Governments and businesses will need to commit $2.4 to $3.4tn in carbon capture technologies by 2050 if climate change targets are to be met.
The International Energy Agency in its 'road map' for carbon capture said 3,400 projects to capture and store greenhouse gas and help reduce discharge from fossil fuels by 50% compared to 2005 levels will be required.
Apart from carbon capture devices, other measures to enhance energy efficiency include renewables and nuclear power.
The author of the report Tom Kerr said the next decade will be a litmus test for industry trialling such new technologies.
According to US Energy Secretary Steven Chu and UK Energy and Climate Change secretary Ed Miliband, both members of the Group of Eight countries, carbon capture technologies are already an important ingredient in their nation’s efforts to combat climate change.
They said the Group of Eight countries are keen to start 20 large-scale demonstration projects by 2010 in order to commercialise the technology by 2020.
Miliband said, however, that the climate change talks in Copenhagen in December this year will offer governments the chance to broker an international treaty on reduction techniques and hopefully endorse and spread the use of carbon-capture technology throughout the world.
The IEA's executive director Nobuo Tanaka said the organisation believes 100 large-scale projects are required by 2020, 850 by 2030 and 3,400 by 2050, and that OECD nations must lead the way with developing the technology to reach these targets.
“OECD countries must lead in the first decade but the technology must very quickly expand to the developing world, where the vast bulk of emissions growth will take place,” Tanaka said.
Kerr said the 100 large-scale projects should each capture a minimum of one million tons of CO2 annually and that the technology should be spread beyond power generators.
Consulting firm McKinsey and Co said that the technology in the demonstration phase costs nearly $133 (€90) to keep a metric ton of emissions from the atmosphere.