Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 7 - Advances in Multivariate Spatial Process Modeling for Environmental Data
Type: Invited
Date/Time: Sunday, August 7, 2022 : 2:00 PM to 3:50 PM
Sponsor: Section on Statistics and the Environment
Abstract #319204
Title: Bayesian Global Flux-Inversion of Carbon Dioxide Fluxes via Multivariate Models Constructed Through Conditioning
Author(s): Andrew Zammit Mangion* and Michael Bertolacci and Noel Cressie and Jenny Fisher and Beata Bukosa and Yi Cao
Companies: University of Wollongong and University of Wollongong and University of Wollongong and University of Wollongong and NIWA and University of Wollongong
Keywords: spatial statistics; MCMC; carbon dioxide; conditional approach; atmospheric chemistry
Abstract:

WOMBAT (the WOllongong Methodology for Bayesian Assimilation of Trace-gases) is a fully Bayesian hierarchical statistical framework for flux inversion of trace gases from flask, in situ, and remotely sensed data. At WOMBAT's core is a bivariate model constructed via two probability models, one for the flux, and one for the mole fraction conditional on the flux, which integrates within it a chemical transport model. WOMBAT allows for modelling correlated error, online bias correction, and the provision of uncertainty quantification on all unknowns that appear in the statistical model. We show, in an observing system simulation experiment (OSSE), that these extensions, often not present in conventional inversion systems, are crucial when the data are indeed biased and have errors that are spatio-temporally correlated. We use WOMBAT to infer CO$_2$ fluxes from Orbiting Carbon Observatory-2 (OCO-2) data. We find that WOMBAT's predictions of out-of-sample retrievals obtained from the Total Column Carbon Observing Network are, for the most part, more accurate than those obtained by other inversion groups.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2022 program