Ensemble smoother with multiple data assimilation for reverse flow routing

Todaro, Valeria; D’Oria, Marco; Tanda, Maria Giovanna; Gómez‐Hernández, J. Jaime

doi:10.1016/j.cageo.2019.06.002

Cited by 21 publications

(9 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The size of the ensemble is critical in the computational cost of the implementation; it should be as small as possible, to save CPU time, but it should be as large as possible to obtain good experimental covariance estimates and prevent filter inbreeding, the appearance of spurious correlations and avoid filter divergence. (These problems could be mitigated for small ensemble sizes with covariance localization techniques [11,38,42].) [11] studied the sensitivity of the EnKF to, among other factors, the ensemble size, and the choice of the initial ensemble, and they showed that prior knowledge of the underlying field, such as the structure of the covariance function, plays an important role in data assimilation.…”

Section: Introductionmentioning

confidence: 99%

“…As will be shown, this power is intrinsic and can be taken advantage of by the ERFF, too. The concept of localization [38,42] is also included in the implementation of the ERFF to reinforce the notion of spatial correlation by training the random forest giving more weight to the observations that are closer to the point being updated. The ERFF results are benchmarked against the r-EnKF.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Ensemble random forest filter: An alternative to the ensemble Kalman filter for inverse modeling

Godoy¹,

Napa-García²,

Gómez‐Hernández³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

The ensemble random forest filter (ERFF) is presented as an alternative to the ensemble Kalman filter (EnKF) for the purpose of inverse modeling. The EnKF is a data assimilation approach that forecasts and updates parameter estimates sequentially in time as observations are being collected. The updating step is based on the experimental covariances computed from an ensemble of realizations and the updates are given as linear combinations of the differences between observations and forecasted system state values. The ERFF replaces the linear combination in the update step with a non-linear function represented by a random forest. In this way, the non-linear relationships between the parameters to be updated and the observations can be captured and a better update produced. The ERFF is demonstrated for the purpose of log-conductivity identification from piezometric head observations in a number of scenarios with varying degrees of heterogeneity (log-conductivity variances going from 1 up to 6.25 (ln m/d) 2 ), number of realizations in the ensemble (50 or 100), and number of piezometric head observations (18 or 36). In all scenarios, the ERFF works well, being able to reconstruct the log-conductivity spatial heterogeneity while matching the observed piezometric heads at selected control points. For benchmarking purposes the ERFF is compared to the restart EnKF to find that the ERFF is superior to the EnKF for the number of ensemble realizations used (small in typical EnKF applications). Only when the number of realizations grows to 500, the restart EnKF is able to match the performance of the ERFF, albeit at triple the computational cost.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Ensemble random forest filter: An alternative to the ensemble Kalman filter for inverse modeling

Godoy¹,

Napa-García²,

Gómez‐Hernández³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Inverse problems have been addressed for many years in hydrology and hydraulics with both deterministic and stochastic approaches. A good review of inverse models in subsurface hydrology can be found in Zhou et al (2014), and some applications in surface hydrology in D' Oria et al (2014), D'Oria et al (2017 or Todaro et al (2019). Probably, subsurface hydrology has been the area with the largest body of research in the subject, from the early deterministic works (i.e., de Marsily et al 1984;Carrera and Neuman 1986) to the later stochastic ones (i.e, Woodbury and Ulrych 1993;Wen et al 1999;Li et al 2012;Capilla et al 1999;Sun et al 2009;Zhou et al 2012).…”

Section: Introductionmentioning

confidence: 99%

Contaminant-Source Detection in a Water Distribution System Using the Ensemble Kalman Filter

Butera

Gómez‐Hernández

Nicotra

2021

J. Water Resour. Plann. Manage.

Self Cite

View full text Add to dashboard Cite

Early detection of a contamination leach into a water distribution system, followed by the identification of the source and an evaluation of the total amount of contaminant that has been injected into the system is of paramount importance in order to protect water user's health. The ensemble Kalman filter, which has been recently applied in hydrogeology to detect contaminant sources in aquifers, is extended to the identification of a contaminant source and its intensity in a water distribution system. The driving concept is the assimilation of contaminant observations at the nodes of the pipeline network at specified time intervals until enough information has been collected to allow the positioning of the source and the estimation of its intensity. Several scenarios are analyzed considering sources at different nodes, with different delays between the beginning of the pollution and the start of the measurements, with different sampling time intervals, and with different observation ending times. The scenarios are carried out in the bench-marking Anytown network demonstrating the ability of the ensemble Kalman filter for contaminant-source detection in real water distribution systems. The use of the ensemble Kalman filter supposed a major 1Butera, Gómez-Hernández, Nicotra December 22, 202 breakthrough in the inverse modeling of subsurface flow and transport, the successful results of its application to the synthetic Anytown network warrant further exploration of its capabilities in the realm of water distribution systems.

show abstract

“…Flow hydrographs are one of the fundamental elements for flood modeling which is recorded as time series at hydrometric or gauged stations. The information of the peak discharge of a hydrograph is also essential for flood-risk assessment and management of water resource systems (Todaro et al 2019). However, due to the lack of hydrometric and/or gauged stations, these hydrographs may not be available everywhere (Kaya et al 2017;Jeslin & Sumam 2021).…”

Section: Introductionmentioning

confidence: 99%

Reverse flood routing in an open channel using a combined model of genetic algorithm and a numerical model

Azizipour

Kashefipour

Haghighi

2021

Water Practice and Technology

View full text Add to dashboard Cite

Flood impact assessment in a river system is done with the help of flood routing and this process helps to determine the status of sensitive points of the route in terms of flood entry and the resulting risks for urban and rural areas. For flood routing a hydrodynamic numerical model should be implemented and this model needs upstream and downstream boundaries. In some casesو the upstream boundary, which is usually hydrograph, is not available due to the lack of facilities and it is necessary to be generated for numerical model implementation. The purpose of this study is to present an integrated method comprising of an optimization model and a hydrodynamic numerical model for flood modeling in order to determine the upstream hydrograph using the measured downstream hydrograph along a river. The routing procedure consists of three steps: (1) generating a hypothetical upstream hydrograph using genetic algorithm method; (2) hydrodynamic modeling using a numerical simulation model for flood routing according to the hypothetical hydrograph which is generated in the first step; (3) compare the calculated and observed hydrograph in downstream by using a fitness function. This recommended procedure was named as Reverse Flood Routing Method (RFRM) and was then applied to Karun River, the largest river in Iran. Comparison of the final generated upstream hydrograph by the RFRM model with the corresponding measured hydrograph at the upstream boundary (here Ahvaz hydrometric station which was assumed as an ungauged river location) shows the high accuracy of the recommended model in this study.

show abstract

Ensemble smoother with multiple data assimilation for reverse flow routing

Cited by 21 publications

References 47 publications

Ensemble random forest filter: An alternative to the ensemble Kalman filter for inverse modeling

Ensemble random forest filter: An alternative to the ensemble Kalman filter for inverse modeling

Contaminant-Source Detection in a Water Distribution System Using the Ensemble Kalman Filter

Reverse flood routing in an open channel using a combined model of genetic algorithm and a numerical model

Contact Info

Product

Resources

About