The urban underground has historically been developed as a series of sporadic responses to surface land shortage or to hide unsightly urban activities (Parriaux et al 2004; 2010). The Deep City project at the École polytechnique fédérale de Lausanne in Switzerland is questioning this piecemeal approach through the development of a methodology to qualify and map the spatial potential of the city subsurface. An important part of this methodology is identifying and critically reflecting upon the spatial configurations that characterize complex urban underground ensembles (like the Montreal underground). Although past research has explored the relationship between form and pedestrian movement in the underground (Zacharias 2000, 2011), the topological relationships of the subsurface to neighbouring surface spaces has not yet been investigated. Furthermore, attempts at characterizing the morphology of underground spaces has been limited to generalizations around very few types (Carmody & Sterling 1993), while ignoring the particularities of the spatial and functional contexts in which such projects are situated. This communication will present a methodology to investigate urban underground spaces within their three-dimensional spatial context drawing from graph theory (Hillier 2007; Sevtsuk 2010) and urban morphology’s approach to investigating city form (Cannigia & Maffei 2001; Oliveira 2013). The objective of the topological analysis is to be able to aid spatial configuration decisions made both in the conceptual phase of underground space design as well as to diagnose and test alternatives to problematic existing spatial configurations of underground spaces and their topological relationship to surface spaces. The morphological analysis seeks to identify patterns that could serve to guide form-based urban legal instruments. The conclusion will present preliminary results of this method taking Montreal’s underground as a case study.