Author: Trine Enemark
Enemark, Trine, 2020 Hydrogeological conceptual model development and testing, Flinders University, College of Science and Engineering
Terms of Use: This electronic version is (or will be) made publicly available by Flinders University in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. You may use this material for uses permitted under the Copyright Act 1968. If you are the owner of any included third party copyright material and/or you believe that any material has been made available without permission of the copyright owner please contact copyright@flinders.edu.au with the details.
Groundwater models are widely applied in groundwater management to guide decision making. The success of groundwater management is directly dependent on a good understanding of the groundwater system. A conceptual model is a summary of our current knowledge about a groundwater system describing the dominating processes and the overall physical structure of the geology. One of the major sources of uncertainties in groundwater model predictions is the conceptual uncertainty that arises when more than one conceptual model can explain the available data. The goal of this thesis is to identify current approaches, unify scattered insights and develop a systematic methodology of hydrogeological conceptual model development and testing, which leads to an improved characterisation of conceptual uncertainty.
Conceptual model development involves formulation of hypotheses about the groundwater system functioning. These are the initial decisions in the modelling that drive the groundwater model predictions and form the basis of the uncertainty analysis. In this thesis we advocate for a systematic model development approach based on mutually exclusive hypotheses. We developed bold hypotheses about the model structure, challenging what was considered possible for the system, in order to give more transparent explanation of which model structures were considered possible.
Conceptual model testing consists of holding the developed models against data to evaluate their validity. Model testing is essential in order to gain confidence in the developed models and remove those models from the ensemble that are inconsistent with the data. We show that model testing does not have to be a time-consuming task but can happen in relatively simple forward models. We advocate for reserving as much data as possible for the model testing exercise rather than using all data for model development in order to be able to explain why no other conceptual models are plausible.
The methodology developed in this thesis is applied to the Wildman River area, Northern Territory, Australia. By acknowledging the existence of conceptual uncertainty, we increase the confidence in the water balance for the area. A second aspect of the investigation is the connectivity of sinkhole-like depressions in the area to groundwater and whether they may act as conduits of groundwater recharge.
The insights gained from this thesis enables more accessible methodology for conceptual model development and testing. By acknowledging and accounting for conceptual uncertainty, more confidence can be gained in groundwater model predictions leading to improved groundwater management.
Keywords: Conceptual models, Model evaluation, Model rejection, Multi-model framework, Conceptual model uncertainty, Groundwater models, Hydrogeology
Subject: Marine Geology and Geophysics thesis
Thesis type: Doctor of Philosophy
Completed: 2020
School: College of Science and Engineering
Supervisor: Okke Batelaan