Reader's Favorites
Media Casualties Mount

Administration Split On Europe Invasion

Administration In Crisis Over Burgeoning Quagmire

Congress Concerned About Diversion From War On Japan

Pot, Kettle On Line Two...

Allies Seize Paris

The Natural

Gore Book Sales Tank, Supporters Claim Unfair Tactics

Satan Files Lack Of Defamation Suit

Why This Blog Bores People With Space Stuff

A New Beginning

My Hit Parade
Instapundit (Glenn Reynolds)
Tim Blair
James Lileks Bleats
Virginia Postrel
Kausfiles
Winds Of Change (Joe Katzman)
Little Green Footballs (Charles Johnson)
Samizdata
Eject Eject Eject (Bill Whittle)

Space
Alan Boyle (MSNBC)
Space Politics (Jeff Foust)
Space Transport News (Clark Lindsey)
NASA Watch
NASA Space Flight
Hobby Space
A Voyage To Arcturus (Jay Manifold)
Dispatches From The Final Frontier (Michael Belfiore)
Personal Spaceflight (Jeff Foust)
Mars Blog
The Flame Trench (Florida Today)
Space Cynic
Rocket Forge (Michael Mealing)
COTS Watch (Michael Mealing)
Curmudgeon's Corner (Mark Whittington)
Selenian Boondocks
Tales of the Heliosphere
Out Of The Cradle
Space For Commerce (Brian Dunbar)
True Anomaly
Kevin Parkin
The Speculist (Phil Bowermaster)
Spacecraft (Chris Hall)
Space Pragmatism (Dan Schrimpsher)
Eternal Golden Braid (Fred Kiesche)
Carried Away (Dan Schmelzer)
Laughing Wolf (C. Blake Powers)
Chair Force Engineer (Air Force Procurement)
Spacearium
Saturn Follies
JesusPhreaks (Scott Bell)

Biting Commentary about Infinity, and Beyond!

« Technology and Psychology | Main | Security Hole in Mac OS 10.3 »

Space Policy Disconnect

NASA was surprised by the reaction to its announcement a few weeks ago that it wasn't going to risk a Shuttle crew to keep the Hubble Space Telescope alive.

It apparently underestimated the popularity of the program. It shouldn't have.

How many people have screen savers of the Shuttle payload bay, or the space station?

How many, in contrast, compute to a background of the Eagle Nebula, or other Hubble images?

Many, particularly in the space science community, were quick to point out the timing of the decision. Was it just a coincidence that, just a few weeks after the president's announcement on January 14 of a new human-exploration space policy, in which NASA's resources would be focused on the goals of sending astronauts to the moon and Mars, Hubble life extension was pronounced to be an unworthy cause on which to risk a Shuttle flight?

Well, actually, it was.

While some of the motivations of the agency in this action remain murky, it's safe to say that, despite the timing, the decision probably wasn't a result of the new space policy. The most likely suspect remains the CAIB report on the loss of the Columbia last February, which was released last fall. While the commission didn't specifically recommend not flying to non-ISS orbits, this was an inferred recommendation from many of the others, given that it's quite possible that the Shuttle crew might have lived, even had the vehicle been written off, had their mission been to the station, where the damage might have been clearly seen and they would have had a safe haven.

But in the timing of all these events and decisions, there seems to be a disconnect in terms of policy. To the degree that the Hubble decision was based on the Gehman recommendations, that decision must now be revisited. Here's why.

The Gehman report was delivered last fall, before the president's January speech. At that time, the space policy of the United States was, among other things, to continue to fly the Space Shuttle as long as possible, until a decision was made to replace it, and its replacement developed. With the loss of Columbia, we had only three orbiters in the Shuttle fleet, making each one very precious if they were to support a program of indefinite duration, particularly given the now-empirical reliability of ninety eight percent (two losses in about a hundred some flights over twenty years). Despite any improvements they're making, that's probably the number that NASA is using to estimate future losses, to be conservative.

What does it mean?

At that reliability, there is a forty percent chance of losing another orbiter (which would cost billions and years to replace) in the next twenty five flights. There's a two in three probability of losing one in the next fifty. That means there's an excellent chance of losing one over the next ten years, at an optimistic flight rate of five per year. Hence the eagerness to follow the Gehman Commission's recommendations and spend hundreds of millions of dollars to solve the problem, because the prevailing policy is to keep the Shuttle fleet flying until it's no longer needed.

OK, now fast forward to January 14th of this year, when the president announces, among other things, that the Shuttle is to be phased out with the planned completion of the International Space Station, in 2010, six years from now.

It's a new policy world. NASA no longer has to worry about sustaining a three-orbiter Shuttle fleet into an indefinite future--they've been told that it only has to fly another thirty flights or so.

In fact, here's the irony.

While its critics are lambasting the agency for sacrificing Hubble on the altar of the new space policy, the new policy in fact would actually justify a Hubble mission. Consider--if it's no longer essential to maintain a three-orbiter fleet into the indefinite future, the two percent risk of losing an orbiter now looks small compared to the value of keeping Hubble going for several more years, until we can be assured of a worthy replacement. If, against the odds, we do lose another orbiter on that mission, the worst case is that it will simply take another couple of years to complete station, at which point we'll still shut down the fleet, if the new policy is to be believed.

So here's the policy disconnect.

We have one part of NASA (the Shuttle program) furiously running off to implement the recommendations of the CAIB, recommendations which were based on a circa-2003 policy made obsolete on January 14th, 2004, with apparently no recognition of the events of that date. We have another part of NASA desperately trying to implement the new, January policy.

It doesn't seem unreasonable to suggest that the expensive and delayed Shuttle return-to-flight activities (the latest estimate is to fly again is March, 2005, almost two years after the Columbia loss, and approaching the ridiculously-long standdown after Challenger) should reflect the new, existing policy, and not the old one in which the Gehman report was written. It's conceivable that, if asked, the commission might not change any of their recommendations, but it's insane not to ask them, given the dramatic change in national space goals since they issued the report.

Accordingly, I propose that the Gehman Commission be reconvened to weigh in on this issue. It need not be a several-month-long process, or disrupt the lives of the commission members to the degree that the first one did. No long investigations are needed, no facts have changed, except that the nation has a new space policy. It would be appropriate to gather the commission members in a room once more, to review their recommendations from last fall, and to reconsider them in the light of the new space policy that the president announced in January. It need not take more than a day or three.

The costs of it would be minimal, particularly considering the ongoing costs of NASA continuing down an expensive and perhaps pointless road, one costing many hundreds of millions of dollar per year, in developing expensive fixes to a system that we have already stated as policy will be phased out in much less than a decade.

Mr. President, Administrator O'Keefe, please reconvene the commission. Please reconcile this apparent disconnect of the policy of yesteryear with the forward-looking policy that you proclaimed over three months ago. It may indicate that the current NASA policy is correct, but it might alternatively save many millions of taxpayer dollars on fruitless fixes to an obsolete program. And it may create many more beautiful images from distant times and distant galaxies, images that satisfy both scientists' curiosities, and peoples' aesthetic souls.

Posted by Rand Simberg at May 18, 2004 12:14 PM

Comments

"The Gehman report was delivered last fall, before the loss of Columbia"

Huh!?!

Typo aside, you do remember that Gehman was asked to review the Hubble policy after pressure by Senator Mikulski, and basically punted?

Posted by Duncan Young at May 18, 2004 12:30 PM

Here's Gehman's take on Service Mission 4.

Posted by Duncan Young at May 18, 2004 12:34 PM

Yes, I'm aware that he punted on the Hubble question. I'm not asking Gehman specifically about Hubble. I'm asking to reconvene the commission to revisit their general recommendations about Shuttle in light of the new policy. The output of that might (or might not) result in a decision to send Shuttle to Hubble, but I'm more interested in the broader policy implications beyond Hubble (e.g., does the current return-to-flight activity make sense in general, given that we're now going to retire the fleet in six years?).

Posted by Rand Simberg at May 18, 2004 12:49 PM

Well, I do agree that some of the recommendations should be revised. Actually the big one in my mind is the 2010 "recertification-or-shutdown" clause. Given the ban on night launchs (which does make some sence, given the inherent susceptibility of the stack to debris) they can only get three (maybe four with a very quick turn-around) launches to ISS per year. There is no way they can finish the station by 2010. To try and do so goes down the same schedule trap that sealed the fates of Challenger and Columbia. That should be revised to "completion of the 14 A mission". In the months that ISS is unavailable, NASA should fly SM-4, with the assigned rescue shuttle for the preceding station mission.

I can understand NASA's caution, though. The next shuttle lost will likely be carrying major station elements, and given the current 2010 shutdown date, it would be impossible to recover from another multiyear standdown. I understand that they are already ending some STS procurement programs. Loss of another shuttle ends the program.

Bruce McCandless suggestion for SM-5 (before the SM-4 cancelation) was intriguing - he recommended that that mission be used as a testbed for telerobotic repair techniques - with human backup.

Posted by Duncan Young at May 18, 2004 01:17 PM

I think not enough credit is being given to guile. The Hubble is of course very popular, and NASA knows this very well.

If NASA were to come out and say that some quiet program were to be cancelled, only the few people directly working on that program would be affected. But if they say that the new budget and plans call for scrubbing the Hubble, then of course everyone will howl. That's how they get more money.

It reminds me of the story of the difference between how the Marines/Navy build air bases and how the Air Forces builds them. The Marines and Navy build the runway and the hangars and then has to go back to beg for more money for barracks and gets the bare minimum to live in. The air force builds the golf course, the gymnasium, the barracks, the clubs, the recreation center, the tv station, etc. Then they run out of money and have to go beg for money for the runway and hangars, and get everything they want.

I don't buy for a moment that this isn't the case concerning Hubble.

Posted by Mike Rentner at May 19, 2004 06:01 AM

Aren't you forgetting something? Station is dependent on shuttle. Station is in a decaying orbit and needs shuttle to lift it back up to safety. You should know that this is part of the March deadline. If shuttle doesn't fly soon after March, then it is game over for both shuttle and station.

Also, isn't clear to you that the real reason for fixing shuttle is to protect the astronauts? If all we cared about was mission priorities and budgets, we would never have bothered to investigate or do return-to-flight; we would just have procurred another shuttle.

Posted by Matthew at May 19, 2004 06:45 AM

Hi Mike,

I can't agree with your analysis of NASA's motives. Sean O'Keefe has staked his professional reputation on cancelling shuttle missions to HST. There simply will not be a manned mission to HST while O'Keefe is in charge. Trust me: I and my coworkers have spent several months with our fingers in the air and the wind is blowing, at best, toward a robotic mission to HST.

O'Keefe is an accountant, not a scientist or an engineer, and he had no idea of the place Hubble holds in the imaginations of so many people. He cancelled missions to HST to avoid another loss of life on his watch and to cut funds that can be redirected to the moon/Mars effort. His underlings, many of whom have been gunning for HST for various political reasons, failed to anticipate the level of outcry from the public and the astronomical community. It was a public relations cock-up, not a double-secret funding scheme.

Posted by C.S. Froning at May 19, 2004 06:53 AM

Given that the Shuttle program had over a hundred consecutive successful launches, that there is increased awareness and vigalance within the program, and that NASA has refocused it's view of the program from "operational" to "experimental," I see no reason why the remaining 25 to 30 flights couldn't be accident free. The 98% success rate is based on old ways.

The new primary mission of NASA is now returning to the Moon and then to Mars. How does Hubble help that?

Also, I've heard from some amateur astronomers that there are now many land based telescopes that give equally good pictures as Hubble. These pictures just aren't released to the public as frequently because they're not public domain.

Posted by mikeholt at May 19, 2004 06:56 AM

Also, I've heard from some amateur astronomers that there are now many land based telescopes that give equally good pictures as Hubble.

This is getting more common as time goes on. BUT 1) they can't do this in all regions of the sky and 2) they can't do it at all wavelengths.

Posted by Angie Schultz at May 19, 2004 07:11 AM

Hubble needs to be replaced, and not having a telescope in space for a couple of years isn't a big deal at all.

And let's build it right this time.

The shuttle ought to be replaced too, and if the space station really needs a shuttle to push it into a better orbit, send a goddamn shuttle without a crew (or some other vehicle)

I don't see the big problem, maybe a PR problem, it sounds bad to lose hubble, but of course we'll eventually have its successor, and focusing on that will probably return better results for the dollar.

Posted by Dustin at May 19, 2004 08:20 AM

The Hubble telescope (HST) is by far the best instrument out there for most near-visible ultraviolet (UV) frequency imaging work. Keck telescope and other ground scopes do work in these frequencies, but the atmosphere isn't as transparent as it is at visible wavelengths which means the HST has a big advantage in UV. The James Webb space telescope won't fill this niche either since it is optimised for infrared observations.

So if HST is scuttled, then we'll lose a significant UV observation tool with no replacement planned.

Posted by Karl Hallowell at May 19, 2004 08:30 AM

Your probability calculations tell me (I'm a math and science teacher, for the record) that you have fallen into the infamous "gambler's fallacy". Basically, the gambler's fallacy goes something like this: if I flip a coin, and it lands on heads, then the next time I flip the same coin, it is more likely to land on tails, since the coin should land on heads and tails in roughly even amounts. In fact, the probability on the second flip is still exactly 50/50. What you have said here is equivalent, though with a smaller probability. You seem to be claiming that, because the shuttle has a 98% (or so) success rate, and the remaining shuttles have made a large number of successful trips, that the probability of them being destroying is increasing with each successful mission. While this may be true from an engineering standpoint (since parts and materials degrade over time), you can not reach that conclusion by looking at straight probabilities.

That having been said, I take some offense at the idea that, since we're planning to replace the shuttle fleet anyway, we can send them up to do more dangerous missions because we don't need them for much longer. I'm sorry, but if the safety of the astronauts is in question, as you indicated, then we should not send them up. The shuttles may be expendable, but the humans are not.

Finally, Hubble is one of the most important scientific devices ever created. I would be willing to wager that more useful scientific data has come from the Hubble in the last week than from the ISS in the last year (and Hubble is cheaper, too). And, despite what a previous poster has said, Hubble can not be replaced by ground-based telescopes. The Earth's atmosphere bends light unpredictably, and there is no way to correct it. Hubble (and other space-based telescopes) is the only available technology capable of doing deep space observations in most parts of the spectrum.

I don't mean to sound so confrontational, but you struck a nerve. I applaud your delving into scientific issues.

Posted by StuTheSheep at May 19, 2004 08:50 AM

No, it's not a gambler's fallacy. If the shuttle has a 98% chance of surviving a single flight, then the chance of it surviving 50 flights is 36%. The chance of survival over 25 flights is .6, or a 40% chance of loss as Rand said.

The real problem is that 98% figure. The error bars around that number are huge, because the sample size is small. We simply don't know what the real reliability of the Shuttle is. We know of two failure modes in the most complex system ever built. We don't know how many others are lurking. It may be that the shuttle is much more reliable than 98%, and we've just been unlucky, or that it's a deathtrap and we've been very, very lucky. NASA does FMEAs to try to quantify the risk, but in something as complex as the shuttle that's obviously an inexact process.

But Rand is overlooking one issue in my opinion, and that's the politics of another shuttle disaster. In my perfect world, Shuttle astronauts would sign up knowing the risks, and if we lost another shuttle we'd simply analyze what went wrong and keep flying them.

But in the real world, I think there is a very, very good chance that the loss of another orbiter would ground the entire program. NASA is just too risk-averse, and the media too scandal-mongering to allow another failure to be accepted as business as usual. Look how long the fleet has been grounded this time - if another orbiter is lost, how long will it take to fly another one, even assuming that the program would continue? Two years? Three? Four? In the meantime, you start running up against the 2010 recertification deadline recommended by the CAIB, which you don't want to do if you're scrapping shuttle anyway.

NASA can inoculate itslef from this outcome a bit by being very up-front about the risks. If it flies only crucial crewmembers, and keeps telling us that this is highly risky but worth doing, then perhaps another accident would be tolerated. But NASA will most likely go the other way - before it flies shuttle again the PR flacks will be running around telling the media how they've taken all the CAIB recommendations to heart, and that they are a new agency, and they're on top of the situation, and everything is much safer, yada yada yada. Then if there is another failure, NASA will look like it was caught napping again.

Posted by Dan Hanson at May 19, 2004 09:43 AM

"There simply will not be a manned mission to HST while O'Keefe is in charge"

That's what I like, a quote with the clear solution contained within.

I appreciate his calculations, and his reasoning, and that he's willing to put his career behind his decisions.

Now, someone fire him. Please.

Posted by W. Ian Blanton at May 19, 2004 09:46 AM

I've responded to some of these comments here.

Posted by Rand Simberg at May 19, 2004 09:47 AM

But Rand is overlooking one issue in my opinion, and that's the politics of another shuttle disaster.

I'm not overlooking it, Dan. I'm trying to change it, with this and other posts, because if we don't, we have no hope of getting off the planet in any significant way.

Posted by Rand Simberg at May 19, 2004 09:59 AM

A couple of minor points:

Rand should look at the specific recommendations in the CAIB report. Although CAIB did not specifically address an HST servicing mission, it did specifically refuse to rule out non-ISS missions, thereby leaving open the possibility of an HST servicing mission. What has happened is that NASA is going beyond the CAIB recommendations and adopting a higher level of safety. It is perfectly justifiable for NASA to do this, but CAIB did not require it.

Also, although only limited info about this has made it into the press, budget concerns--and scheduling concerns--did enter into the HST decision. It was not simply a safety-based decision.

Posted by at May 19, 2004 10:08 AM

The probability problem is a binomial probability situation (I teach COLLEGE math) :-). We are looking out over the next 50 flights and asking:

"What is the probability of 50 successful flights on the next 50 flights, if the probability of an individual successful flight is 98%?"

It is true that if you have had 49 successful flights, then the probability on the 50th flight that you will be successful is 98%. But, we are standing here before the first flight, and that means it is a binomial probability problem.

The math works out to:

(The combination of 50 items taken 50 at a time) multiplied by the probability of a successful flight to the power of the number of successful flights multiplied by the probability of an unsuccessful flight to the power of the number of unsuccessful flights

OR

(50 Cn 50) x (0.98)^50 x (0.02)^(50-50)

which is equal to

1 x 0.3642 x 1

or 36.42%.

So, the chances of 50 flights in a row being successful is 36%, if the probability of an individual flight being successful is 98%.

This costs you in my college class. :-)

Posted by Gene at May 19, 2004 11:17 AM

Yes, when I said forty percent, I was approximating (not inappropriate, given how uncertain the basis for the number is). Though we say 98 percent, we don't really know the number to two significant figures, except that it's probably significantly greater than ninety, and less than a hundred.

Posted by Rand Simberg at May 19, 2004 11:29 AM

Actually, you were exactly right, Rand. You said there was a 40% chance of failure over 25 flights. The chance of success is .98 ^25, or 60.34%. Therefore, the chance of failure is 39.66%. Since there are error bars around the .98 figure, 40% is a perfectly fine estimate.

The 36% number is the chance of success over 50 flights.

The real problem with assigning a safety estimate for a large number of flights is that the uncertainty in the .98 figure propagates over the probability distribution as you increase the number of trials. For example, if the real chance for success is not .98, but .95, then the probably of making 50 flights without an accident drops from 36% to 7%. If the probability is only .90, the chance of flying 50 successful missions becomes vanishingly small (0.5%).

So you have to be very careful when trying to make probablistic estimates over large trials when your starting conditions are uncertain. Small changes in initial conditions lead to radically different outcomes.

So... The real, scientific way to approach this would be to start with some error bars. If we could state with 95% confidence that the real success rate of a shuttle flight is somewhere between .90 and .99, then we could say with confidence that the chance of completing 25 safe flights is somewhere between 7% and 78%. The chance of completing 50 safe flights is between 0.5% and 60%.

Looked at that way, it would appear that the whole safety analysis of future flights is a crapshoot. We simply don't have good enough data to generate a meaningful probability of success over the long term. (unless you can show me that that .98 number is more accurate than just being "2 out of 100 flights failed").

Posted by Dan Hanson at May 19, 2004 01:09 PM

Actually, Rand and Dan, I wasn't trying to correct you guys. :-)

I was writing for Stuthesheep.

40% is CERTAINLY a good enough approximation, given the uncertainties.

Should have made my audience clear. Oops.
Gene

Posted by Gene at May 19, 2004 01:55 PM

Mr. Simberg wrote:
"Accordingly, I propose that the Gehman Commission be reconvened to weigh in on this issue."

It was announced last year that the CAIB was asked by Congress to reconvene approximately one year after issuing its first report. A media search will find references to this.

Posted by Dwayne A. Day at May 21, 2004 09:37 AM

I know that the CAIB was supposed to reconvene after a year, but my point is that a) they shouldn't wait a year--they should reconvene as soon as possible to reconsider in light of the new policy and b) that they should be reconvening for that purpose, rather than to simply see how NASA's doing in following their (now possibly obsolete) recommendations.

Posted by Rand Simberg at May 21, 2004 10:22 AM


Post a comment

Name:


Email Address:


URL:


Comments:


Remember info?