Using Atmospheric δ13C of CO2 observations to link the water and carbon cycles with climate

The ratio of stable carbon isotopes, 13C:12C in atmospheric CO2 (expressed as δ13C) offers unique insights into atmosphere-land CO2 fluxes and the modulating effects of stomatal conductance on this exchange. Photosynthesis discriminates against 13CO2 during uptake. The magnitude of this fractionation is strongly dependent upon ambient CO2 concentrations and water availability, as well as on the mix of C3 and C4 vegetation types. C3 and C4 plants have very different discrimination because of carboxylation pathways, and C3 stomatal conductance varies with water availability because stomata close to reduce transpiration when plants are water stressed. Further, plant stomata respond to ambient CO2 concentrations in order to optimize leaf internal [CO2] while reducing transpirative water loss. Atmospheric δ13C therefore carries information about local and upwind drought conditions and the consequent likelihood of ground-to-atmosphere water transfer via transpiration, and the balance of latent and sensible heat fluxes, as well as about local and upwind distributions of C3 and C4 vegetation and variability therein. δ13C offers a unique lens through which to identify key thresholds and relationships between climate anomalies/change and the modulating climate impacts of plant biosphere response. By unraveling this relationship at local to continental scales, we stand to gain crucial understanding of the drivers of land CO2 uptake variability as well as knowledge of how to predict future climate impacts on the carbon cycle and vice versa. We use a two-step Bayesian inversion model to optimize 1x1 degree and 3-hourly (interpreted at regional and weekly to monthly scales) fields of δ13C of assimilated biomass over North America for the year 2010, using influence functions generated with FLEXPART, driven by National Centers for Environmental Prediction Global Forecast System meteorology. Prior fluxes and fossil fuel, ocean and fire fluxes are from CarbonTracker 2011, and background CO2 and δ13C values are from NOAA/ESRL marine boundary layer and aircraft data. Quasi-daily atmospheric observations are from NOAA/ESRL Global Monitoring Division tall towers and weekly observations are from the Environment Canada observation network. Synthetic data experiments show that, given a prior δ13C of assimilated biomass with zero spatial or temporal variability, the inversion is able to recover broad patterns of C3 and C4 vegetation distributions and seasonality. We investigate correlations between week- to seasonal-scale anomalies in discrimination and climate drivers: temperature, precipitation, potential evapotranspiration, vapor pressure and vapor pressure deficit, Palmer drought severity index, standardized precipitation-evaporation index, ground water from the Grace satellite observations, and USDA crop coverage data.