The atlas of unburnable oil for supply-side climate policies

The Paris Agreement’s goals of limiting warming to well below 2 °C and pursuing 1.5 °C imply stringent limits on cumulative anthropogenic CO2 emissions (the remaining carbon budget). Current emission rates risk exhausting the 1.5 °C budget by around 2030, while the CO2 contained in global fossil fuel resources vastly exceeds this budget, making unabated exploitation incompatible with climate targets. This has spurred research and policy interest in supply-side climate measures and the concept of unburnable fuels. Existing debates on how to allocate which resources remain unextracted have focused on ethical equity principles versus economic efficiency. Prior global estimates indicate that large shares of fossil resources must remain unburned (e.g., ~71% of conventional oil by 2050), but there is little spatially explicit guidance on where not to extract. This study addresses that gap by proposing and applying socio-environmental spatial criteria to identify and prioritize conventional oil resources that should remain unextracted, thereby informing supply-side climate policy while maximizing co-benefits for biodiversity and human well-being.

The paper builds on work quantifying unextractable fossil fuels under climate constraints (e.g., Welsby et al. 2021; McGlade & Ekins 2015) and the growing literature on supply-side climate policy instruments (extraction taxes, moratoria, production quotas). Equity-focused scholarship argues that high-responsibility nations should forego extraction, while efficiency analyses prioritize leaving high-cost, low-rent resources unextracted. However, previous global spatial allocations have largely used cost-based criteria at coarse scales and have not incorporated socio-environmental priorities. Local and national initiatives (e.g., Yasuní-ITT, Costa Rican moratorium, bans in parts of Alaska, Belize, Mexico) illustrate the feasibility and benefits of supply constraints in sensitive areas. Extensive literature documents the biodiversity losses and public health and social harms associated with fossil extraction, reinforcing the rationale for socio-environmental criteria in selecting unburnable resources. The study thus extends prior work by integrating biodiversity hotspots, endemic species richness, protected areas, urban proximity, and Indigenous Peoples in voluntary isolation into a global, spatially explicit framework for prioritizing unburnable conventional oil.

Study scope focuses on conventional oil resources (oil, light tight oil, and natural gas liquids per Welsby et al.’s density threshold), using georeferenced, technically recoverable resource estimates rather than reserves, to guide exploration and long-term planning. Data sources include: USGS World Petroleum Assessment 2000 and 2012 plus 27 regional assessments (2012–2022) and USGS NOGA datasets for the USA; global biodiversity datasets (Critical Ecosystem Partnership Fund biodiversity hotspots; Jenkins et al. terrestrial and marine endemic species richness); protected areas (UNEP-WCMC WDPA); urban areas (MODIS 500 m global urban extent with 10 km buffer); Indigenous Peoples in voluntary isolation (RAISG for Amazonia and Survival International for Southeast Asia); and FAO rural population density (FGGD). Spatial unit of analysis is the USGS WPA 2012 assessment units within geologic provinces, plus USGS USA provinces. Shapefiles were rasterized to match the resolution of the criterion rasters; oil resource volumes per assessment unit were apportioned to pixels by area. CRS used an equal-area projection (Eckert IV). Analyses were conducted in QGIS 3.22.5 (with GRASS 8.2.1) and ArcGIS Desktop/Pro. Analytical framework: 1) Exclusion zones (Boolean overlay) identify top-priority areas strictly off-limits where oil resources coincide with any of: biodiversity hotspots; terrestrial or marine endemic species richness centers; protected areas; urban areas with a 10 km buffer; territories of Indigenous Peoples in voluntary isolation. 2) Conditional prioritization ranks additional areas needed beyond exclusions to meet the unburnable target using three scenarios: (a) social criterion—rural population density, selecting more densely populated areas first; (b) biological criterion—terrestrial endemic species richness; (c) biological criterion—marine endemic species richness. Assumptions: Based on Welsby et al., 71% of conventional oil resources are unextractable under a cost-optimal pathway for a 1.5 °C carbon budget. From a global estimate of 2575 Gbbl of conventional oil resources, 752 Gbbl could be extracted; applying this to the authors’ georeferenced total (2276 Gbbl) yields a required 1524 Gbbl to remain unburned. The study quantifies overlaps between georeferenced resources and socio-environmental criteria to identify the 609 Gbbl in strict exclusion zones, and then uses the ranking scenarios to designate an additional 915 Gbbl to meet the 1524 Gbbl target.

• Georeferenced global conventional oil resources total 2276 Gbbl, unevenly distributed, with major concentrations in the Middle East (648 Gbbl, 28%), USA (402 Gbbl, 18%), Russia and former Soviet states (343 Gbbl, 15%), Gulf of Guinea, offshore Brazil, North Africa, and the North Sea. About 2.8% (64.2 Gbbl) lie within tropical rainforests; 40% of the Amazon basin overlaps with oil resources.
• To align with a 1.5 °C pathway (following Welsby et al.), 752 Gbbl could be extracted; therefore 1524 Gbbl of the 2276 Gbbl must remain unburned.
• Exclusion zones—overlaps with top-priority socio-environmental areas—cover 29.5 million km2 and contain 608.8 Gbbl (26.8%) of conventional oil resources, distributed as: biodiversity hotspots 230.8 Gbbl (10.1%); terrestrial endemic richness centers 71.3 Gbbl (3.1%); marine endemic richness centers 89.0 Gbbl (3.9%); protected areas 185.8 Gbbl (8.2%); urban areas with 10 km buffer 145.7 Gbbl (6.4%); Indigenous Peoples in voluntary isolation 2.3 Gbbl (0.1%). Combining criteria yields 608.8 Gbbl in exclusion zones.
• Regional distribution of exclusion-zone unburnable oil (Gbbl and share of regional resources): Middle East 168.5 (26.0%); USA 107.7 (26.8%); Africa 83.3 (27.9%); Central and South America 117.8 (36.9%); Russia and former Soviet states 45.5 (13.3%); Europe 12.9 (16.6%); China and India 22.2 (25.2%); Other Developing Asia 42.6 (77.6%); Canada 0.6 (2.4%); Australia and other OECD Pacific 7.7 (37.5%).
• Since exclusion zones account for 609 Gbbl, an additional 915 Gbbl must be prioritized as unburnable. Using the social scenario (rural population density), selecting all oil resources coinciding with densities above 1 person/km2 beyond exclusions would meet the remaining 915 Gbbl. Complementary biodiversity-based scenarios similarly prioritize areas of highest terrestrial or marine endemic species richness.
• The approach demonstrates that all oil in exclusion zones can remain unextracted while still meeting the 1.5 °C constraint, with additional resources prioritized to reach the 1524 Gbbl target, thereby maximizing biodiversity and social co-benefits.

The study addresses the allocation problem of which oil resources should remain unextracted under stringent climate targets by offering spatially explicit, socio-environmentally grounded criteria. By first designating strict exclusion zones and then ranking additional areas using population and biodiversity metrics, the framework operationalizes supply-side climate policy in a manner that simultaneously advances biodiversity protection, public health, and Indigenous rights. The results show feasibility: even after protecting all oil within sensitive areas (609 Gbbl), sufficient additional resources can be left unburned to achieve the 1.5 °C-compatible total (1524 Gbbl). This spatial atlas can guide governments, firms, and investors in anticipating regulatory and social risks (e.g., stranded assets) while informing moratoria, phasedown strategies, and priority-setting for conservation-aligned supply constraints. The approach complements demand-side policies, offering a pragmatic pathway to reduce production consistent with carbon budgets while yielding collateral socio-environmental benefits.

The paper delivers a global atlas and methodology for prioritizing unburnable conventional oil based on socio-environmental criteria. It shows that oil resources overlapping biodiversity hotspots, endemic species richness centers, protected areas, urban buffers, and territories of Indigenous Peoples in voluntary isolation (609 Gbbl) can be set aside entirely, and additional resources can be ranked to reach the 1524 Gbbl unburnable target consistent with 1.5 °C pathways. The method is flexible and extendable to other fuels (gas, coal) and criteria (e.g., Arctic/ultra-deep risks, Indigenous consent), and can be refined with finer-resolution reserve/field data and techno-economic factors (prices, costs). Future research directions include integrating equity and techno-economic considerations with socio-environmental priorities, improving georeferenced datasets for reserves and recent extraction, and developing mechanisms to negotiate trade-offs among criteria for policy implementation.

• Data scope and uncertainty: Georeferenced conventional oil resource estimates (2276 Gbbl) differ from non-spatial global estimates (e.g., 2575 Gbbl) due to limited, uneven, and evolving data availability. LTO classification varies across institutions.
• Resources vs. reserves: The analysis uses technically recoverable resources rather than economically recoverable reserves; results are sensitive to future prices, technologies, and policy choices.
• Temporal gaps: Basin-level accounting does not deduct extraction (2000–2022) or fully incorporate discoveries shifting prospective to contingent resources over that period, introducing offsetting over/underestimation effects.
• Spatial resolution: Analyses are at assessment unit/province scale with pixel-based apportionment by area, not at field/reservoir level; this may misrepresent intra-basin heterogeneity relevant for local policy.
• Criteria integration: Marine and terrestrial endemic datasets were not combined due to normalization and taxa differences. The method assumes the Welsby et al. cost-optimal distribution across fuels (e.g., 71% of conventional oil unextractable by 2050).
• Social criteria data gaps: No global datasets on Indigenous Free, Prior and Informed Consent or environmental conflict likelihood, limiting inclusion of consent-based exclusion criteria.
• Economic factors: The analysis does not include basin-specific production costs and prices; using technically recoverable resources may overstate feasible extraction in some basins.