You're previewing the Bennett demo family. Numbers, schedule, and kids are sample data.Start your own
Loading page…
"70% chance of rain." You have read that a thousand times, and there is a good chance you have never once read it correctly. It does not mean 70% of the area. It does not mean 70% of the day.
You open the weather app. It says 70%.
So — what did it just tell you?
Most people, asked to explain it, give one of two answers. Both are wrong, and the National Weather Service says so directly on its own pages. Getting this right takes about four minutes and changes how you read a forecast for the rest of your life.
Wrong answer one: "rain will fall over 70% of the map." People picture the forecast area shaded in, seven tenths blue. That's not it.
Wrong answer two: "it will rain for 70% of the day." People picture the clock instead of the map — about seventeen hours out of twenty-four. That's not it either.
Hold on to both of those. We're coming back for them.
The number is called the probability of precipitation, and forecasters shorten it to PoP. Here is what it is defined to mean:
Pick any single spot in the forecast area — your street, your school gate, one particular garden. The number is the chance that that one spot gets at least a measurable amount of rain during the forecast period.
Not the region. Not the afternoon. A point.
70% chance of rain = at any one point in the area, roughly a 70 in 100 chance of getting at least 0.01 inch of rain in that period.
That 0.01 inch matters more than it looks. It is the smallest amount an ordinary rain gauge can actually record — the official line between "it rained" and "it didn't". So the forecast is answering a strictly yes-or-no question about one spot: does the gauge there catch anything at all?
It says nothing whatsoever about how hard, how long, or how much.
The forecaster has to answer two different questions, and then multiply the answers together.
Confidence times coverage. That's the whole rule, and it's the reason the number can't mean either of the two wrong answers — each of those is only one of the two halves, mistaken for the finished product.
Same forecast on your phone all three days. Completely different weather. Press through and watch what the two halves are doing.
Your app says 70% for tomorrow. Which statement is the one forecasters actually mean?
Sort each one. The reasons are where this actually sticks, so read them even when you get it right.
Tap an item, then tap where it belongs
No, they didn't. Not from that alone.
A 70% forecast is expecting to be dry about three times in every ten. A dry day after a 70% forecast is one of the days the forecast already told you to expect. If it never happened, the forecaster would be systematically overstating the chance.
So you cannot judge a chance-based forecast on one day. You can only judge it on a pile of days.
Here's how it's actually done. Collect every day a forecaster said 70%. Count how many of them ended up with measurable rain at the point in question. If it's near 70 out of 100, that forecaster is doing their job well. If it's 95, they were too gloomy. If it's 40, they were too cheerful.
That's it. It takes hundreds of days and it's the only honest scoreboard there is — which is also why arguing with the weather app about a single afternoon is a waste of breath.
You keep a notebook. Over one year, your forecaster said 70% on 100 days — and rain turned up at your house on 71 of them. What does that tell you?
There's a smaller misreading worth catching too.
The forecast is measured against 0.01 inch — a quarter of a millimetre. That is a very low bar. A shower that dampens the pavement and stops clears it.
So a day forecast at 70% and delivering ninety seconds of drizzle has, officially, rained. The forecast was about the event, and the event happened. If what you actually wanted to know was "will my football match be cancelled?", that is a different question, with a different forecast — how much rain, and when — and the percentage was never going to answer it.
Two towns, same forecast area, both told 30%. Town A gets a downpour all afternoon. Town B stays completely dry. Who was let down by the forecast?
You’re previewing as a parent — nothing here is recorded.
Definition: the National Weather Service defines Probability of Precipitation as the likelihood of occurrence, in percent, of a measurable amount of liquid precipitation (or the water equivalent of frozen precipitation) during a specified period at ANY GIVEN POINT in the forecast area. Measurable means 0.01 inch or greater. (NWS glossary; NWS Jackson product guide glossary.) The rule: PoP = C x A, where C is the forecaster's confidence that precipitation will occur somewhere in the forecast area and A is the fraction of the area expected to receive measurable precipitation if it does. NWS worked examples: 80% certain of rain developing but covering only 50% of the area gives a 40% forecast; a widespread area of precipitation with 100% coverage that the forecaster is only 40% certain will reach the area also gives 40%. (National Weather Service, "What Does Probability of Precipitation Mean?", weather.gov/lmk/pops; NWS Paducah, "Precipitation Probability", weather.gov/media/pah/WeatherEducation/pop.pdf.) The two misreadings named in this episode are named as misreadings by the NWS itself: a 40% probability does NOT mean 40% of the area will be covered by precipitation, and does NOT mean precipitation will be falling 40% of the time. (Same NWS sources; also NWS Peachtree City, weather.gov/ffc/pop.) The three days in the worked table are illustrations of the C x A rule using the NWS's own worked-example form, not records of particular real days. Judging a chance-based forecast by comparing the forecast share against the share that actually occurred, across many days rather than one, is standard forecast verification practice.