Out of the Mountains

by David Kilcullen

  In contrast, TF Ranger in Mogadishu had very little time (less than six weeks) to develop an understanding of the way the city worked—what we’ve called the territorial logic or systems logic of the city. Like LeT, the raiders came from the sea, maneuvering over the ocean in their helicopters before coming in low from the north to strike the target building. Unlike LeT, however, the Americans didn’t nest in the city’s natural flow: they deliberately ignored it. The task force commander, Major General William Garrison, knew the risks of going into Aidid’s stronghold in the Black Sea neighborhood. Most of the United Nations forces in Mogadishu studiously avoided that area, including the Bakara market—the key economic terrain in the city, central to the Habr Gidr’s control of Mogadishu’s urban flow. By assaulting straight into the area, in broad daylight, TF Ranger was directly challenging Aidid’s power base and courting a strong counterpunch.

  As Garrison had warned in a memo to Washington only a few weeks before the battle, “If we go into the vicinity of the Bakara Market, there’s no question we’ll win the firefight, but we might lose the war.”69 Apart from the obvious point that this was not a war but (in its original intent, at least) a humanitarian assistance operation, Garrison’s memo makes it clear that he knew the risks involved. These risks arose not only from ignoring the spatial logic of the city (attacking the places where an enemy was strongest) or the temporal logic of the enemy network (selecting a place where Aidid’s militia could respond quickly by massing combat power at short notice) but also from ignoring the city’s metabolic flow—in particular, the daily qat cycle.

  The leafy green qat plant is chewed as a stimulant across Somalia and the broader Horn of Africa and southwest Arabian region. Q at is very perishable and sours quickly, so over time a complex and informal but highly efficient system has evolved to ensure its timely distribution. The qat system involves a network of distributors and small traders in all major towns and cities and uses aircraft, boats, and trucks to link ports, airports, and distribution hubs with markets and small roadside stalls. This system puts the day’s fresh crop on market stalls across the entire region by mid- to late morning every day, almost without fail. The plant contains an alkaloid compound called cathinone, which has ephedrine- or amphetamine-like qualities. True aficionados—most if not all of whom are men—crave Coca-Cola and other sweet drinks to accompany it. (This, incidentally, is partly why one of the few manufacturing facilities to survive in Somalia throughout the feral chaos of the last decade was a Coca-Cola bottling plant, financed by clan contributions, which opened in Mogadishu in 2004.)70 By midafternoon the daily qat chew is in full swing across the region’s towns and cities, with thousands of young, armed men engaged in argumentative conversation and agitated political discussion that often leads to fights and celebratory (or homicidal) gunfire. By early evening the buzz is over, the qat chewers have crashed, and the city goes quiet. But in the middle of the afternoon most military-age males in town are still violently high, making this perhaps not the best time to attack a nest of heavily armed qat-chewing militiamen.

  The aviators of 160th SOAR are known as Night Stalkers because of their preference—like light infantry and special operations forces the world over—for fighting at night. Infrared and thermal imaging give modern forces of this kind a true nighttime edge and (combined today with real-time imagery from drones, satellites, and surveillance aircraft) can allow rapid and effective maneuver under pitch-black conditions. Indeed, the darker the night, the greater the advantage in strike operations of the sort in which TF Ranger specialized. Again, a midafternoon attack is hardly ideal for a force that’s optimized to fight at night. That General Garrison chose to ignore these issues implies confidence rather than carelessness: as we’ve seen, he knew the risks but counted on his force’s speed of movement and superior combat power, especially its helicopter support, to overcome them. His choice of an afternoon time frame was also almost certainly driven by the perceived need to strike a high-value target—a meeting of Aidid’s two lieutenants and the Habr Gidr leaders, being held in the house that was attacked—before it disappeared. Indeed, this whole style of operations is known in the special ops business as “time-sensitive targeting,” or TST, because of the paramount need for speed.

  This highlights another key difference between Mogadishu and Mumbai: the Mumbai raiders were more or less agnostic as to the individual identity of those they killed. They focused on causing maximum disruption and shock to the city itself, and as described earlier, they killed civilians because of their group identity (their role in the city’s emergency services, their presence at a key urban node, or their nationality) or their media value, rather than targeting particular individuals. In contrast, TF Ranger was going after specific individuals, so the organizing principle of the operation—time-sensitive targeting and individual identity—was completely different. For the LeT raiders the city itself was the target, while individuals were secondary; for TF Ranger the individuals were the target, and this—combined with confidence in their own airborne firepower and mobility—may have led them to discount the effect of their raid on the city as a system; when that system pushed back (and in particular when Aidid’s militia unexpectedly succeeded in disabling their air assets), they quickly lost the initiative and got bogged down in an urban fight in which the locals had clear advantages. If the Mumbai raiders were like a parasite that infested the city and made it convulse, TF Ranger acted instead like a belligerent drunk in a bar brawl, poking someone in the eye and getting punched in return.

  It’s hard to confirm how many of the fighters who confronted TF Ranger in the battle of Mogadishu were members of Aidid’s Somali National Alliance militia and how many were simply local armed civilians: militia affiliations in Somalia were, and are, loose and informal. In any case, the urban density and connectedness of Mogadishu were key factors in the locals’ quick reaction against the raid. As we noted earlier, these days cellphone use is widespread among Somalis, but even in 1993, in the pre-cellphone area, the ability to pass word rapidly in downtown Mogadishu, using radios, runners, and signal fires, was a key local capability. The battle occurred in an area where the terrain and population were intimately familiar to the locals, distances were short, it was easy to move on foot, Aidid’s core group could quickly draw on local allies for reinforcements, and there were multiple routes through the city to and from any given point. The locals could thus react flexibly to American moves—they could aggregate or disperse, their force could shrink or grow in size in response to the changing threat, and they could put ambushes or roadblocks in place ahead of the ground convoy.

  Once Aidid’s militia succeeded in downing the Black Hawks using specially modified RPGs, the Americans became pinned down, and they were forced to concentrate their firepower and air assets around the crash sites. When this happened, instead of exerting a general suppressive effect, TF Ranger was now focusing intense but localized combat power in a tightly limited area, and this relieved the pressure on fighters in other parts of the city, making it easier for them to maneuver with impunity and without detection. Local city dwellers—infuriated by the raids and humiliations of the past few weeks, motivated to repel the violent intrusion of the Americans, and hopped up on qat and Coke—quickly swarmed to the attack from all directions. TF Ranger was indeed in a fight against the whole city.

  Somali Swarm Tactics

  On the other side of that fight, the Somali militia who faced off against Garrison’s troops in the streets of Mogadishu had no formal military training, and many of its members were killed or exiled in the years of feral anarchy that followed. Mohammed Farah Aidid himself died, and his militia was broken up among various clan groups. So it’s impossible to sit down and interview Aidid about his tactical reasoning on the day of the battle, or to directly observe his militia operating in the Mogadishu area. But we can do the next best thing: some units of today’s Somali National Army (SNA) descend from the same clan mili
tias that opposed TF Ranger, and even today these units fight the same way, with almost the same weapons. To the extent that these troops—whose military education consisted almost entirely of on-the-job training in battles among local clan militias in a feral city—represent the self-taught tactics that have proliferated across Somalia over the past twenty years, they can give us an inside view of the fast and deadly swarm tactics that TF Ranger experienced in 1993. They also highlight the differences between true, autonomous urban swarm tactics (as practiced in Somalia) and the superficially similar remote-control system used by the Mumbai raiders.

  Thus in mid-2012 it was a distinct professional pleasure to see these troops in the field during an operation near Afgoye, a town about fifteen miles northwest of Mogadishu in the lower Shebelle River valley. Somali troops, with Ugandan and Burundian forces from AMISOM plus a small number of highly professional unarmed Western advisors, had captured the town from Shabaab a few days before Anna and I arrived, and they were busy consolidating their positions.

  The SNA fighters talked me through the way they operated in the thickets and watercourses on the outskirts of Afgoye. The terrain here was fairly typical of this part of Somalia: an undulating camel-thorn scrub, broken by dry watercourses and the occasional road, farm plot, or cluster of houses, with visibility varying from a few yards to a few hundred feet. They explained how each squad operated with its technical (the gun-carrying truck I mentioned earlier, on which the Somali tactical system is centered) and how the different squads cooperated and coordinated their actions. What was most impressive was the speed and tactical skill with which these fighters—all ex-militia with little formal training—could move and fight.

  These SNA troops had only one basic tactical unit (the mounted squad, comprising a technical carrying a heavy weapon and six to eight fighters with their equipment and supplies) and one tactical formation (the extended line). Their vehicles would move abreast through the bush, about one visual distance apart (a varying space, constantly changing so that each vehicle kept its neighbors just barely in view), and generally avoiding roads. In open terrain the formation would be extremely widely spread, and where visual ranges were shorter it would tighten up. In an urban environment the SNA troops would adopt a variation on this approach, moving whenever possible on several parallel streets at once, picking up their bearings at each intersection in order to stay roughly level with each other. In this way, they achieved the classic tactical goal of moving dispersed but fighting concentrated. Likewise, by expanding the size of their formation as far as the terrain and visibility would allow whenever they were out of actual contact with the enemy, they increased the likelihood that when they did make contact, the flanks of their formation would be wider than the enemy’s position. Like a rugby team playing a running game, their entire approach—the tempo and flow of the way they moved and fought—was designed around creating and exploiting a series of these overlaps.

  Because of the fluid nature of the fighting in the lower Shebelle River valley at this time, most combat actions were encounter battles—engagements where one or both sides are moving or temporarily halted (rather than dug-in in prepared defensive positions). When the SNA encountered a Shabaab group, the SNA vehicles that were first to meet the enemy would immediately halt and lay down heavy suppressive fire. The natural momentum of the advance would cause the other vehicles, not yet in contact with the enemy, to push forward a short distance, perhaps twenty-five or fifty yards, before they had time to react and turn in toward the firefight, putting them naturally in a flanking position. By the time the flanking vehicles did begin to react, there would be no need for radio communication, formal orders, or coordination—each vehicle would simply angle in toward the closest gunfire and, maintaining the extended line, sweep forward until it could see the enemy. This would naturally (again without orders) place these vehicles on the flanks or rear of the enemy, resulting in a quick and automatic encirclement, or near-encirclement, of the Shabaab position.

  Once they could see the enemy, the troops in the back of each technical would dismount and form another extended line on foot about ten yards in front of their gun truck. The vehicle and the soldiers who had dismounted would then sweep forward together, the lightly equipped fighters jogging fast through the bush until they came under fire and were forced to take cover. Because of the technical’s height, the muzzle of its weapon (often a Soviet- or Chinese-made 12.7 mm or 14.5 mm heavy machine gun) could be as high as nine or ten feet above the ground, and it could thus continue firing safely over the heads of the dismounted fighters. The squad leader would coordinate movement using voice commands. He would either fire the heavy weapon himself or stand next to the gunner on the flatbed of the technical, from where he could lean down and direct the driver through a window. Once the dismounted troops were engaged in a direct firefight, the gun truck’s forward advance would stop, and it would lay down fire on the enemy position, allowing other vehicles and their dismounted troops to close in on it (again without orders, simply guided by the sound of the guns). These reinforcements would pile on until the enemy was destroyed or forced to break contact.

  Now, this is an obvious point, but you should understand that I’m putting this into my words, not theirs, and that this is a neatened-up, theoretical description. Real fights are always messy and chaotic, and real fighters rarely do exactly what they’re supposed to do under fire. And yet any leader of irregular cavalry or light infantry (or, indeed, any mounted constabulary officer) of the past century would recognize these simple tactics. Echoing the comment of the special operations officer on Mumbai, any professional soldier in the world would be proud to command troops with this kind of tactical initiative. Indeed, I found only one slight issue on which to fault the SNA tactics: the fact that the squad leader stayed in the vehicle while his troops dismounted to assault. Western tactics would call for the leader to dismount with the troops, carrying a radio to talk back to the vehicle and direct its fire, and leaving a trusted subordinate, as vehicle commander, to maneuver the gun truck.

  But as soon as this thought entered my head, I realized I was looking at the Somali squad in completely the wrong way: I was misapplying the social and economic framework of a professional state-run military to an organization that had evolved from an irregular militia. In the Somali environment of fragmented, semianarchic clan organizations in which these tactics had emerged, the way someone became a squad leader in the first place was to own the technical (an extremely substantial piece of capital equipment). The squad leader became the squad leader precisely because it was his vehicle, so it would have been the height of stupidity for him to dismount and thereby cede control of the gun truck to someone else—let alone to leave someone behind him with a machine gun. He might not have remained the squad leader for long! Moreover, dismounted fighters are cheap and replaceable, but the vehicle is a precious investment that is decidedly not expendable. Seen from this perspective, the SNA’s “mounted swarm” tactics have (like any tactical system) an economic, political, and social logic, as well as a military grammar.

  It actually takes much longer to read these words than to execute a swarming maneuver of this kind. Because each vehicle and its fighters is a semiautonomous unit that needs no formal orders, because the momentum of the advance puts each vehicle in roughly the right position at any given moment, and because the overhead geometry of supporting fire from the vehicle avoids the need for complicated fire control orders, a swarm fight can be incredibly fast and smooth.

  Each dismounted fighter and each vehicle commander need only remember five basic rules. These rules define how the group fights at every scale (the individual, the dismounted squad, the vehicle, and the group of vehicles) and they never change, regardless of the terrain, the tactical situation, or the size of the engagement. They are: “Maintain an extended line abreast,” “Keep your neighbors just in sight, but no closer,” “Move to the sound of the guns,” “Dismount wh
en you see the enemy,” and “When you come under fire, stop and fire back.”

  This explains the speed and flexibility with which the fighters were able to react to TF Ranger’s foray into the Black Sea: the swarm tactics I was observing in 2012 were directly descended from those used by Aidid’s militia in 1993. In systems terms, this kind of autonomous, rule-based maneuver is the essence of a self-synchronizing swarm: like individual birds in a flock, each vehicle and its troops follow a few simple rules to maintain formation and react to the enemy, and like the overall flock, their formation constantly shifts and changes size and shape (without orders) in response to changes in the terrain and the tactical situation. The same rules that bring reinforcements to swell the size of the swarm when it hits a major obstacle also cause it to disperse when there is no imminent threat. In fact, the size, shape, and disposition of the tactical swarm are completely emergent properties of the rule-based swarm maneuver system itself, something that happens without conscious direction or formal control from a central commander.

  This is why, in the battle for Afgoye just past, the Somali brigade commander had simply roamed about the battlefield, armed only with a pistol and carrying the short walking stick that’s the symbol of age and authority across Somalia, and encouraging his troops: he had no need to run a centralized command post, since his fighters fought autonomously by rule, rather than following conscious, formally expressed orders. Rather, the commander’s role was to read the battle, to know where his presence was most needed, and—critically—to think ahead, beyond the current fight, to the next engagement, and the next, and the next. He might also have brought with him, under his personal control, a commander’s reserve of troops, vehicles, and ammunition with which to reinforce weak points or exploit success. A simple handheld radio, connecting him to his most trusted commanders, allowed him to tap into a continuous feed of chatter among those fighting the battle and thus maintain his situational awareness and decide where he needed to be. Strictly speaking, however, such a tactical system should work in complete radio silence, avoiding the need to expose plans to electronic eavesdropping and making the swarm formation relatively invulnerable to jamming or radio deception. Using these tactics, an experienced unit of this type (and remember, these fighters had, in some cases, ten years of nearly continuous irregular warfare under their belts) could—in theory at least—maneuver at a vastly faster tempo than a regular conventional force relying on orders.