If we were starting afresh, we probably wouldn't chose to build an energy infrastructure based around fossil fuels. But like it or not, we are stuck with power stations, cars and homes that use carbon-based energy sources. The problem is that almost all these buildings and vehicles last a long time. If they stay in use, we are committed to large-scale future production of greenhouse gases. But how large? A new paper in Science by Dr Steve Davis and colleagues at Carnegie Institution of Washington in Stanford, California, gives us a clear estimate. Davis says that our existing energy infrastructure will put about 500 gigatonnes (Gt) of CO₂ into the atmosphere during the course of its life (this is about 15 times the world's annual emissions from all sources today).

The paper calculates this number by examining the number of power plants, motor vehicles and homes around the globe and estimating how long they will remain in use. The research team found that in the past, the average electricity-generating station lasted about 35 years before being demolished. Cars typically run for about 17 years before being scrapped, lorries and buses nearer 30. Since we know when all the power plants in the world were constructed and the average age of the planet's vehicles, Davis and his colleagues could estimate how much carbon dioxide will be emitted by existing infrastructure during the remainder of its life.

Put another 500Gt of CO₂ into the atmosphere between now and 2050, and the expected temperature rise will be about 0.5C of extra warming on top of what we have already seen. (Of course there is a very wide range to this forecast because of the uncertainties in the models of how temperature change is related to emissions). Davis and his colleagues make the point that if we stopped building new coal-fired power plants tomorrow and manufactured no new cars or trucks we would therefore keep warming well below the 2C increase which global scientists think is the maximum tolerable. Davis's climate models suggest that CO₂ concentrations in the atmosphere would rise to about 430 parts per million (ppm), a rise of about 40ppm on today's level and well below the 450ppm level that scientists often associate with 2C of warming.

That's the good news - today's energy infrastructure probably isn't enough, by itself, to topple us into wholly unmanageable climate change. The bad news is that this figure assumes that we build no fossil fuel power stations in the future and that all our new vehicles and homes are zero-carbon. That's not going to happen and the scale of the challenge is grimly indicated by the current rate of growth in low-carbon electricity. Of the 1,300 gigawatts of new power station capacity built since 2000, 31% uses coal, 34% gas and 4% oil. This leaves 2% nuclear and 17% renewables. And even this number substantially overestimates the share of future electricity production coming from renewables since both wind and solar power plants only produce a fraction of their maximum output. The wind and the sun aren't available all the time.

In a perspective in Science, Dr Marty Hoffert of New York University looks at how much energy we are likely to need to meet the world's requirements in future. Keeping the world's economy going requires continuously production of about 14,000 gigawatts of energy. That's equivalent to about 10,000 large-scale power plants. As the world economy grows, this is likely to rise to at least twice this level by 2050, even if we achieve major gains in the efficiency with which we use energy. So the challenge is to run down existing carbon-polluting energy sources rapidly and to replace them with atmosphere-friendly equivalents.

The scale of this task is immense. My rough calculation is that the world needs to ramp up its yearly rate of installation of low-carbon energy about 30-fold from today's levels within the next couple of decades.

A few wind turbines aren't going to be enough.