Bayesian MCMC groundwater residence time inference

Knowledge graph centered on Bayesian MCMC groundwater residence time inference with 72 nodes and 203 connections. Top connected: Crested Butte, Greenhouse gas emissions, groundwater hydrology, evapotranspiration, Dissolved solids concentration.

Method synopsis

Bayesian framework using Markov-chain Monte Carlo to infer groundwater residence time distributions from environmental tracer observations while quantifying parameter uncertainties.

Synthesized from method descriptions across 4 papers using this protocol.

Procedure as described in the canonical source

Steps below were extracted from the paper that introduces this protocol — Constraining Bedrock Groundwater Residence Times in a Mountain System With Environmental Tracer Observations and Bayesian Uncertainty Quantification (2023), Water Resources Research. Implementations in other papers (listed below) may differ.

1. MCMC parameter inference
   Applied Bayesian Markov-chain Monte Carlo using Adaptive Differential Evolution Metropolis algorithm with four independent chains of 50,000 discrete samples each. Used Gelman-Rubin R statistic to evaluate convergence with burn-in phase of 10,000 samples.
   Quantities: 4 independent chains, 50,000 samples per chain, 10,000 burn-in samplesDuration: Computational analysis phaseConditions: Convergence criteria using Gelman-Rubin statistic
   Equipment: pyMC3 Python software

Typical Equipment

• pyMC3 Python software

Output Measurements

• residence time distributions
• recharge parameter uncertainties
• mixing fraction estimates

Papers Using This Protocol (4)

Places Where Used