Lawrence Freedman: The Americans have left it too late to send in ground troops before winter

The most controversial issue internationally in contemporary wars is the air campaign, because that leads to the death of civilians, while the most controversial issue domestically is the ground campaign, because that can lead to casualties amongst one's own troops. There is a trade-off between the two: the US always hopes that a successful air campaign will remove the need for a ground campaign, while the more successful a ground campaign, the more focused the application of air power can become.

One way to conduct a ground campaign without risking excessive casualties is to back somebody else's army. This is what the US is now doing with the Northern Alliance. The move has been reluctant. It is not only that this is an alliance more in name than in practice, but its military prowess is, to say the least, unproven.

The US had hoped to create a broader coalition that would make the Northern Alliance's return to Kabul more palatable to the inhabitants and acceptable to Pakistan. Unfortunately a single vision for the country has eluded the many Afghan factions for over a decade, and it is unlikely that one can be forged now in short order.

At the same time, the idea that, with a little support, the Northern Alliance would push through demoralised Taliban ranks has turned out to be optimistic. For the first weeks of the war the Alliance contented itself by alternating bombast with excuses for its lack of progress, arguing that the US needed to switch its air campaign from so-called strategic targets to the Taliban front line.

This argument had merit, in that, after the first few days, demoralisation wasn't exactly what was being achieved by many of the strikes. Even though, compared with recent air campaigns, this one was modest, the impression was given of something more vicious, causing unnecessary loss of life. The Taliban's capacity to operate has undoubtedly been impaired, but neither they nor Osama bin Laden's al-Qa'ida group depend for survival on the structures destroyed.

The importance of the recent switch in US tactics has been obscured by references to "carpet bombing", which dates back to the days of the massive air raids of the Second World War, when large numbers of aircraft were organised to cover a designated area, usually a city, so that nothing underneath could escape destruction.

The real precedent is a similar set of US raids, also using B-52s, against Iraqi forward positions in Kuwait in early 1991 as a prelude to the ground war. These raids were directed against trenches, far from civilian areas. The evidence is that the constant, incessantly noisy bombing did unnerve Iraqi troops, leading many to desert and leaving those that remained in no fit state to fight. There was no point in using expensive "smart" weapons for this purpose, because the objective was not so much to destroy enemy equipment or even to kill troops, but to terrify them.

Whether the much tougher Afghans and their al-Qa'ida allies will crack under this strain as easily as the Iraqi troops is another question, as is the ability of the Northern Alliance to press home any advantage if they do. This helps explain the importance of the arrival of numerous special forces into Afghanistan. They will have a number of roles, including the one canvassed at the start of the campaign, which is to try to find out where Osama bin Laden and his closest aides are hiding out.

Their most important role will be to work closely with the Northern Alliance. Part of this will be to identify the main Taliban troop concentrations and to ensure that air power is directed towards them. They will also, in effect, be sending intelligence on the Northern Alliance back to the Pentagon, for they will be able to report on the combat readiness of these unlikely allies, their equipment needs and the extent to which they will be able to take the war forward. The special forces might also start advising on tactics and training, and even specific operations. This could become controversial: from the Northern Alliance side if they resented interference; and elsewhere if this led to an even closer integration between the Alliance and the US.

Although it is entirely possible that the situation on the ground could change quickly, thoughts now seem to be moving to enabling the Alliance to operate throughout the winter while interfering with the Taliban's ability to do the same. Commando raids, such as the one that took place with much fanfare on 20 October, can provide what are at best irregular supplements. That raid faced serious resistance, although the Pentagon insists that the 12 casualties were largely due to accidents. The main problem is that intelligence has to be very good for this sort of action, and there are not many opportunities to mount them.

Another objective of the current stage in the fighting must be to establish forward operating bases. The old Soviet base at Bagram is an obvious candidate for this, but there are also reports of other, more rudimentary air strips not far from Kabul being considered as a means of bringing air power closer in so that it can respond more quickly to tactical exigencies.

The real issue remains the readiness of the Americans and their allies to introduce substantial ground forces of their own. Public opinion has been prepared for this, and for the potential casualties, but it may now be too late to get in forces of sufficient size to make a real difference prior to winter. So the emphasis may now be on keeping the pressure on the Taliban while preparing American forces to enter Afghanistan in numbers early next year.

Before that, humanitarian issues are likely to come to the fore as the war continues to aggravate what was already a wretched situation. It should be possible to get in relief to the north using the logistical system that is slowly being established by the US in the adjoining central Asian states, but the position in the south is going to depend much more on how Pakistan handles this sensitive border. The US plan could also be affected by another terrorist outrage or by chronic instability in one of the Muslim members of the coalition.

Neither Washington nor London pretended that this was going to be a quick war, and as all the problems come to light, it is clear just how long a haul this might be.

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