B51C-0209
Estimating NEP by micrometeorological and biometric methods in a larch forest in Hokkaido, Japan
We report comparison of NEP of a temperate larch forest between micrometeorological and biometric methods. The study site is a Japanese larch plantation in Hokkaido, Japan ($42°44$^{'}$N, $141°31$^{'}$E). The canopy height was approximately 15 m and tree age was about 45 years old. Canopy LAI reached the maximum at 5.5m$^{2 }$m$^{-2 }$ in July. We measured NEE (-NEP) using eddy covariance method with open- and closed-path systems. We estimated biometric NEP from biomass increment, litter fall of overstory and understory, and heterotrophic respiration measured by chamber method. However, Biometric NEP should have a large uncertainty, because we neglect a fine root production. Daytime NEE tended to be slightly negative for open-path than for closed-path. In contrast, nighttime NEE tended to be slightly less positive for open-path than for closed-path. Consequently, the small difference of half-hourly NEE between from open- and closed-pass provided the large annual bias. High frequency loss corrections using Bandpass covariance method or experimental transfer function improved the daytime bias. However, they made the nighttime bias large. The annual NEPs obtained from open- and closed-path systems were compared with that estimated by the biometric method. NEP estimated by biometric method was close to that estimated by closed-path system.
B51C-0210
Regional Patterns of Ecosystem Response to Climate Change and Elevated CO2 Levels in the Great Plains
This paper will describe the results and procedures used to simulate the potential impact of climate change and elevated CO2 levels on ecosystem properties (soil carbon, trace gas fluxes, plant production, and nitrogen cycling) for grasslands in the Central and Southern Great Plains region. Extensive weather, soils and land use data is currently available at regional scales in the US. This paper will describe the procedure used to use this data as input into the Daycent ecosystem model and simulate the potential impact of elevated CO2, and climate changes on grassland ecosystem properties at the regional scale. Extensive field level elevated CO2 experiments from the Southern and Central Great Plains region were used to test and calibrate the Daycent model's ability to simulate ecosystem responses to elevated CO2 levels. The model results show that elevated CO2 levels will result in increased above and belowground plant production, and plant nitrogen uptake for the whole region, while soil carbon levels are decreased in the Western Shortgrass Steppe parts of the Great Plains and increased in Eastern Tallgrass Prairie region. The enhanced plant nitrogen uptake results from a reduction in nitrogen trace gas fluxes and enhanced nitrogen uptake from the soil. The enhanced plant nitrogen uptake from the soil results from elevated CO2 induced increases in soil water content which increase the soil carbon turnover rate. Increasing air temperatures causes a general pattern of decreased soil carbon levels in the whole region. Increasing temperatures has little impact on plant production in the Shortgrass Steppe Region and small increases in plant production in the Tallgrass Prairie Region.
B51C-0211
Where do Fossil Fuel Carbon Dioxide Emissions from the Western U.S. Go? An Analysis Based on an Atmospheric Model Validated Using Radiocarbon Observations (A Component of NACP-W)
We describe a combined measurement and modeling approach to assess the fate of fossil fuel CO2 emissions released within California. In our analysis, we compared the $^{14}C$ content of California annual C3 grasses with spatially and temporally resolved fossil fuel CO2 concentration estimates obtained using the MM5 atmospheric model and emissions inventories. Annual grasses sampled at the end of the growing season represent a weighted (with photosynthetic uptake) growing season atmospheric CO2 sample. Their $^{14}C$ content can be used to infer the fraction of assimilated carbon associated with fossil fuel since these CO2 emissions are completely depleted in $^{14}C$. We sampled grasses at about 100 sites distributed throughout California, with special foci in the Central Valley, Los Angeles, and San Francisco air basins. Grass $^{14}C$ content ranged from ~10 to ~60\permil in densely and sparsely populated areas, respectively. To analyze these $^{14}C$ data, we applied a model that couples meteorological transport, ecosystem CO2 exchange, and fossil fuel CO2 emissions. The patterns of annual grass $^{14}C$ content are consistent with our estimated fossil fuel CO2 emissions and predicted atmospheric transport. We then applied the coupled model to assess fossil fuel CO2 transport pathways out of the region. Accumulated over the year, about half of the fossil fuel CO2 emitted in California exits to the south within the atmospheric boundary layer (ABL) while most of the remaining CO2 exits in a disperse and lofted plume to the east. These results suggest that atmospheric sampling programs designed to enable continental 'top down' inversions of North American sources and sinks should include an accurate assessment of north to south flow components, transport within the ABL, and exchange between the ABL and free troposphere.
B51C-0212
Quantifying Belowground Carbon Allocation in the Northeastern United States
Forest soils represent a substantial component of the terrestrial carbon cycle and are an important research area for a number of carbon cycle science initiatives. Whereas patterns of aboveground productivity have been relatively well measured and are increasingly included in regional-scale model analyses, belowground estimates are still highly uncertain and progress has been hampered by a variety of methodological difficulties. The lack of data poses a problem because belowground measurements are needed to create a complete carbon budget for terrestrial ecosystems at local, regional and global scales. Ecosystem carbon balances will help identify how and where carbon is being stored, as well as how that might change as forests grow, die back, or transition into different forest types as a result of climate changes. This study focuses on quantifying belowground carbon allocation in the Bartlett Experimental Forest (BEF) of the New Hampshire White Mountains, and examining the degree to which spatial patterns can be related to patterns of soil and canopy nitrogen status. The work is part of a landscape-scale North American Carbon Program (NACP) study currently taking place at the BEF. Belowground carbon allocation can be estimated by subtracting soil respiration from litter (leaf, branch) measurements. Litter and soil respiration are being measured at two scales within the study area. The first includes a 1km2 area around an eddy flux tower at BEF, and is part of the intensive NACP study. Additional plots are distributed throughout the broader landscape to capture a greater degree of variation in vegetation, soils and topography. The goals of the project are (1) to contribute the belowground carbon portion to the total ecosystem carbon budget of BEF, and (2) to extrapolate soil carbon from the plot level to landscape and regional scales using remote sensing of foliar N.
B51C-0213
Estimation and Validation of Regional Biomass Using Vegetation Structure Measurements from ICESAT and Forest Inventory Analysis Data
The ICESAT mission has acquired lidar measurements of land surface vertical structure for several years. Although these have severely limited spatio-temporal resolutions and coverages, they nonetheless are the only consistent, global data of this kind. Our earlier work has linked airborne lidar measurements with a height-structured ecosystem model, the Ecosystem Demography (ED) model for modeling local (10's of kilometers) carbon stocks and fluxes at 1 hectare grid resolution. In this research we explore extending these modeling efforts to link space-based lidar observations of vegetation structure with ED to produce estimates of biomass at regional scales. Sparse ICESAT data are first combined with airborne lidar and MODIS-derived landcover information to provide contextual information for each observation, and to provide appropriate estimates of vegetation structure at 1° resolution. Next, to assess their relationship with known biophysical attributes, we perform exploratory analyses linking various ICESAT lidar waveform metrics with biomass, age and other vegetation data obtained from the U.S. Forest Inventory Analysis (FIA) for 1° grid cells. Lastly, we explore methods for initialization of the ED model using ICESAT data to produce regional estimates of biomass for the eastern half of the US, which are then compared with gridded FIA estimates of biomass.
B51C-0214
Methodologies for Estimating Forest Carbon Stocks From Annual Forest Inventory Data
The USDA Forest Service Forest Inventory and Analysis (FIA) program performs an annual forest inventory that includes measurements of down woody material and soil quality. These measurements are taken on a systematic nation-wide grid of approximately 125,000 plots where each one may represent up to 38,850 ha of forest. Between ten to twenty percent of these plots are measured every year. Tree measurements include species, height, and diameter at breast height. Carbon content of live tree biomass is estimated with carbon conversion constants. Down woody material carbon stocks (coarse and fine woody debris) are estimated using line intercept transects and carbon conversion constants. The soil quality indicator includes volumetric sampling of the forest floor and the collection of mineral soil cores representing depth increments of 0 - 10 cm and 10 - 20 cm. Carbon content of the soil samples is determined by dry combustion. We combined tree, down woody material, and soil measurements collected over three years (2001 - 2003) to evaluate current methodologies for estimating the total carbon sequestered in forests of the United States. Carbon storage by pool is roughly ranked as follows: standing tree biomass > 0 - 10 cm mineral soil > 10 - 20 cm mineral soil > coarse woody debris > the forest floor ~ fine woody debris. The tree measurements actually occur on a denser sampling network than either the down woody material or soil measurements. Addition research is underway to determine effective methodologies for estimating the carbon estimates from these two stocks at all tree measurement locations.
B51C-0215
Novel Technique for Remote Estimation of Gross Primary Production in Crops: Implications for the Synoptic Monitoring of Vegetation Productivity
Accurate estimation of spatially distributed CO2 fluxes is of great importance for regional and global studies of carbon balance. We have found that in irrigated and rainfed crops (maize and soybean) mid-day GPP is closely related to total crop chlorophyll content. We applied a recently developed technique for remote estimation of crop chlorophyll content to assess gross primary production (GPP). The technique is based on reflectance in two spectral channels: the near-infrared and either the green or the red-edge. The technique provided accurate estimations of mid-day GPP in both crops under rainfed and irrigated conditions with root mean square error of GPP estimation of less than 0.3 mg CO2/m2s in maize (GPP ranged from 0 to 3.1 mg CO2/m2s) and less than 0.2 mg CO2/m2s in soybean (GPP ranged from 0 to 1.8 mg CO2/m2s). Validation using an independent dataset for irrigated and rainfed maize showed robustness of the technique; RMSE of GPP prediction was less than 0.27 mg CO2/m2s. Given the substantial improvement in the accuracy of GPP estimation by the models developed in this study, as compared to the currently used methods, it is worthwhile to fully explore the efficacy of these techniques over different crops, at different sites. Further validation of this technique over other crops and vegetation types is required using the green and NIR bands of satellite-based systems such as MODIS, Landsat TM and ETM, and Hyperion (onboard EO-1 satellite) as well as the red-edge and NIR bands of satellite systems such as MERIS and Hyperion. With further validation (using data from already established FluxNet and SpecNet sites), this technique may be found useful in a variety of terrestrial ecosystems. The end result may be an inexpensive yet accurate tool for estimating mid-day GPP.
B51C-0216
Evaluation of the Empirical Piecewise Regression Model in Simulating GPP in the Northern Great Plains
For better understanding the carbon fluxes in the grassland ecosystems, an empirical piecewise regression (PWR) model was developed to estimate gross primary production (GPP) for the grassland ecosystems in the Northern Great Plains and Northern Kazakhstan. The PWR model spatially scales up the localized flux tower measurements across an ecoregion at 1-km resolution. In this study, we compared the PWR GPP and the MODIS GPP with five grassland flux tower measurements. Then we employed cross-validation to evaluate the PWR GPP values. We also compared PWR GPP and MODIS GPP for grasslands for the entire study area. Factors that may explain the spatial pattern of the GPP differences between the two models were explored using decision tree technique. The results indicated that the PWR modeling approach was robust with a good agreement (agreement coefficient d=0.71-0.97) between PWR model and tower measurements. Cross-validation showed a relatively low agreement (d=0.71-0.78) at two influential flux tower sites. We also observed that the PWR GPP was lower than or similar to the MODIS GPP in the east and higher in the west and south. Further analysis suggested that percentage of C4 grasses, soil water holding capacity, percentage of clay, and percentage of crop mixed in the grassland contributed to the GPP difference of the PWR and MODIS models.
B51C-0217
Analysis of Anthropogenic CO2 Signal in ICARTT Observations Using a Regional Atmospheric Model
Atmospheric inversion studies of CO2 sources and sinks typically prescribe the anthropogenic component of observations as the result of forward simulations using surface source estimates. The anthropogenic flux is prescribed as a background flux with low uncertainty. Errors in the spatial and temporal estimates of anthropogenic emissions have been shown to result in a bias in regional surface flux estimates. Improved information about the anthropogenic component of atmospheric CO2 observations will reduce the error in such a top-down approach to inferring surface fluxes. In this study, an analysis of the anthropogenic component to atmospheric CO2 measurements is completed using a regional atmospheric chemistry model and its adjoint. Observations of CO2 are made during the ICARTT (International Consortium for Atmospheric Research on Transport) experiment during the summer of 2004 from aircraft platforms. The STEM-2K1 model and its adjoint are applied to characterize the anthropogenic sources influencing the ICARTT observations. Source signatures are obtained using ratios between CO2 and combustion pollutants as well as model derived airmass markers for source region, source types, and airmass age. Model derived influence functions along with assimilated transport model results of anthropogenic tracers are used to characterize the anthropogenic CO2 emissions in the Midwest during the summer 2004 period.
B51C-0218
Forest Carbon Storage in the Northern Midwest, USA: A Bottom-Up Scaling Approach Combining Local Meteorological and Biometric Data With Regional Forest Inventories
Carbon (C) storage increasingly is considered an important part of the economic return of forestlands, making easily parameterized models for assessing current and future C storage important for both ecosystem and money managers. For the deciduous forests of the northern midwest, USA, detailed information relating annual C storage to local site characteristics can be combined with spatially extensive forest inventories to produce simple, robust models of C storage useful at a variety of scales. At the University of Michigan Biological Station (45$^{o}$35'' N, 84$^{o}$42'' W) we measured C storage, or net ecosystem production (NEP), in 65 forest stands varying in age, disturbance history, and productivity (site index) using biometric methods, and independently measured net C exchange at the landscape level using meteorological methods. Our biometric and meteorological estimates of NEP converged to within 1% of each other over five years, providing important confirmation of the robustness of these two approaches applied within northern deciduous forests (Gough et al. 2005). We found a significant relationship between NEP, stand age ({\it A}, yrs), and site index ({\it I}$_{s}$, m), where NEP = 0.134 + 0.022 * (LN[{\it A}*{\it I}$_{s}$]) (r$^{2}$ = 0.50, P < 0.02). Site index is an integrated measure of site quality, expressed as 50 yr canopy height. We then used stand age and site index data from forests of similar species composition reported in the USDA Forest Inventory and Analysis database (ncrs2.fs.fed.us/4801/fiadb/) to estimate forest C storage at different scales across the upper midwest, Great Lakes region. Model estimates were validated against independent estimates of C storage for other forests in the region. At the local ecosystem-level (~1 km$^{2}$) C storage averaged 1.52 Mg ha$^{-1}$ yr$^{-1}$. Scaling to the two-county area surrounding our meteorological and biometric study sites, average stand age decreased and site index increased, resulting in estimated storage of 1.62 Mg C ha$^{-1}$ yr$^{-1}$, or 0.22 Tg C yr$^{-1}$ in the 1350 km$^{2}$ of deciduous forest in this area. For the state of Michigan (31,537 km$^{2}$ of deciduous forest), average uptake was estimated at 1.55 Mg C ha$^{-1}$ yr$^{-1}$, or 4.9 Tg C yr$^{-1}$ total storage. For the three state region encompassing Minnesota, Michigan, and Wisconsin (97,769 km$^{2}$ of deciduous forest), we estimated average storage in these forests of 1.51 Mg C ha$^{-1}$ yr$^{-1}$, or 14.1 Tg C yr$^{-1}$ total storage. This storage represents ~ 13 % of regional anthropogenic C emissions (US Department of Energy, 2003). This modest rate of C storage by forests in the region may decrease due to changes in forest succession and land-Use, and also in response to climate-driven shifts in the balance between photosynthesis and respiration. Gough C.M., Vogel C.S., Schmid H.P., Su H.-B., and Curtis P.S. 2005. Multi-year convergence of biometric and meteorological estimates of forest carbon storage. Agricultural and Forest Meteorology, In Press.
B51C-0219
Quantifying Representativeness Importance Values for AmeriFlux Sites
We are using a multivariate statistical clustering analysis to determine how well the current distribution of sites in the AmeriFlux network is representative of the dominant combinations of vegetation, soils, and climate which are present in the conterminous US. Statistical indices based on multivariate representativeness and site importance indicate how well the current network of towers "samples" the population of flux environments within the nation. The same empirical approach provides a repeatable rationale for the selection of additional flux tower sites by determining any number of additional locations such that the representation of the overall network is maximized by their addition. A representativeness importance value for each existing eddy covariance tower to the AmeriFlux network can be calculated. We have statistically created a series of nine sets of flux-relevant ecoregions which divide the conterminous U.S. into a set of areas within which the carbon flux from terrestrial ecosystems is expected to be relatively uniform and homogeneous. Starting with digital GIS layers of factors deemed important in regulating carbon fixation and loss from terrestrial ecosystems, we assembled a set of maps of multivariate factors which describe and characterize the flux environment in each map cell. Then, we used a k-means clustering procedure to classify each map cell into a particular group whose cells have sufficiently similar environments. Because there were as many as 30 environmental descriptors, each with nearly 8 million cells, it was necessary to perform the clustering process on a parallel supercomputer. Because the statistical process is quantitative, the similarity of a selected flux-ecoregion to every other ecoregion in the map can be calculated. Maps can be produced that show the degree of similarity to the chosen flux-ecoregion as a series of gray shades. By sequentially selecting flux ecoregions currently containing an AmeriFlux tower, maps showing the geographic area which is represented by measurements from that flux tower will be produced.
B51C-0220
A Landsat Record of North American Forest Disturbance
The Landsat Ecosystem Disturbance Adaptive Processing System (LEDAPS) is generating a decadal, wall-to-wall analysis of forest disturbance and recovery from Landsat satellite imagery for the period 1975-2000. The intent is to provide an accurate, high-resolution view of forest disturbance to support biogeochemical modeling and carbon accounting for the North American Carbon Program (NACP). Through the NASA Science Data Purchase program, substantially cloud-free Landsat MSS, TM, and ETM+ data were selected from the global archive and orthorectified to a UTM map base. The LEDAPS project has calibrated and atmospherically corrected these data (~2100 TM and ETM+ scenes to date) using the MODIS/6S radiative transfer approach. Forest disturbance and recovery is then calculated from the surface reflectance images using change detection techniques. An empirical spectral index (the 'Disturbance Index') is used to classify pixels into classes exhibiting high rates of biomass loss over ten years (disturbance) or high rates of biomass gain (recovery). Initial results from North America show good correlation with areas of known harvest activity (Southeastern US, Maine, Pacific Northwest) and fire activity (Boreal forests). Additional work is concentrating on the use of canopy reflectance models to quantify changes in canopy properties in order to identify more subtle changes due to partial harvest and thinning. Initial versions of the surface reflectance and Disturbance Index products were released during 2005 (http://ledaps.nascom.nasa.gov/ledaps/ledaps_NorthAmerica.html).
B51C-0221
GEMS-EDCM Upscaling of Terrestrial Carbon Dynamics From Sites to Regions
Estimating the dynamic evolution of the magnitude, spatial patterns, mechanisms, and uncertainty of carbon sources and sinks at the regional scale is challenging owing to the spatial and temporal variances and covariance of driving variables and the uncertainties in both the model and the input data. We develop the General Ensemble Biogeochemical Modeling System (GEMS) for upscaling carbon stocks and fluxes from sites to regions with measures of uncertainty. GEMS relies on site-scale biogeochemical models (e.g., the Erosion-Deposition-Carbon Model (EDCM) and CENTURY) to simulate the carbon dynamics at the site scale. The spatial deployment of the site-scale model in GEMS is based on the spatial and temporal joint frequency distribution of major driving variables (e.g., land cover and land use change, climate, soils, disturbances, and management). At the site scale, GEMS uses stochastic ensemble simulations to incorporate input uncertainty and to quantify uncertainty transfer from input to output. Using data assimilation techniques, GEMS simulations can be constrained by field and satellite observations or census data including estimates of net primary productivity from the Moderate Resolution Imaging Spectroradiometer (MODIS), grain yield and cropping practices, and forest inventories. The modeling philosophy embedded in GEMS makes it ideal for incorporating and assimilating information with various uncertainties at a range of spatio-temporal resolutions. The GEMS-EDCM has been applied to quantify the spatial and temporal distributions of the terrestrial carbon sources and sinks in the conterminous United States and in Africa.
B51C-0222
Determination of in Situ Rates of Methane Production and Oxidation From Terrestrial Wetlands
Wetlands are responsible for over 70% of non-anthropogenic methane emissions. We present a method, using the δ13C of CO2 in pore water, to obtain the in situ rates of methanogenesis occurring beneath the wetland surface. This method allows us to distinguish methanogenesis from methane oxidation during escape, both of which contribute to the net methane flux. The δ13C of CO2(aq) - the dominant form of DIC in acidic natural waters - reflects the processes occurring at that location modified by transport of gas from surrounding depths. Methane production and oxidation are imprinted in the δ13C signature of the aqueous CO2 with heaviest values at depth resulting from the fractionation associated with methane production. We measured δ13C profiles with depth along with CO2 and CH$_{4}$ concentrations from Sallie's Fen in Barrington, NH. Although the δ13C profiles varied considerably between locations and seasons, the logarithmic shape of the curves showed that methane production was restricted below a certain depth in the sediment - sometimes as shallow as 30 cm. Using a one-dimensional diffusion-reaction model, we are able to estimate rates of methane oxidation and successfully reproduce features present in the data's seasonal cycle. Features of the data not reproducible by the model indicate the importance of alternate gas transport routes such as ebullition and plant-mediated transport. The model also provides evidence for low-level oxygen availability during the winter-spring transition and narrow zones of very high productivity at depths of 60-70cm during the winter. We suggest that this method provides insight directly into the processes that determine methane fluxes from natural wetlands and has great potential for improving our understanding of the biogeochemistry of these systems.
B51C-0223
Spatial Patterns of Soil Organic Carbon in the United States
The Department of the Interior (DOI) has jurisdiction influencing approximately 22 percent of the land area of the United States. The poster presents estimates of the current stocks of soil organic carbon (SOC) on all lands and Federal lands. The DOI lands have about 22 percent of the nation's SOC, so the average carbon intensity (8.66 kg C m$^{-2}$) about matches the average for all lands (8.81 kg C m$^{-2}$). However the carbon on DOI lands is not evenly distributed. Of the 17.76 Petagrams (1 Pg = 10$^{15}$ grams) of SOC on DOI lands, 13.07 Pg (74 percent) are in Alaska, and 4.69 Pg (26 percent) are in the Conterminous U.S. The Alaska soils are wetter and colder than the national average, and the DOI lands in the conterminous U.S. are warmer and drier than the average. A set of SOC maps is shown, developed by intersecting the State Soil Geographic (STATSGO) database with data on federal lands from the National Atlas. With 22 percent of the nation's soil carbon, the DOI lands are important in a national accounting of greenhouse gas emission and sequestration. Future behavior of these lands is uncertain, but in scenarios of warming or drying, carbon released by respiration may exceed carbon captured by photosynthesis, resulting in a net release of carbon to the atmosphere. If warming stimulates a net release of greenhouse gases, this represents a positive feedback contributing to future global warming, a very unstable condition for the global climate system.
B51C-0224
Estimating Biomass in a Combined SiB and CASA Model Using Data Assimilation
Because direct measurement is not possible, one depends on land surface models to estimate regional carbon fluxes. Weather and climate dominate inter-annual variability in carbon flux, but whether the land surface model produces a long-term carbon source or sink depends sensitively on the assumed initial amount of biomass. Most models assume the initial biomass is in equilibrium with respect to climate, where biomass input from photosynthesis balances biomass losses due to microbial decay. However, biomass is typically not in climate equilibrium because of agriculture, timber harvest, biomass burning, and other external processes, resulting in unacceptable uncertainty in the time-mean simulated fluxes. To improve modeled carbon fluxes, we used data assimilation to estimate initial pool sizes of biomass in the SibCasa model from observed surface fluxes of latent heat, sensible heat, and carbon dioxide from the global flux tower network. SibCasa combines the Simple Biosphere (SiB) biophysical model with the Carnegie-Ames-Stanford Approach (CASA) biogeochemical model to produce a hybrid model capable of estimating net carbon fluxes at a 10-minute time resolution. We use the Maximum Likelihood Ensemble Filter (MLEF) ensemble-based data assimilation technique developed at Colorado State University to calculate optimal estimates of initial pool sizes (and associated uncertainties). The uncertainties are defined in terms of analysis and forecast error covariance matrices, calculated in an ensemble-spanned subspace. We present the SibCasa model and the MLEF technique, and compare estimated initial biomass pool sizes to available observations at various flux tower sites.
B51C-0225
Regional-Scale Surface CO2 Exchange Estimates Using a Boundary Layer Budget Method Over the Southern Great Plains
Concentration gradients of CO2 and H2O at the transition between the atmospheric boundary layer and free troposphere are linked to land surface exchanges at the regional scale. We used atmospheric concentration measurements and a boundary-layer budget model to estimate regional-scale CO2 fluxes, address uncertainties in the model, and suggest approaches to improve model accuracy. The budget model CO2 fluxes were compared against eddy covariance measurements from a 60 m tower at the Atmospheric Radiation Measurement (ARM) Climate Research Facility (ACRF) near Lamont, Oklahoma. The model assumes a slowly evolving boundary layer in which CO2 concentrations are quasi-steady on diurnal and longer temporal scales. We used the mass conservation equation to equate CO2 fluxes across the boundary layer top to surface fluxes. Fluxes across the boundary layer top were calculated as the product of the entrainment velocity and the CO2 concentration gradient between the free atmosphere and the well-mixed boundary layer. We neglected advective and local changes in the diurnally-averaged boundary layer height, and equated entrainment velocity with large-scale subsidence velocities obtained from (1) NCEP reanalysis, (2) the Rapid Update Cycle (RUC) atmospheric model, and (3) a water vapor tracer method. CO2 concentration measurements were made on the 60 m tower and on aircraft at 3700 m, and water vapor data was obtained from meteorological balloon soundings taken above the tower. Fluxes were calculated from diurnally-averaged concentration data from January through November 2004, and averaged over each month. Model flux estimates exhibited monthly trends in CO2 fluxes that matched trends in eddy covariance data. However, during peak surface CO2 exchange in April, model estimates differed from monthly-averaged eddy covariance measurements by 2.3 and 3.2 μmol m$^{-2}$ s$^{-1}$ using the water vapor derived vertical velocities and the RUC vertical velocities, respectively. During a less active vegetation period in August, model estimates differed by 1.7 to 3.0 μmol m$^{-2}$ s$^{-1}$. Preliminary results indicate that neglect of advection and cloud venting may be significant sources of error in the approach at the ARM site. We also attribute the differences between concentration-based model fluxes and eddy covariance fluxes to heterogeneity in the land surface vegetation cover.
B51C-0226
Regional-Scale Estimation of Carbon Fluxes in Complex Terrain Using a Budget Approach During the Airborne Carbon in the Mountains Experiment
Mountain forests represent a large portion of gross primary productivity within the United States and a significant potential net CO2 sink. Therefore, there is a need to develop methods to estimate regional fluxes of CO2 in mountainous terrain. We present results from a combined modeling and observational study of regional CO2 fluxes in mountainous terrain. The focus in this presentation is on the estimation of these fluxes using an approach similar to boundary-layer budgeting but with limited knowledge of the boundary-layer height required. We discuss the challenges that we encounter using the approach in mountainous terrain and possible solutions. We use data from the Airborne Carbon in the Mountains Experiment (ACME), conducted in May and July of 2004. The NCAR C-130 aircraft few over a large region (~350x350 km) of the Colorado Rocky Mountains on ten days, making continuous measurements of CO2, CO, O3, and water vapor concentrations among other measurements. The flights were conducted according to a combination of experimental designs, including morning to afternoon Lagrangian measurements, and morning sampling of nocturnally respired CO2. Applying the budget approach to the aircraft data, we estimated CO2 drawdowns of several ppm in the mountain boundary layer, representing significant CO2 uptake by the forests. These results agree surprisingly well with local flux measurements at a sub-alpine location. To interpret and understand the observations, we use a modeling framework consisting of the Regional Atmospheric Modeling System (RAMS), its adjoint, and a Lagrangian Particle Dispersion Model. The mesoscale model RAMS is run at a 1 km resolution and comparison with available observations shows that the model is able to capture meteorology well under the strongly forced conditions in complex terrain. We prescribe various scenarios of a CO2 flux at the surface and atmospheric conditions resulting in a variety of spatial and temporal behaviors of CO2 concentration in and above the mountain boundary layer. This enables the calculation of surface CO2 fluxes using the same approach as in the observations, while comparison with the prescribed fluxes allows a detailed investigation of the reliability and applicability of the budget approach in mountainous terrain.
B51C-0227
Analysis of Factors Controlling Interannual Variations in Atmospheric CO2 During 1997-2004
Measurements of surface atmospheric CO2 concentrations show that atmospheric CO2 growth rates vary significantly from year to year. Understanding the driving mechanisms of these interannual growth rate variations is important in terms of predicting future levels of atmospheric CO2. In this study, we investigate the relative contributions of interannual variations in terrestrial net primary production, heterotrophic respiration, and fire emissions to interannual variations in atmospheric CO2 during the 1997-2004 period. The geographical and temporal distribution of C fluxes associated with each of these processes is first derived using an updated version of the CASA biogeochemical cycle model that uses multiple satellite datasets as constraints. The CASA fluxes are then used to drive an atmospheric chemical transport model to calculate the resulting atmospheric CO2 concentration anomalies arising from each process. Finally, an inverse modeling methodology using atmospheric CO2 measurements from the NOAA/CMDL network, as well as atmospheric CO measurements from the same network as an additional constraint on fire C emissions, is applied to derive optimal estimates of the geographical and temporal distribution of C flux anomalies associated with each process.
B51C-0228
Leaf-to-Canopy Scaling of Carbon, Energy, and Moisture Fluxes in the Simple Biosphere Model (SiB)
The Simple Biosphere Model (SiB) has traditionally calculated photosynthesis for a single sun-leaf and used an empirical adjustment to extinction law in conjunction with satellite information to adjust carbon flux up to the canopy scale. This 'big leaf' approach has several limitations: modeled photosynthesis reaches light saturation too soon, and modeled Bowen Ratio has typically been too high. We have modified the SiB radiative submodel to simulate radiation balance and photosynthesis for sunlit and shaded leaves in SiB's fully prognostic canopy air space. The leaf-to-canopy scaling is based upon the partitioning of sunlit and shaded leaves, which is determined by canopy type and solar zenith angle. This presents a radical departure from the 'big leaf' radiative submodel used in the past. As a result of these fundamental changes to the model, we found that certain aspects of model parameterization require updating. For example, SiB photosynthesis is limited by the minimum of three assimilation rates: 1) efficiency of the photosynthetic enzyme system, 2) light limitation, and 3) export of photosynthesis products. These rates are not co-limited in individual leaves, but the transition has been shown to be smooth on the canopy scale. Therefore, we have used Maximum Likelihood Ensemble Filter (MLEF) techniques to determine optimum values for co-limitation parameters used in SiB. MLEF was developed at Colorado State University, and is designed to calculate optimal estimates of initial conditions, model errors, and empirical parameters, and also uncertainties of these estimates. The optimal estimates are defined as maximum likelihood values, obtained as results of the minimization of a cost function. The uncertainties are defined in terms of analysis and forecast error covariance matrices, calculated in ensemble-spanned subspace. We apply MLEF techniques to SiB parameters at several flux tower sites to determine if there are uniform modifications to established SiB parameters when the sunlit/shaded radiation submodel is applied.
B51C-0229
Simultaneous carbon flux, tower-based optical measurements and remote sensing in support of NACP scaling efforts
As part of a regional carbon cycling synthesis initiative (NACP) designed to improve methods of scaling plot and tower productivity measurements to regional landscapes we present a complete two year time series of carbon flux, canopy optical measurements and MODIS remote sensing for the USDA Bartlett Experimental Forest eddy-covariance flux tower. The Bartlett flux tower, a recent addition to the AmeriFlux network, is located in the predominantly northern hardwood US Forest Service research station in the White Mountains of New Hampshire . Through the unique tower instrumentation of upward and downward facing quantum (PAR) and pyronometer (full spectrum) optical sensors both above and below the forest canopy we are able to calculate true NDVI, fAPAR and other canopy reflectance indexes throughout the growing season and coincident with tower measurements of net and gross photosynthesis (carbon uptake). We compare these data with estimates from MODIS and explore the potential for refining light-Use efficiency calculations from remote sensing. We also present a unique histeresis relationship between NDVI and fAPAR in which a significant seasonal (green-Up vs. sensence) pattern emerges between these important remote sensing parameters.
B51C-0230
A Regional Atmospheric Continuous CO2 Network In The Rocky Mountains (Rocky RACCOON)
We have established a continuous CO2 observing network in the Rocky Mountains, building on technological and modeling advances made during the Carbon in the Mountains Experiment (CME), to improve our understanding of regional carbon fluxes and to fill key gaps in the North American Carbon Program (NACP). We will present a description of the Rocky RACCOON network and early results from the first three sites. There are strong scientific and societal motivations for determining CO2 exchanges on regional scales. NACP aims to address these concerns through a dramatic expansion in observations and modeling capabilities over North America. Mountain forests in particular represent a significant potential net CO2 sink in the U.S. and are highly sensitive to land-Use practices and climate change. However, plans for new continuous CO2 observing sites have omitted the mountain west. This resulted from expensive instrumentation in the face of limited resources, and a perception that current atmospheric transport models are not sophisticated enough to interpret CO2 measurements made in complex terrain. Through our efforts in CME, we have a new autonomous, inexpensive, and robust CO2 analysis system and are developing mountain CO2 modeling tools that will help us to overcome these obstacles. Preliminary observational and modeling results give us confidence that continuous CO2 observations from mountain top observatories will provide useful constraints on regional carbon cycling and will be valuable in the continental inverse modeling efforts planned for NACP. We began at three Colorado sites in August 2005 and hope to add three to six sites in other western states in subsequent years, utilizing existing observatories to the maximum extent possible. The first three sites are at Niwot Ridge, allowing us to have an ongoing intercomparison with flask measurements made by NOAA CMDL; at Storm Peak Laboratory near Steamboat Springs, allowing us to investigate comparisons between these two relatively nearby sites; and at Fraser Experimental Forest, allowing us to investigate nocturnal respiration rates across a large intermountain valley. Our data are available to the public on the internet in near real time to support quality control, local science, and larger scale synthesis efforts.
B51C-0231
Seasonal variation of carbon-14 in atmospheric carbon dioxide at Point Barrow, Alaska: Observations and modeling
Δ^{14}$C is an isotopic ratio uniquely suited for discriminating between fossil and biosphere carbon emissions, but few long-term measurement series exist. We have measured Δ^{14}$C in about 2 air samples monthly from the Point Barrow Observatory, Alaska (71°N, 157°W) since July 2003 with precision of around 2‰. In this period, Δ^{14}$C decreased by 7‰/year, to ~57‰ in mid-2005. We find a seasonal cycle in Δ^{14}$C with a broad minimum in January-June, a maximum in September and an amplitude of 6‰. Compared with these observations, simulations with the Model of Atmospheric Transport and Chemistry (MATCH) predict a seasonal cycle with broadly similar phase, with seasonality in fossil and biosphere CO2 emissions predicted to make about equal contributions. However, the simulated seasonal cycle has ~70% the observed amplitude, and it has Δ^{14}$C starting to increase earlier in the spring than observed. The shape of the observed cycle suggests a larger biosphere contribution, possibly due to either longer carbon mean residence time (so that respired carbon contains more bomb Δ^{14}$C) or higher net primary production in the region, than assumed in our simulation. We plan to use our measurements from Point Barrow and elsewhere in North America to test the regional carbon balance inferred from inverse modeling using only measurements of the CO2 concentration and δ^{13}$C.
B51C-0232
Regional Drawdown in Carbon Dioxide Observed from Total Column Measurements
We have developed an automated observatory for measuring ground-based column abundances of greenhouse gases. The gas abundances are determined from analysis of high-resolution near-infrared spectra of the direct sun. The laboratory is located in the heavily forested Chequamegon National Forest at the WLEF Tall Tower site, 14 km east of Park Falls, Wisconsin. Using these column CO2 measurements, we evaluate regional-scale drawdown in CO2 during sunny days in northern Wisconsin.
B51C-0233
Adaptive Rule-Based Piece-Wise Regression Models for Estimating Regional Net Ecosystem Exchange in Grassland and Shrubland Ecoregions Using Regional and Flux Tower Data
The scientific understanding of the global carbon cycle requires quantitative documentation, monitoring, and projection of carbon stocks and fluxes at various scales across the landscape. The challenge is to develop predictive models using carbon flux towers at site-specific locations, and to extrapolate these models to landscapes and regions. We use remote sensing and national climate and soil databases within data-driven models to estimate carbon fluxes. To accommodate the study of coupled human-environmental relationships and their influences on carbon dynamics, a coherent suite of models is being developed for agricultural, wooded and wetland ecosystems within predominantly grassland and shrubland ecoregions. In previous work, we have mapped carbon fluxes in terms of Net Ecosystem Exchange (NEE), Gross Primary Production (GPP), and Respiration (Re) in the Northern Great Plains, the Sagebrush Steppes and the Kazakh Steppes at 1-km resolution and 10-day time steps. We now extend this work beyond fairly uniform ecological conditions to accommodate more complex spatial mixtures of ecological types within ecoregions. The models need to adapt to both the complexity of the environmental variables and the land cover patterns. Our rule-based models adapt to local climatic, soil and phenology through the definition of piece-wise regression models. A suite of such models is needed to capture the phenologic and climatic variability across the wide range of shrubland and grassland ecoregions that exist. The result is a multi-year time series of 1-km maps of carbon flux that are suitable for trend and anomaly analysis. We seek sensitive models that permit the effective study of localized carbon dynamics while avoiding over-fitting the available carbon flux tower measurement data. Two critical components of the project are (1) sensitivity and cross-validation studies to evaluate the internal consistencies of the models and (2) intercomparison studies to help isolate methodological artifacts from variability resulting from climate and management of the land.
B51C-0234
Quantifying regional carbon budget: Lessons from studies in Tropics, China and the United States
Using a combination of carbon cycle models, remote sensing, and field measurements, we have estimated regional carbon budget in four geographical regions: Tropics including Amazon and Southeast Asia, China and the United States. We have examined how regional carbon storage has changed as a result of multiple stresses and interactions among those stresses including land-cover change, climate variability, atmospheric composition, precipitation chemistry, and natural disturbances using estimates of carbon fluxes and storage from factorial simulation experiments with ecosystem models. Model estimates along with spatial and temporal patterns of carbon fluxes and storage have been evaluated through comparisons with the results of field studies and forest and soil inventories within these regions. This study based on the previous work represents a major synthesis on our efforts in regional carbon cycle studies in the past decade. We also identify gaps and limitations in existing information that need to be investigated in the future to improve our understanding of processes controlling the regional carbon dynamics and our ability to estimate terrestrial carbon budget at regional level.
B51C-0235
Apparent Trends in Productivity of Monsoon Asia from 1982 to 2002
The rapid economic growth of Monsoon Asia raises concerns about the future of carbon stored in the terrestrial ecosystems of the region, especially in connection with climate change. The regional carbon budget for 1980s suggests that Monsoon Asia as a whole acted as source [Tian et al., 2003], although some parts of the region acted as sink. Here we provide some evidence from satellite data that productivity of the region changed in the manner that suggests similar conclusion. Comparing the period 1982-1992 and the period 1992-2002, we found that the productivity of the territory generally decreased in the forest zone and increased in the non-forest zone of the region. The productivity of a territory strongly depends on the area covered with photosynthetically active vegetation (PAV) and, therefore, we introduce a grid variable, FPAV, which stands for the fraction of a grid cell covered with PAV. (Grid, here, means geographic grid of half-degree resolution.) Deciduous plants are leafless during the dormant season, and so FPAV may vary on seasonal basis. The amplitude enlarges with the ratio between deciduous and evergreen species. The minimal value of FPAV gives the fraction of a grid cell covered with evergreen vegetation. The maximal value gives the fraction of the cell that covered either with deciduous or evergreen vegetation and, thus, tells us which fraction of the cell is vegetated. The changes in FPAV were tracked by using monthly values of AVHRR-NDVI for the period from July 1981 to December 2002 that were recently compiled into a public data set, so called GIMMS-NDVI [see Slayback et al., 2003 and references therein]. We calculated average monthly values of GIMMS-NDVI for two 11-year periods: from 1982 to 1992 and from 1992 to 2002, and used them to evaluate the trends in productivity of the region, characterized by the product of FPAVmax and Pn, where Pn is net primary production of potential natural vegetation, calculated by using TsuBiMo-model, FPAVmax is the maximal value of FPAV. This net increase in vegetated area of non-forest zone implies additional net primary production of 19 MtC/yr. However, this cannot compensate the losses associated with the net decrease in vegetated area of the forest zone. Therefore the total productivity of the region is estimated to drop by 64 MtC/yr. (It is worth to mention here, that comparing the 11-year periods we filter out the effects related to short-term monsoon variability, but this cannot exclude the effects related to "monsoon epochs".)