The Roman Aqueduct of Lisbon: A Utopia or a Viable Enterprise?

Abstract

The first references to the Olisipo (Roman Lisbon) aqueduct date from the 16th century and regard the Água Livre springs and the Belas roman dam. During the 17th and 18th centuries, whenever architects and engineers examined the way to supply Lisbon with water, they reported the existence of remains of an old conduit, part of which has lately been studied and presented in publications as the “Amadora roman aqueduct”. Despite such evidences, the lack of archeological proofs in Lisbon urban area has set back the discussion on the issue, so that some researchers consider the Olisipo aqueduct an utopia. With the support of known literary accounts, in particular the 17th. century description of the roman aqueduct remains by the Italian architect Leonardo Turriano, and of the functional interpretation of archeological roman structures related to water, tests to find out the aqueduct route were developed through an innovative geographic modelling approach, applying various criteria such as an the spatial definition of an equipotential average slope gradient, common in the roman world, aiming to create a viable route for the water to be conveyed to both termini of the ancient town –Porta de Santo André and São Roque – referred to by Leonardo Turriano. Through these tests, the authors have shown the theoretical viability of the Olisipo water supply by the means of an aqueduct connected with the Amadora roman conduit remains. The potential aqueduct route runs close to Palhavã area, which matches the fact that aqueduct structures in this area are referred to in the 1758 Lisbon Parish Reports (Memórias Paroquiais de Lisboa). However, as it does not correspond to the route described by Turriano, the authors have developed new tests taking into account this architect’s topographical references in order to see if the Belas water could reach Porta de Santo André and São Roque. The modelling approach was based on the application of different terrain interpolation methods, with variable detail topographic input data and variable output resolution scales. As a development of previous work, new surfaces of topographic conditioning were generated, namely following an optimal abstract terrain elevation potential and a new conditioned terrain based in historical references and available archaeological evidence. The novel method integrates through a multi scale/resolution approach, a raster optimal path algorithm and a corridor viability scenario. For validation of the methods, several statistical indicators were calculated for the logistical probability functions implemented. Results demonstrate very high levels of significance and inherent adherence from archaeological evidence and generated potential paths to terrain based spatial variables.