predicting the weather..

[thatguy]

I have a possibly stupid question. I want some help understanding weather predictions.

lets say we have a 30% chance of rain/snow - Does that mean it will rain/snow for 30% of the day or just we about 1/3 chance of rain/snow?

I am guessing the percentage has nothing to do with the length of the time it rains or snow.. Just the chance that it will rain/snow..

thanks

Brian

 

[Bird]

Brian, I don't really know for sure myself, but I used to think 50% chance of rain meant it was even odds that it would or would not. However, I have been told that it means 50% of the forecast area will receive rain.

And right now, the forecast is for a 10% chance we'll get a light snow tomorrow.

 

[CharlieTR]

I heard a weather man mention this many years ago and I was sort of suprised to hear what "chance of ..." actually means.

Example: 65% chance of rain means in the last recorded 100 days with the exact weather conditions it rained 65 days of those 100 days.

 

[avorancher]

I thought your question was interesting so I did a lot of internet searching. Turns out that there are many different opinions on this. The link below seems to be the one supported most by "weather-related" postings and not just people in discussion groups.

Listener Question-Forecasting Percentage

I think the "% chance" terminology they use today started when computer-forecasting became popular. I also thought it was weird when I first heard it being used. To me, it seems like 100% chance means that there IS a chance.

 

[Amontilado]

Folks,

The way the NWS defines a "chance of precipitation" is the possibility of a measurable amount of precipitation being recorded, whether it is rain (0.01"+), snow(0.1"+), ice(0.1"+), etc. A lot of the guidance we use is statistical in nature, in that, when a computer model had these conditions over an area, this is the average temperature, high or low temperature, rain coverage, etc that was recorded at a measuring point (usually an airport) that records the weather (these conditions gave a measurable amount of precip xx% of the time, on average).

Steve

 

[johnk]

Thats the only Occupation that you can use all the subjunctives of possible, partly, Potential, mostly, could, and have all the bases covered. If an eartquake occured they would say the potential was there and it could possibly happen again tomorrow. If it doesn't everyones relieved. If it does the weatherman is a hero..........How much they Make???? $$$$$$$$. If I used all the phrases they use in my occupation I'd be fired. Whaen the boss asks if I did it a certain way and I said possibly, could have etc I'd be looking for another job...

 

[Bird]

Thats the only Occupation that you can use all the subjunctives of possible, partly, Potential, mostly, could, and have all the bases covered.

Well, you could be a doctor or lawyer, charge big fees, give absolutely no guarantees, be wrong at least as often as the weatherman, and make even more money.

 

[Amontilado]

Our phaseology is dictated by a stack of binders one to two feet thick. Since I'm so used to using specific wording, it irritates the heck out of me when I hear some of the "weather talent" mix their phrases, usually with the effect of trying to cover all the bases, "in case". JMHO.

Steve

 

[mikim]

yeah Steve --- but we won't get into the old met vs. et "who has to be right what % of the time" debate .....I'll let Brian handle that for me. (I'm assuming here that you're at HUN) ....

 

[the_sandman_454]

The only real forecast that can be relied upon is as follows: Statistically there's roughly a 50% chance the weatherman will be right.

If I was a weatherman I'd simply suffice it to say "There is a 100% chance of between 0 and 10 feet of precipitation today". I'd be a hero, being right 99% of the time...

 

[BB_TX]

The way I have come to look at it is this. If they say 50% chance of rain, I assume 50% of the local area will get rain and 50% will not. Watching the radar that seems to generally be the case. He higher percentage stated, the more general the coverage of the rain is. The lower percentage, the more scattered it is.

 

[Bird]

[Statistically there's roughly a 50% chance the weatherman will be right.

We like to joke about it, and the weathermen are wrong much of the time if you are holding them to an exact temperature, but in reality, they're right, or very close most of the time. If they weren't, we all wouldn't be checking with them every day, would we?

 

[SkyPup]

Well, you could be a doctor or lawyer, charge big fees, give absolutely no guarantees, be wrong at least as often as the weatherman, and make even more money.

Or a politician, by definition -> any person skilled in manipulating group opinions in his or her favor....

 

[Amontilado]

Mike,

I'm down at the lower end of the state, near Mobile. Most of the time, I just sit back and see where the "discussion" goes.

Steve

 

[Amontilado]

Bird,
Exactly. Forecasting the weather is a case of what is an acceptable error. A couple of degrees off on temps for a nice, warm day isn't as much of a problem as being a couple of degrees off when the temps near freezing, and the difference can result in rain, freezing rain, snow or sleet or any combo of the above. Those days are the reason I keep a nice big bottle of aspirin here at work....

Steve

 

[johnk]

If you can't see the big storm approaching on the newest Radar screens and calculate the precipitation being dropped in its path you should have to admit your blindness on the weather channel for all to hear....Maybe stand on a snowbank in a T-Shirt and shorts for punishment for the next weathercast...

 

[thatguy]

thanks for all the comments and thoughts..

As a side story - A few years ago the local weather man (not meteorologist) said almost word for word on the 11 oclock news "the National weather service is calling for snow tonite, but I dont see where they are getting that information"...

Fast forward about 7 hours.. Notice how bright the bedroom is at 6am.. Open blinds and stare out at the 7 to 8 inches of new overnight snow.. He lost all of his credibility at that point for telling me what the weather was going to be ..

thanks

Brian

 

[Amontilado]

JohnK,
If it was only that simple. How about figuring how much precip will fall to the nearest 0.01" over an area encompassing several hundred square miles or more and in what form will it fall. Will the temps above the surface be warm enough for that falling snow to melt partially, completely, or not at all? What will the resultant ratio of liquid water to falling snow be (important to figure out how much measurable snow will be on the ground)? Our observing equipment has gotten better over the years, but with 200-500 miles between upper air stations (the places that release the balloons), plus the the fact that not as much info is returned from a radar as the tv people would let you believe (it can't tell if the precip falling is snow, rain, ice pellets, or even hitting the ground, though large hail is quite obvious if your looking for it, above the surface). A lot of what is happening between the RELIABLE observation points is assumed. We depend upon a lot of TRAINED HUMAN observers to help fill in the blanks, but they can't be everywhere, ans some of the public reports we get are less than credible (all wind damage MUST be because of tornadoes, right????).

As I remind some people, if you think you can do it better, give it a try. Trying to figure out the specifics of what a somtimes capricious person of the fairer *** (Mother Nature)is going to bring is that easy (have given up trying to figure out what's going on in my wife's mind ). It drives the wife nuts when I give a range of what could most likely happen in her little piece of space. Forecasting can be one of the most humbling professions I know, and from some of the stories I've heard, down right dangerous

Steve

 

[CRJCaptain]

I have a possibly stupid question. I want some help understanding weather predictions.

lets say we have a 30% chance of rain/snow - Does that mean it will rain/snow for 30% of the day or just we about 1/3 chance of rain/snow?

I am guessing the percentage has nothing to do with the length of the time it rains or snow.. Just the chance that it will rain/snow..

thanks

Brian

The percentage value has to do with the chance/probability that the precip in question will occur.

Keep in mind that these forecasts are generated by massive computer models that are solving the equations of motion for the entire atmosphere, run ahead in time several days. These equations are what are known in the biz as "non-linear partial differential equations". What does this mean? It means that any error or small deviation in the initial conditions will lead to a large divergence in the model's forecast solution. So in a way,there is a real limitation in how accurate and how far out in time any forecast will ever be. The inaccuracy is not the fault of the forecaster, but is more a function of trying to mathematically predict an inherently chaotic system.

Dean

 

[Amontilado]

CRJCaptain,
Basically correct.

The percentage value has to do with the chance/probability that the precip in question will occur.

Not quite. The chance of precip is the chance a MEASURABLE amount will fall.
A sprinkle or flurry doesn't count in the scheme, nor a light showers where only a TRACE amount (a non-measurable amount) occurs.

Keep in mind that these forecasts are generated by massive computer models that are solving the equations of motion for the entire atmosphere, run ahead in time several days. These equations are what are known in the biz as "non-linear partial differential equations". What does this mean? It means that any error or small deviation in the initial conditions will lead to a large divergence in the model's forecast solution. So in a way,there is a real limitation in how accurate and how far out in time any forecast will ever be. The inaccuracy is not the fault of the forecaster, but is more a function of trying to mathematically predict an inherently chaotic system.
*
Not quite again. The models, both in physical and statistical output is only guidance. We regularly depart when the guidance doesn't quite have things right. The guidance can only tell so much, with the rest being a lot of research and experience. True on the guidance likely deviating the farther out in time it goes. One of the tricks used is to run the same model equations with slightly different starting conditions. Take all the solutions, average them together, and one gets a better forecast, sort of like everyone chiming in on their best way to fix a problem, which I see so much (thankfully) on this site. Plus there is a LOT of sharing of meteorological data/guidance amongst countries (wish politics was this way ).

On a side note, for those in the deep freeze up north. The chile pepper folks in India have come up with a pepper even hotter than the red savina habenaro. Don't have the Chile Pepper Mag issue in front of me for the name, but this variety came in with a heat rating of near 1,000,000 units on the Scoville Scale. Probably a result of global warming, (or another cause ;-) ).

Steve