Year: 2019
Type: article
Abstract: The Bayesian paradigm is becoming an increasingly popular framework for estimation and uncertainty quantification of unknown parameters in geophysical inversion problems. Badlands is a landscape evolution model for simulating topography evolution at a broad range of spatial and temporal scales. Our previous work presented Bayeslands that used the Bayesian inference for estimating unknown parameters in the Badlands model using Markov chain Monte Carlo sampling. Bayeslands faced challenges in terms of computational issues and convergence due to multimodal posterior distributions. Parallel tempering is an advanced Markov chain Monte Carlo method suited for irregular and multimodal posterior distributions. In this paper, we extend Bayeslands using parallel tempering with high‐performance computing to address previous limitations in Bayeslands. Our results show that parallel tempering Bayeslands not only reduces the computation time‚ but also provides an improvement in sampling multimodal posterior distributions, which motivates future application to continental scale landscape evolution models.
DOI: 10.1029/2019gc008465
Authors: Rohitash Chandra, R. Dietmar Müller, Danial Azam, Ratneel Deo, Nathaniel Butterworth, Tristan Salles, Sally Cripps
Institution: The University of Sydney
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2811301477