Date of Award:

5-2015

Document Type:

Dissertation

Degree Name:

Doctor of Philosophy (PhD)

Department:

Civil and Environmental Engineering

Committee Chair(s)

Richard C. Peralta

Committee

Richard C. Peralta

Committee

Thomas E. Lachmar

Committee

Gary P. Merkley

Committee

Wynn R. Walker

Committee

Mac McKee

Abstract

Simulation-Optimization (S-O) models are exemplary for strategic groundwater management planning. The two components of an S-O model serve different purposes. The simulation component models the aquifer and computes state variables resulting from stimuli such as extraction or diversion while the optimizor computes an optimal set of stimuli that maximize an objective function and satisfy a set of constraints. With the increased use of S-O model for larger and more complex real-world study areas, the use of fast and efficient S-O models has risen. Response matrix methods (RMMs) are usually used as substitute simulators within an S-O to solve the posed optimization problem. To increase the chances of a groundwater management plan to be a success employed S-O models should represent sustainability concepts and quantify the ability of the aquifer to replenish from a climatic anomaly such as drought. Aquifers are usually modeled using a single simulation model. However, there exist cases in which stakeholders’ disagreement about the conceptual model leads to different simulation models for the same aquifer.

We propose a hybrid RMM that uses the strengths of existing RMM. The proposed RMM that is most efficient for situations in which optimizable stimuli can vary through consecutive periods of uniform duration interspersed with periods of different duration. Tested scenarios indicate that the hybrid RMM requires as much or 63-89% less computation time than other RMMs.

Ensuring the acceptability of the aquifer states, including seepage losses to surface waters, on sub-annual (less than a year) basis (sustained yield) and quantifying the aquifer’s ability to acceptably recover from an anomaly such as drought are integral for successful groundwater management. We evaluate sustained yield strategies (SYS) and quantify the resilience of a computed SYS for Cache Valley, Utah (USA). We maximize the number of new residents who can have their indoor and outdoor uses satisfied while maintaining acceptable aquifer-surface waters seepage losses, and limiting new residents to projected increases in population (PIiP). Because the optimal solution can be influenced by many factors, we examine the optimal solution as affected by optimizing: (1) population versus pumped volume; (2) optimizing indoor and outdoor water demand (traditional housing development) versus optimizing indoor population (apartment dwellers) then the portion of that population whose outdoor water demand can be met too; (3) optimizing in the presence of temporally-lagged spatially distributed return flow that is a function of optimal groundwater use versus optimizing in its absence; and (4) annual versus quarter-annual time evaluation of acceptability. Results indicate that Cache Valley aquifer has the capacity to sustain the outdoor water demand of 74%- 83% and the indoor water demand of 83%-100% of the PIiP. Aquifer conditions can rebound from a 2-year drought to within 7% and 5% of acceptable levels in three and eight years, respectively, after the drought has ceased. Reducing pumping rates by 25% allows the aquifer conditions to be rebound to 96% of acceptable levels or rates within 3 years.

To circumvent the necessity that stakeholders reach an agreement on a single simulation model for an aquifer, we propose Multi-Conceptual Model Optimization (MCMO). Unlike the traditional S-O models, MCMO computes optimal strategies that simultaneously satisfy analogous constraints and bounds in multiple numerical models differing by more than parameter values. Applying MCMO to Cache Valley (Utah, USA) reveals that protecting local ecosystem limits the increased groundwater pumping to satisfy only 40% of projected water demand increase using both models simultaneously.

Checksum

6dd5d86c4bf7ca4cc5ab74f67c8d2737

Download

Share

COinS

Copyright for this work is retained by the student. If you have any questions regarding the inclusion of this work in the Digital Commons, please email us at .