For years, there's been lots of debate over "peak oil" — the notion that, at some point in the near future, the rate of global oil production will bump up against a hard ceiling. This is generally considered a gloomy prospect.

An oil pump operates in the desert oil fields of Sakhir, Bahrain, at sunset. (Hasan Jamali/Associated Press)

But another, related concept has started to bubble up in energy discussions lately. It's the notion of peak oil demand.

This is the idea that the world's appetite for oil may not be bottomless after all. Thanks to a combination of price increases, improvements in vehicle efficiency, as well as the advent of alternative fuels and electric cars, we may reach a point in the coming decades where the world's demand for oil actually starts declining. And if that happens before peak oil production hits, well, that would come as a relief.

The basic "peak oil demand" argument

The possibility of peaking oil demand recently came up in a long feature in the Economist. Right now, the world consumes about 89 million barrels of oil per day. Yet car travel in wealthy countries appears to be on the wane. New alternatives to gasoline are emerging — from natural-gas-powered trucks to plug-in electric vehicles. And as for developing countries, even China is now adopting stringent fuel-economy standards to curtail air pollution.

Tally it up, and analysts at Citi think the world's thirst for oil could peak in a few years at around 92 million barrels per day (the gray-blue scenario) — as long as vehicle efficiency for cars and trucks keeps improving by about 2.5 percent per year:

(The Economist)

This is an unusual prediction, to say the least. For context, BP expects global oil demand to keep growing from 89 million barrels per day today to around 104 million barrels per day by 2030.

So is Citi's forecast plausible? For some perspective, I asked Stanford's Adam Brandt, who recently co-authored a peer-reviewed study that explored the "peak demand" scenario. The story turns out to be a little more complicated ...

A more complex "peak demand" scenario

Brandt's paper takes a hard look at various historical trends on travel, vehicle efficiency and fuel use and tries to model what would happen if those carried forward. "This isn't meant to be a prediction," he explains, "but more about asking what the implications are of what we know about historical trends." Here's the main chart from his study:

Based on past trends, the study projects that global demand for conventional oil could peak around 2030 or so (the red wedge). And it could peak even sooner if nations adopted even stricter fuel-economy standards or if electric cars caught on faster than expected.

Where things get interesting is in that tan wedge, labeled "unconventional liquids." In Brandt's study, this represents all the possible fuels that could conceivably replace conventional crude oil in our transportation system. It could be natural gas. It could be biofuels. It could be oil from unconventional sources like Canada's tar sands or North Dakota's shale rock. The study is agnostic on this.

The precise contents of that tan wedge will have implications for things like climate change. If the world supplants its thirst for conventional oil with, say, oil from Alberta's tar sands and other "unconventional" oil, then carbon-dioxide emissions will keep rising and heat the planet. If electric cars powered by renewables (say) fill in the gap, the picture looks rosier.

So, Brandt says, we should worry more about that tan wedge rather than the red wedge. "Rather than the current focus on scarcity of conventional oil resources," the study concludes, "we believe that more attention should be focused on understanding and anticipating the economic, environmental and social consequences of adopting the various alternatives to conventional oil."

Will the world hit peak travel?

Clearly Brandt's analysis provides a different picture than Citi's forecast. Barring some huge policy shift, demand for some sort of liquid fuel for transportation is likely to keep growing enormously in the years ahead — the study projects that "transport energy demand" in Asia could grow more than 1,000 percent over the coming century.

Still, Brandt's study does project an overall peak in oil demand sooner than many other forecasts, such as BP's. And that's largely, Brandt explains, because their model suggests that global automobile travel may not keep growing indefinitely. "This is largely a story about land travel," he says.

There are already signs that wealthy countries — the United States, Canada, Australia, Europe, Japan — appear to have hit a peak in driving (see chart). The key question, then, is what happens in fast-growing countries like China and India. But here too, Brandt says, there are reasons to think that their demand for driving could hit a saturation point even as these countries get wealthier.

"You can't just take the rapid growth of the past 10 years and continue this extrapolation indefinitely," he says. For instance: "Eventually you get to the point where people in China are sitting in traffic. There's a long-standing notion of a travel time budget, the idea that there's only so much time people are willing to spend on travel. And that becomes the constraint, rather than income."

Note, by the way, that Brandt is purely talking about driving. As his paper makes clear, there's no evidence that demand for air travel will level off as countries get richer — on the contrary, that seems to keep going up and up indefinitely.

Reasons for skepticism

Now, there are plenty of reasons to be skeptical about the "peak demand" scenarios. For one, as Brandt says, this is more of an exploration of what we currently know about travel demand and technology rather than a hard prediction. But the future won't necessarily resemble the past.

For one, it's possible that fuel efficiency won't improve as rapidly as these forecasts expect. As Chris Nelder has pointed out, stricter fuel-economy laws can help, but much depends on how quickly the cars on the road actually get upgraded. In the United States, the turnover rate for the 240 million vehicles in circulation has been fairly modest these past few years.

What's more, it's entirely possible to imagine technologies that could increase demand for driving in the decades ahead. If Google's autonomous vehicles make motoring around easier and more pleasant, people may do more of it. Likewise, Eric Jaffe of the Atlantic Cities recently discussed evidence that online delivery services could increase overall traffic — by allowing people to outsource their mundane errands and replace that time with other types of driving. Or perhaps the opposite will occur.

So there are a lot of variables at play here. And if the world can't find ways to curb its oil use on the demand side, then we're back to worrying about resource scarcity — and in that case, as Nelder argues here, high oil prices could do the job of cutting our demand.

Further reading:

— Peak oil isn't dead: An interview with Chris Nelder.

— The Economist explores what peak oil demand may mean for the oil supermajors.

— Will driverless cars solve our energy problems — or just create new ones?