The strategy of General Stanley McChrystal, the commander of international forces in Afghanistan, is sound in conception: a military push against the Taliban to extend security accompanied by a step up in civilian reconstruction efforts. Each will be ineffective without the other. Without greater security, humanitarian assistance efforts cannot tangibly improve Afghan lives; without more reconstruction, the local population will see no benefits stemming from the occupation.

It is encouraging that there is co-ordination with the Pakistani military to squeeze the Taliban from both sides of the border. This has been missing in previous Helmand operations, enabling the Taliban to regroup in the Pakistan tribal regions. The focus on Helmand ? the Taliban's stronghold and the producer of more than half of the country's opium crop ? also makes sense. Afghanistan will never be stable until the power of Hamid Karzai's government extends throughout the country. Winning control of Helmand will be a crucial step towards achieving that goal.

It is also encouraging that the focus of this surge will, we are told, be less on killing the Taliban than winning the support of the local population. There are indications this is not just warm words. The US has finally ditched its counterproductive policy of eradicating the poppy crop which catastrophically alienated Afghan farmers.

However, the fact that the strategy is sound does not mean that it cannot fail. The US military surge in Iraq helped marginalise al-Qa'ida. But conditions in Afghanistan are very different. The Iraq surge was essentially an urban operation. The Taliban is a rural force. In Iraq, the US had the support of Sunni tribes. There is no similarly powerful local ally for international forces in Afghanistan. And what we cannot know at this stage is whether enough troops have been assigned to this mission to be able to hold the territory once cleared.

The other big unknown factor is civilian casualties. If this operation results in a large number of deaths of innocent Afghans, it is surely doomed to failure. Recent years have shown that the surest way to kill innocent civilians is through the deployment of air power. Operation Khanjar is primarily a ground offensive, but it is unclear whether the US military has learned that the clumsy use of air support does more harm than good.

The troop surge seems to have met relatively mild resistance thus far. But this is the Taliban's usual approach: retreat, regroup, and then hit back. The US aims to stabilise Helmand in time for next month's Afghan presidential elections. But to prevail against a patient enemy like the Taliban, international forces must be prepared for a much longer stay. They had also better be prepared for the casualties this mission will impose. We have already had a bitter taste of this with the death of two more British soldiers this week, including the highest ranking British officer to be killed in action since the Falklands.

Most important of all, if this twin civilian and military surge is to stand any chance of succeeding, the international forces in Afghanistan must learn from the mistakes of 2001, when the world assumed, quite wrongly, that the battle for the country was essentially won. Eight years on, they must be prepared to stay the course.