Inverting 13CO2 for Terrestrial Carbon Fractionation in North America
C. Alden1, J.B. Miller2 and J.W.C. White1
1Institute of Arctic and Alpine Research (INSTAAR), University of Colorado, Boulder, CO 80309; 719-930-5281, E-mail: caroline.alden@colorado.edu
2Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder, CO 80309
A variety of methods exist for constraining chemical fluxes at the earth’s surface. Inverse modeling is often used to estimate regional-scale fluxes of the greenhouse gas carbon dioxide (CO2), by using atmospheric observations and particle transport models to optimize prior flux estimates. Fluxes of the rare stable carbon isotope in CO2, 13C, have not traditionally been calculated with inverse models. Historically, 13C has been used to partition CO2 fluxes into land and ocean components, because major fluxes of this gas (fossil fuel, ocean, and land) impose distinctive and predictable fractionation patterns upon the stable isotope ratio. As measurement density improves, 13C has the potential for an important new role in carbon cycle research: to illuminate mechanisms driving terrestrial biosphere exchange variability. Calculated time series of the global land flux, disequilibrium flux, and terrestrial discrimination from 1991 through 2008 that are consistent with bottom up net ocean fluxes suggest high interannual variability in terrestrial disequilibrium flux. The primary contributors to this variability likely include discrimination due to plant stomatal opening and the relative contributions of C3 and C4 vegetation to net ecosystem exchange. Discerning 13CO2 flux magnitude and variability on smaller space and time scales could therefore inform our understanding of the dynamics that exist between carbon cycling and environmental/climatic variability, as well as the impacts of biofuels on the 13C budget. We determine “optimal” estimates of CO2 and 13CO2 surface fluxes across the region encompassing 10-80°N and 145-25°W. We use the Lagrangian particle dispersion model FLEXPART to generate partial derivative matrices or “influence functions”, which relate fluxes in each 1x1 degree grid cell in the region of interest to concentrations at several North American NOAA ESRL Global Monitoring Division tower sites. By comparing modeled with observed 13CO2 concentrations at tower sites, we can then optimize our prior flux estimates. We will present comparisons between our simulated values and those observed at the monitoring sites, and we will present the preliminary results of our inversion.
Figure 1. Comparison of Modeled vs Observed 13C at a NOAA ESRL tall tower in Park Falls, WI.