A scheduler building the baseline for a job that starts in spring has to decide how many days the crew will lose to weather each month. Three authors at PMA Consultants write that US contracts commonly leave normal adverse weather with the contractor and grant time only for weather worse than normal. That makes the scheduler’s monthly number the baseline both parties measure against later.
The U.S. Army Corps of Engineers writes that baseline into its contracts as a table of anticipated weather delay days for each month, built from National Oceanic and Atmospheric Administration (NOAA) or similar data. NOAA’s National Centers for Environmental Information (NCEI) publishes the counts that make the table quick to build. Its 1991-2020 U.S. Climate Normals, which are 30-year averages first released in May 2021, give the average number of days per month with 0.10 inch or more of precipitation at more than 15,000 stations.
May averages the most rain days at Chicago O'Hare
Average days per month with 0.10 inch or more of precipitation, 1991-2020 normals
How the Corps clause works
The Corps sets its method out in Engineer Regulation 415-1-15, Construction Time Extensions for Weather, dated October 31, 1989. It defines adverse weather as conditions at a definite time and place that are unfavorable to construction. Unusually severe weather is weather more severe than the adverse weather anticipated for the season or location.
The contract clause carries a 12-month table of anticipated adverse weather delay work days, based on a five-day week. The contractor’s progress schedule has to carry those days in every weather-dependent activity. During the job, the contractor records adverse weather and its effect on the daily quality control report. A day counts as an actual weather delay day when weather stops work on critical activities for 50 percent or more of the scheduled work day.
When a month’s actual delay days exceed the table, the month had unusually severe weather. The regulation says in capital letters that this finding does not by itself earn a time extension for the difference. The contracting officer first checks the progress schedule to see whether the weather delayed work critical to completion, then converts qualifying delays to calendar days and issues a modification at least quarterly.
Pick the threshold
Appendix B of the regulation starts from the scope. The scheduler chooses the weather that could stop the work, such as temperature, precipitation or wind, and takes a numerical limit for each from the specifications. When two phases share a parameter, the regulation uses the less severe limit. Its example is roofing that stops below 40 degrees and concrete that stops below 32, where the threshold becomes 40 unless roofing is minor to the schedule.
For rain, the PMA authors report general agreement among practitioners that 0.1 inch or more during a work day interrupts outdoor work. They add that 1 inch or more is likely to cost the following work day for drying, snow removal or dewatering. They cite the Federal Highway Administration’s Eastern Federal Lands specification, which defines a rain day at 0.10 inch and a drying day as the work day after 1.00 inch or more. NOAA’s monthly normals report both counts directly, in tenths of a day.
Freezing temperature behaves differently. At O’Hare the low reaches 32°F or below on 28.1 days of an average January, according to NOAA. A limit that is met on nearly every day of a month points to winter protection or a different sequence, and it makes a poor delay allowance. The PMA authors also warn against counting one lost day twice, so a day with both rain and high wind counts once.
A worked example
The figures below are an example built on NOAA’s normals for Chicago O’Hare International Airport. A general contractor is scheduling a sitework and foundation package near Chicago on a five-day week. The scheduler defines a lost day as one with 0.10 inch or more of precipitation and adds a drying day for each day with 1.00 inch or more.
- Download the station’s monthly normals file from NCEI and read two columns: MLY-PRCP-AVGNDS-GE010HI, the days with 0.10 inch or more, and MLY-PRCP-AVGNDS-GE100HI, the days with 1.00 inch or more.
- Add them to get calendar days lost.
- Multiply by 5/7 to convert to work days. The PMA authors describe this weekly factor for a five-day schedule.
- Round to whole days, with halves rounded up, because the Corps table uses whole days.
| Month | Rain days (0.10 in. or more) | Drying days (1.00 in. or more) | Calendar days lost | Work days lost (x 5/7) | Work days in the schedule |
|---|---|---|---|---|---|
| Jan | 5.2 | 0.1 | 5.3 | 3.8 | 4 |
| Feb | 4.7 | 0.3 | 5.0 | 3.6 | 4 |
| Mar | 5.5 | 0.4 | 5.9 | 4.2 | 4 |
| Apr | 7.1 | 0.8 | 7.9 | 5.6 | 6 |
| May | 8.1 | 1.1 | 9.2 | 6.6 | 7 |
| Jun | 6.7 | 1.2 | 7.9 | 5.6 | 6 |
| Jul | 5.8 | 0.9 | 6.7 | 4.8 | 5 |
| Aug | 6.5 | 1.2 | 7.7 | 5.5 | 6 |
| Sep | 5.6 | 0.8 | 6.4 | 4.6 | 5 |
| Oct | 6.3 | 0.7 | 7.0 | 5.0 | 5 |
| Nov | 5.3 | 0.5 | 5.8 | 4.1 | 4 |
| Dec | 4.8 | 0.3 | 5.1 | 3.6 | 4 |
| Year | 71.6 | 8.3 | 79.9 | 57.1 | 60 |
O’Hare averages 79.9 calendar days a year that meet this definition. That comes to 57.1 work days before rounding and 60 after, since rounding each month adds a little. May carries the largest allowance at 7 work days. January and February carry 4 each, because O’Hare’s normal precipitation in those months is under 2 inches, compared with 4.49 inches in May.
The regulation asks for the last 10 consecutive years of data where possible and an annual review of the baseline. Normals are 30-year averages that end in 2020. NCEI also publishes 15-year normals for 2006-2020, and it reports that the eastern two-thirds of the contiguous U.S. was wetter in 1991-2020 than in 1981-2010. Where a contract names the data period or the station, the contract governs, and the normals serve as a check at bid time.
Track it month by month
The second table is also an example. The same package runs from April through July with an allowance of 24 work days. The superintendent’s daily logs record 27 lost days.
| Month | Anticipated work days lost | Actual work days lost (example) | Days above the allowance |
|---|---|---|---|
| Apr | 6 | 5 | 0 |
| May | 7 | 11 | 4 |
| Jun | 6 | 4 | 0 |
| Jul | 5 | 7 | 2 |
| Total | 24 | 27 | 6 |
The regulation tests each month on its own, so April and June’s unused days do not cover May. The excess is 4 days in May and 2 in July, a total of 6 work days. Those 6 days support a request for time only where the schedule shows the weather held up critical work. The PMA authors note that the usual practice takes no deduction from the contractor when a month’s weather comes in better than the allowance.
Every excused day has a dollar value on a job with liquidated damages, which the liquidated damages post covers. The float post shows how to watch the critical path across updates, which is the test the Corps applies before it grants weather time.
How to start on your jobs
Pick one active job with weather-sensitive work on the critical path. Find the nearest NOAA station whose normals carry the S completeness flag, which NCEI uses for records with 24 or more years of data. Build the 12-month table from its two precipitation columns and compare it with the weather allowance in the contract or the baseline schedule.
Then add one field to the daily log. Each day the superintendent records whether weather stopped critical work for half the scheduled day or more, so the monthly count matches the definition in the contract.