Unusually Severe Weather: What Actually Counts

Three of the most widely used contract families in American construction hang weather entitlement on this phrase. Not one of them defines it. That gap is where claims are won and lost — and it is a data problem, not a legal one.

The same test, three names

Whichever paper your job runs on, the weather language asks one question: was this worse than what you should have expected here, at this time of year? The wording changes; the test does not.

DocumentThe phrase it usesNotice deadline
AIA A201-2017"abnormal for the period of time" (§15.1.6.2)21 days (§15.1.3)
FAR 52.249-10"unusually severe weather"10 days
FAR 52.249-14"unusually severe weather"Per contract terms
ConsensusDocs 200"adverse weather conditions not reasonably anticipated" (§6.3.1)14 days to claim, +21 to document (§8.4)

Note the notice clocks. They are not the same, and the federal one is the shortest by a wide margin. Ten days on a fixed-price federal job versus twenty-one under A201 is the difference between a claim you still have and a claim you lost while assembling evidence for it.

Why nobody defines it

The omission is deliberate. A numeric threshold written into a national form document would be wrong nearly everywhere it was applied — three inches of February rain is unremarkable in Seattle and a genuine event in Phoenix. So the drafters left the standard comparative and pushed the measurement onto the parties.

The practical consequence: the phrase has no meaning until you supply a baseline. An owner who says "that was a normal spring" and a sub who says "that was the worst spring in a decade" are not disagreeing about the contract. They are disagreeing about a number neither has produced.

The baseline is the whole argument

The customary reference is a 10-year monthly average of daily observations from the nearest official station — long enough to smooth out a freak year, short enough to reflect the current climate. Federal construction contracts often go further and simply hand you the table (see federal contract weather delays), which is why federal weather claims tend to be more arithmetic than argument.

Two mistakes to avoid when building one:

Where we draw the line, and why

Ilystics publishes an abnormal threshold for every month in 45 US metros. It is the 10-year mean of lost workdays for that month plus one standard deviation, capped at the number of workdays in a month. Above that line, the month sits in roughly the worst sixth of years on record for that market — the point at which "unusual" stops being rhetoric.

Be straight about what this is. Mean-plus-one-standard-deviation is not a legal standard. No court or board has adopted it, and no contract we know of specifies it. It is a defensible, disclosed convention for turning a comparative phrase into a number you can put in front of a GC — and its virtue is that the underlying distribution is published alongside it, so anyone can check the work or argue for a different line.

Worked example: Richmond, VA

Ten years of NOAA GHCND observations, scoped to soil-dependent site work — earthwork, grading, excavation, foundations and underground utilities:

MonthAvg. precipTypical lost workdaysAbnormal above
January3.67 in1517
April2.86 in1013
July5.44 in1217
November2.54 in912

Read the January and July rows together, because they contain the entire argument for measuring days instead of inches. July gets about 48% more rain than January and loses three fewer workdays. Summer rain in Richmond arrives as short convective storms onto warm, fast-drying ground. January rain is lighter, slower, and lands on cold saturated clay that stays untrafficable for days after the sky clears.

A claim built on "we got 5.4 inches in July, well above the annual monthly average" invites the response that July is always the wettest month here. A claim built on "we lost 19 workdays against a 12-day norm and a 17-day abnormal threshold" is a different conversation. Same weather, same site, opposite outcome.

Severity is necessary, not sufficient

Proving the month was unusual gets you past the first gate only. Every one of these clauses also requires that the weather actually delayed work:

What to do, in order

  1. Find your notice deadline before you need it. Ten, fourteen or twenty-one days — it is in the contract you already signed, and it starts running without asking you.
  2. Check the supplementary conditions for a table. If the contract fixes anticipated weather days per month, that table is your baseline and the historical argument is over before it starts.
  3. Pull the baseline for your exact site, not your region.
  4. Log daily, including the days after the rain. The drying days are the ones nobody documents and everybody loses.
  5. Send notice early and supplement later. A thin timely notice beats a thorough late one.

Get the baseline this test is measured against

Ilystics generates an Anticipated Weather Exhibit for your exact jobsite: 10-year NOAA baseline, monthly abnormal thresholds, and USDA soil drying analysis. Free, no account.

Generate a Free Exhibit

Already in a dispute? Request a $249 Claim Report for the actual-vs-baseline analysis.

Ilystics provides weather and soil analysis, not legal advice. Section and clause references are to AIA Document A201-2017, FAR 52.249-10 and 52.249-14, and ConsensusDocs 200, and are provided for general orientation only — your contract, its supplementary conditions, and your jurisdiction control. The abnormal threshold described above is an Ilystics analytical convention, not a legal standard. Consult your attorney or claims consultant on entitlement.

Frequently asked questions

What does "unusually severe weather" mean in a construction contract?

It means weather materially worse than the historical norm for that location and time of year. No standard contract defines it numerically — A201, the FAR and ConsensusDocs all use comparative language without a threshold, which is why the baseline you compare against does the real work in a claim.

How much worse than normal does weather have to be?

No contract or regulation sets a number. In practice the reference point is a 10-year monthly baseline from the nearest official station, and the arguable range starts roughly one standard deviation above the mean. A month one day worse than average is not unusual; a month a week worse than average generally is.

Is "unusually severe weather" the same as "abnormal weather"?

Functionally yes. The FAR says "unusually severe weather," AIA A201 requires data showing conditions were "abnormal for the period of time," and ConsensusDocs 200 says "adverse weather conditions not reasonably anticipated." All three ask the same question: worse than what should have been expected here, this month.

Does heavy rainfall alone prove unusually severe weather?

No, and it is the most common way claims fail. Rainfall total is not the same as lost workdays. In Richmond, July averages nearly 50% more rain than January but loses fewer workdays, because summer rain arrives in short storms on ground that dries fast. The claimable unit is the unworkable day, not the inch.