Yuzhmash and Oreshnik Demystified
I blew several hundred dollars on commercial satellite imagery
Update 12/20/24: The analysis in this post is incorrect, none of the visible damage was caused by Oreshnik. I’m leaving it up in an unmodified state for posterity and accountability, but please see this correction article I wrote explaining why I got it wrong:
Mea Maxima Culpa
Before I start, I’d like to thank the following people for giving me feedback which allowed this article to be made so quickly. These people were right, and I was wrong. Simo, magnetro1, Hazarescu, Artem on LostArmour. drakhl on Twitter, 3_bm15 on Twitter,
Note: This post is gaining a lot of traction, so please draw your own conclusions and reach out in the comments or on Twitter with any corrections or mistakes. Metarch on Twitter provided an old report (2023) from LostArmour showing that a couple of these impacts may be from old strikes. I’ve placed an order for a Pleiades pass from October, which will hopefully settle things. Check out LostArmour’s post after reading this one. I’ll make an update post when the October images come in.
It’s been nearly a month since the first use of the Russian “Oreshnik” system took the world completely off guard. But despite multiple angles of the strike on Dnipro, to most, it likely feels like we’re no closer to a clear picture of its capabilities than we were the night it happened. Contradictory analyses have proliferated through social and mainstream media, and it’s unclear what to believe. Various analysts have described the system as a series of “blanks” from a jerry-rigged nuclear ICBM – intended only as a warning shot – or as a new type of cluster munition where each of the 36 kinetic submunitions is roughly equivalent in power to a 250kg bomb (read my thread on this here). Some eyewitness reports of unknown credibility said the plant (or parts of it?) had been “reduced to dust.” Other Ukrainian reports said minimal or even no damage had been done.
When the publicly available (and free) imagery from a pass of the ESA’s Sentinel-2 satellite was published a few days after the strike, it only muddied the waters. The Sentinel is free for a reason, it has a poor 10m resolution (each pixel roughly represents 10 meters). Here’s what it showed:
Commentators noted what appeared to be two large holes in the roof of a structure roughly in the center of the image:
But beyond that, it wasn’t possible to say if any real damage had been done. This was frustrating, and I embarked on a lengthy journey to attempt to purchase commercial satellite data myself to settle the issue once and for all.
The Hunt
Most commercial satellite providers focus on sales to corporations, governments, and institutions. Those companies aren’t interested in one-off sales to individuals (believe me, I tried). One option was accessing Planet Lab’s imagery through Sentinel Hub, and I subscribed ($$$) and purchased 25 sq km of quota for Planet’s SkySat after querying the area around Dnipro and seeing that SkySat (50cm resolution) had made multiple passes over Yuzhmash since the strike. I made a rookie mistake and ordered all the images taken between then and the first week of December, only to find that every one of them had nearly 100% cloud cover. Dnipro gets plenty of snow this time of year, and the forecast showed nothing but cloudy days ahead. Even the passes from the same day the Sentinel-2 did a flyby were from later in the day, when clouds had rolled in. That money had been wasted.
After giving up for a few weeks, I found another provider which offers images from Airbus’s Pleiades-Neo constellation. Pleiades offers even better (30cm) resolution than SkySat and amazingly, it had already taken three separate shots of Yuzhmash on November 22nd, the morning after the strike. I don’t think this is a coincidence. Someone likely requisitioned the constellation (something you can do with many satellites if you have the money) but, for whatever reason, chose to sit on the images rather than publish them. Luckily, the provider I found allows for selecting a very tight polygon for your order, with no minimum. I would be able to select just the territory of the plant, without the order running into the thousands of dollars. I suspected that some of the Oreshnik’s projectiles likely landed outside the territory of the plant itself, but I had no choice but to select a polygon matching the plant’s perimeter, due to the cost. I placed the order and waited around eight hours for it to complete. When done, I was given a 90MB PNG (13,000 x 9,000 pixels in resolution), a 1.1GB GeoTIFF (COG) file, and a JSON file of metadata for download.
Results
I spent the next several hours pouring over this massive image pixel by pixel. At this point, I feel reasonably confident that I’ve identified all damage caused by the Oreshnik on the territory of the plant that lies within the polygon I drew. I also believe we can make much more informed conclusions about the nature of this weapon and and its strategic potential.
But first, a few notes about this image. The satellite’s angle in relation to the plant is 43.7 degrees “off-nadir,” which is good because it allows us to see the height of buildings, but it adds some distortion and prevents us from seeing the north side of structures. The snapshot was taken at 10:18AM local time in Dnipro, in winter, which results in a significant amount of shadows because of the low angle of the sun (something like 20 degrees). I’ll be comparing the images from Pleiades with older, publicly available images from Google and Apple Maps. Ideally, I’d have an older shot from Pleiades from the same angle and at the same time of day to compare with, but that was cost-prohibitive. The publicly available images are at much closer to a 90 degree angle, with the sun high in the sky, so objects will appear flattened and will be more brightly lit.
Also important to note is that the buildings at Yuzhmash are big. Really big. Many are taller than three stories with over 500,000 square feet per floor. Others are over ten stories tall. Keep that in mind when viewing the images. This kind of analysis has a learning curve.
First, a zoomed out view of all the visible damage to the plant. Red circles are high confidence, and yellow are low. Each circle doesn’t necessarily represent one impact, but rather indicates damage of some kind. More on that in a moment.
Zooming out again, we can see what appear to be the paths traced by at least three of the six sets of submunitions. Remember, the Oreshnik releases six Maneuverable Reentry Vehicles (MARVs) of a completely new type, each of which carry six kinetic submunitions. By zooming out again, we can clearly see how the kinetic projectiles from separate MARVs fell:
I haven’t heard anyone suggest an “official” name for the MARVs, so we’ll call them “Cassettes” and refer to the clusters of impacts as Cassettes A, B, and C, from left to right.
Cassette A
First, the southernmost (lowest) building, with our image on the left and Google Maps on the right:
This building is either three or four stories tall and appears to have suffered some damage to its roof on the northernmost section. There may be an impact hole in the center of the image, but it isn’t clear enough to tell.
Moving northwards (up), we see that this next building has suffered total collapse in its central section. This looks like the kind of damage an Iskander or a 250-500kg bomb might do. As we move along, we’ll see that depending on the height of the building and where the kinetic projectile landed, the damage may appear as a small hole in a roof, or might cause a total collapse even in a 5+ story building like this one. Every floor has pancaked all the way to the ground.
Continuing northwards, this building has an impact in the roof but didn’t suffer a catastrophic collapse.
Moving laterally to the East, this roof has some light damage.
And to the south the superstructure of the corner of this building is exposed.
That’s it for Cassette A.
From the damage we see in this first cassette, it’s reasonable to draw the conclusion that the projectiles are moving fast enough to punch a hole straight through a roof and penetrate into the ground, leaving the upper floors of a building relatively intact. If the projectile happens to hit a key element of the building’s foundation, it may cause it to collapse, but this isn’t guaranteed. It is logical to assume that every projectile contains the same energy, and is capable of doing the same damage.
Cassette B
The damage to the southernmost building in Cassette B is extensive and complex. This is the damage that was visible even in the 10m resolution Sentinel-2 imagery. First, a high resolution image with its red circles still included:
Next, side by side with Google Maps:
The damage is severe. An entire section of the structure has collapsed west to east across its length, revealing its inner supports, and two large craters are visible in its center. The visibility of this damage is likely caused by two projectiles striking this three story building in close proximity, leading to a larger structural failure.
The building immediately to the north has a large chunk taken out of its southern edge, all the way to the ground, and some damage to its roof around the center of the image. The boundaries of two tiles in Google Maps overlap here, which is why the right image looks like two different images mashed together.
Lastly, there’s an impact hole and some structural failure in the northernmost building.
Cassette C
There’s much more distortion on this side of the image, but multiple hits are clearly visible.
The roofs of the buildings on the right seem to be damaged and covered in debris, and the center building on the right side seems to be missing its roof entirely.
Moving North, and starting from the left circle and going clockwise, we see a total collapse of the northern edge of the left most building all the way to the ground. Next, a smokestack and its supporting industrial structure has been destroyed, with damage to the exterior of the attached building. Finally, there’s a black section of the roof the southernmost building. This may simply be new asphalt applied to the roof, but I’ve circled it because it’s not visible in any other imagery I’ve seen.
Next, we see the same building with the smokestack and exterior damage has suffered severe collapse in its roof from an impact. This building is very tall – ten stories in places. Here’s what it looks like from Google Streetview:
Takeaways
I believe these images confirm Ted Postol’s assessment that the Oreshnik’s kinetic projectiles deliver the approximate energy of a 250kg bomb, though the nature of this energy is different. The projectiles are moving so quickly that they punch a hole through most anything until they hit the ground. When we see severe enough roof failure to get a good look at the damage, we see an impact crater 100ft in diameter and a lot of dirt. This implies that the projectile is penetrating into or through the foundation of the building. These structures were built by the Soviets and are extremely tough, so if the projectiles are capable of punching through them and reaching earth (or pulverizing the foundations enough that they look like earth), they’re delivering a lot of energy.
Additionally, both the MERVs and the kinetic submunitions are extremely accurate. The gap between their impacts is as small as 25 feet in places, and the >=3 lines of impacts are evenly spaced. This makes the absence of at least one or as many as three of the MERVs puzzling. Either we’re seeing overlapping impacts (“double taps”) from multiple MERVs, or some of them missed the plant entirely. Cassette B is potentially a double tap. It’s also possible there were other targets in Dnipro, outside the plant, or perhaps some of the MERVs malfunctioned and missed the mark. This lines up with the videos of the strike which show one set of projectiles landing well away from the others. It is a prototype weapon after all.
So what is the Oreshnik? Here are three different ways to think about it:
It’s a way to deliver a strategic bomber’s worth of destruction without the strategic bomber.
It gives the Russians the capability to deal the damage of 36x 250kg bombs to any target within several thousand miles in less than 20 minutes, and with no chance of interception.
It can deliver the damage of a few dozen Iskanders with a single missile, and do it from 10 times as far away.
A lingering open question is how much damage, if any, occurred underground. Yuzhmash is infamous for its underground workshops, which the Soviets allegedly designed to withstand a nuclear blast. The Oreshnik’s projectiles certainly seem to have penetrative power, but how deep they go is beyond my pay grade. I’ll leave that problem to the amateur physicists, who hopefully catch wind of this post and use it to inform their calculations. If these projectiles are capable of penetrating as deeply as some analysts have speculated, that changes the picture quite a bit.
The Russians have solved the problem of how to cost-effectively attack dispersed targets with a ballistic missile. That doesn’t make it an all-powerful doomsday device, and it would likely take many Oreshniks to render a large facility like an air base (or Yuzhmash) completely inoperable. But the Russians do have the capability to mass produce them, and have openly stated their intent to do so, which makes the system a potential strategic game changer.
An image of damage to a street outside of Yuzhmash has been removed because it is from an earlier strike. Thank you to @drakhl on Twitter for the correction.
Mike Mihajlovic has some interesting insights at his site. I think what I've seen so far dismisses too readily the effects of kinetic impacts as trivial (see Postol's work although I understand how he got to his conclusion). It's reasonable to think that based on 1st order physics, and that the impactor converts on contact into a ball of plasma, but it's misleading.
It doesn't take into account the detailed penetration dynamics of solids at Mach 10 and the resulting shock generation. Also unaccounted for is the combined impact of multiple shock generators on underground structures.
The key technology that the Russians apparently have is something that is solid at 4000 C. This is not unreasonable. There are known ultra-high temperature ceramics in that regime (e.g. hafnium carbide).
Bloody well done. That's excellent. That's investigative reporting and its science, too.
So where's all the 'turned to dust' ? That was maybe all rubbish. Or maybe when we find where the other impacts were we might know more.
This is great. We need more of it. Thank you. :)