1Digital technological solutions, supposedly turnkey and always effective, are regularly offered to cities. They can however attract many criticisms: for example, if they are "smart cities" - which work with difficulty when they clash with the diversity of urban forms and functions (Caruso, Pumain & Thomas, 2023); or if it is the concept of "mobility-as-a-service" that induces unanticipated effects on accessibility equity (Pangbourne et al., 2020); or "city dashboards" that ultimately offer only a very limited view of urban processes and raise ethical issues (Kitchin & McCardle, 2017). In this vein, it is the concept of Digital Twin that becomes the most fashionable for cities and territories. Would this new instrument be more convincing?

2In the strict sense, the digital twin of a system is a computerized representation that simulates its internal processes and interacts with it in real time by two-way coupling (Batty, 2018). When it applies to the city, the practical meaning of the concept remains more limited, since, for example, urban 3D models are considered twins by the communities of researchers and practitioners who develop them (Batty, 2024a). A better consideration of the complexity of urban systems by simulation models within digital twins remains to be developed (Caldarelli et al., 2023), so that they are part of the long legacy of urban modeling (Batty, 2024b). A systematic review of the literature shows that existing implementations remain mainly prototypes, strongly linked to financing opportunities. They only represent or simulate socio-technical systems at very short time scales (urban climate, buildings, roads, etc.), with a limited spatial scope (localized infrastructure, neighborhood, medium city) for those in operation (Ferré-Bigorra et al., 2022). For example, White et al., 2021 provide residents of the dock district in Dublin with a 3D model in the Unity platform, to test modifications to buildings and green spaces, to report problems in real public space, to experiment with flood scenarios, as well as to understand the impact of street development on pedestrian mobility.

3Still, the concept remains very rich and deserves to be explored in all its diversity. We thus illustrate in figure 1 the variety of approaches that can be understood in the broad sense as urban digital twins, in terms of spatial and temporal scales and functionalities, around the same thematic axis (building and population): at the building scale, high-resolution 3D models are used for heritage conservation (IGN, 2024). At the plot level, the simulation of the buildable envelopes authorized by the Local Urban Planning Plan with the SimPLU3D model (Brasebin et al., 2017) makes it possible to assist urban planning. At the city level, (Bucher et al., 2024) shows that the production of building evolution data requires data enrichment within an integration platform that acts as a twin of the territory. Finally, the simulation of the impact of large national transport projects in the United Kingdom in terms of population and job relocation by the QUANT spatial interaction model is presented by (Batty & Milton, 2021) as a very small-scale twin.

Figure 1. Diversity of potential Urban Digital Twins on a common theme.

Agrandir Original (jpeg, 446k)

(a) 3D model of Périgueux Cathedral, source: (IGN, 2024); (b) simulation of buildable envelopes, source: (SimPLU3D, 2024); (c) spatial interactions on a national scale to simulate the balance between assets, jobs, and accessibility, source: (QUANT, 2021); (d) consideration of user returns on spatial data (Geospatial User Feedback) in the case of building data for the study of peri-urban densification, source: (Bucher et al., 2024).

4A digital twin project in France, whose main component will be a high-resolution 3D model of the entire national territory, is carried out by the National Geographic Institute, jointly with INRIA and CEREMA (IGN, 2024), with the ambition of being a foundation for the design of sustainable policies in urban and territorial planning. In view of the challenges mobilized by the magnitude of such a project and its potential implications for the territories, and more generally the potential impacts of the use of twins on modeled social systems, a critical point of view is important. (Malleson et al., 2024) have recently highlighted very diverse pitfalls for these approaches, both on ethical and technical issues. Here, we are offering to do the same, from the point of view of a journal in the humanities and social sciences.

5Thus, at least four major issues must raise points of vigilance around the deployment of these technologies:

1. The role of private actors, who could, through such projects, secure control over urban data infrastructures. In this context, an absolute open science policy and complete transparency are necessary, in order to ensure the openness of data (as far as possible with regard to ethical issues) as well as the use of free and open tools (platform source codes, algorithms and developed models). In the case of France's twin project, it would be an opportunity to deploy an open public service of territorial simulation, even if private companies were responsible for its implementation.

2. The introduction of disciplinary biases in the definition of “problems to be resolved”. The disciplines involved in the current state of literature being closer to the engineering sciences, the humanities and social sciences could be marginalized. However, these are central to defining a large majority of urban and territorial issues and taking into account ethical issues. Neglecting their role in the construction of territorial models would not be responsible, even dangerous with regard to potentially negative social impacts. A strong interdisciplinarity is therefore required to choose what the functions of the digital twins will be.

3. Systematic validation, quantification of uncertainty and exploration of simulation models involved in twins is a step too often left out. However, ensuring that models do what they say is a necessary prerequisite before any decision-making.

4. Finally, the consideration of long time is a very little considered aspect in the existing examples of territorial twins. Understanding territorial and urban dynamics on extended spatial and temporal scales, through the use of appropriate urban models and theories (Wan et al., 2024), will be crucial to ensure territorial sustainability.

6The journal Cybergeo offers an integrative and inclusive space for scientific exchange, which responds positively to these four challenges: a fervent defender of Open Science since its creation in 1996, the journal supports initiatives such as CybergeoNetworks2 (which aims to build tools of reflexivity and openness for contributors and readers: Kosmopoulos et al., 2023) and encourages submissions of data papers; it is a recognized place of publication that welcomes any contribution in geography, without disciplinary judgment, open to all spatialized SHS, and enhancing interdisciplinarity; it encourages employment via the model papers section of tools for validating simulation models, such as the OpenMOLE platform (Reuillon et al., 2013); finally, Cybergeo has historically been the place of publication of major works on the dynamics of city systems over the long time (Bretagnolle et al., 2006). We welcome any contribution in the direction of a good use of digital twins as true instruments of territorial mediations. The idea is to appropriate this concept before it is overused by technocratic excesses and so that it finds a real scientific and societal utility, beyond a fashion phenomenon linked to requests (or offers) for financing.

Juste Raimbault, October 2024