North American CO_2 exchange: inter-comparison of modeled estimates with results from a fine-scale atmospheric inversion

摘要

Atmospheric inversion models have the potentialto quantify CO_2 fluxes at regional, sub-continental scalesby taking advantage of near-surface CO_2 mixing ratio observationscollected in areas with high flux variability. Thisstudy presents results from a series of regional geostatisticalinverse models (GIM) over North America for 2004, anduses them as the basis for an inter-comparison to other inversionstudies and estimates from biospheric models collectedthrough the North American Carbon Program Regional andContinental Interim Synthesis. Because the GIM approachdoes not require explicit prior flux estimates and resolvesfluxes at fine spatiotemporal scales (i.e. 1° ×1°, 3-hourlyin this study), it avoids temporal and spatial aggregation errorsand allows for the recovery of realistic spatial patternsfrom the atmospheric data relative to previous inversion studies.Results from a GIM inversion using only available atmosphericobservations and a fine-scale fossil fuel inventorywere used to confirm the quality of the inventory andinversion setup. An inversion additionally including auxiliaryvariables from the North American Regional Reanalysisfound inferred relationships with flux consistent with physiologicalunderstanding of the biospheric carbon cycle. Comparisonof GIM results with bottom-up biospheric modelsshowed stronger agreement during the growing relative tothe dormant season, in part because most of the biosphericmodels do not fully represent agricultural land-managementpractices and the fate of both residual biomass and harvestedproducts. Comparison to earlier inversion studies pointed toaggregation errors as a likely source of bias in previous subcontinentalscale flux estimates, particularly for inversionsthat adjust fluxes at the coarsest scales and use atmosphericobservations averaged over long periods. Finally, whereasthe continental CO_2 boundary conditions used in the GIMinversions have a minor impact on spatial patterns, they havea substantial impact on the continental carbon budget, with adifference of 0.8 PgC yr~(-1) in the total continental flux resultingfrom the use of two plausible sets of boundary CO_2 mixingratios. Overall, this inter-comparison study helps to assessthe state of the science in estimating regional-scale CO_2fluxes, while pointing towards the path forward for improvementsin future top-down and bottom-up modeling efforts.