Scaling Regional Land-Atmosphere Fluxes of Carbon Dioxide with Mesoscale Observation Networks: Impact of Land Cover, Management and Disturbance
Open Access
- Author:
- Desai, Ankur Rashmikant
- Graduate Program:
- Meteorology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- March 13, 2006
- Committee Members:
- Kenneth James Davis, Committee Chair/Co-Chair
Toby Nahum Carlson, Committee Member
Raymond Gabriel Najjar Jr., Committee Member
John Corry Wyngaard, Committee Member
Brenton Yarnal, Committee Member - Keywords:
- eddy covariance
land-atmosphere trace gas exchange
regional carbon dioxide flux
bottom-up scaling
dynamic vegetation ecosystem modeling - Abstract:
- Improving techniques to scale observations of land-atmosphere fluxes from site to region to globe is required for better understanding of processes leading to spatiotemporal variability of carbon sources and sinks, rates of change in global atmospheric CO2 burden and associated climate forcing. The Chequamegon Ecosystem-Atmosphere Study regional field study focuses on CO2 and water land-atmosphere fluxes in the forest and wetland covered region of northern Wisconsin, USA. The region is unique for its land-atmosphere observational network density and challenging due to its patchy landscape. The current landscape and exchange of CO2 is strongly mediated by intrinsic abiotic spatial variability and the impact of natural and anthropogenic disturbance. This study investigates how disturbance and land cover variability affects regional carbon exchange and complicates standard observation-based scaling, remote sensing flux estimates and global biogeochemical models. The eddy-covariance method uses high-frequency measurements of atmospheric turbulence and scalar concentration to quantify fluxes on scales of 0.1-100 ha. Evidence is presented showing a large difference in net ecosystem exchange between a mature forest harvested 80 years ago and an old-growth forest with the same pre-European settlement composition. This work is extended to further quantify stand age effects on growing season carbon exchange across 14 sites. A parameter assimilation aggregation scheme is developed that relies on flux towers, tower-observed meteorology, remote sensing and forest inventory statistics to estimate regional carbon exchange. Comparison against fluxes measured from a regionally integrating tall tower (396 m) are encouraging, especially after a tall tower directional sampling bias is diagnosed and removed using footprint decomposition and re-aggregation techniques. A novel ecosystem model, using a statistical representation of sub-grid variability in age and disturbance, is parameterized with ecological measurements and run with assimilated climate data and CO2 mixing ratios to further investigate processes controlling spatiotemporal variability in carbon sources/sinks due to disturbance, cover type, climate and management. The model also provides an additional independent regional flux estimate. Results from all these methods are used to consider the role of regional field observations and bottom-up scaling in informing top-down tracer transport boundary layer budget and inversion methods as well as global climate-carbon models.