Uncertainty quantification of FaIR climate model

Abstract

The FaIR model is a simple climate–carbon-cycle model. It is used for the estimation of temperature change caused by greenhouse gas emissions. The objective of the thesis was the conduction of sensitivity analysis of the FaIR model by using the Markov chain Monte Carlo method and estimation of posterior distributions of 20 internal FaIR parameters. That allowed to estimate the FaIR output uncertainties. The other sources of error (for example uncertainties of the input data) are not considered. The results of the work show improvement of the error estimation comparing to the results calculated by the simple Monte-Carlo method by using the prior distributions of the parameters. In addition, an example use-case is presented as an error validation of the X-Degree Compatibility model which is an economic climate impact model that uses FaIR.