Editor’s Note: Readerly response to a recent news story, “Israeli Rocket Defense System Failing at a Crucial Task, Expert Analysts Say,” where Ted Postol was quoted to say that Iron Dome was not effectively detonating warheads, was so negative, and angered so many people, particularly Israelis, that we asked Professor Postol to explain how he came to his conclusions and to show his data. He gracefully agreed. The following article represents his opinion, and is not necessarily the opinion of MIT Technology Review—and does not represent any collective assessment by MIT or one of its departments, labs, or centers. (That’s because we are editorially independent of the Institute.)
In the early weeks of July 2014 the conflict between Israel and Palestinians in Gaza has again flared up. This has resulted in a new round of large-scale rocket attacks launched by Hamas, operating from Gaza, against Israeli population centers. The last time such large-scale rocket attacks occurred between Hamas and Israel was in November 2012. During the November 2012 conflict a large number of photographs of Iron Dome interceptor contrails were observed in the sky. These contrails revealed that the Iron Dome interceptor rate was very low—perhaps as low as 5 percent or below.
This paper explains why the geometry of the contrails photographed in the sky indicates whether or not an Iron Dome intercept attempt had any chance of intercepting an artillery rocket target.
I will show sample pieces of data indicating that Iron Dome performance was very low in November 2012, and I will show similar data for July 2014, which indicates that Iron Dome performance almost a year and a half later has probably not improved.
At this time, collection of the data for July 2014 is still in progress. However, all the data I have so far collected indicates that the performance of Iron Dome has not improved.
One of the most demanding problems in intercepting an artillery rocket is that the interceptor must destroy the warhead on the rocket. If the interceptor hits the back end of the rocket, all that will happen is damage to the expended rocket motor tube, which is basically an empty pipe. Damaging the back end of the artillery rocket essentially has no effect on the outcome of the engagement. The pieces of the rocket will essentially fall in the same defendant area, and the warhead will almost certainly go on to the ground and explode. These facts mean that the only meaningful definition of a successful intercept is the destruction of the artillery rocket warhead. As will be shown in the discussion to follow, destroying the artillery rocket warhead is considerably more demanding than doing damage to other parts of the artillery rocket—or successfully damaging an aircraft, causing the failure of its mission.
Protecting a population spread over defended areas from the hazards of such missile attacks must involve protection against falling debris, which can cause serious injuries to individuals who are not in protective shelters.
As I will discuss later in this article, Israel does in fact have an extremely effective missile defense. That defense is the early warning system that tells people on the ground a rocket is traveling in their direction, and the shelters that are arranged so that individuals can easily get to protection within tens of seconds of warning. In an article referenced later in this paper, it is shown that during the bombing of London by V-1 and V-2 rockets, seconds of early warning vastly reduced casualties and deaths from individual attacks.
In the particular case of rocket attacks against Israel, the overwhelming number of artillery rocket warheads are in the 10- to 20-pound range, which makes the effectiveness of shelters even greater.
These two factors, the small size of the warheads, and the warning and sheltering system completely explain why there have been no casualties from the rocket attacks.
Assessing Whether or Not an Iron Dome Intercept Attempt Is Successful from Photographs of Iron Dome Contrails
I will first show why the Iron Dome interceptor must approach the target artillery rocket from a frontal direction. I will then show that the Iron Dome interceptor has, for all practical purposes, no chance of destroying the warhead on incoming artillery rockets if the interceptor engages the rocket from the side or from the back.
I will then present photographic evidence of contrails in the sky, indicating that Iron Dome interceptors were mostly chasing or engaging artillery rockets in side-on geometries.
I do not know why the Iron Domes were not engaging most artillery rockets using the proper front-on geometry. However, it is clear that the Iron Dome radar tracking and guidance system is not working, as it is initially sending Iron Dome interceptors to intercept points that then result in the interceptor not being able to achieve the right geometry for a successful engagement against the artillery rockets.
I will show photographs of contrails from November 2012 and from July 2014 indicating that the Iron Domes are still behaving erratically—resulting in continued very low intercept rates.
Assessing the Meaning of Iron Dome Contrails
To understand why the Iron Dome interceptor must approach the artillery rocket from a frontal direction, it is necessary to have a rudimentary understanding of the Iron Dome interceptor.
Figure 1 below shows a conceptual picture of a front-on engagement by an Iron Dome interceptor against a Grad artillery rocket. The blue dashed line emanating from the forward section of the Iron Dome interceptor depicts the line of sight of what is called a “laser fuse.” The purpose of the laser fuse is to create a beam of light that will reflect off the front end of the artillery rocket so that the interceptor can determine that the target artillery rocket is in the process of passing the interceptor. As can be seen from the diagram, the warhead in the Iron Dome interceptor is placed well behind the fuse assembly, a distance of roughly three feet from the laser-fuse aperture. This gives the fuse enough time to determine where the front of the target rocket is, estimate how long it will take for the front of the artillery rocket to pass parallel to the artillery rocket’s warhead, and detonate the Iron Dome warhead.
The timing delay is quite critical to many variables. It must account not only for the location of the target rocket’s warhead but also for the speed of the fragments from the Iron Dome warhead, the miss distance, the off-parallel orientation of the Iron Dome interceptor relative to the artillery rocket, and the high passing speed of the Iron Dome interceptor and the artillery rocket.
Figure 2 shows how the fragments move, under the assumption that the crossing speed of the Iron Dome interceptor and artillery rocket is about 1,200 meters per second and the fragments from the Iron Dome warhead are projected at about 2,100 meters per second perpendicular to the axis of the Iron Dome interceptor. Because the Iron Dome interceptor is moving at 1,200 meters per second relative to the artillery rocket, the additional crossing speed needs to be added to the 2,100-meter-per-second lateral velocity of the fragments. The net direction of the cloud of fragments, as would be seen if an observer were sitting on the artillery rocket, is shown by the pale blue arrow that passes through both the Iron Dome warhead and the artillery rocket’s warhead.
Figure 3 shows the outcome if everything works as intended. However, there is a range of possible outcomes where success is very likely, and beyond that range, the possibility of success diminishes drastically.
As can be seen from the arrow marked “1,500 meters per second” in figures 2 and 3, the higher crossing speed can result in a significant change in the net direction of the cloud of fragments. Thus, the fuse must determine the best time to detonate the warhead based on the crossing speed, the distance of the artillery rocket target as it passes by the Iron Dome interceptor, and the various fusing delays associated with detonating the Iron Dome interceptor’s warhead.
Because of the uncertainties in the exact crossing speed and crossing geometry, even a perfect fuse may fail to put lethal fragments onto the artillery rocket’s warhead.
In addition, unless the distance between the Iron Dome warhead and the warhead of the artillery rocket is small (roughly a meter or so), there will be a greatly diminished chance that a fragment from the Iron Dome warhead will hit, penetrate, and cause the detonation of the artillery rocket warhead.
Thus, a front-on engagement does not guarantee that the Iron Dome interceptor will destroy the warhead on the artillery rocket.
Figure 4 and figure 4A show the consequences of a failure in the fuse timing in what was almost certainly an engagement between an Iron Dome interceptor and the artillery rocket shown on the ground in the photos. As can be seen by inspecting the photograph in figure 4, there is significant damage in the area where the rocket fell. This damage was almost certainly due to the detonation of the rocket’s small warhead. Figure 4A shows the magnified front end of the rocket, where holes can be seen in the expended and empty rocket motor casing that was immediately behind the warhead. This photograph therefore shows an example of what might have been a successful Iron Dome intercept attempt.
In this case, it is nearly certain that the artillery rocket was engaged by an Iron Dome interceptor that was properly approaching the artillery rocket front-on. Unfortunately, the timing commands from the fuse resulted in fragments from the exploding Iron Dome warhead hitting the artillery rocket after the warhead had passed. The relatively low density of holes in the artillery rocket’s afterbody suggests that the encounter also had a relatively high miss distance—possibly several meters.
This photograph illustrates how even when the Iron Dome interceptor is in a proper front-on trajectory, it can still fail to destroy the warhead of a target artillery rocket.
Figures 5, 6, 7, and 8 show detailed vector diagrams that indicate how the Iron Dome interceptor would perform if it engaged an artillery rocket from a variety of directions. In these diagrams the speeds are shown in feet per second, rather than the meters per second used in figures 1, 2, and 3.
Figure 5 shows a nearly front-on engagement direction (again, note that all the vector speeds are now in feet per second). A careful review of the geometry of the engagement will reveal that even a moderately skewed off-frontal direction of approach will drastically reduce the chances that fragments from the Iron Dome warhead could be sprayed onto the warhead of the artillery rocket. This therefore shows that the front-on geometry is very sensitive to small off-frontal errors that could be the result of faults by the master control system in the guidance and control of the Iron Dome interceptor.
This particular diagram (figure 5) demonstrates how important it is for the the master guidance and control system to place the interceptor in the right location before it begins the actual homing process against a target artillery rocket
Figures 6, 7, and 8 show detailed vector diagrams for interceptor engagements that approach the target artillery rocket from the side or from the back. A careful inspection of the geometry of the fuse-sensing beam and the spray pattern of the fragments from the Iron Dome warhead show that there are two very serious problems with these kinds of engagements.
First of all, if the fuse detects the artillery rocket, it has no way of determining where the warhead is on the artillery rocket. Second, it is nearly certain that even if the fuse detonates by chance at a time when the warhead might be in the spray pattern of the Iron Dome warhead, the distance between the Iron Dome warhead and the artillery rocket warhead will in almost all circumstances be very large, resulting in a very low density of fragments at the location of the artillery rocket warhead. Given the small number of fragments that can be dispersed by the Iron Dome warhead, this translates into a very high chance that no fragment will hit the warhead. Making matters even more difficult, the projected area of the warhead is very small, since it will be encountered from the front or back rather than from the side. Moreover, fragments are very likely to hit metal surfaces that are at very low grazing angles relative to the direction of the fragment motion. This will result in fragments’ tending to bounce off the shell or transmit almost no energy to a target. Hence, figures 6, 7, and 8 show that for all practical purposes, the probability that the Iron Dome interceptor can destroy the warhead of the engaged artillery rocket is essentially zero.
What does the data show about Iron Dome’s performance in November 2012 and July 2014?
Figures 9, 10, and 11 show contrails in the sky that indicate that Iron Dome interceptors were attempting to engage target artillery rockets either by chasing them from behind or by attacking them from the side.
The geometries of the engagement are easily established because the artillery rockets are falling at high elevation angles relative to the ground—perhaps 60 to 70 degrees relative to vertical. This reëntry angle is due to aerodynamic drag, which slows up the artillery rocket and eventually causes it to fall at a relatively steep angle.
Figures 12 and 13 show photographs that are supposed to have been taken in July 2014. I have found photographs from November 2012 that have been mislabeled as being from July 2012, so I am in the process of verifying that photographs collected were actually taken in the stated time frames. These two photographs have checked out as being from July 2014.
Figure 14 shows a very rough estimate based on my observations in November 2012, when I saw perhaps no more than 10 to 20 percent of Iron Dome contrails that indicated an engagement geometry that was front-on.
As shown in the performance “guesstimate,” if we assume that the engagement geometry and 20 percent of the engagements were front-on, then at that time I estimated the probability of destroying a SCUD warhead might be between 0.3 and 0.6. Thus, if all other engagements effectively resulted in a zero probability of intercept, then the intercept rate would be roughly
0.2 × (0.3 or 0.6) = 0.06 to 0.12
That is an intercept rate, defined as destruction of the artillery-rocket warhead, of between 6 and 12 percent.
My best estimate is that fewer than 20 percent of the engagements I was able to get data on were actually front-on, and I have no information about the actual miss distances or whether the engagement-attempt geometries were close to antiparallel. Thus the statement that the intercept performance of Iron Dome appears to be probably 5 percent or less.
A sample of such a calculation is shown in figure 14.
Why Are Israeli Casualties from Rocket Attacks so Low?
An article published in the journal Nature in 1993 addressed the debate over the performance of the Patriot missile defense in the Gulf War of 1991. At that time, the same questions were being raised—why was damage so low, and why were there so few casualties? (All reports now indicate that there was only one casualty from the direct effects of the SCUD attacks. This casualty was caused by a Patriot missile that dove to the ground in an attempt to intercept a SCUD missile.)
In the case of the SCUD attacks, there were many fewer rockets launched at Israel (perhaps around 40), but the warheads on the missiles were much larger—about 500 pounds. Nevertheless, many SCUD warheads fell in open areas, doing relatively little damage. In cases where warheads fell near buildings, the civil-defense measures essentially protected the population from the consequences of the SCUD impact.
Figures 15, 16, and 17 show damage in Israel from artillery rocket attacks during November 2012 and July 2014. As can be seen by inspecting the photographs, even when the rockets happen to hit buildings, the damage tends to be quite localized. This does not mean that individuals in the area of the rocket attack would not be injured or killed if they were close enough to the impact site, but it is very clear that the warheads are not of sufficient size to cause casualties or deaths to those who are properly sheltered.
In contrast, figures 17 and 18 show the results of bomb attacks in Gaza in July 2014. The exact yields of the bombs are uncertain, but it appears they are probably in the 1,000- to 2,000-pound category. In these cases, attempts at sheltering the population might well fail, as few shelters can sustain the level of damage that could be inflicted by such large bombs.
So again, this illustrates that the small size of the artillery rocket warheads and the ability to quickly warn populations of these arriving small warheads is an extremely capable defense that works far more effectively than Iron Dome.
Theodore Postol is Professor of Science, Technology and National Security Policy in the Program in Science, Technology, and Society at MIT.
Figure 9 (November 2012)
Figure 10 (November 2012)
Figure 11 (November 2012)
Figure 12 (July 10, 2014)
Two rockets being shot down over Sderot Thursday. (Photo credit: Mitch Ginsburg/Times of Israel)
Figure 13 (July 8, 2014)
Image taken on Tuesday, July 8
A rocket exploded near a road in the Sdot Negev Regional Council, causing damage to the road but no injuries. (July 2014)
Roof of attacked cowshed
Iron Dome also intercepted a rocket launched at the southern town of Netivot, also in the Gaza region. In the Ashdod area, some 10 cows died and many others were hurt after a rocket hit a cowshed on a local moshav, residents said.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
How to fix the internet
If we want online discourse to improve, we need to move beyond the big platforms.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.