Geosimulation Modelling Lab

Geosimulation models help to understand processes in the earth system and interactions between the different components in this system, both natural and human. In addition, they can be used to project the future development of the system. Whereas early-age geosimulation models mainly simulated natural, physical systems, such as water flow, in the past decade models of human-environment interactions have become more widespread. Examples of focus areas in human-environment interaction modelling are: crowd behavior, land use change, flood mitigation strategies, ecosystem service dynamics, and traffic flows.

Building these models is a challenge, firstly because of the limited understanding of human behavior and secondly because this behavior tends to vary over space and time, while physical processes are more constant. At first, human-environment models were used for theory building, mimicking human behavior in non-spatial environments with synthetic data. But current human-environment models are full-fledged geosimulation models replicating real-world case studies, used for decision making. This requires models that are valid now and in the projected future, as management or policy decisions based on erroneous model projections can be costly and/or irreversible.

The overarching goal of this research lab is to better quantify the predictive value of human-environment geosimulation models. Lines of research include:

  • Quantifying and communicating uncertainties in geosimulation models;
  • Developing general methods and tools to calibrate and validate human-environment geosimulation models, borrowing techniques from the natural sciences domain;
  • Assessing how new developments, such as Big Data, volunteered geographic information, and the Internet of Things, can provide inputs for these methods;
  • Finding ways to detect systemic changes in the human-environment system and the drivers behind these changes.