Life-cycle greenhouse gas emissions of energy sources

• 1. Introduction

  ›

  Life-cycle greenhouse gas emissions of energy sources asks a deceptively simple question: how much does a power plant actually warm the planet, from the moment miners pull its fuel from the ground to the day its last turbine shuts down? The answer is rarely what it seems. A solar panel produces no exhaust while it runs, but manufacturing it takes energy and materials. A nuclear plant emits almost nothing during operation, yet building it and mining uranium leaves a footprint. Getting a fair count means tracking every phase, from raw material extraction through construction, operation, and waste management. That full-life accounting is what scientists call a life-cycle assessment, and it generates a single number: grams of carbon dioxide equivalent per kilowatt hour of electricity produced. The Intergovernmental Panel on Climate Change pulled together the findings of hundreds of individual scientific papers in 2014 to produce the most authoritative comparison the world has seen. What it found draws a sharp line between a small group of very high emitters and a much larger cluster of low-carbon alternatives. But the data also carry important caveats about what gets left out of the count, how long plants are assumed to run, and what happens when the simplifying assumptions of laboratory analysis meet the messy realities of an actual power grid.

• 2. Coal at the Top

  ›

  Pulverized coal sits at the very top of the IPCC's 2014 emissions table, with a median life-cycle footprint of 820 grams of carbon dioxide equivalent per kilowatt hour. Its range runs from 740 at the low end to 910 at the high end, making it the worst-performing source in the entire comparison by a wide margin. Natural gas, assessed as a combined-cycle plant, comes in second at a median of 490 grams per kilowatt hour, with a range spanning 410 to 650. The gap between gas and coal is real but narrower than many expect. A 2019 study added another dimension to coal's burden by accounting for the time value of greenhouse gas emissions, a techno-economic angle that, when applied, considerably increases the life-cycle toll from carbon-intensive fuels. Some coal-fired stations operate for as long as 50 years, while others are shut down after 20 years or less, meaning the emissions embodied in building those plants are spread over very different amounts of generated electricity depending on the plant's actual lifespan. The UN Economic Commission for Europe published its own lifecycle analysis in 2021, and under that methodology hard coal using pulverized coal technology without carbon capture reaches as high as 1,000 grams per kilowatt hour. Add carbon capture and storage to the same plant design, and the figure drops dramatically, to around 370 grams per kilowatt hour, though that still leaves it far above any renewable or nuclear option.

• 3. The Low-Carbon Cluster

  ›

  Nuclear power and several renewable sources cluster together at the low end of the IPCC table in a way that often surprises people. Wind onshore records a median of 11 grams per kilowatt hour, wind offshore comes in at 12, and nuclear sits at the same 12 grams per kilowatt hour median. Solar photovoltaic at utility scale carries a median of 48 grams, while rooftop solar comes in slightly lower at 41 grams. Concentrated solar power registers a median of 27 grams. Hydropower's median of 24 grams per kilowatt hour looks clean on paper, but its maximum value reaches an extraordinary 2,200 grams, a reminder that some reservoirs, particularly in tropical regions, generate significant methane from decomposing organic matter. Geothermal power sits at a median of 38 grams, though some geothermal installations in Italy have been found to emit far more, and research into those cases was still ongoing in the 2020s. The UNECE 2021 data offer additional granularity: cadmium telluride photovoltaic panels installed on the ground clock in at just 12 grams per kilowatt hour, while polycrystalline silicon panels on the same ground-mounted configuration reach 37 grams. Nuclear power in the UNECE analysis averages 5.1 grams per kilowatt hour, and in June 2022 Electricite de France published a life-cycle assessment following ISO 14040 that placed the French nuclear fleet's 2019 footprint at less than 4 grams per kilowatt hour.

• 4. What Gets Left Out

  ›

  Life-cycle assessments are generally limited to the construction and operation phases of a power plant, which means some real-world emissions never enter the count. Decommissioning, the process of safely retiring a facility and returning its site to what engineers call greenfield status, is often excluded entirely, and the reasoning varies by technology. For large hydroelectric dams, dam removal is a rare enough practice that little practical data exists, so it is typically omitted. As dams age that is changing: dam removal is becoming more common, and the accounting gap is growing harder to justify. Larger structures like the Hoover Dam and the Three Gorges Dam are designed to last indefinitely with ongoing maintenance, a timeframe that simply cannot be quantified and therefore goes unaccounted. Nuclear is a notable exception to this pattern. The median value of 12 grams of carbon dioxide equivalent per kilowatt hour for nuclear fission that appears in the IPCC 2014 report traces back to a 2012 Yale University review, and that review explicitly includes the contribution of facility decommissioning as an added line item in the full nuclear life-cycle assessment. Thermal power plants raise a separate and subtler issue: even low-carbon stations burning biomass or using geothermal or nuclear energy add heat directly to the planet's energy balance. Wind turbines, for their part, can alter both horizontal and vertical atmospheric circulation. These effects may shift local temperatures slightly, but any influence on global temperature is undetectable against the far larger signal from greenhouse gases themselves.

• 5. Lifetimes and Cutoffs

  ›

  For wind, solar, and nuclear power, the emissions embodied in building the plant are fixed costs that get diluted across every kilowatt hour the plant ever produces. Run the plant longer, generate more electricity, and the per-unit footprint shrinks. That makes lifetime assumptions a consequential variable in the final number. Wind farms are generally estimated to last 30 years, after which any repowering effort would need to carry its own carbon accounting. Solar panels manufactured in the 2010s may have a similar 30-year horizon, but how long panels based on newer materials like perovskite will last is not yet known. Nuclear plants occupy a wider range: some are expected to operate for as long as 80 years, while others face early retirement for safety reasons. More than half the world's nuclear plants are expected to apply for license extensions, and there have been calls for those extensions to be more rigorously scrutinized under the Convention on Environmental Impact Assessment in a Transboundary Context. Studies that estimate lower emissions for a given technology often achieve that result by excluding parts of the life cycle from analysis, while studies at the high end of the range sometimes make unrealistic assumptions about the energy consumed in specific phases. The 2014 IPCC synthesis addressed this by harmonizing results across hundreds of papers rather than relying on any single methodology. One area where the data acknowledges its own limits is efficiency: improvements in technology since the time of publication are not incorporated. Total life-cycle emissions from wind power, for instance, may have fallen since the IPCC figures were compiled.

• 6. Ocean Energy and Open Questions

  ›

  Tidal and wave technologies occupy an unusual position in the IPCC table, listed as pre-commercial and assigned a median of 17 grams of carbon dioxide equivalent per kilowatt hour with a range of 5.6 to 28. Those figures come from a thin body of research, and the available studies have a recognized weakness: they tend to underestimate maintenance impacts, which in marine environments can be substantial. A later assessment examining around 180 ocean technologies found that the global warming potential of ocean energy spans 15 to 105 grams per kilowatt hour, with an average of 53 grams. That average is more than three times the IPCC median, reflecting how much the accounting changes when maintenance is modeled more carefully. A tentative preliminary study published in 2020 specifically examined subsea tidal kite technologies and placed their global warming potential between 15 and 37 grams per kilowatt hour, with a median of 23.8 grams. That median sits somewhat above the IPCC's earlier 17-gram figure for ocean energy as a whole. Biomass presents its own unresolved question at the frontier of the field: whether bioenergy paired with carbon capture and storage can achieve a carbon-neutral or even carbon-negative footprint is still being researched and remains controversial. That question matters because biomass with carbon capture appears in UNECE's 2021 analysis as a potential low-emissions pathway, and how the accounting handles biological carbon cycles will determine whether that pathway lives up to its promise.