Bayesian inference via Parallel Tempering for Badlands
Projects | | Links: BayesLands GitHub page
Bayesian inference via Parallel Tempering for Badlands
Overview
Basin and Landscape Dynamics (Badlands) is a parallel TIN-based landscape evolution model, built to simulate topography development at various space and time scales. The model is capable of simulating hillslope processes (linear & non-linear diffusion), fluvial incision (‘modified’ Stream Power Law, Transport Capacity Law both for sediment erosion/transport/deposition), spatially and temporally varying geodynamic (horizontal + vertical displacements) and climatic forces which can be used to simulate changes in base level, as well as effects of climate changes or sea-level fluctuations. The model uses gFlex package which is designed to solve elastic plate flexure for applications to Earth’s lithosphere.
Comparison of PT-Bayeslands with SC-Bayeslands for the Cr problem for 50,000 iterations. The panels show estimates of the posterior distribution and trace-plot of the precipitation ($\rho$) for the SC-Bayeslands and PT-Bayeslands respectively. \textcolor{blue}{The trace-plots show the accepted or current values of the chain for the given samples}. Note that the results for SC-Bayeslands are taken from earlier single chain experiments (with 5 \% burnin) while PT-Bayeslands features 25 \%burn-in.
