Abstract: Effective management of infrastructural networks in the case of a crisis requires a prior analysis of the vulnerability of spatial networks and identification of critical locations where an interdiction would cause most damage and disruption. This paper presents a preliminary study into how a graph theoretic structural analysis could be used for this purpose. Centrality measures are combined with a dual graph modelling approach in order to identify critical locations in a spatial network. The results of a case study on a street network of a small area in the city of Helsinki indicate that 'betweenness' is the most promising centrality measure for this purpose. Other measures and properties of graphs are under consideration for eventually developing a risk model not only for one but for a group of co-located spatial networks.