the home of online investigations

You can support the work of Bellingcat by donating through the following link:

Buk Launch Site Data in the Dutch Safety Board’s MH17 Investigation

October 17, 2015

By Aric Toler

One of the most contentious and important points of the investigation into the downing of MH17 is the launch site for the Buk missile that caused the tragedy of July 17, 2014. The newly released Dutch Safety Board (DSB) investigative report dismisses all alternative scenarios other than a Buk missile launch, including the possibilities of air-to-air cannon fire and an air-to-air missile. Thus, the remaining and confirmed scenario—downing by a Buk missile carrying a 9N314M warhead—becomes the sole focus of the international investigation. While the DSB did not assign blame (as it is explicitly disallowed from doing this), it did provide vital clues as to who was in possession of the Buk system that downed MH17.

This article will analyze the available information from the DSB report and remarks from the head of the DSB investigation. This information will be compared with the launch sites proposed by Almaz-Antey, Christopher Miller & Roland Oliphant (previous discovered by Ukraine@War and @Wowihay, later confirmed by Bellingcat, and other studies), Correct!v, the Russian Ministry of Defense, and Meduza writer Sergey Parkhomenko. Additionally, this article will parse the reasons for the various conflicting launch sites proposed by Almaz-Antey in the DSB and its own two press conferences.

Flight Path Simulations

The DSB conducted a thorough investigation into the possible flight paths from a Buk missile, taking into account simulations conducted by the Dutch National Aerospace Laboratory (NLR), the Kyiv Research Institute for Forensic Expertise, and the Russian Buk missile manufacturer Almaz-Antey. Additional research from the Netherlands Organisation for Applied Scientific Research (TNO) was used in various models of warhead detonation, though this organization did not carry out their own flight path simulation. Each of the three organizations carrying out flight path simulations used different data sets, as seen below, which assumed that a 9N314M warhead exploded above and to the left of the plane’s nose. As described later in this article, Almaz-Antey’s flight path simulation used the data set provided by the TNO.


The simulation data provided by the NLR is by far the widest, encompassing a 320 square-kilometer area. Speaking with journalists, Tjibbe Joustra, head of the DSB investigation, confirmed that the area that the Buk was fired from was controlled by separatists on July 17, 2014.


The simulation carried out by the Kyiv Research Institute with NLR’s data is much narrower, with only a 4 square-kilometer area:


The Almaz-Antey simulation is not quite as simple as those from the NLR and Kyiv Research Institute. The Russian simulation used the data set provided by the TNO:

Page 145 of DSB report

Page 145 of DSB report

However, as noted in the DSB report, the calculations from Almaz-Antey are consistent with the Kyiv Research Institute and NLR flight path simulations. The Russian Buk manufacturer carried out separate simulations for a 9M38 missile, and another for a 9M38M1 missile.


The red outline with numbered corners should be ignored, as it denotes outdated calculations from the NLR. The Almaz-Antey simulated flight paths for the 9M38 and 9M38M1 missiles are marked by the red and blue outlines (respectively).

The DSB was clear in saying that no other positions, when considering the pattern of damage on MH17, is possible for a Buk missile launch. Thus, all launch sites outside of the combined areas from the three simulated flight paths are impossible.


The three simulated flight paths—including the two from Almaz-Antey with differing missile types—are compatible, with the exception of a small sliver in the northwest of the Almaz-Antey 9M38 missile simulation. However, as will be detailed in the following section, there are major caveats with the Almaz-Antey flight path simulation, as explored in the report’s appendices and Almaz-Antey’s two self-contradicting press conferences.

Almaz-Antey’s Divergent Findings

In viewing the launch site map provided by the DSB, it is easy to initially think that Almaz-Antey is in near-complete agreement with the Ukrainian and Dutch scientists and engineers. However, as Almaz-Antey’s two (self-contradicting) press conferences and Appendix Y of the report illustrate, the Buk missile manufacturer has quite divergent ideas regarding the plane damage pattern and missile detonation point, and thus the launch site location.

The various warhead test locations are listed in Appendix Y, with the most successful and reliable location–provided by Dutch scientists–and by far the least accurate location–provided by Almaz-Antey:


As you see in the table, the azimuth and elevation of the missile in Almaz-Antey’s calculations are drastically different than the accurate models. The difference in azimuth (-27 degrees compared to -72) is due to the differing incoming location, with Zaroshchenske compared to south of Snizhne. The Almaz-Antey calculated elevation of 22 degrees is due to the same reason. According to the rocket trajectories provided by Almaz-Antey, a 9M38M1 missile fired 16km from MH17 will strike the target at an elevation of 22 degrees–the same as the calculations that Almaz-Antey provided to the DSB. The launch site near Zaroshchenske proposed by Almaz-Antey is approximately 16km from the MH17 impact point, and a bit further than 72 degrees away with the calculated azimuth.


The 9M38 missile engine stops after 15 seconds, while the 9M38M1 lasts 20 seconds, altering the missile’s elevation at strike point. In its initial findings, Almaz-Antey found that the 9M38M1 was used with a 9N314M warhead, thus the calculations submitted to the DSB were made under the assumption that the 9M38M1 missile was used. In their first press conference, Almaz-Antey went out of its way to confirm that all of its test data pointed to an 9M38M1 missile with a 9N314M warhead:

aa-press aa-press

Therefore, we can summarize the various Almaz-Antey launch sites after interpreting the detonation point data in Appendix Y. Let’s take a second look at the Almaz-Antey flight path simulations while using the TNO detonation point data (with more credible azimuth and elevation calculations). The northwestern, red zone is for the 9M38 missile, and is quite unlikely due to its close proximity to heavily populated areas. The southeastern blue zone, which closely matches the site proposed by Ukraine, Bellingcat, and others, is calculated for the 9M38M1 missile. The 9M38 missile area was approximately 20 square kilometers, while the 9M38M1 area was 63 square kilometers.


In summary, the detonation point data provided by Almaz-Antey was wildly divergent from the multiple Dutch sets of detonation point data, even when using the correct missile type of the 9M38M1. However, when presented with the detonation point data set and missile type that were confirmed by numerous experiments, Almaz-Antey’s findings agree with the findings of Bellingcat and other analysts.

Comparisons Between All Proposed Launch Sites

The embedded Google Map below compares all available data provided by Almaz-Antey and the DSB on the range of potential launch sites. Four proposed launch sites have been juxtaposed with these areas. Click through the map to view a legend to differentiate the various areas and markers. Thank you to Bellingcat contributor Klement Anders for compiling the area data on the below map.

Clearly, the Miller/Oliphant and Parkhomenko launch sites fit perfectly within the zones calculated by all parties using the most credible detonation point data. The Zaroshchenske sites are, in contrast, not even within the most extreme range of the broad area calculated by the NRL. Additionally, the launch site proposed by Correct!v is not within any calculations from the Dutch, Ukrainians, or Russians.

Aric Toler

Aric Toler started volunteering for Bellingcat in 2014 and has been on staff since 2015. He currently heads up Bellingcat's training efforts and its Eastern Europe/Eurasia research.

Join the Bellingcat Mailing List:

Enter your email address to receive a weekly digest of Bellingcat posts, links to open source research articles, and more.


  1. John Zenwirt

    The truly Fascist “Lake Seliger” is a fake summer camp for young Russians, who listen to hate-the-West PR blaring from various speakers…intense love of Putin and the regime are shouted into them from the first…Lake Seliger is truly fascist training, including introducing the young to various weapons, AK., etc..

  2. Dude

    theresa – October 25th, 2015
    Congratulations! The mindless bigoted, hate filled drivel that you and others have managed to infest this site with…..


    “Welcome to the Russian World – senseless and ruthless” /Alexandre Pouchkine/


    • Dude

      Sorry it should be of course “meaningless” or “nonsensical” instead of “senseless”:

      “Welcome to the Russian World – meaningless and ruthless” /Alexandre Pouchkine/


      • stranger

        The correct citation from Alexander Poushkin is “God, don’t let us see a russian RIOT – meaningless and ruthless” from “Captain’s Doughter” 🙂

        • Dude

          oh yes that’s what you “rousskie” with all your savage orks-“volunteers” (plus regular russian army divisions) brought to Donbass – “a riot, meaningless and ruthless” – and also conpletely set up / manipulated in its entirety by kremlin dwarfs – and you still call it “russian world” (“russki mir”)…

          -pathetic …

          • stranger

            Pushkin wrote about Pugachov’ riot in XVIII… The analogy would be probably color revolutions. Russian world if possible would be only Russia+Belarus+Ukraine

  3. Sean Lamb

    @boggled – some time back
    “They presented a guess that the video was from the day before, but not a proof.
    I admit, there is a possibility they are right, I have yet to find a weather observation in Torez or Snizhne that verifies a Southern wind at the time of the video, but neither have they found a specific observation that it wasn’t.
    So as of now, the video stands as what it is presented to be.
    Taken on July 17th, not the day before.”

    I don’t think we need to worry about the wind direction.
    1. The DSB safety board head confirmed he had seen satellite photos taken at the time of the impact.
    2. The DSB animations all show the BUK firing up into overcast skies to a plane hidden behind a cloud layer – go and view them yourself.
    3. All photos and videos taken in the immediate aftermath show heavy cloud cover.

    It is impossible to argue in good faith that the Svizhne smoke trail photo was taken on the 17th. That doesn’t prove it was taken on the 16th, just that is the most likely explanation as a Su-25 was shot down on the 25th and blamed on “missiles fired from Russia.” From midnight the exclusion altitude was raised to 26 000 feet.

    My final comment is are you really confident that if you fired a Buk missile at 3 different drones flying at top speed at 10 000 m altitude, that the various institutes making claims here would be able to determine the direction that Buk was fired from? Because I don’t, I think they are just making spurious pseudo-scientific claims to support pre-determined locations. That includes the Russian, the Kiev and the Dutch, I think all their claims are pseudo-science.

    It is a certain ironic possibility, but if you believe (as I do) that the rebels almost certainly had a Buk and if you consider believe they would almost certainly have moved in back to Donetsk after the successful strike on the 16th and then moved it to a new location on the 17th. It is quite possible that the place they moved it to was Zaroshchenske – and that the Russians correctly identified a Buk on their aerial photos. So the Dutch are busy “proving” that the missile came from Svizhne – where the rebel’s Buk had been the day before while the Russians are busy “proving” that the missile came from Zaroshchenske, unaware that was where the rebels had moved their battery to. Strange though that may seem, you should not assume perfect information flows between the rebels and all (or any) arms of the Russian government.

    However, neither a Buk from Svizhne or a Buk from Zaroshchenske are going to do a little tour of the south-east Lugansk suburbs in the early hours of the 18th. And that is the Buk we are after.

    • Sean Lamb

      Sheesh: “Su-25 was shot down on the 16th”
      Not: “Su-25 was shot down on the 25th”

    • boggled

      The most complete weather data collected and graphed out I can find nearest there is Mariupol.!dashboard;ws=33802

      You can move back in history for that fateful day.
      Although it is not conclusive, it does show you some South winds were present.
      You can click on the button graphs and pick the variables you want shown.
      I question the graph a little bit because it shows a large gap of clear skies between 12.00 and 18.00, but it is the best collection of data I can find so far.
      Could it be while primary radar was down for scheduled maintenance?
      Mariupol is a land locked sea so I am not sure how much onshore winds in the morning and offshore winds at night would affect the observations that far inland.

      The darker blocks I believe would show recordings of South winds in the observed time frame, if you had click wind direction on the ‘graphs’ button.
      It shows winds in the 12.00 and 15.00 time frame.
      The arrows would show observations at the precise time of scheduled observation, 3 hour blocks looks like what it is for the arrows.
      I am not sure how observation data is collected and made, so those observations could be observations from the previous time frame or for the next.

      Regardless, I believe East winds could even have produced the winds seen in the video.
      The apartments act as a wind block and the winds ‘swirl’ around behind the wind blocks as the air mass moves that way.

      Not proof the video is from that day, just proof it could have been made that day.
      And I think enough proof that Hector and others cannot prove it wasn’t with their ridiculous assertions from forecasts and not actual observations.
      What I produced were observations.
      The wind observation was only important as far as them trying to prove that video was made another day.

      As far as your doubts of the smoke plume.
      Easy enough to prove, take your own digital camera and take two shots on a day with over 75 percent cloud cover of high clouds.
      There were not specifically storm clouds, although they did roll in in different videos that were geolocated around the crash site of the crash and so therefore showing different directions of sky.

      After you take your first image from wide angle, zoom in.
      Can you tell that it is overcast, just hazy, or blue sky on the zoomed in image?
      Make sure you wait until the sun is behind a cloud when you do this.
      The lower light level is a factor that must be included.

      As far as identifying direction of a missile launch, yes I believe with the advances in computers the simulation can be created with all the variables.
      They did it on paper a long time ago and various destruction models have been used and created to test the effectiveness and to improve SAM technology for years now.
      To deny they cannot take all the variables into account and create a computer simulation, one just has to watch the movie Avatar to know how effective computer programmers are at managing variables and making a simulation.
      Do you honestly think the RF signed off on the DSB report if they believe the investigators did not do a blind study into the launch location?
      They could give the investigators only the string and impact locations and ask them to determine the orientation of the missile.
      Then compartmentalize that and have another group look into launch location from orientation numbers and see what they determine.
      The majority agree on the orientation, except for the RF.
      To me that says the RF is working backward from Z and determining the location that makes it possible with impact sites and forcing their math by saying we need to take into account ricochets inside the plane also.
      Is their location Z possible with their placement?
      Of course, they worked backwards to make that location possible.
      The biggest item A-A and the other RF investigators do not account for is the BUK missile’s own targeting profile in the radar seeker and software integrated in that and the proximity fuse.

      I think were the RF investigators manipulate or confuse some is they utilize the programming from the 9m38 missile and not the programming from the 9m38m1.
      The DSB have both and they know what missile it is.
      So they have the programming and the code installed in it.
      One of the variables to take into account and an important one.

      Many witnesses were interviewed in Zaroshchenske, NONE identified a BUK or missile launch or fighting on that day.
      Snizhne and Torez did have multiple witnesses.
      Only so many roads you can travel across with that big rig and BUK on the back, and you would have been noticed and heard.

      In a covert op set up by the Kremlin to destroy a civilian airliner so RF could come in in a peacekeeping operation because Girkin was losing so badly.
      There would not be complete disclosure to all their proxies, just the ones that were part of the operation.
      So no, I do not expect everyone was ‘in the know’ about a BUK being transported and used.

      After the fall of MH17, the Kremlin and her proxies wanted to eliminate all aspects of their involvement, so a trip through Luhansk is highly likely and probable when you consider they would not take the shortest route that would drive through Ukrainian held territory.
      Especially when you considered they wanted to group all the units and troops together and smuggle them out through Russian Federation controlled borders.

      As far as cloud cover, I agree, it was mostly cloudy.
      BUT there were patches of blue sky as seen at 1:02 of this video.
      If winds are correctly blowing from the East, the clouds are moving somewhat from Russia over the crash site and more blue skies are North and West.
      And would have been over various regions at various times previous to that.
      ie, it was not cloudy all day and all locations.
      Hard to tell exactly what is the actual cloudy cover percentage.
      BUT I will agree depending on the direction you faced, some appear overcast, some appear rain clouds, some blue sky.

      Fare thee well

      • Sean Lamb

        “Do you honestly think the RF signed off on the DSB report if they believe the investigators did not do a blind study into the launch location?”

        Yes, if they did a blind study showing that they could identify a launch location by examination of shrapnel marks of airplane shell, they would have mentioned in the report.
        It would be expensive, time consuming and involve the co-operation of a nation that owned a Buk system.

        So of course no one has done a blinded study of such an esoteric question.

        Even if there were a shadow of a possibility that the Svizhne photo was from the 17th (and since the satellite photos taken at the exact time of the launch have been viewed by the DSB this question can be answered definitively by someone even if not by us) there is also the photo in Topez of the Buk being moved back to Donetsk later that evening – also with a clear blue sky. This was despite on the 17th the cloud cover was increasing as the evening came on.
        Both these photos were taken the previous day.

        Why do you think there is no satellite photo of the Buk at Svizhne on the 17th? I believe there is a satellite photo showing some caterpillar tracks in an empty field. There are only 2 possible reasons.

        1. The location didn’t contain a Buk launcher
        2. The location was concealed by cloud
        or – and most likely – both 1 and 2 are correct.

        Either way it kills the Svizhne smoke trail photo dead. Or rather a genuine photo recording the downing of a Su-25 the previous day.

        • boggled

          AS far as the investigation procedure, I suggest you look into this document.

          pages 20-27 are relevant

          As far as people that contributed, I believe both Finland and Ukraine have a large enough supply of these systems and missiles.

          Torez image from later that evening after MH17 was shoot down???
          Is that what your stating?
          And if I may ask, how do you or someone determine that?
          12:30 was the generally guessed at time.
          Who said evening?

          Third possible reason to your ‘only two possible reasons’, satellite took the image after the BUK had killed and moved off?

          You do know in some imaging the clouds are less relevant?
          Certain spectra have differing abilities to see through things.
          You cannot see your skeleton with your unaided eyes, but an x-ray machine can and then produce an image your eyes can view.
          So that is another possible reason.

          Fare thee well

  4. Cowboy Bob

    Is any of this the relevant data? Regardless of where the missile was fired from, it WAS fired by separatists if not actively serving Russian soldiers. Radio and cellular telephone intercepts have proved this well enough to be unimpeachable. The battery commander thought they were firing at a troop transport carrying Ukrainian soldiers. Why no one bothered to confirm the target as is the professional military habit to PREVENT “friendly fire”, a procedure clearly not adhered to nor adopted by the Russian military who are obviously still using battlefield tactics from the time of Zhukov, is certainly a viable line of questioning which should be brought before the International Court for likely war crimes investigation.

  5. Djajakon

    Is it perhaps possible to translate these kind of articles in Russian as well in the future?
    Kind regards

    • Aric Toler

      Hi Djajakon,
      I agree! We will soon, if things go to plan, launch a Russian-language section of Bellingcat. All of our articles–not just the ones about Russia and Ukraine–will hopefully be available на русском in 2016. So, fingers crossed.

      • John Zenwirt

        Launching a Russian language version of this site makes no sense, as no one here reads Russian, right…? Is Bellingcat going to just become an all-Russian site, there are numerous Russian-language topics up @ Bellingcat, why make yourselves incomprehensible to more than 90% of your readers…?

        • boggled

          To me, it makes good sense.
          There is a lot of Kremlin sponsored media/propaganda.
          There are a lot of brainwashed Russians who do not go to English based sites.
          There are also many Ukrainians and other Slavic races who have trouble with English and Russian is easier.
          Since a lot of problems are Kremlin related that are reported here (organized crime has its fingers everywhere), it would be good for the Russians to actually be able to have discussions outside of Kremlin control.
          There are a lot of moderate Russians that do not support Russian policies.

          There are even many Russian conscripts refusing to fight in Ukraine for the Kremlin’s hybrid war.
          Part of that I think is the military is waking up to the idea that something stinks in the Kremlin, and it is not Borscht.
          And I would like to think that BC has helped with that and if it could help more, it would be good for the Russian population in general.

          Most of the Russian language articles are just duplicates of the English ones.
          And if you did not know, you can click on the page in most browsers and have the comments translated to English.
          Then you can chose to respond or not.

          Me personally, I would like to know what type of Russian speakers come and read but do not comment because they do not want to be called English illiterate.
          And what comments they have to make.

          I think one of the biggest goals of global peace should be being able to break the brainwashing of various propaganda orgs and push back on them.

          You do not have to read the articles in Russian, I think I remember one in Spanish as well, because they are duplicated and your not missing anything by not reading them.
          I find it interesting to see the mindset and if Mr. Higgins and crew want to do this, more power to them.
          They do work on crowd funding, and if some Russians themselves contribute to future investigations, I think having them in both Russian and English is a bonus.
          Extra work for the Bellingcat’s crew, but the benefits could be great for both them and the world.

          Mr. Higgin’s has invested a lot of time and work and innovation into his work model and has uncovered some amazing things.

          What would your alternative suggestion be?
          Maybe a second site called Bellingkot for the Russian readers?
          He reaches more population by making it easier for some to read, and some of his crew are very good in alternative languages.
          Maybe it will work for them, maybe it won’t.
          Never know unless you try and apparently there are a good number of Russian speakers commenting so far on the Russian articles.
          Think positive, it may be a good thing overall.

          They will keep up the good work and producing very interesting articles, just going to add another facet to it and see how it works.

          Fare thee well

  6. RHK

    Really good-spirited discussion and clearly on the way to revealing it all. BTW I sat up to the small hours last night reading the CorbettReport view of things broadcast relatively soon after the event. Somehow I don’t think he would put his arguments that way any more after all that has been deduced in the intervening year and a half. Although he professes to be “neutral”, his underlying biasses sometimes seem quaintly obvious. Anything that comes ourt of the American Administration must be a lie! But just listen to Obama’s statement shortly after the event: it fits remerkably well with what we now know to be the probable truth. So this hidden bias is always a problem even for people really trying to be objective – beware! The “but I’m right anyway”-syndrome also creeps in so easily. Corbett’s “cui bono” arguments seem to me to be just too black-and-white. We humans mess up – often accidentally and unintentionally – and then have to cover our tracks in confusion. Could there have been a Ukrainian aircraft trying to hide below MH 17? Would that explain why someone in command in the BUK pressed the button? Who would admit to killing 298 innocent people?

  7. RHK

    Couple of typos – please correct: “comes out of” and “fits remarkably well” Sorry – sight failing a bit!

  8. antitroll

    I wonder why A-A made so different calculations of 9M38 and 9M38M1 near Zaroshchenske vs near Snizhne?

    9M38 seem to fly shorter path when launched near S, but longer path if launched near Z.

    Just some “maskirovska” to fool investigators & press?


Leave a Reply

  • (will not be published)

You can support the work of Bellingcat by donating through the following link: