Jeffrey LewisIsraeli Rocket Motor Test

You might have seen reports this week that there was a large explosion at an Israeli missile facility.  There wasn’t an explosion.  It was a routine test of a solid-rocket motor. Here’s the deal.

On April 20, an Israeli twitter user posted a video of a large smoky cloud to his Twitter account with the comment “A mysterious explosion in northern Israel.”  

https://twitter.com/EdyCohen/status/1384473172645355525?s=20

This video prompted a series of news stories claiming that there had been an explosion at an Isreali defense facility.  Stories appeared in Haaretz (“Powerful Explosion Rocks Sensitive Israeli Missile Factory During Test”) and the Times of Israel (“Big blast at rocket factory jolts central Israel in ‘controlled test’”) among other publications. 

Tomer, a state-owned Israeli defense enterprise, issued a statement that acknowledged the big cloud but didn’t say much else: “This was a controlled test with no exceptional circumstances.” Tomer also seems to have called it “routine.” Despite Tomer’s denial that an explosion had occurred, many news outlets around the world continued to report that an explosion had taken place.  There was plenty of online speculation that Israel had been the target of some sort of sabotage and more than a little glee expressed in certain quarters.

The video does not, however, show an explosion at all — it merely shows a large cloud of smoke produced by what appears to be an intense fire. One might infer an explosion had occurred but there are other aspects to consider.   There is a low rumble during the burn. And it seems to extinguish after a few seconds. At the James Martin Center for Nonproliferation Studies, we spend a lot of time looking at missiles.  This event looks and sounds much more like the last few seconds of a test of a solid-propellent rocket motor than an explosion.  You can watch a video of a US test here to get the idea.

Left: Still image from the April 20, 2021 video; Right: Test of a solid rocket motor in Utah.

This explanation would be totally consistent with Tomer’s admittedly vague assurance that the event was routine.  Tomer is responsible for the production of solid-propellant ballistic missiles in Israel.

As it turns out, it is easy to use open source techniques to confirm that this was a solid motor burn.

It helps to start with a base of knowledge.  We already know where Israel tests its large solid-propellant rocket motors.  Ploughshares recently funded a study at CNS to map Israel’s missile production programs and infrastructure as part of our work understanding missile proliferation in the region. As a part of this project, we tasked some satellites to take newly legal high-resolution images of Israel and in the process caused a minor international incident when a journalist at Haaretz reported that “someone” had tasked a Planet Skysat to image a “highly sensitive site” in Israel.

Well, that “someone” was me (waves) and that “highly sensitive site” was the Sdot Micha Airbase, which among other stuff, is where Israel tests large solid propellant rocket motors. This base of knowledge is what allowed my colleagues and I to work out what happened in just a few hours.

Now, in addition to the high-resolution image (50 cm) that caused such a fuss, Planet offers near-daily imaging of the whole world at a lower resolution (3 m).  So I had access to both a recent high resolution image of the site, as well as almost daily moderate-resolution images for the past month.  We examined these satellite images of the test stand at Sdot Micha and, sure enough, moderate resolution images taken by Planet show a large burn scar indicating a test of a large solid propellant rocket motor that appeared on April 20 — the same day the video was posted. Here is a .gif showing: (1) the high-resolution image from February 22 that caused the international incident, (2) a false color  near-infrared image (NIR) from before the test on April 14, and (3) a second false color NIR image from after showing the burn scar on April 21. 

Image credit: Planet Labs, Inc. cc by-sa-nc 4.0

Oof, that left a mark. The false color image is from April 21, but you can also see the burn scar in an April 20 image. I created a Planet story using several true-color images in case you are a completist.

Now, some of the initial reporting indicated that the explosion had taken place at a Tomer facility near Ramla. There are two missile related facilities near Ramla.  So, just to be sure, we examined both missile-related sites for any sign of an explosion.  They’re just fine.  

The last step is to geolocate the video — to make sure that the video shows the cloud coming from the motor test stand where the burn scar is visible instead of someplace else that went kablooey.

Aram Shabanian and Blake Vincent, two of my colleagues at CNS, volunteered to geolocate the video, which they did in relatively short order.  The low quality of satellite images of Israel in Google Earth complicates geolocating videos, but Aram and Blake pretty quickly figured out the general area. It helped that the satellite imagery suggested where to look.  Blind geolocation can be incredibly difficult, and narrowing down the search area with prior knowledge beforehand can save countless hours.  It also helped that they are clever. 

After some discussion, an online colleague named Samir finally put a pin in the right spot.  I don’t know anything about Samir other than that his work is excellent.  The person taking the video is standing around 31.822, 34.902°, in a kibbutz called Mishmar David, and looking straight at the rocket motor test stand almost due south.

Left: A screenshot of the video of the event; Right: The likely location of the photographer and his/her line of sight toward Sdot Micha.

One method to build confidence in a geolocation is to find a distinct building in the video or image and then see if that same building is visible in the right place in a satellite image. You have to do this separately from the parts of the video or image used to make the geolocation so that it is really a test of the hypothesis with a chance of falsifying the geolocation.  In the video, for example, there is a pretty distinct building in the town that lies almost directly between the person making the video and the cloud (below left).  Although the available images in Google Earth are a few years old and of relatively low resolution, there does seem to be a building in Tal Shahar that might look like that in the correct spot (below center). So a test is to go get a better satellite image of the building and see if it  checks out.  In this case, it did.  A higher resolution image of the same building (below right) does look like a good match. I repeated this process with a couple of other buildings.  Aram, Blake and Samir’s geolocation held up with better imagery.

Left: A distinct building in the video; center: A possible candidate for that building in the correct location in Google Earth; Right: A high-resolution image of the building that appears to match.

So, in the end, it is pretty clear to me that someone in Mishmar David recorded the last few seconds of a solid rocket motor test at Sdot Micha.  The geolocation lines up and satellites captured the burn scar appearing in the right spot on the same day.  So, no, this doesn’t seem to be an act of sabotage. 

Why didn’t Tomer just say it was a solid rocket motor test? The fact that Israel produces large solid rocket motors for, among other systems, the Shavit space launch vehicle is no secret.  There isn’t really any reason to be coy, it’s just … a habit.  A culture that evolved in a past era and has yet to adapt to the current one. You could probably write a book about it.

Harel Dan probably has the best summation of why Tomer would be coy about a test it describes as routine. You might know him as the person who figured out that the radar for a Patriot missile defense battery interferes with Sentinel-1 satellite taking synthetic aperture radar images — and was able to use this interference to geolocate Patriot batteries in Israel. Here’s what he said:

That translates, roughly, as:

“Imagine for a moment that instead of ignoring and encouraging speculation, the Defense Ministry just said, ‘We conducted a solid rocket motor test.’ But no, this policy is like opium for those in power.

Greetings from Democratic Space.”

Huge thanks to Aram Shabanian (@AramShabanian), Blake Vincent (@ColdWarCrush) and Samir (@obretix).  You should follow all three on Twitter.