Anonymised, illustrative composite. A Kelowna, BC residential builder hit an early, hard onset of winter that shut down foundation and framing work well before the crew expected it. The first attempt to claim it as a delay failed for having no baseline. The second didn’t.
At a glance
The Okanagan’s building season is genuinely long by Canadian standards, which is part of why schedules across the region are routinely priced tight against it — there is little slack built in for a season that runs meaningfully shorter than usual. That made this project especially exposed when the working season came up weeks short of what the schedule had assumed.
Foundation and exterior framing work on the project stopped when hard freeze conditions arrived weeks earlier than the crew had scheduled around. The builder had priced and sequenced the job assuming a typical Okanagan working season, and the early onset of winter cost real weeks of productive site time before the crew could resume once conditions allowed.
The builder’s first attempt to claim those weeks as an excusable delay was framed the way most weather claims are framed on a site: it was a rough winter, work stopped, here is the cost. The contract administrator rejected it. Weather happens every year in the Okanagan, and ordinary seasonal weather — even a difficult version of it — is a foreseeable risk a builder is expected to have priced into its schedule already. Nothing in the claim distinguished this season from a typical one.
An excusable weather delay has to be measured against something. Environment and Climate Change Canada publishes 1991–2020 climate normals station by station, and for the Kelowna station the normal frost-free period runs a median of 130 days, with a median last spring frost around May 16 and a median first fall frost around September 23. That is a free, citable, station-specific baseline for exactly the question a delay claim needs answered: was this season typical, or was it genuinely abnormal?
On the rebuilt claim, the builder pulled the actual dates the site log showed work had stopped and compared them against that median. The season’s first hard frost and sustained sub-zero conditions had arrived nearly three weeks before the September 23 median date — not merely an early-feeling winter, but a season measurably outside the normal window the builder had scheduled around, evidenced against a published federal baseline rather than a foreman’s impression.
130 days: the 1991–2020 median frost-free period for the Kelowna station. September 23: the median first fall frost date. Roughly three weeks earlier than that median: how far ahead of schedule the season’s actual freeze arrived, per the site log. Eight weeks: the extension of time ultimately recovered once the early onset, combined with a subsequent stretch of extreme cold days, was measured against the published normal rather than asserted on its own.
The bind is what separated a rejected claim from an eight-week recovery on essentially the same underlying weather event: a claim that just says “it was a bad winter” gives a contract administrator nothing to test it against, and ordinary seasonal weather is deemed priced into the schedule already. A claim that shows the actual dates against ECCC’s published normal for that exact station converts the same facts into evidence of a genuinely abnormal season — which is what a CCDC-based extension-of-time process needs before it will treat a delay as excusable rather than ordinary.
Nothing about the underlying weather changed between the two attempts — the same cold snap, the same lost weeks, the same site log. What changed was whether the claim could show that snap fell outside the range a builder scheduling around this station’s own thirty-year history should reasonably have expected.
The builder now pulls the ECCC normal for the nearest station at the start of every active project and keeps a season log measured against it from day one, so any future weather-delay claim already has its baseline built in rather than assembled after the fact. For how this kind of seasonal risk plays out across the wider Okanagan trades market, see seasonal trade conditions in the Kelowna market, and for the general elements a delay claim needs regardless of cause, see the elements a delay claim needs.
For how AI now helps flag weather-exposed schedule slippage earlier, see AI-assisted weather delay planning.
A 30-minute call is enough to tell you whether AI pays for itself here.
