthetic approach combining detailed ecological field investigations with remote sensing image analysis to provide spatially explicit estimates of forest damage, tree mortality, and biomass loss for U.S. landfalling hurricanes. Analysis results for Hurricane Katrina predicted the death and severe structural damage to about 320 million trees, representing a 100 Tg carbon transfer from live to dead biomass. Under the same wind-field, forest susceptibility to damage was highly tree species dependent, with cypress-tupelo swamp forests exhibiting the greatest resistance. Statistical models were useful for separating storm surge from wind effects on coastal forests. Similar analyses are currently underway for Hurricane Gustav, and will also be presented.
B33E-04
Carbon consequences of forest disturbance and recovery across North America
Forests contain nearly 40% of North American Terrestrial carbon and 70% excluding wetlands and permafrost lands. These forests are thought to be a significant long term sink for atmospheric carbon but are vulnerable to disturbances such as fires, logging and pathogens that can lead to significant transfers of carbon from the land to the atmosphere. Forest Inventory and Analysis (FIA) data for the U.S. provide a valuable resource for estimating long-term forest sinks but spatial and temporal resolutions of these data are coarse, limiting their use for atmospheric carbon modeling studies and process understanding. Our approach to overcome these limitations is to parameterize a carbon cycle model using forest type, biomass, age, and productivity from FIA. We then apply disturbance recovery dynamics to forest age maps derived from analysis of biennial time series from 29 Landsat scenes across the conterminous U.S. for the period 1982-2005 (provided by the NAFD project, S. Goward PI). These scenes were selected in a statistically rigorous manner to represent the forested areas of the U.S. The resulting maps are spatially (30 m) and temporally (~annual) explicit representations of biomass and carbon fluxes resulting from disturbance and regrowth for these representative regions. Finding that these scenes are representative of larger multi-state regions, we estimate forest carbon fluxes for all forested regions across the conterminous U.S. We propagate the uncertainties in the FIA data used for model calibration through to the predicted fluxes using a Monte Carlo approach. We show a conterminous U.S. forest regrowth sink of 0.22 PgC y-1 for the 2005 epoch, not accounting for emissions from biomass removals or local combustion. The largest regrowth sinks are found in the southeast, southcentral, and northwest, with spatially averaged net ecosystem productivity (NEP) of about 130 gC m-2 y-1. Despite low modern disturbance rates in the northeast and northern lakes states, the regrowth sink remains of moderate strength owing to the continued legacy of historical clearing. The prevailing condition of positive NEP is contrasted by very large local sources, particularly in the West where fires and clearcuts create large disturbed patches. These pronounced sources cross over into sinks between 5 and 15 years after disturbance. While broadly consistent with past findings, contemporary disturbance rates of 1 to 2 % per year as derived from this unique Landsat time series tend to exceed rates inferred from the FIA 0-20 year age class, and thus slightly elevate the estimated forest regrowth sink.
B33E-05
Increasing Amount and Spatial Heterogeneity of Soil Organic Carbon After Grazing Exclusion at Semi-Arid Grasslands of Inner Mongolia, P.R. China
Intensive grazing in the grasslands of Inner Mongolia resulted in degradation of steppe vegetation and soil structure followed by a decline of soil organic carbon (SOC), associated with a change of the naturally heterogeneous distribution of vegetation and topsoil properties in semiarid environments to a more homogeneous pattern in the overgrazed steppe systems. The recovery of degraded grasslands in terms of amount and spatial distribution of SOC was assessed by comparison of continuously grazed and ungrazed sites in two major steppe systems of Inner Mongolia, dominated by Leymus chinensis and Stipa grandis. SOC stocks in both steppe types were analyzed across different spatial scales with two sampling grids of different sizes at continuously grazed (CG) and ungrazed (UG79) sites, where grazing was excluded by fencing in 1979. At each site a large grid (150 m x 120 m) and a small grid (2 m x 2 m) containing 100 and 40 sampling points, respectively, were installed and undisturbed soil samples were taken from the topsoil (0 to 4 cm). Semivariograms were calculated to elucidate the spatial structure of SOC. Both UG79 plots show significantly higher concentrations and stocks of SOC and lower bulk densities compared to CG sites. The Stipa dominated site shows higher bulk densities and lower SOC contents and stocks at CG and UG79 than the Leymus dominated site. The spatial pattern of SOC changes from a relatively homogeneous distribution at CG sites to a more heterogeneous allocation after grazing exclusion at UG79 sites at both large and small spatial scales. The observed recovery of SOC at ungrazed plots can be attributed to increased above- and belowground organic matter input after grazing exclusion associated with a higher degree of soil aggregation and reduced soil erosion. This is accompanied by creation of a relatively heterogeneous pattern of SOC through the formation of vegetation patches. Around those “islands of fertility” primary production increases due to higher water infiltration and accumulation of sediments and organic materials leading to higher amounts of SOC under vegetation patches. Field-based data in our study clearly provide evidence for a considerable recovery of degraded semi-arid grasslands of Inner Mongolia after grazing exclusion for almost 30 years.
B33E-06
Budgeting Ecosystem - Atmosphere Carbon Exchange in a Subarctic Birch Forest
The overarching objective of this project is to work towards a better understanding of ecosystem-atmosphere interactions in a composite subarctic landscape with a focus on measurements and modeling of carbon cycling in birch forest environments. In this presentation we document the interactions between the birch forest (Betula pubescens ssp. czerepanovii) ecosystem and the atmosphere both in terms of greenhouse gas and energy exchanges. The study provides new information on climatic controls of interannual variability in annual carbon and energy exchange. This information is complimented with studies of the effects of insect outbreak disturbance on these annual budgets. Carbon flux data produced since 2003 shows that during the first year of measurements the forest acted as a large net sink of atmospheric carbon. However, during the growing season of 2004 the area was severely affected by an extreme outbreak of the autumnal moth (Epirrita autumnata) resulting in total defoliation of the forest over large areas. This caused the same forest stand to act as a net source of CO2 even during the peak growing season. During the summer of 2008, as part of a special campaign under the International Polar Year, the larger scale variability of the subarctic birch forest carbon fluxes was documented. A mobile eddy covariance tower provided seasonal measurements from six different locations in the catchment of lake Tornetrask which can be compared with longer term, inter-annual data from two permanent flux towers operating continuously in the vicinity of the village of Abisko. The sites were chosen to document possible differences in CO2 fluxes depending on the time since last defoliation which was in 2004 in some areas. Also sites were chosen with different types of birch forest (monocormic, polycormic) and at greatly varying distances to the oceanic influence from the Norwegian coast. This poster will present and discuss preliminary CO2 flux data from all these different sites.
B33E-07
Woody Detritus Controls Forest Carbon Budgets: The Disturbance Connection
It is becoming increasingly apparent that the role of woody detritus in forest carbon dynamics can no longer be ignored. Not only is woody detritus a major carbon store in forests, but its importance increases with disturbances. The factors controlling the production and decomposition of this material have been identified. However, how these factors interact and how the balance of these interactions changes from place to place has not been appreciated. The amount and nature of the woody detritus created and left by disturbance profoundly influences the pattern of net ecosystem carbon balance (NECB) over succession. The more woody detritus left by disturbances, the longer NECB remains negative. Since there is a feedback between the accumulation in live wood versus the losses from woody detritus, the average of NECB over succession approaches zero unless the nature of the disturbance regime changes. With increasing rates of disturbance being observed in North America, it is highly likely that losses from woody detritus decomposition will cause a shift from a positive NECB (sink) to a negative NECB over the next few decades. The timing and magnitude of this shift is highly uncertain due to the stochastic nature of disturbance and our very poor understanding of woody detritus decomposition. While remote sensing can decrease the former, the consequences of increased disturbance will not be understood until the latter is improved with systematic field studies.
B33E-08
Effects of Climate Change and Disturbances on Carbon Sequestration of California Ecosystems
The interactions of climate change, wildland fire, and human activities make the calculation of regional ecosystem carbon balance extremely difficult. We used the Integrated Biosphere Simulator (IBIS) and a set of newly available, relatively high resolution (mostly 30- to 60-meter) fire and land cover change data to examine the effects of atmospheric CO2, climate change, fire, logging, and deforestation/devegetation on carbon balance in California's forest land, shrubland, and grassland. Simulations suggested that from the 1950s to the 1990s, the net primary productivity (NPP) of California's natural vegetation increased from 79.4 to 88.9 Tg C yr-1 due to CO2 fertilization and climate change. The net ecosystem productivity (NEP) increased from 6.2 to 8.4 Tg C yr-1. Carbon loss from biomass and soil combustion averaged 0.77 Tg C yr-1. Carbon loss from logging, deforestation, and devegetation averaged 1.01 Tg C yr- 1. The cumulative NEP increase over the 50 years (94 Tg C) offset all the direct carbon losses from disturbances (91 Tg C) during the same period. The model also suggested that net biome productivity (NBP) varied from 11.7 to +26.0 Tg C yr-1, with a mean value of +4.9 Tg C yr-1. The NBP fluctuation was mainly driven by inter-annual climate variability instead of disturbance. The carbon sink response to CO2 fertilization was stronger in high biomass regions (forests) than low biomass regions (shrublands and grasslands). Most southern California ecosystems and some northern California forests tended to lose carbon with a warming trend, whereas forests in relatively high elevation regions tended to gain more carbon. The model also indicated that California forests were not sensitive to precipitation change, but shrublands and grasslands responded positively to precipitation increases.