Reduced-Dimension Groundwater Model Emulation for Scenario Analysis and Decision Support

Bibliographic Details
Main Author: Tracy, Jacob N.
Language:English
Published: University of Cincinnati / OhioLINK 2019
Subjects:
Online Access:http://rave.ohiolink.edu/etdc/view?acc_num=ucin1573574885505114
id ndltd-OhioLink-oai-etd.ohiolink.edu-ucin1573574885505114
record_format oai_dc
spelling ndltd-OhioLink-oai-etd.ohiolink.edu-ucin15735748855051142021-12-18T05:25:47Z Reduced-Dimension Groundwater Model Emulation for Scenario Analysis and Decision Support Tracy, Jacob N. Engineering Emulation Groundwater Decision-Support Optimization Model emulation has become an integral tool in scenario analysis, uncertainty analysis, risk assessment, and calibration of environmental models. When models have relatively long runtimes or many parameters, the computational cost of any modeling exercise involving an exploration of the possibility space scales dramatically. This has traditionally been a limiting factor in methods such as robust optimization, decision-making under deep uncertainty (DMDU), and modeling uncertainty analysis; which all need a suite of model results across a meaningful joint distribution of the possibility space. Model emulation is a field created to remove this limiting factor by creating statistical frameworks trained on model outputs at representative parameter values so that outputs at untried values can be approximated with high fidelity. However, in groundwater modeling the advances have not been as helpful as most techniques focus on emulating complex dynamic models with relatively few outputs. In contrast, groundwater modeling is often not focused on modeling point processes, but rather aquifer processes across a study area with many numerical grid elements as the model outputs. This has historically meant that computationally expensive groundwater analysis has to be completed by aggregating and averaging the outputs to a manageable number, or assuming uniform properties between all model outputs. The former method can result in significant information loss while the latter two methods lose efficacy as the number of elements in the model increase or significant heterogeneity exists in the modeled system. This paper introduces a methodology capable of emulating high-dimensional (tens of thousands of elements) groundwater models while preserving their original spatial resolution by directly modeling the results of dimensionality reduction. This results in a computationally manageable emulation problem while still being able to reproduce the spatial processes of the original physical model. The process is demonstrated in a case study of a model of the aquifer of the Valley of Mexico wherein the spatial preservation of information is crucial as part of a robust optimization framework for planning water supply investments. 2019 English text University of Cincinnati / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=ucin1573574885505114 http://rave.ohiolink.edu/etdc/view?acc_num=ucin1573574885505114 unrestricted This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws.
collection NDLTD
language English
sources NDLTD
topic Engineering
Emulation
Groundwater
Decision-Support
Optimization
spellingShingle Engineering
Emulation
Groundwater
Decision-Support
Optimization
Tracy, Jacob N.
Reduced-Dimension Groundwater Model Emulation for Scenario Analysis and Decision Support
author Tracy, Jacob N.
author_facet Tracy, Jacob N.
author_sort Tracy, Jacob N.
title Reduced-Dimension Groundwater Model Emulation for Scenario Analysis and Decision Support
title_short Reduced-Dimension Groundwater Model Emulation for Scenario Analysis and Decision Support
title_full Reduced-Dimension Groundwater Model Emulation for Scenario Analysis and Decision Support
title_fullStr Reduced-Dimension Groundwater Model Emulation for Scenario Analysis and Decision Support
title_full_unstemmed Reduced-Dimension Groundwater Model Emulation for Scenario Analysis and Decision Support
title_sort reduced-dimension groundwater model emulation for scenario analysis and decision support
publisher University of Cincinnati / OhioLINK
publishDate 2019
url http://rave.ohiolink.edu/etdc/view?acc_num=ucin1573574885505114
work_keys_str_mv AT tracyjacobn reduceddimensiongroundwatermodelemulationforscenarioanalysisanddecisionsupport
_version_ 1723964834103427072