Optimal recovery of regional CO2 surface fluxes by data assimilation of anthropogenic and biogenic tracers

Measurements of atmospheric carbon dioxide (CO2) have led to an understanding of the past and present CO2 trends at global scales. However, many of the processes that underlie the CO2 fluxes are highly uncertain, especially at smaller spatial scales in the terrestrial biosphere. Our abilities to forecast climate change and manage the carbon cycle are reliant on an understanding of these underlying processes. In this dissertation, new steps were taken to understand the biogenic and anthropogenic processes based on analysis with an atmospheric transport model and simultaneous measurements of CO2 and other trace gases.

The biogenic processes were addressed by developing an approach for quantifying photosynthesis and respiration surface fluxes using observations of CO2 and carbonyl sulfide (COS). There is currently no reliable method for separating the influence of these gross biosphere fluxes on atmospheric CO2 concentrations. First, the plant sink for COS was quantified as a function of the CO2 photosynthesis uptake using the STEM transport model and measurements of COS and CO2 from the INTEX-NA campaign. Next, the STEM inversion model was modified for the simultaneous optimization of fluxes using COS and CO2 measurements and using only CO2 measurements. The CO2-only inversion was found to be process blind, while the simultaneous COS/CO2 inversion was found to provide a unique estimate of the respiration and photosynthesis component fluxes. Further validation should be pursued with independent observations. The approach presented here is the first application of COS measurements for inferring information about the carbon cycle.

Anthropogenic emissions were addressed by improving the estimate of the fossil fuel component of observed CO2 by using observed carbon monoxide (CO). Recent applications of the CO approach were based on simple approximations of non-fossil fuel influences on the measured CO such as sources from oxidation of volatile organic carbon species, sinks from oxidation of CO, and sources from forest fires. A revised CO method was developed using STEM simulations of atmospheric reactions and tracers of different combustion sources. Applications of the revised method to the NASA INTEX-NA measurements showed large differences with conventional methods. Application to the INTEX-B measurements resulted in partitioning of continental and offshore oil rig sources around Mexico.