The National Weather Service provided a one-page summary of PoP (“probability of precipitation”) forecasting today that is a good summary of this statistic.  It is important to keep in mind that forecasts are always made for a specific time period and a specific area over which the PoP forecast is valid.

“Probability of precipitation” is what is generally known as a conditional probability, where two probabilities are combined together to get the final number, and one part of the forecast depends on the other.  In this case, the first part is the probability that rain will occur.  This could either be due to uncertainty in the timing of a front or other rain-maker through the area, or an uncertainty about whether or not rain will occur at all under the given weather pattern.

The second probability considers how much of the area will be covered by rain if it does rain.  “If” makes it conditional.  In the case of a front, we might expect that almost everywhere would get rain once it makes it to the forecast area.  In the case of scattered summer showers, the area covered by the rain would be lower, since some areas would likely get missed.

In practice, PoP forecasts are not well understood, even by meteorologists.  My husband John Knox, a professor in the Geography Department, with the help of some of his students, has collected definitions of PoP forecasts from a number of sources.  The variety of explanations of PoP forecasts varies quite a bit from one group to another.  In addition, some groups also consider their audience when they provide PoP forecasts.  I’ve been told that the Weather Channel will put in a 20% forecast if there is any chance of rain at all, even if the actual areal coverage might be a lot less than that.  They may do this because they would rather have their viewers ready rather than be caught unprepared for a sudden shower, however unlikely.  In a study I did for the Southern Region of the NWS many years ago, I found that the NWS employees I studied at the time most often predicted around 30%, because that gave them the lowest total error on their forecasts if they were not certain of what the weather would do.  They were graded on their overall forecast accuracy so this became a way to hedge their bets.  Since this was many years ago, the NWS probably does not use the same evaluation techniques today, although I don’t know that for sure.

PoP forecasts don’t tell you how much it will rain or how long, but they can be useful for determining whether or not to bring an umbrella or a raincoat.

pop forecast slide