Multi-Scale Aspects in the Management of Geologically Defined Geometries

The interactive geological modelling and the geometric-kinetic modelling of geological objects need an efficient database support. In this chapter we present multi-scale aspects in the management of geologically defined geometries. From the computer science side, object-oriented data modelling provides a framework for the embedding of multi-scale information by enabling geoscientists to model different “semantic scales” of the data. In particular, the concepts of inheritance and aggregation implicitly deal with multi-scale aspects. According to these concepts, the database-backed software system GeoStore has been developed. It handles geological data in two different scales: the large-scale data of the Lower Rhine Basin and the small-scale data of the Bergheim open pit mine. We describe the database models for both scales as well as the steps undertaken to combine these models into one view. Furthermore, there are interesting connections between interactive multi-scale modelling and geo-database query processing techniques. In this context, we have developed the “multi-level R*-tree”, an index structure allowing “perspective region queries.” Multi-scale landform data modelling and management imposes manifold requirements. We discuss some of them and conclude that they can be met by a hierarchical data type: a directed acyclic graph structure which, if implemented in an object-oriented database system, is readily extensible. Another important aspect is the interoperability between modelling and data management tools. We describe the tasks to be carried through for accessing and recycling data in different scales contained in heterogeneous databases. In particular, we present the eX-tensible Database Adapter (XDA) which performs network-transparent integration of client-side GIS tools and object-oriented spatial database systems.