Why Hail Damage Comes Down to Minutes, and How Early Warning Changes the Outcome

A storm cell complex can run for two hours, but the individual cell inside it that actually drops hail lives for about 20 minutes. The realistic window to do something about it, to get a seeding agent into the cloud before hail forms, is closer to 10. Everything about hail risk management comes down to whether a system can see that cell forming early enough to act inside that window.

Radar Alone Isn’t Early Enough

Traditional rainfall radar detects a storm once it’s already dense enough to reflect a strong signal, which is often later than operators would like. Selerys’s SKYDETECT closes part of that gap on its own: a short-range X-Band radar with a 6kW antenna, covering a 30km radius in full half-sphere monitoring, built on a storm database spanning 20 years of pattern data. It distinguishes a genuinely developing storm from clouds just passing through, and it installs on a tower, mast, or building in a single day, with mobile mast-mounted units available for temporary or urgent coverage.

But radar still has a physical limit: it needs enough precipitation density to get a clear return. Electrical activity inside a cloud builds before that density arrives.

Reading the Storm Before the Radar Can

That’s the gap EVEARLYS is built to close. It’s a UHF interferometric lightning detector, physically four panels per installation, each 1.3m by 1m and 7kg, giving 180° of field of view per panel, running on as little as four hours of uninterrupted power. It detects more than 99% of total lightning activity, both intra-cloud and cloud-to-ground, at a processing resolution of 10 microseconds.

The value isn’t that EVEARLYS replaces SKYDETECT. It’s that the two read different signals from the same storm. SKYDETECT sees the storm’s shape and movement on radar. EVEARLYS reads how electrically active it already is, which tells operators how mature a cell that’s already visible on radar actually is, before it’s dense enough to register strongly on traditional rainfall radar. Used together, they turn “a storm is forming” into “this specific cell is far enough along that it needs a response now,” which is the distinction that actually matters inside a 10-minute window.

Turning a Warning Into a Launch

Detection only helps if it turns into action fast enough to matter, which is where CIRRUS, Selerys’s coordination software, does its work. It centralizes SKYDETECT, EVEARLYS, and the SOBLI balloon and drone fleet in a single interface, scanning the half-sphere in HD once a minute and tracking cloud cells in real time. It anticipates cloud motion 10 to 20 minutes ahead, using the same storm database that powers SKYDETECT, and issues automated alerts by SMS, email, or phone the moment a cell crosses the threshold that warrants a response.

From there, a launch decision through a SOBLI station reaches its target in under five minutes, well inside the 10-minute window a storm cell actually allows. Every launch and every reading gets logged to a searchable archive going back more than a year, along with flare and launch stock levels per station, so an operator isn’t just reacting in the moment. They’re building a record of exactly which cells got treated, when, and with what result.

What Faster Detection Is Actually Worth

The wine-region hail mitigation results Selerys has published elsewhere, a 74% drop in severe hail frequency and a 26% reduction in average hailstone size, don’t come from the balloons alone. They depend on the balloons arriving inside a window that’s often shorter than the time it takes to make a phone call. Detection speed is what makes that window usable at all. A launch system that can reach a target in five minutes is only as good as the warning that tells it when to fire, and that’s the part of the chain that early warning technology is actually solving.

The Broader Pattern

Hail isn’t becoming less frequent. National data across France, for instance, shows a rising trend in hail days over the past two decades, even in regions with active mitigation programs holding their own local numbers flat. That divergence, worse conditions regionally, stable or improving outcomes locally, is what early warning paired with fast-response seeding is built to produce: not a world without storms, but a shorter list of storms that actually get to do damage.