Results from a Regional Geostatistical Inversion over North America for 2004
Sharon
M
Gourdji, University of Michigan, sgourdji@umich.edu
Kim
L
Mueller, University of Michigan, kimlm@umich.edu
Vineet
Yadav, University of Michigan, vineety@umich.edu
Deborah
Huntzinger, University of Michigan, dnhuntzi@umich.edu
(Presenting)
Adam
Hirsch, National Renewable Energy Laboratory, adam.hirsch@nrel.gov
Arlyn
Andrews, National Oceanic & Atmospheric Administration, arlyn.andrews@noaa.gov
Michael
Trudeau, National Oceanic & Atmospheric Administration, michael.trudeau@noaa.gov
Marc
Fischer, Lawrence Berkeley National Laboratory, mlfischer@lbl.gov
William
Munger, Harvard University, jwmunger@seas.harvard.edu
Steven
Wofsy, Harvard University, swofsy@deas.harvard.edu
Douglas
Worthy, Environment Canada, doug.worthy@ec.gc.ca
Anna
Michalak, University of Michigan, amichala@umich.edu
(Presenting)
This poster presents results from a North American regional grid-scale geostatistical inversion for 2004. CO2 fluxes are estimated at a 1°x1°, 8-day average resolution using 3-hourly concentration data from nine continuous measurement locations operational in 2004, as well as weekly flask data where available. Concentration footprints are derived from the Stochastic Time-Inverted Lagrangian Transport Model using high-resolution winds from the Weather Research & Forecasting (WRF) model. Covariance parameters for the inversion are estimated using the atmospheric data.
Given that regional inversions aim to identify sources and sinks at smaller spatial scales than previous global inverse models, this study first identifies the smallest scales at which fluxes can be reliably interpreted. This is accomplished by using synthetic data inversions where the “truth” from a biospheric model is known and can be compared to inversion results. Fluxes from a real data inversion are then presented, at these scales, in comparison to 13 forward biospheric models collected for the North American Carbon Program Regional Interim Synthesis. Despite a large spread in fluxes from the forward models, inversion results show surprising agreement with the across-model mean, except in agricultural areas where the inversion identifies strong sources in the spring and fall not visible in the bottom-up fluxes. The inversion also identifies eastern temperate and boreal forests as net annual sinks for this year. Overall, this regional geostatistical inversion shows promise in distinguishing between different plausible flux scenarios from biospheric models, especially as the continuous measurement network expands across North America in subsequent years.
Presentation Type: Poster
Poster Session: Carbon Cycle Science
NASA TE Funded Awards Represented:
Michalak, Anna
Contstraining North-American Fluxes of Carbon Dioxide and Inferring their Spatiotemporal Covariances through Assimilation of Remote Sensing and Atmospheric Data in a Geostatistical Framework