Expert: Donald Rosenfeld - 4/24/2008

Question
Hi there,
Back when I was in high school in one of my math classes, a teacher taught us how to very roughly predict rain based on the the weather of previous days (for example, if it's sunny today, and it rained yesterday, there's an x percent chance of rain tomorrow and a y percent chance of no rain tomorrow). This may not be exactly how she explained it, but it's the best my memory can do. My question is: is there a way to roughly predict rain for any given day of the year using some sort of mathematical model? I'd be very curious to know. Thanks in advance!

Answer
Hi CM

There is a method called persistence which simply means that what happened in the last day or two is likely to persist for the next day or two. Then there is a method called climo which basically has a probability distribution based on past historical data that gives you a percent chance of rain for a day in a certain amount of days- as an example, in April there is a hypothetical historical pattern that every day has a 30% chance of rain which means that of three days one will have measurable precipitation.

As for a formula (equation), you can convert the rules above into equations.

And, is there an equation? See the following:

http://www.nws.noaa.gov/mdl/synop/tpb/487body.htm

http://ams.allenpress.com/archive/1520-0434/14/1/pdf/i1520-0434-14-1-38.pdf

http://ams.allenpress.com/archive/1520-0493/113/11/pdf/i1520-0493-113-11-1837.pd...

http://www.freepatentsonline.com/7321827.html