Mea Maxima Culpa

My previous analysis was completely wrong

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, Gio52452 on Twitter, MMetarch on Twitter. Also thanks to Kinez, who has been invaluable in helping me write many articles.

After posting the Pleiades images yesterday, multiple people reached out to me to send me evidence that some of the damage in those images predated the Oreshnik strike. I immediately placed an additional order for an earlier pass (October 22nd) that SkyWatch had available.

In the interest of dispelling the invalid analysis I put out yesterday, I will summarize the results of the second, earlier satellite pass I ordered before covering it in-depth in the rest of this article.

All of the damage seen in the November 22nd pass happened prior to the Oreshnik strike on the 21st, and was the result of attacks using conventional Russian standoff weapons. The November 22nd pass shows no new damage from Oreshnik.

The reasons for my faulty analysis, for which I take total responsibility, are as follows:

• I uncritically accepted the anaylsis of others, who said that the Sentinel-2 pass from November 24th showed new damage. It did not – this damage was the result of a strike in July.

• I failed to adequately search for resources in Russian (something I am capable of), which had thoroughly tracked previous strikes on the plant.

• I didn’t want to spend an additional $125 to order an earlier pass from Pleiades. Had I done this right off the bat, this could have been avoided.

• I instead relied on free and publicly available imagery from Google Maps, provided by Maxar. Someone at Maxar has engaged in a painstaking, but sloppy, coverup of the damage caused by previous strikes, patching a 2024 pass with multiple other ones, despite maintaining the 2024 date in the citation of the images. For this reason, I was much more inclined to accept the damage as new, and as having been caused by Oreshnik.

• I failed to consider that the Ukrainians might have not repaired any damage from the previous strikes.

I apologize to everyone who based their own analysis off of mine.

In the remainder of this article, I will elaborate on why I got it wrong, and dig in to the pass from October, which is a superior image to the one from November. From there, I will:

• Cover damage to Yuzhmash which was not visible in the November pass, due to the inferior angle and clarity.

• Conclusively prove that Maxar is editing their imagery by hand to conceal damage at Yuzhmash.

Finally, I’ll speculate on what I think might be going on here. There are three possible explanations:

1. Oreshnik completely missed Yuzhmash

2. Oreshnik hit the plant, but did no damage which is visible in the satellite imagery I ordered

3. Just like Maxar, Airbus is participating in a coverup of damage to Ukrainian military objects

Note: Images in this article will have an Airbus watermark, which I manually added, and as is required by the provider (SkyWatch) I purchased these images from. I’m still not sure if I’m violating their terms of service by sharing these images like this, but I plan on going back and watermarking the images in the previous article as well to ensure that I’m complaint.

October Pass

First, the full pass from October:

All damage seen in November can be seen here. The angle of this pass is much higher than the one from November, and the image was taken a full hour later, a month earlier in the year, reducing shadows. Here’s a legend to get your bearings as I cover each cluster of damage:

A, B, C, and D are all clearly visible in the November pass. E was not easy to make out because of distortion, but it’s much more clear here, and major damage to an additional (very large) building is visible for the first time because of the new angle. F and G were not possible to make out in the November pass.

First, close-ups of A, B, C, and D:

Because there’s less distortion on the right side of this pass, it’s possible to more clearly see how the smokestack in D appears to have fallen onto the roof of the building its base is attached to when it was destroyed (the smokestack is clearly visible in the Apple Maps images I used in the previous article).

Next, damage that was previously hard to make out, or not visible. In E we can clearly see that there was a total collapse of the roofs of some of the structures:

To the west of E, we can see severe damage to a building that was previously not visible because of the angle, along with a rubbish heap of debris:

This is a big building. Here’s what it looks like in Google Streetview in July of 2015:

Next, F:

And G:

At first I thought the buildings in G had simply been abandoned, and fallen into disrepair, but here’s what they look like on Maxar through Google Maps (copyright indicates 2024):

Establishing a Timeline

So if Oreshnik didn’t cause all this damage, what did? Sources I failed to locate before publishing the last article have documented multiple strikes on Yuzhmash. In some cases, we can use Sentinel-2 to identify which strikes damaged what. In other cases, that damage is simply invisible on Sentinel, but we can rely on local videos. I’m confident we can account for the damage in Clusters A, B, and D.

July 15th, 2022

There are at least two videos of this strike, but they’re from quite far out and don’t allow for the identification of which buildings in the plant were hit. The early period of the war was extremely chaotic, and it’s very difficult to find any reliable sources beyod reports saying targets “in Dnipro” were hit.

Using Sentinel-2, we can see that this is almost certainly the strike which did the damage seen in Cluster A.

Left: July 15th, 2022 (morning). Right: July 18th, 2022

LostArmour confirmed that this is the strike which did the damage seen in Clusters A and C with high resolution imagery of an unknown source:

November 17th, 2022

Given the proximity of this strike to a road, and using Google Streeview as a reference, this is the strike which did some of the damage seen in Cluster D. Video:

Streetview and location in Google Maps (the aerial is oriented east). Note the diagonal parking on the left side in both the video and in street view, and keep in mind that the streetview is a decade old at this point.

A later found a video of the aftermath which confirms the target was the large building in the north of Cluster D, or the other buildings in Cluster D.

LostArmour’s analysis (authors: Simo, magnetro1, Hazarescu, and Artem) from the next summer found that this damage occurred in the next strike, however:

August 15th, 2023

A cruise missile (reported to be 4x Kh-101s) strike hit the plant. It isn’t easy to tell what damage may have been caused by this strike. Good candidates are C, E, F, and G, because damage from A and possibly D may be visible.

Left: August 7th, 2023. Right: August 25th, 2023 (Sentinel-2)

A video of this strike:

Ukrainian and Western media completely failed to cover this strike, from what I can tell. Russian sources, however, did.[1][2]

July 8th, 2024

A combined cruise missile and drone strike on the territory of the plant. This is when the damage seen in section B occurred. Sentinel-2 may be a tiny fraction of the Pleiades, but it’s useful enough once you already know where the damage is.

Left: 7/7/24. Right: 7/9/24

These two impact craters, the same ones I used as proof of Oreshnik’s accuracy, were widely shared in the aftermath of the November strike as evidence of Oreshnik’s damage. To disprove this, all I would have had to do was go back to Sentinel Hub, which I had already spent hours in without noticing that this damage was present back in July. Instead, I accepted this analysis as true. No excuse there.

A video of the strike (Twitter source):

---

I can’t rule out that I’ve missed a strike here, but these are the events I can identify with confidence, using multiple sources for confirmation. Please let me know if you’re aware of one I missed.

A logical question is whether we can assume that the remaining clusters that are unaccounted for were caused by Oreshnik. The October pass rules this out. Multiple structures may have been hit in these strikes, and because Sentinel-2 is so low resolution, there’s no way to confirm the damage that is harder to see in all satellite imagery.

Coverup

LostArmour had already commented (in 2023) on the strange “patches” which are visible in the Maxar imagery that Google Maps uses. I had assumed that these were simply normal boundaries between satellite images. After cross-referencing the October and November passes with this imagery, I can now say conclusively that a manual cover-up has been performed. Every “patch” corresponds to damage we now have high resolution imagery of.

These patches are old archival images which have been superimposed over imagery from 2024 to conceal the damage caused since the beginning of the war. They’re easily identifiable by how a section of the imagery will bleed or blend into another with different coloration (different time of day or year) or a different angle.

Cluster C

Left: October pass. Right: Maxar (Google Maps)

Note that the perfectly parallel roof sections are no longer parallel in the Maxar image, along with being a different color, and the presence of “bleed.” This is a huge patch, running the length of the plant in Google Maps and well beyond its perimiter. But as we’ll see, it’s just one layer of many. It seems absurd to think that the boundary of this patch would precisely start a few meters from the edge of the damage in Cluster B by pure coincidence. Imagery on the left side of this patch must date from before the strike in July of 2022, because the damage from Cluster A isn’t visible, and doesn’t appear to have been individually patched.

Cluster D

This patch, or patches, are much more sloppy. Again, we see that the patch begins precisely where the northernmost damage from Cluster D occurs, hiding the large hole in the roof and the destruction of the smokestack attached to Main Assembly Shop 97. This appears to actually be one patch laid atop the boundaries of two larger images, as both a left and a right boundary are visible on the roof of the Main Assembly Shop. This may have been needed to cover up the damage to the smokestack and the hole in the roof, which might have been caused on different days.

Cluster E

The double patching over Cluster D might have something to do with ensuring the damage in Cluster E was also hidden. We can see bleeding to the right of the Maxar image, which extends all the way up to Cluster D. There’s also some weirdness on the northern edge of the line of buildings in the center. The edge here seems perfectly placed to make sure that no damage to the building on the right can be seen.

Cluster F/G

This patch is the same as the one which runs through Cluster C.

Zooming out, we can see what’s been done:

Each color represents a different layer (edit: it’s possible that two of the colors which are not next to each other represent a single layer with a different layer in between them), taken at a different time. Combined, and overlaid on top of each other, they’ve allowed Maxar to hide any of the damage from the strikes since the beginning of the war. Each boundary between a layer runs precisely through the edge of damage from one of the strikes. As of the time of writing, this is plainly visible on Google Maps, and I invite you to check it out for yourself.

So What About Oreshnik?

As I said in the beginning of this article, there are three possible explanations:

1. Oreshnik completely missed Yuzhmash

2. Oreshnik hit the plant, but did no damage which is visible in the satellite imagery I ordered

3. Just like Maxar, Airbus is participating in a coverup of damage to Ukrainian military objects

Explanations 1 and 2 are possible, but somewhat hard to imagine. It would be very hard to cover up the damage if 36 projectiles traveling at Mach 11 (according to Ukrainian sources!) slammed into random buildings in Dnipro. Similarly, it doesn’t seem to make sense that those projectiles could have hit the plant and done so little damage as for it to be invisible. They were simply moving too fast, and if they were somehow “blanks” with little to no mass, they would have broken up in the atmosphere, as other commentators have pointed out.

I consider explanation 3 the most likely. Some have said that I may simply have been given older images, and offered up the greenery visible in the November pass as proof. I don’t buy this, because Sentinel-2 images from prior years look identical in late November. The November pass isn’t even particularly “green” when you zoom in closely, much of the foliage is clearly brown.

One mysterious detail is that there are not one but three passes from November 21st available from Pleiades. Someone tasked the satellite to make these passes as soon as possible after the strike. There have been none since then, and none in the weeks prior. Is it possible that Pleiades would have done a silent pass before the Oreshnik strike in order to get images of it that could be offered up at a later time as being taken after the strike? This would have required foreknowledge of the strike, which various government officials have confirmed they only had 30 minutes of, and those 30 minutes were in the middle of the night. The imagery can’t be from the previous year, because it includes damage which occurred in the Summer of 2024.

It’s hard to know what to make of all of this. Outside of carefully constructing a photoshopped version of the strike, it’s difficult to imagine how it could have been faked. If that was what happened, ordering the other two passes from that day would be a way to validate it, as it would be nearly impossible to edit three separate passes at different angles and have it all come out consistently.

However, given what we see with Maxar, it is indisputable that western satellite imagery providers are deliberately censoring their products in order to conceal the damage caused by Russian strikes. Someone ordered those patches made. If Maxar is doing it, Airbus is doing it too. The only question is whether they did it in this particular case.

What Now?

I’m done buying satellite imagery. The first post in this series triggered a firestorm, and I feel genuinely awful about being so off the mark. That being said, the opportunity to continue this process is available for anyone equally insane enough to spend their hard-earned money gambling on this.

The four passes over Yuzhmash. I’ve purchased the bottom two.

An obvious next step would be to seek out sources from neutral countries, like China. There are Chinese satellite imagery brokers, but someone who speaks Chinese and can pay in Chinese currency will need to do this. I’ve also mentioned buying the other passes from the 22nd to cross-reference them with each other. This will cost $250 on SkyWatch. Another option is for someone to blow a truly huge amount of money by drawing a much larger boundary (even extending the peremiter of the order a few hundred meters beyond the boundary I drew brings the price to over $1000). This would make it possible to say whether Oreshnik missed the plant, or hit a section of it which wasn’t visible in my passes. Strikes have occurred in the southern section of the plant in the past.

The kitchen sink option would be to wait until there’s a clear day (or days) and task the satellite yourself. It would be hard to imagine Airbus somehow editing the image on the fly before they return it to you. If they reject the request, that’s interesting on its own.

But I’ve wasted enough money on this, and I’m calling it a day. Maybe a motivated person can start up a crowdfunding drive for this, but after performing this analysis, I simply don’t trust satellite providers enough to roll the dice.