Reconciling uncertainties in integrated science and policy models: Applications to global climate change

摘要

In this thesis, tools of data reconciliation are used to integrate available information into scientific and policy models of greenhouse gases. The role of uncertainties in scientific and policy models of global climate change is examined, and implications for global change policy are drawn. A chance-constrained methodology was developed and used to perform inversions on the global methane cycle. Budgets of methane that are consistent with source fluxes, isotopic and ice core measurements were determined. While it is not possible to come up with a single budget for CH4, performing the calculation with a number of sets of assumed priors suggests a convergence in the allowed range for sources. Our results compare favorably with the most recent measurements of flux estimates. For comparison, a similar analysis using bayes monte carlo simulation was performed. Two analyses that attempt to quantify the missing sink were performed. First, a steady state analysis of the carbon cycle was used to determine the pre-industrial inter-hemispheric carbon concentration gradient. Second, a full blown dynamic inversion of the carbon cycle was performed. An advection diffusion ocean model with surface chemistry, coupled to box models of the atmosphere and the biosphere was inverted to fit available measurements of C-12 and C-14 carbon isotopes using Differential-Algebraic Optimization. The model effectively suggests that the 'missing' sink for carbon is hiding in the biosphere. Scenario dependent trace gas indices were calculated for CH4, N2O, and HCFC-22.