Wet soil does not tell us how much water lies beneath our feet. To understand a drought, we need to track several indicators: precipitation, soil moisture and underground water reserves. France’s September 2026 assessment gives a measure of the scale of the deficit, without treating every shower as proof that conditions have returned to normal.
A record that must be read in context
In its bulletin published on 9 October, Météo-France measures a 70% rainfall deficit nationwide for September 2026, compared with the 1991-2020 reference period. It was the third driest September in the 1959-2026 series. The national soil moisture indicator, meanwhile, reached an unprecedented low for a September.
These two findings are not interchangeable. The first concerns the rain that fell over the course of a month; the second describes surface soils. They do not allow us to claim that all French aquifers have reached their lowest level on record. The national average also conceals regional differences: the same bulletin reports above-average rainfall in eastern Languedoc.
Water takes several paths before reaching an aquifer
The BRGM explains that some rain evaporates, runs off or remains in the soil, where plants can use it. Only a fraction continues deeper underground. The speed depends in particular on the rocks it passes through and the depth of the aquifer.
Recharge occurs mainly from autumn to early spring, when evaporation decreases and vegetation draws less water. Some aquifers respond quickly; others change slowly, sometimes over several years. Surface conditions can therefore improve before underground reserves do.
What a shower cannot tell us
According to the BRGM, heavy rainfall can cause runoff without replenishing reserves sufficiently. Making up a deficit that has built up often requires regular, sustained rainfall.
When reading a new assessment, three questions help avoid jumping to conclusions: what kind of water is being discussed, over what area and as of what date? The September bulletin describes September. It is neither a measurement of aquifers at the end of October nor a forecast of water cuts in any particular municipality. Such claims would require other, up-to-date local observations.
