Besides family values, that is?
Both these politicians seem to have a problem with the National Weather Service:
Santorum also accused the weather service’s National Hurricane Center of flubbing its forecasts for Hurricane Katrina’s initial landfall in Florida, despite the days of all-too-prescient warnings the agency had given that the storm would subsequently strike the Gulf Coast.
Governor Cuomo’s attempt to scapegoat the National Weather Service for an inaccurate forecast in advance is not only completely in error—the NWS did an outstanding job—but is a disservice to the public and to the hard-working staff of this federal agency. No forecast of such an historical disaster is going to be absolutely perfect, but no one who lives here can say this event was not well forecast in advance, or that the warning headlines of its impact to come were not well explained in advance…his statement is disinformation, purposeful or ill-informed.
Hey, politicians are politicians, they have to make lots of compromises. But, as a statistician, I’m repulsed by this sort of anti-data attitude coming from either political party.
I have no desire to defend Gov. Cuomo, but if I got the news correctly, the prediction was for 1′ of snow and the reality was 5′. Something you might call Type-M error. Cuomo probably wants to deflect blame for his government low preparedness and push through his project of state weather service, no reason to think his motives are pure. But saying that they predicted the storm while missing the snow accumulation by a factor of 5 is in fact an anti-data attitude. It might be OK to predict an election win with 52% while true result is 60%, but it doesn’t work as well with snowstorms. Just in case it’s not clear: I am not blaming the NWS, for all I know they did the best the current science can do.
D.O.:
According to the link, the National Weather Service wrote, ahead of time, “with feet of snow possible under the lake bands.”
In any case, I’d have been fine with Santorum and Cuomo saying that the weather events were more extreme than predicted. It’s the blaming of the Weather Service that’s uncool.
I find no evidence explicitly stated to either blame nor exonerate the Weather Service. Though, at some threshold of error between the predicted & actual events the weather service should indeed deserve blame, right? Maybe not for any one event but if there’s a trend. As D.O. writes, predicting 1′ & getting 5′ doesn’t exactly sound like a feather in their cap.
The prevailing mood seems to be that the Weather Service is a bunch of hard working, intelligent, scientific guys & hence is above reproach. That seems wrong.
Rahul:
That’s delightfully contrarian of you but you’ll have to take it up with everyone but Santorum and Cuomo. From the Santorum link:
From the Cuomo link:
And more:
Of course no one is above reproach. But given all the above, it’s ridiculous and offensive for Cuomo to characterize this as “when the weather detection system is off, you don’t know a storm is coming, you don’t have a chance to prepare.”
For all I know the NWS may be doing a fantastic job. All I’m asking for is a more data-based assessment of their predictive accuracy.
Unfortunately the general coverage is mostly lacking in any sort of systematic metric of performance. And has lots of emotional appeals from either side: either vilifying the forecasters or praising them.
Politicians apart, what’s a good way to assess the performance of the National Weather Service? I’m curious. What’d be the benchmark anyways. It’s own accuracy in earlier years?
Does anyone do such an assessment regularly? Maybe it’d be interesting checking the relative forecasting accuracies of several national / private forecasters.
Dunno about the USA, but such comprehensive statistical analysis is certainly standard in the UK and shows a steady trend of improvement in predictive performance over as many years as they’ve been doing it. As you would expect from increases in computer power, scientific understanding, and observations.
> But, as a statistician, I’m repulsed by this sort of anti-data attitude coming from either political party.
Santorum is representative of a party that has data problem. Cuomo’s just a d*****bag.
Rahul asked:
“Politicians apart, what’s a good way to assess the performance of the National Weather Service? I’m curious. What’d be the benchmark anyways. It’s own accuracy in earlier years?
Does anyone do such an assessment regularly? Maybe it’d be interesting checking the relative forecasting accuracies of several national / private forecasters.”
Yes, they do. the Weather Service is constantly interested in assessing its performance. The term of art is “skill”, which compares forecast results with deviations from long-term averages. In other words, a forecaster requires no skill to simply forecast the long-term average. You have to do better.
However, I have noticed that a number of papers talk about the better skill values obtained by their work, while the real-life significance seemed small to me.
Any idea, how much has been the improvement over time, if any? Have forecasts got significantly better, say, over the last 20 years?
Some relevant plots:
http://www.randalolson.com/wp-content/uploads/weather-forecast-accuracy-flipped.png
The British Met office actually claims that their 3-day forecast today is more accurate than a 1-day forecast from 1980.
http://www.metoffice.gov.uk/about-us/who/accuracy/forecasts
Yes, the major advance has been the number of days forward that a good forecast makes. The NWS is not as accurate as the European Center for Medium Range Forecasts, because the later has much more computational capacity. All it takes is money.
Rahul OTOH simply is both ignorant and aggressive in claiming his ignorance is the NWS’s problem. Eli can help try Google Scholar with the search term weather forecasting skill
You might be interested in this article in the Boston Globe about fish counts and the governor-elect butting in to say maybe we need to revisit the numbers. Because fishing is the remaining iconic industry here, though there are few actual fishermen. The article itself isn’t terrible but it does the typical give equal time to both sides even when one side is significantly less deserving. When in doubt, question the numbers and the ability of the scientists and professionals involved.
http://www.bostonglobe.com/metro/2014/11/21/how-does-government-count-fish/aCD6gzA2EYdIXMkWElFmHL/story.html?comments=all&sort=OLDEST_CREATE_DT
Urgh. Yeah, screw the catch limits. What do a bunch of pencil-pushing bureaucrats know about fishing? Let fishermen fish! What could possibly go wrong?
Aside from things we’d rather not think about, that is. From “The collapse of the Canadian Newfoundland cod fishery”, http://www.greenpeace.org/international/en/campaigns/oceans/seafood/understanding-the-problem/overfishing-history/cod-fishery-canadian/ –
“On 1 July, 1992 Canada celebrated its 125th anniversary, but the then fisheries minister John Crosbie was not enjoying the party. He had just been confronted by a crowd of hostile cod fishermen in Bay Bulls, a small community in Newfoundland. They had demanded to know why there were no fish left in their waters, and what he was going to do about it. Crosbie angrily responded, “There’s no need to abuse me, I didn’t take the fish from the Goddamn water.” The next day, under police protection, Crosbie announced a two-year moratorium on cod fishing in St. John’s, the Newfoundland capital. A mob of angry fishermen had tried to force their way into the hotel ballroom where he was making his speech, and Crosbie was forced to make a hasty retreat.”
See also –
http://en.wikipedia.org/wiki/Collapse_of_the_Atlantic_northwest_cod_fishery
http://www.heritage.nf.ca/society/moratorium.html
But, at least, no one could call John Crosby a mealy-mmouthed politician. :)