Advanced thematic mapping: GIS/Neural networks application for tracking isoseismic lines

Possibility of making queries on a spatial database, in order to obtain a decisional support is one of the most interesting features of GIS systems. This is possible through the expression of information which are implicit into database and useful for Geoprocessing operations in order to make a sort of data clustering. However, it is not sufficient to represent a priori non-modeling interactions, even if they are present into the informative layers. Case study presented in this paper just concerns this category, taking into account tracking of isoseismic lines on a well-known geographical area. It is very useful in order to generate an affordable map for seismic risk. Proposed procedure, exploiting Neural Networks, can retrieve information about isoseismic lines propagation, starting from information related to examined territory, hypocenter of considered earthquakes, and seismic intensity calculated by standard procedures. Preliminary results we obtained have been used in a GIS software in order to create an Artificial Intelligence informative layer (called OverlayAI). Experimentation carried out shows a preliminary nature and needs further tests and refinement; however, it illustrates useful results to realize an operative plan based on perception of seismic risk in a defined territory.