Overcrowded Dutch natural areas: an agent-based exploration of visitor conflicts

Abstract

Nature-based recreation and tourism brings forward a wide variety of benefits. However, high recreational pressure can also result in various negative impacts, both on nature itself, as well as on the experiences of those visiting natural areas. Visitors can start to experience crowding, which can reduce the quality of their nature visiting experience, and visitor conflicts can be experienced. In order to facilitate sustainable nature-based recreation and tourism, organisation involved in managing natural areas therefore engage in visitor management, which aims at ensuring the quality of the visitor experience and nature itself. Visitor management is becoming increasingly important to natural areas around the world, as visitor numbers are increasing. This is also the case in the Netherlands, where an historic increase in the number of visits to natural areas has been reported since the start of the COVID19 pandemic. For visitor management to be effective, information about visitor behaviour is required. For many natural areas, it is the case that little information about visitor behaviour is known, however. Agent-based modelling allows for the estimation of visitor behaviour that is hard to measure using other methods. Many agent-based models, however, rely on the assumption that human decision-making is rational, and are not based on existing models of human psychology. The Consumat approach to simulation modelling is identified as an exception but has not been applied to the context of visitor behaviour yet. In this thesis, it is investigated how an agent-based model, based on the Consumat framework, can be used to simulate and explore the spatiotemporal behaviour of visitors in natural areas, in order to gain insights into visitor conflicts. The Amerongen Forest in the Netherlands has been selected as the research area for this thesis, and GAMA has been used to develop the model. The behaviour of hikers, cyclists, mountain bikers, and horse riders was simulated, and encounters between these different types of visitors were recorded. The model output suggests that many encounters in the research area occur near the main entrances to the area, along frequently used marked hiking trails, and near landmarks situated on these trails. Furthermore, encounters between visitors of different activity types are suggested to occur in places where there is an overlap or intersection between the networks of various activity types. The inclusion of the concepts of familiarity and satisfaction allows for the further exploration of visitor behaviour. A sensitivity analysis was performed, and scenarios were developed. Furthermore, through expert interviews, it was determined that, generally, the outcomes of the model can be regarded as plausible. However, further research, and in particular, empirical research into the behaviour of visitors in natural areas, is required in order to further calibrate and statistically validate the model.