Jeffrey LewisUSA 193 Risk Calculation

In my post yesterday, Don’t Shoot Down USA 193, I voiced my “suspicion … that the expected risk to the ISS crew will be surprisingly competitive with the expected risk to persons on the ground” from the falling satellite.

My friend, the geographer Tim Gulden, did the math and it turns out my suspicion seems quite plausible.

Overall, Tim finds the chance that the debris will fall in an area with a population density of one or more persons per 1/4 hectare — about the size of the contamination zone — to be no more than one-half of one percent. That’s about 5 chances in 1,000. The probability that the debris will come down in an area with 3 or more persons per quarter hectare is about 2 chances in 1,000.

———— Original Message ————
Subject: Risk from Hydrazine onboard USA 193
Date: Mon, 18 Feb 2008 12:06:58 -0500
From: Tim Gulden
To: Nancy Gallagher, Jeffrey Lewis

I have done some really simple calculations of the risk of the hydrazine tank onboard USA193 actually landing in a densely enough populated area that people might be exposed. These calculations are based on the LandScan 2005 data set (developed by Oak Ridge National Labs):

The method here is to choose a threshold of population density and then see what percentage of the earths surface under the satellites orbit (between 58.5 degrees north and 58.5 degrees south) has a population density that high or higher.

The “contamination zone” for the hydrazine tank is stated as having a radius of about 27 meters — that is very nearly 1/4 of a hectare (which is an area 100 meters square).

The simplest thing to look at, then, is the percentage of the area under the orbit that has a density of 4 people per hectare or more. That corresponds to a very low density suburban type environment (one family every 100 meters on average). Approximately one half of one percent of the area under the orbit has this density or higher. That means that there are about five chances in 1000 that the tank lands in an area with this kind of density.

If we step the density up to a real urban density of 100 people per hectare, that is to say 25 people per contamination zone, this probability drops to about 0.02% or two chances in 10,000.

At 12 people per hectare, or 3 people per contamination zone, the probability is about 2 in 1000 — a risk of similar magnitude to the 1 in 1000 risk that NASA suggested for a debris strike from the interception hitting the International Space Station (and presumably killing its three crew members).

I am assuming that people stay where they live. The actual behavior of people toward this fallen object will matter a great deal more than it does with most falling debris. If people are inclined to move toward the tank, and stay near it, they could be killed even if it lands half a kilometer away. If, on the other hand, people are inclined to leave the area with the toxic smelling gas (apparently ammonia-like), then the people would not likely be killed even if they lived near enough to be exposed.

Another important caveat here is that these data are fine-grained, but not quite as fine-grained as would be ideal. The LandScan grid cells are 30 arc seconds square. At the equator, that translates to about 0.85 square kilometers. At the northern and southern extremes of the orbit it is about half a square kilometer. There are 100 hectares in a square kilometer and about 400 potential contamination zones in a kilometer. There are 100 hectares in a square kilometer and about 400 potential contamination zones in a kilometer. My calculations assume that people are evenly distributed within each cell — when in fact they are likely to be clustered. This would tend to drive the probability of hitting an occupied contamination zone downward. To the extent that low-density areas are typified by villages and extended family households, this could be really significant. I am guessing that clustering within cells could drive the probability of exposure down by a factor of 5 or more.

Of course, the error bars on both calculations are very large. The ISS estimate is extrapolated from Griffin’s comments about a NASA analysis that we cannot see for ourselves. Moreover, Tim did not model the probability that the tank remains intact — merely where it was likely to come down.

I continue to maintain that the Bush Administration should make public the real calculations of the risk. NASA published such estimates, by the way, for the Compton Gamma Ray Observatory — placing the probability that someone would be killed in an uncontrolled deorbit at 1 in 1,000. (Thanks to Yousaf Butt at the Union of Concerned Scientists for pointing this out.)

The unnecessary secrecy will feed the perception that Administration officials are being less than truthful about the President’s motivation for the intercept.


  1. Geoffrey Forden (History)

    I get 3.5%. There are serious flaws to the analysis you just posted. You know I’m not making this stuff up, right? Other people have used my same method to calculate the chances of debris hitting somebody and published their methods in peer-reviewed journals. And, when I stick in the “effective area” into their formulas, I get exactly the same amount. This methodology was not controversial when it didn’t involve an ASAT so I assume that most people object to it now because they don’t like the policy implications. The funny thing is, we all agree that the shoot down is a bad idea.

    you can now read my report on the proposed shoot-down at

  2. CKR (History)

    “The unnecessary secrecy will feed the perception that Administration officials are being less than truthful about the President’s motivation for the intercept.”

    But that is how this administration has consistently operated, particularly in the arms control area.

    It would have made a lot of sense for the Secretary of Defense to make a clear statement: this is where we think the satellite will come down, these are the risks, this is our plan, this is why we’re doing it.

    Heck, if they did that, they could almost get away with some lies, although their track record makes that less likely.

    Or it could be the Pres himself. Didn’t he make the decision?

    A generalized calculation like this one is not terriby informative. (Although many thanks to the people who are willing to spend their time enlightening the rest of us.)

    Only when the orbit of the satellite and its likely path of decay, along with the likely scattering of debris from The Shot™ and its paths of decay are known can a real risk assessment can be done.

  3. Mark Gubrud

    I get one chance in a thousand of a fatality, assuming the tank comes down intact and full of hydrazine slush.

    The number Geoffrey Forden calculates is the EXPECTED CASUALTY number, where being a “casualty” simply means being within 27 m (30 yards) of the impact point of the intact tank full of hydrazine slush. This number can be calculated very simply without breaking the population down, as Tim Gulden did, into urban, suburban and rural subgroups. In fact, you don’t even need to distinguish between land and ocean.

    However, you can’t ignore the fact that people don’t just distribute randomly on the surface of Earth; they live in towns, and cities and spend most of their time close to other people in houses, workplaces, etc. Geoffrey is doing his calculation on the assumption that humans are uncorrelated with each other. Taking into account the clustering of people means that the likelihood of any casualty is less than the expected number of casualties, since more of the possible incident scenarios affect more than one person than would be predicted by a random distribution of people.

    The “effective area” correction that Geoffrey refers to is negligible since the area of a 27m circle plus one person is almost exactly the area of a 27m circle by itself. This “methodology” is more relevant to the case of small pieces of debris where the area of a person makes a larger contribution to the effective area. In the present case, it makes no significant difference.

    The expected casualty number is calculated by multiplying the area of a 27 m radius circle by the average population density within the regions which can possibly be hit by the satellite. This latitude band actually contains nearly the entire world’s population. You could do better by integrating over the population by latitude since the satellite spends more time at high and low latitudes than in the middle. But this would not change the result much.

    The calculation is simple.

    Population of Earth: 6.7 billion
    Surface area between +-58.5 latitude (including ocean): 435 million km^2
    6.7e9/4.35e8 = 15.4
    Average density of people on surface below satellite orbit: 15.4 persons/km^2
    Area of 27 m circle: 0.00229 km^2
    15.4 * 0.00229 = 0.035
    Expected number of persons within 27 m of random impact point = 0.035

    That is Geoffrey’s number. It is not the probability that one or more persons will be casualties. To say that there is a 3.5% chance of someone being a casualty somewhere on the basis of this calculation is simply incorrect.

    Rather, it is the average number of people who would be casualties, per satellite crash, if this kind of crash occurred repeatedly. Please note that the expected casualty number is not 3.5 persons, it is 0.035 persons.

    The chance that one or more persons will be casualties of a single crash is lower than this 3.5% due to the fact that people even in low-density areas cluster in groups. We can estimate this effect by asking how many other persons are within 30 yards of any given person on Earth, averaged over everyone at all times. Tim Gulden estimates that clustering reduces the casualty probability by a factor of 5; I chose the more cautious estimate of 3.

    This means we have about 1% probability that at least one person will be within 30 yards of the impact. This result is consistent with Tim’s numbers, because it is the sum of the probabilities for one person, two persons, three, etc.

    However, this is still only the probability of at least one casualty, which is not the same as the probability of at least one fatality.

    Based on the toxicity of hydrazine (it is a chemical poison but it is not nerve gas) and its noxiousness (acrid smell, burning lungs and skin), some people may be attracted by the impact but most will try to flee the scene and all will try to avoid breathing the hydrazine cloud. In addition, the public can be warned and emergency teams readied to respond to the scene of the impact. Therefore, ambulatory persons are unlikely to be killed by the cloud. Some people may be immobile to begin with or may be incapacitated as a result of being hit by a piece of debris. Since the debris directly affects a much smaller area than the cloud, we can estimate at most a 10% chance that a person within 30 yards of the impact site would be killed.

    The conservative, cautious estimate, then, is about one chance in a thousand that someone, somewhere (at least one person) will be killed, assuming no action and that the tank lands somewhere intact and full of hydrazine slush.

    Note that the ASAT nature of this test has nowhere been invoked in calculating this result.

  4. Andy Grotto (History)

    Geoffrey — on page 2 of your compelling analysis, you say that the risk of death or injury is 3%, but you cite 3.5% in your comment above. Please clarify.

    Also, on page 9 of your analysis, first paragraph last sentence reads, “As will
    be shown below, the chances that some satellite will be hit by the debris created in this engagement are much, much smaller and last for a much, much shorter time.” An enticing promise, but I don’t see where you go into this in the remaining two pages of the article. Am I missing something? Also, I’d be particularly interested to know how the estimated risk to human life caused by a USA-193 debris field (i.e, to the ISS) compares to the 3%/3.5% figure.

  5. Tim Gulden (History)

    I think Mark gets at some of the fundamental differences between Goeff’s estimate and mine. We are estimating slightly different things using very different methods — I find is reassuring that our results are as close as they are. Geoff’s estimate is of the expected number of casualties, whereas my estimate is the probability of having at least one casualty — these are not quite the same thing.

    Perhaps more interestingly, the standard method is really geared toward small things hitting small targets. The fact that people are clustered doesn’t really matter much if you are dealing with lots of objects that are on the order of a meter across hitting people who take up about a square meter on the ground. There comes a point where clustering breaks down — except at certain rock concerts you seldom see more than two people occupying the same square meter — and even then they don’t share it for very long. Thus, you are looking at the possibility of a bunch of small, essentially independent things hitting a bunch of small, essentially independent, targets.

    This case is different. The 250 or so square meters that might be contaminated are all contiguous. On the scale of a meter, you can assume that everyone has their own impact zone. On the scale of 250 square meters, this does not hold and the distribution of people on the ground (the spatial autocorrelation of people, if you like) starts to matter a great deal. The only way to get at this is with real data — the tendency of people to cluster has to do with human nature, the nature of economic activity, etc. It has lots of interesting fractal qualities, etc. I was able to leverage ORNL’s estimates (which are not perfect, but they are as good as we get) down to something a bit smaller than a square kilometer, but clustering within square kilometers is still a matter of guesswork.

    The estimate of half a percent assumes that there is a probability of one that someone will be in the contamination zone if the tank lands in an area with a density of four or more people per hectare. That is certainly an overestimate because it is an estimate for an area that is nearly a kilometer square. The people will be clustered together in part of that kilometer — leaving a portion of it (possibly a large portion of it) unpopulated. That is where my downward adjustment (which I don’t include in the half percent) comes from. I think a factor of five is reasonable, but I don’t object to Mark’s suggestion of three.

    Geoff has correctly pointed out that it is possible for a person in a lower density area to be hit. There are a lot of such low density areas, but the chances of landing near a person in one of them goes down with the density. My calculations indicate that this could double the risk of landing near a person — though this also depends on clustering. I am also less confident in the ONRL data at these very low densities and suspect that they are assigning very low density to areas that actually have a small village every few kilometers. If so, that would drive down the doubling estimate.

    Bottom line is that I think an estimate of 1 or 2 in 1000 is pretty reasonable — that being 0.5% divided by 5 (for clustering) times something less than two (for lower density areas). The estimate is different than the one that the standard method gives, but that is because the situation is different. We aren’t talking about 250 one meter pieces, but one 250 square meter patch.

  6. Carey (History)

    Taking the expected number of fatalities (not casualties) as being between 1 in 100 and 1 in 1000, then the estimated cost of the SM-3 shot (about $60 million) represents a willingness to spend between $6 billion and $60 billion per fatality for fatality-avoidance, if indeed hazard reduction is the true principal motivation for the shot.

    There are perhaps cases where regulation mandated hazard reduction (which precludes explicit cost-benefit analysis) has imposed relative expenditures this great or even greater. But it is, I submit, quite anomalous for an organization with freedom of action – especially the U.S. military – to commit to this level of expenditure for similar levels of hazard reduction.

  7. Andy Grotto

    Carey — the decision wasn’t made by the military; it was made by the President. And I can’t imagine any politician saying, “Hmmm, do I pay $60 million out of a $1 trillion budget to try and eliminate the risk that one of my spy satellites lands on a retirement home? Nah…”

    Also, for the conspiracy theorists out there who think this is really about demonstrating an ASAT capability, why would you choose an out-of-control satellite humming along at a speed that would puts the closing velocity at a speed that’s 50% faster than in controlled tests to “demonstrate” this capability? It just doesn’t make any sense.

    Prediction: if the operation is a success, then the U.S. Congress will do nothing because there’s little political gain to be had in questioning both the motive (public safety) and the execution of the operation. But if it fails, the Congress will raise hell about the motive and the reasoning behind it.

  8. Lao Tao Ren (History)

    Just a note of thanks to Dr. Forden for his real neat papers (on his website) detailing the Shoot down and the Chinese ASAT test.

    It really makes a big difference where / how the bird is hit what happens.

    I am happy to see long term debris risk to be pretty minimal, but for what it is worth, it is still an ASAT / ABM test as far as I am concerned, with all the risk of saber rattling does in a very sensitive time for Sino-US relations (over the Kosovo Unilateral Declaration of Independence and diplomatic recognition of the new regime, the upcoming Taiwan referendum, etc.).

    I broached the issue earlier of whether or not there is another way to bring the bird down in a controlled fashion —- with another kind of intercept that nudges it out of orbit.

    Does that option exist?

    BTW, for the readers from China who got an earful about their ASAT test… I hope you can see that good decisions and good making processes are just as hard to do in the USA as it is in China.

    No country has a monopoly on being able to make not so good decisions.

  9. Glenn Burke (History)

    Andy, to answer “why would you choose an out-of-control satellite …” my answer would be, because it’s the perfect excuse (cover) to give it a try.

    That said, I personally don’t trust much of anything this (to use someone else’s term) maladministration does. However, they could well be operating under a sort of extension of the one-percent principle: if something sufficiently bad might happen, even with infinitesimal probability, they feel obliged to worry about it.

    I have neither the data nor expertise to judge the numbers that have been presented here. I assume they are reasonably accurate. But they have seemed to ignore the possibility of extreme clustering, and the fallout – political, and of human life – that could result.

    What, for instance, would be the length of advanced warning and accuracy? What of the effect in an enclosed space? An occupied sports stadium in a developed country could presumably be evacuated in advance. An apartment building in the middle of the night in southern Asia? What might be the effect of an evacuation on short notice (a time period I don’t know) of an area of unknown size (at least I don’t know)? Panic and deaths? These extreme, if improbable, circumstances may be entering into the equation.

    One thing that gives me pause about this reasoning is that, given the concern about debris, it would seem prudent for the ‘shot’ to wait for the orbit to degrade more significantly. That we don’t – does that mean we just want multiple chances? We’re unsure of our capability and don’t want to be embarrassed, and want the easiest possible shot(s)? (I use ‘we’ with reservation…) I don’t know, but I think the political fallout from a ‘worst case’ scenario cannot be discounted from the political calculation.

    Then again, maybe Bush is just marking his territory like my late pit bull did.

  10. Mike M (History)

    It is worth accounting for the fact that everyone now knows it is coming and they will have 30 to 45 minutes of warning depending on how slow a news day it is. This will further reduce the chance of anyone being hit; they will have time to react and go to the basement or some other protected area once the impact area has been estimated and released to the public.

  11. yousaf

    It would help put the conspiracy theorists’ minds at rest if the probability of the tank remaining intact during an uncontrolled re-entry were released.

    I think it is very likely that the tank would explode even in an uncontrolled re-entry since the auto-ignition temp of hydrazine is ~150C in a metal container, and, further, there is surely plenty of static discharge around during re-entry. I had posted the link to the source in a previous post but here is some of the info again:

    Hydrazine Autoignition Temperature varies with contact surface:
    23C (74F) in contact with iron rust;
    132C (270F) in contact with black iron;
    156C (313F) in contact with stainless steel;
    270C (518F) in contact with glass.


    Sealed containers may rupture when heated. Vapors can flow along surfaces to distant ignition source and flash back. Above the flash point, explosive vapor-air mixtures may be formed. Sensitive to static discharge.

    Given this data, who thinks the tank will survive and why?

  12. Mark Gubrud

    Andy, I don’t know why you think a conspiracy needs to be involved for the motive behind this to be demonstrating an ASAT capability, but you ask a good question: Why bet on a weapon not designed or tested for the mission, if you want to make a demonstration? Isn’t the risk of failure too high? It sure seems like a bad bet, something only “someone with a cowboy mentality” would do. Perhaps that’s why General Cartwright stressed the decision was made by the president.

    You are probably right that success will succeed and failure will fail. And Bush is a failure already, so what’s there to lose? No downside to this, as Cartwright kept insisting. Except that it will then be officially open season for ASAT testing in space.

    My conspiracy theory is that Bush thought the safety issue was a good cover for his taking a shot at a satellite, arms control, China, liberals and Europeans just to show that he is still licensed to do so. A conspiracy of one – although Bush was probably sold the idea by somebody else, a Lowell Wood type perhaps.

  13. Wang Ting (History)

    I spend severl minutes to understand Time and Mark’s method.

    An example could be used to tell the difference of the two methods.

    If we presume, the 6.7 million people are at one point, although that is not going to happen, the expect fatalities number is the same:

    Area of 27m circle / ( Surface area between +-58.5 latitude ) * Population = 3.5%,

    while probability of having at least one fatalities is:

    Area of 27m circle / ( Surface area between +-58.5 latitude ) = 5.26e-9%.

    Actually, in this case, “at least one fatalities” is equivalent to “6.7 million fatalities”.

    The “at least one fatalities” and “expect fatalities number” are not the same thing, these two number give us a different way to understand the potential casualties.

    By the way, “a risk of similar magnitude to the 1 in 1000 risk that NASA suggested for a debris strike from the interception hitting the International Space Station”, does not mean that such collision will kill 3 crew. Only the fragment hit fatal device will make the mission failure. The probability of that is very low. I personaly agree with Griffin’s statements that the risk to ISS is very low.

    Geoffrey —— Your estimation about FY fragments liftime might not correctly, you probably get that from T.S Kelso’s website. However, one of his assumption is probably wrong. According to NASA, ESA and my calculation, most of FY debris will decay within 100 years.

  14. rocketnerd

    I don’t think that you should underestimate the health implications associated with a large scale hydrazine exposure. In addition Mark’s comment that describes its effects as “acrid smell, burning lungs and skin,” it is a known carcinogen, and causes permanent tissue damage to your lungs if inhaled in sufficient (small) concentration.

    I’m not questioning your analysis. I am just stating that as someone who has worked around hydrazine, and been involved in a satellite debris recovery, if this landed in your neighborhood, you’d be having a pretty bad day.

  15. James (History)

    If an satellite falls out of the sky and hurts someone, there’s some political embarrassment that is quickly dispelled by money and the observation that “what goes up must come down, eventually.” Accidents happen, they’ve happened before, they’ll happen again.

    But if damage is done to another space object, or even if the debris from the collision harms someone on the ground, then the US is liable in a way they would not have been before: that the US deliberately took an action that led to the injury. Citing a hypothetical risk of a thousand to one as a contrary scenario isn’t much of a defense against a genuine injury that would not have happened if the US hadn’t decided to change a situation whose outcome they could not predict in the first place.

    This political factor is the only real risk in the scenario. On a personal level, we are all far more at risk of getting struck by lightning either way. And I think the US government has done a very good job of getting people to split hairs over this fuel tank, whose contents aren’t really much more toxic than what millions of farmers spray in their fields every spring. Those tanker trucks overturn all the damned time.

    So if the stated motive is bogus, there must be an ulterior motive, which is probably a combination of allowing Republicans to brag about “saving NMD” and making sure the satellite’s remains are not available for analyis. And like most of Bush’s endeavors, it’s a matter of too much risk for far too little gain.

  16. J.Lo (History)

    Carey’s point about the military isn’t entirely off-base, Andy. The president made the decision, but with the recommendation of “national and homeland teams”, according to James Jeffrey. It seems a pretty safe bet that he had neither knowledge nor opinion about the matter prior to the recommendation and merely approved based on his confidence in his own team. What we do not know and likely will not know is how these teams played it—as a safety matter or as an ASAT test. I don’t know enough about either to make a useful guess, but the decision itself was almost certainly not taken without expert influence toward a predetermined conclusion.

    However, I would slightly amend Carey’s otherwise interesting cost-benefit analysis for projected fatality risk. While the intercept indicates a level of spending for fatality avoidance, the only verifiable value put to this particular fatality avoidance opportunity is US$60 million.

  17. Andy (History)

    Well, after being away in the wilds all weekend, I’m finally back to find this fascinating development and the excellent dialog here and in the other two posts.

    I would like to return, for a bit, to some of Gen. Cartwright’s comments and propose a possible “worst case” scenario.

    First, Gen. Cartwright suggested that the intercept had two objectives. Here’s the second:

    Second objective is to hit the tank, the hydrazine tank, and rupture it so that we can off-gas this hydrazene as early as possible so the least amount of it returns to the Earth….

    So, what if the tank isn’t hit? What if the hit is simply a graze that doesn’t catastrophically break apart the satellite? He says later on:

    We have to be able to assess if parts of the satellite came apart is the — which part is which, and that’s a very difficult thing to do. In other words, if the satellite grazed but did not directly impact, how do you decide whether you should take a second shot? And we’ll work our way through that, but it’ll be a conscious decision that we’ll take.

    So, what he’s essentially saying is that there’s a possibility the tank won’t be damaged and that a grazing shot could break the satellite into a few large pieces instead of many small ones. His question on taking a second shot is a good one to consider and I would add another – would they even be able to tell which piece contained an intact hydrazine tank and is the SM-3 capable of hitting that piece on a second shot with several large “decoys” (other pieces of the sat) nearby? If a second shot cannot be make, Gen. Cartwright asserts the situation will still be better:

    If we graze the satellite, we’re still better off because likely we’ll still bring it down sooner and therefore more predictably. If we hit the hydrazine tank, then we’ve improved our potential to mitigate that threat. So the regret factor of not acting clearly outweighed the regret factors of acting. And as long as that’s the case, we felt that the responsible activity was to go ahead and try to engage the satellite.

    Bringing down the satellite sooner does not, it seems to me, equal bringing it down more predictably, particularly if there are several large chunks coming down and not a single satellite. But I hasten to add that is my layman’s opinion and would like to hear expert opinion from the likes of Dr. Higgins.

    Furthermore – and here’s my worst-case scenario – what if the shot does graze the satellite, altering its orbit slightly and leaving the tank intact. Further, what if this slightly altered orbit brings the tank down in a populated area? That is, admittedly a lot of “ifs” but we could conceivable be looking at a situation where our actions cause the very event they were intended to prevent. Add in all the other negative political effects from taking the shot and we’d have what we call in the military a Golf Foxtrot.

    It seems to me the administration is pursing a course of action entailing many low-probability bad consequences plus the usual unknown and unintended consequences in an attempt to mitigate a single low-probability event. In my book that’s inherently dangerous, particularly given the potential political consequences of using an nascent “ASAT” weapon, whatever the reality of the SM-3’s capabilities.

  18. Mark Gubrud

    rocketnerd – There is a potential for injury and long-term health effects from heavy exposure to hydrazine fumes for a short period. The risk of carcinogenesis from such an exposure is very small as compared with chronic exposure.

    Long-term health effects make a not insignficant contribution to the risk of a fatality, but I think this is adequately allowed for if one assumes that a person caught within the “two football fields” that might be affected has a 10% chance of dying because he or she “lingered” on the scene.

  19. SmellyTerror (History)

    I find it very hard to believe the tank would survive re-entry. Multiply that by the minimal chance of an intact tank actually killing anyone. Now compare that figure with the chances of missing the sat or failing to destroy the tank with a missile, and the chances of all of this just making things worse anyway. You have something that just does not seem in any way worth 1. the money, 2. the diplomatic fallout and 3. the degree to which they’ll look like ass if this attempt fails.

    Something does not add up. So yeah, I can see where “conspiracy theories” gain some traction on this one, given that the stated reasoning seems so unlikely. Personally, though, I’m keen to lay the blame on the testicle factor, which is always possible when a small number of people get to put their heads together on a decision like this with no real oversight. I can easily imagine that a test launch involving a spittoon featured somewhere in all of this.

    Oh sure, maybe other considerations played a part, and I don’t doubt the survival of secret junk was discussed, and there were probably people talking up the ASAT testing option, and hell, maybe someone really was worried about that one one-thousandth of a human likely to get toasted (what is that? An ear?). But ultimately I think that someone, somewhere, said “to hell with the numbers, we’re the United States. Just shoot the fricking thing down”.

  20. Lao Tao Ren (History)

    Time for a bit of comic relief on this site!

    When the risk analysis was done for the Shoot Down, did they factor in the possibility that the Aegis cruisers may run over some fishing boats —- like the Atago just did?


    Storm erupts over warship in sea smash

    Miwa Suzuki

    Wednesday, February 20, 2008

    A Japanese naval destroyer with state-of-the-art radar collided yesterday with fishermen off the Pacific coast, leaving a father and son missing after their small boat was split in half.
    Tokyo immediately offered apologies for the early morning incident, which comes in the wake of defense scandals.

    TV footage showed a half-sunken white-colored part of the tuna-fishing ship in the ocean as divers searched the area. The stern was floating separately with its red bottom pointing up.

    The 165-meter Atago destroyer, returning from a visit to Hawaii, crashed into the fishing boat, less than 10 times its size, off the Nojimazaki Cape, 80 kilometers south of Tokyo.

    The Atago is the newest and largest of Japan’s five vessels equipped with the US-developed Aegis combat system.

  21. FOARP (History)

    Well, it seems that this great demonstration of America’s shield against missile attack is postponed, at least for today, due to ‘heavy seas’.

    I guess the bad guys are only planning to attack the US on days where the sea-state in the Pacific is calm or better.

  22. Andrew Higgins (History)

    I’ve put up on my website a small analysis my grad students and I did regarding the generation of very high velocity, forward-launched fragments in ASAT events, such as Geoff Forden found with Fengyun 1C.

    Similar forward-projected high velocity fragments had been observed with the 1985 Solwind intercept as well.

    This might give you some idea of the shock dynamics involved in these type of impacts, and how fragments can get launched into orbits with perigees 1000’s of km greater than the original satellite. Think of this as a refined version of Geoff’s “ping-pong ball” model.

    You can see the last couple of pages for my predictions regarding the USA 193 intercept.

  23. tina (History)

    BBC reporting:
    The US has struck a defunct spy satellite with a missile fired from a warship, military officials say.

  24. Andy (History)

    Well, now that it’s over, all I can say is…

  25. John Debo (History)

    Mercury, Gemini, Apollo, and the shuttle all use/used hydrazine (UDMH), so the hazards to general population seem overrated. Also—good article on space debris hazard in 3/2008 Smithsonian Air & Space

  26. Mike (History)

    Okay. That was some cool video of the missile tracking the satellite and the explosion.

  27. Will (History)

    No comment on the “direct” hit? C’mon give us a little love!

  28. meleager

    “C’mon give us a little love!”

    Probability = .003%

    it might get back to the (wished for) next Democrat administration

  29. yousaf

    as of right now the prob. that the tank was breached is only 80-90%:

    GEN. CARTWRIGHT: But I would tell you, from watching and from participating, that we’re in the very, very high — 80, 90 percent sure that the tank was breached.


    So if there is 20% chance it was not hit, and the identical reasoning were to be re-applied, they ought to keep shooting the debris until the cumulative probability goes below 1-in-a-thousand which apparently was bad enough for NASA to forcibly re-enter CGRO.

    If we take Geoff’s 3.5% injury/death probability number then the probability that the possibly existing hydrazine tank can still poison a football field is as bad as ~0.7% (as there no certainty that the tank exploded already)….

    In the next day or so we should know the definitive result. Anything less than a 97% certainty that the tank was breached will mean it was failed interception as the prob.‘s are still too large.

  30. JimO (History)

    “The simplest thing to look at, then, is the percentage of the area under the orbit that has a density of 4 people per hectare or more. That corresponds to a very low density suburban type environment (one family every 100 meters on average). Approximately one half of one percent of the area under the orbit has this density or higher. That means that there are about five chances in 1000 that the tank lands in an area with this kind of density.”

    These odds can be modified by the strange feature of satellite orbits, that a satellite does not spend equal time over any square-unit-of-choice area between its northern and southern extremes. It tends to spend more time over squares near those extremes than those near the equator. In this case, with an inclination of 58 degrees, the satellite would spend more time over regions that tend to be less populated.

    Now, do satellites fall to Earth randomly along their orbits, when they decay at last? I don’t know, but they do tend to experience air drag more severely over the equator, because of the equatorial bulge. I’d be interested in a statistical analysis of the latitude of terminal decay (purely vertical velocity of surviving fragments) as a function of orbital inclination — but I don’t think such a study exosts, or may even be possible.

    Maybe all these second-order quibbles just cancel out.

  31. JimO (History)

    Gubrud: “I get one chance in a thousand of a fatality, assuming the tank comes down intact and full of hydrazine slush.”

    Thanks for doing the work. This figure is precisely the one given by NASA in 2000 when they terminated the Compton Gamma Ray Observatory mission when its attitude control system began misbehaving, and made a deliberate de-orbit rather than risk the control systen failing further, leading to an uncontrolled entry. The threshhold they announced at the time was that the heavy structural elements of the satellite raised the chance of human casualty to 1 in a 1000.

    I share the common curiosity of what was the figure calculated by NASA for USA-193.

    I understand that a gag order re any related subjects, on the NASA team, has just been lifted. Let’s ask them again.