Artificialization of soils: state of play of a complex challenge - Soil artificialisation: a concept in motion

Rémi Guidoum, June 2022

Fondation pour la Nature et l’Homme (FNH)

After 30 years of action, the Fondation pour la Nature et l’Homme is giving a new ambition to its Think Tank activity, which was born about ten years ago. Starting with the question: « What would a government determined to make the ecological transition do?", the Foundation’s Think Tank focuses on the « blind spots » of public policies by untangling the subjects that are bogged down or even unexplored. Its ambition: to create the conditions for a next five-year period of social and ecological transformation.

Based on a complex notion, recently redefined by the « Climate and Resilience » law and fuelled by historically disparate data, the debates relating to soil artificialisation are often difficult to access for the uninitiated, and sometimes lead to confusion. Despite its new definition, the notion of soil artificialisation remains trapped in a binary vision between artificialised and non-artificialized soil.

This dichotomy is maintained by monitoring tools that focus on the surface of the land, without characterising it in depth or measuring the ecological impacts actually induced on the ground. However, not all areas classified as artificial are equal from an ecological point of view, just as land uses considered as non-artificial are not necessarily favourable to biodiversity. The policy of sober land use must therefore be equipped with tools that allow for a detailed description of situations, in order to be able to reduce the consumption of space while favouring biodiversity throughout the territory. In addition to presenting the definitions and impacts, this contribution provides a summary of the distribution of built-up land in metropolitan France (geography and uses)

To download : tt-contribution-sols.pdf (7.7 MiB), artificialisation-des-sols-synthese-en-francais-1.pdf (4.7 MiB), egs_2021_28_calvaruso_03-29.pdf (2.6 MiB), primeur326.pdf (2.7 MiB)

Agricultural land with ecological functions

The first difficulty in dealing with the subject of soil artificialisation is probably that of its definition. As noted by the INRA-IFSTTAR collective scientific expertise in December 20171 , ‘soil artificialisation is a recent concept, initially corresponding to a concern to quantify the loss of land available for agricultural use through land use changes’. The perspective was then broadened to include changes taking place not only to the detriment of agricultural areas, but also to the detriment of natural and forest areas: this gave rise to the notion of ‘natural, agricultural or forest area’, which brings together concerns about preserving agricultural potential, renewable resources and habitats for living organisms. Thus, artificialised land is defined in a negative way as land that is no longer a natural, agricultural or forestry area due to a change in occupation or use. The term « consumption of ANFE » is used to refer to changes in land use or occupation due to urbanisation. It is the monitoring of this indicator that remains, to this day, the tool for measuring the phenomenon of land artificialisation.

Following the Biodiversity Plan published by the Government in 2018, the law of 22 August 2021, known as the « Climate and Resilience » law, clarified the definition of soil artificialisation in law and introduced the objectives of « Zero Net Artificialisation » (ZAN) by 2050, as well as halving the consumption of ENAF by 2031, compared to the previous ten years. Soil artificialisation is thus defined in the law as « the lasting alteration of all or part of the ecological functions of a soil, in particular its biological, hydric and climatic functions, as well as its agronomic potential by its occupation or use ».

A new definition that remains trapped in a binary vision

In principle, this new definition introduces a significant change of perspective, since it no longer focuses primarily on the occupation or use of the land, but on the alteration of its ecological functions and agronomic potential. Nevertheless, the ‘Climate and Resilience’ law did not go all the way in this definition, i.e. to move away from the binary vision of artificialisation to propose a gradual vision of the state of degradation (or, to use the terms of the law, alteration) of the ecological functions of soils. In fact, the realities of land use and occupation are very diverse, with equally diverse consequences for their ecological functions. These consequences also depend on the characteristics of each soil and its position in the landscape grid. This complexity is poorly reflected by a binary distinction between land considered as artificial and land considered as non-artificial. The « Climate and Resilience » law specifies that from 2031 onwards this distinction between artificial and non-artificialized soils will be made in planning and urban development documents according to the following criteria:

Thus, it can be seen that despite the desire to focus on the alteration of ecological functions, the distinction is ultimately made on the basis of the occupation and use of the land - on the basis of an a priori association between occupation, use and the level of alteration of ecological functions - and not on the consequences actually observed as a result of these occupations and uses. The ecological functionality of soils is thus prejudged on the basis of the presence or absence of certain elements: impermeable cover, buildings, vegetated surface, etc. While this « proxy » approach may be appropriate for certain situations, it is blind to the composition of the soil. Characterising soils, and therefore being able to assess their quality in relation to the use one wishes to make of them (urban, agricultural, etc.), is however particularly important in order to prioritise and plan developments intelligently. For example, having a soil quality map makes it possible to concentrate certain essential urban developments on areas of lesser agronomic or ecological quality, and thus not to sacrifice the best agricultural land and the richest natural habitats, for lack of sufficient knowledge of the terrain. Conversely, such measures would make it possible to observe the state of alteration of the ecological functions of certain agricultural and forest soils, even though they are considered not to be artificial. Consequently, it is necessary to supplement the monitoring of artificialisation with efforts to characterise soils, such as those developed in the framework of the MUSE project « integrating the multifunctionality of soils in urban planning documents », led by Cerema 2.

Beyond the fact that they are still based on land use and occupancy, the criteria listed in the law to distinguish between artificial and non-artificial soils are relatively general. The law therefore refers the publication of a detailed nomenclature to a decree in the Council of State, which will be used to monitor artificialisation from 2031. The artificialisation of land will then no longer be monitored through the prism of ENAF consumption, as is currently the case, but through this specific nomenclature, which is intended to serve as a distribution key between surfaces identified by photo-interpretation of satellite images. This nomenclature was made public by a decree of 29 April 20223 , and will be the subject of a dedicated analysis in a forthcoming contribution by the FNH think tank.

Data sources, sources of confusion

Debates on the artificialisation of land are sometimes marked by a certain amount of confusion about the situation, with figures that can vary from one to two. This situation is due in particular to the fact that several sources of data exist concerning the artificialisation of land, which use different methods.

The most recent data source, available since March 2020 4 , is that developed by Cerema via the « land files » method. These data are now used by the State services for the annual monitoring of artificialisation and therefore in the context of the objectives of reducing land consumption established by the ‘Climate and Resilience’ law. The data from the land files are made available to the public via a national portal dedicated to land artificialisation. a(]

Prior to the use of these « land files », the measurement of the consumption of ENAF was historically carried out using two other methods: the Corine Land Cover tool and the Teruti-Lucas survey. These three methods produce clearly different results concerning the annual flows of land artificialisation: these differences have been a source of confusion in the debates concerning the rate of land artificialisation in France.

The European CORINE Land Cover (CLC) programme, which stands for « Coordination of Information on the Environment - Land Cover », was initiated in 1985 and is managed by the European Environment Agency5. It is a biophysical inventory of land use and its evolution, carried out by interpreting satellite images. The images produced by CLC make it possible to map homogeneous units of at least 25 ha with a minimum width of 100 metres, and to monitor changes of at least 5 ha. Five vintages of the CLC inventory have been produced to date: 1990, 2000, 2006, 2012 and 2018. The CLC inventory is produced in 39 European countries as part of the European Union’s Copernicus 6 Earth observation programme. For France, the services of the Ministry of Ecological Transition (MTE) and, since 2018, of the National Institute of Geographic and Forestry Information (IGN) produce the CLC inventory data.

The Teruti survey, for « Land Use », is an annual survey conducted by the services of the Ministry of Agriculture and historically aimed at monitoring changes in the occupation and use of agricultural land. The first survey of agricultural land - based on cadastral maps - was launched in 1946. In 1962, the use of aerial photographs was introduced in order to update the cadastral plans. The survey was then based on a combination of aerial photographs and field surveys of a sample of plots7. In 1982, the Teruti survey extended the sample used from agricultural land to the entire national territory. This survey is based on the observation of a sample of representative points, the results of which are statistically extrapolated to the departmental, regional and national scales 8. In 2005, this method was adapted to that of the European Lucas (Land Use/Cover Area frame statistical Survey), giving rise to the Teruti-Lucas survey. The resolution of a Teruti-Lucas point varies from one territory to another: a point represents approximately 94 ha in Paris and the inner suburbs, but 178 ha in the other metropolitan departments (between 2006 and 2014), and 40 ha in Guadeloupe, Martinique and Réunion (France Stratégie, 2019, op cit.).

The term ‘land files’ refers to a database produced by Cerema from tax data, derived from the land tax 9. These data are grouped together in the information system of the General Directorate of Public Finance, called « MAJIC » for « updating cadastral information ». The data thus produced by Cerema has made it possible to monitor the annual flow of artificial land since 2009. The principle of this method is to monitor changes in land use by analysing the changes recorded in property tax returns. This method allows for a fairly fine resolution because each parcel is subdivided into one or more « tax subdivisions » that must be declared by the owner in one of the 13 categories proposed in the nomenclature. This method is not only declarative and several control mechanisms are in place to ensure the reliability of the data. The main limitation of this method lies in the fact that it can only process cadastral parcels. No data is therefore collected on the non-cadastral domain, which includes all public roads (streets, squares, national and departmental roads, communal roads and rural paths). The non-cadastral domain represents approximately 4% of the territory of mainland France. Cerema (2019) nevertheless highlights several elements that limit the scope of this bias: on the one hand, some elements of the public domain ‘including large-scale elements (peripheral national roads, certain motorways, etc.) are still registered’; on the other hand, artificialisation most often occurs before the parcel is returned to the unregistered domain (a road is created, artificialising a parcel, and then is returned to the unregistered public domain).


Far from being a binary phenomenon, land artificialisation manifests itself in various ways, the consequences of which depend on the nature of the activities implemented and the initial context. It therefore needs to be monitored with tools adapted to this complexity, allowing the study of soils not only at the surface but also at depth, and considering the ecological impacts at the landscape level.

Beyond the necessary reduction in the consumption of space, the policy of land sobriety is a global project, which must integrate at all levels the issues of biodiversity and social justice. In this perspective, the question of the drivers of land artificialisation is central, in order to be able to treat not only the symptoms but also the causes of the phenomenon.

Thus, the FNH Think Tank will soon propose a synthesis of the main drivers of land artificialisation in France, before launching a reflection on possible responses in terms of public policies. This work will seek to propose measures that contribute to achieving the objectives of reducing the consumption of space, without increasing inequalities or harming biodiversity. For example, it will be a matter of thinking about public action capable of reconciling land sobriety, access to housing and services, and the promotion of biodiversity in all areas.

  • 1 INRA - IFSTTAR, Résumé de l’expertise scientifique collective, Sols artificialisés et processus d’artificialisation des sols : déterminants, impacts et leviers d’action, Décembre 2017

  • 2 Déterminer les fonctions que remplissent les sols et la multifonctionnalté des sols (Cerema)

  • 3 Décret n° 2022-763 du 29 avril 2022 relatif à la nomenclature de l’artificialisation des sols pour la fixation et le suivi des objectifs dans les documents de planification et d’urbanisme | Légifrance

  • 4 Cerema, L’artificialisation et ses déterminants d’après les fichiers fonciers, avril 2020

  • 5 CORINE Land Cover | Données et études statistiques

  • 6 À propos de Copernicus -

  • 7 France Stratégie, Objectif “Zéro artificialisation nette” : quels leviers pour protéger les sols ? Juillet 2019

  • 8 TERUTI LUCAS | Fiche donnée | Portail de l’artificialisation

  • 9 Cerema, Mesure de l’artificialisation à l’aide des Fichiers fonciers : définition, limites et comparaison avec d’autres sources, juin 2019