Afghanistan and the Future of Warfare: Implications for Army and Defense Policy
Stephen Biddle
Strategic Studies Institute
U.S. Army War College
122 Forbes Avenue
Carlisle, PA 17013-5244
717-245-4126 voice 717-245-3820 fax
[email protected]
20 October 2002
Wartime experience always shapes the postwar policy debate, and the campaign in Afghanistan has had an unusually prompt effect. The global war on terrorism is barely joined. Sporadic fighting is still ongoing in Afghanistan itself. Yet the apparent success of American arms there last fall and winter is already playing a major role in a series of debates over the future of the American military, the conduct of the continuing war, and even the shape of American foreign policy as a whole.
In particular, many now argue that in 2001-2, a novel combination of special operations forces (SOF), precision-guided munitions (PGMs), and an indigenous ally destroyed the Taliban’s military, toppled their regime, and did so while neither exposing Americans to the risk of heavy casualties nor expanding the American presence in a way that might spur nationalist insurgency. In this new “Afghan Model,” it is argued, small teams of elite commandos on the ground provided the targeting information needed for precision weapons to reach dispersed, concealed opponents. Until the commandos arrived, high-altitude bombing could do little against a country with few large, fixed targets. But once Coalition bombing was guided by friendly eyes on the ground, many claim, it became possible for airpower to annihilate the Taliban infantry and armor that had stymied the Northern Alliance for the preceding six years of civil warfare, enabling even an unsophisticated, outnumbered ally to liberate the entire country in a matter of weeks.[1]
Many now believe that this Afghan Model could be used elsewhere with comparable effect. It has been widely reported that senior civilian Defense officials advocate its use against Iraq.[2] Many see it more broadly still – as a new “American way of war” applicable across a wide range of future conflict types. After all, almost any likely opponent has local enemies who could serve as indigenous allies, either domestically or in neighboring states threatened by our enemies’ ambitions – and we can employ special forces and precision weaponry almost anywhere. If this combination works elsewhere with anything like its effectiveness in Afghanistan, then it might enable us to win the wars of the future with minimal U.S. casualty exposure and political footprint.[3] If so, the policy implications could be sweeping. It might, for example, make sense to restructure the U.S. military around the requirements of this new way of war, with larger special operations forces, modernized, longer-range aircraft and missiles, deep reductions in conventional Army and Marine Corps ground forces, major reallocations of Service roles and missions, and sweeping changes in joint military doctrine.[4] Modernization programs for systems originally designed to wage large-scale theater warfare, such as the F-22 Raptor air superiority fighter, the Comanche armed reconnaissance helicopter, or the V22 tilt rotor transport aircraft could all be challenged as backward-looking failures to reflect the new possibilities.[5] The case for expanding the ongoing war on terrorism to topple regimes in Iraq or elsewhere would be strengthened if the Afghan Model were seen as a cheap, effective way to leverage local allies in pursuit of American aims.[6] Most broadly, a powerful, widely-applicable means of waging war at global distances with minor risks would add major impetus to proposals for neo-imperialist conceptions of American foreign policy, in which unilateral U.S. military might underwrites a Pax Americana and advances American values and interests around the world.[7]
Others disagree.
Many argue that the Afghan Model’s success in 2001-2 was idiosyncratic – a product of unique local circumstances unlikely to recur in future conflicts. [8] Any real world outcome is a product of many contributing causes, some of which are replicable and others not; critics see non-replicable features of the Afghan theater or the Taliban opponent as necessary preconditions for the outcome observed there. As Secretary of Defense Donald Rumsfeld put it: “It is hard for me to imagine another Afghanistan. If you think about that situation, it is kind of distinctive. Now it doesn’t mean that some of the things that are working there won’t work elsewhere, but the totality of it is distinctive. I don’t think we’re going to run around with a cookie mold and repeat this.”[9] A related critique rests on the politico-military disadvantages of relying so heavily on local proxies who may not share our aims. Many such analysts point to the fighting at Tora Bora, where Afghan allies with different interests than ours apparently failed to press the attack and may have allowed Osama bin Laden to escape. To secure U.S. interests, they argue, will often require that U.S., not foreign, troops do the heavy lifting.[10]
If so – if Afghanistan was unique or if reliance on proxies is politically unsound – then the policy implications are very different than those seen by Afghan Model proponents. In particular, the case for expanding the war to Iraq is weakened (as this could require major U.S. ground forces rather than the Afghan Model’s small commando teams), the rationale for a sweeping redesign of the U.S. military loses its most trenchant empirical support, and the military demands of a neo-imperialist foreign policy become much more burdensome.
To date, however, this debate has been conducted on the basis of only the most preliminary impressions of the war’s actual conduct. The purpose of this paper is thus to provide a more extensive body of evidence on the way the Model’s elements were actually employed in Afghanistan, the nature and behavior of its targets, and the circumstances of the fighting – and to use this evidence to assess the Model’s role in the Taliban’s collapse on a more systematic basis.
The results suggest that none of the main current views offer a sound account of the war’s actual conduct. The Afghan Model did not in fact work as its proponents now suggest; nor was its success due chiefly to unique properties of Afghanistan or the Taliban.
To account for what is now known thus requires a different explanation of the campaign’s outcome – one resting neither on precision firepower per se nor on Afghan idiosyncrasies.[11] I argue below that the best explanation is that the Afghan campaign was actually far less different or unusual than most now suppose: it was a surprisingly orthodox air-ground theater campaign in which heavy fire support decided a contest between two land forces. Of course, some elements were quite new: the fire support came almost exclusively from the air; the air strikes were directed mostly by commandos whose methods, equipment, and centrality to the outcome were unprecedented; and the ground armies were mostly not countrymen of the commandos and air forces who provided the firepower.[12] In an important sense, though, the differences were less salient than the continuities: the key to success in Afghanistan as in traditional joint warfare was the close interaction of fire and maneuver, neither of which was sufficient alone and neither of which could succeed without significant ground forces trained and equipped at least as well as their opponents. In Afghanistan, our allies provided these ground forces; where others can do so, the Afghan Model can be expected to prevail. Hence Afghanistan is not unique. But not all future allies have armies trained and equipped to their enemies’ standards. Without this, neither the bravery of our special operations forces nor the sophistication of our PGMs can ensure a quick Afghan-like collapse of a resolute opponent.
This in turn implies that we should neither restructure our military to wage Afghan-style wars more efficiently, nor reflexively commit conventional US ground forces in every conflict. Where our allies’ ground forces are good enough, the Afghan Model will work and we should use it. But we will not always be so lucky. In Iraq, for example, the lack of a credible, trained opposition bodes ill for an Afghan-style campaign without major American ground forces. Deep cuts in US ground capability could thus be very risky in spite of our strengths in air power and special operations forces. More broadly, though, we should be wary of suggestions that precision weapons have so revolutionized warfare that traditional ground forces or close combat are now superceded. Where our allies are able, they may provide most of the ground troops for us, but what Afghanistan really shows is that the wars of tomorrow – like those of yesterday – will require tight integration of fires and ground maneuver at close quarters to exploit technology’s effects. Precision weapons make this combination more powerful, but against resolute opponents, neither air power nor conventional ground forces alone will suffice any time soon.
I base these findings on a new collection of primary source evidence centered on a series of 46 interviews with key American participants in the conflict, ranging from Special Forces Sergeants to the Major General who commanded CJTF Mountain during Operation Anaconda, and including subjects from the Special Operations Command, the U.S. Army, the U.S. Air Force, and the Central Intelligence Agency.[13] These interviews were complemented with official written documentation on the conduct of the war and direct physical inspection of the Anaconda battlefield in Afghanistan’s Shah-i-kot valley, together with available secondary source accounts, chiefly from the print news media. This body of evidence cannot be considered complete; a definitive history of the Afghan campaign would require years of research on a much broader range of issues. Rather, my intention here is to focus on one key issue – the new Model’s role in the Afghan campaign and its implications for the future – and to muster as much evidence as can be produced in the near term, so as to make initial findings available sooner than a definitive history would permit, but with a stronger foundation in the evidence than the debate to date has offered.
I present the analysis in five steps. First I outline the key events in the Afghan campaign. Next I assess in turn the two main current explanations of the Taliban’s fall: first, the school that sees the causes as local and idiosyncratic; second, the school that situates them in replicable features of the Afghan Model.[14] I then present an alternative explanation, and show how it accounts more successfully for the available evidence. Finally, I discuss the resulting implications for the Army and for defense policy more broadly.
Key Events
The bombing campaign began the night of October 7, 2001, and focused initially on destroying the Taliban’s limited air defense and communications infrastructure. American and British SOF teams had been conducting scouting missions in Afghanistan beginning at least a week prior to the first air strikes; by October 15, teams designated to make contact with the major Northern Alliance warlords had been inserted and begun preparations for combined offensive action against the Taliban.[15]
[Figure 1, map graphic, about here]
Some of the first major combat actions occurred in the mountains south of Mazar-e-Sharif, as SOF teams working with Northern Alliance Generals Abdul Rashid Dostum and Atta Mohammed fought their way north toward Mazar up the Dar-ye Suf and Balkh river valleys. The terrain and conditions here were extraordinary: at elevations of up to 6,400 feet, movement was restricted to winding mountain trails in which sheer rock faces were sometimes separated from thousand-foot drops by no more than a three-foot width of rocky path. With no vehicles able to negotiate such trails, commandos hauling loads of over 40 pounds of equipment per man were given Afghan mountain ponies with wooden saddles and told to ride along with Dostum’s troops. Luckily, the SOF team commander assigned to Dostum had been a high school rodeo rider in Kansas, but none of the other Americans had ever been on a horse before. Their knees in their chests, balancing heavy rucksacks on their backs, they were instructed by their commander to keep their downhill foot out of the stirrups and to lean uphill so if the pony lost its balance they would fall onto the trail as the pony went into the gorge. On particularly rocky stretches the team commander ordered his men to travel with weapons out and a round chambered to shoot immediately any pony that bolted before it could drag its rider to his death over the rocks.[16]
The first combat action in this sector came when Dostum’s group took the village of Bishqab on October 21; this was followed by engagements at Cobaki on October 22, Chapchal on October 23, and Oimetan on October 25. The key battle came when Dostum’s troops encountered hostile forces occupying old Soviet-built defensive positions at the hamlet of Bai Beche some 16 kilometers southeast of Keshendeh-ye Pa’in.[17] On November 5, Dostum’s cavalry overran these defenses; shortly thereafter Atta’s forces captured Ac’capruk on the Balkh river, and the door swung open for a rapid advance to Mazar, which fell to Atta and Dostum’s forces on November 10.[18]
The fall of Mazar unhinged the Taliban position in northern Afghanistan. Taliban defenders near Bamiyan resisted briefly before surrendering the city on November 11; Kabul fell without a fight on November 13. A force of some 5,000 Taliban and al Qaeda survivors were then encircled in the city of Konduz, where they surrendered following a 12-day siege on November 26.[19]
Many of the prisoners taken at Mazar were subsequently detained at the 19th century mud fortress of Qala-e-Gangi west of the city. On November 25, a revolt at the prison killed one American and gave rise to a two-day struggle before control of the prison was reestablished on the 27th.[20]
With the fall of Kabul and Konduz, attention shifted to the Taliban’s stronghold of Kandahar in the south. SOF teams in support of Hamid Karzai’s forces advanced on the city from the north; teams in support of Gul Agha Shirzai advanced from the south. The result was a series of battles at Tarin Kowt and Sayed Slim Kalay north of the city on November 18 and December 2-4, respectively, and along Highway 4 south of Kandahar from December 2-6. On the night of December 6, Mullah Omar and the senior Taliban leadership fled the city and went into hiding, ending Taliban rule in Afghanistan.[21] Allied forces subsequently tracked a group of al Qaeda survivors thought to include Osama bin Laden to a series of redoubts in the White Mountains near Tora Bora. The redoubts were taken in a 16-day battle ending on December 17, but many al Qaeda defenders escaped death or capture and fled across the border into Pakistan.[22]
In March, a second concentration of al Qaeda holdouts was identified in the Shah-i-kot valley and surrounding mountains east of Gardez. In Operation Anaconda, a combined offensive by two battalions of U.S. regular infantry from the 101st Airborne and 10th Mountain divisions, supported by allied Afghan and Western troops and special operations forces from seven nations, descended on the al Qaeda defenders, killing many, dispersing the rest, and bringing to a close the major combat operations in the country as of this writing. [23]
Existing Views: The Taliban’s Fall as Idiosyncratic
Were these results the products of local idiosyncrasies? Afghanistan is certainly an unusual place, and the Taliban were unusual opponents.[24] Several of their peculiarities might in principle have been necessary preconditions for the Afghan Model’s success, including:
• Poor enemy morale or motivation;
• Poor enemy military training and expertise;
• Lack of popular support for the Taliban regime;
• The ease of defection in Afghan culture;
• Surprise;
• Taliban dependency on fragile sources of outside support; or
• Availability of contiguous, secure territory for re-supplying and re-equipping proxy forces. •
Yet on closer inspection, the evidence suggests a weaker causal role for these idiosyncrasies than some now suppose.
Atypically poor enemy morale
The native Afghan Taliban, for example, were fighting only to prevent the extradition of a foreigner, Osama bin Laden, whom many of them resented anyway. They might be expected to display only limited commitment to such modest stakes, and hence poor morale and combat motivation would hardly be surprising. Future opponents, by contrast, for whom much more could easily be at stake, might be better motivated. If poor enemy morale were a necessary precondition for the Afghan Model’s success, then the Model might not succeed elsewhere. [25]
Yet much of the actual fighting was shouldered not by Afghan Taliban but by foreigners, and especially by al Qaeda – and these foreign troops were much better motivated than their Afghan comrades. Our opponents in this campaign were not a monolithic or homogeneous military. Instead, their three major components (the indigenous Afghan Taliban, the predominantly foreign al Qaeda, and other, non-al Qaeda foreign allies of the Taliban) differed in many important respects.[26] Among the most important of these differences was combat motivation: whereas the indigenous Afghan Taliban quit the field quickly, the foreigners did not. On the contrary, there are numerous reports of foreign fighters threatening to kill any Afghan Taliban who fled the front lines.[27] At Bai Beche, Konduz, Sayed Slim Kalay, Highway 4, and elsewhere in the theater, al Qaeda fighters launched counterattacks to try to close with American and allied forces.[28] In Operation Anaconda, al Qaeda defenders not only stood their ground against overwhelming American firepower, they actually reinforced their positions in the midst of the battle: their fighters were willing to advance into the teeth of a fierce bombardment to enter the Shah-i-kot Valley from safer positions elsewhere and seek battle with our forces.[29] These are not the actions of unmotivated or unwilling troops. The non-Afghan Taliban have shown themselves to be resolute combatants with the motivation to stay in the field when there seemed to be something to be gained militarily from doing so.
These better-motivated foreigners were responsible for much of the fighting in the campaign’s key actions. Though comprising under 25 percent of the regime’s overall troop strength, the hard-core foreigners were allocated disproportionately to the critical sectors of the front and bore an increasing fraction of the combat load beginning in late October.[30] By early-to-mid November, American special forces teams were reporting predominantly al Qaeda, rather than Afghan, opposition in the critical battles for Mazar-e-Sharif and Kandahar.[31] The Afghan Taliban’s tendency to quit the field following their initial defeats doubtless evened the numerical odds for the decisive battles in November and December, but the key engagements at Bai Beche, Sayed Slim Kalay, and Highway 4 were all fought against motivated, resolute opponents – as were the later actions at Tora Bora and Anaconda. If poor enemy morale were a necessary precondition for the Afghan Model to succeed, then the Model should thus have failed after its initial victories in October, yet it did not.
Atypically poor enemy training
The Afghan Taliban were often very poorly trained soldiers. Many had little or no formal military instruction, and Afghan ranks swelled and shrank with the seasons and the fortunes of war as troops went home to their villages or took up arms depending on the crop cycle and apparent military need.[32] In fact, former Supreme Allied Commander in Europe Gen. Wesley Clark has referred to them as “the least competent adversary we’ve faced since the Barbary Pirates.”[33] Other enemies might well be better trained; if the Afghans’ poor skills were a necessary precondition for their defeat, then their collapse might be hard to replicate elsewhere.
Yet here, too, the distinction between Afghan and foreign Taliban is important. The al Qaeda soldiers in the key battles were much better trained than the Afghan Taliban. Unlike the latter, which were often closer to a civilian militia than a professional army, the al Qaeda troops were committed, full-time soldiers. Osama bin Laden’s infamous training camps served primarily to prepare these troops for combat on Afghanistan’s front lines, and taught them a curriculum not radically different from orthodox Western armies’.[34] While the result may not always have been up to contemporary Western standards, there is no reason to suppose that al Qaeda’s skills or training were not at least comparable to those of most other likely opponents, few of whom are trained to Western standards either.[35] The Afghan Model’s ability to prevail against this better-trained opposition suggests that atypically poor enemy skill was not a necessary precondition for success.
Atypical lack of popular support for the enemy’s war effort
Many Afghans hated the Taliban and chafed under the harshness of their regime.[36] This surely weakened their Afghan soldiers’ morale, and presumably reduced the material and intelligence assistance a state might otherwise expect from its citizens. Enemies with stronger popular political support could well be much harder to defeat.
Yet the Taliban’s unpopularity was hardly unique. Many other potential opponents’ regimes are at least as unpopular: Saddam, for example, hardly governs by the will of the governed. If unpopularity makes a military unable to resist American arms, then this hardly makes Afghanistan unrepresentative of the future.[37]
Nor were the Taliban uniformly hated. In southern Afghanistan, their traditional stronghold, they remained popular throughout the war and indeed to this day. In fact, in November the conventional expectation was that Kandahar and the south would prove much harder to conquer than Mazar and the north, given the former’s greater sympathy for Mullah Omar and his regime. [38] These concerns proved unfounded, but not because the southern public failed to support the government. Moreover, the foreign Taliban were apparently undeterred by public opinion in Afghanistan – they were waging a jihad whose legitimacy in their minds was unrelated to the political views of Afghans, and they fought hard regardless of the latter. On balance, there is little reason to assume that the Taliban’s degree of popular support was either unique or an essential precondition for their collapse.
Atypical enemy willingness to defect
The acceptability of defection in Afghan military culture is certainly unusual to Western eyes. Not only do many Afghans consider it normal and acceptable to switch sides in war, but most of their senior commanders know one another personally from their service against the Soviets. Hence it is probably easier to persuade an opponent to defect in Afghanistan than in much of the rest of the world.[39] And there were certainly many defections in fall and winter of 2001-2. If the Taliban’s loss of combat power resulting from these defections were a necessary precondition for our success, then the 2001-2 outcome might well be an anomaly.
Yet the overwhelming majority of defections in this campaign occurred after the military tide had turned, not before. The largest mass turnovers came with the fall of the northern cities of Mazar-e-Sharif and Konduz on November 10 and 26, respectively.[40] The key battle in the north, however, was east of Keshende-ye Pa’in near Bai Beche on November 5. Prior to this, the Northern Alliance and its accompanying SOF teams had been forced to work their way methodically up the Dar-ye Suf and Balkh river valleys, fighting their way from one Taliban defense line to the next. The Taliban, however, had chosen to make a stand at Bai Beche, pushing all available reserves forward to reinforce these positions. Once Dostum broke through their lines, the Taliban thus had nothing left to back them up and the result was a rout: “the dam broke,” as one Special Forces soldier described it, with the Taliban retreating in disorder and unable to rally or establish new blocking positions.[41] The Taliban’s consequent loss of the Dar-ye Suf corridor made Mazar-e-Sharif untenable, and when it fell their entire position in northern Afghanistan collapsed.[42] By the time the major defections occurred in mid-to-late November, the military outcome in the north had thus been determined. By contrast, before the battle at Bai Beche, the SOF commanders operating in the north reported few if any defections to the forces they served with.[43] Moreover, even after Bai Beche, few al Qaeda voluntarily defected, and in fact they continued to resist stubbornly around Kandahar in the south through December 6, and fought hard at Tora Bora and Anaconda well into 2002.[44] Some were captured when overrun or encircled, and many tried to blend into the population and slip away once their defenses had been breached, but most of the voluntary defections were by Afghan Taliban – and even for them, most stayed and fought at least until the military tide had turned. Defections in the Afghan campaign were thus important, but they were effects not causes; consequences of military failure not determinants of it.
Atypical susceptibility to surprise
The Afghan Model surely surprised the Taliban. Though neither SOF nor PGMs were new, they had never been combined in quite this way before, presenting a novel problem to defenders.[45] Other potential opponents, however, have now seen the Model in use and are surely studying means of thwarting it. The Taliban’s ignorance of the Afghan Model is thus probably unique to Afghanistan – our next opponent will not be caught as unsuspecting as the Taliban were in October.
Yet while the Taliban were initially surprised, they adapted quickly. As early as October 22, American SOF teams were reporting that Taliban vehicles in their sectors had been smeared with mud to camouflage them.[46] By Bai Beche they were already making aggressive use of overhead cover and deliberate concealment.[47] In the fighting north of Kandahar and along Highway 4 in December, al Qaeda defenses were well-camouflaged, dispersed, and making use of natural terrain for expedient cover.[48] This pattern continued through Operation Anaconda in March, by which time al Qaeda forces were practicing systematic communications security, dispersal, camouflage discipline, use of cover and concealment, and exploitation of dummy fighting positions to draw fire and attention from their real dispositions.[49]
Others may adopt different countermeasures, but those the Taliban turned to by early November were the natural choices one could expect an adaptive opponent to try to implement when faced with capabilities like the Afghan Model’s. Again and again, from Western Front armies’ response to artillery and machine guns in 1916-18, to the German Wehrmacht’s response to Allied air supremacy in northwest Europe in 1944, to the Vietcong and North Vietnamese armies’ response to U.S. firepower in Vietnam, armies facing overwhelming firepower have dispersed, exploited cover and concealment, restricted their radio transmissions, and adopted camouflage.50 In fact, this pattern parallels the approach suggested by several U.S. marine corps officers assigned to oppose an operational concept resembling the Afghan Model in a pre-911 exercise.[51]
Much of the actual fighting in Afghanistan thus came against an opponent who had already begun to adopt the standard countermeasures that one would expect to see against the kind of firepower we employed there; the effects of surprise were attenuated rather quickly in this campaign. Yet the loss of surprise did not prevent us from continuing to prevail on the battlefield. Surprise per se thus does not appear to have been a necessary precondition for success in Afghanistan.
Atypical enemy dependence on vulnerable sources of outside support
The Taliban had certainly enjoyed important outside support. They were largely a creation of the Pakistani intelligence service, the ISI, which continued to supply them with materiel, tactical advice, and intelligence throughout the civil war. Shortly after September 11, Pakistani President Pervez Musharaff threw his support to the United States and pledged to end ISI assistance to the Taliban, who could thus have expected to lose a major source of outside assistance. Relative to other potential opponents, the Taliban might thus be considered unusually vulnerable to such a cutoff.
Yet aid withdrawal was probably not decisive. The key battles were fought in early November and December. Pakistani aid continued to reach the Taliban until at least October 12, and may have continued even after that.[52] Most militaries stockpile supplies against the possibility of interdiction, hence supply cutoffs typically take time to affect battlefield outcomes. Even if all outside sources had dried up instantly on October 12, it is unlikely that such a short period of isolation could have induced catastrophic military consequences for the Taliban. And in fact, American inspection of captured Taliban positions indicates that they remained well-stocked (especially with ammunition) through at least March 2002.[53] Nor does the withdrawal of Pakistani intelligence or tactical advice appear to have had decisive effects. Sympathetic civilians have regularly provided al Qaeda with information on Allied movements with or without help from the ISI; this problem was aggravated with the return of Afghan Taliban fighters to their home villages following their initial defeats, placing thousands of still-hostile observers in villages across the country.[54] Enemy tactics, moreover, typically improved, not decayed, as the war has progressed and more sophisticated al Qaeda fighters replaced less-skilled Afghans – withdrawal of ISI tactical assistance hardly undid the Taliban tactically.[55] In all, there is little basis to see the Pakistani aid cutoff as a necessary precondition for success.
Atypical availability of safe havens for training and equipping indigenous allies
The Northern Alliance controlled an unusually large contiguous territory in which they could be re-supplied, retrained, and reequipped: perhaps 10-15 percent of Afghanistan’s total land area was under their control prior to American intervention.[56] Other potential allies might not enjoy such extensive or contiguous safe havens. If re-supply, retraining, and re-equipment were necessary preconditions for the Afghan Model’s success, then this territorial advantage might make the 2001 outcome idiosyncratic.
Yet here, too, the tide turned before its effects could become decisive. Most of the key SOF teams were calling air strikes and supporting Northern Alliance offensives within hours of arriving in Afghanistan – the fighting was well under way before the Americans could do any significant retraining, and before any major re-supply could be accomplished.[57] Once the battle was joined, moreover, the high tempo of operations made it impossible for American SOF to pause long enough to conduct training. Some food and other humanitarian aid arrived quickly, but few weapons and little ammunition were dropped prior to the critical battle at Bai Beche: Dostum’s first significant lethal aid delivery was not until 2:00 a.m. on November 5, that is, the predawn hours of the day of the battle itself. And even then, this aid was limited to small arms, ammunition, and rocket propelled grenades – no heavy weapons or armored vehicles were provided.[58] Most of the key battles were thus fought by Afghan allies with skills and equipment little different from what they had in the civil war; Afghanistan’s suitability for retraining and re-equipping played little role in the outcome.
Existing Views: Afghan Model Lethality
There is thus substantial evidence that the campaign was not merely a fluke, hence its implications may have some real meaning for the future. But are those implications necessarily the ones that Afghan Model proponents now see? In fact, they are not – because the Model did not actually work in Afghanistan the way most now think it did. And its actual performance implies a less radical departure from past military practice than many Afghan Model proponents argue.
For the Afghan Model to be a widely applicable template suitable for restructuring the U.S. military, several key features must have obtained in Afghanistan:
1. SOF and standoff sensors must have been able to find key targets for precision engagement;
2. PGMs must have been able to kill the targets found, at standoff ranges;
3. The indigenous allies’ role must have been undemanding.
The first two conditions are logically necessary for the efficacy of the PGMs around which the Model is built – strike systems must be able to find and kill the critical targets at long range, before enemies can close with and overwhelm the small SOF teams on the ground. The third, however, is at least as important. We can almost always find local allies, but armies vary widely in skill, equipment, and motivation; some will be capable of demanding tasks, but others not. Much of the Model’s apparent power in Afghanistan was due to the perception that our local allies there were not terribly capable; if SOF and PGMs can make conquerors out of such modest material then they can do so in many other places, it is often supposed. In particular, the key roles implied for the allies are to screen U.S. SOF from hostile patrols or infiltrators too small to make good targets for PGM attack; to mop up surviving remnants; and to occupy abandoned territory (it is often politically important that local forces be the ones to take control of the ground that the PGMs effectively clear). If the allies’ mission becomes much more demanding than this, then both the generality and the novelty of the Afghan Model come into question – after all, allies like the Israelis or the British have always been capable of impressive feats with or without American PGMs; the claim that the Afghan Model is a revolutionary departure rests on the perception that it turned an overmatched militia into a juggernaut.
In Afghanistan, however, these three conditions were not always met. SOF and standoff sensors found many targets, but could not ensure acquisition of properly concealed opponents. PGMs were very lethal, but against well-entrenched positions they could not destroy enough to prevent survivors from halting unsophisticated attackers. Our indigenous allies’ role was undemanding at times, but far more challenging at others. In particular, significant numbers of properly prepared al Qaeda positions eluded SOF and remote target acquisition and survived even extensive air strikes. To take such positions required orthodox combined arms, fire-and-maneuver tactics – even with the intense precision fire support available in this campaign. Such tactics are difficult to execute, and require skills and leadership many potential indigenous allies lack; in fact, not all our Afghan allies in 2001-2 were up to the job. Where the available ground forces faced skilled enemies but had the necessary skills themselves, American firepower brought decisive victory. But where our ground elements lacked the needed skills or motivation, offensives against capable enemies stalled and outcomes were frequently in doubt in spite of American air support. In Afghanistan, friendly ground forces were mostly as good as their enemies, hence our theater offensive as a whole overwhelmed the Taliban and drove their regime from power. But the variations in tactical behavior in this campaign produced important variations in local outcomes – and these differences teach important lessons about how future wars with different balances of ground force skill and motivation might turn out.[59] These lessons in turn imply a different understanding of cause and effect in Afghanistan, a different model for the future of warfare, and different implications for Army and defense policy.
To show why, I address each of the three key conditions in turn.
1. Target Acquisition in Afghanistan
In the campaign’s early stages, enemy targets were exposed and relatively easy to acquire. The Afghan Taliban commonly deployed on ridge crests, silhouetting themselves for SOF observation. Afghan defenders made little or no effort to camouflage or conceal their positions. They engaged in casual, exposed movement in the target area, often emerging to inspect bomb craters while still under surveillance. Armored vehicles and crew-served weapons were deployed without revetments or apparent efforts at cover or concealment. As a result, their positions were easily identified for PGM targeting, sometimes from extraordinary distances.
At Bishqab on October 21, for example, US SOF observers acquired Taliban targets at ranges of 8-10 kilometers.[60] At Cobaki on October 22, Taliban observation posts were easily spotted at 1500-2000 meters.[61] At Zard Kammar on October 28, Taliban defensive positions were visible from a mile away.[62] At Ac’capruk on November 4, exposed Taliban combat vehicles and crew-served weapons on hillsides west of the Balkh river could be identified from SOF observation posts on the Koh-i-Almortak ridge line some 4-5 kilometers distant.[63] At Polanyi Canyon near Bamiyan on November 7, Taliban positions were plainly visible from more than four kilometers away.[64] Taliban counterattacks, as at Tarin Kowt on November 18, were conducted in the open with no evident attempt at concealment, dispersal, or covering fire to keep U.S. or allied observers’ heads down or to complicate target acquisition.[65]
This changed as the campaign continued and as the opposition shifted increasingly from Afghan Taliban to al Qaeda. In the November 5 battle at Bai Beche, for example, al Qaeda’s defensive works greatly complicated American target acquisition. While American SOF observers knew they faced an interconnected trench system with the enemy somewhere within it, and while particular strongpoints could be identified for engagement, al Qaeda’s local cover and concealment were good enough to prevent the SOF from locating the entirety of the enemy’s individual fighting positions, many of which could not be singled out for precision attack.[66]
By the December fighting along Highway 4 south of Kandahar, not even this was available. In fact, concealed al Qaeda defensive positions among a series of culverts and in burned-out vehicle hulks alongside the road remained wholly undetected by any friendly element until their fire drove back an AMF (Afghan Military Forces – our Northern and Southern Alliance allies) advance. An al Qaeda counterattack in the same sector using a system of wadis for cover approached undetected to within 100-200 meters of AMF and American SOF positions along the highway before opening fire on friendly forces.[67]
At the village of Sayed Slim Kalay north of Kandahar between December 2-4, concealed al Qaeda defenders likewise remained undetected until they fired upon unsuspecting U.S. and AMF attackers. An al Qaeda counterattack using local terrain for cover maneuvered into small arms range of friendly defenders before being driven back.[68]
At Operation Anaconda in March 2002, an intensive pre-battle reconnaissance effort focused every available surveillance and target acquisition system on a tiny, ten-by-ten kilometer battlefield. Yet fewer than 50 percent of all the al Qaeda positions ultimately identified in the course of the fighting on this battlefield were discovered prior to ground contact. In fact, most fire received by U.S. forces in Anaconda came from initially unseen, unanticipated al Qaeda fighting positions.[69]
How could such things happen in an era of persistent reconnaissance drones, airborne radars, satellite surveillance, thermal imaging, and hypersensitive electronic eavesdropping equipment? The answer is that the earth’s surface remains an extremely complex environment with an abundance of natural and manmade cover and concealment available for those militaries capable of exploiting it.
Figure 2 provides a concrete illustration of this problem in the form of a photograph of an al Qaeda fighting position from Objective Ginger on the Anaconda battlefield.[70] The yellow arrow indicates the al Qaeda defenders’ location; without the arrow, there would be no visible sign of a combat position even from the nearly point blank range at which this photograph was taken. Overhanging rock in turn provides cover and concealment from overhead surveillance systems. In principle one might hope to observe resupply movement or al Qaeda patrols into or out of such positions, or to overhear radio communications from its occupants. Al Qaeda soldiers wearing the flowing robes of local herdsmen and traveling in small parties among the mountains, however, are nearly impossible to distinguish at a distance from the noncombatants who tend goats or travel through such areas as a matter of routine.71 And defenders able to operate under radio listening silence while communicating using runners, landlines or other non-broadcast means can reduce signals intercepts to a level that makes identification of specific fighting positions very problematic. Against such targets, it is far from clear that any surveillance technology coming any time soon will ensure reliable target acquisition from standoff distances.
[Figure 2, Objective Ginger fighting position, about here]
Nor are positions such as this one rare anomalies or atypical of Afghan terrain more generally. Figures 3 and 4 show broader samples of the Shah-i-kot battlefield on which Anaconda was fought, including the features known as “The Whale” (after a similar rock formation at the U.S. National Training Center at Ft. Irwin California) and Objective Ginger, respectively.[72] Almost any of the dozens of shadows, crevices, or folds in the earth scattered across these landscapes could house positions like the one shown in Figure 2. And this is just a tiny subset of even the Anaconda battlefield, which is itself a tiny subset of Afghanistan as a whole. The natural complexity of such surfaces offers any adaptive opponent with the necessary training and skills a multitude of opportunities to thwart even modern remote surveillance systems. Against such opponents, remote surveillance will still detect some targets, and remote sensors remain crucial assets, but the only sure means of target acquisition is direct ground contact: a ground force whose advance threatens objectives that the enemy cannot sacrifice and thus must defend compels them to give away their locations by firing on their attackers. Skilled attackers can eventually locate any defensive position by observing the source of the fire directed at them – and this, in fact, is how the majority of the al Qaeda positions at Anaconda were found.
[Figures 3 and 4, The Whale, Shah-i-Kot Valley; and Objective Ginger, Shah-i-kot Valley, about here]
Nor is this problem unique to Afghanistan or its mountainous terrain. Militarily exploitable cover is commonplace in almost all likely future theaters of war. For targets who observe radio listening silence, as al Qaeda now does, foliage, for example, degrades all current remote sensor technologies; urban areas provide overhead cover, create background clutter, and pose difficult problems of distinguishing military targets from innocent civilians.[73] Each is widely available. More than 26 percent of Somalia’s land area is wooded or urban, as is more than 20 percent of the Sudan’s, 34 percent of Georgia’s, or 46 percent of the Philippines’.[74] This cover, moreover, is often distributed in small, widespread patches. On the Goodwood battlefield of 1944 in Normandy, for example, over 80 percent of all one-kilometer grid squares now contain at least some forest or urban cover (though only 26 percent of the total land area is covered).75 In most countries, the central geostrategic objectives are urban areas; even where the bulk of the national land area is open desert (as in Iraq), the cities are both the key terrain and an ample source of cover from overhead sensors (Baghdad alone covers more than 300 square kilometers).[76] Among the most important themes in the history of modern tactics is the growth of methods for exploiting such cover to reduce vulnerability to modern firepower – the “empty battlefield” that has characterized the modern era is a product of skilled armies’ ability to find cover sufficient to thwart standoff target acquisition and to exploit this cover to perform meaningful military missions without excessive exposure.[77]
This is not to suggest that modern sensors are useless or that anyone anywhere can fight effectively from cover opaque to standoff observation. To find and exploit cover while taking or holding ground requires a very demanding set of tactical skills. Historically, armies have varied widely in their ability to do this, and in fact the Afghan campaign itself displays substantial variation in the Taliban’s ability to do so: the Afghan Taliban early in the fighting were systematically unable to do so; only the foreign Taliban and al Qaeda encountered later in the campaign proved able to contest territory from covered, concealed positions. The less skilled the opponent, the more exposed they will be to remote target acquisition. And no army can disappear utterly: diligent reconnaissance will always uncover part of an enemy’s dispositions; the better the sensors, the more they will find, and today’s sensors can find enough to be a crucial contributor to success in theaters like Afghanistan. But this is not to say that they can find enough – on their own – to break a skilled, resolute opponent by standoff fires alone. Even today’s best sensors are still far from an ability to acquire most or all of a hostile force that has learned to exploit the natural complexity of the earth’s surface for cover and concealment – as our experience against al Qaeda in 2001-2 demonstrates.
2. PGM Lethality in Afghanistan
The second logical requirement for Afghan Model effectiveness is PGM lethality: standoff weaponry must be able to kill the targets acquired. As with target acquisition, however, lethality varied widely in Afghanistan as a function of the targets’ tactics. Where the Taliban presented exposed or massed targets in the open, PGMs were extremely lethal. At Tarin Kowt on November 18, for example, Taliban forces tried to recapture the village by advancing in a column of vehicles up an exposed road. Frightened AMF defenders were prepared to abandon the village, but precision air strikes called in by American commandos located on an overlooking ridgeline decimated the Taliban column, whose survivors fled the scene in disorder.[78] Taliban reserves ordered forward to reinforce their defenses at Bai Beche were caught moving in the open along Highway 326 between Shulgerah and Keshendeh-ye Pa’in and were slaughtered by American airpower; officers who surveyed the scene afterward said it brought to mind the infamous “Highway of Death” leading out of Kuwait City in the 1991 Persian Gulf War.[79] Taliban armored vehicles and heavy weapons left parked on the surface were routinely destroyed by pinpoint PGM attacks as at Bagram on October 20, Oimetan on October 25, Ac’capruk on November 4-7, or Polanyi Canyon on November 7.[80] Crude foxholes dug in the open without overhead cover or effective concealment were regularly annihilated by precision bombing with satellite-guided 2,000 pound JDAMs (Joint Direct Attack Munitions).[81]
Where fighting positions were properly prepared, however, they were much harder to destroy – even with modern PGMs. At Bai Beche, for example, al Qaeda’s fighting positions offered ample overhead cover and reinforced firing positions from which attackers could be engaged without exposing the garrison. Though the defenders could not all be located individually within this trench system, American commandos knew the system’s extent and thus called for heavy bombing across the entire position for more than two days. Yet even after all this, enough defenders survived to thwart Dostum’s initial attack.[82]
In the Qala-e-Gangi fortress uprising, the renegade prisoners were quickly driven out of the above-ground prison yard and isolated in a handful of small underground chambers whose locations and perimeter were well-known. These were then pounded by allied airpower: entire ammunition payloads of multiple AC-130 gunships and no fewer than seven 2000-pound JDAMs were expended against this tiny area. Yet the defenders survived and continued to resist until succumbing only to the medieval technology of flooding by cold water.[83]
In Operation Anaconda, well-prepared al Qaeda positions survived repeated aerial attack by U.S. PGMs. On Objective Ginger, for example, American infantry inadvertently disembarked from their assault helicopters almost on top of an unseen al Qaeda position on March 2; after being pinned down for much of the day, they were extracted that night. American troops then spent much of the next 10 days fighting their way back toward the Ginger hilltop from more secure landing zones well to the north. In the meantime, American aircraft pounded the hill. Yet in spite of over a week of sustained heavy bombing, al Qaeda positions on Ginger survived to fire upon U.S. infantry when the latter finally reached and overran the objective. One dug-in al Qaeda command post was found surrounded by no fewer than five JDAM craters, yet its garrison survived and resisted until overrun by U.S. infantry.[84]
This is not to suggest that precision firepower is not extremely lethal, or that even well dug-in al Qaeda defenses did not suffer heavy losses from precision engagements. But the evidence does indicate that a combination of cover and concealment can allow defenders, though battered, to survive modern firepower in sufficient numbers to mount serious resistance.
Nor is this the first time that properly prepared defenses have survived massive firepower, precise or otherwise. French defenses at Verdun in 1916 endured a two-day German artillery barrage equal to about 1,200 tons of explosives – in nuclear parlance more than a kiloton, or more explosive power than the W48 tactical nuclear warhead – yet enough of the entrenched defenders survived this maelstrom to halt the German assault.[85] In 1917, German defenses at Messines absorbed more than a kiloton of explosive power per mile of frontage, yet still halted the ensuing British offensive.[86] At Cassino on March 15, 1944, German positions in the village were struck by 300 tons of bombs in a single day, yet defeated the associated Allied infantry advance.[87] On July 18, 1944, more than 4,500 Allied aircraft, three corps’ worth of artillery, and naval gunfire from two Royal Navy cruisers and the monitor Roberts deposited more than 8,700 tons of explosives – over eight kilotons of firepower – on just seven kilometers of German frontage in less than three hours in Operation Goodwood.[88] Yet the entrenched Germans halted the subsequent British armored advance, killing more than one-third of all the British armor on the continent in the process.[89] Firepower on such scales is tremendously destructive, and each of these defenders suffered heavily under such barrages. But even fantastic volumes of firepower alone cannot annihilate defenses outright. Precision allows crushing firepower to be delivered using vastly fewer platforms, but to expect precision to accomplish what literally nuclear-scale fires have not been able to attain in the past is to ask too much of new technology. The village of Cassino was struck by far less accurate weapons than the al Qaeda defenders of Objective Ginger, but this tiny 500 by 700 meter Italian hamlet was still hit with the equivalent of more than 300 2000-pound JDAMs, which was more than enough to reduce every building in the village to rubble – but not enough to exterminate its defenders.[90] The problem at Verdun, Messines, Cassino, or Operation Goodwood was not any inability to turn defenses into crater fields or reduce specific buildings to rubble without today’s precision – the problem was that resolute defenders can survive even within crater fields and rubble piles to mount serious resistance. Firepower is critical, but against resolute, well-prepared defenders it has rarely been sufficient by itself in the past; taken together, Bai Beche, Qala-e-Gangi, and Operation Anaconda suggest that it may not be now, either.
3. The Role of Indigenous Allies in Afghanistan
The third logical precondition for a broadly applicable Afghan Model is that it be effective without placing heavy demands on indigenous allies’ tactical sophistication. Here, too, experience in Afghanistan varied importantly.
Early in the campaign, it was not uncommon for Taliban defenders to withdraw after SOF-directed standoff bombing destroyed key positions. At Polanyi Canyon, for example, the AMF advance following initial air strikes found only abandoned (or destroyed) Taliban positions and met no residual resistance.[91]
This, too, changed after the early engagements. The remainder of the campaign saw extensive close combat. Moreover, this close combat was neither trivial nor wholly one-sided: many of the outcomes were close calls, involving initial reverses, heavy AMF losses, or both.
At Bai Beche on November 5, for example, the dug-in al Qaeda defenders refused to withdraw in spite of over two days of heavy American air strikes. To dislodge them, Dostum’s AMF cavalry was ordered to charge the position. The first attempt was driven back. The American SOF attached to Dostum’s forces observed this reverse and began calling renewed airstrikes against the al Qaeda positions in anticipation that Dostum would eventually order a second assault. In the process, however, a SOF warning order to the cavalry to prepare for another push was mistaken by the cavalry as a command to launch the assault, with the result that the cavalry began its attack much sooner than intended. The surprised Americans watched the Afghan cavalry break cover and begin their advance just as a series of laser-guided bombs had been released from American aircraft in response to the SOF calls for air support. The SOF commander reported that he was convinced they had just caused a friendly fire incident: the bomb release and the AMF cavalry advance were way too close together for official doctrinal limits, and the air strike would never have been ordered if the SOF had known that the cavalry was then jumping off for the second assault. As it happened, the bombs landed seconds before the cavalry arrived on the position. In fact, the cavalry galloped through the enormous cloud of smoke and dust that was still hanging in the air after the explosions, emerging behind the enemy defenses before their garrison knew what was happening. The defenders, seeing Dostum’s cavalry to their rear, abandoned their positions in an attempt to avoid encirclement.[92]
The result was an important victory – in fact, the victory that turned the tide in the campaign. But the battle involved serious close combat (cavalry overrunning prepared, actively resisting defenses), and the outcome was a very close call. Dostum’s assault profited from an extremely tight integration of movement with suppressive fire – far tighter, in fact, than either Dostum’s troops or their supporting SOF would ever have dared arrange deliberately. Luck thus played an important role in the outcome. Dostum might well have carried the position eventually even without the good fortune of an extraordinary integration of fire and movement; this was clearly a crucial battle and he would presumably have redoubled his efforts if the second attempt had failed. But as fought, the outcome involved an important element of serendipity.
Nor was Bai Beche unique in demanding hard fighting at close quarters. As noted above, al Qaeda counterattackers reached small arms range of U.S. and AMF forces before being driven back at Sayed Slim Kalay and at Highway 4.[93] At Konduz in late November, al Qaeda counterattackers penetrated AMF positions deeply enough to compel supporting American SOF teams to withdraw at least three times to avoid being overrun.[94] In Operation Anaconda, AMF forces associated with the warlord Mohammed Zia and supported by American SOF were assigned to drive al Qaeda defenders from the “Tri-cities” area (the villages of Shirkankeyl, Babakuhl and Marzak); they were instead pinned down under hostile fire from prepared defenses in the surrounding mountainsides and eventually withdrew after they proved unable to advance. Only after the al Qaeda defenders pulled back under joint, multinational attack by allied airpower, U.S. infantry, and multinational SOF were the AMF able to enter the Tri-cities and adjoining ridgelines.[95] At Tora Bora, massive American bombing proved insufficient to compensate for the AMF’s unwillingness to close with dug-in al Qaeda defenders in the cave complexes of the White Mountains; many now see this ground force hesitancy as having allowed Osama bin Laden and much of his command structure to escape capture and flee into neighboring Pakistan.[96]
Among these examples, the fighting along Highway 4 in December is particularly instructive. Shirzai’s Afghans here were divided among two factions. The first, commanded by Haji Gul Alai, were very capable troops by Afghan standards. They used terrain for cover and concealment, maintained good intervals between elements in the advance, moved by alternate bounds, exploited suppressive fire to cover moving elements’ exposure, and were able to exploit the effects of American air strikes by coordinating their movement with the bombing (which many AMF factions could not). The second faction, by contrast, were much less skilled: the attached SOF commander characterized them as “an armed mob – just villagers given weapons.” Their tactics consisted of exposed, bunched-up movement in the open, with no attempt to use the terrain to reduce their exposure, and little ability to employ supporting or suppressive fires. At the Arghestan Bridge on December 5, this second faction launched an assault on a dug-in al Qaeda position south of the Kandahar airport. Driven back repeatedly, they proved unable to take the position, in spite of U.S. air support. Only after these troops were withdrawn and Haji Gul Alai’s forces took over the assault the following day could the al Qaeda positions be taken.[97]
What conclusions should be drawn from this experience?
Will American infantry be needed any time the United States uses force?
Not necessarily: where our Afghan allies met their enemies on something like an equal tactical footing, U.S. air power and small SOF teams made all the difference. With such support, unsophisticated allies overwhelmed unsophisticated opponents at Polanyi Canyon and Tarin Kowt – and better-skilled allies beat comparably skilled, resolute opponents at Bai Beche and Highway 4.
But does this mean that American air power can turn unskilled, unsophisticated allies into conquerors even when facing enemies with superior skill and resolution, as many now suppose?
No.
By and large, the Afghan campaign pitted ground forces of roughly comparable ability and equipment: the net result of years of civil warfare had been a stalemate in which the Taliban had proved unable to dislodge the Northern Alliance from the roughly 15 percent of Afghanistan that it held prior to September 11. But though the typical combat units on each side were thus about equally matched (as the stalled pre-intervention battle lines imply), both armies were actually diverse mixtures of better- and worse-trained, more- and less-motivated troops – and this diversity offers a couple of valuable opportunities to observe instances of unequally-skilled forces in combat. In such unequal fights as the first day at Arghestan Bridge and the assault on the Tri-cities in Anaconda, the results suggest that where our allies are substantially overmatched tactically, American air power and SOF support per se may not be enough to turn the tide. Even where the balance of skills was closer to parity, as at Bai Beche, the results could still be perilously close calls.
If SOF-directed air power had simply annihilated Taliban forces at standoff range, as some now suppose, then even a radically unsophisticated army would have been sufficient to walk in and occupy the blasted ruins. But in Afghanistan, though air power could destroy most of a hostile force, it could not annihilate well-prepared defenses outright. Nor could it defeat well-directed, skillfully concealed assaults by itself. The result was a series of close combat actions, rather than a war fought exclusively at standoff ranges. And in these actions, even a few al Qaeda survivors, properly motivated and armed with modern automatic weapons, could mow down large numbers of unsophisticated indigenous soldiers caught exposed in the open. To survive long enough to take advantage of the tremendous firepower leveled by American air support thus still required the fundamental combat skills of cover, concealment, dispersed small unit maneuver and local suppressive fire. Novice troops without these fundamentals either found themselves pinned down and unable to advance or driven back with heavy losses. Better-skilled troops able to advance under fire, on the other hand, found the precision firepower of American air strikes to be a decisive advantage. And where Afghan Taliban irresolution or exposure left an undefended battlefield following American air strikes, then even unskilled allies did indeed prove able to occupy the abandoned ground.
The results thus suggest that where the troops on the ground are comparable – either both skilled or both unskilled – American precision fires can make the difference. But where an unsophisticated ally is pitted against an enemy with the skills and motivation to survive precision engagement and fight back when attacked, then poorly trained allies will be unable to take advantage of the enormous potential that precision fires bring. Even with precision air support, indigenous allies thus need a combination of skill, motivation, and equipment at least broadly comparable to their enemies’ to prevail.
A New Explanation: The Afghan Campaign as Orthodox Theater Warfare
Neither of the main views in the current debate is thus wholly satisfactory. On the one hand, the campaign’s idiosyncrasies were less decisive than often assumed. On the other hand, the systematic lessons most commonly drawn are at odds with the war’s actual conduct – the fighting did not unfold as the Afghan Model supposes.
A more satisfactory accounting would be to view the campaign neither as a fluke nor as a military revolution, but rather as a surprisingly orthodox example of modern joint theater warfare – albeit one with unusually heavy fire support for ourselves and our allies. While the Afghan campaign was unique in some respects and new in others, it is easy to exaggerate the war’s distinctiveness. There were important continuities between Afghanistan and prior military experience – and these continuities were at least as important for the campaign’s outcome as its differences.
The campaign, after all, centered on a struggle between two sizeable armies. Although public attention has tended to focus on the role of Western air power and the handful of commandos who directed it, the purpose of that Western role was to tip the scales in an ongoing, preexisting land war. The Taliban and Northern Alliance forces whose struggle this air power was meant to influence deployed together some 60-80,000 troops (let alone the Southern Alliance forces that arose as the fighting moved south).[98] These troops were striving to take and hold ground in very orthodox ways prior to the American intervention, and al Qaeda continued to contest control of key geographic points through the end of Operation Anaconda in March.[99] In fact, American commanders were surprised by al Qaeda’s willingness to stand their ground and fight for the Shah-i-kot valley at Anaconda; CJTF Mountain had expected them to melt away in guerilla fashion once struck rather than reinforcing a dug-in territorial defense-in-place as they actually did.[100] The campaign as a whole, moreover, produced extensive close combat, the great majority of which was waged by conventional infantry.
Of course, the outcome of this territorial struggle was affected profoundly by SOF-directed precision air power – this ultimately made the difference between stalemate and victory. But it did not do so by itself. Tightly integrated ground maneuver was necessary to exploit its effects once the Coalition began to encounter chiefly foreign Taliban and al Qaeda opponents with significant skill and motivation. To overcome skilled, resolute opposition required both precision fire and maneuver; neither alone was sufficient in Afghanistan.
This, however, was hardly new.
In fact, the necessary synergy between fires and maneuver has been at the heart of most great power military doctrines since at least 1918. The history of modern tactics is largely a story of the discovery of its importance in the crucible of the Western Front and the subsequent spread of such methods across nations and over time.[101] The German tactical system that first broke the Western trench stalemate in March 1918, for example, was centered on new methods for combining suppressive fire and movement via dispersed formations that used local terrain for cover.[102] Allied armies had learned similar lessons from hard experience in the trenches and adopted similar methods by war’s end (albeit with less complete implementation in the field).[103] The German tactics that wrought blitzkrieg in 1939-41 were partly a matter of exploiting the tank’s new mobility, but largely a product of a heavy emphasis on combining fire and movement in balanced, combined arms formations.[104] Germany’s opponents had allowed their skills in integrating fire and maneuver to atrophy in the interwar years, but under the pressure of hard fighting these skills were gradually recovered, contributing centrally to the end of blitzkrieg after 1941 and the gradual triumph of allied arms by 1945.[105] Israel’s success against poorly-trained Arab infantry in 1956 and 1967 led them to de-emphasize orthodox fire and maneuver in favor of an imbalanced stress on the shock effect of mounted maneuver per se; they paid for this imbalance with their heavy losses against better-prepared Egyptian infantry in the Sinai in 1973.[106] American doctrinal writers initially responded to the Israelis’ 1973 experience by substituting a one-sided emphasis on anti-tank fires for the Israelis’ converse overemphasis on maneuver; by 1982 this imbalance was remedied by a return to orthodox fire-maneuver synergy in the new doctrine of AirLand Battle.[107]
Nor was it new that the fires and the maneuver were provided by different nations in Afghanistan. In Vietnam, our late-war, post-withdrawal policy was precisely to support a South Vietnamese proxy with American air power, much as we did with our AMF proxies in Afghanistan. Of course, the ultimate outcome was much more favorable in Afghanistan. In part this is due to air power’s greater lethality in 2001-2. But the difference is also due in part to the different balance of maneuver skills on the ground in these two conflicts. In Afghanistan, our proxies were generally a match for their enemies tactically, and in key battles like Bai Beche they were able to integrate their maneuver closely with our fires. In Vietnam, by contrast, the southern Army of the Republic of Vietnam (ARVN) was generally less skilled and motivated than its northern foes – and American air support was correspondingly less decisive.[108]
As in Afghanistan, however, there were exceptions to the general rule in Vietnam, too, and these exceptions offer some instructive parallels. In the 1972 Easter Offensive, for example, North Vietnamese attackers who had heretofore waged a mostly guerrilla war were thrown into a conventional, large-unit armored assault. With little experience in such tactics, they had great difficulty combining arms, coordinating attacks, or exploiting terrain for cover. American air power thus found massed, exposed targets in the open and inflicted heavy losses, enabling our ARVN proxies to contain the offensive.[109] By contrast, in the South Vietnamese Lam Son 719 offensive in February 1971 it was the ARVN which found itself in an unaccustomed role, launching its first major airmobile assault without American advisors (who were forbidden to accompany them to their objectives in Laos); the North Vietnamese were here implementing very familiar tactics, waging just the sort of conventional defense they had practiced in more than a decade of continuous warfare against American and French attackers. American air strikes here enabled the ARVN to get as far as the Laotian village of Tchepone, but ARVN maneuver skills left much to be desired and losses were correspondingly heavy; the ensuing ARVN withdrawal left much of their equipment behind and proved so harrowing that no similar offensive was ever again attempted.[110] These actions from the early 1970s were actually much more like Afghanistan than many now suppose: where our Afghan allies met opponents with superior tactical skills, even lavish, radically-precise air support proved insufficient; where they met their enemies with something like equal skills and motivation, American firepower made all the difference.
By contrast, many now believe that in Afghanistan we turned a ragtag militia into conquerors who subsequently overwhelmed a superior enemy by simply walking forward in the wake of our precision bombing. This belief is largely responsible for the general perception of military revolution in Afghanistan – and if the war had really been fought this way, then the perception would be right. But the war was not actually fought this way. And what did happen was much closer to the longstanding historical precedent on the need for integrating fire and maneuver to overcome skilled, resolute opponents. Of course, this is not to suggest that nothing has changed since 1918. In particular, the form that fire support has taken has changed dramatically since then – and the increases in firepower’s range, precision, round-for-round lethality, responsiveness, and flexibility have obviously been dramatic in recent years. In a sense, these changes have been revolutionary if considered chiefly in terms of their effect on the way fires have been provided.
Moreover, the increasing lethality of standoff precision engagement has made the combination of fire and movement much more powerful where both elements are present. Tight integration of JDAMs or laser-guided bombs with skilled ground maneuver is far more effective today than were 77 mm field guns and German stosstruppen in 1918. This is an important development and has greatly increased the real military power of the United States today relative to any plausible foe.[111]
But what new technology has not done is to enable us to beat opponents using either fire or maneuver alone. The Afghan campaign – like the Western Front in World War I, the invasions of France, Poland, or Russia in 1939-41, the Northwest European campaign of 1944-45, the Mideast Wars of 1956-1982, or any of dozens of other, similar examples – was a joint air-land struggle in which the ability to combine fire and maneuver by diverse arms made the difference between success and failure. Seen in this larger context of the underlying determinants of campaign outcomes in modern warfare, this continuity in the need for a demanding system of integrated fire and maneuver is at least as significant for the future of warfare – and for the policy decisions that turn on this – as the accompanying changes in the technology of precision firepower.
Stephen Biddle
Strategic Studies Institute
U.S. Army War College
122 Forbes Avenue
Carlisle, PA 17013-5244
717-245-4126 voice 717-245-3820 fax
[email protected]
20 October 2002
Wartime experience always shapes the postwar policy debate, and the campaign in Afghanistan has had an unusually prompt effect. The global war on terrorism is barely joined. Sporadic fighting is still ongoing in Afghanistan itself. Yet the apparent success of American arms there last fall and winter is already playing a major role in a series of debates over the future of the American military, the conduct of the continuing war, and even the shape of American foreign policy as a whole.
In particular, many now argue that in 2001-2, a novel combination of special operations forces (SOF), precision-guided munitions (PGMs), and an indigenous ally destroyed the Taliban’s military, toppled their regime, and did so while neither exposing Americans to the risk of heavy casualties nor expanding the American presence in a way that might spur nationalist insurgency. In this new “Afghan Model,” it is argued, small teams of elite commandos on the ground provided the targeting information needed for precision weapons to reach dispersed, concealed opponents. Until the commandos arrived, high-altitude bombing could do little against a country with few large, fixed targets. But once Coalition bombing was guided by friendly eyes on the ground, many claim, it became possible for airpower to annihilate the Taliban infantry and armor that had stymied the Northern Alliance for the preceding six years of civil warfare, enabling even an unsophisticated, outnumbered ally to liberate the entire country in a matter of weeks.[1]
Many now believe that this Afghan Model could be used elsewhere with comparable effect. It has been widely reported that senior civilian Defense officials advocate its use against Iraq.[2] Many see it more broadly still – as a new “American way of war” applicable across a wide range of future conflict types. After all, almost any likely opponent has local enemies who could serve as indigenous allies, either domestically or in neighboring states threatened by our enemies’ ambitions – and we can employ special forces and precision weaponry almost anywhere. If this combination works elsewhere with anything like its effectiveness in Afghanistan, then it might enable us to win the wars of the future with minimal U.S. casualty exposure and political footprint.[3] If so, the policy implications could be sweeping. It might, for example, make sense to restructure the U.S. military around the requirements of this new way of war, with larger special operations forces, modernized, longer-range aircraft and missiles, deep reductions in conventional Army and Marine Corps ground forces, major reallocations of Service roles and missions, and sweeping changes in joint military doctrine.[4] Modernization programs for systems originally designed to wage large-scale theater warfare, such as the F-22 Raptor air superiority fighter, the Comanche armed reconnaissance helicopter, or the V22 tilt rotor transport aircraft could all be challenged as backward-looking failures to reflect the new possibilities.[5] The case for expanding the ongoing war on terrorism to topple regimes in Iraq or elsewhere would be strengthened if the Afghan Model were seen as a cheap, effective way to leverage local allies in pursuit of American aims.[6] Most broadly, a powerful, widely-applicable means of waging war at global distances with minor risks would add major impetus to proposals for neo-imperialist conceptions of American foreign policy, in which unilateral U.S. military might underwrites a Pax Americana and advances American values and interests around the world.[7]
Others disagree.
Many argue that the Afghan Model’s success in 2001-2 was idiosyncratic – a product of unique local circumstances unlikely to recur in future conflicts. [8] Any real world outcome is a product of many contributing causes, some of which are replicable and others not; critics see non-replicable features of the Afghan theater or the Taliban opponent as necessary preconditions for the outcome observed there. As Secretary of Defense Donald Rumsfeld put it: “It is hard for me to imagine another Afghanistan. If you think about that situation, it is kind of distinctive. Now it doesn’t mean that some of the things that are working there won’t work elsewhere, but the totality of it is distinctive. I don’t think we’re going to run around with a cookie mold and repeat this.”[9] A related critique rests on the politico-military disadvantages of relying so heavily on local proxies who may not share our aims. Many such analysts point to the fighting at Tora Bora, where Afghan allies with different interests than ours apparently failed to press the attack and may have allowed Osama bin Laden to escape. To secure U.S. interests, they argue, will often require that U.S., not foreign, troops do the heavy lifting.[10]
If so – if Afghanistan was unique or if reliance on proxies is politically unsound – then the policy implications are very different than those seen by Afghan Model proponents. In particular, the case for expanding the war to Iraq is weakened (as this could require major U.S. ground forces rather than the Afghan Model’s small commando teams), the rationale for a sweeping redesign of the U.S. military loses its most trenchant empirical support, and the military demands of a neo-imperialist foreign policy become much more burdensome.
To date, however, this debate has been conducted on the basis of only the most preliminary impressions of the war’s actual conduct. The purpose of this paper is thus to provide a more extensive body of evidence on the way the Model’s elements were actually employed in Afghanistan, the nature and behavior of its targets, and the circumstances of the fighting – and to use this evidence to assess the Model’s role in the Taliban’s collapse on a more systematic basis.
The results suggest that none of the main current views offer a sound account of the war’s actual conduct. The Afghan Model did not in fact work as its proponents now suggest; nor was its success due chiefly to unique properties of Afghanistan or the Taliban.
To account for what is now known thus requires a different explanation of the campaign’s outcome – one resting neither on precision firepower per se nor on Afghan idiosyncrasies.[11] I argue below that the best explanation is that the Afghan campaign was actually far less different or unusual than most now suppose: it was a surprisingly orthodox air-ground theater campaign in which heavy fire support decided a contest between two land forces. Of course, some elements were quite new: the fire support came almost exclusively from the air; the air strikes were directed mostly by commandos whose methods, equipment, and centrality to the outcome were unprecedented; and the ground armies were mostly not countrymen of the commandos and air forces who provided the firepower.[12] In an important sense, though, the differences were less salient than the continuities: the key to success in Afghanistan as in traditional joint warfare was the close interaction of fire and maneuver, neither of which was sufficient alone and neither of which could succeed without significant ground forces trained and equipped at least as well as their opponents. In Afghanistan, our allies provided these ground forces; where others can do so, the Afghan Model can be expected to prevail. Hence Afghanistan is not unique. But not all future allies have armies trained and equipped to their enemies’ standards. Without this, neither the bravery of our special operations forces nor the sophistication of our PGMs can ensure a quick Afghan-like collapse of a resolute opponent.
This in turn implies that we should neither restructure our military to wage Afghan-style wars more efficiently, nor reflexively commit conventional US ground forces in every conflict. Where our allies’ ground forces are good enough, the Afghan Model will work and we should use it. But we will not always be so lucky. In Iraq, for example, the lack of a credible, trained opposition bodes ill for an Afghan-style campaign without major American ground forces. Deep cuts in US ground capability could thus be very risky in spite of our strengths in air power and special operations forces. More broadly, though, we should be wary of suggestions that precision weapons have so revolutionized warfare that traditional ground forces or close combat are now superceded. Where our allies are able, they may provide most of the ground troops for us, but what Afghanistan really shows is that the wars of tomorrow – like those of yesterday – will require tight integration of fires and ground maneuver at close quarters to exploit technology’s effects. Precision weapons make this combination more powerful, but against resolute opponents, neither air power nor conventional ground forces alone will suffice any time soon.
I base these findings on a new collection of primary source evidence centered on a series of 46 interviews with key American participants in the conflict, ranging from Special Forces Sergeants to the Major General who commanded CJTF Mountain during Operation Anaconda, and including subjects from the Special Operations Command, the U.S. Army, the U.S. Air Force, and the Central Intelligence Agency.[13] These interviews were complemented with official written documentation on the conduct of the war and direct physical inspection of the Anaconda battlefield in Afghanistan’s Shah-i-kot valley, together with available secondary source accounts, chiefly from the print news media. This body of evidence cannot be considered complete; a definitive history of the Afghan campaign would require years of research on a much broader range of issues. Rather, my intention here is to focus on one key issue – the new Model’s role in the Afghan campaign and its implications for the future – and to muster as much evidence as can be produced in the near term, so as to make initial findings available sooner than a definitive history would permit, but with a stronger foundation in the evidence than the debate to date has offered.
I present the analysis in five steps. First I outline the key events in the Afghan campaign. Next I assess in turn the two main current explanations of the Taliban’s fall: first, the school that sees the causes as local and idiosyncratic; second, the school that situates them in replicable features of the Afghan Model.[14] I then present an alternative explanation, and show how it accounts more successfully for the available evidence. Finally, I discuss the resulting implications for the Army and for defense policy more broadly.
Key Events
The bombing campaign began the night of October 7, 2001, and focused initially on destroying the Taliban’s limited air defense and communications infrastructure. American and British SOF teams had been conducting scouting missions in Afghanistan beginning at least a week prior to the first air strikes; by October 15, teams designated to make contact with the major Northern Alliance warlords had been inserted and begun preparations for combined offensive action against the Taliban.[15]
[Figure 1, map graphic, about here]
Some of the first major combat actions occurred in the mountains south of Mazar-e-Sharif, as SOF teams working with Northern Alliance Generals Abdul Rashid Dostum and Atta Mohammed fought their way north toward Mazar up the Dar-ye Suf and Balkh river valleys. The terrain and conditions here were extraordinary: at elevations of up to 6,400 feet, movement was restricted to winding mountain trails in which sheer rock faces were sometimes separated from thousand-foot drops by no more than a three-foot width of rocky path. With no vehicles able to negotiate such trails, commandos hauling loads of over 40 pounds of equipment per man were given Afghan mountain ponies with wooden saddles and told to ride along with Dostum’s troops. Luckily, the SOF team commander assigned to Dostum had been a high school rodeo rider in Kansas, but none of the other Americans had ever been on a horse before. Their knees in their chests, balancing heavy rucksacks on their backs, they were instructed by their commander to keep their downhill foot out of the stirrups and to lean uphill so if the pony lost its balance they would fall onto the trail as the pony went into the gorge. On particularly rocky stretches the team commander ordered his men to travel with weapons out and a round chambered to shoot immediately any pony that bolted before it could drag its rider to his death over the rocks.[16]
The first combat action in this sector came when Dostum’s group took the village of Bishqab on October 21; this was followed by engagements at Cobaki on October 22, Chapchal on October 23, and Oimetan on October 25. The key battle came when Dostum’s troops encountered hostile forces occupying old Soviet-built defensive positions at the hamlet of Bai Beche some 16 kilometers southeast of Keshendeh-ye Pa’in.[17] On November 5, Dostum’s cavalry overran these defenses; shortly thereafter Atta’s forces captured Ac’capruk on the Balkh river, and the door swung open for a rapid advance to Mazar, which fell to Atta and Dostum’s forces on November 10.[18]
The fall of Mazar unhinged the Taliban position in northern Afghanistan. Taliban defenders near Bamiyan resisted briefly before surrendering the city on November 11; Kabul fell without a fight on November 13. A force of some 5,000 Taliban and al Qaeda survivors were then encircled in the city of Konduz, where they surrendered following a 12-day siege on November 26.[19]
Many of the prisoners taken at Mazar were subsequently detained at the 19th century mud fortress of Qala-e-Gangi west of the city. On November 25, a revolt at the prison killed one American and gave rise to a two-day struggle before control of the prison was reestablished on the 27th.[20]
With the fall of Kabul and Konduz, attention shifted to the Taliban’s stronghold of Kandahar in the south. SOF teams in support of Hamid Karzai’s forces advanced on the city from the north; teams in support of Gul Agha Shirzai advanced from the south. The result was a series of battles at Tarin Kowt and Sayed Slim Kalay north of the city on November 18 and December 2-4, respectively, and along Highway 4 south of Kandahar from December 2-6. On the night of December 6, Mullah Omar and the senior Taliban leadership fled the city and went into hiding, ending Taliban rule in Afghanistan.[21] Allied forces subsequently tracked a group of al Qaeda survivors thought to include Osama bin Laden to a series of redoubts in the White Mountains near Tora Bora. The redoubts were taken in a 16-day battle ending on December 17, but many al Qaeda defenders escaped death or capture and fled across the border into Pakistan.[22]
In March, a second concentration of al Qaeda holdouts was identified in the Shah-i-kot valley and surrounding mountains east of Gardez. In Operation Anaconda, a combined offensive by two battalions of U.S. regular infantry from the 101st Airborne and 10th Mountain divisions, supported by allied Afghan and Western troops and special operations forces from seven nations, descended on the al Qaeda defenders, killing many, dispersing the rest, and bringing to a close the major combat operations in the country as of this writing. [23]
Existing Views: The Taliban’s Fall as Idiosyncratic
Were these results the products of local idiosyncrasies? Afghanistan is certainly an unusual place, and the Taliban were unusual opponents.[24] Several of their peculiarities might in principle have been necessary preconditions for the Afghan Model’s success, including:
• Poor enemy morale or motivation;
• Poor enemy military training and expertise;
• Lack of popular support for the Taliban regime;
• The ease of defection in Afghan culture;
• Surprise;
• Taliban dependency on fragile sources of outside support; or
• Availability of contiguous, secure territory for re-supplying and re-equipping proxy forces. •
Yet on closer inspection, the evidence suggests a weaker causal role for these idiosyncrasies than some now suppose.
Atypically poor enemy morale
The native Afghan Taliban, for example, were fighting only to prevent the extradition of a foreigner, Osama bin Laden, whom many of them resented anyway. They might be expected to display only limited commitment to such modest stakes, and hence poor morale and combat motivation would hardly be surprising. Future opponents, by contrast, for whom much more could easily be at stake, might be better motivated. If poor enemy morale were a necessary precondition for the Afghan Model’s success, then the Model might not succeed elsewhere. [25]
Yet much of the actual fighting was shouldered not by Afghan Taliban but by foreigners, and especially by al Qaeda – and these foreign troops were much better motivated than their Afghan comrades. Our opponents in this campaign were not a monolithic or homogeneous military. Instead, their three major components (the indigenous Afghan Taliban, the predominantly foreign al Qaeda, and other, non-al Qaeda foreign allies of the Taliban) differed in many important respects.[26] Among the most important of these differences was combat motivation: whereas the indigenous Afghan Taliban quit the field quickly, the foreigners did not. On the contrary, there are numerous reports of foreign fighters threatening to kill any Afghan Taliban who fled the front lines.[27] At Bai Beche, Konduz, Sayed Slim Kalay, Highway 4, and elsewhere in the theater, al Qaeda fighters launched counterattacks to try to close with American and allied forces.[28] In Operation Anaconda, al Qaeda defenders not only stood their ground against overwhelming American firepower, they actually reinforced their positions in the midst of the battle: their fighters were willing to advance into the teeth of a fierce bombardment to enter the Shah-i-kot Valley from safer positions elsewhere and seek battle with our forces.[29] These are not the actions of unmotivated or unwilling troops. The non-Afghan Taliban have shown themselves to be resolute combatants with the motivation to stay in the field when there seemed to be something to be gained militarily from doing so.
These better-motivated foreigners were responsible for much of the fighting in the campaign’s key actions. Though comprising under 25 percent of the regime’s overall troop strength, the hard-core foreigners were allocated disproportionately to the critical sectors of the front and bore an increasing fraction of the combat load beginning in late October.[30] By early-to-mid November, American special forces teams were reporting predominantly al Qaeda, rather than Afghan, opposition in the critical battles for Mazar-e-Sharif and Kandahar.[31] The Afghan Taliban’s tendency to quit the field following their initial defeats doubtless evened the numerical odds for the decisive battles in November and December, but the key engagements at Bai Beche, Sayed Slim Kalay, and Highway 4 were all fought against motivated, resolute opponents – as were the later actions at Tora Bora and Anaconda. If poor enemy morale were a necessary precondition for the Afghan Model to succeed, then the Model should thus have failed after its initial victories in October, yet it did not.
Atypically poor enemy training
The Afghan Taliban were often very poorly trained soldiers. Many had little or no formal military instruction, and Afghan ranks swelled and shrank with the seasons and the fortunes of war as troops went home to their villages or took up arms depending on the crop cycle and apparent military need.[32] In fact, former Supreme Allied Commander in Europe Gen. Wesley Clark has referred to them as “the least competent adversary we’ve faced since the Barbary Pirates.”[33] Other enemies might well be better trained; if the Afghans’ poor skills were a necessary precondition for their defeat, then their collapse might be hard to replicate elsewhere.
Yet here, too, the distinction between Afghan and foreign Taliban is important. The al Qaeda soldiers in the key battles were much better trained than the Afghan Taliban. Unlike the latter, which were often closer to a civilian militia than a professional army, the al Qaeda troops were committed, full-time soldiers. Osama bin Laden’s infamous training camps served primarily to prepare these troops for combat on Afghanistan’s front lines, and taught them a curriculum not radically different from orthodox Western armies’.[34] While the result may not always have been up to contemporary Western standards, there is no reason to suppose that al Qaeda’s skills or training were not at least comparable to those of most other likely opponents, few of whom are trained to Western standards either.[35] The Afghan Model’s ability to prevail against this better-trained opposition suggests that atypically poor enemy skill was not a necessary precondition for success.
Atypical lack of popular support for the enemy’s war effort
Many Afghans hated the Taliban and chafed under the harshness of their regime.[36] This surely weakened their Afghan soldiers’ morale, and presumably reduced the material and intelligence assistance a state might otherwise expect from its citizens. Enemies with stronger popular political support could well be much harder to defeat.
Yet the Taliban’s unpopularity was hardly unique. Many other potential opponents’ regimes are at least as unpopular: Saddam, for example, hardly governs by the will of the governed. If unpopularity makes a military unable to resist American arms, then this hardly makes Afghanistan unrepresentative of the future.[37]
Nor were the Taliban uniformly hated. In southern Afghanistan, their traditional stronghold, they remained popular throughout the war and indeed to this day. In fact, in November the conventional expectation was that Kandahar and the south would prove much harder to conquer than Mazar and the north, given the former’s greater sympathy for Mullah Omar and his regime. [38] These concerns proved unfounded, but not because the southern public failed to support the government. Moreover, the foreign Taliban were apparently undeterred by public opinion in Afghanistan – they were waging a jihad whose legitimacy in their minds was unrelated to the political views of Afghans, and they fought hard regardless of the latter. On balance, there is little reason to assume that the Taliban’s degree of popular support was either unique or an essential precondition for their collapse.
Atypical enemy willingness to defect
The acceptability of defection in Afghan military culture is certainly unusual to Western eyes. Not only do many Afghans consider it normal and acceptable to switch sides in war, but most of their senior commanders know one another personally from their service against the Soviets. Hence it is probably easier to persuade an opponent to defect in Afghanistan than in much of the rest of the world.[39] And there were certainly many defections in fall and winter of 2001-2. If the Taliban’s loss of combat power resulting from these defections were a necessary precondition for our success, then the 2001-2 outcome might well be an anomaly.
Yet the overwhelming majority of defections in this campaign occurred after the military tide had turned, not before. The largest mass turnovers came with the fall of the northern cities of Mazar-e-Sharif and Konduz on November 10 and 26, respectively.[40] The key battle in the north, however, was east of Keshende-ye Pa’in near Bai Beche on November 5. Prior to this, the Northern Alliance and its accompanying SOF teams had been forced to work their way methodically up the Dar-ye Suf and Balkh river valleys, fighting their way from one Taliban defense line to the next. The Taliban, however, had chosen to make a stand at Bai Beche, pushing all available reserves forward to reinforce these positions. Once Dostum broke through their lines, the Taliban thus had nothing left to back them up and the result was a rout: “the dam broke,” as one Special Forces soldier described it, with the Taliban retreating in disorder and unable to rally or establish new blocking positions.[41] The Taliban’s consequent loss of the Dar-ye Suf corridor made Mazar-e-Sharif untenable, and when it fell their entire position in northern Afghanistan collapsed.[42] By the time the major defections occurred in mid-to-late November, the military outcome in the north had thus been determined. By contrast, before the battle at Bai Beche, the SOF commanders operating in the north reported few if any defections to the forces they served with.[43] Moreover, even after Bai Beche, few al Qaeda voluntarily defected, and in fact they continued to resist stubbornly around Kandahar in the south through December 6, and fought hard at Tora Bora and Anaconda well into 2002.[44] Some were captured when overrun or encircled, and many tried to blend into the population and slip away once their defenses had been breached, but most of the voluntary defections were by Afghan Taliban – and even for them, most stayed and fought at least until the military tide had turned. Defections in the Afghan campaign were thus important, but they were effects not causes; consequences of military failure not determinants of it.
Atypical susceptibility to surprise
The Afghan Model surely surprised the Taliban. Though neither SOF nor PGMs were new, they had never been combined in quite this way before, presenting a novel problem to defenders.[45] Other potential opponents, however, have now seen the Model in use and are surely studying means of thwarting it. The Taliban’s ignorance of the Afghan Model is thus probably unique to Afghanistan – our next opponent will not be caught as unsuspecting as the Taliban were in October.
Yet while the Taliban were initially surprised, they adapted quickly. As early as October 22, American SOF teams were reporting that Taliban vehicles in their sectors had been smeared with mud to camouflage them.[46] By Bai Beche they were already making aggressive use of overhead cover and deliberate concealment.[47] In the fighting north of Kandahar and along Highway 4 in December, al Qaeda defenses were well-camouflaged, dispersed, and making use of natural terrain for expedient cover.[48] This pattern continued through Operation Anaconda in March, by which time al Qaeda forces were practicing systematic communications security, dispersal, camouflage discipline, use of cover and concealment, and exploitation of dummy fighting positions to draw fire and attention from their real dispositions.[49]
Others may adopt different countermeasures, but those the Taliban turned to by early November were the natural choices one could expect an adaptive opponent to try to implement when faced with capabilities like the Afghan Model’s. Again and again, from Western Front armies’ response to artillery and machine guns in 1916-18, to the German Wehrmacht’s response to Allied air supremacy in northwest Europe in 1944, to the Vietcong and North Vietnamese armies’ response to U.S. firepower in Vietnam, armies facing overwhelming firepower have dispersed, exploited cover and concealment, restricted their radio transmissions, and adopted camouflage.50 In fact, this pattern parallels the approach suggested by several U.S. marine corps officers assigned to oppose an operational concept resembling the Afghan Model in a pre-911 exercise.[51]
Much of the actual fighting in Afghanistan thus came against an opponent who had already begun to adopt the standard countermeasures that one would expect to see against the kind of firepower we employed there; the effects of surprise were attenuated rather quickly in this campaign. Yet the loss of surprise did not prevent us from continuing to prevail on the battlefield. Surprise per se thus does not appear to have been a necessary precondition for success in Afghanistan.
Atypical enemy dependence on vulnerable sources of outside support
The Taliban had certainly enjoyed important outside support. They were largely a creation of the Pakistani intelligence service, the ISI, which continued to supply them with materiel, tactical advice, and intelligence throughout the civil war. Shortly after September 11, Pakistani President Pervez Musharaff threw his support to the United States and pledged to end ISI assistance to the Taliban, who could thus have expected to lose a major source of outside assistance. Relative to other potential opponents, the Taliban might thus be considered unusually vulnerable to such a cutoff.
Yet aid withdrawal was probably not decisive. The key battles were fought in early November and December. Pakistani aid continued to reach the Taliban until at least October 12, and may have continued even after that.[52] Most militaries stockpile supplies against the possibility of interdiction, hence supply cutoffs typically take time to affect battlefield outcomes. Even if all outside sources had dried up instantly on October 12, it is unlikely that such a short period of isolation could have induced catastrophic military consequences for the Taliban. And in fact, American inspection of captured Taliban positions indicates that they remained well-stocked (especially with ammunition) through at least March 2002.[53] Nor does the withdrawal of Pakistani intelligence or tactical advice appear to have had decisive effects. Sympathetic civilians have regularly provided al Qaeda with information on Allied movements with or without help from the ISI; this problem was aggravated with the return of Afghan Taliban fighters to their home villages following their initial defeats, placing thousands of still-hostile observers in villages across the country.[54] Enemy tactics, moreover, typically improved, not decayed, as the war has progressed and more sophisticated al Qaeda fighters replaced less-skilled Afghans – withdrawal of ISI tactical assistance hardly undid the Taliban tactically.[55] In all, there is little basis to see the Pakistani aid cutoff as a necessary precondition for success.
Atypical availability of safe havens for training and equipping indigenous allies
The Northern Alliance controlled an unusually large contiguous territory in which they could be re-supplied, retrained, and reequipped: perhaps 10-15 percent of Afghanistan’s total land area was under their control prior to American intervention.[56] Other potential allies might not enjoy such extensive or contiguous safe havens. If re-supply, retraining, and re-equipment were necessary preconditions for the Afghan Model’s success, then this territorial advantage might make the 2001 outcome idiosyncratic.
Yet here, too, the tide turned before its effects could become decisive. Most of the key SOF teams were calling air strikes and supporting Northern Alliance offensives within hours of arriving in Afghanistan – the fighting was well under way before the Americans could do any significant retraining, and before any major re-supply could be accomplished.[57] Once the battle was joined, moreover, the high tempo of operations made it impossible for American SOF to pause long enough to conduct training. Some food and other humanitarian aid arrived quickly, but few weapons and little ammunition were dropped prior to the critical battle at Bai Beche: Dostum’s first significant lethal aid delivery was not until 2:00 a.m. on November 5, that is, the predawn hours of the day of the battle itself. And even then, this aid was limited to small arms, ammunition, and rocket propelled grenades – no heavy weapons or armored vehicles were provided.[58] Most of the key battles were thus fought by Afghan allies with skills and equipment little different from what they had in the civil war; Afghanistan’s suitability for retraining and re-equipping played little role in the outcome.
Existing Views: Afghan Model Lethality
There is thus substantial evidence that the campaign was not merely a fluke, hence its implications may have some real meaning for the future. But are those implications necessarily the ones that Afghan Model proponents now see? In fact, they are not – because the Model did not actually work in Afghanistan the way most now think it did. And its actual performance implies a less radical departure from past military practice than many Afghan Model proponents argue.
For the Afghan Model to be a widely applicable template suitable for restructuring the U.S. military, several key features must have obtained in Afghanistan:
1. SOF and standoff sensors must have been able to find key targets for precision engagement;
2. PGMs must have been able to kill the targets found, at standoff ranges;
3. The indigenous allies’ role must have been undemanding.
The first two conditions are logically necessary for the efficacy of the PGMs around which the Model is built – strike systems must be able to find and kill the critical targets at long range, before enemies can close with and overwhelm the small SOF teams on the ground. The third, however, is at least as important. We can almost always find local allies, but armies vary widely in skill, equipment, and motivation; some will be capable of demanding tasks, but others not. Much of the Model’s apparent power in Afghanistan was due to the perception that our local allies there were not terribly capable; if SOF and PGMs can make conquerors out of such modest material then they can do so in many other places, it is often supposed. In particular, the key roles implied for the allies are to screen U.S. SOF from hostile patrols or infiltrators too small to make good targets for PGM attack; to mop up surviving remnants; and to occupy abandoned territory (it is often politically important that local forces be the ones to take control of the ground that the PGMs effectively clear). If the allies’ mission becomes much more demanding than this, then both the generality and the novelty of the Afghan Model come into question – after all, allies like the Israelis or the British have always been capable of impressive feats with or without American PGMs; the claim that the Afghan Model is a revolutionary departure rests on the perception that it turned an overmatched militia into a juggernaut.
In Afghanistan, however, these three conditions were not always met. SOF and standoff sensors found many targets, but could not ensure acquisition of properly concealed opponents. PGMs were very lethal, but against well-entrenched positions they could not destroy enough to prevent survivors from halting unsophisticated attackers. Our indigenous allies’ role was undemanding at times, but far more challenging at others. In particular, significant numbers of properly prepared al Qaeda positions eluded SOF and remote target acquisition and survived even extensive air strikes. To take such positions required orthodox combined arms, fire-and-maneuver tactics – even with the intense precision fire support available in this campaign. Such tactics are difficult to execute, and require skills and leadership many potential indigenous allies lack; in fact, not all our Afghan allies in 2001-2 were up to the job. Where the available ground forces faced skilled enemies but had the necessary skills themselves, American firepower brought decisive victory. But where our ground elements lacked the needed skills or motivation, offensives against capable enemies stalled and outcomes were frequently in doubt in spite of American air support. In Afghanistan, friendly ground forces were mostly as good as their enemies, hence our theater offensive as a whole overwhelmed the Taliban and drove their regime from power. But the variations in tactical behavior in this campaign produced important variations in local outcomes – and these differences teach important lessons about how future wars with different balances of ground force skill and motivation might turn out.[59] These lessons in turn imply a different understanding of cause and effect in Afghanistan, a different model for the future of warfare, and different implications for Army and defense policy.
To show why, I address each of the three key conditions in turn.
1. Target Acquisition in Afghanistan
In the campaign’s early stages, enemy targets were exposed and relatively easy to acquire. The Afghan Taliban commonly deployed on ridge crests, silhouetting themselves for SOF observation. Afghan defenders made little or no effort to camouflage or conceal their positions. They engaged in casual, exposed movement in the target area, often emerging to inspect bomb craters while still under surveillance. Armored vehicles and crew-served weapons were deployed without revetments or apparent efforts at cover or concealment. As a result, their positions were easily identified for PGM targeting, sometimes from extraordinary distances.
At Bishqab on October 21, for example, US SOF observers acquired Taliban targets at ranges of 8-10 kilometers.[60] At Cobaki on October 22, Taliban observation posts were easily spotted at 1500-2000 meters.[61] At Zard Kammar on October 28, Taliban defensive positions were visible from a mile away.[62] At Ac’capruk on November 4, exposed Taliban combat vehicles and crew-served weapons on hillsides west of the Balkh river could be identified from SOF observation posts on the Koh-i-Almortak ridge line some 4-5 kilometers distant.[63] At Polanyi Canyon near Bamiyan on November 7, Taliban positions were plainly visible from more than four kilometers away.[64] Taliban counterattacks, as at Tarin Kowt on November 18, were conducted in the open with no evident attempt at concealment, dispersal, or covering fire to keep U.S. or allied observers’ heads down or to complicate target acquisition.[65]
This changed as the campaign continued and as the opposition shifted increasingly from Afghan Taliban to al Qaeda. In the November 5 battle at Bai Beche, for example, al Qaeda’s defensive works greatly complicated American target acquisition. While American SOF observers knew they faced an interconnected trench system with the enemy somewhere within it, and while particular strongpoints could be identified for engagement, al Qaeda’s local cover and concealment were good enough to prevent the SOF from locating the entirety of the enemy’s individual fighting positions, many of which could not be singled out for precision attack.[66]
By the December fighting along Highway 4 south of Kandahar, not even this was available. In fact, concealed al Qaeda defensive positions among a series of culverts and in burned-out vehicle hulks alongside the road remained wholly undetected by any friendly element until their fire drove back an AMF (Afghan Military Forces – our Northern and Southern Alliance allies) advance. An al Qaeda counterattack in the same sector using a system of wadis for cover approached undetected to within 100-200 meters of AMF and American SOF positions along the highway before opening fire on friendly forces.[67]
At the village of Sayed Slim Kalay north of Kandahar between December 2-4, concealed al Qaeda defenders likewise remained undetected until they fired upon unsuspecting U.S. and AMF attackers. An al Qaeda counterattack using local terrain for cover maneuvered into small arms range of friendly defenders before being driven back.[68]
At Operation Anaconda in March 2002, an intensive pre-battle reconnaissance effort focused every available surveillance and target acquisition system on a tiny, ten-by-ten kilometer battlefield. Yet fewer than 50 percent of all the al Qaeda positions ultimately identified in the course of the fighting on this battlefield were discovered prior to ground contact. In fact, most fire received by U.S. forces in Anaconda came from initially unseen, unanticipated al Qaeda fighting positions.[69]
How could such things happen in an era of persistent reconnaissance drones, airborne radars, satellite surveillance, thermal imaging, and hypersensitive electronic eavesdropping equipment? The answer is that the earth’s surface remains an extremely complex environment with an abundance of natural and manmade cover and concealment available for those militaries capable of exploiting it.
Figure 2 provides a concrete illustration of this problem in the form of a photograph of an al Qaeda fighting position from Objective Ginger on the Anaconda battlefield.[70] The yellow arrow indicates the al Qaeda defenders’ location; without the arrow, there would be no visible sign of a combat position even from the nearly point blank range at which this photograph was taken. Overhanging rock in turn provides cover and concealment from overhead surveillance systems. In principle one might hope to observe resupply movement or al Qaeda patrols into or out of such positions, or to overhear radio communications from its occupants. Al Qaeda soldiers wearing the flowing robes of local herdsmen and traveling in small parties among the mountains, however, are nearly impossible to distinguish at a distance from the noncombatants who tend goats or travel through such areas as a matter of routine.71 And defenders able to operate under radio listening silence while communicating using runners, landlines or other non-broadcast means can reduce signals intercepts to a level that makes identification of specific fighting positions very problematic. Against such targets, it is far from clear that any surveillance technology coming any time soon will ensure reliable target acquisition from standoff distances.
[Figure 2, Objective Ginger fighting position, about here]
Nor are positions such as this one rare anomalies or atypical of Afghan terrain more generally. Figures 3 and 4 show broader samples of the Shah-i-kot battlefield on which Anaconda was fought, including the features known as “The Whale” (after a similar rock formation at the U.S. National Training Center at Ft. Irwin California) and Objective Ginger, respectively.[72] Almost any of the dozens of shadows, crevices, or folds in the earth scattered across these landscapes could house positions like the one shown in Figure 2. And this is just a tiny subset of even the Anaconda battlefield, which is itself a tiny subset of Afghanistan as a whole. The natural complexity of such surfaces offers any adaptive opponent with the necessary training and skills a multitude of opportunities to thwart even modern remote surveillance systems. Against such opponents, remote surveillance will still detect some targets, and remote sensors remain crucial assets, but the only sure means of target acquisition is direct ground contact: a ground force whose advance threatens objectives that the enemy cannot sacrifice and thus must defend compels them to give away their locations by firing on their attackers. Skilled attackers can eventually locate any defensive position by observing the source of the fire directed at them – and this, in fact, is how the majority of the al Qaeda positions at Anaconda were found.
[Figures 3 and 4, The Whale, Shah-i-Kot Valley; and Objective Ginger, Shah-i-kot Valley, about here]
Nor is this problem unique to Afghanistan or its mountainous terrain. Militarily exploitable cover is commonplace in almost all likely future theaters of war. For targets who observe radio listening silence, as al Qaeda now does, foliage, for example, degrades all current remote sensor technologies; urban areas provide overhead cover, create background clutter, and pose difficult problems of distinguishing military targets from innocent civilians.[73] Each is widely available. More than 26 percent of Somalia’s land area is wooded or urban, as is more than 20 percent of the Sudan’s, 34 percent of Georgia’s, or 46 percent of the Philippines’.[74] This cover, moreover, is often distributed in small, widespread patches. On the Goodwood battlefield of 1944 in Normandy, for example, over 80 percent of all one-kilometer grid squares now contain at least some forest or urban cover (though only 26 percent of the total land area is covered).75 In most countries, the central geostrategic objectives are urban areas; even where the bulk of the national land area is open desert (as in Iraq), the cities are both the key terrain and an ample source of cover from overhead sensors (Baghdad alone covers more than 300 square kilometers).[76] Among the most important themes in the history of modern tactics is the growth of methods for exploiting such cover to reduce vulnerability to modern firepower – the “empty battlefield” that has characterized the modern era is a product of skilled armies’ ability to find cover sufficient to thwart standoff target acquisition and to exploit this cover to perform meaningful military missions without excessive exposure.[77]
This is not to suggest that modern sensors are useless or that anyone anywhere can fight effectively from cover opaque to standoff observation. To find and exploit cover while taking or holding ground requires a very demanding set of tactical skills. Historically, armies have varied widely in their ability to do this, and in fact the Afghan campaign itself displays substantial variation in the Taliban’s ability to do so: the Afghan Taliban early in the fighting were systematically unable to do so; only the foreign Taliban and al Qaeda encountered later in the campaign proved able to contest territory from covered, concealed positions. The less skilled the opponent, the more exposed they will be to remote target acquisition. And no army can disappear utterly: diligent reconnaissance will always uncover part of an enemy’s dispositions; the better the sensors, the more they will find, and today’s sensors can find enough to be a crucial contributor to success in theaters like Afghanistan. But this is not to say that they can find enough – on their own – to break a skilled, resolute opponent by standoff fires alone. Even today’s best sensors are still far from an ability to acquire most or all of a hostile force that has learned to exploit the natural complexity of the earth’s surface for cover and concealment – as our experience against al Qaeda in 2001-2 demonstrates.
2. PGM Lethality in Afghanistan
The second logical requirement for Afghan Model effectiveness is PGM lethality: standoff weaponry must be able to kill the targets acquired. As with target acquisition, however, lethality varied widely in Afghanistan as a function of the targets’ tactics. Where the Taliban presented exposed or massed targets in the open, PGMs were extremely lethal. At Tarin Kowt on November 18, for example, Taliban forces tried to recapture the village by advancing in a column of vehicles up an exposed road. Frightened AMF defenders were prepared to abandon the village, but precision air strikes called in by American commandos located on an overlooking ridgeline decimated the Taliban column, whose survivors fled the scene in disorder.[78] Taliban reserves ordered forward to reinforce their defenses at Bai Beche were caught moving in the open along Highway 326 between Shulgerah and Keshendeh-ye Pa’in and were slaughtered by American airpower; officers who surveyed the scene afterward said it brought to mind the infamous “Highway of Death” leading out of Kuwait City in the 1991 Persian Gulf War.[79] Taliban armored vehicles and heavy weapons left parked on the surface were routinely destroyed by pinpoint PGM attacks as at Bagram on October 20, Oimetan on October 25, Ac’capruk on November 4-7, or Polanyi Canyon on November 7.[80] Crude foxholes dug in the open without overhead cover or effective concealment were regularly annihilated by precision bombing with satellite-guided 2,000 pound JDAMs (Joint Direct Attack Munitions).[81]
Where fighting positions were properly prepared, however, they were much harder to destroy – even with modern PGMs. At Bai Beche, for example, al Qaeda’s fighting positions offered ample overhead cover and reinforced firing positions from which attackers could be engaged without exposing the garrison. Though the defenders could not all be located individually within this trench system, American commandos knew the system’s extent and thus called for heavy bombing across the entire position for more than two days. Yet even after all this, enough defenders survived to thwart Dostum’s initial attack.[82]
In the Qala-e-Gangi fortress uprising, the renegade prisoners were quickly driven out of the above-ground prison yard and isolated in a handful of small underground chambers whose locations and perimeter were well-known. These were then pounded by allied airpower: entire ammunition payloads of multiple AC-130 gunships and no fewer than seven 2000-pound JDAMs were expended against this tiny area. Yet the defenders survived and continued to resist until succumbing only to the medieval technology of flooding by cold water.[83]
In Operation Anaconda, well-prepared al Qaeda positions survived repeated aerial attack by U.S. PGMs. On Objective Ginger, for example, American infantry inadvertently disembarked from their assault helicopters almost on top of an unseen al Qaeda position on March 2; after being pinned down for much of the day, they were extracted that night. American troops then spent much of the next 10 days fighting their way back toward the Ginger hilltop from more secure landing zones well to the north. In the meantime, American aircraft pounded the hill. Yet in spite of over a week of sustained heavy bombing, al Qaeda positions on Ginger survived to fire upon U.S. infantry when the latter finally reached and overran the objective. One dug-in al Qaeda command post was found surrounded by no fewer than five JDAM craters, yet its garrison survived and resisted until overrun by U.S. infantry.[84]
This is not to suggest that precision firepower is not extremely lethal, or that even well dug-in al Qaeda defenses did not suffer heavy losses from precision engagements. But the evidence does indicate that a combination of cover and concealment can allow defenders, though battered, to survive modern firepower in sufficient numbers to mount serious resistance.
Nor is this the first time that properly prepared defenses have survived massive firepower, precise or otherwise. French defenses at Verdun in 1916 endured a two-day German artillery barrage equal to about 1,200 tons of explosives – in nuclear parlance more than a kiloton, or more explosive power than the W48 tactical nuclear warhead – yet enough of the entrenched defenders survived this maelstrom to halt the German assault.[85] In 1917, German defenses at Messines absorbed more than a kiloton of explosive power per mile of frontage, yet still halted the ensuing British offensive.[86] At Cassino on March 15, 1944, German positions in the village were struck by 300 tons of bombs in a single day, yet defeated the associated Allied infantry advance.[87] On July 18, 1944, more than 4,500 Allied aircraft, three corps’ worth of artillery, and naval gunfire from two Royal Navy cruisers and the monitor Roberts deposited more than 8,700 tons of explosives – over eight kilotons of firepower – on just seven kilometers of German frontage in less than three hours in Operation Goodwood.[88] Yet the entrenched Germans halted the subsequent British armored advance, killing more than one-third of all the British armor on the continent in the process.[89] Firepower on such scales is tremendously destructive, and each of these defenders suffered heavily under such barrages. But even fantastic volumes of firepower alone cannot annihilate defenses outright. Precision allows crushing firepower to be delivered using vastly fewer platforms, but to expect precision to accomplish what literally nuclear-scale fires have not been able to attain in the past is to ask too much of new technology. The village of Cassino was struck by far less accurate weapons than the al Qaeda defenders of Objective Ginger, but this tiny 500 by 700 meter Italian hamlet was still hit with the equivalent of more than 300 2000-pound JDAMs, which was more than enough to reduce every building in the village to rubble – but not enough to exterminate its defenders.[90] The problem at Verdun, Messines, Cassino, or Operation Goodwood was not any inability to turn defenses into crater fields or reduce specific buildings to rubble without today’s precision – the problem was that resolute defenders can survive even within crater fields and rubble piles to mount serious resistance. Firepower is critical, but against resolute, well-prepared defenders it has rarely been sufficient by itself in the past; taken together, Bai Beche, Qala-e-Gangi, and Operation Anaconda suggest that it may not be now, either.
3. The Role of Indigenous Allies in Afghanistan
The third logical precondition for a broadly applicable Afghan Model is that it be effective without placing heavy demands on indigenous allies’ tactical sophistication. Here, too, experience in Afghanistan varied importantly.
Early in the campaign, it was not uncommon for Taliban defenders to withdraw after SOF-directed standoff bombing destroyed key positions. At Polanyi Canyon, for example, the AMF advance following initial air strikes found only abandoned (or destroyed) Taliban positions and met no residual resistance.[91]
This, too, changed after the early engagements. The remainder of the campaign saw extensive close combat. Moreover, this close combat was neither trivial nor wholly one-sided: many of the outcomes were close calls, involving initial reverses, heavy AMF losses, or both.
At Bai Beche on November 5, for example, the dug-in al Qaeda defenders refused to withdraw in spite of over two days of heavy American air strikes. To dislodge them, Dostum’s AMF cavalry was ordered to charge the position. The first attempt was driven back. The American SOF attached to Dostum’s forces observed this reverse and began calling renewed airstrikes against the al Qaeda positions in anticipation that Dostum would eventually order a second assault. In the process, however, a SOF warning order to the cavalry to prepare for another push was mistaken by the cavalry as a command to launch the assault, with the result that the cavalry began its attack much sooner than intended. The surprised Americans watched the Afghan cavalry break cover and begin their advance just as a series of laser-guided bombs had been released from American aircraft in response to the SOF calls for air support. The SOF commander reported that he was convinced they had just caused a friendly fire incident: the bomb release and the AMF cavalry advance were way too close together for official doctrinal limits, and the air strike would never have been ordered if the SOF had known that the cavalry was then jumping off for the second assault. As it happened, the bombs landed seconds before the cavalry arrived on the position. In fact, the cavalry galloped through the enormous cloud of smoke and dust that was still hanging in the air after the explosions, emerging behind the enemy defenses before their garrison knew what was happening. The defenders, seeing Dostum’s cavalry to their rear, abandoned their positions in an attempt to avoid encirclement.[92]
The result was an important victory – in fact, the victory that turned the tide in the campaign. But the battle involved serious close combat (cavalry overrunning prepared, actively resisting defenses), and the outcome was a very close call. Dostum’s assault profited from an extremely tight integration of movement with suppressive fire – far tighter, in fact, than either Dostum’s troops or their supporting SOF would ever have dared arrange deliberately. Luck thus played an important role in the outcome. Dostum might well have carried the position eventually even without the good fortune of an extraordinary integration of fire and movement; this was clearly a crucial battle and he would presumably have redoubled his efforts if the second attempt had failed. But as fought, the outcome involved an important element of serendipity.
Nor was Bai Beche unique in demanding hard fighting at close quarters. As noted above, al Qaeda counterattackers reached small arms range of U.S. and AMF forces before being driven back at Sayed Slim Kalay and at Highway 4.[93] At Konduz in late November, al Qaeda counterattackers penetrated AMF positions deeply enough to compel supporting American SOF teams to withdraw at least three times to avoid being overrun.[94] In Operation Anaconda, AMF forces associated with the warlord Mohammed Zia and supported by American SOF were assigned to drive al Qaeda defenders from the “Tri-cities” area (the villages of Shirkankeyl, Babakuhl and Marzak); they were instead pinned down under hostile fire from prepared defenses in the surrounding mountainsides and eventually withdrew after they proved unable to advance. Only after the al Qaeda defenders pulled back under joint, multinational attack by allied airpower, U.S. infantry, and multinational SOF were the AMF able to enter the Tri-cities and adjoining ridgelines.[95] At Tora Bora, massive American bombing proved insufficient to compensate for the AMF’s unwillingness to close with dug-in al Qaeda defenders in the cave complexes of the White Mountains; many now see this ground force hesitancy as having allowed Osama bin Laden and much of his command structure to escape capture and flee into neighboring Pakistan.[96]
Among these examples, the fighting along Highway 4 in December is particularly instructive. Shirzai’s Afghans here were divided among two factions. The first, commanded by Haji Gul Alai, were very capable troops by Afghan standards. They used terrain for cover and concealment, maintained good intervals between elements in the advance, moved by alternate bounds, exploited suppressive fire to cover moving elements’ exposure, and were able to exploit the effects of American air strikes by coordinating their movement with the bombing (which many AMF factions could not). The second faction, by contrast, were much less skilled: the attached SOF commander characterized them as “an armed mob – just villagers given weapons.” Their tactics consisted of exposed, bunched-up movement in the open, with no attempt to use the terrain to reduce their exposure, and little ability to employ supporting or suppressive fires. At the Arghestan Bridge on December 5, this second faction launched an assault on a dug-in al Qaeda position south of the Kandahar airport. Driven back repeatedly, they proved unable to take the position, in spite of U.S. air support. Only after these troops were withdrawn and Haji Gul Alai’s forces took over the assault the following day could the al Qaeda positions be taken.[97]
What conclusions should be drawn from this experience?
Will American infantry be needed any time the United States uses force?
Not necessarily: where our Afghan allies met their enemies on something like an equal tactical footing, U.S. air power and small SOF teams made all the difference. With such support, unsophisticated allies overwhelmed unsophisticated opponents at Polanyi Canyon and Tarin Kowt – and better-skilled allies beat comparably skilled, resolute opponents at Bai Beche and Highway 4.
But does this mean that American air power can turn unskilled, unsophisticated allies into conquerors even when facing enemies with superior skill and resolution, as many now suppose?
No.
By and large, the Afghan campaign pitted ground forces of roughly comparable ability and equipment: the net result of years of civil warfare had been a stalemate in which the Taliban had proved unable to dislodge the Northern Alliance from the roughly 15 percent of Afghanistan that it held prior to September 11. But though the typical combat units on each side were thus about equally matched (as the stalled pre-intervention battle lines imply), both armies were actually diverse mixtures of better- and worse-trained, more- and less-motivated troops – and this diversity offers a couple of valuable opportunities to observe instances of unequally-skilled forces in combat. In such unequal fights as the first day at Arghestan Bridge and the assault on the Tri-cities in Anaconda, the results suggest that where our allies are substantially overmatched tactically, American air power and SOF support per se may not be enough to turn the tide. Even where the balance of skills was closer to parity, as at Bai Beche, the results could still be perilously close calls.
If SOF-directed air power had simply annihilated Taliban forces at standoff range, as some now suppose, then even a radically unsophisticated army would have been sufficient to walk in and occupy the blasted ruins. But in Afghanistan, though air power could destroy most of a hostile force, it could not annihilate well-prepared defenses outright. Nor could it defeat well-directed, skillfully concealed assaults by itself. The result was a series of close combat actions, rather than a war fought exclusively at standoff ranges. And in these actions, even a few al Qaeda survivors, properly motivated and armed with modern automatic weapons, could mow down large numbers of unsophisticated indigenous soldiers caught exposed in the open. To survive long enough to take advantage of the tremendous firepower leveled by American air support thus still required the fundamental combat skills of cover, concealment, dispersed small unit maneuver and local suppressive fire. Novice troops without these fundamentals either found themselves pinned down and unable to advance or driven back with heavy losses. Better-skilled troops able to advance under fire, on the other hand, found the precision firepower of American air strikes to be a decisive advantage. And where Afghan Taliban irresolution or exposure left an undefended battlefield following American air strikes, then even unskilled allies did indeed prove able to occupy the abandoned ground.
The results thus suggest that where the troops on the ground are comparable – either both skilled or both unskilled – American precision fires can make the difference. But where an unsophisticated ally is pitted against an enemy with the skills and motivation to survive precision engagement and fight back when attacked, then poorly trained allies will be unable to take advantage of the enormous potential that precision fires bring. Even with precision air support, indigenous allies thus need a combination of skill, motivation, and equipment at least broadly comparable to their enemies’ to prevail.
A New Explanation: The Afghan Campaign as Orthodox Theater Warfare
Neither of the main views in the current debate is thus wholly satisfactory. On the one hand, the campaign’s idiosyncrasies were less decisive than often assumed. On the other hand, the systematic lessons most commonly drawn are at odds with the war’s actual conduct – the fighting did not unfold as the Afghan Model supposes.
A more satisfactory accounting would be to view the campaign neither as a fluke nor as a military revolution, but rather as a surprisingly orthodox example of modern joint theater warfare – albeit one with unusually heavy fire support for ourselves and our allies. While the Afghan campaign was unique in some respects and new in others, it is easy to exaggerate the war’s distinctiveness. There were important continuities between Afghanistan and prior military experience – and these continuities were at least as important for the campaign’s outcome as its differences.
The campaign, after all, centered on a struggle between two sizeable armies. Although public attention has tended to focus on the role of Western air power and the handful of commandos who directed it, the purpose of that Western role was to tip the scales in an ongoing, preexisting land war. The Taliban and Northern Alliance forces whose struggle this air power was meant to influence deployed together some 60-80,000 troops (let alone the Southern Alliance forces that arose as the fighting moved south).[98] These troops were striving to take and hold ground in very orthodox ways prior to the American intervention, and al Qaeda continued to contest control of key geographic points through the end of Operation Anaconda in March.[99] In fact, American commanders were surprised by al Qaeda’s willingness to stand their ground and fight for the Shah-i-kot valley at Anaconda; CJTF Mountain had expected them to melt away in guerilla fashion once struck rather than reinforcing a dug-in territorial defense-in-place as they actually did.[100] The campaign as a whole, moreover, produced extensive close combat, the great majority of which was waged by conventional infantry.
Of course, the outcome of this territorial struggle was affected profoundly by SOF-directed precision air power – this ultimately made the difference between stalemate and victory. But it did not do so by itself. Tightly integrated ground maneuver was necessary to exploit its effects once the Coalition began to encounter chiefly foreign Taliban and al Qaeda opponents with significant skill and motivation. To overcome skilled, resolute opposition required both precision fire and maneuver; neither alone was sufficient in Afghanistan.
This, however, was hardly new.
In fact, the necessary synergy between fires and maneuver has been at the heart of most great power military doctrines since at least 1918. The history of modern tactics is largely a story of the discovery of its importance in the crucible of the Western Front and the subsequent spread of such methods across nations and over time.[101] The German tactical system that first broke the Western trench stalemate in March 1918, for example, was centered on new methods for combining suppressive fire and movement via dispersed formations that used local terrain for cover.[102] Allied armies had learned similar lessons from hard experience in the trenches and adopted similar methods by war’s end (albeit with less complete implementation in the field).[103] The German tactics that wrought blitzkrieg in 1939-41 were partly a matter of exploiting the tank’s new mobility, but largely a product of a heavy emphasis on combining fire and movement in balanced, combined arms formations.[104] Germany’s opponents had allowed their skills in integrating fire and maneuver to atrophy in the interwar years, but under the pressure of hard fighting these skills were gradually recovered, contributing centrally to the end of blitzkrieg after 1941 and the gradual triumph of allied arms by 1945.[105] Israel’s success against poorly-trained Arab infantry in 1956 and 1967 led them to de-emphasize orthodox fire and maneuver in favor of an imbalanced stress on the shock effect of mounted maneuver per se; they paid for this imbalance with their heavy losses against better-prepared Egyptian infantry in the Sinai in 1973.[106] American doctrinal writers initially responded to the Israelis’ 1973 experience by substituting a one-sided emphasis on anti-tank fires for the Israelis’ converse overemphasis on maneuver; by 1982 this imbalance was remedied by a return to orthodox fire-maneuver synergy in the new doctrine of AirLand Battle.[107]
Nor was it new that the fires and the maneuver were provided by different nations in Afghanistan. In Vietnam, our late-war, post-withdrawal policy was precisely to support a South Vietnamese proxy with American air power, much as we did with our AMF proxies in Afghanistan. Of course, the ultimate outcome was much more favorable in Afghanistan. In part this is due to air power’s greater lethality in 2001-2. But the difference is also due in part to the different balance of maneuver skills on the ground in these two conflicts. In Afghanistan, our proxies were generally a match for their enemies tactically, and in key battles like Bai Beche they were able to integrate their maneuver closely with our fires. In Vietnam, by contrast, the southern Army of the Republic of Vietnam (ARVN) was generally less skilled and motivated than its northern foes – and American air support was correspondingly less decisive.[108]
As in Afghanistan, however, there were exceptions to the general rule in Vietnam, too, and these exceptions offer some instructive parallels. In the 1972 Easter Offensive, for example, North Vietnamese attackers who had heretofore waged a mostly guerrilla war were thrown into a conventional, large-unit armored assault. With little experience in such tactics, they had great difficulty combining arms, coordinating attacks, or exploiting terrain for cover. American air power thus found massed, exposed targets in the open and inflicted heavy losses, enabling our ARVN proxies to contain the offensive.[109] By contrast, in the South Vietnamese Lam Son 719 offensive in February 1971 it was the ARVN which found itself in an unaccustomed role, launching its first major airmobile assault without American advisors (who were forbidden to accompany them to their objectives in Laos); the North Vietnamese were here implementing very familiar tactics, waging just the sort of conventional defense they had practiced in more than a decade of continuous warfare against American and French attackers. American air strikes here enabled the ARVN to get as far as the Laotian village of Tchepone, but ARVN maneuver skills left much to be desired and losses were correspondingly heavy; the ensuing ARVN withdrawal left much of their equipment behind and proved so harrowing that no similar offensive was ever again attempted.[110] These actions from the early 1970s were actually much more like Afghanistan than many now suppose: where our Afghan allies met opponents with superior tactical skills, even lavish, radically-precise air support proved insufficient; where they met their enemies with something like equal skills and motivation, American firepower made all the difference.
By contrast, many now believe that in Afghanistan we turned a ragtag militia into conquerors who subsequently overwhelmed a superior enemy by simply walking forward in the wake of our precision bombing. This belief is largely responsible for the general perception of military revolution in Afghanistan – and if the war had really been fought this way, then the perception would be right. But the war was not actually fought this way. And what did happen was much closer to the longstanding historical precedent on the need for integrating fire and maneuver to overcome skilled, resolute opponents. Of course, this is not to suggest that nothing has changed since 1918. In particular, the form that fire support has taken has changed dramatically since then – and the increases in firepower’s range, precision, round-for-round lethality, responsiveness, and flexibility have obviously been dramatic in recent years. In a sense, these changes have been revolutionary if considered chiefly in terms of their effect on the way fires have been provided.
Moreover, the increasing lethality of standoff precision engagement has made the combination of fire and movement much more powerful where both elements are present. Tight integration of JDAMs or laser-guided bombs with skilled ground maneuver is far more effective today than were 77 mm field guns and German stosstruppen in 1918. This is an important development and has greatly increased the real military power of the United States today relative to any plausible foe.[111]
But what new technology has not done is to enable us to beat opponents using either fire or maneuver alone. The Afghan campaign – like the Western Front in World War I, the invasions of France, Poland, or Russia in 1939-41, the Northwest European campaign of 1944-45, the Mideast Wars of 1956-1982, or any of dozens of other, similar examples – was a joint air-land struggle in which the ability to combine fire and maneuver by diverse arms made the difference between success and failure. Seen in this larger context of the underlying determinants of campaign outcomes in modern warfare, this continuity in the need for a demanding system of integrated fire and maneuver is at least as significant for the future of warfare – and for the policy decisions that turn on this – as the accompanying changes in the technology of precision firepower.
Comment