Probabilistic carbon flux upscaling in a complex northern forest ecoregion
K
J
Davis, Penn State, kjd10@psu.edu
(Presenting)
P
V
Bolstad, U. Minnesota, pbolstad@umn.edu
R
Anderson, U. Montana, ryan.anderson@ntsg.umt.edu
K
Cherrey, Penn State, kcherrey@psu.edu
B
D
Cook, U. Minnesota, brucecook@umn.edu
A
R
Desai, U. Wisconsin, desai@aos.wisc.edu
F
A
Heinsch, U. Montana, faithann@ntsg.umt.edu
R
Kolka, USDA Forest Service, rkolka@fs.fed.us
S
W
Running, U. Montana, swr@ntsg.umt.edu
N
Z
Saliendra, USDA Forest Service, nsaliendra@fs.fed.us
P
Weishampel, U. Minnesota, weish004@umn.edu
Flux upscaling describes efforts to turn limited-area flux measurements into spatially and temporally comprehensive maps of ecosystem-atmosphere carbon exchange. This approach is frequently utilized, but rigorous tests are difficult to construct. The Chequamegon Ecosystem-Atmosphere Study (ChEAS) is an effort to test flux tower upscaling in a forested region similar in character to broad expanses of North American forests. We are testing the value of a a regional, high-density data set including a regional flux tower network, ground-based biomass, chamber flux and sap flux measurements, lidar-derived forest structural data, and high-resolution airborne and space-based passive remote sensing in order to evaluate the flux-tower upscaling hypothesis in detail, and illustrate the value of additional data and model complexity in performing precise and accurate upscaling.
We have found that net ecosystem-atmosphere exchange of CO2 and water varies dramatically across the landscape due to changes in vegetation type, soil conditions and canopy stocking, variations that are not represented well by the MODIS-based methods that are currently the default for flux upscaling. This “failure” in the MODIS land products is caused by a dominant scale for landscape variability that is only a few hundred meters, thus unresolved by MODIS. Flux upscaling in this region is improved by higher resolution vegetation classification based on products such as Landsat, ASTER or AVIRIS, by continuous integration of flux tower and MODIS data. Further, we have found that net flux depends strongly upon forest structure, particularly the stocking or density of trees in the forest. Woody plant stocking that is specified uniformly for each vegetation class ignores a large source of variability in fluxes in the region. Measurements of above-ground biomass using both ground-based and airborne LiDAR observations capture these differences in woody biomass and should enable significant improvements in regional flux upscaling.
This study is working to develop a framework for improved regional upscaling of carbon fluxes that can be applied beyond the ChEAS region, including quantitative estimates for the degree of improvement in the regional upscaling as a function of the data and model complexity implemented in the region.
NASA Carbon Cycle & Ecosystems Active Awards Represented by this Poster: