Simple enough. Surely that’s a statistic that NASA or the Air Force has lying around somewhere. So I do some digging and come up with…
Nothing.
Nada.
Zip.
Zilch.
Now maybe it was my Google-fu that failed me, and the data is sitting on a backwater server somewhere collecting dust. The best information I could find was the Spaceport Weather Data Archive which has, among other nifty things, a historical log of lightning strikes in the area of the shuttle pads. So I passed the link along and went on my merry way.
It bugged me though, that of all the data NASA tracks and distributes, historical launch probabilities are not among them. Even if the only use was to satisfy the curious, I’d pay 3 3/4 Empire State Buildings* for it. So I decided to do it myself.
I began with the most visible and most well attended launches out of the Eastern Range: the space shuttle. I started working backward through Wikipedia. I had seen that the entry for STS-135 included a launch attempt table. Unfortunately, in true Wikipedia fashion, there is no consistency among articles on similar topics. After a few missions, these handy tables petered out and I was down to digging scrub dates out of sparse paragraphs. Then I got lucky while double checking a reference: Bill Harwood at Spaceflightnow.com had compiled information for every launch up to the final flight of Columbia. It was all over but the number crunching.
The verdict? As I had expected, April is the worst month to launch a shuttle. The spring weather patterns in Florida are against it. The next worst was December. This surprised me. I had expected April and May to be close together. December, on the other hand, tends to be a fairly benign month for weather. So what’s going on?
First of all, despite there being 135 launches to work with, this isn’t actually a lot of data to try to see trends over a 30 year period. April is tied for the most launch attempts in a month at 29, while NASA only tried to launch a shuttle 13 times in May. Of the 29 April attempts, ten were scrubbed due to weather, but ten were outrightly successful on-time launches. December only had 6 weather scrubs, of 18 total attempts. Ideally, NASA would have attempted launches an equal number of times per month to make my math easy. This would have allowed a pattern to jump out. Reality, alas, is another matter.
Second, the rules governing a launch are ever-changing, and depend on what vehicle is flying and what its mission is. The shuttle had very stringent launch requirements. Not only did the weather at the pad have to be acceptable, the weather at the runway 7 miles away had to be good in case of an abort. Additionally, the weather at at least one of the Trans-Atlantic Landing sites had to be OK!** (Yes, the rain in Spain could ground the space plane.) Early in the shuttle program launch windows were typically long enough that a popup thunderstorm could be waited out, but as NASA focused on building the International Space Station those hours long windows dropped to ten minutes, increasing the likelihood of a scrub.
Increasing scientific knowledge has also changed the rules. It’s been known since Apollo 12 was struck twice during liftoff that lightning could be dangerous. The rule then simply became “don’t launch if bad weather is 10 miles away.” Finally, after a Navy satellite was lost after Atlas-Centaur 67 was struck and subsequently destroyed in 1987, major work went into understanding the interaction between lightning and rockets. This has let launches go ahead when, to a layman observer, the weather looked poor but was actually safe. Those in attendance for the final two shuttle launches and their attendant cloud decks understand this quite well.
In order to fully answer Emily’s question it would be helpful to add data from each launch, manned and unmanned, out of Cape Canaveral over the last 50-plus years. Unfortunately that data is scattered, if it even exists for some of the earlier launches. Now, at least, it’s tabulated for shuttle.
I'm including the spreadsheet I used to generate the above results here. Have a look. Chew on it. And, hey, let me know if you happen to have any spark (heh heh) of insight you'd like to add!
----------
*Comparative size of the VAB.
**In the data, I classified scrubs due to poor TAL site weather as “technical” since it’s a scrub due to an idiosyncrasy of the shuttle program, not a reflection of weather at the launch site interfering with an attempt.
