Using surface and space-based observations of CO2 for inverse modeling of carbon fluxes
Using the GEOS-Chem chemical transport model and a Bayesian inversion approach, we compare the information provided by surface and space-based measurements of atmospheric CO2 for constraining estimates of CO2 surface fluxes. The impact of combining data from the GLOBALVIEW-CO2 surface network with vertical profile retrievals of CO2 from the Tropospheric Emission Spectrometer (TES) satellite instrument on the Aura spacecraft is investigated, with a focus on the North American biosphere. Although the accuracy and precision of the TES CO2 data are lower than those of the surface data and the sensitivity of TES peaks in the mid-troposphere, we demonstrate that incorporating TES observations with surface data can provide additional constraints to reduce the uncertainty in estimates of CO2 sources and sinks.
Development of an Integrated Hyperspectral Imager and 3D-Flash LADAR for Terrestrial Characterization
The characterization of terrestrial ecosystems using remote sensing technology has a long history with using multi-spectral imagers for vegetation classification indices, ecosystem health, and change detection. Traditional multi-band imagers are now being replaced with more advanced hyperspectral imagers, which offer finer spectral resolution and more specific characterization of terrestrial reflectances. Recently, 3- dimensional (3D) imaging technologies, such as radar interferometry and scanning laser rangers, have added a vertical dimensional to the characterization of ecosystems. The combination of hyperspectral imagery with 3D LADAR allows for detailed analysis of terrestrial biomass, health and species identification. Recognizing the need, and the technical feasibility of this type of environmental assessment, the National Research Counsel has advocated two future NASA satellite missions to measure terrestrial ecosystem health and structure, the DESDynI and HyspIRI missions. These programs will orbit synthetic aperture radar, LADAR and hyperspectral imagers. To mitigate program risk it is desirable and prudent to first demonstrate the integration of these instruments on an airborne platform. Although systems developed for separate purposes have been flown on a single aircraft, the requirements and performance of a dual sensor system has not yet been developed nor integrated as a single unit. We demonstrate a development pathway from an aircraft platform with an integrated sensor suite, using a hyperspectral imager and a laser ranger for a comprehensive remote sensing characterization of terrestrial ecology.
Integrating Landscape-scale Forest Measurements with Remote Sensing and Ecosystem Models to Improve Carbon Management Decisions
Managing forests to increase carbon stocks and reduce emissions requires knowledge of how management practices and natural disturbances affect carbon pools over time, and cost-effective techniques for monitoring and reporting. This study improves upon the methodology to collect and integrate the multi-tier monitoring data from the North American Carbon Program (NACP) with management decisions by systematically scaling up intensive forest carbon measurements to land management areas (or landscapes), and reconciling these estimates with ecosystem models and decision-support systems that are driven by remote sensing and national inventories. We propose to use spatial analysis techniques and an ecosystem process model (PnET- CN) to scale up and map observations from flux towers, landscape biometrics, and inventories to areas of approximately 2500 km2 around flux tower sites. The NASA-CASA model is used to derive estimates for the same areas from remote sensing observations by the MODIS sensor, and biophysical maps. We compare and reconcile the top-down and bottom-up approaches, then use the mapped estimates of productivity and biomass that embed consequences of land disturbances and forest age structure as input to decision-support tools. Key information for the decision-support tools includes (1) estimates of carbon stocks and quantified impacts of management activity; (2) estimates of net ecosystem production (NEP) and changes in carbon pools; and (3) estimates of forest/atmosphere carbon fluxes and relevant effects from various environmental controls. This work is relevant to land managers and climate change policy because it supports a need to estimate and report carbon stocks and changes in carbon stocks to state, regional, national, and private greenhouse gas registries. This work builds upon a foundation of work begun in 2001 by the U.S. Forest Service to implement a forest carbon monitoring and observation system at intermediate or "Tier 3" sites as described in the North American Carbon Program (NACP) science plan.
Modelling Trace Gas Gluxes From Soils Along Slope Transects in Eastern Canadian Forest Ecosystems
Trace gases exchange between forest soils and atmosphere is important for Green House Gases (GHG) budget at both national and global scales. However, this part has not been included in the Canadian national GHG inventory yet. Although several sites have flux measurements, accurate larger scale fluxes are difficult to estimate due to the nature of spatial and temporal variability for gas generation and consumption processes. This study aims to simulate trace gas fluxes (mainly CH4 and N2O) at different landscape scales, using the process-based Forest-DNDC model. As the first part of the work, model has been parameterized and validated against flux measurements along slope transects at two deciduous forest stands at Mt. St. Hilaire (MSH) and Morgan Arboretum (MA), near Montreal, Quebec. The preliminary results suggest that both N2O and CH4 fluxes have two peaks during a year corresponding to snow melting and summer rain, respectively. The upland, riparian zone and wetland area at the two sites are sources of N gases dominated by N2O and N2. Both upland area and riparian zone act as a CH4 sink, while wetland a net source of CH4. Forest successional stage, tree species composition and the decomposition status of forest floor are possible controls on N gaseous flux at site level. The validation results proved that Forest-DNDC is able to capture trace gas exchange in different forest soils. The modeled patterns and magnitudes of fluxes were basically in agreement with observations for all studied plots. Further works for parameter optimization with inversion techniques are expected to improve the model performance before extrapolating the model to larger scale.
Interaction of Insect Defoliation, Wildfires and Climate Change on Carbon Dynamics
We assess and predict the interactive effects of gypsy moth defoliation, fire management, and climate change on carbon uptake, forest productivity, species composition, and tree mortality in the New Jersey Pine Barrens. This effort will combine carbon flux measurements, a forest landscape disturbance model, and field monitoring data. We will determine how interactions among these disturbances affect current management and potential carbon management goals. The LANDIS-II forest landscape simulation model in this study uses three model extensions or modules: the Dynamic Fire System (DFS) extension, the Biomass Succession extension, and an insect defoliation extension. Parameterization of the DFS and the Biomass Succession extension uses new and existing data sources for the study area. This includes flux tower data from three upland forest types, for annual net ecosystem exchange of carbon taken before and after defoliation as well as during prescribed burns. An intensified grid of FIA-type plots around each tower (up to 24 plots per tower) provides additional biometric information. The study conducted a field mortality survey and canopy foliar analysis to understand the process of forest decline with insect defoliation. This project provides a predictive framework for working through landscape to regional management scenarios in areas with multiple, interacting management priorities that can be applied across the US, especially in areas where both insect and fire disturbances occur.
Atmospheric Measurements of Co-variation Carbonyl Sulfide and Carbon Dioxide Provide Information on Carbon Cycle Processes
Carbonly sulfide (OCS), an analog of carbon dioxide, is emerging as a useful atmospheric tracer of the terrestrial carbon cycle. Previous studies have shown that OCS is taken up by leaves and soils. The principle source of OCS to the atmosphere is oxidation of sulfur compounds produced in the oceans. Flask samples and IR absorption spectroscopy are used to measure OCS concentration in the global atmosphere. Over the continents there is strong depletion of OCS in the atmospheric boundary layer during the growing season. We show that this uptake is consistent with the proposed linkage of OCS up take with gross primary production (GPP) and we show that the ratio of OCS to CO2 depletion is consistent with estimates of GPP and ecosystem respiration based on eddy correlation studies in the same region. We have incorporated OCS into a framework for carbon cycle modeling that is being used for top-down data assimilation analysis of the carbon cycle. We use this to demonstrate the added value of OCS measurements to to carbon cycle inversions.
Estimation Of The Carbon Exchange At Different Northern Peatlands Using Modis Vegetation Indices
Satellite remote sensing has the potential for extracting information related to the phenology of carbon exchange at regional and global extents. A good relationship between remote sensing-derived phenology and eddy covariance (EC) flux tower-derived phenology is important for the extrapolation of the carbon exchange phenology over large regions. Yet, these relationships remain largely unexplored for northern hemisphere peatlands. In this study, we evaluate the potential of MODIS-derived vegetation indices for the estimation of the carbon exchange and phenology for two peatlands in Canada. MODIS vegetation indices, including the normalized difference vegetation index (NDVI) and the simple ratio (SR) were compared with EC flux derived gross ecosystem productivity (GEP) and net ecosystem exchange (NEE) time series from an ombrotrophic bog and a treed fen in Canada, between 2002 and 2006. Several phenological metrics including the start of the growing season (SGS), the annual peak photosynthetic activity, the carbon uptake period (CUP), the spring increase and the autumn decrease rate and the area under the time-series curve (small integral and large integral) were extracted from these datasets using the TIMESAT software. Our preliminary results show that: 1. the NDVI-derived spring increase rate, the SR-derived SGS date, and the SR-derived small integral are related to the EC tower-derived GEP (r2 =0.81, r2 =0.90, r2 =0.86, respectively); 2. The SR-derived spring increase rate, the SR-derived SGS date, and the SR-derived small integral are related to the EC tower- derived annual NEE (r2 =0.88, r2 =0.98, r2 =0.86, respectively); 3. The SR index is more linearly related to the EC-tower derived GEP and NEE than the NDVI and shows potential for the estimation of daily carbon exchange measures; 4. Different pixel resolutions (250 m vs. 1000 m) do not significantly affect the SR- NEE and SR-GEP relationships. Future work will use these findings to map carbon exchange and phenology for peatlands across Canada. Additionally, the principal drivers behind the spatio-temporal variability in the carbon dynamics at these sites will be explored.
Using DayCENT to Simulate Carbon Dynamics in Conventional and No-till Agriculture
This study investigated how conventional tillage(CT) and no-till(NT) practices may affect the long-term dynamics of carbon. In this paper we examine the role of decomposition in influencing the carbon budget in a long-term cropping system and briefly discuss the potential for agricultural tillage practices to mitigate increases in atmospheric carbon dioxide concentrations. We employ the DayCENT model, eddy covariance measurements and a chamber experiment to investigate the net ecosystem exchange(NEE), heterotrophic respiration(Rh) and soil organic carbon(SOC) in the growing and non-growing seasons from 2000 to 2008 in southern Ontario, Canada. In order to provide information on the impact of tillage applications on decomposition, we ran simulations for CT and NT plots using a modified DayCENT model for a 9-yr time period. We parameterized and validated an equilibrium simulation for 5000 years to ensure that the model was stable, followed by a simulation of the study years 2000-2008 using observed meteorological data and known site properties and management practices. Our investigation suggests that no-till cropping will result in the largest reduction in soil CO2 emissions, while rewetting in relatively dry soils will result in a significant increase in soil CO2 emissions. Furthermore, The NEE also was strongly influenced by environmental factors, agricultural schedule and plant species, however there are no significant differences between CT and NT plots in our studies. Simulations with a modified DayCENT model indicated that favorable growing conditions during the 9 years could account for a 10 g C/m2/year increase in soil organic carbon at a site in Elora, Ontario.
NOrth AMerica Soil (NOAM-SOIL) Database
NOAM-SOIL is being created by combining the CONUS-SOIL database with pedon data and soil geographic
data coverages from Canada and Mexico. Completion of the in-progress NOrth AMerica Soil (NOAM-SOIL)
database will provide complete North America coverage comparable to CONUS. Canadian pedons, which
number more than 500, have been painstakingly transcribed to a common format, from hardcopy, and key-
entered. These data, along with map unit polygons from the 1:1,000,000 Soil Landscapes of Canada, will be
used to create the required spatial data coverages. The Mexico data utilizes the INEGI 1:1,000,000 scale soil
map that was digitized by U. S. Geological Survey EROS Data Center in the mid 1990's plus about 20,000
pedons. The pedon data were published on the reverse side of the paper 1:250,000 scale Soil Map of Mexico
and key entered by USDA and georeferenced by Penn State to develop an attribute database that can be linked
to the 1:1,000,000 scale Soil Map of Mexico based on taxonomic information and geographic proximity. The
essential properties that will be included in the NOAM-SOIL data base are: layer thickness (depth to bedrock or
reported soil depth); available water capacity; sand, silt, clay; rock fragment volume; and bulk density. For
quality assurance purposes, Canadian and Mexican soil scientists will provide peer review of the work. The
NOAM-SOIL project will provide a standard reference dataset of soil properties for use at 1km resolution by
NACP modelers for all of North America. All data resources, including metadata and selected raw data, will be
provided through the Penn State web site: Soil Information for Environmental Modeling and Ecosystem
Management (www.soilinfo.psu.edu). Progress on database completion is reported.
The Value of Agricultural Data in Carbon Cycle Studies: A Case Study from Ontario
Agricultural activities can play a significant role in the terrestrial carbon cycle by either serving as a source or sink of carbon dioxide (CO2). Specifically, cultivation can act as a sink when carbon is photosynthetically fixed and sequestered into soils during crop growth. Ontario is one of the major field crop producing regions in Canada, accounting for 55.2% of total national production of corn and 60% of winter wheat during 2006. As a result, understanding Ontario's agricultural net primary production (NPP) is of great importance to resource managers and policy makers when managing risks and opportunities arising from climate change. In this study, Ontario's agricultural data were used in order to determine the following: (1) relationships between Ontario's crop NPP and climatic variables such as temperature, precipitation, as well as climate indices; and (2) uncertainties in a carbon flux model - the Vegetation Photosynthesis and Respiration Model (VPRM). VPRM is a data-driven, diagnostic biospheric carbon model designed for data assimilation purposes. To calculate crop NPP, two different published approaches that relied on using known yield information along with a series of crop-specific parameters and coefficients were applied. Then, environmental factors and spatiotemporal changes of crop NPP were examined statistically through exploring historical weather records of the province. With respect to the NPPs' interannual variability, comparison of their times series and that of climate indices such as the multivariate El-Nino-Southern Oscillation (ENSO) index and others during 1987 to 2007 found that NPP values were fairly positively correlated with the East Atlantic (EA) pattern. In fact, this teleconnection pattern showed stronger correlations than ENSO and North Atlantic Oscillation (NAO). Comparisons of the crop data against VPRM estimates revealed biases in VPRM simulations. The model underestimated fractions of cropland areas and overestimated annual gross primary productivity (GPP). We discuss reasons for these errors.
Modeling canopy-level productivity: is the "big-leaf" simplification acceptable?
The "big-leaf" approach to calculating the carbon balance of plant canopies assumes that canopy carbon fluxes have the same relative responses to the environment as any single unshaded leaf in the upper canopy. Widely used light use efficiency models are essentially simplified versions of the big-leaf model. Despite its wide acceptance, subsequent developments in the modeling of leaf photosynthesis and measurements of canopy physiology have brought into question the assumptions behind this approach showing that big leaf approximation is inadequate for simulating canopy photosynthesis because of the additional leaf internal control on carbon assimilation and because of the non-linear response of photosynthesis on leaf nitrogen and absorbed light, and changes in leaf microenvironment with canopy depth. To avoid this problem a sunlit/shaded leaf separation approach, within which the vegetation is treated as two big leaves under different illumination conditions, is gradually replacing the "big-leaf" strategy, for applications at local and regional scales. Such separation is now widely accepted as a more accurate and physiologically based approach for modeling canopy photosynthesis. Here we compare both strategies for Gross Primary Production (GPP) modeling using the Boreal Ecosystem Productivity Simulator (BEPS) at local (tower footprint) scale for different land cover types spread over North America: two broadleaf forests (Harvard, Massachusetts and Missouri Ozark, Missouri); two coniferous forests (Howland, Maine and Old Black Spruce, Saskatchewan); Lost Creek shrubland site (Wisconsin) and Mer Bleue petland (Ontario). BEPS calculates carbon fixation by scaling Farquhar's leaf biochemical model up to canopy level with stomatal conductance estimated by a modified version of the Ball-Woodrow-Berry model. The "big-leaf" approach was parameterized using derived leaf level parameters scaled up to canopy level by means of Leaf Area Index. The influence of sunlit/shaded leaf separation on GPP prediction was evaluated accounting for the degree of the deviation of 3-dimensional leaf spatial distribution from the random case. More specifically, we compared and evaluated the behavior of both models showing the advantages of sunlit/shaded leaf separation strategy over a simplified big-leaf approach. Keywords: canopy photosynthesis, leaf area index, clumping index, remote sensing.
Early Field Results for a new Method of Measuring CO2 Efflux
Soil respiration vastly outweighs the total anthropogenic contributions of CO2 to the atmosphere. Permafrost soils alone contain between twenty to sixty percent of all soil carbon. Accelerated decomposition of soil carbon due to global warming could lead to a critical positive feedback loop that pushes global greenhouse gas concentrations ever higher and accelerates global warming. For this reason soil CO2 monitoring at high latitudes is important, but continues to pose instrumental and methodological challenges. Using a flow-through probe that 'forces' diffusivity along with a mathematical solution adapted from heat flow research we have developed an instrument requiring only 3.5W of power and containing no moving parts. This method allows for the conversion of single point concentration time series' to fluxes in real time. Deployable in wide range of environments, including below snow measurements this instrument allows for continuous long- term measurement of soil respiration in permanent and temporary monitoring campaigns.All relevant theory to this method along with data gathered from two sites will be presented here. Site one is a wind-swept study site with little snow accumulation, and site two has a considerable snowpack accumulation, where we tested our technique for measuring soil and sub snowpack fluxes during the late winter and spring. Each site offers a different monitoring challenge and required slightly different instrumental and mathematical approaches. Here we present initial findings and lessons learned from these two successful field deployments.
Verification of the Flux Solution Technology and Quantification of Errors Associated With Calibration and Implementation
Soil systems are among the largest contributors of carbon to the atmosphere, emitting sixty GT each year. Soil flux monitoring has become increasingly important for determining the magnitude and climatic sensitivity of this CO2 source, and specific monitoring challenges are associated with high latitude environments where monitoring is difficult. Our Flux Solution technology is based on a novel mathematical approach for measuring soil surface gas surface emission. Instead of headspace accumulation chambers, vertically spaced gas wells, or other hardware-intensive approaches commonly used for land based measurements, our flux calculations are made simply using a timeseries of concentration measurements at a single point. This is done using a mathematical approximation originally developed for geophysical heat flow research, which we have adapted to gas diffusion. For soil CO2 work, we have embodied this mathematical technique in a simple, reliable, and inexpensive flow-through membrane diffusion probe, tuned to impose a very specific known diffusive regime at the soil surface. Before such an instrument can be deployed on a large scale however, the errors associated with this technique, as well as its correspondence with established methods, need to be rigorously determined. This poster presents some initial analysis of various lab and field deployments. In all cases we establish a characteristic relationship between the Flux Solution measurements and those of accepted techniques by using RMS comparison, regression analysis, contrasting their temporal behaviors, comparing the amplitude of the diurnal cycle and various other methods. The sensitivity of this correspondence to changes in soil parameters such as diffusivity and moisture content is thoroughly examined, as well as the measurement errors associated with our technique and calibration mechanism. Early results are very promising, and this type of rigorous study helps focus further development of the Flux Solution technology.
Influence of sunlight and exchange across the sediment water interface on nitrogen and carbon within agricultural streams
Carbon and nitrogen cycling within agricultural streams is controlled by a combination of biogeochemical and physical transport processes. Within streams, net metabolism (balance between photosynthesis and respiration) varies as a function of sunlight, resulting in diurnal fluctuations in dissolved oxygen and carbon dioxide. Here, we explored the impact of photosynthesis versus respiration on carbon and nitrogen. The experiment involved 4 chambers, 2 chambers with light and 2 chambers without light. For each set, one chamber was closed bottom and one chamber was open to the bed sediments. Prefiltered water in combination with a conservative tracer were injected into each chamber for 12-hours. Our results show that biological activity on the bed sediments alters carbon lability and carbon and nitrogen concentrations. Using our tracer data, we will assess the impacts of light variation in the surface waters and quantitatively examine N and C production / removal during transport from the surface into the subsurface.
B13C-15 [Moved to B23A]
Non-Growing Season Dynamics of Nitrous Oxide Emissions From Cropped Land in Southern Ontario, Canada
As atmospheric nitrous oxide (N2O) is increasing at a rate of 0.3% per annum and has a global warming potential 300 times greater than that of carbon dioxide (CO2), it is crucial to understand the dynamics of anthropogenic emissions of this greenhouse gas. Although agriculture represents a small proportion of Canadian land use, it is the most significant source of both human-derived and natural N2O emissions in this country. More than 50% of annual N2O emissions from agricultural soils in northern latitudes may take place in the non-growing season (NGS). Southern Ontario is one of the most intensively farmed regions of Canada. Greater understanding of NGS dynamics of N2O flux will help to develop more accurate N2O modeling methodologies for this region and may help in the refinement of agricultural practices which reduce N2O emissions. This study investigated field and lab NGS dynamics of agricultural N2O flux in southern Ontario, where anaerobic conditions, such as those caused by snowmelt, icemelt and rain, promote the reduction of soil nitrate (NO3-) to N2O. Gas samples were collected from permanent soil gas collars in nine field sites following winter and spring thaw events over the NGS of 2007-2008; otherwise samples were collected at biweekly intervals from these collars, or by means of snow gas chambers when the sites were snow-covered. The field sites were situated in three working farm fields where soils were subjected to typical tillage, amendment and cropping practices. Field data were supplemented by experimental results from soil cores taken from one of two fields previously seeded to corn, and subjected to simulated freeze-thaw cycles of average frequency, duration, and amplitude found in this region. The experimental data show a highly significant positive correlation between N2O flux and soil temperature at 5 cm depth (p < 0.0005). Overall, the field data show a positive correlation between N2O flux and soil temperature (p < 0.025). The less significant relationship observed in the field data may be due to the relatively low sampling frequency, the timing of sampling, and/or confounding variables that influenced N2O flux, such as in situ variability of soil macropore and labile carbon distribution. As such, the laboratory experiment was able to provide a more controlled and intensive examination of the dynamics between N2O flux and climate during the key thaw periods in this agricultural region.
Spatial and Temporal Alterations on Carbon and Water Cycles Due to Grazing
Grasslands are vital in the carbon cycle, as large amounts of carbon are stored in the soils of the prairie. As climate change affects the carbon cycle, it is essential for the agricultural communities to understand the impacts of these changes on farming practices such as grazing and meat production. The objective of this study is to determine the effect of grazing on the carbon cycle by characterizing the surface boundary layer of both a grazed field and an ungrazed field. Data were collected from open path eddy covariance systems over Rannells Flint Hills Prairie Preserve in north-central Kansas, one over an ungrazed field and one over a grazed field. Cospectra of fluxes of CO2, heat, water, and momentum for July 2007 were compared to assess the size of eddies contributing energy to each field. For CO2, the cospectra for both the ungrazed and the ungrazed field were similar. For all of the other fluxes, lower frequency eddies contributed more energy in the grazed field than the ungrazed field. By using a footprint model, the contributing source areas were determined for fluxes from May through October of 2007. The grazed field had a larger distance of contribution in both stable and unstable atmospheric conditions. Implications of this study include the alterations on fields and impacts on the carbon and water cycles as a result of grazing.
Importance of CO2 Evasion From Small Boreal Streams
Boreal lakes are known to be super-saturated with CO2 and significant sources of CO2 to the atmosphere. In this study it is shown that small boreal streams are relatively more super-saturated with CO2 and represent an important source for the evasion of CO2 to the atmosphere from boreal catchments. Small boreal streams account for an annual average of 3.2% and a seasonal high of 10.5 % of the total CO2 evaded from streams + lakes. A seasonal pattern exists with evasion from streams being maximal during the summer, coinciding with maximum primary production, and least during the spring. Evasion of CO2 is not expected during the winter when streams are ice- and snow-covered. The evasion of CO2 by small boreal streams represents 87% of the total CO2 exported by the streams and 10% of all forms of C (DOC + DIC) exported. Knowledge of this important transfer and loss mechanism for CO2 will allow for more accurate modeling of boreal catchment C budgets.
The Net Impact of Hydroelectric Reservoir Creation on Greenhouse Gas Emissions: A Study of the Eastmain-1 Reservoir in the Eastern James Bay region of Quebec, Canada
In order to satisfy present and future energy demands and to minimize greenhouse gas (GHG) emissions, there is a growing need to develop energy sources that are not based on combustion. In the boreal regions of Canada, there is a huge potential for hydroelectricity production. However, in most cases, large areas of the boreal ecosystem must be inundated to create hydroelectric reservoirs. Previous studies have established that reservoirs emit GHGs, but these studies have typically focused on emissions some years after reservoir creation. The critical question that has not been asked is 'what is the net change in the exchange of GHG that results directly from the creation of the reservoir?' - i.e. 'what is the net difference between the landscape scale exchange of GHGs before and after reservoir creation, and how does that net difference change over time from when the reservoir was first created to when it reaches a steady-state condition?'. The Eastmain-1 (EM-1) hydroelectric reservoir, located in the James Bay region of Quebec was created in late 2005 and provides a tremendous opportunity to study the impacts of reservoir creation on GHG emissions which are still largely unknown for this type of land conversion. The creation of the EM-1 hydroelectric reservoir required the flooding of over 600 km2 of the boreal ecosystem along the Eastmain River, of which 65% was occupied by forest, 14% by peatland, and 21% by lakes and rivers. In order to assess the impacts of the creation of the reservoir on GHG emissions, three eddy covariance (EC) tower flux sites were established in a black spruce forest, peatland and on an island in the reservoir itself to measure continuous net ecosystem exchange (NEE) of CO2. Together, these represent the dominant terrestrial pre-flooded (forest and peatland) and post-flooded (reservoir) environments. The forest and reservoir EC systems were installed and operational by the end of summer 2006 with the peatland site coming on-line summer of 2008. Through the use of the forest and peatland analogue sites, the EC results will be used to evaluate the pre-flooded vs. post- flooded CO2 fluxes, and thus the net impact of the EM-1 reservoir creation in terms of CO2 emissions. By measurement and modeling, we will provide an estimate of the change in GHG source the atmosphere would see, an estimate of the net emissions that can be used for intercomparison of GHG contributions with other modes of power production and a basis on which to develop biogeochemically sound, verifiable, and transparent estimates for GHG accounting. This presentation will provide an overview of the project and its goals and will discuss preliminary results from the EC and terrestrial measurement campaigns.