Notes and definitions

This year’s Energy Outlook attempts to account explicitly for the impact of climate change on ‎economic activity as well as the mitigation costs associated with decarbonizing the energy system. ‎There is considerable uncertainty in the economic and scientific literature as to how to model these ‎impacts and so any estimates of these effects, including those contained in the Outlook, are ‎imperfect and almost certainly incomplete. That said, we have judged that it is better to use the ‎research that is available than to make no attempt to include it in our analysis. ‎

The economic literature on climate change has traditionally quantified the relationship between ‎climate change effects and economic activity using climate-economy integrated assessment models ‎‎(IAMs). A more recent strand of empirical literature analyses the economic impact of climate change ‎based on estimates of how past changes in temperature in different parts of the world have affected ‎GDP. One of the benchmark studies of this literature, Burke et al. (2015) uses the IPCC ‎Representative Concentration Pathways (RCP) scenarios to assess the non-linear impact of ‎temperature changes on GDP across 166 countries. They find that GDP per capita is a concave ‎function of temperature, peaking at an annual average temperature of 13°C and declining strongly at ‎higher levels.‎

The illustrative estimates of the impact from climate change on GDP contained in the Outlook are ‎based on the models from Burke et al. The assumed temperature profiles implied by the three main ‎scenarios are based on the RCP scenario which most closely approximate the trajectories for carbon ‎emissions from energy use in each of the scenarios. For Rapid this is RCP 2.6; Net Zero – RCP 1.9; and ‎BAU – RCP 4.5. The economic impacts from these implied temperature increases are computed ‎relative to a counterfactual scenario, in which future temperatures are assumed to be held constant ‎at recent (1980-2010) average levels.‎

The median climatic change impacts derived using Burke’s methodology suggest that, for BAU, the ‎implied increase in global temperatures would decrease global GDP by close to 5% by 2050. The ‎estimated impacts for Rapid and Net Zero are somewhat smaller, reflecting the lower path of carbon ‎emissions. The regional impacts are distributed according to the evolution of their temperatures ‎relative to the concave function estimated by Burke et al. Regions that are already relatively warm ‎are likely to experience negative impacts on GDP, while colder regions could potentially benefit from ‎relatively warmer weather. ‎

These climate change impacts are hugely uncertain and incomplete as the Burke et al framework ‎focuses only on temperature changes on GDP, and does not incorporate other climate change ‎effects (such as rising sea levels, more frequent and stronger storms, floods, droughts or loss of ‎biodiversity) or other sources of economic disruption, such as large-scale human migration. ‎

The mitigation costs of actions to decarbonize the energy system are also very uncertain, with ‎significant variations across different external estimates. Most estimates, however, suggests that ‎these costs increase with the stringency of the mitigation effort, suggesting that they are likely to be ‎bigger in Rapid and Net Zero, than in BAU. Estimates published by the IPCC (AR5 – Chapter 6) ‎suggest that for scenarios consistent with keeping global temperatures increases to well below 2°C, ‎median estimates of mitigation costs range between 2-6% of global consumption by 2050. ‎

Given the huge range of uncertainty surrounding estimates of the economic impact of both climate ‎changes and mitigation, and the fact that all three of the main scenarios include both types of costs ‎to a greater or lesser extent, the Outlook is based on the illustrative assumption that these effects ‎reduce GDP in 2050 by around 5% in all three scenarios, relative to the counterfactual in which ‎temperatures are held constant at recent average levels. ‎

Importantly, if the scenarios were extrapolated beyond 2050, the Burke methodology would imply ‎GDP growth and prosperity in BAU would get progressively worse, leading to significantly lower ‎levels of well-being than in Rapid and Net Zero. ‎

References

• Burke, M., Hsiang, S. & Miguel, E. Global non-linear effect of temperature on economic production. ‎Nature 527, 235–239 (2015)‎
• The global aggregate mitigation cost estimates in terms of GDP losses are taken from IPCC AR5 – ‎Chapter 6.

The world’s scientific community has developed a number of “integrated assessment models” (IAMs) ‎that attempt to represent interactions between human systems (the economy, energy, agriculture) ‎and climate. They are “simplified, stylized, numerical approaches to represent enormously complex ‎physical and social systems” (Clarke 2014). These models have been used to generate a large number ‎of scenarios, exploring possible long-run trajectories for GHG emissions and climatic changes under a ‎wide range of assumptions.‎

The Intergovernmental Panel on Climate Change (IPCC) carries out regular surveys of this scenario ‎modelling as part of its assessment work. The most recent survey was carried out in support of the ‎‎2019 IPCC Special Report on Global Warming of 1.5°C (SR15). A total of 414 scenarios from 13 ‎different modelling frameworks were compiled and made available via an online portal.‎

Some of the scenarios are now quite dated and, in some cases, scenario results are already ‎significantly out of line with recent historical data and so were excluded from our analysis. From the ‎remaining model runs, we extracted 112 scenarios that were judged to be consistent with the Paris ‎Agreement Long Term Temperature Goal. They were further divided into two subsets: “well below ‎‎2°C” (69 scenarios); and “1.5°C with no or low overshoot” (43 scenarios). A more detailed note on ‎the scenario selection methodology is available at www.bp.com/energyoutlook . For each of these ‎two subsets of scenarios, the ranges of outcomes for key variables are described in terms of medians ‎and percentile distributions.‎

It is important to note that the scenario dataset represents “an ensemble of opportunity” – a ‎collection of scenarios that were available at the time of the IPCC survey and which were produced ‎for a variety of purposes. “It is not a random sampling of future possibilities of how the world ‎economy should decarbonise” (Gambhir et al, 2019). That means that the distributions of IPCC ‎scenarios cannot be interpreted as reliable indicators of likelihood of what might actually happen. ‎Rather, the distributions simply describe the characteristics of the scenarios contained in the IPCC ‎Report.‎

The sample ranges included in the section ‘Global energy system at net zero’, are ‎based on those IPCC scenarios in our sample which embody net carbon emissions from energy and ‎industrial use falling below 1 Gt before 2100. This was the case for 84 of the scenarios from the ‎sample of 112 scenarios. For each of these scenarios, the size and structure of the energy system is ‎considered at the point at which carbon emissions fall below the 1 Gt threshold. The earliest point at ‎which this ‘net zero’ state is reached in the sample of scenarios is around 2045 and the median ‎scenario in 2070. ‎

References 

• Clarke L. et al (2014). Assessing Transformation Pathways. In: Climate Change 2014: Mitigation of ‎Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the ‎Intergovernmental Panel on Climate Change
• Gambhir A. et al (2019). Energy system changes in 1.5 °C, well below 2 °C and 2 °C scenarios Energy ‎Strategy Reviews 23‎

This year’s Energy Outlook includes estimates of the investment requirements implied for each of ‎the three main scenarios for upstream oil and natural gas, renewables and carbon capture use and ‎storage (CCUS). ‎

Oil and gas investment

Upstream oil and natural gas capital expenditure (excluding operating costs) profiles in each of the ‎three main scenarios are calculated based on the investment required at an individual asset level to ‎meet the shortfall between estimates of the demand for oil and natural gas and a hypothetical ‎supply of oil and natural gas where no new investment is undertaken in new fields. Both the asset-‎level database and decline rates are derived from Rystad. The average base decline rate up to 2050 ‎for oil and natural gas is estimated at 4.3% p.a. and 4.8% p.a., respectively. When including fields ‎that have already been sanctioned, the decline rate is mitigated to 4.1% p.a. and 4.5% p.a., ‎respectively.‎

The hypothetical supply baselines for oil and natural gas assume that no new investment is ‎undertaken in new fields beginning with the 2020 supply baseline. It assumes that continuous ‎investment at producing and sanctioned fields takes place including infill wells and costs related to ‎maintaining the facility. Additionally, projects that have already been sanctioned (up to almost 7 ‎Mb/d by 2025 and 400 Bcm by 2027 for oil and natural gas, respectively) are assumed to be ‎completed in the next few years. ‎

A set of non-producing, unsanctioned, dispatchable assets needed to meet the oil and gas shortfalls ‎in our three main demand scenarios is defined. Based on Rystad data, the asset-level investments ‎required to bring those assets online are estimated. Finally, the capital spending on new assets is ‎added to the capex of the producing and under development assets. Investment in producing and ‎under development assets are assumed to be equal in all scenarios.‎

Investment in wind and solar energy, and CCUS

For wind and solar energy, the deployment rate of each technology in each scenario is estimated. ‎Investment costs are assigned to each based on their historical costs and learning curves. The ‎investment costs of solar and wind energy are broadly aligned with their historical learning curves, ‎around 8% for wind and 20% for solar.‎

For carbon capture and storage, the cost of investment in 2018 for different technologies – iron & ‎steel, cement, hydrogen, power sector, chemical sector, fertilizers – is taken from a range of sources. ‎It is assumed that the investment costs decline over time as a reflection of technology progress. The ‎annual investment cost reduction varies from a minimum of 1.3% to a maximum of 1.9%, depending ‎on the technology. ‎

The total investment in CCUS is based on deployment and costs. These also include variables costs ‎and, in particular, the cost of transportation and storage of carbon emissions which vary between ‎regions and over time.‎

References 

• Rystad Energy’s UCube global upstream database, August 2020‎

Extracting relevant scenario ranges consistent with the Paris Agreement goals

Review of existing methodologies

Scenarios are used by the private sector, international organizations and academia alike to explore how the future could evolve under an internally consistent set of assumptions. Full details are available here.

• Unless noted otherwise, data definitions are based on the BP Statistical Review of World ‎Energy

• All comparison data, including scenarios from IPCC and other energy forecasters, have been ‎rebased to be consistent with the bp Statistical Review

• Primary energy comprises commercially-traded fuels, and excludes traditional biomass ‎

• The primary energy values of nuclear, hydro and electricity from renewable sources have ‎been derived by calculating the equivalent amount of fossil fuel required to generate the ‎same volume of electricity in a thermal power station, the thermal efficiency assumption is ‎time varying, with the simplified assumption that efficiency will increase linearly to 45% by ‎‎2050. For more information see the Stats Review methodology

• Gross Domestic Product (GDP) is expressed in terms of real Purchasing Power Parity (PPP) at ‎‎2015 prices‎

• Transport includes energy used in road, marine, rail and aviation ‎

• Industry includes energy combusted in manufacturing; construction; the energy industry ‎including pipeline transport; and for transformation processes outside of power generation ‎

• Non-combusted includes fuel that is used as a feedstock to create materials such as ‎petrochemicals, lubricant and bitumen ‎

• Buildings includes energy used in residential and commercial building, plus agriculture, fishing ‎and IEA’s non-specified sector “Other” ‎

• Power includes inputs into power generation (including combined heat and power plants)

• OECD is approximated as North America plus Europe plus OECD Asia ‎

• China refers to the Chinese Mainland ‎

• Other Asia includes all countries and regions in non-OECD Asia excluding mainland China and ‎India

• Oil unless noted otherwise includes: crude (including shale oil and oil sands); natural gas liquids ‎‎(NGLs); gas-to-liquids (GTLs); coal-to-liquids (CTLs); condensates; and refinery gains ‎

• Liquids includes all of oil plus biofuels

• Renewables unless otherwise noted includes wind, solar, geothermal, biomass, biomethane and ‎biofuels and excludes large-scale hydro

• Non-fossils includes renewables, nuclear and hydro

• Hydrogen demand includes its consumption in transport, industry, buildings and power, but does ‎not include the demand as a non-combusted feedstock (such as use for fertilizer or methanol ‎production)‎

• Gas includes natural gas and biomethane

• References to carbon emissions consider only CO₂ emissions from fuel combustion ‎

• Plastics includes synthetic fibres

• Equinor: Renewal Scenario, Energy Perspectives 2019, June 2019‎

• IEA: Sustainable Development Scenario, International Energy Agency, World Energy Outlook ‎‎2019, Paris, France, November 2019 ‎

• IHS Markit: Accelerated Carbon Capture and Storage and Multitech Mitigation: IHS Markit 2020 ‎low emission cases, January 2020‎

• Shell: Sky Scenario, February 2018‎

• BP p.l.c., bp Statistical Review of World Energy, London, United Kingdom, June 2019‎
• International Energy Agency, Energy Balances of Non-OECD Countries, Paris, France, 2019‎
• International Energy Agency, Energy Balances of OECD Countries, Paris, France, 2019‎
• United Nations, Department of Economic and Social Affairs, Population Division (2019). World ‎Population Prospects 2019, Online Edition. Rev. 1‎