Grid Electricity Emission Factors by Country (Ember 2025) — Complete Reference Dataset

What grid emission factors measure

A grid emission factor is the average mass of CO₂-equivalent greenhouse gases released per unit of electrical energy delivered through a national or regional power network. Expressed as kilograms CO₂e per kilowatt-hour (kg CO₂e/kWh) or grams CO₂e per kilowatt-hour (g CO₂e/kWh), it is the multiplicative bridge between a meter reading on an organisation’s electricity bill and a Scope 2 line item in its greenhouse gas inventory. Multiply the kilowatt-hours consumed by the grid’s emission factor and you have the location-based Scope 2 emissions attributable to that consumption.

The factor is constructed from the total CO₂e emissions released by all power generation feeding the grid in a given year, divided by the total electrical energy delivered to end consumers. Hydroelectric, nuclear, wind, solar, and geothermal generation contribute zero direct emissions to the numerator; coal, lignite, oil, and gas-fired generation contribute very large quantities; biomass contributes a small amount under standard accounting conventions. The mix of these technologies — combined with transmission and distribution losses on the way to the consumer — determines the grid’s average carbon intensity.

Grid factors update annually. Each year, Ember refreshes its Yearly Electricity Data, reflecting the prior calendar year’s actual generation mix. National regulators publish on parallel cycles: DEFRA (now DESNZ) releases UK conversion factors each June; the US EPA refreshes eGRID approximately every two years. A grid factor that was correct for FY2024 reporting is not the correct factor for FY2025 reporting, even when the country has not changed its underlying energy policy. The factor changes because the actual generation mix changes — wind output varies year to year, hydro reservoirs fill and drain, coal plants retire, gas plants ramp.

Location-based vs market-based — the Scope 2 distinction that matters

The GHG Protocol Scope 2 Guidance, published in 2015 as an amendment to the Corporate Standard, established a dual-reporting framework that almost every modern climate disclosure regime now codifies. Filers compute their Scope 2 footprint using two parallel methods and disclose both. The values on this page support the first method only; the second method requires data this dataset does not contain.

Both methodologies produce defensible Scope 2 totals, and both are required for most international filers. The location-based total tells you what the grid actually emitted to deliver the electricity you consumed, regardless of any contractual claim you have made about that electricity. The market-based total tells you the emissions you have contracted to be associated with your consumption, which can be lower than the grid average if you have purchased renewable supply, and can be higher than the grid average if your supplier’s residual mix is dirtier than the national average.

Complete dataset — all countries, all regions

All values: kilograms CO₂-equivalent per kilowatt-hour (kg CO₂e/kWh), location-based Scope 2 methodology. Source column distinguishes Ember 2025 (38 countries), DEFRA 2025 (United Kingdom only), and EPA eGRID 2023 (United States national average only). See §5 for why three sources are used and how they differ.

Europe

North America

Asia-Pacific

Middle East & Africa

Latin America

Bold green column (Factor) = the kg CO₂e/kWh value for direct multiplication against kilowatt-hour consumption to obtain location-based Scope 2 emissions in kg CO₂e. Source indicates the authoritative publication from which the figure is drawn — see §5 for why three different sources are required to cover the dataset and how the methodologies differ. Highlighted rows mark grids below 0.10 (clean, green tint) and above 0.70 (carbon-intensive, amber tint) for rapid orientation.

Notable factors — the grid story behind the numbers

Several values in the dataset deserve practitioner-level commentary because the headline figure obscures meaningful structural detail. The following narratives are not editorial colour — they are the context a verifier or sustainability analyst needs to interpret what a corporate Scope 2 disclosure actually means against a country baseline.

France — 0.041 kg CO₂e/kWh

France’s structurally low grid factor reflects roughly seventy per cent nuclear generation as the baseload, supplemented by hydropower in the south and east and a growing wind and solar contribution. The factor moves modestly between years primarily as a function of nuclear fleet availability — when reactor outages cluster, gas-fired generation fills the gap and the factor rises measurably. A French data centre or office sits in the bottom decile of European grid intensity by virtue of physical location alone, before any procurement decisions are made.

Norway — 0.028 kg CO₂e/kWh

The lowest grid factor in the dataset. Norwegian electricity is approximately ninety-nine per cent hydropower, with the remaining fraction split between wind and a small thermal contribution for grid stability. Two practitioner cautions apply. First, Norway is a net exporter of electricity to neighbouring grids; the location-based factor describes physical generation feeding consumption inside Norway, not the carbon content of electricity that may transit elsewhere. Second, organisations operating across the Nordic synchronous grid should be aware that Sweden, Norway, Finland, and Denmark price-couple under Nord Pool but report distinct location-based factors — apply the country factor to country consumption, never blend.

Poland — 0.589 kg CO₂e/kWh

Poland’s grid remains coal-heavy by structure: lignite and hard coal contribute the majority of generation, with gas, wind, and a fast-growing solar PV fleet making up the remainder. The factor is declining year over year as coal capacity retires and renewable capacity scales, but the trajectory from 2020s baseline to a target consistent with EU 2030 reductions is steep. For corporate filers, a Polish facility appears as a high-intensity Scope 2 line by structure, not by management failure — the operational lever for improvement is contractual (PPAs, market-based renewable supply), not facility-level efficiency alone.

South Africa — 0.699 kg CO₂e/kWh

The highest factor in the dataset. South Africa’s Eskom-dominated grid is approximately eighty per cent coal, with the remainder split between nuclear (Koeberg), pumped hydro, and a recently expanding wind and solar fleet. The factor has been broadly stable for several years; structural change requires both new generation capacity and resolution of the long-running grid constraint and load-shedding regime. Multinationals with South African operations face a structural Scope 2 burden that disclosure alone cannot reduce; the legitimate decarbonisation path is wheeling renewable supply under the country’s renewable independent power producer programme or signing corporate PPAs with new-build solar and wind.

United Kingdom — 0.177 kg CO₂e/kWh (DEFRA 2025)

The DEFRA 2025 release (DESNZ, June 2025) carries a 0.177 kg CO₂e/kWh location-based factor for grid-supplied electricity, a fall of approximately fifteen per cent from the 0.207 figure in DEFRA 2024. The drop reflects continued coal retirement (the UK’s last coal-fired power station closed in September 2024), increased wind output across the reporting year, and a higher contribution from imports across the cross-Channel and North Sea interconnectors. The DEFRA 2025 factor uses an AR5 GWP basis — see §5 for why this matters less for grid factors than for fuel combustion factors. UK SECR-regulated entities must use the DEFRA value rather than the Ember value; this is a regulatory mandate, not a methodology preference.

China — 0.526 kg CO₂e/kWh

China’s grid factor is on a structural decline trajectory as enormous solar PV and wind capacity additions are absorbed into the generation mix. The 2026 value reflects the actual delivered mix in the prior calendar year, including coal (still the majority), gas, hydro, nuclear, wind, and solar. Practitioners should note that this single national grid-average factor masks substantial regional variation across China’s seven dispatch regions; for site-level accuracy where available, regional grid factors are preferable, though they are not in this dataset.

India — 0.670 kg CO₂e/kWh

India’s grid sits in the carbon-intensive band, reflecting a coal-dominated baseload offset by rapid renewable additions. The factor is declining year over year, but coal’s share of installed capacity is still the structural anchor. For Indian operations of multinational corporates, the GHG Protocol Scope 2 dual-reporting expectation applies the same as in any other market; the Indian Renewable Energy Certificate (IREC) regime supports market-based reporting where instruments are retired against consumption.

Source methodology — three data sources explained

This dataset combines three authoritative publications because no single publisher provides the optimal source for every country in scope. The choice of source per country reflects regulatory mandate (UK SECR requires DEFRA), national data quality (US EPA eGRID is more granular than the corresponding Ember national figure), and geographic coverage (Ember covers the rest). The page is transparent about which source applies to which country precisely because the methodologies differ in important ways.

Ember Yearly Electricity Data (2025 release) — primary source for 38 countries

Ember — an independent global energy think tank — publishes its annual Yearly Electricity Data, compiling electricity generation by source for over two hundred geographies from multi-country datasets (the US EIA, Eurostat, the Energy Institute, the UN) and national statistics. Ember derives each country’s grid carbon intensity — grams of CO₂ per kilowatt-hour of generation — by applying standard per-technology emission factors to the reported generation mix. The dataset is refreshed through the year and the long-format yearly CSV is updated in the first quarter; the 2025 release covers the 2024 calendar year.

The Ember factors published on this page are grid-average carbon-intensity values, applied here as location-based Scope 2 factors. They are appropriate for Scope 2 location-based reporting under the GHG Protocol Scope 2 Guidance and for ESRS E1, CDP, SBTi, ISO 14064-1, and IFRS S2 disclosures. Ember publishes its data under a CC BY 4.0 licence, which permits reuse — including commercial and corporate reporting use — with attribution. Like the DEFRA and EPA grid factors on this page, these are generation-based (production) intensities: applied to metered consumption they give the location-based Scope 2 figure, with transmission and distribution losses accounted separately under Scope 3 Category 3.

DEFRA 2025 — primary source for the United Kingdom

The UK Department for Energy Security and Net Zero (DESNZ, formerly BEIS) publishes annual greenhouse gas conversion factors for company reporting, branded as “DEFRA factors” by long-standing convention. The DEFRA 2025 release (June 2025) carries a UK grid factor of 0.177 kg CO₂e/kWh, a fifteen per cent reduction from DEFRA 2024’s 0.207 figure. The reduction is driven by the closure of the last UK coal-fired power station, increased wind output, and shifts in interconnector flows.

UK organisations subject to the Streamlined Energy and Carbon Reporting (SECR) regulations and the Energy Savings Opportunity Scheme (ESOS) must use DEFRA factors for compliance reporting. This is a regulatory mandate. UK reporters who use Ember factors in place of DEFRA factors are technically non-compliant with the SECR methodology requirement, even if the Ember value happens to be similar. See the DEFRA standards page for the regulatory framing, and the DEFRA dataset reference page for the broader DEFRA factor set.

DEFRA’s pre-aggregated CO₂e factors embed AR5 GWP-100 values, per the DEFRA 2025 Introduction tab, Row 35. This differs from Ember’s approach, which derives carbon intensity from each country’s generation mix.

EPA eGRID 2023 — primary source for the United States

The US Environmental Protection Agency publishes the Emissions & Generation Resource Integrated Database (eGRID), the authoritative US-specific source for grid emission factors. eGRID is published annually each January, roughly a two-year lag from the data year; the current edition is US EPA eGRID 2023 rev2 (data year 2023, published 2025). The US national average value of 0.350 kg CO₂e/kWh published here is the eGRID 2023 national figure.

The national average masks very substantial variation across the US grid. eGRID publishes factors at multiple scales: NERC region (eight regions), eGRID subregion (twenty-six subregions), state, and balancing authority. For US operations where geographic location is known and the consumption is not spread randomly across the country, the eGRID subregion factor is materially more accurate than the national average and is the preferred basis for Scope 2 reporting where available. See §9 below.

Why GWP basis matters less here than for fuel combustion factors

An important methodological distinction: grid emission factors are dominated by CO₂ from power generation, with much smaller methane and nitrous oxide components. Fuel combustion factors at the source level (a kilogram of natural gas burned) carry meaningful CH₄ and N₂O contributions whose relative weight in the CO₂e total depends on the GWP version applied. For grid emission factors at the consumption level, the CO₂ component dominates the total to the point where the AR4-vs-AR5-vs-AR6 GWP choice produces a materiality of typically less than one per cent on the final factor.

This is why Ember, DEFRA, and EPA can co-exist on this dataset without producing methodologically incoherent totals when applied to a multi-country footprint. Each source uses a slightly different GWP basis for the small non-CO₂ component (Ember applies per-technology factors to the generation mix; DEFRA 2025 uses AR5 explicitly; EPA uses a mix of AR4 and AR5 depending on the gas), but the total is dominated by the CO₂ portion in every case. For corporate Scope 2 reporting, the practical implication is that mixing factors from these three sources across a multi-country footprint is acceptable and is the standard approach. See the IPCC AR6 GWP page for the broader GWP context.

Year-over-year changes — 2025 to 2026 delta

Grid factors update annually. The most material changes between the 2025 and 2026 reference cycles are summarised below. A material change is defined here as greater than five per cent on the headline factor; below that threshold, year-on-year movement is generally within normal grid mix variation and does not warrant restatement of prior-year disclosures unless the inventory total exceeds the GHG Protocol significance threshold.

Headline movements

The United Kingdom posted the largest material reduction in the 2025 cycle, from 0.207 (DEFRA 2024) to 0.177 (DEFRA 2025), a fifteen per cent fall. The driver mix is well documented in the DEFRA 2025 methodology annex: full coal closure, increased wind contribution to the generation total, and a higher import share via the cross-Channel and North Sea interconnectors. For UK SECR filers, the DEFRA 2025 factor is mandatory for FY2025 reporting; FY2024 disclosures should not be restated unless the change crosses the GHG Protocol materiality threshold for the inventory total.

China’s headline factor continues to decline as the world’s largest renewable capacity rollout absorbs into the generation mix, with the 2026 value of 0.526 representing a roughly four per cent year-on-year fall. Indonesia and India remain at the carbon-intensive end of the dataset, with India’s modest year-on-year decline reflecting coal-plus-renewables additions rather than coal retirement. South Africa remains broadly stable.

General trend

The OECD-wide grid average is on a downward trajectory, driven by coal retirement in Europe, North America, and parts of East Asia, and by renewable capacity additions across all major OECD economies. Emerging-market grids show a mixed picture: some declining (China, parts of South Asia) as renewables scale, others stable or rising as electrification outpaces clean capacity additions. The next Ember Yearly Electricity release (2026, covering CY2025) will provide the next benchmark; GreenCalculus will update the MasterBrain and this page within thirty days of its publication.

Application — the Scope 2 calculation formula

The base conversion formula

The same 500,000 kWh consumption produces 248.5 tCO₂e in Singapore and 349.5 tCO₂e in South Africa — a 1.4-fold difference driven entirely by the underlying generation mix, before any operational decisions about energy efficiency or procurement.

Transmission and distribution losses for Scope 3 Category 3

The grid factors published here are end-use factors: emissions per kWh of electricity delivered to the consumer’s meter. They do not include the upstream transmission and distribution (T&D) losses that occur between the generator and the consumer. Those losses are typically four to ten per cent depending on the country, and they are accounted under GHG Protocol Scope 3 Category 3 (fuel- and energy-related activities not included in Scope 1 or 2).

Practitioners must not substitute the Scope 2 grid factor directly into a Scope 3 Category 3 line. The Scope 3 line requires a separate T&D loss factor, typically published by the same regulator (DEFRA, EPA) alongside the headline grid factor. Using the headline grid factor as the Scope 3 Category 3 figure double-counts the consumer’s delivered electricity emissions.

Multi-country operations — apply per country, never average

Organisations operating in multiple countries must apply the country-specific factor to country-specific consumption, then sum the resulting tCO₂e values. Computing a weighted-average factor across the portfolio and applying it to total consumption produces a different and incorrect total. The error is small when consumption is concentrated in one country and grows with portfolio diversity; for organisations operating across the clean-grid–dirty-grid spectrum, the difference between correct and averaged calculations can exceed twenty per cent.

Framework requirements — which Scope 2 approach each standard mandates

The GHG Protocol Scope 2 dual-reporting framework is now embedded in most major climate disclosure regimes, but the specifics — which methodology is required, which is optional, and which national source is mandatory for which jurisdiction — vary. The table below documents the position as of May 2026.

UK SECR is the only framework on this list that explicitly mandates a particular national source — DEFRA — rather than accepting any authoritative published value. For multinational filers reporting under multiple frameworks simultaneously, the UK consumption portion must be computed using DEFRA factors regardless of what factor is used for the rest of the portfolio. This is one of the few hard methodology constraints in the cross-framework reporting space.

Subgrid and country subdivision considerations

National grid factors are appropriate for many corporate inventory contexts and are required in many regulatory contexts. They are not, however, the most accurate basis where consumption is concentrated in a known subregion of a large, electrically-heterogeneous country. Three jurisdictions in this dataset have material subnational grid variation and warrant explicit treatment.

United States — eGRID subregions

The US national average of 0.350 kg CO₂e/kWh published here masks a roughly six-fold range across EPA eGRID subregions. The Western Electricity Coordinating Council (WECC) subregions — particularly the California ISO area — sit substantially below the national average due to the California renewable portfolio standard, hydropower from the Pacific Northwest, and rapidly growing solar PV. The SERC subregions covering parts of the Southeast sit above the national average due to coal and gas dominance. The RFC region (much of the industrial Midwest and Mid-Atlantic) sits near the national average. For US operations where the facility’s eGRID subregion is known, the subregion factor is the preferred basis for Scope 2 reporting; the national average should be used only when the boundary spans subregions or when the more granular figure is unavailable.

Australia — National Electricity Market and Western states

Australia’s 0.525 national average reflects three structurally distinct grids: the National Electricity Market (NEM) covering the eastern and southern states, the Wholesale Electricity Market (WEM) covering Western Australia’s South West Interconnected System (SWIS), and the various non-interconnected systems in the Northern Territory and the WA North West. State-level factors published by the Australian Government’s Department of Climate Change, Energy, the Environment and Water (DCCEEW) under the National Greenhouse Accounts Factors are the appropriate granular source for Australian operations.

Canada — provincial variation

Canada’s 0.191 national average masks very large provincial variation. Quebec, British Columbia, Manitoba, and Newfoundland and Labrador have hydropower-dominated grids with very low factors comparable to Norway. Alberta and Saskatchewan have coal-and-gas-dominated grids with factors several times higher than the national average. Ontario sits in between. For Canadian operations, the Environment and Climate Change Canada provincial factors are preferred to the single national-average figure where the provincial location is known.

Common reporting errors

Methodology, boundaries, and uncertainty

Rounding policy

Ember, DEFRA, and EPA publish at three decimal places of precision (0.XXX kg CO₂e/kWh). GreenCalculus publishes at three decimal places on this page, matching the default precision of the live — shortcode and the Scope 2 calculators. The MasterBrain stores higher internal precision and rounds to three decimal places for display; trailing zeroes (for example “0.300”) are significant.

Scope boundary — what is and is not included

The factors on this page are end-use, location-based, grid-average factors for grid-connected electricity consumption. They include direct CO₂ emissions from power generation feeding the grid, weighted by the actual generation mix, plus the small contribution of CH₄ and N₂O from combustion processes converted to CO₂e via the source’s chosen GWP basis. They do not include:

• Transmission and distribution losses — accounted separately under Scope 3 Category 3 with a country-specific loss factor.
• On-site generation — solar PV, wind, diesel back-up, combined heat and power on-site is Scope 1, not Scope 2, and uses fuel-specific emission factors.
• Purchased steam, heat, and cooling — covered by separate factors for each energy carrier; not captured by the electricity grid factor.
• Upstream fuel emissions for the generation fleet — accounted under Scope 3 Category 3 (well-to-tank).

Uncertainty

Grid carbon-intensity factors carry country-level uncertainty driven by the quality of national generation reporting; ranges of approximately five to fifteen per cent are typical, with wider ranges in countries where generation reporting is incomplete or where significant unmonitored off-grid generation operates. Because Ember derives its intensities from reported generation by source, the uncertainty tracks the completeness of each country’s underlying energy statistics. DEFRA does not publish formal uncertainty ranges with its UK factor but the methodology annex documents the input data sources. EPA eGRID publishes detailed methodology with uncertainty estimates by subregion.

For corporate Scope 2 reporting, point values are used; uncertainty propagates into the inventory total but is generally not the dominant source of inventory uncertainty. Activity data uncertainty — whether the kilowatt-hour figure on the bill accurately reflects consumption, and whether all sites have been captured in the inventory boundary — is usually the larger contributor in practice.

Market-based factors are not in this dataset

Market-based factors are supplier-specific or instrument-specific by definition and cannot be reproduced as a country-level reference set. The appropriate sources are the supplier’s own published emission factor (often included on the energy bill), or the residual mix factor for the market, published by the Association of Issuing Bodies (AIB) for European markets and by Green-e or equivalent registries for North American markets. Practitioners should request the supplier-specific factor in writing and retain the document as the audit object.

Geographic applicability

Grid factors are country-specific by construction (subgrid where applicable, per §9). They are not interchangeable across borders. A French operation using a German factor, or a Singapore operation using a Malaysian factor, is producing a methodologically incorrect Scope 2 disclosure even when the values are numerically close. Apply the country factor to country consumption, every time.

Provenance and implementation chain

Every value on this page traces to a published primary source. This section documents how those values flow from the underlying publications into the MasterBrain data layer, into calculator runtime, and out via the public Data Library API.

Data access — REST API and CSV

The 2026 grid emission factor dataset is available as a machine-readable REST API endpoint and as a flat CSV download. Both surface the same dataset shown in the regional tables above, sourced from Ember 2025, DEFRA 2025, and EPA eGRID 2023, and citable under the source-specific terms documented below.

Citation guidance

If you use this dataset in a published tool, report, or academic work, please cite the underlying primary source(s). The GreenCalculus implementation reference is optional but appreciated.

Ember Climate. Ember Yearly Electricity Data (2025 release). Available at: ember-energy.org/data/yearly-electricity-data. Licensed under CC BY 4.0.

Primary source for 38 of 40 countries

UK Department for Energy Security and Net Zero (2025). Greenhouse Gas Conversion Factors for Company Reporting 2025. London: DESNZ. Used here for United Kingdom only.

Primary source for GB

US Environmental Protection Agency (2024). Emissions & Generation Resource Integrated Database (eGRID) 2024. Washington, DC: EPA. Used here for United States national average only.

Primary source for US

Cite this dataset (GreenCalculus compilation). A versioned, machine-readable snapshot is permanently archived on Zenodo with a citable DOI:

Say, Jeremiah (2026). Grid electricity emission factors by country (machine-readable) (v2025.69). GreenCalculus. Zenodo. https://doi.org/10.5281/zenodo.20612874

GreenCalculus dataset DOI

What this API does not expose

The Data Library API exposes the location-based country-level factors only. Subgrid data (eGRID subregions, Australian states, Canadian provinces), market-based and residual mix factors, transmission and distribution loss factors, and historical year-on-year datasets are not exposed via the public API. The MasterBrain runtime layer carries some of these for internal calculator use; external API access requires a commercial arrangement.

Related GreenCalculus resources

Frequently asked questions

Version history