Constraints on the Global Carbon Budget from Variations in Total Column Carbon Dioxide

Citation

Keppel-Aleks, Gretchen (2011) Constraints on the Global Carbon Budget from Variations in Total Column Carbon Dioxide. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/5BZZ-BW98. https://resolver.caltech.edu/CaltechTHESIS:05312011-112642236

Abstract

Diagnosing the patterns and trends in the flux of carbon dioxide, CO₂, between the land or ocean and the atmosphere is necessary to predict the response of the carbon cycle to climate change. Atmospheric observations of the vertically averaged mixing ratio of CO₂, (CO₂), provide a new tool that complements existing observations of boundary layer CO₂ in constraining surface fluxes of CO2. My dissertation explores how variations in (CO₂) arise and how these variations can be used to estimate surface fluxes.

This thesis takes advantage of (CO₂) measurements from the Total Carbon Column Observing Network (TCCON). This global network uses ground-based Fourier transform spectrometers to obtain direct solar spectra in the near infrared, from which (CO₂) is retrieved. Because variations in atmospheric CO₂ are relatively small, it is essential that the data achieve high precision and accuracy to be useful for carbon cycle science. Using a retrieval algorithm I developed to remove transient interference from clouds and aerosols, precise measurements are achievable under a range of meteorological conditions, allowing the inclusion of data from partially cloudy days.

At midlatitude TCCON sites, (CO₂) varies substantially on diurnal, synoptic, and seasonal timescales. A comparison of diurnal variations in (CO₂) with flux tower observations of net ecosystem exchange in northern Wisconsin suggests that local ecosystem fluxes account for only 10%–15% of variation in (CO₂) on hourly timescales. I use an atmospheric transport model with imposed surface fluxes of CO₂ to examine further the sensitivity of (CO₂) to surface fluxes and find that, as in the observations, simulated (CO₂) is relatively insensitive to local phenomena. Large variations in local fluxes and local physics (e.g., convection) produce only small changes in atmospheric (CO₂) patterns. Patterns in (CO₂) are most sensitive to the large-scale north–south flux gradient, as zonal variations in fluxes are smoothed by transport and therefore have little impact on the total column.

Rapid temporal variations in midlatitude (CO₂) arise due to transport across north-south gradients, and can be used to infer information about large-scale spatial patterns in CO₂. Here, I use the correlation between synoptic-scale variations in (CO₂) and dynamical tracers, such as potential temperature, to infer spatial gradients in large-scale (CO₂) from sparse ground-based data. These estimated gradients, as well as the amplitude of the seasonal cycle in (CO₂), can be used as diagnostics to evaluate flux models. In simulations with one such model, the Carnegie Ames Stanford Approach (CASA) ecosystem fluxes, gradients inferred from TCCON data are 75% larger than simulated CO₂ gradients during summer, while the seasonal cycle amplitudes in (CO₂) at midlatitude TCCON sites are between 20% and 40% larger than in simulations. Given that (CO₂) is insensitive to local fluxes, the mismatch between observations and simulations points to an underestimation of northern hemisphere ecosystem fluxes. Simulated (CO₂) diagnostics are consistent with TCCON data if boreal net ecosystem exchange is increased by 40%, a finding that suggests boreal ecosystem parameterizations must be reevaluated.

This work demonstrates that variability in (CO₂) is driven by large-scale phenomena rather than local fluxes, and has important implications for interpreting (CO₂) measurements from satellites such as GOSAT and OCO-2. These results suggest that total column measurements will provide a strong constraint on large-scale flux estimates. At the same time, extracting information about regional fluxes from (CO₂) observations alone will be challenging. Coupling column and surface CO₂ observations will yield improved flux estimates as surface CO₂ observations will constrain regional flux patterns, superimposed on top of the large-scale flux distribution revealed by total column observations.

Thesis Files