Long-duration electricity storage needs for coping with Dunkelflaute events in Europe

Wind droughts are more severe in summer than in winter.

“Wind droughts are more frequent and severe in summer than in winter33,44,47.”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceSupported
• cites [44] Ohlendorf, N. & Schill, W.-P. Frequency and duration of low-wind-power events in germany. Environ. Res. Lett. 15, 084045 (2020). sourceDisputed

  Section 3.2 says there is no clear seasonal trend in the most extreme low-wind-power events; under one event definition, the longest winter events even exceed those in summer. The paper finds summer events more frequent, but that does not establish the attributed claim that they are more severe. (LLM, conf=0.86)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured
• cites [47] Cannon, D., Brayshaw, D., Methven, J., Coker, P. & Lenaghan, D. Using reanalysis data to quantify extreme wind power generation statistics: a 33 year case study in Great Britain. Renew. Energy 75, 767–778 (2015). sourceSupported

Coping with prolonged periods of low availability of wind and solar power (Dunkelflaute) is a key challenge for realizing decarbonized energy systems based on renewable energy.

“Of particular concern are long-lasting extreme weather events, also referred to as “VRE droughts” or “Dunkelflaute”, which are characterized by an overall very low availability of VRE sources21–23.”

• cites [21] Raynaud, D., Hingray, B., François, B. & Creutin, J. D. Energy droughts from variable renewable energy sources in european climates. Renew. Energy 125, 578–589 (2018). sourceNot checked — no full text of the source
• cites [22] Kittel, M. & Schill, W.-P. Measuring the Dunkelflaute: how (not) to analyze variable renewable energy shortage. Environ. Res. Energy 1, 035007 (2024). sourceSupported
• cites [23] Van Der Most, L., Van Der Wiel, K., Benders, R. M. J., Gerbens-Leenes, P. W. & Bintanja, R. Temporally compounding energy droughts in European electricity systems with hydropower. Nat. Energy 9, 1474–1484 (2024). sourcePartly supported

  The Introduction describes energy droughts as periods of low renewable electricity production, and the Discussion and conclusions identifies prolonged shortages as a challenge for energy management. But the paper does not characterize these periods as necessarily involving low availability of both wind and solar power; its discussion of compounding conditions notes that some occur with increased solar radiation. No page numbers are available in our copy. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

With a rising share of VRE, the European power sector becomes increasingly exposed to weather variability.

“With a rising share of VRE, the European power sector becomes increasingly exposed to weather variability9,10.”

• cites [9] Staffell, I. & Pfenninger, S. The increasing impact of weather on electricity supply and demand. Energy 145, 65–78 (2018). sourceSupported
• cites [10] Collins, S., Deane, P., Ó Gallachóir, B., Pfenninger, S. & Staffell, I. Impacts of inter-annual wind and solar variations on the European power system. Joule 2, 2076–2090 (2018). sourceSupported

Experience with converting existing gas storage to hydrogen or building new underground hydrogen storage is so far limited.

“Yet, experience with converting existing gas storage to hydrogen or building new underground hydrogen storage is so far limited98.”

• cites [98] Talukdar, M., Blum, P., Heinemann, N. & Miocic, J. Techno-economic analysis of underground hydrogen storage in Europe. iScience 27, 108771 (2024). sourceSupported

From a techno-economic perspective, the production of hydrogen or its derivatives using renewable electricity, followed by storage and reconversion to electricity, currently appears most promising for enabling long-duration energy storage.

“From a techno-economic perspective, the production of hydrogen or its derivatives using renewable electricity, followed by storage and reconversion to electricity, currently appears most promising for enabling long-duration energy storage36,41–43.”

• cites [36] Dowling, J. A. et al. Role of long-duration energy storage in variable renewable electricity systems. Joule 4, 1907–1928 (2020). sourcePartly supported

  In “Technological Options for Long-Duration Storage” (page unavailable), the paper describes making hydrogen from electricity, storing it, and converting it back to electricity; it also discusses methane made from hydrogen as an alternative. “System Sensitivities to Region, Technology Mix, and Cost” (page unavailable) calls the hydrogen pathway competitive at current costs, but does not establish that hydrogen or its derivatives are the most promising long-duration storage option. (LLM, conf=0.87)

  Experimental — accuracy not yet measured
• cites [41] IEA. The future of hydrogen (IEA, 2019).Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites [42] Victoria, M., Zhu, K., Brown, T., Andresen, G. B. & Greiner, M. The role of storage technologies throughout the decarbonisation of the sector-coupled european energy system. Energy Convers. Manag. 201, 111977 (2019). sourcePartly supported

  The Conclusions report that hydrogen storage is preferred over batteries for smoothing multi-day wind fluctuations in the modelled European power system, and the Electric batteries section describes reconverting stored hydrogen to electricity with fuel cells. But the paper does not establish that hydrogen or its derivatives are the most promising long-duration option generally: its Other storage technologies section says several alternatives were excluded, including synthetic methane. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites [43] Hunter, C. A. et al. Techno-economic analysis of long-duration energy storage and flexible power generation technologies to support high-variable renewable energy grids. Joule 5, 2077–2101 (2021). sourcePartly supported

  The ‘Technologies to support high-variable renewable energy grids’ section describes renewable-powered electrolysis, hydrogen storage, and reconversion to electricity. The ‘Economic analysis of current and future technology costs’ section finds geologic hydrogen storage among the lowest-cost options for 120-hour storage under both current and future cost assumptions. It does not establish that hydrogen or its derivatives are the most promising option across long-duration storage generally: the finding is duration-specific, natural gas with carbon capture is also among the lowest-cost options, and the cited analysis concerns hydrogen rather than its derivatives. Page numbers are unavailable in our copy. (LLM, conf=0.85)

  Experimental — accuracy not yet measured

Long-duration storage is a central option for coping with VRE droughts because it can shift renewable surplus energy to periods of low wind and solar availability over long time scales, i.e., several days up to seasons.

“Particularly long-duration storage, sometimes also referred to as “ultra-long-duration storage”34, is a central option as it can shift renewable surplus energy to periods of low wind and solar availability over long time scales, i.e., several days up to seasons34–36.”

• cites [34] Brown, T. & Hampp, J. Ultra-long-duration energy storage anywhere: methanol with carbon cycling. Joule 7, 2414–2420 (2023). sourceSupported
• cites [35] Schill, W.-P. Electricity storage and the renewable energy transition. Joule 4, 2059–2064 (2020). sourcePartly supported

  The grid-scale storage section (p. 5) says storage shifts renewable surplus to periods of positive residual load, and the Conclusions (p. 12) call long-term storage a key element at very high renewable shares and say it can address seasonal mismatches. The paper does not establish the claim’s specific lower time scale of several days; it also qualifies the storage need by the availability of other firm low-carbon resources. (LLM, conf=0.82)

  Experimental — accuracy not yet measured
• cites [36] Dowling, J. A. et al. Role of long-duration energy storage in variable renewable electricity systems. Joule 4, 1907–1928 (2020). sourceSupported

The iterative procedure of VMBT overcomes shortcomings of previous research.

“The iterative procedure of VMBT overcomes shortcomings of previous research22.”

• cites [22] Kittel, M. & Schill, W.-P. Measuring the Dunkelflaute: how (not) to analyze variable renewable energy shortage. Environ. Res. Energy 1, 035007 (2024). sourceSupported

Wind droughts are more frequent in summer than in winter.

“Wind droughts are more frequent and severe in summer than in winter33,44,47.”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceSupported
• cites [44] Ohlendorf, N. & Schill, W.-P. Frequency and duration of low-wind-power events in germany. Environ. Res. Lett. 15, 084045 (2020). sourceSupported
• cites [47] Cannon, D., Brayshaw, D., Methven, J., Coker, P. & Lenaghan, D. Using reanalysis data to quantify extreme wind power generation statistics: a 33 year case study in Great Britain. Renew. Energy 75, 767–778 (2015). sourceSupported

Events identified by higher thresholds may last substantially longer, up to several weeks or months.

“In contrast, events identified by higher thresholds may last substantially longer, i.e., up to several weeks or months33.”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceSupported

The selection of weather years has a notable impact on least-cost long-duration storage use.

“Next, our analysis shows that the selection of weather years has a notable impact on least-cost long-duration storage use, which confirms previous research34,36,82,87.”

• cites [34] Brown, T. & Hampp, J. Ultra-long-duration energy storage anywhere: methanol with carbon cycling. Joule 7, 2414–2420 (2023). sourcePartly supported

  The passage on 71 years of weather data (labelled Abstract in our copy; no page given) says storage capacities were optimized across those years to account for interannual variability. The “Methanol for inter-annual storage” passage also describes different storage use across high- and low-wind years. The paper does not compare selections of weather years or establish how large their impact is on least-cost storage use. (LLM, conf=0.78)

  Experimental — accuracy not yet measured
• cites [36] Dowling, J. A. et al. Role of long-duration energy storage in variable renewable electricity systems. Joule 4, 1907–1928 (2020). sourceSupported
• cites [82] Diesing, P. et al. Exploring the demand for inter-annual storage for balancing wind energy variability in 100% renewable energy systems. Energy 312, 133572 (2024). sourceNot checked — no full text of the source
• cites [87] Gøtske, E. K., Andresen, G. B., Neumann, F. & Victoria, M. Designing a sector-coupled European energy system robust to 60 years of historical weather data. Nat. Commun. 15, 10680 (2024). sourceSupported

Interconnection among European countries can reduce least-cost long-duration energy storage capacity.

“Interconnection among European countries can reduce least-cost long-duration energy storage capacity.”

The weather year 1996/97 contains the most extreme pan-European drought in the data.

“The figure further shows the least-cost long-duration storage operation for the weather year 1996/97, in which we find the most extreme pan-European drought in the data (Supplementary Fig. 1).”

To reduce the build-out of capital-intense hydrogen gas turbines for covering peak residual load events, battery storage is sometimes used in our least-cost system.

“To reduce the build-out of capital-intense hydrogen gas turbines for covering peak residual load events, battery storage is sometimes used in our least-cost system.”

The correlation between the most extreme winter drought events and least-cost storage capacities differs between countries.

“The correlation between the most extreme winter drought events and least-cost storage capacities differs between countries.”

Extreme pan-European droughts affect several European countries simultaneously.

“Notably, several European countries are affected simultaneously33.”

• Also stated in: § Least-cost long-duration storage capacities decrease with geographical balancing · plain

  “This is in line with our renewable availability time series analysis, which finds that the most extreme pan-European drought affects many countries simultaneously (Supplementary Fig. 1).”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Countries that heavily rely on wind power, such as the UK, may face the most extreme renewable technology portfolio events also in summer, driven by wind droughts that are generally more pronounced in summer in Europe.

“Countries that heavily rely on wind power, such as the UK, may face the most extreme renewable technology portfolio events also in summer, driven by wind droughts that are generally more pronounced in summer in Europe33,47.”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

• cites [47] Cannon, D., Brayshaw, D., Methven, J., Coker, P. & Lenaghan, D. Using reanalysis data to quantify extreme wind power generation statistics: a 33 year case study in Great Britain. Renew. Energy 75, 767–778 (2015). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

In our model, least-cost storage deployment shows a substantial degree of inter-annual variation.

“In our model, least-cost storage deployment shows a substantial degree of inter-annual variation.”

The least-cost long-duration storage capacity is not defined by diurnal variability patterns in summer but rather by major winter droughts.

“Importantly, the least-cost long-duration storage capacity is not defined by these diurnal variability patterns in summer but rather by major winter droughts.”

In Central European countries, such as France or the UK, storage capacity increases substantially in years with more severe droughts.

“In Central European countries, such as France or the UK, storage capacity increases substantially in years with more severe droughts, illustrated by the steep slope of the fitted regression lines in Fig. 2.”

Nuclear power can partially reduce storage needs.

“While nuclear power can partially reduce storage needs, the storage-mitigating effect of fossil backup plants in combination with carbon removal is limited.”

In our model, there are complex interactions between long-duration storage and other flexibility options in the power sector.

“In our model, we find complex interactions between long-duration storage and other flexibility options in the power sector.”

Under unconstrained geographical balancing in scenario (4), median long-duration storage energy reduces by 5%.

“Under the assumption of unconstrained geographical balancing in scenario (4), median long-duration storage energy reduces by 5%”

• AssumptionThe authors state this as an assumption: Scenario (4) assumes unconstrained geographical balancing.

Without geographical balancing in scenario (1), low nuclear capacities mitigate least-cost long-duration storage investments only in those countries where nuclear power is deployed.

“Without geographical balancing in scenario (1), low nuclear capacities mitigate least-cost long-duration storage investments only in those countries where nuclear power is deployed”

Compared to the island setting in scenario (1), the inter-annual variation is higher in scenarios (2) and (3) with policy-oriented geographical balancing.

“Compared to the island setting in scenario (1), the inter-annual variation is higher in scenarios (2) and (3) with policy-oriented geographical balancing.”

At DACCS costs of 100 EUR per t CO2, significant energy storage capacity needs of over 30 TWh remain required in the least-cost solution.

“However, significant energy storage capacity needs of over 30 TWh remain required in the least-cost solution.”

• AssumptionRests on an assumption: Oil can be easily stored above ground at low costs and transported via trucks, avoiding the need for extensive capital-intensive grid infrastructure.

The substitution between storage and VRE particularly affects onshore wind.

“This substitution particularly affects onshore wind.”

Increasing degrees of interconnection reduce aggregate storage energy needs across all scenarios.

“Fig. 4 shows that increasing degrees of interconnection reduce aggregate storage energy needs across all scenarios.”

The highlighted extreme drought events consist of sequences of shorter but more severe droughts within contiguous periods of well-below-average renewable availability.

“The highlighted extreme drought events consist of sequences of shorter but more severe droughts within contiguous periods of well-below-average renewable availability.”

In scenario (4) (unconstrained exchange of electricity and hydrogen), the median long-duration storage energy investment is 91 TWh, corresponding to 1.8% of yearly demand.

“Scenario (4) models the counterfactual of unconstrained exchange of electricity and hydrogen between countries. This results in the lowest storage energy investments with a median (maximum) value of 91 (159) TWh, corresponding to 1.8 (3.2)% of yearly demand.”

These patterns generally apply across all investigated cost scenarios.

“These patterns generally apply across all investigated cost scenarios.”

In countries such as Romania or Spain, more severe droughts hardly increase storage capacities, driven by a less seasonal or even summer-peaking electricity demand.

“In contrast, more severe droughts hardly increase storage capacities in other countries, particularly in Romania or Spain, driven by a less seasonal or even summer-peaking electricity demand in such countries.”

In scenario (3) (including hydrogen exchange within policy-oriented limits), the median long-duration storage energy is 170 TWh, equivalent to 3.4% of yearly demand.

“Additionally including hydrogen exchange within policy-oriented limits in scenario (3) further decreases the median (maximum) long-duration storage energy to 170 (351) TWh, equivalent to 3.4 (7.1)% of yearly demand.”

In the model, lower VRE or higher storage energy costs lead to overbuilding wind and PV to reduce reliance on storage.

“In our model, lower VRE or higher storage energy costs lead to overbuilding wind and PV to reduce reliance on storage.”

In an additional sensitivity for Germany, significant long-duration storage capacities remain part of the least-cost solution even in cases with much higher capacities of nuclear or other firm zero-carbon technologies exceeding by far the peak nuclear generation capacity ever realized in Germany.

“In an additional sensitivity for Germany, we show that significant long-duration storage capacities remain part of the least-cost solution even in cases with much higher capacities of nuclear (or other firm zero-carbon technologies) exceeding by far the peak nuclear generation capacity ever realized in Germany”

Only for very low DACCS costs of 100 EUR per t CO2, system costs reduce considerably by 4.0%.

“Likewise, only for very low levels of DACCS costs of 100 EUR per t CO2, system costs reduce considerably by 4.0%.”

• AssumptionRests on an assumption: Oil can be easily stored above ground at low costs and transported via trucks, avoiding the need for extensive capital-intensive grid infrastructure.

The storage-mitigating effect of fossil backup plants in combination with carbon removal is limited.

“While nuclear power can partially reduce storage needs, the storage-mitigating effect of fossil backup plants in combination with carbon removal is limited.”

In more policy-relevant TYNDP interconnection scenarios between European countries, least-cost long-duration storage investments substantially increase in the model to 351 TWh, or more than 7% of yearly European electricity demand.

“In more policy-relevant TYNDP interconnection scenarios between European countries, the least-cost long-duration storage investments substantially increase in our model to 351 TWh, or more than 7% of yearly European electricity demand.”

For each modeled weather year, there is a clear positive correlation between the most extreme renewable droughts in a given country and that country's long-duration storage energy need.

“For each modeled weather year, there is also a clear positive correlation between the most extreme renewable droughts in a given country and that country’s long-duration storage energy need (Fig. 2a).”

Extreme droughts define long-duration storage operation and investment.

“We find that extreme droughts define long-duration storage operation and investment.”

In scenario (2), the maximum long-duration storage energy capacity is 385 TWh, or 7.8% of yearly demand.

“Scenario (2) allows for policy-oriented TYNDP 2022 cross-border exchange of electricity, which reduces (increases) the median (maximum) long-duration storage energy capacity to 206 (385) TWh, or 4.2 (7.8)% of yearly demand.”

In such cases, the discharging of long-duration storage is interrupted, and batteries and pumped hydro storage are used instead for balancing.

“In such cases, the discharging of long-duration storage is interrupted, and batteries and pumped hydro storage are used instead for balancing (e.g., visible on December 16, 22, 25, or 26 in Fig. 6b).”

In scenario (1) (energy islands), the median long-duration storage energy capacity aggregated across all modeled countries is 232 TWh, corresponding to around 4.7% of yearly European electricity demand.

“The resulting median (maximum) long-duration storage energy capacity aggregated across all modeled countries is 232 (378) terawatt hours (TWh). This corresponds to around 4.7 (7.6)% of the yearly European electricity demand.”

The effect of geographical balancing on reducing long-duration storage energy investments is even more pronounced when including cross-border exchange hydrogen in scenario (3).

“This effect is even more pronounced when including cross-border exchange hydrogen in scenario (3).”

In the scenario with high levels of nuclear power, higher shares of demand during extreme droughts can be met by nuclear, which reduces least-cost long-duration storage investments in the model-based analysis.

“In the scenario with high levels of nuclear power, higher shares of demand during extreme droughts can be met by nuclear, which reduces least-cost long-duration storage investments in our model-based analysis.”

Only at very low DACCS costs of 100 EUR per t CO2, 250 GW of oil-fired backup capacities are deployed in the least-cost system in Europe.

“Only at very low DACCS costs of 100 EUR per t CO2, 250 GW of oil-fired backup capacities are deployed in the least-cost system in Europe.”

• AssumptionRests on an assumption: Oil can be easily stored above ground at low costs and transported via trucks, avoiding the need for extensive capital-intensive grid infrastructure.

In scenario (1), the maximum long-duration storage energy capacity aggregated across all modeled countries is 378 TWh, corresponding to around 7.6% of yearly European electricity demand.

“The resulting median (maximum) long-duration storage energy capacity aggregated across all modeled countries is 232 (378) terawatt hours (TWh). This corresponds to around 4.7 (7.6)% of the yearly European electricity demand.”

As all drought events are balanced to the fullest extent possible in the copperplate scenario (4), the lowest inter-annual variation is found in this case.

“As all drought events are balanced to the fullest extent possible in the copperplate scenario (4), we find the lowest inter-annual variation in this case.”

In our model, the reducing effect of geographical balancing on long-duration storage investments remains limited for policy-oriented interconnection levels in scenarios (2) and (3) in 1996/97.

“In our model, the reducing effect of geographical balancing on long-duration storage investments remains limited for policy-oriented interconnection levels in scenarios (2) and (3) in 1996/97.”

The finding that interconnection reduces storage capacity holds for the pan-European renewable energy system including 33 countries optimized here, particularly considering the most pronounced simulated drought events in a large dataset of 35 weather years.

“we show that this finding also holds for the pan-European renewable energy system including 33 countries optimized here, particularly considering the most pronounced simulated drought events that occur in a large dataset of 35 weather years.”

For all other DACCS costs, system cost savings range only around 1.0%.

“For all the other DACCS costs, system cost savings range only around 1.0%.”

• AssumptionRests on an assumption: Oil can be easily stored above ground at low costs and transported via trucks, avoiding the need for extensive capital-intensive grid infrastructure.

Low levels of nuclear power reduce the least-cost long-duration storage energy capacity across all investigated interconnection scenarios only to a minor extent.

“Low levels of nuclear power reduce the least-cost long-duration storage energy capacity across all investigated interconnection scenarios only to a minor extent (Fig. 8).”

Long-duration storage may also be used for diurnal cycling in summer in the least-cost system.

“Finally, long-duration storage may also be used for diurnal cycling in summer in the least-cost system (Fig. 6c).”

The operation of long-duration storage coincides with VRE droughts.

“One of the principal insights from our time series and model-based analysis is that the operation of long-duration storage coincides with VRE droughts.”

Despite additional energy losses from conversion, this may lead to charging and discharging cycles of batteries or pumped hydro while hydrogen gas turbines generate electricity.

“Despite the additional energy losses from conversion, this may lead to charging and discharging cycles of batteries or pumped hydro while hydrogen gas turbines generate electricity (e.g., on December 12, 16, and 17 in Fig. 6b).”

Among the weather years in our data, 1996/97 marks the period with the highest aggregated long-duration storage energy investments across all interconnection scenarios in the least-cost system configurations.

“Among the weather years in our data, 1996/97 marks the period with the highest aggregated long-duration storage energy investments across all interconnection scenarios in the least-cost system configurations.”

In scenario (4), the maximum least-cost storage energy capacity declines by only 7%.

“The maximum least-cost storage energy capacity declines by only 7%.”

• AssumptionThe authors state this as an assumption: Scenario (4) assumes unconstrained geographical balancing.

In a perfectly interconnected Europe allowing unconstrained spatial smoothing of the most extreme drought in the data, 159 TWh of long-duration storage energy capacity result in the least-cost solution of the baseline parameterization, corresponding to around 3% of yearly electric load.

“Even in a perfectly interconnected Europe, which allows for unconstrained spatial smoothing of the most extreme drought in the data, 159 TWh of long-duration storage energy capacity result in the least-cost solution of our baseline parameterization, corresponding to around 3% of yearly electric load.”

Geographical balancing has a particularly strong storage-mitigating effect in 1988/89, as the largest droughts in major countries hardly overlap and co-occur with periods of high demand only to a limited extent.

“For instance, geographical balancing has a particularly strong storage-mitigating effect in 1988/89, as the largest droughts in major countries hardly overlap and co-occur with periods of high demand only to a limited extent (Supplementary Fig. 6).”

Normalized long-duration storage energy needs are generally much lower in Spain than in Central or Northern European countries.

“Normalized long-duration storage energy needs are generally much lower in Spain than those in Central or Northern European countries.”

Short-duration flexibility options are also used in our model to balance brief events of renewable surplus generation within longer drought events.

“Short-duration flexibility options are also used in our model to balance brief events of renewable surplus generation within longer drought events.”

During long-lasting drought events, average renewable availability is well below its long-run mean but still remains well above zero.

“During these long-lasting events, average renewable availability is well below its long-run mean, but still remains well above zero.”

During periods of significantly lower availability within droughts, storage is often discharged at full capacity and the storage state-of-charge declines strongly, particularly when co-occurring with peak-demand periods.

“During these periods, storage is often discharged at full capacity and the storage state-of-charge declines strongly, particularly when co-occurring with peak-demand periods.”

Additional cross-border exchange of renewable hydrogen, as planned in the European Union, contributes to reducing least-cost long-duration storage needs.

“Our analysis shows that additional cross-border exchange of renewable hydrogen, as planned in the European Union, contributes to reducing least-cost long-duration storage needs.”

In scenario (2) (policy-oriented TYNDP 2022 cross-border exchange of electricity), the median long-duration storage energy capacity is 206 TWh, or 4.2% of yearly demand.

“Scenario (2) allows for policy-oriented TYNDP 2022 cross-border exchange of electricity, which reduces (increases) the median (maximum) long-duration storage energy capacity to 206 (385) TWh, or 4.2 (7.8)% of yearly demand.”

In scenario (4), the maximum long-duration storage energy investment is 159 TWh, corresponding to 3.2% of yearly demand.

“Scenario (4) models the counterfactual of unconstrained exchange of electricity and hydrogen between countries. This results in the lowest storage energy investments with a median (maximum) value of 91 (159) TWh, corresponding to 1.8 (3.2)% of yearly demand.”

Allowing cross-border electricity trade in scenario (2) enables nuclear power exports to balance renewable droughts domestically and in other countries, which reduces long-duration storage energy capacity by 37% median and 35% maximum in the model.

“Allowing cross-border electricity trade in scenario (2) enables nuclear power exports to balance renewable droughts domestically and in other countries, which further reduces long-duration storage energy capacity by 37 (35)% in our model.”

Extreme renewable energy droughts identified for the policy-oriented renewable capacity portfolios simulated here may last several weeks or even months.

“Extreme renewable energy droughts identified for the policy-oriented renewable capacity portfolios simulated here, which may last several weeks or even months”

Compared to the default scenario without nuclear power, high nuclear effects result in a decrease of median long-duration storage energy capacity by 26% and maximum by 26% in scenario (1).

“Compared to the default scenario without nuclear power, these effects result in a substantial decrease of the median (maximum) storage energy capacity by 26 (26)% in scenario (1).”

Under unconstrained geographical balancing in scenario (4), maximum long-duration storage energy slightly increases by less than 1% because of interactions with bioenergy in the least-cost system.

“while the maximum storage energy even slightly increases by less than 1% because of interactions with bioenergy in the least-cost system.”

• AssumptionThe authors state this as an assumption: Scenario (4) assumes unconstrained geographical balancing.

For moderate deviations from the default setting of future wind and solar PV costs, least-cost storage deployment is hardly affected in the parameterization.

“For these scenarios, least-cost storage deployment is hardly affected in our parameterization.”

These operational patterns of nuclear power generally apply across all investigated weather years.

“These operational patterns generally apply across all investigated weather years (Supplementary Fig. 15 and Supplementary Fig. 16).”

Different weather years result in widely varying least-cost long-duration storage investments, particularly in scenarios with constrained interconnection.

“Different weather years result in widely varying least-cost long-duration storage investments, particularly in scenarios with constrained interconnection.”

For higher DACCS costs, which appear more likely given current knowledge, the system cost effect is only marginal.

“For higher DACCS costs, which appear more likely given current knowledge92, the system cost effect is only marginal.”

• cites [92] Young, J. et al. The cost of direct air capture and storage can be reduced via strategic deployment but is unlikely to fall below stated cost targets. One Earth 6, 899–917 (2023). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Our findings prove robust against the addition of moderate levels of flexible nuclear generation, which mitigate the least-cost long-duration storage investments in the model only to a minimal extent.

“Our findings prove robust against the addition of moderate levels of flexible nuclear generation, which mitigate the least-cost long-duration storage investments in our model only to a minimal extent.”

A complementary operation of long-duration storage with short-duration batteries and other flexibility options can mitigate least-cost long-duration storage and battery charging and discharging capacities to some extent, hence decreasing total system costs.

“A complementary operation of these technologies can mitigate least-cost long-duration storage and battery charging and discharging capacities to some extent, hence decreasing total system costs.”

The most pronounced renewable energy droughts, which drive long-duration storage needs, primarily occur in winter and not in summer in an interconnected European energy system.

“Our analysis also indicates that the most pronounced renewable energy droughts, which also drive long-duration storage needs, primarily occur in winter and not in summer in an interconnected European energy system.”

In a setting with substantially higher nuclear generation capacity, the storage-mitigating effect is more pronounced.

“In a setting with substantially higher nuclear generation capacity, the storage-mitigating effect is more pronounced.”

The least-cost long-duration storage capacity considering TYNDP interconnection levels in the corresponding summer-to-summer weather years (1994/95, 1995/96, 2007/08, 2008/09, and 2009/10) ranges from 103 to 239 TWh.

“Our analysis shows that the least-cost long-duration storage capacity considering TYNDP interconnection levels in the corresponding summer-to-summer weather years (1994/95, 1995/96, 2007/08, 2008/09, and 2009/10) ranges from 103 to 239 TWh.”

Yet, substantial long-duration storage investments remain optimal in the least-cost system configuration.

“Yet, substantial long-duration storage investments remain optimal in the least-cost system configuration.”

The maximum long-duration storage energy capacity in Europe identified in the least-cost system configuration, driven by the most extreme renewable drought in the winter of 1996/97, exceeds the next highest storage need determined for the weather year 1984/85 by 42%.

“Furthermore, the maximum long-duration storage energy capacity in Europe identified in the least-cost system configuration, driven by the most extreme renewable drought in the winter of 1996/97, exceeds the next highest storage need determined for the weather year 1984/85 by 42%.”

The least-cost storage capacity required to balance the most extreme drought in winter 1996/97 is 47% higher than the upper bound of the 103-239 TWh range.

“Notably, we find the least-cost storage capacity required to balance the most extreme drought in winter 1996/97 is 47% higher than this range’s upper bound.”

A sensitivity analysis for weather year 1996/97 shows that least-cost deployment of fossil backup capacity in combination with carbon abatement via DACCS could result in minor system cost decreases, but only if DACCS costs are extremely low.

“A sensitivity analysis for the weather year 1996/97, which includes the most extreme simulated renewable drought in the data, shows that the least-cost deployment of fossil backup capacity in combination with carbon abatement via DACCS could result in minor system cost decreases, but only if DACCS costs are extremely low.”

In Europe, cross-border interconnection can mitigate energy storage needs, particularly by balancing differences in wind power availability across countries.

“In Europe, such cross-border interconnection can mitigate energy storage needs, particularly by balancing differences in wind power availability across countries32.”

• cites [32] Roth, A. & Schill, W.-P. Geographical balancing of wind power decreases storage needs in a 100% renewable european power sector. iScience 26, 107074 (2023). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Variable renewable energy technologies are likely to form the backbone of transitioning to net zero in most European countries.

“Therefore, these variable renewable energy (VRE) technologies are likely to form the backbone of transitioning to net zero in most European countries1,8.”

• cites [1] Jacobson, M. Z. et al. 100% clean and renewable wind, water, and sunlight all-sector energy roadmaps for 139 countries of the world. Joule 1, 108–121 (2017). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

• cites [8] Brown, T. et al. Response to ‘burden of proof: a comprehensive review of the feasibility of 100% renewable-electricity systems’. Renew. Sustain. Energy Rev. 92, 834–847 (2018). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Transitioning to a renewable European energy system may require significant long-duration storage capacities in the order of several hundred TWh, particularly for coping with extreme, yet rarely occurring, Dunkelflaute events.

“This finding leads to the conclusion that transitioning to a renewable European energy system may require significant long-duration storage capacities in the order of several hundred TWh, particularly for coping with extreme, yet rarely occurring, Dunkelflaute events.”

In cases with high nuclear capacities, substantial long-duration storage investments remain optimal in the least-cost solution.

“our model-based analysis suggests that also in cases with high nuclear capacities, substantial long-duration storage investments remain optimal in the least-cost solution.”

The results are very robust against introducing conventional electricity generation and DACCS, except if this option could be realized at extremely low costs.

“That is, our results are very robust against introducing conventional electricity generation and DACCS, except if this option could be realized at extremely low costs.”

• AssumptionRests on an assumption: Oil can be easily stored above ground at low costs and transported via trucks, avoiding the need for extensive capital-intensive grid infrastructure.

The substitution of wind and solar capacity by nuclear power yields a higher residual demand, which is met by firm technologies including long-duration storage.

“The substitution of wind and solar capacity by nuclear power thus yields a higher residual demand, which is met by firm technologies including long-duration storage.”

If the expansion of the electricity or hydrogen networks in scenarios (2) or (3), as envisaged by European transmission system operators, cannot be realized, least-cost long-duration storage energy capacities will be between those of scenarios (1) and (2).

“If the expansion of the electricity or hydrogen networks in scenarios (2) or (3), as envisaged by European transmission system operators, cannot be realized, our results suggest that least-cost long-duration storage energy capacities will be between those of scenarios (1) and (2).”

Long-duration storage is not necessarily required to continuously discharge during the entire drought event.

“This means that long-duration storage is not necessarily required to continuously discharge during the entire drought event.”

Hydrogen-based storage integrates solar PV surplus energy, which reduces the need for short-duration battery capacity that otherwise would be required for this purpose.

“Here, hydrogen-based storage integrates solar PV surplus energy, which reduces the need for short-duration battery capacity that otherwise would be required for this purpose.”

Storage capacity needed to deal with the 1996/97 event cannot be compensated any further by additional interconnection, given that geographical balancing has already been leveraged across all of Europe to the greatest extent possible.

“That is, storage capacity needed to deal with this event cannot be compensated any further by additional interconnection, given that geographical balancing has already been leveraged across all of Europe to the greatest extent possible.”

The limited reducing effect of geographical balancing in 1996/97 is because the storage-defining drought event in this year is temporally highly correlated across many countries and also coincides with demand peaks in several countries.

“This is because the storage-defining drought event in this year is temporally highly correlated across many countries and also coincides with demand peaks in several countries (see teal boxes and gray lines in Fig. 5).”

The slight differences in storage needs between this study and previous single-country analyses are likely driven by the storage-mitigating effect of geographical balancing across countries included in the modeling.

“The slight differences in storage needs are likely driven by the storage-mitigating effect of geographical balancing across countries, which we include in our modeling.”

Nuclear displaces bioenergy and some battery discharging capacity, resulting in a net firm capacity gain far less than the 78 GW additional nuclear capacity.

“One reason for this is that nuclear displaces bioenergy and some battery discharging capacity, resulting in a net firm capacity gain far less than the 78 GW additional nuclear capacity.”

The higher inter-annual variation in scenarios (2) and (3) is because the correlation of storage-defining droughts in individual countries varies across space and weather years.

“This is because the correlation of storage-defining droughts in individual countries varies across space and weather years33.”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

The seasonal demand effect by far outweighs the differences between summer and winter droughts.

“This seasonal demand effect by far outweighs the differences between summer and winter droughts.”

Long-duration storage is an indispensable technology for coping with very pronounced renewable droughts in fully renewable energy systems with significant power supply and demand seasonality.

“Our model-based findings highlight that long-duration storage is an indispensable technology for coping with very pronounced renewable droughts in fully renewable energy systems with significant power supply and demand seasonality.”

Even during extended extreme droughts, renewable generation potential substantially above zero remains.

“Further, even during extended extreme droughts, renewable generation potential substantially above zero remains33.”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

The long-duration storage capacity that is built to cope with winter droughts can have repercussions on the least-cost battery capacity and dispatch decisions for dealing with summertime solar PV variability.

“In other words, the long-duration storage capacity that is built to cope with winter droughts can have repercussions on the least-cost battery capacity and dispatch decisions for dealing with summertime solar PV variability.”

The substantial increase in storage capacity with more severe droughts in Central European countries is because electricity demand has a pronounced winter peak in these countries.

“This effect is because electricity demand has a pronounced winter peak in these countries (Fig. 3).”

In contrast, the unconstrained balancing in the copperplate scenario (4) substantially mitigates overall storage investments, as energy can be imported from countries that are not or less affected by renewable droughts.

“In contrast, the unconstrained balancing in the copperplate scenario (4) substantially mitigates overall storage investments, as energy can be imported from countries that are not or less affected by renewable droughts.”

Significant hydro reservoir and pumped hydro storage capacities in Spain substitute long-duration storage capacities for dealing with extreme droughts to some extent.

“One driving factor for this are significant hydro reservoir and pumped hydro storage capacities in Spain, which substitute long-duration storage capacities for dealing with extreme droughts to some extent.”

The lesser effect of interconnection on storage needs in 1987/88 can be explained by a large temporal overlap of droughts in individual countries.

“This can be explained by a large temporal overlap of droughts in individual countries (Supplementary Fig. 7).”

Large-scale adoption of hydrogen-based long-duration storage, comprising electrolyzers, caverns, and hydrogen turbines for reconversion, will likely have long lead times because of supply chain and permitting bottlenecks.

“Importantly, large-scale adoption of hydrogen-based long-duration storage, comprising electrolyzers, caverns, and hydrogen turbines for reconversion, will likely have long lead times because of supply chain and permitting bottlenecks100,101.”

• cites [100] Odenweller, A., Ueckerdt, F., Nemet, G. F., Jensterle, M. & Luderer, G. Probabilistic feasibility space of scaling up green hydrogen supply. Nat. Energy 7, 854–865 (2022). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

• cites [101] Odenweller, A. & Ueckerdt, F. The green hydrogen ambition and implementation gap. Nat. Energy 10, 110–123 (2025). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Since unconstrained geographical balancing cannot be realized in reality, scenario (3) shows the most policy-relevant long-duration storage need required for dealing with the most extreme simulated Dunkelflaute event in the data.

“However, since unconstrained geographical balancing cannot be realized in reality, scenario (3) shows the most policy-relevant long-duration storage need that is required for dealing with the most extreme simulated Dunkelflaute event in the data.”

• AssumptionRests on an assumption: Unconstrained geographical balancing cannot be realized in reality.

There is a sizable need for long-duration storage in a fully renewable European power sector, irrespective of the extent of the assumed interconnection.

“Our model results suggest that there is a sizable need for long-duration storage in a fully renewable European power sector, irrespective of the extent of the assumed interconnection.”

The inter-annual variation in least-cost storage deployment is driven by differences in renewable energy droughts and demand patterns across years.

“This is driven by differences in renewable energy droughts and demand patterns across years (compare Fig. 3).”

Extreme renewable energy droughts define the least-cost operational and investment needs for long-duration storage determined in the model.

“define the least-cost operational and investment needs for long-duration storage determined in our model.”

Additional renewable capacity decreases residual demand that otherwise would require long-duration storage discharge.

“the additional renewable capacity decreases residual demand that otherwise would require long-duration storage discharge.”

In our model, 159 TWh, or 3.2% of yearly European electricity demand, is the minimum need for long-duration storage capacity for a reliable, fully renewable European energy system.

“Hence, in our model, 159 TWh or 3.2% of the yearly European electricity demand, is the minimum need for long-duration storage capacity for a reliable, fully renewable European energy system.”

Realizing the energy capacity levels determined by the model appears technically feasible, considering that around 1700 TWh of natural gas storage is currently installed in Europe, even when taking into account the lower volumetric energy density of hydrogen compared to methane.

“Overall, realizing the energy capacity levels determined by our model appears technically feasible, considering that around 1700 TWh of natural gas storage is currently installed in Europe99, even when taking into account the lower volumetric energy density of hydrogen compared to methane.”

• cites [99] GIE. Gie storage database (GIE, 2021).Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The 159 TWh figure indicates a lower bound for storage energy capacity required in a fully renewable European energy system.

“This number indicates that there is a lower bound for storage energy capacity required in a fully renewable European energy system.”

For Europe, the power sector impacts of VRE droughts are most pronounced in winter, particularly in the winter of 1996/97.

“For Europe, the power sector impacts of VRE droughts are most pronounced in winter, particularly in the winter of 1996/9733.”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Extreme VRE droughts often occur in European winter.

“Extreme VRE droughts often occur in European winter, may last up to several weeks, and span across the turn of years33.”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Summer droughts generally matter less than winter droughts in Central and Northern European countries.

“Since peak electricity demand periods usually occur in winter in Central and Northern European countries, summer droughts generally matter less than winter droughts.”

• AssumptionRests on an assumption: Peak electricity demand periods usually occur in winter in Central and Northern European countries.

Including more specific operational constraints would likely reduce nuclear's storage-mitigating effect.

“Including more specific operational constraints would likely reduce it.”

Additional demand from cooling would affect short-duration flexibility options rather than long-duration storage, considering typical diurnal cooling patterns and their partial coincidence with solar PV generation peaks.

“we expect the additional demand from cooling would affect short-duration flexibility options rather than long-duration storage, considering typical diurnal cooling patterns and their partial coincidence with solar PV generation peaks.”

Periods that qualify as drought events based on lower thresholds are likely brief and severe with very low availability.

“Periods that qualify as drought events based on lower thresholds are likely brief and severe with very low availability.”

Alternative demand profiles with more pronounced heating-related seasonality are likely to cause higher least-cost long-duration electricity storage investments than modeled here.

“Alternative demand profiles with more pronounced heating-related seasonality are likely to cause higher least-cost long-duration electricity storage investments than modeled here.”

The TYNDP 2022 draws on three weather years (1995, 2008, and 2009) for its long-term scenarios for 2050.

“For instance, the TYNDP 202218, which is a crucial European planning tool to determine cross-border infrastructure needs, draws on three weather years (1995, 2008, and 2009) for its long-term scenarios for 2050.”

• cites [18] ENTSO-e. TYNDP 2022 Scenario Building Guidelines. Tech. Rep. (European Network of Transmission System Operators for Electricity, 2022).Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Underground cavern and porous storage energy potentials vary substantially across countries, including newly built and retrofitted storage facilities from natural gas infrastructure.

“Underground cavern and porous storage energy potentials vary substantially across countries, including newly built and retrofitted storage facilities from natural gas infrastructure98,110.”

• cites [98] Talukdar, M., Blum, P., Heinemann, N. & Miocic, J. Techno-economic analysis of underground hydrogen storage in Europe. iScience 27, 108771 (2024). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

• cites [110] Caglayan, D. G. et al. Technical potential of salt caverns for hydrogen storage in Europe. Int. J. Hydrog. Energy 45, 6793–6805 (2020). sourceNot checked — no full text of the source

Under imperfect foresight, long-duration storage needs may exceed those determined in the analysis.

“Under imperfect foresight, which applies to operators in the real world, long-duration storage needs may exceed those determined here.”

• AssumptionThe authors state this as an assumption: The model used in this analysis assumes perfect foresight.

As the analysis largely abstracts from inter-hourly flexibility constraints, nuclear's potential to substitute long-duration storage needs as found here can be interpreted as an upper bound of this storage-mitigating effect.

“As we largely abstract from inter-hourly flexibility constraints, nuclear’s potential to substitute long-duration storage needs as found here can be interpreted as an upper bound of this storage-mitigating effect.”

The summer-to-summer modeling approach better captures the effects of compound drought events spanning across the turn of years, as compared to models using single calendar years.

“We thus argue that our summer-to-summer modeling approach better captures the effects of compound drought events spanning across the turn of years, as compared to models using single calendar years.”

Extreme VRE droughts span across the turn of years.

“Extreme VRE droughts often occur in European winter, may last up to several weeks, and span across the turn of years33.”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Scenario (4) identifies minimal long-duration storage needs, including unconstrained geographical balancing of VRE droughts.

“Finally, the copperplate scenario (4) identifies minimal long-duration storage needs, including unconstrained geographical balancing of VRE droughts, which can be interpreted as unavoidable or “no-regret” investments in a fully renewable European energy system.”

Market actors are unlikely to invest in such rarely utilized long-duration storage capacity without additional deployment incentives.

“Market actors are unlikely to invest in such rarely utilized long-duration storage capacity without additional deployment incentives.”

In countries with high wind shares in their capacity portfolio, the most extreme renewable portfolio drought events may occur in summer.

“In countries with high wind shares in their capacity portfolio, the most extreme renewable portfolio drought events may thus occur in summer.”

It appears unlikely that including hydro droughts would qualitatively alter the main findings, as the number of affected countries and the overall contribution of hydro power to European electricity supply remains limited.

“At the same time, it appears unlikely that including hydro droughts would qualitatively alter our main findings, as the number of affected countries and the overall contribution of hydro power to European electricity supply remains limited.”

Extreme VRE droughts may last up to several weeks.

“Extreme VRE droughts often occur in European winter, may last up to several weeks, and span across the turn of years33.”

• cites [33] Kittel, M. & Schill, W.-P. Multi-threshold time series analysis enables characterization of variable renewable energy droughts in Europe. Commun. Earth Environ. 7, 242 (2026). sourceNot checked yet

  Checked before — its earlier verdict is held back until a new check runs.

Considering significant uncertainty concerning realizable DACCS costs and implementation levels, the role of fossil fuel-based backup capacity with emission abatement for mitigating long-duration storage remains limited in a climate-neutral European energy system.

“Considering the significant uncertainty concerning realizable DACCS costs and implementation levels, our analysis suggests that the role of fossil fuel-based backup capacity with emission abatement for mitigating long-duration storage remains limited in a climate-neutral European energy system.”

If methanol-based alternative storage came with higher roundtrip efficiencies or lower costs than assumed for hydrogen-based long-duration storage, least-cost storage capacities aggregated across all countries are likely to increase compared to the results.

“Further, if this alternative storage technology came with higher roundtrip efficiencies or with lower costs than assumed here for hydrogen-based long-duration storage, least-cost storage capacities aggregated across all countries are likely to increase compared to our results.”