Thursday, November 17, 2011, 4:30AM - 5:30AM
The stable isotope of Carbon, 13C, has been used globally to partition ocean from terrestrial CO2 fluxes, and regionally as an indicator of C3 plant discrimination and the relative abundances of C3 and C4 vegetation. In North America, the spatial and temporal density of atmospheric observations enables, for the first time, inversions of land surface fluxes of CO2 and the 13C signature of those fluxes (multiplied together: isofluxes). The traditional Bayesian inversion is performed in two steps: first, we optimize North American CO2 fluxes. Optimized CO2 fluxes are then combined with plant discrimination to create isofluxes, which are optimized simultaneously with terrestrial disequilibrium fluxes. We are left with estimates of the isotopic signature of photosynthesis, terrestrial disequilibrium flux, and related error. Influence functions (i.e. influence of surface fluxes on measurements made “downwind”) are generated with FLEXPART, driven by National Centers for Environmental Prediction Global Forecast System meteorology. Prior isofluxes (CarbonTracker 2010 posterior net ecosystem exchange multiplied by biosphere 13C from SiB, and disequilibrium flux from CASA impulse-response functions and historical atmospheric CO2 observations) are of 3-hour, 1x1 degree resolution. Unoptimized fire and fossil fuel emissions are also from the CarbonTracker system, and background CO2 and 13C values are taken from NOAA/ESRL marine boundary layer and aircraft data. We invert for isofluxes using quasi-daily observations from tall towers in Park Falls, Wisconsin (WLEF). At this site the air is thought to be influenced predominately by corn production. Our aim is to disentangle the competing influences of agriculture, drought, and other climatic stressors on spatial and temporal variability in atmospheric 13C.