A study by Christophe McGlade and Paul Ekins demonstrates that in order to limit global warming to 2oC above pre-industrial levels, a third of oil reserves, half of gas reserves and over 8o% of coal reserves should remain unused until at least 2050.

The study also finds that the outcome is broadly similar even if Carbon Capture and Storage (CCS) is widely deployed from 2025.

The data also breaks down where in the world it is most economically optimal to leave these untapped reserves, which could indicate to investors which particular fossil fuel companies and projects to invest in, divest from or engage with.

The authors used an integrated assessment model containing estimates of the types, quantities and geographical locations of existing fossil fuel reserves and resources to map what should remain unused until at least 2050 if there is to be a reasonable chance of keeping within the 2oC limit.

The authors don’t assume that other resources should automatically remain unused within the overall emissions limits, as they may be cheaper to produce than reserves and new resources may be developed at a particular time to maintain flow rates, without increasing the total amount used.

Regional distribution of unburnable reserves under 2 degree scenario, with and without CCS
 2oC with CCS2oC without CCS
FSU, the former Soviet Union countries; CSA, Central and South America; ODA, Other developing Asian countries; OECD, the Organisation for Economic Co-operation and Development. A barrel of oil is 0.159m3; %, Reserves unburnable before 2050 as a percentage of current reserves.
  Oil Gas Coal Oil Gas Coal
Country or regoinBillions of barrels%Trillions of cubic metres%Gt%Billions of barrels%Trillions of cubic metres%Gt%
Africa 23 21% 44 33% 28 85% 28 26% 4.4 34% 30 90%
Canada 39 74% 0.3 24% 5 76% 40 75% 0.3 24% 5.4 82%
China and India 9 25% 2.9 63% 180 66% 9 25% 2.5 53% 207 77%
FSU 27 18% 31 50% 203 94% 28 19% 36 59% 209 97%
CSA 58 39% 4.8 53% 8 51% 63 42% 5 56% 11 73%
Europe 5 20% 0.6 11% 65 78% 5.3 21% 0.3 6% 74 89%
Middle East 263 38% 46 61% 3.4 99% 264 38% 47 61% 3.4 99%
OECD Pacific 2.1 37% 2.2 56% 83 93% 2.7 46% 2 51% 85 95%
ODA 2 9% 2.2 24% 10 34% 2.8 12% 2.1 22% 17 60%
United States of America 2.8 6% 0.3 4% 235 92% 4.6 9% 0.5 6% 245 95%


The authors calculated the economically-optimal fossil fuel use solution, based on the estimated production costs of each of the oil and gas resource categories within each country.

They used data from previous models that quantify the volumes of oil, gas and coal produced globally under a range of future emissions trajectories. The different greenhouse gas emissions profiles from each fuel were then converted to approximate temperature rise trajectories (using the MAGICC climate model) and mapped against cumulative production from 2010 to 2050. From over 1,000 scenarios, 379 resulted in a temperature rise less than 2oC.

To test assumptions, the authors constructed further sensitivity scenarios (using the TIAM-UCL integrated assessment model) spanning a broad range of assumptions on production costs, demand projections, technology availability, etc.

They then compared total production with the projections from their scenario that remain within 2oC. Other regional distributions of unburnable reserves are possible while still remaining within the 2oC limit, but would have a higher impact on social welfare.


Over 430 barrels of oil and 95 trillion cubic metres of gas currently classified as reserves should remain unburned until at least 2050 (even when CCS is available). The Middle East, although using over 60% of its oil reserves, carries over half of the unburnable oil globally, and would need to leave over 260 billion barrels in the ground. The Middle East also holds half of unburnable global gas reserves, with Former Soviet Union countries accounting for another third: they would only be able to use half their current reserves.

Coal is by far the least-used fossil fuel under the 2oC scenario, with 82% of global reserves remaining unburned before 2050. The USA and Former Soviet Union countries would each use less than 10% of their current reserves, meaning over 200 billion tonnes of coal reserves remaining unburned. Coal reserve use is 25% higher in China and India, but they still should also leave nearly 200Gt of their current coal reserves unburned.

For oil, in Canada, while in situ production of natural bitumen can continue, because it’s a carbonintensive form of oil, the energy required to produce it must be rapidly and totally decarbonised. This occurs particularly through the use of CCS, so without CCS, all bitumen production ceases by 2040. Similar results are seen for Venezuelan extra-heavy oil, where 99% resources are deemed unburnable, even with CCS, meaning any increase in unconventional oil production at all is likely incommensurate with efforts to limit global warming to 2oC.

Due to the expense of CCS, its relatively late introduction (2025) and the assumed maximum rate at which it can be built, it has a modest effect in this model: burnable coal reserves are only 6% higher with CCS, and oil and gas just 2% higher.

In the Arctic Circle, the authors estimate there to be 100 billion barrels of oil and 35 trillion cubic metres of gas that were not being produced as of 2010. As neither scenario models using these resources as part of the most economically efficient global balance, the authors believe that all Arctic resources should be classified as unburnable.

Download the issue

  • Download report

    RI Quarterly vol. 9: Investing in a low-carbon world

    December 2015