Hindsight, Twenty-Twenty

With the report of the investigation into the Columbia (search) disaster to be released in the next few weeks, many are speculating that it will say that the crew of the ship could have been saved, had NASA (search) been more diligent in gathering data and understanding the true situation.

I've written on this subject before, and pointed out that there was the potential here for a flight director's nightmare, and there certainly may have been a subconscious reluctance to look for bad news, because they knew that broken tiles meant a lost Orbiter and dead astronauts. The problem here was not the shuttle design per se, but the fact that our orbital infrastructure is essentially nonexistent, and we sent Columbia off into the wilderness, where a breakdown meant death.

That was the reality, as NASA understood it. Was it "possible" to somehow rescue the Columbia crew, given early understanding of the problem, and a sped-up Atlantis (search) launch, and heroics on orbit?


Was it realistic, or sensible, to make such an attempt?

Almost certainly not, but that's where we have to put astronauts' lives in the balance against, well, other things.

Many have compared this to what happened a third of a century ago, and claim that NASA missed an "Apollo XIII moment." But that's an oversimplistic comparison -- there are many differences between Apollo XIII and Columbia.

First, the obvious one, of course, is that in the case of the latter, Houston didn't know "we have a problem" until the vehicle started to come apart over the western United States. The critique here implies that that's the only difference, and that had they known right after launch, the crew could have been saved, as it was then. This ignores several other significant differences.

Apollo XIII was fortunate enough to have a great deal of margin, in both time and equipment. Had the oxygen tank exploded on the way back from the moon, the crew would have died, regardless of derring do and heroics on the ground. Their vehicle was essentially in good shape, and didn't require repair -- just reconfiguration. And for all of that, it was still a very close thing. While we should be thankful that we could save that crew, it's had a bad side effect of presenting NASA as being omnipotent in the public mind, and capable of overcoming any adversity (despite abundant evidence to the contrary in the three decades since). The fact was, it takes nothing away from the skill and brilliance of actions at mission control to say that they were also damned lucky.

We now know that Columbia was badly broken during launch, and that the fate of the crew and vehicle were sealed once the decision was made to press to main engine cutoff, and orbit. They might have been saved by a trans-Atlantic abort, but there was no time at all to gather the information to make that decision, which occurs during launch itself. There was nothing in the vehicle that could be reconfigured, or duct-taped, that would repair what was a broken primary and vital subsystem (the thermal protection system) -- one that absolutely had to work in order to bring the vehicle back.

Some will argue (and apparently do, according to the article linked) that the vehicle could have been reconfigured to buy more time, but that's not sufficient, because of the next, perhaps even more important difference, because it raises very uncomfortable ethical issues.

Apollo XIII had nothing to lose.

Other than lost sleep and fatique among mission controllers, there was no cost or risk to doing everything possible to bring that command module back from its misbegotten journey. Because all of the resources needed were already (at least in theory) aboard the vehicle, there was no need to send anything else up to it.

What the second guessers are proposing in the case of the Columbia disaster was to hasten the launch of Atlantis. Even they admit that without doing this, the chances of a rescue were probably non-existent. That means that we would have had to launch (and risk) another vehicle (one-third of our remaining fleet, not counting the doomed Columbia) and, at a minimum, another two-person crew.

One of the reasons that we fly shuttles so seldomly is that the turnaround process takes a long time. Since the loss of Challenger, the procedures to do so have become even more stringent, further reducing the flight rate. Prior to Challenger, NASA was still deluding themselves that they might eventually get to a flight rate of a couple dozen per year. Since then, due to increased safety concerns, six in a year is a good year, and many years have been less. Rushing to launch is exactly the opposite of the philosophy of maximizing flight safety, and many might argue that it would in itself be playing Russian roulette with few empty chambers.

So here are the options, assuming that NASA had been as diligent in getting the data as its critics would have had them do.

1) Let them come in as they are, and cross your fingers. This was essentially the option taken, except with fingers in normal configuration (that is, always slightly crossed, shuttle flights being what they are, but digits twisted no more than normal,) because they didn't have the data.

2) Attempt to do an on-orbit repair with available crew and equipment. Despite Apollo XIII fantasies, this was never a realistic option. Even if they had the equipment and materials available (they didn't), it still would have necessitated finger crossing on a planetary scale.

3) Try to extend life support as long as possible, just in case, put together a tile repair kit of some kind on the ground, change the next Atlantis mission to go repair the vehicle, launch it on or close to schedule, and bring it back with a minimal crew. Oh, by the way, this probably results in dead crew, because "as long as possible" isn't long enough, but offers the possibility of returning the vehicle, albeit at the risk of two more astronauts. This is probably the one that's most rational, given the value of the hardware, but would be totally politically unacceptable.

4) The option suggested by the critics: extend life support as long as possible, and plan to launch Atlantis in time to get there before they run out of air. This is the highest risk, because now you're rushing the launch. In other words, we've probably already lost a quarter of our shuttle fleet. Its crew is likely going to die regardless of what's done. But in our determination to save them, we're going to risk a third of the remaining fleet and more astronauts.

The interesting question to me is not which of the four options have the highest probability of getting the crew back (that's probably option four), but which one has the highest probability of ending the manned space program -- a subject that is rarely very far from a senior Johnson Space Center official's mind.

Option four is a real roll of the dice. If it works, NASA is a hero again, a la Apollo XIII. If it fails (worst case, Atlantis is lost due to the rush), we've lost several astronauts, and half of the shuttle fleet. Is the shuttle program viable with only two vehicles, particularly given the circumstances in which the others were lost?

On the other hand, having lost the crew and the vehicle, few people are talking seriously about ending manned space flight or even just the shuttle program (though some, inevitably, are).

I'm not smart enough to predict what the public reaction would have been to any of these scenarios, and I doubt that anyone else is either, though many no doubt think they are.

There's an old saying that it's easier to get forgiveness than permission. In a sense, by remaining institutionally ignorant of the vehicle's plight (whether willfully or subconsciously), NASA spared itself a horrible dilemma, and one that they must now be grateful that they didn't have to confront, painful as the loss of Columbia and crew must be.