BAYESIAN HIERARCHICAL MODELING FOR INTERRING THE CASUAL RELATIONSHIP BETWEEN HUMAN ACTIVITIES AND CLIMATE CHANGE IMPACTS
SAMUEL BAUGH – LAWRENCE BERKELEY NATIONAL LAB
While the impacts of heat waves, droughts, and floods have been increasing along with rising greenhouse gas concentrations, the complex structure of natural variability in the climate system makes it challenging to precisely quantify the extent to which human activities are responsible for observed changes. The statistical methods used by high-profile scientific bodies to address this connection have been observed in recent findings to underestimate the magnitude of variability, resulting in potentially misleading over-confidence. To address this issue, I propose a physically-informed basis function representation of the global covariance structure within a regularized Bayesian hierarchical framework to avoid over-fitting the limited amount of data and to propagate the associated estimation uncertainty to the final inference. When validated against climate model simulation data, this method achieves better-calibrated credible intervals than methods relying on the estimation of potentially uncertain principal components. Incorporating the physically-informed basis representation into a mixture model additionally allows for the error in the climate simulations informing the natural variability component to be assessed and accounted for in the framework. Motivated by the need for policymakers and the public at large to understand the extent to which human activities are responsible for climate impacts at specific locations, an extension of this model leverages the global covariance structure to provide more precise quantification of causal connections at fine spatial scales. Future extensions include the incorporation of deep learning techniques to understand more complex distributions and non-linear causal relationships.