Clay shrink–swell: crossing RGA 2026, CatNat orders and soil drought

The official RGA zoning is spatial: it describes exposure zones that do not follow municipal boundaries. Published drought CatNat orders, by contrast, mention the municipalities concerned for a given period and natural phenomenon. To compare these two types of information at national scale, the RGA zoning has to be brought onto a common spatial unit.

This is what fiñv does: the official zoning is summarized at municipal scale so it can be compared with CatNat orders, hydroclimatic trajectories, recent weather indicators, geology and topography.

This operation makes the datasets comparable, but it simplifies reality. A single municipality may contain several RGA exposure levels, several geological formations, contrasting soil contexts and very different buildings. The municipal result therefore does not describe the soil beneath each house.

The point of this aggregation is not to say that a municipality is dangerous. It is to identify, using a shared institutional unit, the territories where signals converge, diverge or justify a finer local analysis.

The official exposure zoning for clay shrink–swell changes in 2026 with the order of 9 January 2026, which amends the order of 22 July 2020 defining the zones exposed to the clay shrink–swell phenomenon [2][9].

Around 240,000 RGA-related claims were declared between 2018 and 2022, representing 58% of all RGA claims since 1989 [1]. Areas of medium and high exposure now account for 55% of mainland France, compared with 48% in the previous zoning [1][2].

The new zoning therefore does not mean that clays, marls or superficial formations changed within a few years. It reflects a new reading of exposure to the phenomenon in a climate, insurance and regulatory context that has become more critical.

A drought CatNat order is an administrative decision published in the French Official Journal. For a given municipality, period and natural phenomenon, it mentions whether the state of natural disaster is recognized or not recognized.

It is not an inventory of observed damage. A municipality that is not recognized may still have experienced cracks or building damage. Conversely, a municipality recognized under CatNat does not mean that all buildings are affected, or that all claims will automatically be compensated.

CatNat orders are therefore used here as a published institutional signal, not as a direct measurement of drought, cracks or building vulnerability.

The map layers do not describe the same object. They should be read as complementary signals, not as equivalent evidence.

• RGA zoning: official exposure to the clay shrink–swell phenomenon.
• Geology: useful lithological context, but not enough to conclude on local geotechnical susceptibility.
• SWI and hydroclimatic indicators: modelled signals, useful for comparing trajectories between territories.
• CatNat orders: published administrative response, not an exhaustive inventory of damage.
• Buildings: an essential link in the damage chain, still absent from this first municipal-scale reading.

When several signals point in the same direction, the municipality deserves closer attention. When they diverge, the gap is itself informative: it may reveal a scale effect, an indicator limitation, a local specificity, or a difference between the physical phenomenon and the published administrative response.

The uniform SWI does not describe the actual soil of a municipality. It provides a modelled meteorological signal, comparable between territories, within the drought CatNat framework.

The uniform SWI deliberately neutralizes local differences in soil and vegetation. It therefore does not aim to represent each municipality faithfully; it is used to compare the meteorological effect between territories within a homogeneous administrative framework.

This indicator is modelled by Météo-France’s SIM system on a grid of roughly 8 km. It is relevant for comparing hydroclimatic trajectories at broad scale, but it does not measure moisture at the level of a building, parcel or foundation.

It therefore becomes misleading if read as a local measurement. Its value is comparative: identifying soil drought trajectories and comparing them with RGA zoning, published drought CatNat orders and geological context.

The map does not only compare states. It also distinguishes two time scales: long-term hydroclimatic trends and recent deviations. The objective is not to say that a municipality is “at risk” because one indicator is high, but to examine whether the recent period extends, amplifies or contradicts an already established trajectory.

This distinction matters for clay shrink–swell. Some signals, such as thermal indicators, may reinforce an already marked geography. Other hydrological signals may instead appear as recent breaks or spatial redistributions.

These indicators remain municipal-scale and relative readings. They do not replace a local soil measurement or a building assessment. They help identify whether a territory follows a long-term dynamic, shows a recent deviation or presents a more ambiguous situation.

The fiñv typology is an exploratory classification built through rules for crossing and grouping public signals brought to municipal scale. It does not come from BRGM, CCR or Météo-France, and it is not an official regulatory classification.

It is used to identify territorial situations: convergence of sensitivity or constraint signals, increase in RGA zoning without a recently published drought CatNat order, marked CatNat history without a robust hydrological constraint, or municipalities where recent weather indicators reinforce an already sensitive context.

This typology has not been calibrated on an exhaustive database of building damage. It should therefore be read as a tool for territorial exploration, not as an explanatory or predictive model of claims.

RGA damage does not depend only on soil and drought. It also depends on buildings: foundation type, construction age, presence or absence of a basement, extensions, networks, drainage, nearby vegetation, surface sealing around the building and maintenance.

This is why a municipality with high aggregated RGA exposure does not automatically correspond to widespread damage. Conversely, a locally sensitive zone may concentrate vulnerable buildings even if the municipal signal remains moderate.

The sources below do not play the same role. Some provide official data, others provide hydroclimatic indicators or geographic reference datasets. The fiñv processing makes these datasets comparable at municipal scale.

fiñv processing. The municipal aggregation and the typology displayed in the map are exploratory treatments: they make heterogeneous public datasets comparable at municipal scale, without replacing the official RGA zoning or local expertise.