Key Points:

Worried about "peak oil"? The International Energy Agency's annual report, The World Energy Outlook 2008, admits for the first time that "although global oil production in total is not expected to peak before 2030, production of conventional oil ... is projected to level off towards the end of the projection period".

When the Guardian's environmental columnist, George Monbiot, pressed IEA director Fatih Birol on that opaque phrase, the actual date turned out to be 2020.

The IEA's previous reports, which assured everyone that there was plenty of oil until 2030, were based on what Birol called "a global assumption about the world's oilfields": that the rate of decline in the output of existing oilfields was 3.7 per cent a year.

It turns out that production in the older fields is really falling at 6.7 per cent a year.

There are still some new oilfields coming into production, but this number means that the production of conventional oil - oil that you pump out of the ground or the seabed in the good old-fashioned way - will peak in 2020, 11 years from now.

Birol assumes, or rather pretends, that new production of "unconventional oil" (oil extracted from tar sands or oil shales) will allow total production to match demand for another decade until 2030, but this is sheer fantasy.

So what are we to make of this news? Monbiot uses Birol's admission to launch an impassioned appeal for the rapid development of non-oil alternative sources of energy.

That is obviously urgent if we are close to "peak oil", but this may not be as great a crisis as it seems. It may not be a bonanza for the oil-producing countries, either.

The IEA presumes that demand for oil will rise indefinitely, so the price of oil only gets higher after "peak oil", but in technology nothing is forever.

Set into the front doorstep of my house (and most other 19th century houses in London) is an iron contrivance called a boot-scraper.

It is a device for scraping the horseshit off your boots before coming into the house, and the iron blade is worn into a shallow curve by a half-century of use.

Nineteenth-century cities depended on horses to move people and goods around.

London in the 1890s had 11,000 horse-drawn taxis and several thousand buses, each of which required 12 horses a day.

Add all the private carriages and the tens of thousands of horse-drawn carts, wagons and drays delivering goods, and there were at least 100,000 horses on the streets of London every day - each producing an average of 10kg of manure. Two thousand tonnes of manure a day. There were flies everywhere, and if you didn't shovel the manure up quickly, it dried up and blew into your eyes, your hair, your nose, your clothes.

As the cities grew, even more horses were needed and the problem grew steadily worse.

One writer in the Times in 1894 estimated that in 50 years the streets of London would be buried under 3m of manure.

In fact, within 35 years the streets of London were almost completely free of horses, and filled with automobiles instead.

They created a different kind of pollution, but at least you didn't step in it.

The same fate is likely to overtake oil-fuelled vehicles in the next 35 years.

The shift will be driven by concerns about foreign exchange costs and energy independence, and increasingly by the need to curb greenhouse gas emissions.

It is starting with ever-tightening standards for fuel efficiency.

That will be followed by the first mass-market generation of electric vehicles, due in the next two or three years.

The coup de grace will be delivered by third-generation biofuels, probably produced from algae that do not use valuable agricultural land, that are fully competitive with oil in price and energy content.

We will never get back the eight wasted years of the Bush Administration, and it may now be too late to avoid drastic climate change, but Barack Obama is clearly going to try. You do not appoint Steve Chu as your energy secretary and Carol Browner as your "climate tsarina" if you intend to evade the issue.

So American oil consumption is going to start falling quite fast, quite soon.

The same is true elsewhere. Indeed, it is a safe bet that the demand for oil is going to fall faster than the supply over the next 10 or 15 years, even if we are already at or near "peak oil", for the annual decline in oil production just after the peak is actually quite shallow - around 2 per cent - in the classic Hubbert curve.

And if demand falls faster than supply, the price will also collapse.

Ladies and gentlemen, place your bets ...

* Gwynne Dyer is a London-based independent journalist whose articles are published in 45 countries.