Ecocide Methodology

The data about oil and gas licences, provided by MapStand, is dated October 2025. As a consequence, the accuracy of our results will decrease over time, because the owners, operators and status of the licences (active, inactive, exploration, production) may change in the future.

Entries marked as “unknown” point to incomplete data in the MapStand database (for instance exploration or production status unknown).

We identified overlaps between oil and gas licences and protected areas in 99 countries. Due to inconsistencies and contradictory data from public oil and gas registers, licences in 16 countries (Canada, Japan, Iran, Kazakhstan, Pakistan, Romania, Turkey, Venezuela, China, Ecuador, Botswana, Zambia, South Sudan, Niger, Mali and Chad) were excluded from our analysis. In total, our analysis focused on 120 unique countries.

We could not fact-check every single data point on every licence (status, shore, type) and every protected area provided by MapStand and WPDA.

However, we fact-checked (and complemented when necessary) information on all licences operated by the top 10 operators in our ranking (see below,“Operators analysis”). We also fact-checked the 100 largest overlaps included in our dataset - which together covered more than 50% of the total overlap surface.

We identified and calculated the size (in km²) of the overlaps between protected areas and all types of upstream oil and gas licences. The licence size refers to the total area in which a government allows hydrocarbon companies to perform exploration or production activities. A licence area is never fully covered with infrastructure on the ground.

We could not assess the location and size of all oil and gas infrastructure within 11,792 unique overlaps, including 7,021 unique protected areas impacted.

The overlap size that we calculated does not represent the size of actual damage caused to the environment and biodiversity, as:

• In some exploration licences drilling and land clearing operations have not yet begun.
• In some production and/or exploration licences, activity on the ground is ongoing in the part of the licence which does not overlap with protected areas.
• When production/exploration is ongoing in a part of the licence overlapping with a protected area, the area actually covered by infrastructure on the ground (drilling pads, roads, pipelines, offshore platforms) is always smaller than the size of the overlap between the licence area and the protected area.

Thus, the overlaps listed in the “Fueled Ecocide” project represent, for most of the surface that we calculated, an actual or potential threat to biodiversity, and for a small part of this surface, an actual damage to biodiversity.

To assess the theoretical (or maximum) threat oil and gas activities could pose on protected nature, we conducted a global analysis of the overlaps between all types of oil and gas upstream licences (exploration, production and unknown) and protected areas.

Global figures include the total overlap (in km²) between licence size and protected area size and consider both production and exploration licences.This follows the standard methodology as suggested by experts and used in several similar studies.

This methodology has been used by UNESCO in its own reporting on natural resources extraction within World Heritage Sites and by the NGO Leave it in the ground (LINGO) in its report mapping fossil fuels licences across protected areas.

Experts, UNESCO and NGOs have chosen this approach because the IUCN and UNESCO say that there should be no oil and gas production or exploration in their protected areas, and because a licence overlapping with a protected area represents a potential or actual threat to biodiversity.

This methodology is suitable and recommended for a global overview of potential threats to protected areas worldwide, but it was not suited to identify the companies most responsible for encroaching on protected areas.

Operators analysis

To be able to identify the oil and gas companies responsible for operating on protected areas, we focused on production licences only. These production licences, which represent 11% of the total overlap size with protected areas (against 54% for exploration licences and 36% for licences listed as “unknown”), have allowed ongoing drilling and production activities which are most likely to negatively impact the environment and biodiversity.

We identified the ten operators with the largest area overlap, and manually fact-checked each of their licences with publicly available information using press articles, governmental geo-data and publications, national oil and gas registers and corporate filings.

As explained in the “Overlaps definition and methodology” section, the infrastructure on the ground (oil fields, access roads, pipelines) typically has a much smaller footprint than the overlap size, and is sometimes located in the part of the licence which does not overlap with the protected area.

Therefore, the overlap area corresponding to each operator represents an estimation of the protected area “at risk” because of the company’s oil and gas operations. Some of these protected areas may still remain relatively unimpacted by the company’s active production activities.

In order to determine the most affected countries, we analysed all types of oil and gas licences (production, exploration and unknown) overlapping with internationally recognized protected areas, which are the most precious on the planet: UNESCO World Heritage Sites and Biosphere Reserves, Indigenous Lands, Ramsar wetlands, IUCN areas categories I to VI, and Natura 2000 areas protected by European Union regulations. We chose to include IUCN areas I to VI because the IUCN discourages all industrial activity inside those areas.

As explained in the “Overlaps definition and methodology” section, infrastructure on the ground (oil fields, access roads, pipelines) typically has a much smaller footprint than the overlap size, and is sometimes located in the part of the licence which does not overlap with the protected area.

Therefore, the overlap area corresponding to each country represents an estimation of the protected area “at risk” because of oil and gas operations. Some of these protected areas may still remain relatively unimpacted by active oil and gas production activities.

The general analysis, operator-specific analysis and country-specific analysis, described above, focused on three main metrics:

1. the area of overlap (in km²) between oil and gas licences and protected areas
2. the number of unique protections impacted (i.e. allowing for duplicates where areas are protected by two or more protections)
3. the number of unique protected areas impacted (i.e. correcting for duplicate protections)

We generated a “master file” spreadsheet, as well as a global report and country-specific analysis in HTML format, in order to facilitate the work of reporters and allow them to dive deeper into specific data for a country or company.

The results are available in the result section presented on this website, as well as in publications linked in the publication page.

Before any data cleaning, all geospatial datasets were reprojected from WGS 84 (EPSG:4326), which expresses locations in degrees of latitude and longitude, to the Equal Earth Greenwich projection (EPSG:8857).

The Equal Earth projection is a pseudocylindrical equal-area map projection designed to show the world's landmasses in their true proportional sizes, avoiding the distortions of more common projections like Mercator.

Because it uses meters as units, it enables precise surface and area measurements while maintaining a visually balanced global representation. This made it suitable for calculating the actual size and overlap of extraction blocks across continents and seas.

Mapstand's geospatial dataset included information on hydrocarbon extraction blocks, such as operator, location, and licence type. While generally well structured, key fields were often incomplete or unreliable.

Here are the main issues that we identified:

• Lack of operator information.
• Unreliable or missing country values.
• Poor data quality for: Canada, Japan, Iran, Kazakhstan, Pakistan, Romania, Turkey, Venezuela, China, Ecuador, Botswana, Zambia, South Sudan, Niger, Mali and Chad.
• Several blocks overlapped partially or completely, an acknowledged Mapstand error requiring correction.

We downloaded (October 2025) geographic information on protected areas and other effective area-based conservation measures (OECMs) listed in the WDPA database from the Protected Planet website.

We found the WDPA's data to be mostly clean and accurate. We only used QGIS to fix some of the polygon geometries and apply the Equal Earth Greenwich projection (EPSG:8857) to our data. We also excluded a few protected areas that we found to be outdated.

Our aim was to produce a reliable, non-overlapping global dataset of active hydrocarbon licences. All cleaning was carried out in QGIS and included the following steps:

1. Keep only blocks with identified operators.
2. Filter for hydrocarbon-related licences with status Current or Pending Award.
3. Re-assign country values.
4. Clean overlapping polygons.
5. Exclude blocks located in countries with unreliable data.

Re-assigning country to oil and gas licences

We used the Flanders Marine Institute(VLIZ) dataset Union of World Country Boundaries and Exclusive Economic Zones (EEZs) (Version 4 - 2024-10-10 , methodology here), a verified global maritime and land borders dataset, to assign country values to each oil and gas licence. This dataset combines both land borders and maritime boundaries (EEZs), which define each country's rights at sea.

Blocks spanning more than one country were labelled as shared. In cases where Mapstand and the boundary dataset disagreed, we manually reviewed company information and source data to determine the most accurate country assignment.

Handling Overlapping Blocks

To avoid double-counting surface area and misleading totals, we removed overlaps between extraction blocks. Our aim was to identify true overlaps, understand why they occurred, and correct them while preserving licence information.

Correcting drawing imperfections

Many apparent overlaps were caused by small drawing errors along shared borders rather than true surface duplication. These were corrected by aligning nearby edges so that consecutive polygons met exactly. All nearby edges were snapped to each other to eliminate false overlaps. A tolerance of 200 metres was selected after testing smaller values.

After fixing geometries, remaining polygon interactions were classified into four categories:

1. Touches (acceptable): Blocks sharing borders only.
2. Containments (full overlaps): Smaller blocks fully contained within larger ones.
3. Partial overlaps (real overlaps): Blocks sharing part of their surface.
4. Self-intersections (duplicates): Identical blocks appearing more than once.

Each category required different computational or manual handling.

Identifying true overlaps

To isolate real surface overlaps, we used QGIS spatial joins:

1. We first counted all intersections (touches, partial overlaps, containments, and duplicates).
2. We then counted only border touches using the touches predicate.
3. Touches and duplicates were subtracted from total intersections, leaving only true overlaps.

Distinguishing between partial and total overlaps

To separate full containments from partial overlaps, we ran an additional spatial join using the within predicate.

Handling total overlap cases

In most containment cases, smaller blocks (often production licences) were nested within larger exploration or research blocks held by the same operator. To retain all licence information while removing surface overlap:

We subtracted the smaller blocks from the larger ones using the Difference tool. The resulting geometries were merged back together. Attributes of the larger blocks were preserved. Final geometries were validated to ensure no gaps, slivers, or topological errors remained.

Handling partial overlap cases

Partial overlaps were the most complex to resolve, as they did not follow a single pattern. Some overlaps were extremely small and likely caused by geometric imprecision rather than real duplication. This indicated that the initial geometry-snapping step could not eliminate all accidental overlaps, requiring additional criteria.

We decided to exclude overlaps where the overlapping area was smaller than 1 km² or covered less than 5% of the smaller block's area. This was a conservative filtering.

To apply the 1 km² and 5% thresholds, we first created a layer containing all pairs of blocks that genuinely shared surface area.

Starting from the partially overlapping blocks identified earlier, we intersected the layer with itself. This produced a row for every pair of blocks that overlapped in any way, including both directions (A∩B and B∩A) as well as self-intersections (A∩A).

Self-pairs were removed, and duplicate pairs were standardised using a combined, ordered pair identifier. We then aggregated results by pair identifier, producing one geometry per unique overlapping pair. This resulted in a clean table of all unique block pairs with real surface overlap.

Calculating overlap metrics

Before calculations, area fields stored as text in the original data were converted to numeric values. For each block pair, we then calculated overlapping area (km²), identified the smaller block in the pair, calculated the overlap ratio (share of the smaller block affected), calculated an overlap percentage (for review) and an intersection-over-union score (to compare blocks of different sizes).

Any pair was kept if the overlapping surface exceeded 1 km² or if the overlap covered more than 5% of the smaller polygon. Everything below these thresholds was classified as accidental geometry noise.

These significant overlaps were manually reviewed and resolved in the same way as the full containments: by subtracting the smaller surface from the larger one using the Difference tool, and then reintegrating the cleaned geometries back into the main dataset.

Creating the final dataset of oil and gas licences

After resolving all containments and partial overlaps, the cleaned geometries were merged back into a single unified layer. At this stage, no blocks overlapped, no area was double counted, and every polygon had a unique, reliable footprint. Combined with the earlier country-reassignment and filtering steps, this produced a final dataset in which all blocks had consistent geometry, correct attribution, and trustworthy surface measurements suitable for mapping, statistical analysis, and reporting. The final dataset remains in EPSG:8857 to preserve accurate area calculations.

We used PostGIS - an open source software program that adds support for geographic objects to the PostgreSQL object-relational database - to overlay oil and gas licences from Mapstand on protected areas from WDPA. We found this approach to be more replicable and less computer-intensive than QGIS.

We created a dedicated SQL server where we uploaded both our datasets. We then created geographic indexes for our data, and intersected the two tables based on whether their geometries overlapped with each other. As part of this process, we also confirmed that the two datasets used the correct projection (EPSG:8857) and only presented valid geometries.

For each overlap, we then added additional columns with the following calculations:

• The protected area in km².
• The area of the oil and gas licence in km².
• The area of overlap between the protected area and the oil and gas licence in km².
• The percentage of the protected area covered by the overlap area.
• The percentage of the oil and gas licence covered by the overlap area.

We also included the geometry of the overlap area between the protected area and the oil and gas licence.

We cleaned the resulting master file by running the following steps:

1. We created a new column with the clean category of protected area, by re-categorising each category of protected area into the following categories:
   • RAMSAR: Wetlands of international importance.
   • INDIGENOUS LAND: Territories managed by or for indigenous communities.
   • UNESCO SITES: World Heritage and Biosphere reserves.
   • IUCN I-VI: International standards for protected area management, from IUCN Ia and Ib to IUCN VI.
   • NATURA 2000: areas protected by member states of the European Union (EU) as per EU regulations regarding protected areas.
   • OTHER: Any designations not captured by the above logic.
2. We extracted, harmonised and cleaned the licence operator and licence ownership data. We also assigned the label UNKNOWN to any missing operator, owner, or source data.
3. We standardized licence status into three categories: EXPLORATION, PRODUCTION, or UNKNOWN.
4. We renamed all column headers and converted the text of categorical fields (Country, ISO codes, Area Names, Designations, etc.) to uppercase, to ensure consistency.

Finally, as a conservative step, we excluded all active licences that had an overlap with protected areas lower than 0.01 km² (= 10 000 m²).

Note on overlapping surfaces