Hydrologic data assimilation using particle Markov chain Monte Carlo simulation: Theory, concepts and applications

J.A. Vrugt, C.J.F. ter Braak, C.G.H. Diks

Research output: Contribution to journalArticleAcademicpeer-review

98 Citations (Scopus)

Abstract

During the past decades much progress has been made in the development of computer based methods for parameter and predictive uncertainty estimation of hydrologic models. The goal of this paper is twofold. As part of this special anniversary issue we first shortly review the most important historical developments in hydrologic model calibration and uncertainty analysis that has led to current perspectives. Then, we introduce theory, concepts and simulation results of a novel data assimilation scheme for joint inference of model parameters and state variables. This Particle-DREAM method combines the strengths of sequential Monte Carlo sampling and Markov chain Monte Carlo simulation and is especially designed for treatment of forcing, parameter, model structural and calibration data error. Two different variants of Particle-DREAM are presented to satisfy assumptions regarding the temporal behavior of the model parameters. Simulation results using a 40-dimensional atmospheric “toy” model, the Lorenz attractor and a rainfall–runoff model show that Particle-DREAM, P-DREAM(VP) and P-DREAM(IP) require far fewer particles than current state-of-the-art filters to closely track the evolving target distribution of interest, and provide important insights into the information content of discharge data and non-stationarity of model parameters. Our development follows formal Bayes, yet Particle-DREAM and its variants readily accommodate hydrologic signatures, informal likelihood functions or other (in)sufficient statistics if those better represent the salient features of the calibration data and simulation model used.
Original languageEnglish
Pages (from-to)457-478
JournalAdvances in Water Resources
Volume51
DOIs
Publication statusPublished - 2013

Fingerprint

Markov chain
data assimilation
simulation
calibration
particle
uncertainty analysis
parameter
filter
sampling

Keywords

  • rainfall-runoff models
  • stochastic parameter-estimation
  • ensemble kalman filter
  • global optimization
  • differential evolution
  • streamflow simulation
  • automatic calibration
  • metropolis algorithm
  • genetic algorithm
  • input uncertainty

Cite this

@article{c3a9a3cc37614fa9a511d6ce7b3bd68a,
title = "Hydrologic data assimilation using particle Markov chain Monte Carlo simulation: Theory, concepts and applications",
abstract = "During the past decades much progress has been made in the development of computer based methods for parameter and predictive uncertainty estimation of hydrologic models. The goal of this paper is twofold. As part of this special anniversary issue we first shortly review the most important historical developments in hydrologic model calibration and uncertainty analysis that has led to current perspectives. Then, we introduce theory, concepts and simulation results of a novel data assimilation scheme for joint inference of model parameters and state variables. This Particle-DREAM method combines the strengths of sequential Monte Carlo sampling and Markov chain Monte Carlo simulation and is especially designed for treatment of forcing, parameter, model structural and calibration data error. Two different variants of Particle-DREAM are presented to satisfy assumptions regarding the temporal behavior of the model parameters. Simulation results using a 40-dimensional atmospheric “toy” model, the Lorenz attractor and a rainfall–runoff model show that Particle-DREAM, P-DREAM(VP) and P-DREAM(IP) require far fewer particles than current state-of-the-art filters to closely track the evolving target distribution of interest, and provide important insights into the information content of discharge data and non-stationarity of model parameters. Our development follows formal Bayes, yet Particle-DREAM and its variants readily accommodate hydrologic signatures, informal likelihood functions or other (in)sufficient statistics if those better represent the salient features of the calibration data and simulation model used.",
keywords = "rainfall-runoff models, stochastic parameter-estimation, ensemble kalman filter, global optimization, differential evolution, streamflow simulation, automatic calibration, metropolis algorithm, genetic algorithm, input uncertainty",
author = "J.A. Vrugt and {ter Braak}, C.J.F. and C.G.H. Diks",
year = "2013",
doi = "10.1016/j.advwatres.2012.04.002",
language = "English",
volume = "51",
pages = "457--478",
journal = "Advances in Water Resources",
issn = "0309-1708",
publisher = "Elsevier",

}

Hydrologic data assimilation using particle Markov chain Monte Carlo simulation: Theory, concepts and applications. / Vrugt, J.A.; ter Braak, C.J.F.; Diks, C.G.H.

In: Advances in Water Resources, Vol. 51, 2013, p. 457-478.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Hydrologic data assimilation using particle Markov chain Monte Carlo simulation: Theory, concepts and applications

AU - Vrugt, J.A.

AU - ter Braak, C.J.F.

AU - Diks, C.G.H.

PY - 2013

Y1 - 2013

N2 - During the past decades much progress has been made in the development of computer based methods for parameter and predictive uncertainty estimation of hydrologic models. The goal of this paper is twofold. As part of this special anniversary issue we first shortly review the most important historical developments in hydrologic model calibration and uncertainty analysis that has led to current perspectives. Then, we introduce theory, concepts and simulation results of a novel data assimilation scheme for joint inference of model parameters and state variables. This Particle-DREAM method combines the strengths of sequential Monte Carlo sampling and Markov chain Monte Carlo simulation and is especially designed for treatment of forcing, parameter, model structural and calibration data error. Two different variants of Particle-DREAM are presented to satisfy assumptions regarding the temporal behavior of the model parameters. Simulation results using a 40-dimensional atmospheric “toy” model, the Lorenz attractor and a rainfall–runoff model show that Particle-DREAM, P-DREAM(VP) and P-DREAM(IP) require far fewer particles than current state-of-the-art filters to closely track the evolving target distribution of interest, and provide important insights into the information content of discharge data and non-stationarity of model parameters. Our development follows formal Bayes, yet Particle-DREAM and its variants readily accommodate hydrologic signatures, informal likelihood functions or other (in)sufficient statistics if those better represent the salient features of the calibration data and simulation model used.

AB - During the past decades much progress has been made in the development of computer based methods for parameter and predictive uncertainty estimation of hydrologic models. The goal of this paper is twofold. As part of this special anniversary issue we first shortly review the most important historical developments in hydrologic model calibration and uncertainty analysis that has led to current perspectives. Then, we introduce theory, concepts and simulation results of a novel data assimilation scheme for joint inference of model parameters and state variables. This Particle-DREAM method combines the strengths of sequential Monte Carlo sampling and Markov chain Monte Carlo simulation and is especially designed for treatment of forcing, parameter, model structural and calibration data error. Two different variants of Particle-DREAM are presented to satisfy assumptions regarding the temporal behavior of the model parameters. Simulation results using a 40-dimensional atmospheric “toy” model, the Lorenz attractor and a rainfall–runoff model show that Particle-DREAM, P-DREAM(VP) and P-DREAM(IP) require far fewer particles than current state-of-the-art filters to closely track the evolving target distribution of interest, and provide important insights into the information content of discharge data and non-stationarity of model parameters. Our development follows formal Bayes, yet Particle-DREAM and its variants readily accommodate hydrologic signatures, informal likelihood functions or other (in)sufficient statistics if those better represent the salient features of the calibration data and simulation model used.

KW - rainfall-runoff models

KW - stochastic parameter-estimation

KW - ensemble kalman filter

KW - global optimization

KW - differential evolution

KW - streamflow simulation

KW - automatic calibration

KW - metropolis algorithm

KW - genetic algorithm

KW - input uncertainty

U2 - 10.1016/j.advwatres.2012.04.002

DO - 10.1016/j.advwatres.2012.04.002

M3 - Article

VL - 51

SP - 457

EP - 478

JO - Advances in Water Resources

JF - Advances in Water Resources

SN - 0309-1708

ER -