the home of online investigations

Bellingcat Report – Origin of Artillery Attacks on Ukrainian Military Positions in Eastern Ukraine Between 14 July 2014 and 8 August 2014

February 17, 2015

By Bellingcat Investigation Team

Translations: Русский

 

Summary

Ukrainian armed forces positioned near the Russian-Ukrainian border were attacked by artillery fire in the summer of 2014. Between 9 July and 5 September 2014, the Ukrainian Border Service and the National Security and Defense Council reported more than 120 artillery attacks from Russia. Russian officials, however, have denied the existence of any artillery attacks on Ukraine originating from its territory.

The pressure of sustained artillery attacks through early August led Ukrainian armed forces to lose control of hundreds of kilometers of border territory. Google Earth satellite images of eastern Ukraine from July, August, and September 2014 have enabled the Bellingcat investigation team to find evidence of these artillery attacks, determine their origin, and compare them with local sources.

After reviewing and analyzing these satellite images as well as videos from social media, local media reports, and the shifting maps of the ongoing conflict, the Bellingcat investigation team has determined that there is compelling evidence that artillery attacks on Ukrainian territory and against Ukrainian armed forces originated from the territory of Russia.

A PDF version of this report can be found here.

Deutsche Version des Berichts.

Introduction

In July 2014, Ukrainian armed forces launched an “anti-terrorist operation” against pro-Russian separatists and made significant territorial gains in eastern Ukraine, regaining control over a large portion of the Russian-Ukrainian border.

The situation in the eastern regions of Ukraine on 11 July 2014. Image courtesy of the National Security and Defense Council of Ukraine.

The situation in the eastern regions of Ukraine on 11 July 2014. Image courtesy of the National Security and Defense Council of Ukraine.

 

However, on the morning of 11 July 2014, the situation suddenly changed. A massive and unexpected artillery attack on units in the Zelenopillya region killed 30 and wounded over 100 soldiers. In the days and weeks that followed, the units at the border were subjected to dozens of additional artillery attacks. By late July 2014, the massive bombardment reversed Ukrainian gains and contributed to the encirclement of portions of the Ukrainian armed forces. A separatist offensive across eastern Ukraine began simultaneously with the artillery attacks, leading to the separatists capturing much of the Russian-Ukrainian border along with vast areas of the Donetsk and Lugansk oblasts.

Artillery attacks on Ukrainian forces near the border proved to be an important factor in turning the conflict in favor of the separatists. Using in-depth analysis of satellite imagery and social media sources, this Bellingcat investigation brings new evidence to help clarify to the origins of these artillery attacks against Ukrainian armed forces.

Research methods and sources

The starting point for this Bellingcat investigation was the distribution of updated satellite imagery from Google (DigitalGlobe satellite imagery) with a panchromatic resolution of 0.5m from the territory of eastern Ukraine and its border regions with Russia (17 July to 31 August 2014 satellite images). Additionally, the Bellingcat investigation team analyzed videos shared on social media (YouTube and VKontakte) and geolocated the events captured in these videos to key sites involved in the artillery attacks on Ukraine.

From the satellite imagery, the Bellingcat investigation team located artillery impact crater fields from artillery fire and conducted extensive analysis on a crater-by-crater basis. In the examination of each area, our team created a novel analysis method based on internationally-recognized ‘on-the-ground’ procedures to determine the trajectory of the artillery fire.

The satellite images from eastern Ukraine show two main types of craters, low-angle fuze quick craters (with distinctive ‘side spray’ areas projecting diagonally from a central crater) and high-angle shell craters (triangular-shaped craters that spread outwards towards the origin of fire):

Low-angle fuze quick craters (e.g. low-angle artillery or Multiple Rocket Launcher Systems fire)

Low-angle fuze quick craters (e.g. low-angle artillery or Multiple Rocket Launcher Systems fire)

High-angle shell craters (e.g. mortars, high-angle Multiple Rocket Launcher Systems fire)

High-angle shell craters (e.g. mortars, high-angle Multiple Rocket Launcher Systems fire)

Both types of craters are suitable for determining the trajectory of artillery fire. In reality, the trajectory of the projectile is determined by a variety of factors, such as the type and hardness of the ground struck, wind direction and speed, and the type of projectile. Our team adopted a simple linear trajectory in our analysis, but the actual trajectory may vary somewhat due to the aforementioned factors.

The Amvrosiivka attack 14 July 2014

In a 14 July 2014 summary of the “anti-terrorist operation,” Ukrainian media reported that an attack took place on positions of the Ukrainian armed forces in the vicinity of Amvrosiivka. It was suspected that the origin of this attack was from the territory of Russia.

On the satellite image below, taken on 16 July 2014, an extensive crater field south of Amvrosiivka is visible, located at the coordinates 47°45’52.38″N 38°30’47.65″E.

The crater field near Amvrosiivka from the 14 July 2014 attack, positioned at 47°46'1.07" N 38°30'43.16" E. Google Earth satellite image date 16/07/2014.

The crater field near Amvrosiivka from the 14 July 2014 attack, positioned at 47°46’1.07″ N 38°30’43.16″ E. Google Earth satellite image date 16/07/2014.

The Bellingcat investigation team counted a total of 330 craters in this crater field, and determined the observable direction of each impact crater based on the analysis methods previously described. From these 330 craters, the average trajectory was calculated, and was determined to be 193.97°, i.e. from the south-south west (180 ° being due south). When screening for possible firing positions from this trajectory, the Bellingcat investigation team found a firing position 14.6km from the crater field. Burn marks are visible at this location, which is on Russian territory and approximately 750m from the border near the Russian village of Seleznev at the coordinates 47°38’13.52″N 38°28’9.69″E.

Crater field and firing position, viewing north-north-east

Crater field and firing position, viewing north-north-east

Based on our crater analysis, we judge that these were very likely the five firing sites that caused the craters near Amvrosiivka. There are several pieces of evidence that lead us to believe that these strikes were performed by five 122mm BM-21 Grad or 9K51M Tornado-G Multiple Rocket Launcher Systems (MLRS) as opposed to BM-27 Uragan, the BM-30 Smerch, or other types of field artillery.

It is well known that BM MLRS create large areas of smoke behind them during and after firing, of which many examples are available online. This area of intense smoke and heat results in burned/singed ground behind the MLRS, and are visible from Google map satellite images. Such burned areas are visible at the firing position considered for the Amvrosiivka attack, and so we judge that these were very likely MLRSs.

Firing position near Seleznev. Position: 47°38'14.38"N, 38°28'9.77"E. Google Earth satellite image date 16/07/2014.

Firing position near Seleznev. Position: 47°38’14.38″N, 38°28’9.77″E.
Google Earth satellite image date 16/07/2014.

Analyzing the satellite imagery of the firing position also gave us evidence for determining the type of MLRS most likely used to conduct the artillery attack.

The visible tracks that lead to the site come from further inside Russian territory.

The visible tracks that lead to the site come from further inside Russian territory.

The tracks of the vehicles that entered and exited the field to reach their firing positions are visible from the satellite imagery. This leads us to believe that there was no cross-border (Ukraine to Russia) movement of military equipment for this particular location.

In the Russian military three MLRS systems are primarily used (see Figure Russian MLRS systems).

Russian MLRS systems

Russian MLRS systems

The measurement of the average track width from the satellite images yielded an average value of 2.40m. As the resolution of the satellite images is 0.5m, the tolerance for this measurement is 1.9 to 2.9m. In comparison, the widths of the BM-21/Tornado-G, BM-27, and BM-30 are 2.4, 2.8, and 3.1m, respectively. The track width determined from satellite imagery is suggestive of a smaller vehicle, but because of the potential error due to imagery resolution, this is not conclusive.

Using the satellite images, we were able to determine the turning radii of the vehicles. The smallest turning radius was measured as 14m. The BM-27 and BM-30 MLRS have two steered front axles, so if they were used in this instance, we would expect to see evidence of two overlapping sets of wheel tracks in the turns leading to or from the firing position. However, the satellite image shows only the traces of one clear, single track. Based on this evidence, we conclude with high probability that a BM-21 Grad or 9K51M  Tornado-G was used for this attack.

The spacing between, and the sizes of the types of MLRS units judged to be at the firing position

The spacing between, and the sizes of the types of MLRS units judged to be at the firing position

The visible traces of the vehicles also fit well for the typical spacing of BM-21/Tornado-G systems in battle formation (see the image above for the spacing of the units visible from the satellite image).

Our investigation of this attack shows that the average trajectory of the craters pointed toward the direction of the suspected firing positions with a high degree of accuracy; the maximum-to-minimum deviation of the analyzed crater angles is only 0.2%. Additionally, a measurement in the other direction, extrapolating the trajectories of the visible burn marks at the firing position, revealed that the trajectories led directly to the center of the artillery crater field.

 

Artillery attacks in the Chervonopartyzansk region 14 July to 8 August 2014

In early July 2014, units from the 72nd Motorized Brigade, the 79th Airmobile Brigade, the 24th and 51st Mechanized Brigades, and elements of the 3rd Separate Special Forces Regiment were tasked with securing the Russian-Ukrainian border south of the separatist-held towns from Marynivka to the Izvaryne checkpoint. This area, covering over 150km of border, was known as “Sector D”.

From the end of the ceasefire on 1 until the 11 July 2014, Ukrainian units advanced quickly and found relatively little resistance until they reached the Chervonopartyzansk/Gukovo checkpoint. Then, the situation changed dramatically. The Zelenopillya artillery attack against Ukrainian forces on 11 July was followed by artillery attacks against large conglomerations of Ukrainian units in the area close to Dolzhanskaya-Capital mine and Panchenkove, Chervonopartyzansk, Khmelnytskyi, Biryukove, Dolzhanskyi border control point, and the Dyakove region. As the world diverted their attention on the horror of the Malaysian Airlines Flight 17 (MH17) shoot down on 17 July 2014, the situation sharply deteriorated for the Ukrainian units on the eastern border.

On 22 July 2014, Dmytrivka, a town on the supply road to Sector D, came under attack from separatists. For the next two weeks, transport along the roads east of these towns greatly limited resupply and reinforcement for the trapped units (see examples here, here, and here). From interviews and media reports, it was evident that conditions were dire. Despite this, the units on the border continued to hold their positions.

On 3 August, a breakthrough by Ukrainian units in the west unblocked a potential resupply route (between Dyakove and Dmytrivka) to the beleaguered units on the border. That same day, pro-Russian forces sharply increased the number of attacks against the trapped units on the border. Out of ammunition, on 4 August, over 400 Ukrainian soldiers from the 72nd Mechanized Brigade fled across the border to escape the constant bombardment. By 8 August, the remainder of the 72nd and 79th brigades had broken out of the encirclement, and around 1,000 survivors were able to regroup with other units near Savur-Mohyla.

Massed artillery attacks upon Ukrainian units in the border region east of Dmytrivka were undoubtedly a key factor in the retreat of Ukrainian units from Sector D. However, until now, the trajectories and firing positions of the artillery attacks that led to this defeat have not been analyzed. In the following section, we analyze two of the largest visible artillery attacks close to Chervonopartyzansk and Sverdlovsk, where Ukrainian units were hit by hundreds of visible artillery shells during July and August 2014.

The Dolzhanskaya-Capital mine/Panchenkove attacks 14 July 2014 to 8 August 2014

On 14 July 2014, consistent accounts emerged in both Ukrainian and Russian media outlets that clashes took place in the vicinity of the mine called ‘Dolzhanskaya-Capital’, close to the Ukrainian town of Panchenkove.

Crater fields near Dolzhanskaya-Capital mine

Crater fields near Dolzhanskaya-Capital mine

Significant disruptions to the mine’s power system affected pumps and the ventilation system, leading to the evacuation of 278 miners.

The satellite image from 8 August 2014 shows the cause of the power outage: large amounts artillery fire that destroyed several electricity poles in the vicinity of the mine. Analysis of the damaged electricity poles and their connections suggested that the only lines leading directly to the mine were affected. Therefore there was no evidence to suggest that power to nearby towns was disrupted due to this attack.

The Bellingcat investigation team researched media reports from mid-July to confirm attacks from this time period. Official Ukrainian government sources reported artillery attacks on 23 July and 1 August, and Ukrainian media sources also reported attacks in this area on 16 and 24 July, and 1-4 August. The following comparison of maps provided by separatist sources shows the course of the battles in July and August throughout the region and also suggests several specific dates for attacks on Ukrainian forces.

attles southeast of Sverdlovsk from 11 to 24 July 2014 (map from pro-russian sites)

Battles southeast of Sverdlovsk from 11 to 24 July 2014 (map from pro-Russian sites)

Battles southeast of Sverdlovsk from mid-July to early August 2014 (map from pro-russian sites)

Battles southeast of Sverdlovsk from mid-July to early August 2014 (map from pro-Russian sites)

The Bellingcat investigation team analyzed a number of videos shared on YouTube and VKontakte (VK), Russia’s most popular social networking site. From these sources, it was established that at least two attacks occurred on 16 July 2014.

Next, our team analyzed the crater field near the Dolzhanskaya-Capital mine outside of the village Panchenkove by using satellite imagery from 8 August 2014. A total of 813 craters were measured.

Crater fields near the Dolzhanskaya-Capital mine

Crater fields near the Dolzhanskaya-Capital mine

Analysis of the calculated trajectories in the attacked crater field reveals that there were six separate attacks from five different directions.

Crater fields near the Dolzhanskaya-Capital mine showing six separate firing trajectories

Crater fields near the Dolzhanskaya-Capital mine showing six separate firing trajectories

With maximum differences of 300m at a shelling distance of 15km, all of the calculated trajectories can be traced to five separate firing positions. Every one of these – with the exception of the position near Chervonopartyzansk, Ukraine, which is close to the border – is clearly within the territory of Russia.

The firing positions, from north to south, were located in the following positions:

1) Near Nizhnyaya Kovalevka, Russia at 48°07’51.4″N, 39°54’02.4″E

2) Near Chervonopartyzansk, Ukraine at 48°03’03.2″N, 39°49’52.2″E

Firing position near Nizhnyaya Kovalevka, Russia

Firing position near Nizhnyaya Kovalevka, Russia

Firing position near Chervonopartyzansk, Ukraine

Firing position near Chervonopartyzansk, Ukraine

3. Near Pavlovka, Russia at 47°56’28.1″N, 39°49’12.0″E

4. Near Ukrainskiy, Russia at 47°58’23.0″N, 39°51’01.5″E

Firing position near Ukrainskiy, Russia

Firing position near Ukrainskiy, Russia

Firing position near Ukrainskiy, Russia

Firing position near Ukrainskiy, Russia

5. Near Gukovo, Russia at 48°03’33.1″N, 39°57’22.6″E

The firing position near Gukovo is particularly noteworthy because numerous videos have been shared on social media showing artillery fire. The firing position shown in these videos will be examined separately in the following section.

Crater fields near the Dolzhanskaya-Capital mine – located firing positions

Crater fields near the Dolzhanskaya-Capital mine – located firing positions

The Bellingcat investigation team has also found that at all of the firing positions except one (Chervonopartyzansk), vehicle tracks show both an arrival from and departure to areas within Russia. In other words, in four out of five cases there were no observable tracks linking the firing positions to Ukrainian territory, and all the observable tracks near the firing sites were exclusively within the territory of Russia.

The firing position near Gukovo

On 17 July 2014, a series of videos were shared on YouTube and VK showing an attack with MLRS launchers. The attack occurred on 16 July 2014 in the vicinity of Gukovo, Russia. This attack has already been located by a number of blogs and articles, for example by the Interpreter Magazine.

The Bellingcat investigation team has found six videos on social networks that show the events of 16 July 2014 near Gukovo. Two of these videos were captured from the edge of the Kovalevsky pond on the afternoon of 16 July 2014.

Video001

г. Гуково (РФ), град бьет по Украине
“City of Gukovo (RF), a Grad strikes at Ukraine”
added 16.07.2014
Camera location 48°4’30.76″N 39°55’36.94″E

Alternative link: https://www.youtube.com/watch?v=6SIDxVxdIjw

Video001 shows several MLRS salvos west of the pond. This video has been successfully geolocated a number of times by various journalists, bloggers, and open source investigators.

Video002

Обстрел ГРАДами территории Украины. г. Гуково, Ростовская область.
“Bombardment of the territory of Ukraine with Grads, city of Gukovo, Rostov oblast“
16.07.2014, 17.05 (MSK)
added 16.07.2014
Camera location 48° 4’15.08″N 39°55’24.16″E

Alternative link: https://www.youtube.com/watch?v=zRCjUPa7qHU

Video002 shows burning fields northwest of the same pond that appears in Video001. It was suspected at the time that this fire came from MLRS launchers.

Other Grad rocket volleys were documented in four other videos on the same day in the evening. Each of these videos was captured exactly at sunset, leading to our determination that the videos were captured at approximately 19:15 (MSK, local time).

An important feature of this collection of videos is that each was captured from different angles facing the firing position. The Bellingcat investigation team was able to precisely determine the location of the firing position from the exact locations of the cameras and the angle and positions of the Grad launchers in each video.

Video003

Град в Гуково
“Grad in Gukovo”
added 16.07.2014
Camera location: 48°03’59.0″N 39°55’36.2″E

Alternative Link: https://www.youtube.com/watch?v=edE_FshuF2U

Video003 was captured from a cherry tree on a residential plot on Bazarnaya Street 84 in Gukovo. The identifying features of this video include buildings in the center of the video, the cherry tree as camera location, and the angle and distance from the firing position.

Camera location Video003 – yellow-green lines in the middle point towards the firing position

Camera location Video003 – yellow-green lines in the middle point towards the firing position

Video004

Залп российских ГРАДов из Гуково в сторону Украины
“Salvo of Russian Grads from Gukovo in the direction of Ukraine“
added 16.07.2014
Camera location: 48° 4’29.80″N 39°55’36.07″E

Alternative link: https://www.youtube.com/watch?v=8F9WFkA9C08

Video004 shows the firing position on the southwest side of Kovalevsky pond. The identifying landscape features in this video are the solitary tree on the right side of the camera’s perspective, the big bush in roughly the middle of the perspective, and the utility poles on the far left side of the perspective.

Camera location of Video004 – light green lines in the middle point toward the firing position

Camera location of Video004 – light green lines in the middle point toward the firing position

Video005

Град Гуково,Ростовской обл.,Россия по Украине Russian attacked Ukraine from Gukovo, Rostov Oblast
added 16.07.2014
Camera location: 48° 4’24.30″N 39°55’37.12″E

Alternative link: https://www.youtube.com/watch?v=6ufW8UxbkpM

The firing position in Video005 is visible from the southwest side of Kovalevsky pond. The identifying landscape features in this video are the same as Video004: a big bush, and utility poles.

Camera location of Video005 – blue lines in the middle point toward the firing position

Camera location of Video005 – blue lines in the middle point toward the firing position

Video006

Обстрел территории Украины из Гуково, россия
“Shelling of the territory of Ukraine from Gukovo, Russia”
Added Jul 16, 2014 at 10:18 pm
Camera location: 48° 3’33.15″N 39°57’22.60″E

Alternative link: https://www.youtube.com/watch?v=TsmcdunS0LA

Video006 was captured from a residential building on Botanicheskaya Street 7a in Gukovo. The identifying features of this video include a building on the right side of the video, another building on the left side, the road below, and the angle and distance from the firing position.

Camera location of Video006 –dark purple lines show the direction of the firing position

Camera location of Video006 –dark purple lines show the direction of the firing position

The intersections of the viewing directions to the firing position give the exact location

The intersections of the viewing directions to the firing position give the exact location

On the satellite image below, taken 8 August 2014, the intersection of the lines clearly shows two visible burns, marking the firing position.

The intersection of viewing directions and the location of Grad firing positions

The intersection of viewing directions and the location of Grad firing positions

The positions of the cameras in the previous videos clearly point to a firing position north of Platovo. These firing positions are visible from satellite imagery of the burned fields (image above).

As evident from the satellite imagery, there are two visible firing positions for four Grad MLRS launchers at 48°05’25.0″N 39°54’45.3″E. The Bellingcat investigation team has measured the direction of the traces of fire and calculated a trajectory for each firing position.

Trajectories analyzed from the direction of the burn marks

Trajectories analyzed from the direction of the burn marks

The calculated trajectories of the burn marks point towards two different crater fields on Ukrainian territory (image below). After analysis of the individual craters in the artillery strike site, we found that the calculated trajectories of the firing positions fit almost exactly the two trajectories determined from the artillery crater analysis of the crater field near Dolzhanskaya-Capital mine.

Trajectories from the firing position north of Platovo to the target area around the Dolzhanskaya-Capital mine

Trajectories from the firing position north of Platovo to the target area around the Dolzhanskaya-Capital mine

Thus, from the evidence of crater field analysis, social media geolocation, and local media reports, the Bellingcat investigation team has determined that the artillery attacks on the Ukrainian armed forces positions near the Dolzhanskaya-Capital mine were conducted on 16 July 2014 from firing positions near Gukovo, within the territory of Russia.

Video footage of the aftermath of the attacks

On 12 August, the separatists, via their so-called “Information and Analysis Agency South East” posted two videos entitled “Destroyed position of the ‘heroic’ 72nd Brigade APU near Sverdlovsk, Lugansk region.”

The Bellingcat investigation team has located the camera position of these videos as being in the crater field in the vicinity of the Dolzhanskaya-Capital mine.

The same destroyed military vehicles are clearly visible in the 8 August 2014 satellite image from Google Earth, taken just after the Ukrainian forces pulled out of this border area.

Camera location for the scene from the video at 1:05

Camera location for the scene from the video at 1:05

Other shots of the videos

Other shots of the videos

These videos show the extent of the destruction of the Ukrainian forces’ equipment based close to the Dolzhanskaya-Capital mine. Many heavily damaged trucks and armored vehicles are visible in the video, including a truck, several BMP-2 units, a self-propelled artillery unit, BM-21 ‘Grad’ launchers, and tanks.

The Khmelnytskyi attack 25 July 2014

On 28 July 2014, the Russian news agency Ruptly published a video on YouTube entitled”Ukraine: Battle aftermath litters after Sverdlovsk militia pummels 72nd Motorized Brigade.”

Shot from Ruptly Video

Shot from Ruptly Video

Several abandoned and damaged armored vehicles are visible in this video. The video description also gives us a date for the attack:

Ukrainian Army BMP-2, MT-LB and rocket launchers were left abandoned near Sverdlovsk on Saturday, after the 72nd Motorized Brigade suffered heavy losses during mortar shelling from the Lugansk People’s Militia on Friday.

Lugansk People’s Militia has said that heavy losses were suffered by the Ukrainian army in both hardware and personnel after the Kiev troops were hit by the 12mm [sic] mortar shells.

The video was released on 28 July 2014, which means that the day of the attack (the previous Friday, as mentioned in the description) was the 25 July 2014.

Using the video footage, the Bellingcat investigation team has located the camera location of the Ruptly video as being a crater field close to the village of Khmelnytskyi, south of Sverdlovsk.

Comparison shot from Ruptly video and the 15 August 2014 satellite image from Google Earth

Comparison shot from Ruptly video and the 15 August 2014 satellite image from Google Earth

Analyzing the satellite imagery, The Bellingcat investigation team found a total of 209 craters. We were also able to determine three primary attack trajectories.

Result of the crater analysis at the village Khmelnytskyi, south of Sverdlovsk

Result of the crater analysis at the village Khmelnytskyi, south of Sverdlovsk

The trajectories of nine craters point directly east (green line). Two other trajectories (red and cyan lines) originate in a southeasterly direction. All three trajectories clearly point to firing positions within Russian territory.

Firing position 1 and military camp (left) and Firing position 2 and military camp (right) Both Google Earth satellite images from 08/08/2014

Firing position 1 and military camp (left) and Firing position 2 and military camp (right)
Both Google Earth satellite images from 08/08/2014

These firing positions are directly connected to the military camp of the Russian army near Pavlovka and a smaller firing position close to Malyy.

Conclusion

The Bellingcat investigation team used internationally-recognized methods and satellite imagery to analyze a total of 1,353 artillery craters in eastern Ukraine and determine their trajectories. We located firing positions that closely matched these trajectories, all of which were inside Russian territory with one exception (which was less than 2km from the Russian border).

Three artillery attack case studies were investigated in this report: Amvrosiivka (14 July 2014), between Dolzhanskaya-Capital mine and the village of Panchenkove (16 July to 8 August 2014), and Khmelnytskyi (25 July 2014). Artillery crater analysis concluded that there were a total of ten primary attack trajectories across all the case studies. From each of these attack trajectories a firing position was identified. Nine of these firing positions were – without any doubt – within the territory of Russia, with three within 400 to 800m of a military camp. The one position in Ukraine was near Chervonopartyzansk, within the territory of Ukraine 1.2km south, and 1.5km north of the border with Russia.

The target trajectory of three firing positions was determined through analysis of scorch marks created by MLRS fire. All three of these trajectories led precisely to the target area of the crater field, allowing our team to calculate the trajectories in both directions (i.e., from crater field to firing position and from firing position to crater field). Vehicle tracks are also visible and showed clearly that they originated from, and led back to Russian territory near seven firing positions.

The firing position for an attack on 16 July 2014 in a field north of Gukovo, Russia was documented in a series of videos found on social media. Four of these videos show the same instance of 122mm MLRS fire from different perspectives. By comparing the video locations and views, the Bellingcat investigation team was able to geolocate the exact firing position. With a small deviation (300m at a shelling distance of 15km), the firing position geolocated through video analysis was the same as that from the crater analysis.

This study has provided compelling evidence that a series of artillery attacks on Ukrainian territory were conducted between 14 July and 8 August 2014 from firing positions within Russia. The first attack on 14 July 2014 originated from Russian territory near the Russian village of Seleznev and was directed towards positions of the Ukrainian armed forces south of the Ukrainian village of Amvrosiivka. Four attacks were conducted near Gukovo, Russian on Ukrainian armed forces close to the Dolzhanskaya-Capital mine and the village of Panchenkove between 16 July and 8 August 2014. Two additional attacks were conducted on this area from Russian territory north of Gukovo. On 25 July 2014, an artillery attack was conducted on the positions of the Ukrainian 72nd Mechanized Brigade near Khmelnytskyi village, south of Sverdlovsk. These attacks also came from locations within Russia, specifically from a military base in Pavlovka and a position east of the village of Malyy.

Acknowledgments

Based on the original work of Sean Case

The Investigation team:

Timmi Allen

Sean Case

Eliot Higgins

Veli-Pekka Kivimäki

Iggy Ostanin

Aric Toler

 

Editorial support provided by Nathan Patin

 

This report was created collaboratively using Slack.com

 

Bellingcat Investigation Team

The Bellingcat Investigation Team is an award winning group of volunteers and full time investigators who make up the core of the Bellingcat's investigative efforts.

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.

96 Comments

  1. Factism

    I thing, it would be nice to provide some extra information on the crater analysis data like standard deviation associated with the scattering of the trajectory angle and mean error. In that way, you can state that the deviation between deduced angle and the suspected firing positions lies within a certain confidence bound (i.e. 95%), which may be quantitatively very convincing.

    Reply
  2. Mark Fielding

    I posted this on reddit and you didn’t reply, maybe you could answer?

    I hope you don’t mind me asking some questions on your very interesting report. I have some friends who work for the UN mine action teams so they do this at work, I sent them the link to your report

    1) The Sri Lanka report says Crater analysis from satellite imagery is problematic, and site visits are needed to confirm presence and origin of the identified possible craters. What was your sample size of site visits?

    2) The report says ‘Unfortunately, budgetary constraints do not allow analysis of all potential artillery sites at present, and it is unlikely sufficient imagery exists for a complete review.’ So how much additional imagery did you need?

    3) UN standards on crater examination are in line with (copied from) US standards. http://www.globalsecurity.org/military/library/policy/army/fm/6-121/fm612_9.htm These standards don’t permit side spray method for use with photographs. How did you amend the method to allow for side spray?

    4) Splinter groove method uses photographic evidence as step 1 but then you still have to inspect the craters. Again what was your sample size?

    5) Panchromatic 0.5 satellites have a standard scale of 1:2500 . Therefore 1mm of error translates to 25m on the ground. What image enhancing tools did you use to ensure no loss of resolution for accurate calculation of crater depth?

    6) The depth of the crater will vary hugely between marshland and volcanic rock. Even between sand, loam, clay etc the difference will be pronounced. What form of soil analysis did you use and how did you apply this to such a wide area? There were a few days between the shelling and the pictures being taken, what were the weather patterns during those days and how did you adjust your calculations for them?

    7) Craters have internal and external craters based on impact and explosion. What crater enhancing models did you use to recreate the relationship between internal and external craters?

    8) To calculate the trajectory you need to know what the material was that you hit, this will come from your soil sample. In addition the angle the shell hit the ground at (how did you calculate that from the photographs?) and the velocity of the shell (did you recreate the launch to get an accurate reading?).

    9) The report on Sri Lanka say again that they know the direction of mortar fire because it is a peninsular and they know there is no navy involved and you can see the mortars in the pictures. In your pictures the artillery has gone so did you look at the complete diameter of the range when geolocating before deciding on the Russian location?

    10) Who is interested in using this? And for what? Have you submitted to the UN?

    Reply
    • Factism

      Of course the image measuring approach will be less accurate than actual site visits regarding the evaluation of a single crater. The overall accuracy however will dramatically increase with the number of available samples (say, the standard deviation of a single measurement is +-10° than the error for all 330 samples will be less than 1° within confidence bounds of 95%). I believe, this method is pretty reliable to determine the direction since it simply relies on the symmetry of the surface imprints with regard to the flight direction (irrespective of whether high or low-angle fire). On the other hand it is seems not suitable to determine the firing distance (which was no claim of the present study). Therefore knowledge about elevation angle and impact energy are required which in turn relies on information about depth and material properties of the crater. Obviously the accuracy is limited by resolution of the available image data and degradation of the visible crater edges by weather.

      Reply
  3. Mark Fielding

    I just thought that I would add a caveat to say why these questions are important. An organisation such as the UN uses this in mine clearance for example. Unexploded artillery shells behave something like land mines and have to be removed. The security cost of having 1 man on he ground, just security cost is $500k per annum and this requires a whole team of experts, mathematicians, administrators etc. The saving to the UN Secretariat plus UNOPS who administer the programme would be an absolute minimum of $200m per year if a method could be devised out of country.

    Any submission from anyone though has to be reviewed and shown that either it is in line with international best practice or has a replacement which is as good as/ better than current best practice.

    But inventing international best practice would be pretty cool

    Reply
    • Walker

      This analysis is not meant to stand up in court. It is meant as a best effort to answer questions for the public to see what is going on.

      Reply
  4. Keu

    Very interesting report. I have also checked some locations on the google maps. I noted this launch site:
    https://www.google.fi/maps/@47.8298847,38.7891929,363m/data=!3m1!1e3
    It seems that they could be back tracked to the Russian side through here:
    https://www.google.fi/maps/dir/47.7693381,38.7945336//@47.769278,38.7945302,454m/data=!3m1!1e3

    and then there seems to be depot area near by:
    https://www.google.fi/maps/@47.8120047,38.8530628,444m/data=!3m1!1e3
    https://www.google.fi/maps/@47.8157132,38.8667171,445m/data=!3m1!1e3
    The area also seems to have its own launch site with the burned ground:
    https://www.google.fi/maps/@47.814417,38.8738891,445m/data=!3m1!1e3

    Interestingly there are some abandoned fortifications around the depot:
    https://www.google.fi/maps/@47.7724823,38.8030136,443m/data=!3m1!1e3
    https://www.google.fi/maps/@47.7748254,38.8133413,444m/data=!3m1!1e3
    https://www.google.fi/maps/@47.8327782,38.8538321,375m/data=!3m1!1e3
    https://www.google.fi/maps/@47.8361549,38.830148,376m/data=!3m1!1e3
    https://www.google.fi/maps/@47.8333075,38.8617296,416m/data=!3m1!1e3

    There are signs of heavy fighting to the W and NW towards Torez and Snizhne.

    Reply
  5. Max

    There where many readers before, thanking you for your work. Let me know if you need help, for example by assisting translations to german if needed.

    Well done… and keep on doing this

    Reply
  6. Frank

    I just find all this analysis laughable. It’s all probably well done and meticulous. There is great creativity going on here too, and I think those you do it deserve commendation.
    It’s just that in toto it serves as nothing but obfuscating noise around the simple fact that there is no longer an effective border there, and probably never was. It all rests on this inapplicable notion of the State , of the Nation, with some kind of post-industrial mechanized border, where a line of utterly trustworthy Ukrainians and utterly trustworthy Russians , patrol along steadfastly up and down some immutable stretch of perfectly demarcated no-man’s land, as though it existed since the dawn of man.
    Really, garbage…..

    Reply
    • Factism

      I would totally agree with your relaxed position about borders and stuff – until people start killing each other.

      Reply
  7. Frank

    Just to add….the scenario could easily have been that some cousin with a relative who just got killed in Ukraine could have called someone else up with access to this equipment….that’s how it works amongst these people….corruption, protection, oligarchs……it’s a mess. The borders don’t and didn’t matter to anyone there.

    Reply
    • Max

      Just to reply… the scenario you added (the cousin an his relative) is no garbage, it is pure crap.

      What the guys have done here, is just investigate one of the mentioned stories about the withdrawel of ukrainian forces from the russian boarder in last summer. They say, there is some evidence, that the mentioned story is plausible. These guys where not creative in narating a story, the just took what proof can be found to a story of somebody else.

      What you are doing, well let me say in a polite way, you just deliver proof of a giant nothing between your ears instead of giving some facts that could turn the story around. Your not mentioning any mistake they made, you just invented a new story about a cousin with relatives… Worthless to mention, that the story you try to tell has no chance of getting close to the truth.

      Hey, you got the chance. Deliver some facts about that cousin and his relatives, and maybe you’ll think about a plausble story of how the cousin and his relatives could have done their business along the whole boarder. I am eager to get those info’s.

      Reply
  8. TTSlim

    I find this study laughable. Not only do the statistical methodology question go unanswered.. but the practicalities are that this is at best a directional analysis of the likely source ‘direction’ of artillery fire. So if we are to accept that methodological approach we may be able to guess with some certainty that the artillery came from over ‘that way’. And then the wishful thinking analysis comes into play… if you are hoping to find something… there will always be a statistically ‘valid’ method for defending your findings… all in the interest of sharing the truth ?? Its obvious you can’t handle the truth.

    This is a war between people who never trusted each other…and more importantly people who want to support the people they believe to be their brethren…and a view unpalatable to the western strategic interests in this area… that of self-determination. In all probability… and I use probability because we cannot rule out anything without real factual evidence… shells have been fired from Russian soil…but does the really mean the Russians did it…or is it more likely that pro-Russian separatists retreat into Russia when it suits them tactically..knowing Ukraine would never follow into Russian territory knowing it would be a pretext for overt Russian involvement.

    And that Russian authorities turn a blind eye to this … just as much as they are turning a blind eye to the NATO/US training, intelligence, satellite imagery and instructions being provided to the Ukrainian Army and Fascist Volunteer Defenders.

    So I ask why this report with such a poor regard for accuracy…
    Sharing the truth.. definitely not
    Influencing people to believe your point of view…highly likely …
    Another link in the western propaganda machine… hmm possibly
    Trying to prove this organisation’s relevance in this market … definitely

    I leace you all with just one thought… given NATO eagerness to strategically contain Russia as both a military and economic power…and its significant arsenal of satellite intelligence..and the might of the US war machine… why has all of the rhetoric to date been heresay and opinion.. not a bit of hard evidence… which they purport to have but won’t release… is it perhaps because it will raise a lot more questions than it will answer.

    The only sad truth to this matter is that Russia, the US and Ukraine are all the bad guys in this scenario…and the good guys are the Ukrainian people who are suffering… some paying with hunger and the inadequacies of a falling economy in Western Ukraine…and others with their lives and displacement in Eastern Ukraine… and lets not forget … that the pro-Russian separatists are still Ukrainians..how a government can consider its own people as expendable in the name of territorial integrity is beyond me… surely there are existing models of democracy that can cater for a yearning for self-determination.

    Your report has done nothing for the Ukrainian people.

    Reply

Leave a Reply

  • (will not be published)