Between 2013 and 2021, the edges of 500 French agricultural plots changed their plant cover. On these narrow margins that border cultivated fields, competitive and ruderal species retreated in favor of species more tolerant to stress, heat, and aridity, driven by a temperature increase of 1.2 °C and a 14% drop in soil moisture. The ANSES-INRAE analyses published in 2025 draw on meteorological data from Météo-France in particular. The problem: some public subsidies that pay for the upkeep of these margins are calibrated to fixed geographic zones, defined for a given programming period, while plant communities are shifting at a pace that administrative reviews struggle to follow.

The essentials

  • The studied margins saw the advancement of species with a stress-tolerance strategy at the expense of competitive and ruderal species: agricultural life is moving, public payments stay in place.
  • On 500 monitored plots between 2013 and 2021, temperature rose by 1.2 °C and soil moisture fell by 14% on average (ANSES-INRAE, 2025).
  • The Common Agricultural Policy finances agro-environmental practices by plot and by zone: there is no automatic mechanism to adjust support when species shift northward or to higher altitudes.
  • Possible approaches exist—environmental service payments adapted to climate trajectories, mobile contracting—but they remain experimental in France.
  • The challenge by 2030–2050: designing financial instruments capable of keeping pace with moving biology, otherwise we risk paying to maintain biological deserts while leaving future refuges without funding.

Field margins as sentinels of change

A field margin is seemingly little: a strip of grass and wild plants between the crop and the ditch, measuring anywhere from a few decimeters to a few meters across. Yet these fringes harbor a disproportionate fraction of agricultural biodiversity—pollinating insects, seed-eating birds, small mammals. They filter runoff, slow erosion, and serve as corridors between woodland massifs.

The ANSES-INRAE study focuses on these precise features, tracked over eight years across 500 sites distributed throughout France. The diagnosis is clear: the plant communities of the margins are changing in composition. Species employing competitive and ruderal strategies are gradually retreating, while those able to withstand drought are advancing. The pace of replacement is a gradual substitution, consistent with measured climate changes: 1.2 °C higher on average across the plots, soil moisture down 14% on average.

What strikes about these data is their granularity. We’re not talking about a climate model projecting national averages. We’re talking about repeated botanical surveys on actual plots, with identified species, over eight years. The plant substitution is measured, not extrapolated.

Payments drawn up before the shift

The Common Agricultural Policy allocates a portion of its aid to biodiversity-friendly practices. Agro-environmental and climate measures—MAEC in the sector’s vocabulary, that is, contracts between a farmer and the state to adopt environmentally beneficial practices—can support certain infrastructure like hedgerows depending on the territory, but the eligibility of surfaces coded as margins under MAEC depends on the declaration code, the type of MAEC, and the zoning of the PAEC (Agro-environmental and Climate Plan). France’s strategic plan for the CAP 2023–2027 maintains this mechanism and has even expanded it slightly.

The design problem is as follows. These contracts are awarded according to geographic zonation: this region, this type of habitat, this Natura 2000 perimeter. For a localized MAEC, only elements situated at least partly within the PAEC zoning can be committed.

Life moves. Thermophilic species climb northward and to cooler altitudes. Over 2013–2021, the plant communities of monitored margins at the national scale experienced a functional shift and an increase in species richness, with advancing drought-tolerant species. The zonations are administratively defined for the programming period, with possible updates according to campaigns, but this adjustment cycle remains out of step with the pace at which species are shifting.

Result: a farmer may receive aid for maintaining a margin whose ecological value is collapsing, while another whose plot has become a biological refuge in fact does not meet eligibility criteria. MAEC aid is governed by 2023–2027 programming and by zonations established for that programming, not a map from 2015.

Agricultural financial instruments resist movement

This mismatch is not administrative oversight. It reflects a structural tension in the design of public environmental policies.

Zone-based payments rest on a logic of stability: you identify precious ecosystems, map them, pay for their conservation. This framework worked well in a stable climate, where “precious today” was roughly equivalent to “precious tomorrow.” The CAP was built on this assumption. Revising its zonations requires lengthy negotiations, official surveys, multi-year administrative decisions.

Climate, meanwhile, advances at its own pace. Between two revisions of the national strategic plan, the plant communities of a margin may have changed composition twice over. The mismatch between biological speed and administrative speed lies at the heart of the problem the ANSES-INRAE study makes visible.

There is an instructive precedent in other conservation policies. Multilateral funding for ecological restoration faces a similar constraint: restoration models work, but financial instruments are designed for fixed geographic targets, not for moving living systems. The problem of French agricultural margins is a local and measured version of this same rigidity.

The Ministry of Ecological Transition has been working, for several years, on the concept of environmental service payments—direct compensation to farmers who produce measurable collective goods: biodiversity, water retention, carbon storage. The idea is more flexible than classical MAEC: you pay for a result, not a practice in a given zone. But deployment remains limited. Current experiments cover only a fraction of the affected surfaces, and their scaling faces the challenge of measurement costs: evaluating biodiversity plot by plot is expensive.

Observations from farmers on the ground

The shift of species in margins goes beyond the statistical findings of researchers alone. Farmers in the Paris Basin and the Loire Valley have reported in recent years the disappearance of species they encountered since childhood—some weeds, some insects—and the appearance of others, less familiar to them.

These practical observations converge with the ANSES-INRAE survey data. Farmers who maintain their margins carefully, by avoiding too-short mowing and leaving refuge areas, observe that the effort is not in vain: margins managed this way retain higher plant diversity than those abandoned or mowed short. Management matters. But it matters more where climate conditions still allow multiple plant communities to coexist—that is, in cool zones, at higher altitudes, or in areas where soil moisture persists.

These cool zones are precisely those that agricultural zoning protects least well, because they were not priorities when agricultural biodiversity was abundant everywhere. They are becoming refuges by default, without public policy having anticipated this role.

Funding refuges in motion by 2050

The question posed by the ANSES-INRAE study goes beyond the French case. It poses a design challenge for all agro-environmental policies in temperate countries: how do you fund the conservation of agricultural life when that life moves faster than the instruments protecting it?

Three trajectories are emerging for the coming decades, though none is yet clearly charted in France.

The first rests on accelerated periodic revision of zonations. You keep the logic of zone-based payments, but shorten the update cycle of maps to every three or four years rather than every ten. Standardized botanical surveys, like those the ANSES-INRAE study carried out on 500 plots, could automatically feed into a revision of eligible perimeters. The National Museum of Natural History and networks of naturalists have the tools to do this at the national scale. The cost is real, but moderate compared to the cost of impoverished agricultural biodiversity.

This trajectory is reformist: it keeps the existing framework and adapts it to the pace of change.

The second trajectory bets on measured environmental service payments. The farmer is paid because an evaluation of their plot attests to a service delivered: plant diversity above a threshold, a maintained pollinator insect population, documented water retention. This model follows life by construction, since measurement happens on site. The United Kingdom has embarked on a transition to this type of instrument with its Environmental Land Management schemes after leaving the CAP. Initial results are encouraging on farmer participation, but the cost of measurement and monitoring remains the main obstacle to scaling.

In France, local experiments, notably within agro-environmental projects in water-catchment zones, are testing comparable mechanisms without yet reaching a scale that would allow assessment of their effectiveness for biodiversity.

The third trajectory is more structural. It consists of explicitly integrating the climate-driven mobility of species into the design of national CAP strategic plans. Concretely: identify at the design stage zones likely to become refuges within ten or twenty years, according to regional climate projections available from Météo-France and INRAE, and pre-qualify these zones for conditional agro-environmental aid. Aid is triggered because the refuge potential is documented and the farmer commits to maintaining or improving reception conditions. This approach requires cooperation between climatologists, ecologists, and public policy designers that France has not yet institutionalized at this scale.

It also assumes political acceptance of preventive funding—paying for what is not yet there—which runs counter to usual budgetary logic.

These three trajectories are not mutually exclusive. Accelerated zoning revision can coexist with experiments in measured payments and gradual integration of climate projections in the next CAP revision after 2027. The European Union has inscribed in its “Farm to Fork” strategy the objective of covering 10% of agricultural surfaces in high-diversity landscape elements by 2030—hedgerows, margins, flower strips. Reaching this target in a context of species movement would require revising the very definition of what we seek to protect: no longer fixed habitat types, but ecological functions—thermal refuge, connectivity, water retention—capable of moving with life.

The survey of 500 plots offers precisely the type of longitudinal monitoring missing to build these instruments. Replicating this protocol at larger scale, on shorter cycles, and sharing data with designers of national strategic plans: that is the transmission chain between field science and public policy that the coming years must consolidate.

The open question is one of pace. The post-2027 framework must be prepared in 2027 for implementation scheduled for January 1, 2028, pending adoption of legislation by the European Union. Within this calendar, the approaches emerging from current experiments—measured payments, dynamic zonations, integration of climate projections—have a chance of being scaled. The eight years of ANSES-INRAE data constitute a weighty argument for including them.

Sources

  1. ANSES-INRAE, Study on field margins, monitoring 2013-2021 (2025)
  2. Ministry of Agriculture and Food Sovereignty, National Strategic Plan for the CAP 2023–2027 (plan-strategique-national.agriculture.gouv.fr)
  3. Ministry of Ecological Transition, Environmental Service Payments Scheme, framing note (ecologie.gouv.fr)
  4. European Commission, “Farm to Fork” Strategy, 2030 biodiversity objectives (agriculture.ec.europa.eu)
  5. National Museum of Natural History, National Biodiversity Observatory, monitoring protocols for agricultural plant communities (mnhn.fr)