A novel heteroscedastic error model for a fully distributed groundwater flow model

Syed Md Touidul Mustafa, Jiri Nossent, Gert Ghysels, Marijke Huysmans

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In hydrological modelling, it is often observed that the error on the simulated flow values is heteroscedastic. In order to obtain homoscedastic errors, researchers have been applying transformations of the flow values (e.g. a Box-Cox transformation) within their model calibration procedure or have introduced an error model within a Bayesian framework for parameter estimation and uncertainty analysis. Transient numerical groundwater flow models are also affected by different sources of uncertainty. Nevertheless, the possible heteroscedasticity of the errors on the simulated groundwater levels has been mostly neglected when performing uncertainty analysis for these types of models. Therefore, we introduce a novel heteroscedastic error model for groundwater level predictions within a general and flexible Bayesian framework. In this way, we can consider the heteroscedasticity of the groundwater level error along with the parameter uncertainty. Moreover, we can also analyze these two sources of uncertainty in combination with input uncertainty. The proposed methodology is applied on a fully distributed physically-based groundwater flow model of an overexploited aquifer in Bangladesh. The results of the study confirm that the heteroscedasticity of the groundwater level error should be considered and that it has an effect on the model predictions and parameter distributions. It is also shown that the observation coverage of the parameter uncertainty band increases from 1.5 % to 8.5 % when the heteroscedasticity is explicitly taken into account along with model parameter uncertainty.
Original languageEnglish
Title of host publicationProceedings of the 9th International Congress on Environmental Modelling and Software, June 24-28, Fort Collins, Colorado, USA
EditorsM. Arabi, O. David, J. Carlson, D.P. Ames
PublisheriEMSs
Number of pages1
Publication statusPublished - Jun 2018
Event9th International Congress on Environmental Modelling and Software
: "Modelling for Sustainable Food-Energy-Water Systems"
- Fort Collins, United States
Duration: 24 Jun 201828 Jun 2018

Conference

Conference9th International Congress on Environmental Modelling and Software
Abbreviated titleiEMSs 2018
CountryUnited States
CityFort Collins
Period24/06/1828/06/18

Fingerprint

groundwater flow
groundwater
uncertainty analysis
hydrological modeling
prediction
aquifer
calibration
methodology
parameter

Keywords

  • groundwater flow model
  • Bayesian approach
  • heteroscedastic
  • uncertainty quantification

Cite this

Mustafa, S. M. T., Nossent, J., Ghysels, G., & Huysmans, M. (2018). A novel heteroscedastic error model for a fully distributed groundwater flow model. In M. Arabi, O. David, J. Carlson, & D. P. Ames (Eds.), Proceedings of the 9th International Congress on Environmental Modelling and Software, June 24-28, Fort Collins, Colorado, USA iEMSs.

A novel heteroscedastic error model for a fully distributed groundwater flow model. / Mustafa, Syed Md Touidul; Nossent, Jiri; Ghysels, Gert; Huysmans, Marijke.

Proceedings of the 9th International Congress on Environmental Modelling and Software, June 24-28, Fort Collins, Colorado, USA. ed. / M. Arabi; O. David; J. Carlson; D.P. Ames. iEMSs, 2018.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Mustafa, SMT, Nossent, J, Ghysels, G & Huysmans, M 2018, A novel heteroscedastic error model for a fully distributed groundwater flow model. in M Arabi, O David, J Carlson & DP Ames (eds), Proceedings of the 9th International Congress on Environmental Modelling and Software, June 24-28, Fort Collins, Colorado, USA. iEMSs, 9th International Congress on Environmental Modelling and Software
, Fort Collins, United States, 24/06/18.
Mustafa SMT, Nossent J, Ghysels G, Huysmans M. A novel heteroscedastic error model for a fully distributed groundwater flow model. In Arabi M, David O, Carlson J, Ames DP, editors, Proceedings of the 9th International Congress on Environmental Modelling and Software, June 24-28, Fort Collins, Colorado, USA. iEMSs. 2018
Mustafa, Syed Md Touidul ; Nossent, Jiri ; Ghysels, Gert ; Huysmans, Marijke. / A novel heteroscedastic error model for a fully distributed groundwater flow model. Proceedings of the 9th International Congress on Environmental Modelling and Software, June 24-28, Fort Collins, Colorado, USA. editor / M. Arabi ; O. David ; J. Carlson ; D.P. Ames. iEMSs, 2018.
@inproceedings{3d8d18a630ea43d5a6228545598ad551,
title = "A novel heteroscedastic error model for a fully distributed groundwater flow model",
abstract = "In hydrological modelling, it is often observed that the error on the simulated flow values is heteroscedastic. In order to obtain homoscedastic errors, researchers have been applying transformations of the flow values (e.g. a Box-Cox transformation) within their model calibration procedure or have introduced an error model within a Bayesian framework for parameter estimation and uncertainty analysis. Transient numerical groundwater flow models are also affected by different sources of uncertainty. Nevertheless, the possible heteroscedasticity of the errors on the simulated groundwater levels has been mostly neglected when performing uncertainty analysis for these types of models. Therefore, we introduce a novel heteroscedastic error model for groundwater level predictions within a general and flexible Bayesian framework. In this way, we can consider the heteroscedasticity of the groundwater level error along with the parameter uncertainty. Moreover, we can also analyze these two sources of uncertainty in combination with input uncertainty. The proposed methodology is applied on a fully distributed physically-based groundwater flow model of an overexploited aquifer in Bangladesh. The results of the study confirm that the heteroscedasticity of the groundwater level error should be considered and that it has an effect on the model predictions and parameter distributions. It is also shown that the observation coverage of the parameter uncertainty band increases from 1.5 {\%} to 8.5 {\%} when the heteroscedasticity is explicitly taken into account along with model parameter uncertainty.",
keywords = "groundwater flow model, Bayesian approach, heteroscedastic, uncertainty quantification",
author = "Mustafa, {Syed Md Touidul} and Jiri Nossent and Gert Ghysels and Marijke Huysmans",
year = "2018",
month = "6",
language = "English",
editor = "M. Arabi and O. David and J. Carlson and D.P. Ames",
booktitle = "Proceedings of the 9th International Congress on Environmental Modelling and Software, June 24-28, Fort Collins, Colorado, USA",
publisher = "iEMSs",

}

TY - GEN

T1 - A novel heteroscedastic error model for a fully distributed groundwater flow model

AU - Mustafa, Syed Md Touidul

AU - Nossent, Jiri

AU - Ghysels, Gert

AU - Huysmans, Marijke

PY - 2018/6

Y1 - 2018/6

N2 - In hydrological modelling, it is often observed that the error on the simulated flow values is heteroscedastic. In order to obtain homoscedastic errors, researchers have been applying transformations of the flow values (e.g. a Box-Cox transformation) within their model calibration procedure or have introduced an error model within a Bayesian framework for parameter estimation and uncertainty analysis. Transient numerical groundwater flow models are also affected by different sources of uncertainty. Nevertheless, the possible heteroscedasticity of the errors on the simulated groundwater levels has been mostly neglected when performing uncertainty analysis for these types of models. Therefore, we introduce a novel heteroscedastic error model for groundwater level predictions within a general and flexible Bayesian framework. In this way, we can consider the heteroscedasticity of the groundwater level error along with the parameter uncertainty. Moreover, we can also analyze these two sources of uncertainty in combination with input uncertainty. The proposed methodology is applied on a fully distributed physically-based groundwater flow model of an overexploited aquifer in Bangladesh. The results of the study confirm that the heteroscedasticity of the groundwater level error should be considered and that it has an effect on the model predictions and parameter distributions. It is also shown that the observation coverage of the parameter uncertainty band increases from 1.5 % to 8.5 % when the heteroscedasticity is explicitly taken into account along with model parameter uncertainty.

AB - In hydrological modelling, it is often observed that the error on the simulated flow values is heteroscedastic. In order to obtain homoscedastic errors, researchers have been applying transformations of the flow values (e.g. a Box-Cox transformation) within their model calibration procedure or have introduced an error model within a Bayesian framework for parameter estimation and uncertainty analysis. Transient numerical groundwater flow models are also affected by different sources of uncertainty. Nevertheless, the possible heteroscedasticity of the errors on the simulated groundwater levels has been mostly neglected when performing uncertainty analysis for these types of models. Therefore, we introduce a novel heteroscedastic error model for groundwater level predictions within a general and flexible Bayesian framework. In this way, we can consider the heteroscedasticity of the groundwater level error along with the parameter uncertainty. Moreover, we can also analyze these two sources of uncertainty in combination with input uncertainty. The proposed methodology is applied on a fully distributed physically-based groundwater flow model of an overexploited aquifer in Bangladesh. The results of the study confirm that the heteroscedasticity of the groundwater level error should be considered and that it has an effect on the model predictions and parameter distributions. It is also shown that the observation coverage of the parameter uncertainty band increases from 1.5 % to 8.5 % when the heteroscedasticity is explicitly taken into account along with model parameter uncertainty.

KW - groundwater flow model

KW - Bayesian approach

KW - heteroscedastic

KW - uncertainty quantification

UR - http://documentatiecentrum.watlab.be/owa/imis.php?module=ref&refid=305816

M3 - Conference contribution

BT - Proceedings of the 9th International Congress on Environmental Modelling and Software, June 24-28, Fort Collins, Colorado, USA

A2 - Arabi, M.

A2 - David, O.

A2 - Carlson, J.

A2 - Ames, D.P.

PB - iEMSs

ER -