NOAA Adverse Weather Days for Construction

In this report, NOAA adverse weather days for construction are the days in your date range that pass a line built from NOAA daily history at one station. A day on the ordinary side of that line is not marked. This is general information, not legal advice.

What the count is, and what it is not

NOAA does not publish a construction day count for a job site. The National Centers for Environmental Information publish the daily station record. This report reads that record and does the comparison itself.

The comparison is not a fixed rain cutoff, and it is not a monthly allowance written into a schedule. Some schedules use a count of expected bad days in a month. This report stays on each day. It asks whether that day's reading at the chosen station was outside the spread of the same dates in earlier years.

The PDF lists the marked days. It does not decide whether your contract treats a marked day, or an unmarked day, as extra time.

Where the daily history comes from

Every reading is a Global Historical Climatology Network daily (GHCN-Daily) observation, retrieved through NOAA's Climate Data Online service. These are the official NCEI pages for that record:

• GHCN-Daily product page
• Climate Data Online
• GHCN-Daily readme (element names, units, and the quality flag)

The readme names the elements this report uses. Maximum temperature is TMAX, in tenths of degrees Celsius. Minimum temperature is TMIN, in tenths of degrees Celsius. Precipitation is PRCP, in tenths of millimeters. Average daily wind speed is AWND, in tenths of meters per second. The report converts those to degrees Fahrenheit, inches, and miles per hour.

The same readme says a blank quality flag means the value did not fail a quality-assurance check. This report keeps a reading only when that flag is blank. It also drops a converted reading outside the range it accepts: temperature colder than −80°F or hotter than 140°F, precipitation below 0 inches or above 15 inches, and wind below 0 mph or above 100 mph. A dropped reading is left blank. It is not stored as zero.

How the nearest station is chosen

The report looks for a GHCN-Daily station near the project coordinates. Distance is the great-circle distance in miles. The search starts at 15 miles. It widens to 30 miles, then to 60 miles, only when no station inside the smaller distance qualifies.

A station qualifies for temperature and precipitation when both of these are true:

1. On the dates in the report, TMAX, TMIN, and PRCP are each present on at least 80 percent of the days.
2. The same 80 percent test passes for at least 5 of the 10 earlier years described below.

The end date has to be on or before the latest day NOAA has published at that station for these elements. A later end date disqualifies the station. When a wind reading exists, that latest day is the oldest of the latest maximum-temperature, minimum-temperature, precipitation, and wind dates, so a wind record that stops early can disqualify a station whose temperature and precipitation are more current.

Wind is separate. AWND is usable on a window when it is present on at least 80 percent of the days and no more than half of those wind readings are 0 mph. Wind is included only when the report dates pass that test and at least 5 of the baseline years pass it too.

Inside the search distance that has a qualifying station, a station with that usable wind baseline is chosen over a closer station that has temperature and precipitation only. The report then uses the closest station that has the usable wind baseline. If no station in that distance has one, the report uses the closest station that passed the temperature and precipitation test, and the PDF leaves wind out.

How the 10 baseline years are built

The baseline is the daily history for the same calendar window, not a single monthly normal and not a 30-year climate normal. The report checks the 10 years immediately before the year of the start date. A report that starts in 2026 checks 2016 through 2025. A year that fails the 80 percent test is skipped. At least 5 years have to pass. Every year that passes is used, up to those 10. The PDF names the years that were used.

Each earlier window is the report's start date and end date, each moved back the same number of years. A window that runs from December 20 into January 10 still runs from December 20 into January 10. It is not folded into one calendar year. February 29 moves to February 28 when that earlier year is not a leap year.

When wind is included, the years in the baseline are the years that also passed the wind test, and there are at least 5 of them. When wind is left out, the years are the ones that passed for TMAX, TMIN, and PRCP.

For each kept year, the report keeps every daily reading in that window. A 30-day report across 8 kept years can contribute up to 240 daily values for each element. Missing days are omitted from the average. They are not filled with zero, and they are not replaced with a neighboring day.

How the line is calculated

For each element, the report takes the mean of the daily readings it kept. The standard deviation is the square root of the average squared difference from that mean. The average uses the number of readings, not one less than that number.

The line for each element is:

• Maximum temperature: mean plus one standard deviation
• Minimum temperature: mean minus one standard deviation
• Precipitation: mean plus one standard deviation
• Wind, only when wind is included: mean plus one standard deviation

These numbers are an illustration of that arithmetic, not a real station. If the kept daily precipitation readings have a mean of 0.20 inches and a standard deviation of 0.15 inches, the line is 0.35 inches. If the kept maximum temperatures have a mean of 78°F and a standard deviation of 6°F, the line is 84°F. If the kept minimum temperatures have a mean of 40°F and a standard deviation of 5°F, the line is 35°F.

Why only days past that line count as abnormal

The line is the historical spread for those dates at that station. Weather inside the spread is already in the record. Rain, heat, cold, or wind that stays on the ordinary side of the line is not an abnormal day in this report. The report marks a day only when at least one reading crosses the line:

• Maximum temperature is greater than its line. A maximum equal to the line is not marked.
• Minimum temperature is less than its line. A minimum equal to the line is not marked. This is the cold side of the baseline, not a temperature above the average.
• Precipitation is greater than its line. A total equal to the line is not marked.
• When wind is included, average wind is greater than its line. A wind speed equal to the line is not marked.

One crossing is enough. Using the illustration above, 0.35 inches of precipitation is not marked, and 0.36 inches is marked. A maximum of 84.0°F is not marked, and 84.1°F is marked. A minimum of 35.0°F is not marked, and 34.9°F is marked.

A missing element does not mark the day, and it does not block a mark from an element that is present. A day can be wet and still unmarked, because the question is whether that day passed the historical line, not whether the crew lost time. Work can stop on an unmarked day. A marked day can also be a day the crew kept working. The table does not record either fact.

What the PDF contains

After a $199 payment you download one PDF. For the location and dates you enter, it names the GHCN-Daily station, the station id, the station coordinates, the distance from the project, and the station record's data-coverage figure. It lists the baseline years actually used, then the mean, the standard deviation, and the threshold for maximum temperature, minimum temperature, and precipitation. Wind is in that table only when the wind test above passed. Otherwise the wind row says it is not reported.

The rest of the PDF is a day-by-day table of the readings, with the days that crossed a line marked, a count of those days, a count of days with incomplete data, a short methodology note, and a limitations note. The methodology note states the same rule: a day is marked when it passes mean plus one standard deviation, or, for minimum temperature, when it falls below mean minus one standard deviation.

This is general information, not legal advice. The PDF compares station readings with the baseline above. It does not determine the cause of a delay.