I started writing this in January and sort of got stuck on it, along with some other things going on. I’m writing here to sort my thoughts out about something, so it may not be the most coherent.
City-fortresses don’t really exist anymore. If they did, what might they look like? Why would anyone want one?
Bottom Line
Fortifying the city as-is will require billions of dollars, a reserve system, and would probably create a tradeoff between mass resettlement and maintaining accustomed levels of economic activity. Two historically novel problems are the volume of traffic required for commerce and the threat of stand-off attack. The fortress would have no strategic depth to mitigate the effects of air and missile attacks, which are much easier to carry out on the margin than even in the mid-20th century.
Contemporary Urban Defense
Defended cities are very difficult for attacking forces to seize and secure. Modern defenders rely on the “depth” provided by urban terrain: as the attacker moves into a city, he becomes more and more vulnerable. This effect generalizes to offensives in general, but urban terrain multiplies it considerably. The problem is not new. As we are about to see, the size of modern cities makes it impractical to defend everywhere at once, and so the defense is not homogeneous; rather, it consists of a combination of outposts, patrols, strongpoints, and mobile response forces. The defender is much aided in his efforts by concealment of different types offered by the terrain — which is not only visual but human and electromagnetic.
The low-intensity counterpart to such hardened defense is the urban police department, with its precincts, patrols, and networks of informants.
Deliberate defense as outlined above works well against a concerted attack. Police methods allow urban administrations to suppress crime and exert control when anyone, more or less, can show up and start living on the territory. No continuous fortification would be more effective than standard depth-based defenses against a large combined-arms attack, and police methods work fine for dealing with criminal infiltrators — at least to the satisfaction of contemporary governments. If someone wanted to fortify a city — in order to completely control traffic in and out, or to prevent depredations by armed groups too large for police but too small to seize the city by main force — what would be required?
The Fortress and the Fortified City
Modern fortresses already exist; these are military installations and especially the “Forward Operating Bases” used by expeditionary militaries. However, I must note the differences between a fortress and a city that has been fortified. A fortress has primarily military functions. It serves as both a logistical support area and as a base for combat patrols by ground, air, and watergoing forces. They inevitably serve some social-economic purposes as well, and given enough time and enough peace these eventually predominate. But such activities are purely ancillary while the base serves its military purpose. If need be, the garrison commander can halt them indefinitely.
Cities serve primarily economic and social functions, whatever their strategic or operational importance. The upshot of this is that cities are far larger in population and territory than a “comparable” fortress. Also, the need for trade outside its walls (and there will be walls) means that the amount of traffic in and out of a fortified city will dwarf that of a forward operating base, the modern fortress.
Remember the Alamo
As an example, I am going to use the city of San Antonio. It is a mid-size (population 1.5 million) city in good health economically, situated in uncomplicated terrain, and with no waterways.
Build the Wall
A continuous protective obstacle is the sine qua non of a fortified city. The wall, if you will, exists to restrict entry to the protected area. It does this in two related ways: by physically impeding the trespasser, and by requiring the trespasser to prove intention in crossing the barrier. Of these, the latter is more important.
The necessity of forcing a potential trespasser to betray his intentions cannot be overemphasized. In a classic defense, the defender has already decided to kill the approaching enemy, and merely needs to determine how he will most efficiently accomplish this. An obstacle will usually help, but the enemy approaching, breaching, or bypassing the obstacle does not affect the defender’s purpose. In this scenario, the barrier itself establishes the nature of the threat. If it can be bypassed in plausibly good faith, it doesn’t work. So it must be continuous.
That being said, we assume that the security situation outside the perimeter is both considerably worse and less definite than that outside of an American city currently. So the barrier will need to be a substantial piece of engineering. Let’s look at some real-world examples.
Building the Wall
Israel-Egypt Border
Israel-Gaza Border
Bagram Airfield, Afghanistan
You’ll notice some commonality: a wall and a road behind the wall (to provide mobility for mounted patrols and response forces). The Israel-Egypt barrier, which runs through a desert, is the least imposing. The layered nature of the Bagram and Gaza walls is visible; the Gaza wall has a quite wide “exclusion zone” of at least several hundred meters (visible on the left of the image). In all three instances, there is a clear field of fire on at least one side of the wall: the Gaza wall has “enemy” structures right up against it but is cleared on the “friendly” side. You can’t tell from the photo, but the Bagram wall is cleared on both sides. This provides both standoff from large explosions and a kill zone for anyone approaching the barrier or attempting to exploit a breach. The “density” or security level per unit length of both the Bagram and Gaza walls is much higher than the section of the Egypt barrier shown, and hardened observation posts are visible. I’ll also point out that Egypt and Israel cooperate on border security; the environment on the outside of the Gaza and Bagram walls is assumed hostile.
The Gaza wall, which includes an “underground barrier“, will cost about $876 million to cover 56km, or about $15.6 million/km. This appears to include outposts and ground clearance but doesn’t include the costs of manning. The comparatively sparse Israel-Egypt wall/fence cost $380 million for 230km or $1.6 million/km. I will assume these are pretty good estimates for an upper and lower bound on barrier construction.
The walls should have an anti-vehicle barrier in front of them, since one of the most straightforward ways to breach a wall is to drive a bunch of explosives right up to it and set them off. A minefield sounds impressive but isn’t really ideal for this. Long-term minefields require tedious and sometimes dangerous maintenance — periodic re-seeding at the very least — to mitigate the effects of weather and wildlife. They easily cause unintended damage and injury, and can be surreptitiously compromised. Protective minefields also require quite large numbers of mines. A “high density” minefield — defined as 85% chance of encounter for a given man or vehicle traveling through the field — requires 960 mines for a 550m frontage. An anti-vehicle ditch has a 100% chance of encounter and won’t blow off a photogenic teenager’s limbs. “Dragon’s Teeth” and other bollard obstacles also have a 100% encounter rate, but aside from obstructing fields of fire more than an enfiladable ditch, it allows smaller vehicles like cross-country motorbikes through. This may be acceptable, especially if terrain isn’t conducive to digging or maintaining a vehicle ditch, and I would guess although I don’t know that maintenance costs would be lower.
But Where?
There are two obvious candidates for a perimeter in our example: the outer Highway 1604 loop and the inner I-410 loop. The outer loop is 151 kilometers in circumference, and the I-410 loop is 84 kilometers. A “max density” perimeter on the outer loop would be around $2.4 billion at $15.6million/km, while a “low density” perimeter on the inner loop would cost $134 million at $1.6million/km.
The inner loop is shorter, obviously. This also means that response forces can reach any given point on the perimeter faster. However, you’re also leaving significant parts of “the city”, including an airport, outside your wall. And as we’ll see in a minute, this greatly complicates entry control because much of the city’s business involves mass daily movement back and forth across I-410. Some of this could be fixed by mass resettlement, but it’s always good to avoid that.
The ideal setup would probably something similar to the Israel-Egypt barrier on the outer perimeter, possibly with an anti-vehicle obstacle. This barrier would be vulnerable to surreptitious compromise along the lines of a few guys with wirecutters getting through, but the goal of the outer barrier is to create depth and establish hostile intent.
While it probably shouldn’t be as imposing as the Gaza wall, the inner perimeter probably needs to be more substantial, since that wall may actually need to provide defensive cover and be more secure against infiltrators. It will be much lower, to allow for observation and direct fire from behind it. It should also impede approach by not only vehicle but by foot, both to prevent assault and to control refugee inflows in an emergency.
Transmural Traffic
As hinted at in the first section, one of the biggest challenges to building a fortified city will be entry control. The traffic into and out of a modern city is far greater than even even the largest hardened military installations.
Using the traffic map from the Texas State Department of Transportation along with some of the information on this page allows us to understand the magnitude of the problem.
The graphic and table below (which includes ancillary routes near the marked checkpoints) gives us an idea of the traffic (both ways). Outer perimeter uses number, inner perimeter uses letters.
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 |
| 151783 | 144681 | 261565 | 55644 | 12151 | 20295 | 44917 | 18983 | 3220 | 48144 | 56411 | 42615 |
| A | B | C | D | E | F | G | H | I | J | K | L |
| 274823 | 133769 | 647420 | 54113 | 22309 | 88493 | 22009 | 32998 | 61922 | 87824 | 117138 | 49373 |
Total Daily Traffic Across Outer Perimeter: 860,409
Total Daily Traffic Across Inner Perimeter: 1,592,191
There is nearly twice as much traffic across the inner perimeter as the outer perimeter.
Entry Control
According to the information here, the busiest United States border crossing (El Paso) is crossed between 4-5 million times per year. This is inbound traffic only, while the numbers above for San Antonio include traffic in both directions. If Fortress Alamo places no restrictions whatever on outbound traffic, giving us about 430,000 entries across the outer perimeter each day, that still dwarfs the ~13,500 entries or so at the busiest US border crossing. In fact, you can see from above that several entry points into San Antonio dwarf the El Paso border checkpoint numbers, even after halving to estimate inbound traffic only.
This means that entry control will require a massive investment of manpower and materiel. It also means that the city authorities will probably want to take measures to systematically lower the amount of transmural traffic — such as by resettling commuters inside the perimeter and removing economically-or-otherwise unproductive populations outside of it. But that is a separate discussion.
The default security level of Fortress San Antonio in the absence of a specific threat would be fairly cursory monitoring of traffic into the outer perimeter comparable to current US border control measures, and increased but still minimal monitoring of traffic into the inner wall along with a pass system. The city can’t hope to eliminate all unwanted visitors using such controls. However, it can act as a deterrent and chokepoint, and prevent large armed bodies from simply moving onto its territory. Specific critical facilities in the city center will still have to be locally hardened and guarded.
The El Paso sector of US Customs and Border Patrol is responsible for monitoring 429 kilometers of the US-Mexico border and operates 11 border stations. I can’t find internal details of CBP’s budget, but the El Paso Sector employs 2,400 agents and an unspecified number of support personnel. Again, I don’t know the breakdown between “border patrol” (the wall) and “customs” (checkpoints). The Fortress San Antonio border/wall is shorter than the El Paso Sector US-Mexico border (and offers interior lines) but inbound traffic is much greater as we saw above. 2,400 agents, then, is a reasonable upper estimate for entry control.
Perimeter Security
Most extant governments have a highly regularized Border Patrol — anyone responsible for securing the border receives standardized training and a regular paycheck from the central government, and everyone else is generally prohibited from enforcing entry laws other than by filing reports with the proper authorities. Since perimeter security in the fortress city — at least along the outer cordon — consists of a screen or perhaps a “tripwire” for commitment of larger response forces in the face of a concerted threat, the 2,400-man border patrol force mentioned is probably adequate for perimeter security as well as entry control.
For administrations comfortable with less regularity, there is historical precedent to franchise perimeter security. Probably the city government would offer a bounty to kill or capture intruders under the supervision of an appointed officer, and those who live near the perimeter would thus be incentivized to prevent infiltration. This would likely be cheaper than a regular border patrol, and would give otherwise marginal populations something useful to do.
These modernized march-lords and Border Reavers would presumably also have the same disadvantages as their historical counterparts. They would be incentivized to fight beyond the specified borders of state (which might not be a bad thing) or figure that they can make easier pickings towards the interior of the polity (absolutely a bad thing). The city would need enforce a deliberate policy of keeping the Borderers less well-armed than the regular response forces, and possibly even less well-armed than the average citizen. Their loyalty would probably be more negotiable than a regular border force. Modern technology could mitigate some of this: detailed census and registration of inhabitants would prevent border ruffians from turning in city citizens or their remains for bounty.
Internal Security & Response Forces
San Antonio employs about 2300 police officers, and we can probably add the 1300 Bexar County sheriff’s deputies to that as well. Given the assumption above that border control will be left to a separate department, this might be sufficient. On the other hand, the combined police, sheriff, and border forces would only be about 6,000, or one security force member for every 250 inhabitants. In terms of police, this is a similar proportion to Singapore. However, Singapore also maintains a large, separate standing military.
The Singapore Army is a conscript-heavy force of around 72000 soldiers, or about one for every 70 citizens. Fortress San Antonio would need about 21000 personnel to field a proportionate force, probably one active division of several brigades, an air wing, and support units including state-level air defense. Like Singapore, this would require a draft of most young men and create a large reserve force of former conscripts essentially consisting of every fit male above a certain age.
Life in Fortress Alamo would be qualitatively more militarized than at present.
Air & Missile Defense
It’s only about thirty miles from the center of the city to the “outer perimeter” outlined above, so virtually any medium-range air defense system could cover the entire city. Of course, waiting for an enemy aircraft or munition to cross the perimeter might not be wise: the tiny patch of airspace over the city indicates the biggest problem with a modern city-fortress, the lack of strategic depth.
Technical limitations here pale in comparison to the diplomatic/political problem of what to do about the vast expanse of airspace outside of the city’s perimeter from which an attack could be launched. It wouldn’t be hard to justify expansion into the surrounding territory on this basis alone.
An interesting note: since the Patriot radar only covers a 120 degree arc, the city would need at least three such systems (realistically four) to get complete coverage. On the other hand, for an attacker such a limited arc would be ideal for suppressing air traffic and high-angle munitions from the city without wasting radar emissions looking the other way. This is an interesting illustration of Clausewitz’s concept of the “diverging” nature of the defense vs the “converging” nature of the offense.
-
- Full coverage on the defense (Aegis, S-300)
-
- Limited-arc defense (Patriot, SA-11/17 TELAR)
-
- Full Coverage Attack
-
- Limited-arc attack
Conclusion
The enormous amount of traffic involved in a modern city’s economy makes a “fortress city” difficult. The financial and intangible costs incurred make doing so infeasible as long as any given city can rely on a larger state for protection; this could be considered a sort of economy of scale. Even if the measures above were taken, it would still be possible for an adversary to infiltrate individuals or small groups even into the city center.
The need to create standoff from air or missile attack is probably an intractable problem. It didn’t occur to me to look into Singapore until near the end of writing this, and I’m very curious now how they handle their air defenses. (The European sovereign cities aren’t as interesting since they’re more or less consciously at the mercy of surrounding states).
Inky
Interesting musings. A couple of assorted thoughts:
In the very beginning you mention urban defense. Conceptually, urban terrain is very hostile towards attacker, and the anti-ISIS operations prove it. But the thing is that the nature of urban defense runs contrary to the functioning of the city itself. Granted, any kind of military activity does, a city is a fantastically complex entity, but urban defense essentially turns city into less of a “human computer” with all of it’s inputs and outputs, and more into a veritable living being, trying to expunge a foreign body. But it comes at a price of a loss of what constitutes a functioning city. A city can become a death trap for the attacker, but in the process it becomes a husk of oneself.
And this is why, it seems to me, our-world entities that are closest to city fortresses — Singapore, and in my opinion, Israel, have such a proactive approach to defense. They have to, since a concerted attack by a military leaves them very vulnerable — no matter how high a wall you will build, modern HE can level it in a matter of minutes, and in any military action they stand to lose more than the attacker, and those losses could not be easily replenished. They are like boxers that can land a nasty punch, but can’t take one. And that’s why for continued survival they absolutely must hit first, and hit hard. Which inevitably makes such party an aggressor, which is not a good thing to be in a modern diplomatic environment.
And this, I think, makes for a biggest problem: the very nature of the modern successful city is hyper-connectendess. Like, look at the Singapore, it’s a world hub of trade, commerce, finance, all kinds of information and material currents. (otherwise, how could it sustain itself? It has no other resources). This way a city state a) is very dependent on the continuity of flow of resources of information and matter — any impediment to them will severely impede it’s economic functioning b) is very dependent on the diplomatic relations with the surrounding countries for much of the same reasons — what are you going to do if no one would trade with you, and you don’t even have all that big stockpiles? Which is the reason why Singapore has to maintain very careful international relationships AND a strong military force to be reasonably secure, and it doesn’t even have any powers that actively want their demise.
That’s why I think for the city-states that have actually fought, you have to look to Israel. Yes, it’s way to big to be considered a city state, but it’s a very small country and has almost the same level of strategic depth as a city has (and anyway, a significant part of it’s economic and human capital is concentrated in the center of the country) and it has hostile (very hostile in the past) neighbors. And actually, thinking about that made me realize just how hard it is to make it as a state in a really hostile environment. It’s nothing short of a miracle that it survived and even apart from that it had one of the best human capitals in the world, tremendous financial support and US having their back whenever a big fight erupted. The situation as it stands now seems more like what you describe and the Israel’s response to the challenge of living in a moderately hostile environment seems to be: secure the borders that is possible to secure and act proactively to disrupt the hostile action on those that cannot be secured completely. And that action can be anything from recruiting informers from the local population up to assassinations of particularly troublesome individuals. And it took decades. But it did work in the end. Prosperity helped too, allowing them to build an ever bigger arsenal of ever more sophisticated tools to maintain an upper hand over the would-be infiltrators.
tl;dr: surviving as a fortress city in any prolonged term seems only possible if you are rich, have a technological superiority over your would-be enemy and can make it work with your diplomacy (or you don’t give a damn, in which case, good luck).
sfoil
Starting to write this really made me grapple with the sheer scope of the problem involved. Like you say: it’s not limited to a tactical/operational issue. It’s a radical departure from how the world is right now. The background assumption is that it’s a matter of military technology, but the real problem is economic patterns. I think something similar is going on with the (lack of) border control in developed nations right now. This implies that an entity capable of maintaining a modern economic standard without such flows has significant military implications.
The biggest problem with what I wrote is equivocation about threat capabilities and intents. Who is the adversary, what can they do and what do they want? Is it a big state who wants to annihilate the city, a rival who wants to shift a balance of power in its favor, or either one trying to improve their economic situation? As you said, urban warfare can “kill” a city, since cities are primarily economic entities. This problem might be made worse by the fact that warfare for economic reasons has gone out of style. On the other hand, John Spencer points out that this is harder than it seems, probably because of geography.
I really ought to look into Singapore more. I suspect that they’re deceptively reliant on sea power given both the geography and social history of their region.
I think you are wrong about the need for technological superiority. Singapore has superior human capital to Indonesia and Malaysia but I don’t think there’s anywhere near the kind of capabilities gap that exists between Israel and its neighbors. I don’t think their existence as an independent state benefits their neighbors economically either. On the other hand, they’re nowhere near as belligerent as Israel.
I think one of the biggest changes we’re going to see is the proliferation of “cool toys” (particularly PGMs) to minor actors. The end result will be more technological parity between forces in every conflict. A major spur for thinking about this is that I see certain people who share this vision predicting that this will make small sovereign states more viable. I’m not sure that their logic is sound — in fact, I suspect it is magical thinking.