Lawrence Freedman: Why our ground troops have to invade Afghanistan

'Soon the bombing will risk doing no more than re-arranging the sand or hitting the wrong targets'

It is now clear that the air campaign in Afghanistan cannot last more than a few days before the commanders run out of targets. The bombing was never going to be decisive, which is why the focus is now switching to the next and much more difficult stage – the ground war.

The limited scope of the air war can be illustrated by a comparison with the Kosovo conflict. During the initial weeks, in addition to cruise missiles, Nato used about 500 aircraft, flying just under 300 sorties per day. The numbers were raised significantly before the attacks made much impression on either Slobodan Milosevic's will or the Serb oppression of Kosovar Albanians.

The first night's strikes against Afghanistan, in contrast, saw some 31 targets attacked, using only 15 aircraft in attack sorties plus 15 cruise missiles, supported by 20 aircraft. Until the final assessments are in, the Pentagon will not want to pronounce the minimal Taliban air defences and air force (at most 30 aircraft) destroyed or the complete degradation of the Taliban command and control facilities. None the less, further attacks will soon risk doing no more than rearranging the sand, or, even worse, catching the wrong targets, as may already have happened with some UN workers.

Neither Washington nor London has claimed that the strategic objectives can be achieved by air attacks against a political system that is crude and rugged in its infrastructure. As Donald Rumsfeld, the US Secretary of Defence, has observed, they will not "rock the Taliban back on their heels".

Although, in Kosovo, ground attacks had been explicitly ruled out at the start of hostilities, this was recognised as an error. The critical ground force was provided not by Nato but by the Kosovo Liberation Army. As they drew Serb forces into the open, they rendered them vulnerable to Nato aircraft. In this case the KLA role is being played by the Northern Alliance (NA).

It has been reported that up to 15,000 troops are close to the northern city of Mazar-i-Sharif, facing the main thrust of the NA offensive, with a similar number guarding Kabul. There are possibly another 30,000 fighters elsewhere, some under the control of fickle warlords who might change loyalties, especially if the NA makes progress. Integrated with the Taliban forces are hundreds of the few thousand non-Afghan graduates of the al-Qa'ida camps.

If the Taliban leave these forces where they are, they could be in for a battering from American aircraft. Once the aircraft are able to operate in daylight, Taliban trenches and artillery pieces will be easily picked out. After a few days of this sort of treatment, the Taliban might be unable to resist an NA offensive. Rather than face this, a strategic decision may be made to retreat to the mountains. That would mean relinquishing political control of the country in return for an opportunity to repeat the successful battle against Soviet occupation.

Any decision in Kabul to stand and fight may require bin Laden to put more of his fighters into the defence. This would probably be greeted with relief by the coalition who do not relish the thought of having to mount protracted special-forces operations to smoke out the terrorists, especially if they were joined by large numbers of Taliban fighters.

At one level, the current strategy promises an early success if it sees the Taliban in flight and the NA triumphant in Kabul. At another level, it ties the coalition too closely to the NA and its agenda, just as the Kosovo campaign did to the KLA, leading to misgivings elsewhere in the region, especially Pakistan. Furthermore, unless it performs considerably better than it did when last in power in the mid-1990s, the NA may not attract the loyalties of Afghans and may soon face the old problem of holding only cities and roads while swathes of the country are occupied by hostile fighters.

This is why the coalition should aim to get its own forces into Afghanistan in substantial numbers over and above the special forces deployed for individual operations. The best way to accomplish this would be to capture an airstrip in a convenient, defensible location and use it to ferry in troops. Unlike Kosovo, preparations appear to be advanced, with elements of the 101st Airborne Division being prepared to follow the 1,000 soldiers from the 10th Mountain Division already en route to Uzbekistan. This would release the coalition from appearing as simply the air arm of the NA. At the same time, Britain and other allies, including France and Germany, are ready to add several thousand each of their troops to give the "invasion" a truly international appearance.

Such a military presence on the ground would strengthen the coalition's political hand during deliberations to form the post-Taliban regime. To hasten that day they might be able to move to cut off any obvious lines of retreat and, to the extent that all else fails, a base will have been established for protracted counter-guerrilla operations.

The author is Professor of War Studies, King's College, London