Thursday, September 18, 2008

The Intelligence War in Afghanistan

The war in Afghanistan is a counterinsurgency.
Counterinsurgency is above all an intelligence war - a matter of pinpointing the right targets at the right place at the right time so they can be captured or killed.
Intelligence matters because without it, the coalition effort is reduced to what the man in the humvee can learn about what's going on around him as he drives around.  Often this seat of the pants, "feeling-out" intelligence, or lack thereof,  results in a humvee hitting a roadside bomb, and people are often injured or killed.  Good intelligence allows security forces to create a new way of doing things.
Then it is too bad that seven years into the war, much of the intelligence effort in Afghanistan is rudimentary and fragmented and bound by serious shortcomings.
There are in essence three types of intelligence used to stop attacks.
The most basic kind, called 'pattern analysis',  charts past hostile acts and predicts when and where future hostile acts will take place.  For instance, insurgents plant roadside bombs on this bend in the road during the full moon.  So the next time the full moon hits, be wary of this corner or, even better, set an ambush and take out the bomber-setters when they arrive to emplace the bomb.
This is run of the mill, common intelligence; easy to collect and easy to graph, because bombs that hit friendly forces are easy to record.
The second kind of intelligence allows Americans and Afghans to kill High Value Targets (HVTs) - the people who plan and direct the minions who set the roadside bombs.
In order to be effective, intelligence needs to predict this HVT will be at a certain place at a certain time.  This allows soldiers or special operations people to swoop in and capture or kill him.
In Iraq the problem was never knowing who was the HVT.  It was predicting when and where he is expected to be somewhere to catch him.  HVTs tend to move every night to avoid this trap.  Many snatch raids came up empty because the HVT was not where he was expected to be.
This is called 'actionable intelligence' - the ability to predict he will be at his house at 11pm tonight.  Plain old intelligence saying a man is an HVT is a useful start, but it is not 'actionable'.
So obviously, intelligence is not just information, it is secret information.  If the HVT knows the Americans know he plans to be somewhere, he will change his plans.  The information is useless if it is not secret.
In this scheme of things, the intelligence effort in Afghanistan is a mess.
On the simple matter of 'pattern analysis', of predicting where a roadside bomb will be set, the Afghan forces are just beginning to set up a system for tracking past acts, in oder to predict future acts.  The Americans already do.  This is intelligence of the most basic kind.  But because so few Afghans can even write, it is a major task for the Afghan security forces to get rolling, and is only just beginning.
On the more difficult matter of generating actionable intelligence, the nature of Afghanistan itself presents insurmountable hurdles.  HVTs may be expected to stay overnight in the house of a village elder, but how do you identify the house if there are no maps of the village, no street names nor house numbers?  How do you get in quickly enough if the insurgents can see the dust plume from your humvees ten miles off and run away?  And how do you fly in more quickly if there are too few helicopters to run snatch operations at a consistently high level?
It would require an agent on the ground to walk up to the house with a GPS, get the grid coordinates, and then send in a helicopter team to snatch the target - that is, if you first knew he was expected that night.
Sadly, none of those elements are in place on a large scale in Afghanistan.  So the HVTs can plot terror acts in comparative peace.
The Afghan National Army does have good contacts in the village, because many of its soldiers grew up in them.  Identifying HVTs and the terror teams they "command" is no problem.  Knowing who they are, though, is not enough.  Knowing where they are, where they will be at what time in the future, and then doing something about it is currently an insurmountable problem in many parts of the country.  So, in provinces you have cells of 20-90 men, under known leaders in specific districts, but there is little anyone can do about it.
Then there is a third level of intelligence - real intelligence.  That is infiltrating a working terror network with an agent in place, who can send back reports on operations, intentions and capabilities. 
That is what the US and the Soviet Union tried to develop against each other with varying degrees of success throughout the cold war.  It is the ultimate aim and gold standard of intelligence.
And it appears American and the Afghan forces aren't even on the road to that destination, yet alone actually there.