The challenge of observing renaturation: feedback from the use of Land Use and Large-Scale Land Use in Île-de-France

FNAU (National Federation of Urban Planning Agencies) report 64

Jean Bénet, Damien Decelle, June 2025

Institut Paris Région (IAU), Agence pour l’Environnement et la Maîtrise de l’Energie (ADEME)

To download : fnau-64-institut_paris_region.pdf (190 KiB)

Observing artificialisation and its symmetry, renaturation, at the heart of implementing the ZAN objective

Achieving the ‘Zero Net Artificialisation’ objective relies on a cascade implementation from the national to the local level through urban planning documents, as opposed to a project-based approach that would require each operation impacting the land to compensate for the artificialisation generated. The corollary is the ability to monitor compliance with land artificialisation reduction targets, decade by decade, using observation tools. This monitoring is a regulatory requirement, which is reflected both in the assessment of the urban planning document – to be carried out for PLU(i) and SCoT plans, as well as regional plans – and in the triennial report on artificialisation monitoring. Any deviation between observed and planned land take is therefore likely to lead to a revision of the urban planning document in order to put the territory back on track. While the definition and tools for monitoring land take (and, initially, land consumption) have been the subject of much discussion, the issue of renaturation has been little addressed. There are several possible explanations for this: on the one hand, achieving ZAN relies primarily on efforts to reduce land use, i.e. to reduce artificialisation; on the other hand, renaturation is seen as a process symmetrical to artificialisation, and its monitoring follows the same principles. However, the recent nature of this concept and the still limited number of examples of operations mean that there is a lack of perspective on the monitoring of renaturation and raise questions:

· What are the current dynamics of renaturation, both quantitatively and qualitatively?

· Do existing tools allow this process to be measured satisfactorily? The first elements of comparison between the two land use monitoring databases in Île-de-France – Land Use Mode and Large-Scale Land Use – provide a preliminary answer to these questions. This work highlights a misleading renaturation from a quantitative point of view, and an unsatisfactory one from a qualitative point of view, emphasising the need for tools to evolve, but also the potentially limited contribution of renaturation to achieving Zero Net Artificialisation compared to current artificialisation dynamics.

A measurement of trompe l’oeil renaturation in Île-de-France based on the MOS and OCSGE

Renaturation in Île-de-France between 2017-2018 and 2021 ranges from 279 hectares (8% of the land consumption observed by the MOS) to 666 hectares (19% of the artificialisation observed by the OCSGE). While these differences in figures can be attributed in part to the subtle difference in the definition of renaturation between the ‘land consumption’ approach and the ‘artificialisation’ approach, and to methodological differences between the two tools, the MOS and the OCSGE have common limitations when it comes to observing renaturation. In fact, nearly 60% of the renaturation identified by the Land Use Mode is originally linked to temporary land use, whether it be the restoration of construction sites, open-air storage (often linked to agricultural activities), or waste storage facilities. The counting of the renaturation of these facilities once they have been filled raises questions, as they are mainly developed within ENAFs and their restoration is a requirement from the moment they open: their renaturation therefore offsets their own creation, mirroring the special status of material extraction facilities in the counting of artificialisation. For the OCSGE, nearly 37% of the renaturation observed takes place in areas considered to be ‘transition zones’. The majority of these transition zones are considered to be renatured because they fall into the ‘unused’ category ¹ category, which, all flows combined, represents 38% of the renaturation observed. More than true renaturation, this is a nomenclature bias that equates unused areas with low vegetation to non-artificialised areas, even though this category often reflects a temporary lack of information on the use of an area: this may be, for example, land developed in a ZAC (joint development zone) where construction work has not yet begun, a ‘renatured’ space that will subsequently be ‘re-artificialised’. Nearly a third of the renaturation observed in the MOS concerns urban open spaces (parks, gardens, abandoned infrastructure, vacant lots), reflecting spontaneous fallow dynamics, but also the ambiguity of the classification of certain open spaces with low vegetation (e.g. airport and aerodrome runways, vegetated brownfield sites in urban areas). While these observations in many cases demonstrate the increasing naturalness of these spaces, it is difficult to define in a consistent manner the time required or the tipping point for such renaturation to be effective, or to ensure the improved functionality of the soils concerned. 38% of the renaturation observed by the OCSGE in Île-de-France concerns areas that can be considered as ‘open land’ in origin, whether this renaturation involves an increase in vegetation cover (transition from low to high vegetation within urbanised areas: 18% of the renaturation observed), a change in use (with the disproportionate role of ‘no use’ for areas with low vegetation), or both. This observation reflects the same difficulties in classifying open spaces with low vegetation as the MOS, but accentuated by the overlap between cover and use in the artificialisation nomenclature. Overall, whether in the MOS or the OCSGE, the renaturation actions observed involving the demolition of a building or soil de-impermeabilisation are extremely rare and quantitatively negligible at the regional level.

A path to be found between the necessary reliability of tools and a project-based approach

While a reading of the raw renaturation figures from the MOS and OCSGE suggests a potentially significant quantitative contribution of renaturation to the achievement of ZAN in Île-de-France, a more careful analysis of these databases paints a different picture: apparent net artificialisation is more a measure of artificialisation stripped of the ‘noise’ of temporary land uses than a real difference between permanent artificialisation and renaturation flows. Most of the remaining renaturation observed concerns land that was originally open ground, through a process of abandonment or ‘nomenclature effects’ for which improved functionality is not guaranteed. These initial observations point to several ways out of the rut:

· a need to improve the reliability of tools, in terms of the ‘yo-yo’ effects of classifying temporary land uses, but also in terms of classifying open spaces with low vegetation (urban wasteland, abandoned land, vacant lots);

· the dissemination of precautions for the use of databases for monitoring changes in land use, which must remain tools and require analysis and corrections to reflect the reality of the territory;

· a certain flexibility in the application of ZAN, which must be based on the project of land use restraint and soil restoration in the territory rather than on observation tools.

Sources