National carbon intensity figures are widely cited in Scope 2 reporting. France's, for example, sits around 42 gCO₂eq/kWh, driven by a dominant nuclear fleet. Germany's is several times higher, reflecting a larger share of fossil generation. These figures are real. They are also, for most electricity consumers, not the number that describes what they are actually consuming.
The electricity mix is not uniform across a country's territory. It varies by location, by hour, and by season, because electricity is produced and consumed in a physical grid where geography constrains what flows where. An organization that uses a national average to account for its electricity emissions is using a number that may significantly over- or understate the carbon content of the electricity it actually draws.
This article explains why.
The electricity grid is not a reservoir from which any consumer draws an identical mix. It is a network of transmission lines with physical capacity limits. When a solar farm produces electricity, that output reaches nearby consumers more easily than it reaches consumers hundreds of kilometers away.
This is the principle of deliverability: the ability of electrons generated at a given location to reach a given consumption point depends on network topology and available capacity. When congestion occurs, grid operators must manage imbalances, sometimes by curtailing renewable output in surplus zones, sometimes by dispatching more thermal generation in deficit zones.
The result is that two industrial consumers on the same national grid, located 600 km apart, may be implicitly consuming electricity from meaningfully different generation mixes at the same moment.
A national carbon intensity figure aggregates all generation across a territory and divides by total consumption. The calculation is straightforward, and the result is a useful macro indicator: for policy benchmarking, for country-level comparisons, for high-level Scope 2 disclosures using the location-based method.
But it collapses spatial information that exists in the grid and that affects actual emissions.
Consider two simplified cases:
Case 1: Renewable surplus zone. A region hosts significant solar capacity. Around midday, local generation exceeds local demand. In principle, surplus electricity could be exported to neighboring areas. In practice, this is often not possible: distribution networks were designed to deliver electricity from large central plants to end consumers, not to export decentralized generation upstream. As a result, grid operators must curtail the excess output, meaning they instruct generators to reduce production even though electricity could have been consumed somewhere else. This is curtailment, and it is becoming more frequent as decentralized solar capacity grows. In 2025, half of all solar installed capacity in Europe was connected to distribution networks rather than transmission networks. The carbon intensity of electricity in a surplus zone, during these periods, is lower than any national average would suggest.
Case 2: Import-dependent zone. A region lacks significant local generation. It depends on imports over long transmission paths. During high-demand periods, when lines are stressed, local operators may need to dispatch peaker units, typically gas turbines, to maintain voltage stability. Carbon intensity in that zone, at that moment, may be substantially higher than the national average.
Neither of these realities shows up in a national figure.
Several factors determine how carbon intensity varies geographically within a grid:
Local generation mix. The share of nuclear, gas, solar, and hydro in the regional generation portfolio is the primary driver. A region near a run-of-river hydro complex will have structurally lower intensity during wet seasons. A region relying on combined-cycle gas plants as its baseload source will have structurally higher intensity.
Decentralized generation. An increasing share of renewable capacity, particularly solar, is connected directly to distribution networks rather than to the high-voltage transmission grid. These generators cannot easily export surplus electricity beyond their local grid node. Their output is highly local by design, which makes the distribution network a relevant unit of analysis for carbon intensity, not just transmission-level zones.
Network congestion. Transmission constraints limit the ability of low-carbon electricity to flow from surplus to deficit zones. When a line is congested, the constrained zone must rely more heavily on local, often higher-carbon, generation. The European grid regularly experiences congestion between national zones and between internal bidding zones.
Time of day and season. Solar generation is zero at night and peaks around midday. Nuclear is dispatchable but not infinitely flexible. The local residual load, what must be served by dispatchable, often thermal, generation after solar output is subtracted, changes continuously and predictably within each day. Carbon intensity at the local level follows this pattern far more sharply than national averages suggest.
Grid losses. Transmitting electricity over long distances involves resistive losses. Higher losses mean more generation is required to deliver a given quantity of electricity. Local consumers, closer to generation sources, face lower effective losses, which is an implicit carbon benefit that national averages don't capture.
The GHG Protocol Scope 2 Guidance recognizes two methods for attributing electricity emissions:
The location-based method is where geographic granularity has the most direct impact. If an organization uses a national average emission factor, it is applying an approximation. If it uses a local emission factor that reflects its actual grid node, it is describing the carbon content of the electricity it physically consumed with more accuracy.
For organizations with large, fixed consumption sites, data centers, industrial facilities, manufacturing plants, the difference is not marginal. A data center located near significant solar capacity may have a carbon intensity well below the national average during midday hours. That gap translates directly into reported Scope 2 emissions if the accounting methodology captures it.
For organizations making procurement decisions, choosing when to run flexible compute workloads, when to charge storage, when to defer energy-intensive processes, local intensity signals provide actionable information that national averages cannot.
The practical challenge is that local carbon intensity data, at meaningful granularity, has historically been difficult to obtain. Grid operators publish national and sometimes zonal figures. They do not typically publish intensity estimates at the distribution substation level, the point where most commercial and industrial consumers actually connect.
Modeling local intensity requires combining several data sources: generation dispatch data (from TSO feeds and market platforms), network topology information, historical congestion patterns, and emission factors per generation technology. The output is not a single national figure but a spatially differentiated signal that varies by location and by hour.
This is the type of signal that makes it possible to answer: not "what is this country's carbon intensity right now" but "what is the carbon intensity of the electricity mix serving this substation, at this hour?"
The case for geographic granularity in carbon intensity data is not primarily about regulatory compliance. It is about accuracy.
An organization that reports its Scope 2 emissions using a national average is making an accounting choice that introduces error, in a direction that is unknown without local data. It may be overstating its emissions (if it is located in a low-carbon zone) or understating them (if it is located in a high-carbon zone). It has no way to know.
Local carbon intensity data replaces that uncertainty with a figure that corresponds to the physical reality of the electricity the organization consumed: where it was, when it was, and what was on the grid at that moment.
Carbon intensity is not a national characteristic of electricity. It is a local and temporal one. Organizations that account for it at that resolution are measuring something real.
