Revising estimates of global GPP using new information from eddy covariance and satellite datasets
Yi
Wang, UC, Irvine, ywang17@uci.edu
(Presenting)
Yufang
Jin, UC, Irvine, yufang@uci.edu
Michael
L
Goulden, UC, Irvine, mgoulden@uci.edu
James
T
Randerson, UC, Irvine, jranders@uci.edu
Obtaining precise estimates of contemporary GPP at regional and global scales is an important first step towards developing realistic prognostic models of the global carbon cycle. Major technical and theoretical advances have improved our understanding of GPP over the past decade. The proliferation of eddy covariance measurements and the systematic organization of these data through Fluxnet provide important new constraints on the distribution of GPP across ecosystems. Concurrently, the availability of high quality remote sensing products makes it possible to monitor biosphere continuously at a global scale on time scales of several days. Here we revise global estimates of global GPP using a light-use-efficiency (LUE) model. We used EVI and NDVI measurements from MODIS to estimate fPAR. PAR was derived as a product of adjusted surface shortwave radiation measurements from GEOS version 4 and conversion factor from shortwave radiation to PAR based on work by Frouin and Pinker (1995). LUE was optimized using Ameriflux GPP estimates and a site-level GPP dataset compiled by Luyssaert et al. (2007). Scalars for temperature were applied using reanalysis observations from GEOS 4. In our analysis we tested different model structures and evaluated their success at predicting GPP at an independent set of measurement sites. We find that the global estimates of GPP of 120 Pg C/yr that are widely used in carbon cycle figures and assessment reports may need to be revised upward.
Presentation Type: Poster
Poster Session: Carbon Cycle Science
NASA TE Funded Awards Represented:
Randerson, James
Global fire emissions derived from Terra and Aqua satellites