Bayesian characterization of uncertainties surrounding fluvial flood hazard estimates

Sharma, Sanjib; Ghimire, Ganesh R.; Talchabhadel, Rocky; Panthi, Jeeban; Lee, Benjamin Seiyon; Sun, Fengyun; Baniya, Rupesh; Adhikari, Tika B.

doi:10.1080/02626667.2021.1999959

Cited by 4 publications

(4 citation statements)

References 74 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Given the temporal and spatial variability of flood events as well as the complexity of watersheds, the pursuit of a “perfect” model that can incorporate all hydrologic and hydraulic processes and handle different scenarios is encouraging, but this pursuit faces multiple challenges. The challenges in simulating flooding processes, including the limitations of governing mathematic principles, estimation of parameters, measurement of driving forces, and computational efficiency, are important issues that modelers need to take into consideration in order to provide reliable and robust predictions about flooding (Jafarzadegan et al., 2021; Kobarfard et al., 2022; Liu & Merwade, 2018; Merwade et al., 2008; Pappenberger et al., 2005, 2006; Sharma et al., 2022; Teng et al., 2017). Additionally, “equifinality” in hydrologic and hydraulic modeling may lead to multiple models or different model configurations to yield similar results that match the observations equally well (Beven & Binley, 1992; Refsgaard et al., 2012; Von Bertalanffy, 1972).…”

Section: Introductionmentioning

confidence: 99%

“…Given the limitations of the default EM algorithm in the BMA analysis discussed above, the Markov Chain Monte Carlo (MCMC) sampling method (Robert et al., 1999) is proposed in this study to estimate the BMA parameters. It has been shown that the application of MCMC method for hydrology can provide a full view of the unknown parameter's posterior probability distribution (Gaume et al., 2010; Nguyen et al., 2021; Reis & Stedinger, 2005; Sharma et al., 2022; Vrugt et al., 2008; Wang et al., 2017; Zhao et al., 2021), which can provide an explicit representation and quantification for the parameter uncertainty. Specifically, this method can be implemented by different algorithms to generate a sequence of stochastic samples that converge to the target PDF of the unknown parameter (Luengo et al., 2020).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Improving Bayesian Model Averaging for Ensemble Flood Modeling Using Multiple Markov Chains Monte Carlo Sampling

Huang,

Merwade

2023

Water Resources Research

View full text Add to dashboard Cite

As all kinds of numerical models are emerging in hydrologic and hydraulic engineering, Bayesian model averaging (BMA) is one of the popular multi‐model methods used to account for various uncertainties in the flood modeling process and generate robust ensemble predictions. The reliability of BMA parameters (weights and variances) determines the accuracy of BMA predictions. However, the uncertainty in BMA parameters with fixed values, which are usually obtained from Expectation‐Maximization (EM) algorithm, has not been adequately investigated in BMA‐related applications over the past few decades. Given the limitations of the default EM algorithm, Metropolis‐Hastings (M‐H) algorithm, which is one of the most widely used algorithms in Markov Chain Monte Carlo (MCMC), is proposed to estimate BMA parameters. Both numerical experiments and one‐dimensional Hydrologic Engineering Center‐River Analysis System models are employed to examine the applicability of M‐H algorithm with multiple independent Markov chains. The performances of EM and M‐H algorithms are compared based on the daily water stage predictions from 10 model members. Results show that BMA weights estimated from both algorithms are comparable, while BMA variances obtained from M‐H algorithm are closer to the given variances in the numerical experiment. Moreover, the normal proposal distribution used in M‐H algorithm can yield narrower distributions for BMA weights than those from the uniform proposal distribution. Overall, MCMC approach with multiple chains can provide more information associated with the uncertainty of BMA parameters and its performance is better than the default EM algorithm in terms of both deterministic and probabilistic evaluation metrics as well as algorithm flexibility.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Improving Bayesian Model Averaging for Ensemble Flood Modeling Using Multiple Markov Chains Monte Carlo Sampling

Huang,

Merwade

2023

Water Resources Research

View full text Add to dashboard Cite

show abstract

“…Given the temporal and spatial variability of flood events as well as the complexity of watersheds, the pursuit of a "perfect" model that can incorporate all hydrologic and hydraulic processes and handle different scenarios is encouraging, but this pursuit faces multiple challenges. The challenges in simulating flooding processes, including the limitations of governing mathematic principles, estimation of parameters, measurement of driving forces, and computational efficiency, are important issues that modelers need to take into consideration in order to provide reliable and robust predictions about flooding (Jafarzadegan et al, 2021;Kobarfard et al, 2022;Liu and Merwade, 2018;Merwade et al, 2008;F Pappenberger et al, 2005;Florian Pappenberger et al, 2006;Sharma et al, 2022;Teng et al, 2017). Additionally, "equifinality" in hydrologic and hydraulic modeling may lead to multiple models or different model configurations to yield similar results that match the observations equally well (Beven and Binley, 1992;Refsgaard et al, 2012;Von Bertalanffy, 1972).…”

Section: Introductionmentioning

confidence: 99%

“…Given the limitations of the default EM algorithm in the BMA analysis discussed above, the Markov Chain Monte Carlo (MCMC) sampling method (Robert et al, 1999) is proposed in this study to estimate the BMA parameters. It has been shown that the application of MCMC method for hydrology can provide a full view of the unknown parameter's posterior probability distribution (Gaume et al, 2010;Nguyen et al, 2021;Reis Jr and Stedinger, 2005;Sharma et al, 2022;Vrugt et al, 2008;Wang et al, 2017;Zhao et al, 2021), which can provide an explicit representation and quantification for the parameter uncertainty. Specifically, this method can be implemented by different algorithms to generate a sequence of stochastic samples that converge to the target probability density function of the unknown parameter (Luengo et al, 2020).…”

Section: Introductionmentioning

confidence: 99%

Improving Bayesian Model Averaging for Ensemble Flood Modeling Using Multiple Markov Chains Monte Carlo Sampling

Huang

Merwade

2023

Preprint

View full text Add to dashboard Cite

As all kinds of physics-based and data-driven models are emerging in hydrologic and hydraulic engineering, Bayesian model averaging (BMA) is one of the popular multi-model methods used to account for various uncertainty sources in the flood modeling process and generate robust ensemble predictions. The reliability of BMA parameters (weights and variances) determines the accuracy of BMA predictions. However, the uncertainty in BMA parameters with fixed values, which are usually obtained from Expectation-Maximization (EM) algorithm, has not been adequately investigated in BMA-related applications over the past few decades. Given the limitations of the commonly used EM algorithm, Metropolis-Hastings (M-H) algorithm, which is one of the most widely used algorithms in Markov Chain Monte Carlo (MCMC) method, is proposed to estimate BMA parameters. Both numerical experiments and one-dimensional HEC-RAS models are employed to examine the applicability of M-H algorithm with multiple independent Markov chains. The performances of EM and M-H algorithms are compared based on the daily water stage predictions from 10 model members. Results show that BMA weights estimated from both algorithms are comparable, while BMA variances obtained from M-H algorithm are closer to the given variances in the numerical experiment. Moreover, the normal proposal used in M-H algorithm can yield narrower distributions for BMA weights than those from the uniform proposal. Overall, MCMC approach with multiple chains can provide more information associated with the uncertainty of BMA parameters and its performance is better than the default EM algorithm in terms of multiple evaluation metrics as well as algorithm flexibility.

show abstract

Estimating flood recurrence uncertainty for non-stationary regimes

Gomes,

Marques,

Souza

2023

RBRH

View full text Add to dashboard Cite

Assuming non-stationarity in flood frequency models is still controversial due to uncertainty in estimates. In this study, a hierarchical Bayesian framework for flood frequency analysis is presented without assuming the stationarity hypothesis. We account data and model uncertainty in all modelling steps and use the Pardo River, Brazil, as study case. Results showed the presence of increasing trends in floods in Pardo River. The stationary model underestimated floods compared to the non-stationary model. Physical-based covariates models performed better than time-based showing the importance of adding physical covariates to explain the trend behavior. The presented model is adaptable to other case. Finally, this study provided guidance for the flood recurrence estimation under non-stationary conditions.

show abstract

Bayesian characterization of uncertainties surrounding fluvial flood hazard estimates

Cited by 4 publications

References 74 publications

Improving Bayesian Model Averaging for Ensemble Flood Modeling Using Multiple Markov Chains Monte Carlo Sampling

Improving Bayesian Model Averaging for Ensemble Flood Modeling Using Multiple Markov Chains Monte Carlo Sampling

Improving Bayesian Model Averaging for Ensemble Flood Modeling Using Multiple Markov Chains Monte Carlo Sampling

Estimating flood recurrence uncertainty for non-stationary regimes

Contact Info

Product

Resources

About