Like most Northern Alliance generals, Abdul Rashid Dostum's experience with Russian bombers during the Soviet invasion of Afghanistan in the 1980s left him skeptical about calling for help from the air. You never knew when--or even where--the bombs would hit.
But on Nov. 8, he had no choice. It had been days since his rebels won their first victory in Mazar-i-Sharif, and he was watching Al Qaeda fighters amassing to retake the northern city of Kunduz.
"We need some air," he told a young U.S. Air Force special operations lieutenant.
Within 20 minutes, the eyes of Afghanistan's most feared warlord widened as a succession of fireballs erupted over an expanse the size of a football field, killing 259 Al Qaeda fighters and taking out a command center, artillery and armored vehicles.
"You've got to be kidding," Dostum said. He hadn't expected the strike for a day or more.
Such a precise, rapid air assault could not have occurred in any previous war. And not just because of the technology.
The episode encapsulates much of what went right for the coalition forces in Afghanistan, a combination of factors that had never been pulled together successfully before. This included a unique, on-the-ground collaboration between Americans and a disparate, half-trained collection of local fighters; nearly seamless ground-air teamwork, thanks in part to the debut of a laser range finder that let the troops below direct pilots to their targets; and an assortment of "smart" bombs and rejiggered "dumb" bombs that turned even the ungainly B-52 bomber into a virtual attack fighter.
As such, the tale is one of the success stories combat commanders have sent to the desk of Defense Secretary Donald H. Rumsfeld, who will decide how the lessons of the Afghan war could be used on future fronts in the war on terrorism.
They will be balanced against the apparent failures--foremost, the escapes of hundreds or thousands of Taliban and Al Qaeda fighters due to the willingness of some members of the proxy army to look the other way. There were civilian casualties of the U.S. bombing, as in every war, touching off a fierce debate within the Pentagon over whether the military was reckless or actually so obsessed with avoiding "collateral damage" that it missed targets and prolonged the war. And there was the decision to make airlifted food rations the same color as cluster bombs.
All of it is being sifted for guidance on how and where to proceed next in the war on terrorism.
Somalia and Iraq, for example, offer the tantalizing prospect of enabling the Pentagon to reprise a combination of U.S. air power and proxy forces on the ground. Pentagon planners, fearful of reliving a failed 1993 mission in which 18 U.S. soldiers were killed and one was dragged through the streets, are considering using Ethiopian rebels to help engage suspected Al Qaeda sites in the vast ungovernable expanses of Somalia.
The proxy strategy that gave Dostum access to a B-52 bomber was ill-formed when the campaign began. Defense officials feared the Northern Alliance could not keep up with the pace of U.S. air gains, and disaffected southern Pushtuns would refuse to raise arms against their ethnic tribesmen in a ground war led by northern ethnic rivals.
"It wasn't really totally thought out that they were going to become a surrogate army," said a senior Air Force official involved in crafting the campaign. "The strategy evolved as it came along. The idea of using the Northern Alliance as surrogates was always there, but the confidence level of whether they could do it wasn't always assured."
The strategy has its shortcomings. For one thing, local forces don't always have the same agenda. More than once, Taliban leader Mullah Mohammed Omar has been reported to have negotiated a surrender with anti-Taliban forces only to escape. Some Pentagon officials have speculated that Afghan allies on the ground, more interested in control than capture, simply looked the other way as Omar and others fled.
In Iraq, the presumed proxy army would be the anti-government Kurds, who have spent much of their time fighting among themselves and lack the military might mustered by the Northern Alliance. The State Department recently rescinded funding for one major anti-government group, the Iraqi National Congress, over financial mismanagement.
"The model sort of falls apart," said Michele Fluornoy, a military analyst at the Center for Strategic and International Studies and author of "To Prevail," a book on the anti-terrorism campaign. "I don't think we can contemplate the use of the model in Iraq without the use of ground forces on a major scale."
A Proving Ground for Alliances, Weapons
Then again, in Afghanistan, the Northern Alliance was initially ridiculed by some western military analysts. The Americans learned a great deal about harnessing the strengths of imperfect indigenous fighters. Hawks note that the Kurds weren't effectively used in the Persian Gulf War campaign and argue that they could prove useful in a war to topple Saddam Hussein's military machine, with its weapons of mass destruction and an army that far outnumbers the Taliban.
The Afghan war also became a proving ground for a wide range of technology. Early on, for example, a debate raged over the usefulness of unmanned bombers, pilotless surveillance craft that had never before been fitted with bombs.
"That's over," Fluornoy said. "We know it's incredibly useful. They allowed us to get very close to take out time-sensitive targets without risking American lives."
The Gulf War was known for introducing precision-guided bombs, but only 1 in 10 explosives dropped by U.S. warplanes in that war was "smart," compared with 2 out of 3 in Afghanistan.
The "smart" CBU-103 cluster bomb Dostum saw in Kunduz also got its first battle testing in Afghanistan. It is simply the "dumb" version of a bomb used in the Gulf and Kosovo wars, fitted with course-correcting tail fins. The tail device corrects the drift of the bomb canister even in high winds and determines the proper location for the bomblets to disperse. Because they were delivered from more than 15,000 feet, strong winds tended to blow them off course.
Cluster bombs are designed to disperse in the air to maximize the damage to large groups of soldiers or equipment. The type used in Afghanistan breaks into 202 soda can-sized bomblets.
After testing an updated model for two decades, the Pentagon is now preparing to deploy a smarter cluster bomb, the senior Air Force official said. The new "sensor-fused" weapon uses fewer, more powerful bomblets and is designed to strike large vehicles and troop concentrations. What make it smarter are a parachute and a heat-seeking mechanism that allows it to literally follow tanks and other moving targets. In early versions, the first bomblet to reach its target set off the others, causing a string of explosions but striking only the first target. New versions have apparently corrected that problem.
Limits to Uses of Cluster Munitions
The newest cluster munitions might be effective against lines of tanks in Iraq but are too powerful to be used on small targets such as the terrorist cells in Somalia or Yemen, military analysts say.
"If you were worried about, say, half a million North Korean soldiers lunging across the DMZ [demilitarized zone] or if you were worried about Iraq trying to grab Kuwait while nobody was looking, then this certainly would be a weapon of choice," said John Pike, a munitions expert at Globalsecurity.org. "But not if you're talking about terrorist training camps."
In the "no kidding" category of lessons learned, top Pentagon strategists privately acknowledge that it was a mistake to make the food ration dropped by U.S. cargo planes the same color (yellow) as cluster bombs. About 5% of cluster bomblets land unexploded. They later go off with the touch of a hand, maiming or killing civilians who mistake them for food.
Publicly, defense officials have said such confusion was not a problem because the bombs were dropped among heavy troop concentrations, while food rations have been dropped largely in the rural countryside. However, civilians have been known to pick them up months after combatants have cleared a battlefield. Three decades after American warplanes rained cluster bombs near the Vietnamese border with Laos, reports of civilian casualties persist. Most of the victims are children.
The deaths of civilians, whether by unexploded ammunition or misdirected bombs, are a lesson learned in every war. But the more sophisticated the weaponry, the less chance of collateral damage, military officials point out.
Smart Bomb Found to Be Highly Accurate
Another smart bomb likely to be widely used in anti-terrorism strikes is the Joint Direct Attack Munition. Military officials say they're cheap, accurate and unaffected by weather. According to Air Force figures, 85% of the weapons have destroyed their targets, and the accuracy of individual bombs has exceeded the accuracy specified by the manufacturer, Boeing Co., striking within 43 feet.
The bomb, which was dropped 650 times in Kosovo, has struck Afghan soil 4,600 times.
Like smart cluster bombs, the JDAM merely adds a satellite-based guidance system to "dumb" bombs--the Pentagon prefers the term "gravity" bombs--as large as 2,000 pounds. They can be launched up to 15 miles from the target and, compared to older precision bombs costing up to 10 times as much, they're affordable. Under a 1995 contract, Lockheed Martin and Boeing charged the Air Force about $18,000 per bomb.
Unlike pricier and more finicky laser-guided munitions, they can be dropped in any weather. In Kosovo, Air Force officials complained that persistent cloud cover often made laser-guided weapons useless. JDAMs use the satellites of the Global Positioning System. Heavy bombers like the B-1 and B-2 use GPS rather than laser guidance.
For all those reasons, "the JDAM is the weapon of choice," Pike said.
Smart bombs such as the JDAM have given new life to the B-52, which first rolled off the Boeing assembly line in Seattle in 1951 to carry nuclear warheads and ended production a decade later. Supporters of the B-52 say it's only coincidence that the vintage bomber was used in both "friendly fire" incidents in which U.S. bombs killed American soldiers and Afghan allies in Mazar-i-Sharif and near Kandahar.
The incidents are still being investigated, but a senior Pentagon official with knowledge of the air campaign said the errant strikes were more likely because of the ground forces mistakenly calling in their own coordinates.
"It was human error, it sounds like, one way or another," he said.
CIA, Defense Dept. Find Ways to Cooperate
The reviews on intelligence-gathering in Afghanistan have been mixed. The bombing strikes became more devastating as ground spotters put their eyes and lasers on pro-Taliban targets, Rumsfeld has said. One field evaluation forwarded to the Pentagon cited a newfound "synergy" between the CIA and the Department of Defense. CIA spotters and interrogators worked alongside U.S. special operations troops to ferret out bombing targets from Mazar-i-Sharif to the eastern caves and bunkers of Tora Bora and Zhawar Kili.
A cumbersome target-approval process that reportedly allowed Taliban leader Omar literally to escape from the sights of U.S. forces was smoothed out to the point where he would have been hit under the same circumstances later in the campaign, Fluornoy said.
"Historically, the two cultures have clashed," she said of the CIA and the Pentagon. "They've refused to share information. They've refused to conduct operations jointly. This is really a cultural breakthrough."
Yet human intelligence reports have proved unreliable, placing Omar and accused terrorist mastermind Osama bin Laden in various spots, often at the same time. Some critics suggest that double-agent allies using American satellite phones have deliberately pointed warplanes to civilian sites.
Intelligence officials have apparently had no luck infiltrating the Al Qaeda network and argue that fundamentalist Muslim terrorism groups are nearly impossible to penetrate. Yet critics note that John Walker Lindh, the youthful Al Qaeda fighter from Marin County now charged with conspiracy to kill U.S. citizens and providing aid to terrorists, inveigled his way into Al Qaeda and met Bin Laden within two years.
The war confirmed the widely held view that the U.S. intelligence apparatus suffers by valuing technology over linguists and analysts, said Lt. Col. Ralph Peters, a retired military intelligence officer.
"The fundamental lesson is that we don't do the human side well enough," Peters said. "We have these networks of satellites and electronics that cost literally . . . tens of billions of dollars each year, and yet we'll pay the young person who's analyzing the data $25,000 a year. We need a much more balanced intelligence system."