Agent-based modelling of groundwater systems

Abstract

Groundwater plays a vital role in sustaining ecosystems, food and water security, and will aid human adaptation to climate change. Scenario-based modelling can make unrealistic long-term forecasts as the two-way feedbacks between social and hydrogeological systems are not explicitly included. The goal of this thesis is to develop a modelling framework that enables interdisciplinary, transparent, and collaborative model-based assessments that capture the coevolution of coupled systems of people and groundwater. To achieve this goal, principles of agent-based modelling and complex systems were adapted to groundwater problems, as a novel approach to combine socio-economic and groundwater flow modelling. FlowLogo, a new software based on a finite-difference solution to the governing equation of groundwater flow implemented in NetLogo, a widely-used open-source ABM platform, is introduced. Its capabilities are demonstrated using a hypothetical groundwater management problem and scenario analysis of a suite of policy levers and agent traits. Opportunities and limitations of using FlowLogo as a decision-support tool in a number of groundwater hotspots are discussed. We further test and demonstrate the use of FlowLogo through the “Groundwater Commons Game” (GCG), an agent-based model of irrigated agriculture rooted in principles of human cooperation and collective action, grounded on the largest dataset of cultural values in existence—The World Values Survey 6. Using this model, we simulate the long-term efficacy of groundwater conservation policies in three major aquifer systems currently facing unsustainable demands—the Punjab (India/Pakistan), the Central Valley (USA) and the Murray-Darling Basin (Australia). These simulations reveal 'tipping points' where collective attitudes towards groundwater conservation shift abruptly with changes in cultural values and enforcement provisions. A second study examines the temporal dynamics of social norms and advances our understanding of the factors that may accelerate transitions to regulatory compliance, such strategies used to enforce and promote compliance and the prevalence of role models. Finally, the foundations for participatory groundwater modelling are proposed, where scientists and stakeholders may interact with live computational experiments to test policy interventions and evaluate solutions.