An improved Kalman Smoother for atmospheric inversions

[摘要] We explore the use of a fixed-lag Kalman smoother for sequential estimationof atmospheric carbon dioxide fluxes. This technique takes advantage of thefact that most of the information about the spatial distribution of sourcesand sinks is observable within a few months to half of a year of emission.After this period, the spatial structure of sources is diluted by transportand cannot significantly constrain flux estimates. We therefore describe anestimation technique that steps through the observations sequentially, usingonly the subset of observations and modeled transport fields that moststrongly constrain the fluxes at a particular time step. Estimates of eachset of fluxes are sequentially updated multiple times, using measurementstaken at different times, and the estimates and their uncertainties are shownto quickly converge. Final flux estimates are incorporated into thebackground state of CO₂ and transported forward in time, and the finalflux uncertainties and covariances are taken into account when estimating thecovariances of the fluxes still being estimated. The computational demands ofthis technique are greatly reduced in comparison to the standard Bayesiansynthesis technique where all observations are used at once with transportfields spanning the entire period of the observations. It therefore becomespossible to solve larger inverse problems with more observations and forfluxes discretized at finer spatial scales. We also discuss the differencesbetween running the inversion simultaneously with the transport model andrunning it entirely off-line with pre-calculated transport fields. We findthat the latter can be done with minimal error if time series of transportfields of adequate length are pre-calculated.

[学科分类] 大气科学