Conceptualising and Implementing an Agent-Based Model of an Irrigation System

Abstract

The literature on irrigated agriculture is primarily concerned with irrigation techniques, irrigation water-use efficiency, and crop yields. How human and non-human agents co-shape(d) irrigation landscapes through their activities and how these actions impact long-term developments are less well studied. In this study, we aim to (1) explore interactions between human and non-human agents in an irrigation system; (2) model the realistic operation of an irrigation system in an agent-based model environment, and; (3) study how short-term irrigation management actions create long-term irrigation system patterns. An agent-based model (ABM) was used to build our Irrigation-Related Agent-Based Model (IRABM). We implemented various scenarios, combining different irrigation control methods (time versus water demand), different river discharges, varied gate capacities, and several water allocation strategies. These scenarios result in different yields, which we analyse on the levels of individual farmer, canal, and system. Demand control gives better yields under conditions of sufficient water availability, whereas time control copes better with water deficiency. As expected, barley (Hordeum vulgare, Poaceae) yields generally increase when irrigation time and/or river discharge increase. The effect of gate capacity is visible with yields not changing linearly with changing gate capacities, but showing threshold behaviour. With the findings and analysis, we conclude that IRABM provides a new perspective on modelling the human-water system, as non-human model agents can create the dynamics that realistic irrigation systems show as well. Moreover, this type of modelling approach has a large potential to be theoretically and empirically used to explore the interactions between irrigation-related agents and understand how these interactions create water and yields patterns. Furthermore, the developed user-interface model allows non-technical stakeholders to participate and play a role in modelling work.