High-frequency atmospheric CO2 measurements
should become increasingly available by the end of this decade from a variety
of sources, including low-Earth orbiting satellites. If of sufficient accuracy,
these should allow the functioning of the global carbon cycle to be monitored
at fine time/space resolutions using atmospheric transport inversions. Since
traditional direct inversion methods (e.g., Bayesian synthesis) become
computationally infeasible at these resolutions, we use an approximate method,
variational data assimilation, to estimate surface CO2 fluxes at
spatial resolutions ranging from 10x10 degrees to 1x1 degrees and at time
resolutions ranging from 2 weeks to 1 hour. We assess its performance using simulated
data, including the effects of realistic transport and data errors.
Author: D.F. Baker, S. Doney, and D. Schimel (dfb at ucar dot edu)
Filesize: 12.88 Kb