They do seem to have improved accuracy by a lot, especially when I think to how it was back when I was a kid. Though once in awhile, they can really screw it up yet.

In my mind, a really key advance in weather forecasting is communication. In the old days, I had to wait until the 6pm weathercast came on the TV, and try to remember everything that was said. Now, I can dial up the *right now* weather, local weather forecast, and even track storms and fronts as they approach me. A forecast isn't going to tell me if I'm about to be hit by a gully-washer or a hailstorm, but online NEXRAD helps a lot.

No question that the science and prediction accuracy of forecasting has improved, but the internet has really empowered me with those predictions.

I like live radar images,

The forecast for yesterday morning was 1" to 2" of snow. We got 8+".

One of our local station's weather guy refers to any given forecast during a major weather change as coming from the "American Model", or this model or that model (all are based on satellite and other factors). Turns out that apparently for N. TX, the most accurate weather model they like is the European Model.

Weird isn't it...

The European model is somewhat more sophisticated, and may be more costly.

https://www.youtube.com/watch?v=Myms0a-bzbw

Where you run the models doesn't really determine what produces the most accurate forecasts.

Jamie, that was a good article -- thanks for posting it. Weather reporting is more important than just knowing if you'll need an umbrella or not! Once or twice a year, the chief meteorologist of our local station gives a 10 hour weather class to seniors. I took it a year ago. It was fascinating to see all the data she and her team pull together. She showed us a bunch of technical websites. My favorite shows wind around the whole globe. Not that it tells me very much, but it's pretty: earth surface wind

I'm a fan of the National Weather Service and the important work they do. More is possible with the greater data and satellite observation availability and the continuing refinement of the many computer models they use and compare results of. But remember they're dealing with something that is almost the poster child for the phrase "variability and unpredictablility". That will never change and will always limit the accuracy of forecasts.

Edit to add - I think for each region, the cumulative knowledge of the meteorologists is of course important but they also develop a feel over time about which models are more indicative when and under what circumstances. It isn't always just one and not the other, as I understand it.

The "feel" was mentioned by one of our night/weekend weather people at our local NWS location. Normally they post what they think will be the forecast but until the person retired or moved to another location several times they mentioned that their "feel" was for more or less gunk than this area received. Especially important when tornados can develop out systems that may not show on any weather map. Last major outbreak here was one of those "feel" forecasts.

I think "feel" is a somewhat obsolete skill in local weather forecasting. In the olden days, local weather forecasts were done with a minimum of measured inputs, and experienced local observers knew what to pay attention to, and what to ignore. A lot of what they knew, they knew by looking out the window. But the wealth of physical information that forecasters now have, and one in France will have almost as much about Texas as one in Texas, means that predictive models are vastly better informed and offer vastly better insights than local weather people. Local weather people are now "reporters", rather than "predictors". They don't make forecasts, but just relay the forecasts that come from the national and international teams who are armed with the real data and the computing power. The relevant education, training, and experience is in those teams who manage the models. More power to them. Now, those teams do need to know what to pay attention to, and what to ignore, but the models are becoming intelligent enough that there is some peril in ignoring things that the models were programmed to be smart about. That is, the "feel" goes into the models, making them smart enough not to require "feel".

Yes, Friday morning we got like 5" when Wednesday it was forecasted as nothing much more than an inch. But they can only predict the likely scenarios on how the storm tracks. I don't know if they underestimated on how fluffy the snow would be in cold weather, it was very lightweight snow but they usually have greatly improved in predicting exactly what type the snow or rain will be like, the characteristics of the precipitation. Much advanced over the past than just look what is happening to the West and guessing what it will be like in a couple of days.

I seem to remember back in the 80's into the 90's local news would forecast the worst possible scenario of every snowstorm just to be on the "safe" side, and maybe hopefully scare viewers into watching. I think weather reports are more responsible now with greater accuracy, I think meteorologists have taken higher standards with the advanced accuracy. But weather is always fluid.

The European model seems more accurate in my experience too.