1. Reconciling estimates of the contemporary North American carbon balance among an inventory-based approach, terrestrial biosphere models, and atmospheric inversions D.J. Hayes 1 *, D.P. Turner 2, G. Stinson 3, A.D. McGuire 4, Y. Wei 1, T.O. West 5, L.S. Heath 6, B. deJong 7, B.G. McConkey 8, R.A. Birdsey 9, W.A. Kurz 3, A.R. Jacobson 10, D.N. Huntzinger 11, Y. Pan 9, W.M. Post 1, and R.B. Cook 1 1 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA; 2 Department of Forest Ecosystems and Society, Oregon State University, Corvallis, OR, USA; 3 Pacific Forestry Centre, Canadian Forest Service, Victoria, BC, Canada; 4 U.S. Geological Survey, Alaska Cooperative Fish and Wildlife Research Unit, University of Alaska Fairbanks, Fairbanks, AK, USA; 5 Joint Global Change Research Institute, Pacific Northwest National Laboratory, College Park, MD; 6 USDA Forest Service, Durham, NH, USA; currently on secondment to the Global Environment Facility, Washington, DC, USA; 7 El Colegio de la Frontera Sur (ECOSUR), Mexico; 8 Agriculture and Agri-Food Canada, Ottawa, ON, Canada; 9 USDA Forest Service, Newtown Square, PA, USA; 10 NOAA Earth System Research Lab, Boulder, CO, USA; 11 Department of Civil and Environmental Engineering, The University of Michigan, Ann Arbor, MI, USA Reconciling estimates of the contemporary North American carbon balance among an inventory-based approach, terrestrial biosphere models, and atmospheric inversions D.J. Hayes 1 *, D.P. Turner 2, G. Stinson 3, A.D. McGuire 4, Y. Wei 1, T.O. West 5, L.S. Heath 6, B. deJong 7, B.G. McConkey 8, R.A. Birdsey 9, W.A. Kurz 3, A.R. Jacobson 10, D.N. Huntzinger 11, Y. Pan 9, W.M. Post 1, and R.B. Cook 1 1 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA; 2 Department of Forest Ecosystems and Society, Oregon State University, Corvallis, OR, USA; 3 Pacific Forestry Centre, Canadian Forest Service, Victoria, BC, Canada; 4 U.S. Geological Survey, Alaska Cooperative Fish and Wildlife Research Unit, University of Alaska Fairbanks, Fairbanks, AK, USA; 5 Joint Global Change Research Institute, Pacific Northwest National Laboratory, College Park, MD; 6 USDA Forest Service, Durham, NH, USA; currently on secondment to the Global Environment Facility, Washington, DC, USA; 7 El Colegio de la Frontera Sur (ECOSUR), Mexico; 8 Agriculture and Agri-Food Canada, Ottawa, ON, Canada; 9 USDA Forest Service, Newtown Square, PA, USA; 10 NOAA Earth System Research Lab, Boulder, CO, USA; 11 Department of Civil and Environmental Engineering, The University of Michigan, Ann Arbor, MI, USA I.Overview The North American carbon sink is generally considered to account for a large, but highly uncertain, portion of the northern extratropical land based sink of the late 20 th Century, with estimates ranging from 15% to 100% (see the SOCCR report; King et al., 2007). This uncertainty is owing to a number of sources, including limitations related to: a) methodologies, b) data on key driving forces, c) knowledge of long-term ecosystem responses to these driving forces. Here, we compare “top-down” (atmospheric inverse models) and “bottom-up” (terrestrial biosphere, or “forward”, models) model estimates of sources and sinks of carbon across the North American continent with those based on national forest and agricultural inventories for Canada, the U.S. and Mexico. I.Overview The North American carbon sink is generally considered to account for a large, but highly uncertain, portion of the northern extratropical land based sink of the late 20 th Century, with estimates ranging from 15% to 100% (see the SOCCR report; King et al., 2007). This uncertainty is owing to a number of sources, including limitations related to: a) methodologies, b) data on key driving forces, c) knowledge of long-term ecosystem responses to these driving forces. Here, we compare “top-down” (atmospheric inverse models) and “bottom-up” (terrestrial biosphere, or “forward”, models) model estimates of sources and sinks of carbon across the North American continent with those based on national forest and agricultural inventories for Canada, the U.S. and Mexico. ~ 25 – 30% of FF emissions II.Inventory-based Data Analysis Here, we develop a novel approach using existing inventory-based stock change data sets for estimating net ecosystem exchange (NEE; the vertical exchange of CO 2 between the land and the atmosphere) over North America at the level of the “reporting zone” for the early 21 st Century (ca. 2000 – 2006). The approach uses a conceptual model of the C budget based on the “atmospheric flow” approach (Figure 1). Component fluxes include Forest Lands sector (vegetation and soil C stock change and wood harvest removals and production emissions) and Crop Lands sector (crop harvest and soil C stock change) indicators. Carbon removal due to growth and emissions due to primary consumption are accounted for in the Forest Land or Crop Land sector of the ‘producing’ zone. Our accounting reflects the assumption that some amount of the carbon in harvested products is not necessarily emitted directly from within the sector (Forest Lands or Crop Lands) or even reporting zone where it is produced, but rather in the “Other Lands” sector that the consumers (i.e., humans and livestock) occupy. We use data on stock change to account for both product storage and “inherited” emissions from the decay of products produced from wood harvested prior to our study period; and product pools are adjusted by international imports and exports. Thus, the approach notably retains information on the spatial and temporal distribution of fluxes at the sub-regional scale, which can then be compared with modeled estimates using common scales and flux indicators. II.Inventory-based Data Analysis Here, we develop a novel approach using existing inventory-based stock change data sets for estimating net ecosystem exchange (NEE; the vertical exchange of CO 2 between the land and the atmosphere) over North America at the level of the “reporting zone” for the early 21 st Century (ca. 2000 – 2006). The approach uses a conceptual model of the C budget based on the “atmospheric flow” approach (Figure 1). Component fluxes include Forest Lands sector (vegetation and soil C stock change and wood harvest removals and production emissions) and Crop Lands sector (crop harvest and soil C stock change) indicators. Carbon removal due to growth and emissions due to primary consumption are accounted for in the Forest Land or Crop Land sector of the ‘producing’ zone. Our accounting reflects the assumption that some amount of the carbon in harvested products is not necessarily emitted directly from within the sector (Forest Lands or Crop Lands) or even reporting zone where it is produced, but rather in the “Other Lands” sector that the consumers (i.e., humans and livestock) occupy. We use data on stock change to account for both product storage and “inherited” emissions from the decay of products produced from wood harvested prior to our study period; and product pools are adjusted by international imports and exports. Thus, the approach notably retains information on the spatial and temporal distribution of fluxes at the sub-regional scale, which can then be compared with modeled estimates using common scales and flux indicators. Figure 1. Conceptual diagram of the continental-scale carbon budget based on data available from the inventory-based approaches that estimate carbon stock changes, fluxes and transfers among forest, crop and other lands. Data on carbon exchange for each sector are summarized according to “reporting zones”, which refer to political state units in the case of the U.S. and Mexico, and UNFCCC reporting zones for Canada. III.Model-data Processing “Off-theshelf” flux estimates from existing forward (n=16) and inverse (n=7) model runs (Table 1) were collected for the Interim Synthesis activities under the North American Carbon Program (NACP). These estimates were processed to allow comparison at the spatial and temporal scales of the inventories (Figure 2). Processing of contributed output files used monthly indicator variables, which were standardized to a common 1x1 degree spatial resolution of annual values. For comparison with inventory estimates, the model data were aggregated by reporting zone: within each reporting zone, the areal extent of forest and croplands were determined using land cover types from the GLC-2000 data set reclassified into 3 categories: forest, crop, and other. Annual indicator variables from each model for each land cover type were calculated by area-weighting according to the area of each land cover type in that zone. III.Model-data Processing “Off-theshelf” flux estimates from existing forward (n=16) and inverse (n=7) model runs (Table 1) were collected for the Interim Synthesis activities under the North American Carbon Program (NACP). These estimates were processed to allow comparison at the spatial and temporal scales of the inventories (Figure 2). Processing of contributed output files used monthly indicator variables, which were standardized to a common 1x1 degree spatial resolution of annual values. For comparison with inventory estimates, the model data were aggregated by reporting zone: within each reporting zone, the areal extent of forest and croplands were determined using land cover types from the GLC-2000 data set reclassified into 3 categories: forest, crop, and other. Annual indicator variables from each model for each land cover type were calculated by area-weighting according to the area of each land cover type in that zone. Table 1. Participating Models Figure 2. Forest Lands and Crop Lands sector reclassification in each inventory reporting zone for model- data processing of 1 o x1 o gridded carbon flux estimates. IV.Results The inventory-based total NEE estimate of a -327 ± 252 TgC yr -1 sink for NA was driven primarily by CO 2 uptake in the Forest Lands sector (-248 TgC yr -1 ), largely in the Northwest and Southwest regions of the U.S., and in the Crop Lands sector (-297 TgC yr -1 ), predominantly in the Midwest U.S. states. These sinks are counteracted by the carbon source estimated for the Other Lands sector (+218 TgC yr -1 ), where much of the forest and crop products are assumed to be returned to the atmosphere (through livestock and human consumption). The ecosystems of Mexico are estimated to be a small net source (+18 TgC yr -1 ) due to land use change. We compare these inventory-based estimates with results from a suite of terrestrial biosphere and atmospheric inversion models, where the mean continental-scale NEE estimate for each ensemble is -511 TgC yr -1 and -931 TgC yr -1, respectively (Table 2). Depending on the approach, the estimates of the NA sink presented here represent between 18% and 52% of continental fossil fuel emissions over this same time period. In the modeling approaches, the Other Lands sector was generally estimated to be a carbon sink, driven in part by assumed productivity enhancement from rising atmospheric CO 2 concentrations and/or lack of consideration of the carbon source from harvested forest and crop product consumption or decomposition. Additional fluxes (woody encroachment, wetlands, riverine DOC export) not measured by the inventories (Table 3), though highly uncertain, could add an additional -239 TgC yr -1 to the inventory-based NA sink estimate, thus suggesting some convergence with the modeling approaches The data and the methodologies used in this study suggest considerable spatial variability in NEE estimates across sectors and reporting zones (Figure 3). The spatial patterns are driven both by the estimated direct, vertical surface fluxes as well as the lateral transfer of carbon between sectors in the form of harvested products (Figure 4). IV.Results The inventory-based total NEE estimate of a -327 ± 252 TgC yr -1 sink for NA was driven primarily by CO 2 uptake in the Forest Lands sector (-248 TgC yr -1 ), largely in the Northwest and Southwest regions of the U.S., and in the Crop Lands sector (-297 TgC yr -1 ), predominantly in the Midwest U.S. states. These sinks are counteracted by the carbon source estimated for the Other Lands sector (+218 TgC yr -1 ), where much of the forest and crop products are assumed to be returned to the atmosphere (through livestock and human consumption). The ecosystems of Mexico are estimated to be a small net source (+18 TgC yr -1 ) due to land use change. We compare these inventory-based estimates with results from a suite of terrestrial biosphere and atmospheric inversion models, where the mean continental-scale NEE estimate for each ensemble is -511 TgC yr -1 and -931 TgC yr -1, respectively (Table 2). Depending on the approach, the estimates of the NA sink presented here represent between 18% and 52% of continental fossil fuel emissions over this same time period. In the modeling approaches, the Other Lands sector was generally estimated to be a carbon sink, driven in part by assumed productivity enhancement from rising atmospheric CO 2 concentrations and/or lack of consideration of the carbon source from harvested forest and crop product consumption or decomposition. Additional fluxes (woody encroachment, wetlands, riverine DOC export) not measured by the inventories (Table 3), though highly uncertain, could add an additional -239 TgC yr -1 to the inventory-based NA sink estimate, thus suggesting some convergence with the modeling approaches The data and the methodologies used in this study suggest considerable spatial variability in NEE estimates across sectors and reporting zones (Figure 3). The spatial patterns are driven both by the estimated direct, vertical surface fluxes as well as the lateral transfer of carbon between sectors in the form of harvested products (Figure 4). Table 3. Inventory Uncertainty Table 2. Model vs. Inventory Comparison Figure 3. Mean area-weighted average annual NEE (gC m -2 yr -1 ), 2000 to 2006 for the Forest Lands, Crop Lands and Other Lands sectors, along with all land (total), in each reporting zone, from inventory-based estimates against mean results from the sets of terrestrial biosphere (forward) models and inverse models. Figure 4. The balance between product harvest and product emissions (TgC yr -1 ), 2000 to 2006, for each reporting zone from the inventory-based estimates, shown for a) forest harvest products (croplands masked), b) crop harvest products (forest lands masked) and c) all products. A negative value represents a net producing (exporting) zone and a positive value represents a net consuming (importing / emitting) zone. V.Discussion Ultimately, confidence in our ability to understand and predict the role of the North American carbon cycle in the global climate system will increase as the estimates from these different approaches begin to more closely converge and are combined in more fully integrated modeling systems. V.Discussion Ultimately, confidence in our ability to understand and predict the role of the North American carbon cycle in the global climate system will increase as the estimates from these different approaches begin to more closely converge and are combined in more fully integrated modeling systems. *Contact: hayesdj@ornl.govhayesdj@ornl.gov *Contact: hayesdj@ornl.govhayesdj@ornl.gov B43F-0369 2011 NASA CC&E Joint Science Workshop – October 3-7, 2011