Mobility in ancient times: Combining land and water costs Pedro Trapero Fern andez C adiz University, Spain ARTICLE INFO Keywords: GIS Network analysis Spatial analysis Roman transport Horizontal factor Guadalquivir river ABSTRACT Mobility in Roman Times can be modelled using GIS tools. These resources allow us to understand how com- munications could be done. In this paper, we debate about a better way to use these models in cases where we have a navigable river in our studied territory. Normally these water spaces are considered as limiters of land movement, because of the real nature of Least Cost Path tools, that will always consider them as natural paths. We argue that in roman mentality the rivers are the best areas for movement and in our current development we could not integrate them with land ones. The main problem is the principal factor in river mobility, the current upstream and downstream. We present a tool in ArcGIS to calculate the Horizontal Factor of the river's current. This allows us to test the integration and results of a combined land and river analysis in different scenarios, like the connexion between centre and node, like a city to the villa. Finally, we argue about the need of this kind of integration to better understand mobility with the ancient mentality, allowing us to consider other ways of cost, not only in terms of time but also in the increased cost value of product transportation. 1. Introduction In this paper, we propose a method to include cost analysis in fluvial areas alongside traditional land movement analysis. In addition, we debate the need for better models for understanding ancient forms of communication. Our methods integrate water and land costs as a single process in GIS to create a cost distance and backlink raster for Least-Cost Path (LCP) analysis (Brughmans, 2013). These tools are well known in mobility studies of ancient times (Lo Cascio and Tacoma, 2017). In the movement of people and goods we will always need to be limit by physical conditions like gradients, generating natural paths in inland transport or currents and wind in maritime ones (Malmberg, 2015). Due to the incremented cost of trade and the interchange of goods, the best routes would always have been navigable, either by river or by sea. To move an indeterminate number of people or objects in Roman times, the chosen medium would not have been the fastest in terms of time, but the low value costs in terms of economic costs (di Salvo, 1992). Our model tries to represent this reality and for this purpose we need to modify the tools we use in GIS. The best example is when we need to somehow restrict the routes generated because of the existence of rivers and streams. An ordinary LCP would use the centre of the water basin, without considering the difficulty of walking or crossing flooded terrain. This is why in many land cost-distance analyses, rivers and streams are considered as boundaries or restrictions to passing (Abu Izzeddin, 2014). Some recent high-quality research has considered river and sea transport separately (Melles et al., 2011; Gustas and Kisha, 2017; Verbrugghe, 2017; Milheira, 2019); the same is the case of modelling sea communi- cations alone (Safadi, 2016; Jarriel, 2018; Gambash, 2017). In this paper, we ask whether this is the best way of modelling rivers in cost analysis? It is contradictory to restrain models in such areas as if they were not relevant paths. Nevertheless, rivers were like highways as means of communication in ancient times. To understand ancient history, we need to avoid being restricted by a tool and instead use it to represent the ancient mentality about mobility, economy and landscape (Grau, 2006; Lag ostena, 2016; Ruiz Gil et al., 2019). We propose a clear and verifiable model to integrate both costs into one analysis. We think that this issue has not already been solved owing to one problem. River transport has many factors and conditions, the most important of which is movement direction into the river, because the velocity significantly changes upstream and downstream (Werther and Kr€ oger, 2017). For this reason, we need to know the direction of the river and consider at least a horizontal factor in the creation of the cost distance raster. There are other conditions taken into account, gradient, currents, depths and winds, but we consider that in a navigable river the main one would be the horizontal factor. To solve this problem, we develop a tool in Model Builder for ArcGIS and share it in this paper for other researchers. We present in this paper a model to generate the angle of water movement in a river. In addition, we seek to demonstrate how and why land and river cost analyses should be integrated. We argue that it is E-mail address: pedro.trapero@uca.es. Contents lists available at ScienceDirect Digital Applications in Archaeology and Cultural Heritage journal homepage: www.elsevier.com/locate/daach https://doi.org/10.1016/j.daach.2021.e00192 Received 5 December 2020; Received in revised form 19 April 2021; Accepted 24 May 2021 2212-0548/© 2021 Elsevier Ltd. All rights reserved. Digital Applications in Archaeology and Cultural Heritage 22 (2021) e00192