On the heels of celebrating 5 years as a podcast, TRC delivers another great episode. Adam takes a look at area 51 in the wake of the CIA admitting it exists, Elan investigates the accuracy of weather reports and Darren complains about Vector cereal.
Download direct: mp3 file
If you like the show, please leave us a review on itunes.
SHOW NOTES
Area 51
Area 51 and its purpose declassified: No UFOs, but lots of U-2 spy planes – NBC News
Weather Forecasting Accuracy
Cracked – 6 Most Statistically Full of Shit Professions
Freakonomics – How Valid are TV Weather Forecasts?
http://weather.slimyhorror.com/ – BBC Analysis
Timeanddate.com – Forecast Accuracy
Forecastadvisor.com – New York
How Stuff Works – Why can’t scientists accurately predict the weather?
How Stuff Works – How far in advance should I check the weather forecast?
Wonderful, wonderful intro song 🙂
Thanks, Matthew!
Three wee comments:
– Solid episode. I especially liked the song and Darren’s smack-down on Vector.
– I was hoping you guys would explicitly mention one aspect that I think is the main reason that weather forecasting is so difficult (as far as I can tell), which is a lack of data. If we fed enough high-quality data into our models, then assuming we have the computing power, we would have near-perfect predictions. We generally know the physical mechanisms at play. Of course, that data’s expensive and difficult to obtain, but I don’t think the problem is usually our models, per se.
As far as precipitation probabilities go, I’ve long-since wondered what those numbers actually mean, and I found this answer from NOAA: http://www.srh.noaa.gov/ffc/?n=pop
– I don’t know how often people tell you this, but I love the outtakes that you guys include!
Loved the intro song!
I’ve seen some French TV stations equivocating their longer term predictions by assigning a confidence factor to the forecast, e.g. 4 out 5 that their two day prediction will be right.
Also, the probability of precipitation is not the result of using the Monte Carlo method. There is a formal formula: http://en.wikipedia.org/wiki/Probability_of_precipitation
Oops, I see that the POP was covered by the previous commenter.
I was waiting for someone to bring this up during the weather discussion, but, unless I missed it, no one did:
Most weather forecasts are for a fairly large area. Down here in North Carolina, a forecast might cover at least a hundred square miles. When it rains, especially in summer, it’s not raining everywhere in that area. So some spots get rain, others don’t.
Just last month, I drove for 45 minutes through a torrential, apocalyptic storm with massive flash floods. Cars were stuck, trees were falling. Near home, I hit the edge of the storm and it stopped raining. Two minutes later, it was completely dry and sunny. Not a drop had fallen. Kids were playing, parents were pushing babies in strollers. It was surreal, like an M. Night Shyamalan movie. What would you call that? Rain? Not rain?
Related to that, forecasts will often say “scattered showers.” Does this mean scattered geographically or temporally (or both)? That it’s raining in some places but not others? Or that it’s raining now, but not later?
P.S. As the guy who said Cristina had a sexy voice, uh, thank you for the birthday episode!