The ordinary rain module set the major event aside to treat the one that recurs. The other half must now be taken up, that of sites where the hazard has a name, a season and a map: the cyclone coast, the flood plain, the seismic zone. On those sites the question is not whether the event will come, it is what state it will find the works in. And that second question, unlike the first, is decided entirely inside the project.
The starting point is that the vulnerability of works under construction has nothing to do with that of the same works finished. A completed building in a cyclone zone has been designed for the reference event, it carries its roof fixed down, its openings tested, its envelope closed and its services protected. The same building halfway through shows a bare frame, open floors, façades not yet hung, materials stored in the open and lifting plant deployed. It is not the same object in front of the same wind, and the gap between the two is counted in orders of magnitude, not percentages.
From that follows a result intuition first refuses. One readily imagines a site's worst moment is the eve of acceptance, when all the value is in place. That is wrong as soon as the hazard is seasonal and vulnerability varies strongly. What is at stake is not installed value but the product of that value and the share an event would destroy, and that product passes through a peak in mid-programme, when value is already considerable and the permanent protections are not yet there. Maximum exposure and maximum value do not fall on the same day.
That offset explains a frequent and never settled disagreement with the way amounts are tracked. The drifting sum module showed that value in place updates with progress, that is along a curve that only rises. The exposure curve rises and then comes back down. Following the first while believing one is following the second leads to reassurance precisely in the period when concern is due, and to concern at the end, when the works have become the resistant object they were designed to be.
The practical consequence is that a site's exposure in an exposed zone is steered by the calendar, and by almost nothing else. Closing an envelope before the season opens, deferring the delivery of sensitive stock, choosing to dismantle or anchor lifting plant, sequencing sections so that none crosses the season in its most open state: these are programme decisions, taken by people whose trade is to hold a date and whose effect on risk no indicator measures. That is what makes these trade-offs silent: they are made every day, they weigh heavily, and nobody in the room looks at them in that light.
It must be added that exposure built this way is concentrated, which the accumulation on one site module described from another angle. A cyclone draws no distinction between packages: it finds the whole site on the same day, with its stocks, its temporary installations, its plant and its works, and it produces a single event where an ordinary year would have produced twenty independent ones. That concentration distinguishes an exposed zone from one that is not far more than the frequency of the hazard itself.
That leaves what can honestly be said about prevention during this period. It does not consist in making unfinished works resistant, which is not possible: it consists in reducing what stands in the vulnerable state at the moment the hazard is probable. Tie down, clear out, close up, bring forward, push back. That vocabulary is the site manager's and not insurance's, which is the sign that the subject is settled on site. The prevention survey module describes the outside eye that brings it to light; that eye still has to come before the season, and not after.
A €180m hotel complex is being built on a cyclone coast where the season runs from June to November. Value in place on July 1 is €22m in the first year, €96m in the second and €168m in the third. For the reference event the prevention engineer takes a destroyed share of 15% while the site is at foundations, 60% once the frame is up and the envelope still open, and 12% once the envelope is closed. The second season finds the site with the frame up and the envelope not hung; the third finds it with the envelope closed. Where does the worst moment sit, and what is bringing the envelope forward by four months worth?
The calculation is three multiplications, and it is their comparison that surprises. The first season exposes €22m at 15%, that is €3.3m. The second exposes €96m at 60%, that is €57.6m. The third exposes €168m at 12%, that is €20.2m. The worst moment is therefore the second season, when only 53% of the value is in place, and it is worth nearly three times the last season, where installed value is nonetheless three quarters higher. The intuition placing the worst moment on the eve of acceptance is contradicted by a factor of 2.9. On the calendar trade-off the answer is of the same order: bringing envelope closure forward by four months takes the second season from 60% to 12% destroyed share, that is from €57.6m to €11.5m, and so removes €46.1m of exposure. That figure deserves comparing with what it actually costs to mobilize the façade trades earlier, which is a construction management discussion settled in site euros. The non-obvious point is who decides: this decision belongs to the programme, it is taken in a progress meeting, and nothing in ordinary tracking of amounts makes it appear as the project's heaviest risk decision. Tracking value in place, which only rises, would on the contrary have pointed to the third season as the most worrying.
- 01In an exposed zone the question is not whether the event will come but what state it will find the works in, and that second question is decided entirely inside the project.
- 02Works under construction and the same works finished are not the same object in front of the same wind: the gap is counted in orders of magnitude, not percentages.
- 03Maximum exposure and maximum value do not fall on the same day: what is at stake is value multiplied by destroyed share, and that product passes through a peak in mid-programme.
- 04Value in place only rises, exposure rises then falls: following the first while believing one follows the second reassures exactly when concern is due.
- 05Exposure is steered by the calendar, and the trade-offs that move it are taken in progress meetings by people whose effect on risk no indicator measures.