Advanced Training Course on Modelling Land-Use Change (Amsterdam, 26-27 November 2018)
After two days of intense work the goals were accomplished, the fear of models was gone and our understanding of land use models notably increased, up to the point where some of the PhD fellows realized they knew a lot less than they thought. Above all, it became clear that the uncertainties inherent in models are a strong motivation to continue using them for research to improve their practice, rather than a reason to avoid them.
The Training Course started off with an introductory lecture about the pros and cons of land use models given by Professor Peter Verburg, one of the COUPLED supervisors, where the importance of understanding land system transitions was explained. The presentation gave just enough theoretical information for the exciting practical session that followed to be of use – even to those with no modelling background.
Using the exercise ‘Your land’ we analysed the trade-offs of land use choices. The lesson: Different configurations deliver completely different outcomes of ecosystem services! Finding an adequate solution for our desired functional landscape was not so easy, helping us to realise the challenges policy-makers face when managing their territories. Interestingly, there was more than one type of solution, as is also often the case in “real world” scenarios.
After that session, when exploring the DynaCLUE model, we were surprised by how small changes to various economic and political aspects can completely change the land use patterns given by the model depending on the local land characteristics. It helped showcase modelling as a useful tool to understand socio-ecological systems.
Professor Peter Verburg presenting DynaCLUE
In the following session, Koen Tieskens presented the utility of system dynamic models and how probability functions and Bayesian networks can be used to understand the functioning of telecoupled systems. He also introduced us to the fascinating world of agent-based models and how these can be used to study systems that are highly influenced by agents’ behaviour. Even for the PhD-students least familiar with this method, the communication advantages of these models were obvious. We even dared to practice some coding in the software NetLogo. Without any doubt, the lecturers were positively surprised by the bravery of ESRs to experience with tools that are completely different from the ones they are familiar with.
In another session, Professor Nynke Schulp, one of the COUPLED supervisors, described the impacts of land use change on ecosystem services. Thanks to her insightful presentation, we understood that the spatial configuration of a landscape could matter even more than the simple area occupied by a given land use. It became very clear to us that measuring ecosystem services is an almost impossible task, so modelling is a well-suited tool to help quantify them. Moreover, she explained the utility of scenario analysis to study potential alternative futures and to explore management options.
And for the ones still not convinced about the usefulness of models, e.g. due to uncertainties, Professor Jasper van Vliet gave an overview of sensitivity analysis and the validation exercises used to reduce uncertainties. No more excuses for avoiding using models, now that the tools to improve them are on the table!
Effective inter-disciplinary collaborations are expected from these fruitful days. COUPLED PhD fellows with a focus on empirical studies saw the opportunity to work with other fellows studying larger-scale phenomena and vice versa. Undoubtedly, the number of research ideas has increased as the number of bikes in Amsterdam city.
Gratitude goes to Peter Verburg and Nynke Schulp at the Institute for Environmental Studies, VU University Amsterdam.