Mike Bithell, Department of Geography

Currently we have a number of global models that represent large scale change of non-human systems. Atmosphere-ocean models have existed for many years, forest models have followed, and global model of the animal kingdom have now become available. Yet research into the modelling of people at global scale has been distinctly lacking, despite the critical role of society in driving environmental change, and the continuing globalization of social processes. Increased computational power and new methods for representing people at individual scale now make such models possible: these models will be needed to make sense of the flood of societal data that is beginning to be available, and to guide policy, since process-based understanding is needed to make progress. The size of data structure needed for building these models is formidable, as will be the amount of output they may be able to generate. Initialisation, parametrization and testing will require detailed long time series across multiple different types of dataset. However, the available data is fragmented, uneven in size and quality, variable in complexity and velocity, and we have little idea what is important, and what can be safely discarded. Multiple challenges will be present for storing, analysing, characterising uncertainty and presenting the outputs of the models of this type, not least in the issues of who will own, interpret and allow access to the information that may be generated. On the other hand we need to do better than the current pure-economic driving of policy at global scale. Multiple sets of richly detailed models of society that can give insight into the impact of global change on individual's lives, and the way collective behaviour is instrumental in constructing that change, may help us to do better.