Reading a Forecast for the Work, Not the Weekend
Why this matters
The forecast most people read is a picnic forecast: a high, a low, an icon and "40 percent chance of rain". That is enough to decide whether to carry an umbrella and nowhere near enough to decide whether a coat of paint will cure, a tear-off can be dried in, or a bead of sealant will skin before the shower arrives. Every job that weather can ruin has its own limits, set by the product data sheet or the equipment rating, and each one reads a different column of the same hourly forecast. The skill this card teaches is reading the right column, at the right hour, with the right idea of what the numbers mean.
Get the right forecast first
A point forecast, for the site. Use an hourly forecast for the job address, not the city name. In the US, the National Weather Service publishes hourly point forecasts for any location; many weather apps repackage the same or other model data. A city forecast can miss a valley that runs several degrees colder at dawn, a lakeshore that stays windier, or a hill that gets the shower first.
Note when it was issued. Forecasts are updated through the day. The one you read at 4 pm yesterday is not the one the crew should be working from at 6 am. Read it the afternoon before to set the board, and again at first light to confirm it.
Trust timing less the further out it is. A forecast five days out is reasonable on the broad picture (a front is coming, it will be cold) and weak on the hour it arrives. Two days out the hour is usable, and the morning of the job it is the best you will get. Plan the week on the daily outlook and commit the day on the hourly one; the scheduling side of that is in How to Build Weather Holds Into the Schedule.
Know what each number is measured as. Forecast temperature is air temperature at about head height in the shade. It is not the temperature of a sunny wall, a shaded deck or a roof at dawn, which can be far warmer or colder; see Air, Surface and Material Temperature: Three Different Limits. Forecast wind is for a standard height of about 33 feet (10 m) above open ground, so a roof edge sees roughly the forecast and a sheltered yard often sees less.
What the columns mean
Temperature. The hourly value, not the daily high or low. The daily low happens around sunrise and the high in mid-afternoon; your work happens in between.
Dew point. The temperature at which the air's moisture condenses. When a surface is at or below the dew point, water forms on it. The gap between air temperature and dew point (the spread) tells you how close the air is to wetting things. What Dew Point Predicts covers the idea in full, and the coating rule built on it belongs to The Surface Five Degrees Above Dew Point Rule for Coatings.
Relative humidity. The air's moisture as a percentage of what it could hold at that temperature. It rises as the air cools even when the moisture does not change, which is why it peaks at dawn. Some products, mainly coatings and moisture-cure materials, state humidity limits directly.
Sustained wind and gusts. Sustained is the average speed; gusts are the brief peaks. Equipment and product limits usually say which one they mean, and that distinction is the subject of Sustained Wind Versus Gusts and Which One a Limit Means.
Probability of precipitation (PoP). The most misread number on the page. As the National Weather Service defines it, PoP is the chance that measurable precipitation, at least 0.01 inch, will fall at any given point in the forecast area during the forecast period. A 40 percent PoP for the afternoon does not mean it rains for 40 percent of the afternoon, and it does not mean 40 percent of the area gets wet. It means that at your site there is a 40 percent chance of at least a hundredth of an inch sometime in that period. Watch which period: a daily PoP covers 12 hours, an hourly PoP covers one hour.
Precipitation amount (QPF). The expected amount, usually in hundredths of an inch. PoP tells you whether, QPF tells you how much if it does. A high PoP with 0.01 inch is a sprinkle; a modest PoP with half an inch is a downpour you might miss.
Alerts. A watch means conditions are favorable for a hazard; a warning means it is occurring or imminent; an advisory is a lesser hazard that still matters. Thunder in the forecast at any probability triggers the plan in The Lightning Stop Work SOP, not a judgment call on the roof.
Worked example: one Tuesday, three jobs
A residential contractor has three crews on the board for Tuesday. The hourly point forecast, read at 6 am, is below. The columns each job reads are marked under the table.
HOUR TEMP DEWPT RH WIND GUST POP QPF
F F % mph mph % in
7a 41 38 89 3 6 5 0.00
8a 45 39 79 4 8 5 0.00
9a 50 40 68 6 10 5 0.00
10a 55 41 59 8 14 10 0.00
11a 59 42 53 10 18 10 0.00
12p 62 44 52 12 22 20 0.00
1p 63 47 56 14 26 30 0.01
2p 61 50 67 15 28 50 0.05
3p 58 52 80 12 22 60 0.10
4p 55 52 90 9 16 60 0.08
PAINT reads: TEMP, TEMP minus DEWPT, POP, QPF
ROOF reads: POP, QPF, GUST
CAULK reads: TEMP, POP
Crew 1, exterior paint on a south wall. Say the paint's data sheet reads: air and surface at least 50 F, surface at least 5 F above dew point, and no rain for 4 hours after application. Air reaches 50 F at 9 am, when the spread is 10 F (50 minus 40), and the south wall in the sun will read warmer than the air, so the start condition is met at 9 am. The problem is the back end. The first measurable rain in the forecast is 0.01 inch at 1 pm with a 30 percent PoP, rising to 50 percent at 2 pm. Paint applied at 9 am needs to stay dry until 1 pm. Anything applied after 9 am needs dry weather into the hours where the PoP is 30 to 60 percent. A wall takes this crew most of a day. The paint window on this Tuesday is effectively zero hours of useful work. The crew moves to an interior job, and the paint day goes to the next forecast with a full dry afternoon.
Crew 2, a tear-off on a garage roof. Tearing off exposes the deck, and the crew's rule is that nothing is torn off that cannot be dried in (underlayment fastened over it) before the rain chance reaches 20 percent. That happens at noon. Gusts also climb past 20 mph at noon, which matters for sheets of underlayment being carried and laid. The garage takes about 3 hours to tear off and dry in. Starting at 8 am, when the surface frost is gone and gusts are 8 mph, the crew is dried in by about 11 am, an hour before the PoP reaches 20 percent at noon. Shingling can continue until gusts rise, and the deck is covered either way. If the morning had started an hour late, the honest call would have been to tear off half and dry it in, not to race the whole roof.
Crew 3, exterior caulking at window perimeters. Say the sealant's data sheet requires 40 F or above, and the bead needs a few hours to skin before rain. Air is 41 F at 7 am and rising, so temperature is met from the start, though the crew checks the joint surfaces with an infrared thermometer because a shaded north frame at 7 am can read below the air. The job is 3 hours. Starting at 8 am, the last bead goes in around 11 am, two hours before the first measurable rain chance at 1 pm. If the data sheet's skin time is longer than two hours, the crew starts with the windows on the side that will be finished first and covers the last elevation's work if the shower comes early.
What the three readings have in common. The same forecast said "go" to one crew, "go with a hard stop" to another and "not today" to the third. None of those calls came from the daily icon, which showed a cloud and a raindrop.
When the morning read disagrees with last night's
Say the board was set at 4 pm Monday from a forecast that showed 10 percent at 1 pm and no measurable rain before 4 pm. The 6 am read above has the first measurable rain at 1 pm, three hours earlier. That shift is information on its own: a front or a band of showers is arriving faster than modeled, and fronts that speed up once often keep doing it. When the morning read moves rain earlier, plan to the earlier time and add a margin; when it moves rain later, keep last night's plan and treat the extra hours as slack, not as permission to start a bigger scope.
During the day, the hourly forecast is the wrong tool for the next hour. Radar shows where precipitation is now and which way it is moving, and a crew lead who checks it at the start of each work block sees a shower 20 miles out before the forecast does. The forecast plans the day; the radar runs the last hour of it.
Where people misread it
- Treating a 30 percent PoP as safe. At your site, it means roughly a one-in-three chance of getting wet in that period. For a job that is ruined by any rain, that is a real risk. For a job that only cares about an inch of rain, it barely matters. The job decides, not the number.
- Reading the daily high as working temperature. The high on this Tuesday was 63 F at 1 pm; the crews worked from 41 F to 59 F.
- Ignoring the trend. RH rises from 53 to 90 percent between 11 am and 4 pm, and the spread closes from 17 F (59 minus 42) to 3 F (55 minus 52). That is evening moisture arriving early ahead of rain, and it is what makes a coating applied at noon blush or stay soft into the night.
References
- National Weather Service, definition of probability of precipitation and hourly point forecasts
- Product data sheets for application temperature, dew point, humidity and rain-safe limits
- See related: Reading Weather and Demand for the Week
- See related: Weather as a Work Condition: The Rule for Every Trade
- See related: How to Check Tomorrow's Board Against the Hourly Forecast