Article ID Journal Published Year Pages File Type
429380 Journal of Computational Science 2014 5 Pages PDF
Abstract

Understanding and characterizing sources of uncertainty in climate modeling is an important task. Because of the ever increasing sophistication and resolution of climate modeling it is increasingly important to develop uncertainty quantification methods that minimize the computational cost that occurs when these methods are added to climate modeling. This research explores the application of sparse stochastic collocation with polynomial edge detection to characterize portions of the probability space associated with the Earth's radiative budget in the Community Earth System Model (CESM). Specifically, we develop surrogate models with error estimates for a range of acceptable input parameters that predict statistical values of the Earth's radiative budget as derived from the CESM simulation. We extend these results in resolution from T31 to T42 and in parameter space increasing the degrees of freedom from two to three.

Related Topics
Physical Sciences and Engineering Computer Science Computational Theory and Mathematics
Authors
, , ,