Spatio-temporal processes are processes where something varies in space and time, and can range from environmental processes (weather, air pollution) to human behaviour. Modelling such processes helps in understanding this variation, to describe it, or to predict under new conditions (unobserved locations, future moments).
Several different model representations will be introduced for time series modelling, spatial modelling, and spatio-temporal modelling. These include stochastic models, models based on physical concepts expressed through partial differential equations, and agent-based models. Choosing and calibrating models and estimating parameters will be discussed in general, for the linear and non-linear case, and for one-dimensional and higher dimensional cases. Tools for the comparison and evaluation of models will be discussed and compared. Practical exercises are given with R, and with domain-specific implementations of transport models.
The course will be in English if one or more students do not master German.
The course capacity is 30 persons (room availability); to fill this capacity, geoinformatics students get preference over non-geoinformatics students.