Wednesday, September 18, 2013

Crater Analysis

The diplomat added that the trajectory points directly at known Syrian military bases. "There isn't a shred of evidence in the other direction," he said.
Not a whole lot to say except that UN report includes some sort of crater analysis that is apparently a high point of the evidence arrayed against the Syrian government.

Let me be the first to say, I have no idea what happened and this is not an attempt to implicate either party.  It is an attempt to pour some much needed cold water on a hot situation.

I've conducted over a dozen crater analyses in a combat zone and I have a dead enemy mortar team at the end of that.  As a result, I've come to this situation with some skepticism and experience.  There are several questions we need to ask about the analysis that was conducted.

*Note* I also used some pics from this site to see what the alleged chemical weapons craters looked like.

Question 1: was the site tampered with at all?
It's a fact that it took multiple weeks before the team arrived at the crater site.  I've seen some pictures of Syrians around the suspected crater and obviously it was several days since the attack.  The dirt is again packed tightly.  In a couple it's quite clear that the shrapnel was gathered up and stored in a particular site.
That suggests that sites were tampered with.  How do we know that the rocket fins were not moved to a different direction?  Who was there, what was done on site?  Since it's a chemical weapons site, you'd assume there was minimal tampering, but the shrapnel in a pile suggests otherwise.
Question 2: did they actually inspect the site?
It's not possible to do this from a picture.  It just isn't.  I had a sergeant tell me I was 180 degrees off in where the crater was based on some video evidence.  When I told him he could come down to the next crater analysis, he conceded a lack of real interest.
It's very easy to play armchair quarterback on this, but without actually standing over the crater and looking over the exact site, you can easily make mistakes.  Things just look different standing exactly on top of it.  Does it make sense that the round came this way?  Could it have come in anything other than a high-angle arc?  You cannot answer that question from a hotel room or even a TOC through a video feed or pictures.  Get a radar if you are going that route.
Question 3: was there anyone amongst the inspectors who has done crater analysis before?
I can imagine that there were a number of chemical weapons inspectors present.  But how many have actually seen a rocket crater before?  Were they surprised to see so much rocket fin still present in the crater?  Does this particular model carry any HE or is it all chemical gas?  Did the craters appear to be coming from a single back azimuth or was it relatively random?
It's super easy to fuck this up.  I got good at it from a combination of necessity and mentoring.  I imagine most police work is similar.  You just need to learn where to look and how to look before you can find what you need to be good at it.  If they sent a bunch of chemos there, it'd be like giving a little child who's never seen it, a basketball and a basketball hoop.  He may get perfect form on his own, but probably not without someone else's guidance or a lot of time and energy invested.
There wasn't much time or energy invested, so I hope someone from the UN that went there knew what to look for.
Question 4: did the inspectors come with a preconceived notion of where the rockets came from?
I cannot tell you how many times I saw shitty crater analysis because "we know" it's from over there.  Mortars make a V back towards their point of origin (POO), artillery usually does the opposite.  Rockets can do either depending on the type and angle of attack.  But at the end of the day, to the untrained eye it's a hole in the ground.  Give it a week and it really is just that.  Especially if someone digs through it.  It's very easy to see what you want to see, and if someone "helps" you think that, it can seem overwhelmingly obvious.  Especially when you're taking fire.
One of the first ones that I did, I confidently assessed that my predecessor had gotten the POO correctly and we laid some suppressing artillery fire on that spot.  My sergeant, though, had the presence of mind to ignore my ignorance and start combing through the actual crater.  He pulled out the fin a little and I saw to my amazement that I was off by about 90 degrees.  I looked again and it became so obvious that I was wrong and had suppressed the wrong site.  But you live and learn.  You get your hands dirty first.  You investigate from a position of ignorance and assemble facts.

I will remain skeptical of their assessment because I know others who have been really wrong and I have been wrong as well.  I hope that this investigation can overcome its weaknesses

PF Khans


  1. I agree with PF Khans that this stuff really is rocket science. Fortunately, there is no need to rush judgement.

  2. Several of your points are addressed in the actual report. Here's the link:

  3. I agree that the determination trajectory based on the pattern of a crater can be complex. But rocket science it is not. I tried it in Nam once and did not get precise results - in my defense the impact area was a rice paddy. But the FO teams could do it well even on mortars or high angle howitzer fire. And the 122mm rockets they threw at us were not that hard to figure out.

    The UN document Andy linked to indicated the projectiles used were some 140mm rockets (M14 with cyrillic markings), and some 330mm rockets. Assuming there was just a small explosive charge in the rocket to release the sarin and no HE warhead then the crater would be fairly small and the rocket itself not destroyed. Wouldn't that make it easier to determine direction? I agree there are a lot of factors involved: high or low angle of fire, type of fuse, soil characteristics, and many more. It would seem to be easier where they impacted in urban or semi-urban areas. Especially those rockets that impacted walls of a house, in that case there is only one general direction the rocket could have come from ruling out the back azimuth and getting it 180 out.

    The key thing for me though is the following statement from the ABCnews article: "The diplomat cited five key details, including the scale of the attack, the quality of the sarin, the type of rockets, the warheads used and the rockets' trajectory." So the analysis of the trajectory is only one factor out of five that the UN is using to lay the blame on Assad's troops.

    By the way, I understood that UN Observers in peacekeeping situations typically are well trained in crater analysis. Perhaps FDChief could throw some light on that since he served in the Sinai.

  4. Is it normal that binary chemical agents don't use a burster charge (which could provoke hostiles to mistake them as HE shots)?
    Do they begin to mix with the firing shock or with the impact?

  5. I am not NBC or CW trained and am in the dark as to when the binary agents mix. But as I understood it there was a burster charge in the Syrian case. That small charge is used to turn the chemical agent into a gaseous mist.

    1. The burster charge is to aerosolize the sarin and to distribute it.

  6. Points well taken; CA is shifty enough that one of the FA branch's highest priorities in the Forties was adapting radar to backtrack projectiles to their firing points; radar is far and away the instrument of choice for counterbattery work.

    That said, Occam's Razor suggests that this probably WAS the government's strike; a false flag operation could have been just as effective with fewer casualties in pointing fingers at Assad...

    The larger geopolitical pressures suggest that the U.S. government would insist on government culpability and Russia would defend the Assad regime regardless of the UN investigation, so in a sense while all of PFK's points here are correct they are as unlikely to have any more impact on the political situation surrounding Syria than did the actual events in the Gulf of Tonkin in August, 1964...

  7. From reading the UN's report, I think they did a pretty solid job on a lot of what they were asked to do.

    They determined quite well that there was a chemical attack, where it was, who it affected, and they made a decent case for what weapon systems were used.

    Their appendix on the munitions used does not really make me feel super confident about their analysis. They have a couple craters that clearly lead back to Syrian government compounds, but this whole area has been getting shelled for weeks from what I've heard. Finding rockets and craters that have similar tail fins is not a guarantee that they are in fact the same weapon system, or that they were used in this attack. There are a lot of assumptions being made about how the attack was carried out and from where.

    Don't get me wrong, I but the UN's case. Chief is right that Occam's Razor and common sense points away from a false flag operation, but that's not a reason to cut corners or take serious leaps of faith.

    I recall a mortar round that hit our base and blew up a large chunk of a HESCO barrier. Upon inspection I found a larger caliber tail fin sticking awkwardly out of the crater. It looked almost like a failed detonation and so we did not try to actively mess with the crater. It looked like the Taliban in the area had gotten a serious upgrade in their weapon system and as a result this caused a mini-freak out on the base.
    Turned out, the ANA had decided to fill this barrier with mostly pre-assembled shrapnel and tail fins and then sprinkled in dirt to fill it up rather than filling it totally with dirt. So the fin we saw was not related to the actual strike.

    Stupid shit happens way too often to say anything conclusively about the blasts without repeated trips or some radar data. It's just too easy to make mistakes and turn suggestions into fact. This sort of forensics is very susceptible to suggestion and deserves greater scrutiny as a result.

    All that said, again, the chemical work they did looks insurmountable. They did a good job as far as I can tell on that end. Common sense suggests it was government forces that did this, but what I read in no way proves it.

    BTW, chief, radars only work on flat terrain. They are less helpful when there are mountains in the way. You need to have an upward and downward trajectory for the radar to work and if one or the other is blocked by something, radar is a really fancy box of can't-do-shit. That was my experience in any event.

    PF Khans

    1. Old counter-mortar/arty radars used to scan the horizon (thus the trouble with mountains/valleys) to pick up the projectiles easily.
      This isn't the only way business can be done, though. You can also trust very much on the doppler effect and brute computing (processing) power.

      The real problem nowadays are the guided or course-correcting munitions which can conceal their exact origin with some early manoeuvres.

      There are countermeasures to this, of course - but this gets really expensive and elaborate.

      By the way; acoustic sensors are still not out of fashion. IIRC the British bought new Land Rover-mobile sets only a decade ago or so. These sets don't have the problem with manoeuvring munitions if they listen for the muzzle bang, not the supersonic crack of howitzers shells and rockets.

    2. Eh, if you can mask the round with a mountain or a building until about its apex, there's not much to distinguish it from any other object flying through the air. Or at least the latest American systems I used couldn't handle it. Of course, we're talking about elevation changes of several thousand meters in a short space, so this isn't exactly the most common terrain. At the bottom of Hindu Kush mountains, your only shot is looking at craters.

      Or I guess the only economic option is to do that. I suppose if you had Special Forces bucks you could just loiter a Predator on station for a week and get a visual confirmation instead of trying to use geometry and physics.

      And the acoustic stuff is interesting but gets rather muddied once you start adding your own shots into the mix. We tried putting one on our TOC, although, they are intended for vehicles. Every time we fired our mortars, it would call out that we were getting shot at.

      PF Khans

    3. Oh, don't I know it. I wish I had a nickel for every time we tried to use our own radar guys to observe our own rounds - so that gave them 50% of the computation cold, since they had the survey data for our firing position - and kept getting "round unobserved". Throw in an unknown enemy position and clutter from terrain...let's just say that counterbattery radar was designed to work like water off a cat's ass on the North German plain...

      Terrain does funky things with acoustics, too; the sound can either bounce off something hard (like a hillside) or get lost in something soft (like a woodline).

      So, yeah; sometimes you gotta go with what you've got.

      But perhaps the hardest lesson I had to learn as a troop was that you can present your higher with everything you've got and if your higher has it fixed in his mind that "this is the way it is" all your evidence isn't going to make a bit of difference. I think this case is one of those times. The McCains are going to seize on the crater evidence as "hard proof" that the government shelled these areas. The Putins are going to point out the same flaws you did and suggest that there is still a question over responsibility. In reality the "right" answer is "probably this but there's still a lot of questions" but that will satisfy the politics of nobody but those of us leaning on the bar watching the geopolitical drunks in the corner...

    4. Sounds like you are assuming that crater analysis is the only evidence for the CW rockets coming from a government position.

  8. @PFK - " I suppose if you had Special Forces bucks you could just loiter a Predator on station for a week and get a visual confirmation instead of trying to use geometry and physics."

    Yeah, or instead of visual, maybe infrared of the rocket plume at launch. Old technology AFAIK but as you say very expensive.

  9. Audio? I did not know it was being used to locate artillery or mortar positions. I read that large city police organizations for years have been trying to adapt it to determine precise location of gunshots. And that the Army was experimenting with it. Did they actually field a system in IQ or AF?

    I had heard or read somewhere that the Luftwaffe used audio to supplement their GCI system against the British night bombers. And supposedly it was fairly effective in figuring out azimuth and even in some cases altitude when their radar was being jammed or spoofed.

    I know the Red Army used seismometers to determine Wehrmacht artillery positions on the Russian steppe. It was fairly accurate there. Especially as the German big guns stayed fairly close to roads so you did not need a lot of triangulation. The road itself could be considered one factor. This was even more effective against the huge guns on rail carriages. One back azimuth from a seismometer and using the intersection with the RR tracks they could put massive counterbattery fire on that train.

    1. FA counterbattery and countermortar sections have acoustic sensors called UTAMS:

      "Another acoustic system is the US Army Unattended Transient Acoustic MASINT Sensor (UTAMS), developed by the U.S. Army Research Laboratory, which detects detect mortar and rocket launches and impacts. UTAMS has three to five acoustic arrays, each with four microphones, a processor, radio link, power source, and a laptop control computer. UTAMS, which was first operational in Iraq, first tested in November 2004 at a Special Forces Operating Base (SFOB) in Iraq. UTAMS was used in conjunction with AN/TPQ-36 and AN/TPQ-37 counter-artillery radar. While UTAMS was intended principally for detecting indirect artillery fire, Special Forces and their fire support officer learned it could pinpoint improvised explosive device (IED) explosions and small arms/rocket-propelled grenade (RPG) fires. It detected Points of Origin (POO) up to 10 kilometers from the sensor.

      Analyzing the UTAMS and radar logs revealed several patterns. The opposing force was firing 60 mm mortars during observed dining hours, presumably since that gave the largest groupings of personnel and the best chance of producing heavy casualties. That would have been obvious from the impact history alone, but these MASINT sensors established a pattern of the enemy firing locations.

      This allowed the US forces to move mortars into range of the firing positions, give coordinates to cannon when the mortars were otherwise committed, and to use attack helicopters as a backup to both. The opponents changed to night fires, which, again, were countered with mortar, artillery, and helicopter fires. They then moved into an urban area where US artillery was not allowed to fire, but a combination of PSYOPS leaflet drops and deliberate near misses convinced the locals not to give sanctuary to the mortar crews.

      Tower-mounted UTAMS array component of UTAMS in the Rocket Launch Spotter (RLS) system

      Originally for a Marine requirement in Afghanistan, UTAMS was combined with electro-optical MASINT to produce the Rocket Launch Spotter (RLS) system useful against both rockets and mortars.

      In the Rocket Launch Spotter (RLS) application,[16] each array consists of four microphones and processing equipment. Analyzing the time delays between an acoustic wavefront’s interaction with each microphone in the array UTAMS provides an azimuth of origin. The azimuth from each tower is reported to the UTAMS processor at the control station, and a POO is triangulated and displayed. The UTAMS subsystem can also detect and locate the point of impact (POI), but, due to the difference between the speeds of sound and light, it may take UTAMS as long as 30 seconds to determine the POO for a rocket launch 13 km away. In this application, the electro-optical component of RLS will detect the rocket POO earlier, while UTAMS may do better with the mortar prediction."

  10. It's pre-WWI tech.

    The newest addition to acoustic sensors was signal processing of the supersonic crack using multiple microphones. Supersonic projectiles give their fligth path away with it and this is how most deployed counter-sniper sensors work.

    The obvious countermeasure is to shoot from not too far away and use some of the Russian subsonic sniper rifles with a suppressor and other measures to eliminate basically all but the bullet's signatures.
    This would leave radars such as the Arena system as the only practical sensor left for counter-sniper sensors.

  11. FDChief:

    The electro-optical component of the Rocket Launch Spotter (RLS) system that you mentioned as part of UTAMS is an infrared sensor per this link:

    And that link also mentions it was originally developed as part of the Tactical Aircraft Directed Infra-Red Countermeasures (TADIRCM) system. TADIRCM was fielded as a defense against shoulder launched SAMs according to this link:

    But I have to believe there are more sensitive IR sensors with a longer range fielded already in more strategic platforms than tactical aircraft. Maybe RC-135s flying figure eights over Turkey or Jordan just beyond the Syrian border could detect those 140mm and for sure the 330mm rocket launches and pinpoint their points of origin and impact. Or perhaps they are mounted on the Predator drones that PFKhans mentioned. Or they could be on a longer loiter drone like Global Hawk. Or U-2s. Or ____ fill in the blanks. Obama was not depending on the UN crater analysis to make his case.

    1. Do you realise Cold War satellites meant to war about ICBM and SLBM launches regularly created false alarms when bombs were dropped in war zones or similar actions created a suitable heat signature?

      One shouldn't trust electronics so much when it's about war or peace.

    2. I agree S O.

      The same false alarms came from other sources also, fires, factory stacks, etc. Lucky those alarms were double checked by humans instead of letting some doomsday machine react.

      One should not trust any single source of intel, whether electronic or otherwise. The same could be said of evidence in judicial proceedings.