$20B/yr to air condition troops in Iraq and Afghanistan

When I've presented my work on thermal stress and economic productivity (also see herehere, here and here for related work) most people's first response is, "so... poor countries should use more air conditioning?" to which is say "Yes, but..." and then discuss the fact that air conditioning isn't exactly cheap if you have $1000/year to live on (so this investment may not always be worth it for poor individuals). To do this, I usually point out that and AC costs at least $100 to buy and $10/month (at least) in electricity costs.  But my casual back of the envelope estimates of operational costs might be way off.  I had been assuming that electricity and ACs could be obtained in a poor country for the same price I can get these goods in New York.  I was probably being too optimistic.

Listening to NPR today, I heard this report:
The amount the U.S. military spends annually on air conditioning in Iraq and Afghanistan: $20.2 billion. 
That's more than NASA's budget. It's more than BP has paid so far for damage during the Gulf oil spill. It's what the G-8 has pledged to help foster new democracies in Egypt and Tunisia. 
"When you consider the cost to deliver the fuel to some of the most isolated places in the world — escorting, command and control, medevac support — when you throw all that infrastructure in, we're talking over $20 billion," Steven Anderson tells weekends on All Things Considered guest host Rachel Martin. Anderson is a retired brigadier general who served as Gen. David Patreaus' chief logistician in Iraq. 
Why does it cost so much? 
To power an air conditioner at a remote outpost in land-locked Afghanistan, a gallon of fuel has to be shipped into Karachi, Pakistan, then driven 800 miles over 18 days to Afghanistan on roads that are sometimes little more than "improved goat trails," Anderson says. "And you've got risks that are associated with moving the fuel almost every mile of the way." 
Anderson calculates more than 1,000 troops have died in fuel convoys, which remain prime targets for attack. Free-standing tents equipped with air conditioners in 125 degree heat require a lot of fuel. Anderson says by making those structures more efficient, the military could save lives and dollars.
This suggests the annual price of AC for each of our 70,000 troops is $31,428.57 per soldier, two orders of magnitude over my ~$200 back of the envelope estimate. I'm certain that this price is not the actual consumer price that we would observe for air conditioners being used by residents in the long-run, but it's so much larger than my previous estimate that I may have to reconsider how effective I think AC expansion is for mitigating the economic impact of high temperatures.


  1. A question I've seen elsewhere, but can't recall the answer: to heat space (e.g. house) you need that many Joules. To cool space via AC, you can use fewer Joules (at least I've heard that said, but haven't bothered think about it).

    If true, that suggests that (once you've reconfigured all your infrastructure) it could actually be cheaper (energetically) to live in hot countries; and similarly, if the average year-round temperature goes up.

  2. A good idea. Unfortunately, thermodynamics isn't so generous. Heating and cooling a room 1 degree involves moving the same amount of thermal energy into or out of the room. This sounds like it should make heating and cooling a room symmetric processes. But the technology we have for accomplishing these two tasks is very different. Usually we heat spaces by releasing chemical energy through combustion (think about a burning a log in the fireplace). In this case, all of the chemical energy that is released is transformed into thermal energy that heats the room. If you ignore whatever goes out the chimney, this part of the process is 100% efficient. But when we cool a room with air conditioning, we don't achieve anything near that efficiency.

    Air conditioners are heat pumps that move thermal energy from a cool space into a hot space (against the thermal gradient). This is opposite of what "nature" would like to do: move thermal energy from the hot space to the cool space (with the thermal gradient). Because the air conditioner has to fight against this natural process of thermal energy diffusing back into the cool room (one might say the AC has to *ahem* fight entropy....), it has to expend additional energy. We supply this energy to air conditioners via electricity and it is released outdoors into the environment in the form of additional heat. If you touch the outside of an air conditioner (the part that sticks out of your window) it will feel warm (or hot) both because it is transferring thermal energy from inside the room to the outdoors and because it is consuming additional energy in the process, which it dissipates as heat. Thus, an AC is far from the 100% efficiency of our burning log (my guess is that it's not more than 30-40% efficient, since that's what an ideal Carnot engine can achieve at sea level, but I'm not certain).

    So, from a thermodynamic standpoint, cooling is much more difficult than heating. However, cost involves more than just thermodynamic efficiency. One can imagine a world in which logs become very rare and expensive, so the price of our simple heating technology might rise. But then again, if electricity is rare (as seems to be the case in parts of Iraq and Afghanistan), then air conditioning might become relatively more costly.