How are these numbers calculated?

Each metro is run through the same model as our paid exhibits: ten years of daily NOAA GHCND station observations, walked day by day, with a soil-specific drying clock. A Monday-to-Friday day counts as lost when precipitation exceeded 0.1 inch, the ground had not finished drying, or the ground was frozen.

Why do two metros with similar rainfall differ so much?

Soil. Rain leaves a sand site almost immediately and sits on a clay site for days. Two metros with the same annual rainfall can differ by 50 or more lost workdays a year purely on drainage class, which is why rainfall totals alone are a poor proxy for schedule impact.

Can I use these figures in a bid?

They are sound planning figures for a metro. For a bid or a claim, run your actual jobsite ZIP: the estimator uses the station nearest the site and your soil rather than a metro centroid, and returns a documented exhibit.

Weather Delay Days by Metro

How many workdays earthwork and foundation crews lose to weather, metro by metro — from ten-year NOAA station baselines, scored against each location’s soil rather than rainfall totals alone. Enclosed and vertical work is not modelled.

Why rainfall totals mislead

An inch of rain costs a sand site nothing and a clay site three days. The column that matters below is not rainfall — it is lost workdays, which is rainfall put through a soil-specific drying model. Metros are ranked worst first.

Metro	Soil	Worst month	Lost workdays that month	Rainfall / yr
Minneapolis, MN	Silt	January	20	33.6″
Pittsburgh, PA	Clay	January	19	42.3″
Detroit, MI	Clay	January	19	40.0″
Chicago, IL	Clay	January	19	35.7″
Milwaukee, WI	Silt	January	18	35.3″
Portland, OR	Silt	January	18	43.2″
Cleveland, OH	Clay	January	17	31.8″
Newark / North Jersey, NJ	Silt	January	17	55.2″
Columbus, OH	Silt	January	17	43.0″
New Orleans, LA	Clay	July	17	76.0″
Seattle, WA	Silt	November	17	38.4″
Atlanta, GA	Clay	January	16	56.1″
Birmingham, AL	Clay	January	16	57.7″
Indianapolis, IN	Silt	January	16	46.0″
Boston, MA	Silt	January	16	43.4″
Hartford, CT	Silt	January	16	44.8″
Nashville, TN	Clay	January	15	52.5″
Richmond, VA	Clay	January	15	49.1″
Charlotte, NC	Clay	January	15	46.3″
Washington DC / NoVA, DC	Clay	January	14	46.5″
Baltimore, MD	Clay	January	14	44.8″
Raleigh-Durham, NC	Clay	January	14	51.0″
Philadelphia, PA	Silt	January	14	48.8″
Memphis, TN	Silt	January	14	58.9″
Houston, TX	Clay	January	14	56.5″
Boise, ID	Silt	December	14	15.6″
St. Louis, MO	Silt	January	13	44.2″
Dallas-Fort Worth, TX	Clay	May	13	44.4″
Oklahoma City, OK	Clay	May	12	33.0″
Salt Lake City, UT	Silt	January	12	18.0″
Sacramento, CA	Clay	January	12	20.5″
Kansas City, MO	Silt	January	11	31.1″
Austin, TX	Clay	May	10	32.7″
Denver, CO	Silt	May	10	14.2″
Tampa, FL	Sand	August	10	53.6″
San Antonio, TX	Clay	May	9	25.3″
Jacksonville, FL	Sand	July	9	56.0″
Orlando, FL	Sand	July	9	60.3″
Virginia Beach, VA	Sand	May	7	61.0″
Miami, FL	Sand	August	7	55.1″
Charleston, SC	Sand	August	7	46.1″
Savannah, GA	Sand	June	6	48.2″
Los Angeles, CA	Sand	January	4	15.5″
Phoenix, AZ	Sand	February	2	8.6″
Las Vegas, NV	Sand	January	1	3.4″

Minneapolis, MN currently tops the list: 20 unworkable workdays in January. Rainfall alone would not have told you that — several wetter metros lose fewer days because the water drains.

Note how often the worst month is a winter one rather than the wettest. Frontal rain arrives spread out and each event collects its own drying tail; summer storms cluster, and their tails overlap. The same rainfall costs more when it is spaced out, which is why totals mislead. The full write-up is here, including where these figures disagree with a published state contract baseline.

How to read these

Each metro page carries the full monthly table: the ten-year mean (the number to carry in a bid), the median, and an abnormality threshold set at the mean plus one standard deviation. Exceeding a mean is the expected case in roughly half of all periods and is not by itself evidence of abnormal weather; the threshold is the stronger basis for a claim.

These are monthly figures and should not be summed into a year. Each month assumes a site in the condition it is in that month. Adding twelve together would assume a jobsite never improves — no stone haul roads, no stabilised pads, no ditching — which is wrong for any project past its first few months. Over the two-to-four week window a delay claim actually covers, the effect is negligible. Over a year it is not, so we do not publish an annual figure.

Every page names the NOAA station it used and its distance, the soil basis and the effective drying constant, so the figures can be checked rather than taken on trust.

A metro average is not a jobsite

These use a metro centroid and a representative soil. Your site has a specific station, a specific soil, and specific dates. Enter a ZIP and the estimator runs the same model against all three and returns a contract-ready exhibit. Free, no account.

Run it for your ZIP

Baselines rebuilt September 15, 2026. Figures model soil-dependent site work — earthwork, grading, excavation, foundations and underground utilities. Enclosed and vertical work is not modelled. Planning figures from station observations and a deterministic soil model, not site-specific geotechnical findings.