The Weatherman

[martyp]

The 5-days-out accuracy has now reached where 3-days-out had been about 20 years ago. . . . However, any human forecaster who simply goes by what a computer suggests rarely performs as well as one who knows the local tendencies.

This has been my experience in the San Francisco Bay Area. It's pretty accurate 5-7 days out and gets refined for local microclimates by local forecasters.

 

[Nords]

Edit add: Maybe Nords will pipe up. IIRC his spouse was in the Navy weather bureau.

My spouse is a Navy-trained meteorologist and oceanographer, with master's degrees in both subjects. (For you Navy & Air Force vets: "METOC".) She was also the Joint Typhoon Warning Center Operations Officer for three years, which is almost as thankless a task as being a submarine operations officer.

I've been learning from her for over 30 years, and I'm still learning. I have learned that when she's worried about a weather pattern, then I'd better start worrying too. And if she says "Holy cow, look at the eye on that satellite image!" then I just pull out the hurricane checklist.

You are perhaps a victim of selective memory. As has been mentioned by other posters, you'll remember the mistakes far more clearly than all the days that they got it right. And you don't even know about all of the other locations where they also got it right.

My first recommendation: stop watching the TV station's meteorologist. Some of them aren't even certified meteorologists ("We're working on it!") and anyway they're just reading the National Weather Service forecast. They barely even bother to do their own forecasts, as if the station's lawyers would allow that potential liability.

Second, if you're worried about freezing your sewage lines then stop pushing the envelope on the season. Winterize as soon as you get an indication that you're heading for a freeze in another week or two, not just at the very last minute. You're simply putting your sewage lines at risk through overconfidence in an inexact science. The military spends bezillions of dollars on meteorologists and equipment but still takes prudent avoidance measures as early as possible.

The most accurate weather forecast is a concept called "persistence". "Persistence" says that tomorrow's weather is going to be the same as today's weather. That's all it says, but it's over 50% accurate all year 'round. (In Hawaii it's accurate for about 320 days out of 365.) Every professional NWS meteorologist curses persistence while secretly wishing that they were as good.

Finally, your best RV forecast is going to come from your local NWS office's website nearest to your RV. They're familiar with regional differences that could be worse than the national forecast.

NOAA is perpetually short of satellites because they're expensive-- and because they occasionally blow up on the launch pad or die an early death due to some space-related casualty. If there's no satellite coverage then you need aircraft overhead with sensors or someone on the ground launching weather balloons... otherwise the forecast really sucks.

Hurricane forecasting has improved tremendously in the last decade, due to both improved models and faster computing power. I forget the precise degree of the improvement but they're going out past five days, and the error cone is less than half of the size that it used to be. Of course the Hurricane Hunters help a lot, although even us submariners think that those aviators are nuts.

Tornado forecasting is very very hard, and if you have too many false alarms then people start ignoring the warnings. The best improvement in tornado forecasting has been Doppler radar (airports and meteorological stations) and faster warning systems (like text messages). In other words, it hasn't improved much in the last decade.

Oh, good, I've saved my Broadside cartoon of the "Combat METOC". Jeff Bacon was a Navy METOC but still managed to make O-6 rank...

 

[ls99]

Thanks Nords. Pretty much along the lines of the old and not bold Alaskan pilots. Don't push the envelope.

 

[athena53]

I just find percentages annoying. I'm an actuary. I know about uncertainty and standard deviations and all that, but "30% chance of precipitation" doesn't tell me much. If it does rain, well, they said it might. Does anyone keep track of what % of the time rain occurs, to what the predicted % was? (EG out of 100 days when they predicted a 30% chance, how many days did it actually rain?)

I guess I'm still disgruntled from earlier this year when I had a 35-mile charity bike ride. They predicted good weather right up to the night before. When I drove to the event I thought I saw some threatening clouds. The rain started just as we got out on the road and never stopped. I turned back early- still managed 24 miles- but was miserable. Got into my car and turned the heat on full blast and it still took me about an hour to warm up after a hot shower and dry clothes.

 

[H2ODude]

Excellent point re percentages. I actually remember the days when they did NOT use them, and then when they did, I thought what a wonderful way to excuse your errors. It's rare to see anything much more the 70%, even when it's raining outside! They will use 0%, but only when there's a high pressure parked right over the area.

 

[MikeWillRetire]

Best weatherman ever:

 

[imoldernu]

Finally, your best RV forecast is going to come from your local NWS office's website nearest to your RV. They're familiar with regional differences that could be worse than the national forecast..

Thank you for the excellent explanation, and particularly the part about "persistence". Subconsciously it seems, the broad overview of the weather fronts and the movement on the radar maps is a better predictor of what to expect than the little map on the TV that shows a 50 mile radius.

Some of the online weather channels seem better than others, giving forward and backward trends. I do use NOAA... and Accuweather or Underground Weather, but as Nords stated, they all use the same basic National service.

As to the OP... didn't winterize and all is ok.... (neighbor checked)... I guess what set me off on the rant was the fact that there are so many RV's and no weather station that I've seen has ever acknowledged the problem, and a Google search of RV sites that discuss freezing ends in no consensus.

So the temperature DID go down to 28 degrees for a few hours.. neighbor checked... all ok. It takes me a few hours to clear the lines, water heater, drains, outside hoses etc...and pump in the RV antifreeze, plus an extra several hours travel, so it's not just a five minute deal...
Stayed home, visited with the kids, and now can maybe squeeze an extra week or so at camp.

 

[Redbugdave]

I just find percentages annoying. I'm an actuary. I know about uncertainty and standard deviations and all that, but "30% chance of precipitation" doesn't tell me much. If it does rain, well, they said it might. Does anyone keep track of what % of the time rain occurs, to what the predicted % was? (EG out of 100 days when they predicted a 30% chance, how many days did it actually rain?)

Most people do not interpret rain percentages the right way. For a forecast of 30% this afternoon...that means for a particular area, (whatever counties are listed), and for a particular time frame, (this afternoon), you will see measurable rainfall, (0.01 inch or more), over 30% of that area. That also means that 70% of the area will not see any rainfall. 70% means 70% of an area will get rain and 30% of the area will not. You don't know exactly where in the area given will get the rain and the rain could be only be a small amount such as 0.01". Using the baseline of 0.01" precipitation for an area is the quantifying factor for verification on a forecast.

 

[timo2]

George Carlin 'The Hippy Dippy Weatherman' - YouTube

 

[Walt34]

Most people do not interpret rain percentages the right way. For a forecast of 30% this afternoon...that means for a particular area, (whatever counties are listed), and for a particular time frame, (this afternoon), you will see measurable rainfall, (0.01 inch or more), over 30% of that area. That also means that 70% of the area will not see any rainfall. 70% means 70% of an area will get rain and 30% of the area will not. You don't know exactly where in the area given will get the rain and the rain could be only be a small amount such as 0.01". Using the baseline of 0.01" precipitation for an area is the quantifying factor for verification on a forecast.

Interesting. I had always interpreted "a 30% chance of rain" to mean that "under these weather conditions 30% of the time it rains and 70% of the time it does not". But perhaps that's just another way of saying what you did.

The distinction matters a great deal if you're riding a bicycle or motorcycle.

 

[ls99]

Funny story. As a Kayaker, I do keep and eye on weather for two or three days out.
Rain is a minor consideration, wind is a major.

I have gotten thoroughly drenched when 20% rain was predicted, and I foolishly left my raingear in the suburban. I now better understand percentages based on Redbugdave's explanation. Like Walt34 I had misunderstood what it means.

Now I will have to try and guess in a 20% forecast, which 20% of the area is gonna get it.

Edit add: Some lines from my days in Alaska on weather.

Q: so whats it gonna be tomorrow?
A: There will be weather.

Stuck on mountaintop, waiting for the cloud, er I mean the fog to diasappear.
Cheif scientist: I think it is lifting.
me: why?
Chief scientist: It can't get anny lower.

What is the difference between a cloud and fog?

A: Depends, are you in it, or looking at it from a distance.

 

[LOL!]

I am unconcerned about percentages for rain for a few reasons.

1. When they say chance of rain is 40% in our area (South Texas), that just means that somewhere it will be raining, but that doesn't mean it will rain on me. It might even mean that 40% of the area will get some rain or sprinkles.

2. I can look at animated radar and see exactly where it is it raining and project whether the rain or front or whatever will impact my location and when. That is, I can do my own weather report whenever I want to. Before I walk the dog, I know if it will be raining 10 minutes from now or 30 minutes from now. I can even see when there will be a 30 minute zone of no rain for a quick outing between rain clouds.

Sure there are forecasts that say it will rain on Tuesday and Wednesday of this week in my area. Once again, it may not rain at my house or in my neighborhood, but it will probably rain in my county. I will know on Tuesday if it will be raining at my house and that is good enough for me.

Bottom line: Become your own weather person. Look at animated radar.

 

[Redbugdave]

Lol...That's what I do too.

The NWS came out with a product called the NowCast. It was supposed to address what you just described as what you are doing, now. A short term forecast tells what was happening where and where it is going to move in the next half hour or so. But, the NWS did not follow up with it's product and it has pretty much gone by the wayside. Some offices may do a good job with it, I don't know. Most do not.

My common sense way of interpreting the forecasts:
First off...Go ahead and make your plans since the forecasts change big time beyond two days out. In good weather...no rain, etc...The forecast you see may be good out to two days. Beyond that I don't hold much value. If rain is forecast beyond day two...add a day or so to the event since it seems the forecast brings it in too soon most of the time. For bad weather like thunderstorms, about 15 or 20 minutes is about the accurate length of time the forecast is good for. Most warnings will go out almost an hour, but severe thunderstorms are in constant flux and changes by the minute. So, you need to look at the radar yourself to figure if you are in trouble or not.

If the forecast is a good one...the human Weatherman gets the credit. If it's a bad forecast the human Weatherman simply says the models were bad and changes the subject. Many of the TV Weather people are simply just rereading the NWS forecast and most not meteorologists. By not giving a quantifying figure like percentages they are off the hook for any accuracy. Many of the TV people get their foot in the door broadcasting the weather before moving onto sports or whatever pays the best or there is a slot. Here today gone tomorrow with the way their contracts are written.

I know this sounds gritty...but that's my opinion and I'm sticking with it.

 

[davismills]

I feel your pain, Imoldernu. I had to rush out Saturday and winterize my motor home as well. The weather was forecast with a freeze warning and I'm terrified of the pipes bursting. To make matters worse, both house batteries were dead as well. Long day!!!

 

[DUFUS]

What's to be expected from a group of people that instead of saying it could rain tonight, they say there is a "possibility of a precipitation event occurring during the overnight hours"?
If I wanted to be exposed to that kind of mufti-syllabic masturbation I wouldn't have ER'd. I 'd have stayed at Megacorp and attended more meetings with upper level management, or sneaked into the executive washroom more often.

Now back to the news desk for some awkward banter...

 

[Fred123]

Most people do not interpret rain percentages the right way. For a forecast of 30% this afternoon...that means for a particular area, (whatever counties are listed), and for a particular time frame, (this afternoon), you will see measurable rainfall, (0.01 inch or more), over 30% of that area. That also means that 70% of the area will not see any rainfall. 70% means 70% of an area will get rain and 30% of the area will not. You don't know exactly where in the area given will get the rain and the rain could be only be a small amount such as 0.01". Using the baseline of 0.01" precipitation for an area is the quantifying factor for verification on a forecast.

I don't believe this is correct; the rain percentage is the chance that rain will occur at any given point in the forecast area. See the NOAA FAQ for more details.

 

[Walt34]

I don't believe this is correct; the rain percentage is the chance that rain will occur at any given point in the forecast area. See the NOAA FAQ for more details.

I got a "link not found" error.

 

[Lsbcal]

It is clear that quantum mechanical effects have been ignored in current weather predictions. Also what about the Heisenberg Uncertainty Principle?

Before I go out running I always calculate this stuff as a correction factor. If you don't you are at risk.

 

[aja8888]

Here's a really nice weather site with lots of features:

Beautiful Weather Graphs and Maps - WeatherSpark

I plugged in Egypt, Texas and used the dashboard as an example.

 

[Fred123]

I got a "link not found" error.

Operator error. Try here.

 

[ls99]

It is clear that quantum mechanical effects have been ignored in current weather predictions. Also what about the Heisenberg Uncertainty Principle?

Before I go out running I always calculate this stuff as a correction factor. If you don't you are at risk.

The reuslts must me amalgamated with the butterfly effect.

 

[Walt34]

It is clear that quantum mechanical effects have been ignored in current weather predictions. Also what about the Heisenberg Uncertainty Principle?

Before I go out running I always calculate this stuff as a correction factor. If you don't you are at risk.

The reuslts must me amalgamated with the butterfly effect.

Before I go out on a motorcycle ride I just look out the window.