Development of an integrated modelling framework for the assessment of river management policies

Abstract

Many rivers in developed countries are entering a state where demands on the system are met with diminishing reliability and there are significant negative external effects related to the use of the resource. Careful management is required to ensure that irreversible damage to the economic, environmental and social systems reliant upon these resources are minimised. In an effort to better manage these rivers, water resource managers are no longer turning only to ‘hard’ infrastructure or technology based approaches in isolation. Other, ‘softer’ approaches focusing on changing attitudes, behaviours and the perception of the value of water are being considered by managers in conjunction with traditional policies of using infrastructure to alter natural processes. In order to determine which combination of policies is most likely to meet management objectives, river managers require a model capable of representing the relevant system dynamics. Hard management policies have traditionally been modelled using simulation models, whereas policies aimed at altering the decision processes of the actors within the system have commonly been comparatively static in nature, and often use a whole-of-system optimisation approach. These two approaches to assessing different types of management policies do not integrate particularly well. This paper outlines a proposed conceptual model for an exploited river system and a framework with which to implement the proposed model that will allow for analysis of a combination of both hard and soft management policies. The proposed conceptual model represents an exploited river system at three levels, those of the regulator, the water user, and the environment. A river manager represents the regulator level of the system. The river manager has both environmental and economic objectives. The set of management policies available to the manager totry and meet these objectives is set by the modeller. Sharing information between the river manager and irrigator allows management policies aimed at altering water use behaviour to be tested along with more traditional infrastructure-based policies implemented directly by the river manager. By representing irrigators at an individual level, the proposed conceptual model allows for overarching assumptions of water user behaviour used in the past to be replaced with more realistic behavioural models. Using multiple instances of the irrigator agent, and an explicit spatial representation of environmental processes, facilitates heterogeneity of water user behaviours within the system. In order to implement the proposed conceptual model, a model framework has been developed using an agent based architecture. The framework is developed using Cormas, an agent based modelling platform using the SmallTalk object oriented programming language. The basic Cormas classes have been further developed to represent the agents in the conceptual model and their behavioural processes. The framework allows the modeller to specify the manner in which the agents interact so that the proposed conceptual model can be customised. This allows the modeller to identify scenarios where certain river management policies are likely to be more or less effective.