The first thing an underwriter does when facing a site in a climate-exposed area is to look at a map. Several exist, they do not say the same thing, and confusing them is the field's most frequent opening error. A regulatory zoning map, a hazard map produced by a technical service and an insurer's rating zone map are three distinct objects, built for three uses, and none of the three was made to answer the question being put to it.
Regulatory zoning is a planning decision. It says what may be built, under what conditions, and it carries prohibitions and requirements. It is the product of an administrative procedure, of consultation and sometimes of litigation: its grain is that of the parcel and its logic that of controlling development. It does not say what an already built structure will lose in a flood, and it never claimed to. A site outside a regulated zone is not a site outside the hazard, it is a site where building is allowed.
The hazard map says something else and says it truly within its frame: here is the water depth expected for such a reference scenario. Its limit lies in the words "such a scenario": it is computed for a given event, with protective structures assumed to hold, and it is silent on what exceeds that scenario. Yet what exceeds is precisely what concerns an insurer, since that is where its burden is made. A map that stops at the reference scenario is mute on the tail of the distribution.
The insurer's rating zoning, finally, is a commercial instrument as much as a technical one. It carves the territory into classes, often on an administrative grain, because that is the grain available in the files. That grain bears no relation to the peril's. One commune can contain a valley-floor district and a plateau, and receive a single class that overrates one and underrates the other. This is not a design flaw, it is the price of an instrument that must apply to millions of contracts without a survey.
From these three gaps comes a recurring situation worth recognizing: the out-of-zone site that ends up under water. Three mechanisms produce it. Runoff, where the water comes not from a watercourse but from a sealed slope and a saturated network, and which no river flood map represents. Failure of a protective structure, embankment or pumping station, whose good working order the map assumed. And rising groundwater, which arrives from underneath and appears on almost no flood map.
One must add the reverse case, less spectacular and more costly in business lost: the in-zone site that risks nothing. A building whose first usable floor is three meters above natural ground, with no basement, its technical plant on the roof, is in a flood zone and has almost no vulnerability to flooding. Declining that risk, or loading it on the strength of the zoning alone, means leaving to a competitor who took the time to look a risk better than the average of its class.
The conduct to adopt requires neither software nor budget. One asks on which scenario and with which structures the map consulted was built. One looks at local topography rather than at the color of the parcel: where the water comes from, where it goes, what stands upstream. One records the level of the first usable floor and the position of technical plant, which decide vulnerability far more than the zone does. And one keeps in mind the sentence that sums up this lesson: a zoning map answers the planner's question, and an underwriter is not a planner.
An underwriter examines two sites of the same food group, presented together at renewal. Site A, 12 million euros in values, is classed in the blue zone of its commune's flood risk prevention plan; its workshops are 2.80 meters above natural ground, it has no basement, its refrigeration units are on the roof. Site B, 9 million in values, lies outside any regulated zone; it stands at the foot of a hillside urbanized over the past fifteen years, its receiving dock sits below road level, and its electrical cabinets are at ground level in a semi-buried room. The committee proposes loading A by 40% and accepting B at base rate. What must be objected?
The decision exactly inverts the two vulnerabilities, and it does so because it reads a map made to answer another question. A's blue-zone classification is a planning decision: it says building is allowed there under conditions, and it says nothing about what the workshops will lose. Yet the three items that actually decide A's vulnerability are in the file and all point the same way: first floor at 2.80 meters, no basement, technical plant on the roof. Over almost the whole distribution of plausible water depths, this site takes nothing worse than yard cleaning. Loading it by 40% means charging a good risk the average of its class, and letting it leave for whoever took the trouble to look. Site B poses the symmetric problem and it is the file's real subject. Outside a regulated zone does not mean outside the hazard: the zone describes a watercourse overflowing, and what threatens B is not a watercourse but a hillside urbanized over fifteen years, that is, a slope whose absorption capacity was removed as it was built on. Runoff appears on no river flood map, it is no less real, and the file's two other items say it will cost: a dock below road level funnels water toward the building, and electrical cabinets in a semi-buried room mean fifteen centimeters of water are enough to stop the site for as long as replacement takes. What has to be asked before rating is neither a map nor a model. It is B's upstream topography, the record of runoff episodes reported at the town hall, and the level of technical plant on both sites. Three items obtained in one visit, which reverse the committee's decision.
- 01Regulatory zoning, hazard maps and rating zones are three objects made for three uses, and none answers the underwriter's question.
- 02A hazard map holds for a reference scenario with protective structures assumed to hold: it is mute precisely on the tail of the distribution.
- 03Out of zone does not mean out of hazard: runoff, structure failure and rising groundwater appear on almost no flood map.
- 04The reverse case costs in business lost: a high first floor, no basement and no ground-level plant is a good risk its zone drives away.
- 05The level of the first usable floor and the position of technical plant decide vulnerability far more than the color of the parcel.