URBAN GROWTH SIMULATION USING URBAN DYNAMICS AND CITYGML: A USE CASE FROM THE CITY OF MUNICH

Abstract. Urban dynamics modelling using system dynamic (SD) approaches aims to provide an understanding of the major internal forces within an urban area, such as population development. SD models provide valuable information for decision and policy making. Urban systems are strongly related to the urban space, which is well described by geospatial data. Therefore, the connection of SD and geospatial data is advantageous, both for feeding spatial information into SD models and for further spatial analyses and for visualizing the results of SD in geographic context. This paper describes a new approach to combine an SD model with a semantic 3D city model. Our approach shows that a bidirectional data exchange between semantic city models and SD models improves the predictions generated by the SD models. Furthermore, we show that automatic modification of the semantic city model by the output of the SD model allows for 3D visualization and further analysis of future scenarios.

Since semantic 3D city models and SD models have complex data structures, and since the models have evolved in very different domains, integrating the models is a complex task. In order to facilitate the integration process, we developed a conceptual model, which describes the data structures of the semantic city model and of the SD model as well as the bidirectional relations between the models using the concept of model weaving. The approach was tested using the SD tool Vensim and a CityGML data set from the city of Munich for an urban densification use case.