To calculate the warming effects of hydrogen, methane, and carbon dioxide emissions, we use the traditional GWP metric but account for constant emissions rather than a pulse of emissions. We first use the absolute global warming potential (AGWP) components, which compute the cumulative radiative forcing of a climate forcer over a specified time horizon in watts per square meter per kilogram per year ((W m-2) / (kg yr-1)). For carbon dioxide and methane, we use the Intergovernmental Panel on Climate Change (IPCC) formulations of AGWP, as shown in Eqs. (1) and (2), respectively (Myhre at al., 2013; Forster et al., 2021). Input parameters and their sources can be found in Table 1. 1AGWPCO2(H)=ACO2a0H+∑i=13aiτi1-exp⁡-Hτi 2AGWPCH4H=1+f1+f2×ACH4τ1-exp⁡-Hτ While these equations are appropriate for climate forcers with primarily direct radiative effects, hydrogen's radiative effects are entirely indirect. Therefore, we use the AGWP equations recently derived specifically for hydrogen based on sophisticated chemistry–climate modeling experiments, which explicitly account for its three main indirect effects and their varying temporal dynamics (methane, tropospheric ozone, and stratospheric water vapor) (Warwick et al., 2022). The equations are shown in the following (Eqs. 3–8) and provide the same output information of cumulative radiative forcing per time horizon ((W m-2) / (kg yr-1)) as in Eqs. (1) and (2). More details on their derivation are available in Warwick et al. (2022). Input parameters and their sources can be found in Table 1. 3AGWP1H2,i(H)=AiaiτiτH2Ctp-τi1-exp⁡-tpτi-τH2τH2-τiτH21-exp⁡-tpτH2-τi1-exp⁡-tpτi 4AGWP2H2,i(H)=AiaiτiτH22C1-exp⁡-tpτH2τH2-τi×τH2exp⁡-tpτH2-exp⁡-HτH2-τiexp⁡-tpτi-exp⁡-Hτi 5AGWP3H2,i(H)=Aiaiτi2τH2C1-exp⁡-tpτi-τH2τH2-τiexp⁡-tpτH2-exp⁡-tpτiexp⁡-tpτi-exp⁡-Hτi 6AGWPH2,iH=AGWP1H2,iH+AGWP2H2,iH+AGWP3H2,iH 7AGWPH2,CH4H=1+f1+f2AGWPH2,CH4H 8AGWPH2H=AGWPH2,CH4H+AGWPH2,O3H+AGWPH2,H2OH

To account for a constant emission rate of each forcer, as opposed to just a pulse of emissions, we consider a new pulse of emissions every year. Assuming linearity, the summation of the cumulative radiative forcing (AGWPi) from past and current pulses for each year is equal to the cumulative radiative forcing from a constant emission rate (AGWPci). To account for multiple forcers emitted from each technology, we add up the individual AGWPcis for each time horizon. Finally, to compare the climate impacts from hydrogen technologies to their fossil fuel technology counterparts, we simply divide their AGWPc values (comparable to how GWP is calculated). The results are then presented as a ratio of climate impacts (using cumulative radiative forcing as a proxy) as a function of time between two different technologies (i.e., hydrogen alternatives vs. fossil fuel technologies). A value greater than one indicates that the alternative technology (in this case hydrogen) has larger climate warming impacts at time horizon H than the original technology (and vice versa for less than one). In our analysis, we present the results as a percent change in climate impacts (cumulative radiative forcing) from the original technology, such that 1 = 0 % change (or equal), 0.5 = 50 % decrease, 2 = 100 % increase, etc.

This concept – an extension of AGWP and GWP that considers a constant emission rate (as opposed to a one-time pulse) and calculates the relative climate effects over time (as opposed to one specified time horizon, such as over 100 years) – is further documented and discussed in Alvarez et al. (2012), where it is called the technology warming potential. Several studies have used this metric to assess the climate impacts of different technologies that emit multiple greenhouse gases with varying atmospheric lifetimes, to show how the climate impacts of specific technologies change over time relative to one another (Alvarez et al., 2012; Camuzeaux et al., 2015; Ocko and Hamburg, 2019). However, given hydrogen's unique AGWP equations resulting from its varying indirect effects, we do not use the specific formulas derived in Alvarez et al. (2012) but rather follow the calculation chain described above.

To account for uncertainties in our analysis, we follow the approach of Warwick et al. (2022). First, we consider uncertainties in hydrogen's atmospheric lifetime, which (given the uncertainty in the strength of hydrogen's soil sink) are arguably the greatest source of uncertainty in hydrogen's atmospheric impacts overall (Paulot et al., 2021; Warwick et al., 2022). Compared with a central estimate of hydrogen's atmospheric lifetime of 1.9 years (Warwick et al., 2022), we use a lower-end estimate of 1.4 years (Warwick et al., 2022) and a higher-end estimate of 2.5 years (Paulot et al., 2021). Second, we apply a ± 20 % uncertainty to hydrogen's GWP (AGWPH2(H) / AGWPCO2 (H)) due to uncertainties in radiative forcing scaling factors and CO2's radiative effects (Warwick et al., 2022).

In order to assess the absolute warming impact from future hydrogen demand levels based on varying hydrogen emission rates, we apply the simple approach used by Paulot et al. (2021) to approximate long-term temperature responses to hydrogen emissions. This method uses the best estimates of the long-term increase in global surface temperature (equilibrium climate sensitivity; ECS) and radiative forcing from a doubling of CO2 concentrations and assumes that hydrogen would have a similar efficacy. The Coupled Model Intercomparison Project Phase 6 (CMIP6) models suggest a best estimate of 3.78 ± 1.08 ∘C for the ECS and a 3.93 W m-2 effective radiative forcing for a doubling of CO2 (Forster et al., 2021). This suggests a climate efficacy of 0.96 ∘C (W m2)-1. To estimate temperature responses to hydrogen emissions, we multiply this efficacy by the hydrogen effective radiative efficiency estimated in Paulot et al. (2021) per unit of emission per year (0.84 mW m-2 (Tg yr-1)-1) and the hydrogen emissions per year (emission inputs discussed in Sect. 2.2). To account for uncertainties, we use a ± 40 % uncertainty in the hydrogen effective radiative efficiency, which is comparable to the uncertainty arising from both soil sink impacts on hydrogen's atmospheric lifetime and the uncertainty in radiative forcing scaling factors and carbon dioxide's radiative effects (discussed above). Note that for the temperature analysis, we do not consider additional temperature impacts from methane emissions associated with the natural gas supply chain utilized in the production of blue hydrogen, as we want to focus on the absolute impacts from hydrogen emissions in particular.

The emissions from hydrogen applications that we consider in our analysis are hydrogen emissions (e.g., leakage, venting, and purging) from green hydrogen production and consumption as well as both hydrogen and methane emissions (e.g., leakage, venting, purging, and flaring) from blue hydrogen production and consumption. We do not consider CO2 emissions from incomplete CCUS technologies to retain simplicity and be conservative, but this would increase the climate impacts of blue hydrogen consumption depending on the efficiency and the permanence of storage. We also do not consider greenhouse gas emissions from hydrogen infrastructure build-out.

For hydrogen emissions, there is a paucity of quantitative data addressing in situ hydrogen leakage along the value chain, with empirical measurements to date focused on safety concerns (i.e., large leaks), primarily in confined spaces (Kobayashi et al., 2018). While there are many methods of hydrogen gas sensing (e.g., optical, acoustic, thermal, and electrochemical) and several types of sensors exist (Najjar, 2019), there are currently no commercially available sensors that can detect hydrogen emissions at levels well below the threshold for hydrogen gas flammability which is required to characterize emissions in the open.

It is, however, very likely that hydrogen is emitted throughout the value chain, but it is unclear – given lack of data – which components contribute most and least to emissions. Research suggests that loss rates from electrolyzers could be high, and (based on the first principles of moving a small gas molecule) it is likely that transport of hydrogen is a major source (van Rujiven et al., 2011; Cooper et al., 2022; Frazer-Nash Consultancy, 2022). Fluid dynamics theory suggests that hydrogen can leak 1.3 to 3 times faster than methane (the main component of natural gas) (Swain and Swain, 1992), although a recent study focused on low-pressure distribution pipes suggested that small leaks of methane and hydrogen may occur at similar rates if the path to leakage is convoluted (Mejia et al., 2020). Previous work also suggests that liquified hydrogen could have high emission rates from boil-off (Sherif et al., 1997).

Total value-chain emissions will ultimately depend on the configuration of the pathway from production through to end use, and there can be very little confidence in any published estimates of hydrogen emissions from a future hydrogen economy in the absence of empirical data. Of the previous studies that have made assumptions regarding the total hydrogen emissions for the purpose of assessing environmental impacts from a potential hydrogen economy, estimates range from 0.3 % to 20 % for minimum to maximum emissions (Schultz et al., 2003; Tromp et al., 2003; Colella et al., 2005; Wuebbles et al., 2010; van Ruijven et al., 2011; Bond et al., 2011; Cooper et al., 2022; Frazer-Nash Consultancy, 2022). All studies acknowledge major uncertainty in the estimates due to a lack of data, and several do not include all components of the value chain (e.g., production, compression, storage, and end-use applications). Some studies have made assumptions on total value-chain emissions citing these previous studies, typically using a range of 1 % to 10 % (Prather, 2003; Derwent et al., 2001, 2020; Paulot et al., 2021; Warwick et al., 2022). Therefore, we follow the published literature and incorporate a hydrogen emission rate of 1 % (best case) to 10 % (worst case) per amount of hydrogen consumed.

For blue hydrogen production, methane is needed as both a feedstock and a heat source, and it can be emitted along the supply chain (upstream and midstream) before it is used for producing hydrogen. The amount of methane needed to produce a unit mass of hydrogen will depend on the composition of the natural gas, the efficiency of the reformer, and how much is needed as feedstock and fuel combined. The amount needed is not well documented in the published literature and, based on public documents and private communications, can range anywhere from 2.5 to 4.5 times the mass of hydrogen (Budsberg et al., 2015; Thibault et al., 2020). In this analysis, we use a central estimate that 3 times the mass of hydrogen is needed in the form of methane. This value is on the lower end of all estimates but in the middle with respect to published values; this makes methane emission assumptions from blue hydrogen applications potentially conservative.

For methane emission estimates (including venting, purging, and flaring) upstream of hydrogen production, we use a range of 1 % (best case) to 3 % (worst case) per unit methane consumed. This is based on the latest understanding of upstream natural gas leakage from oil and gas production as well as the distribution of natural gas (Alvarez et al., 2018).

Table 2 shows the hydrogen and methane emissions used in this study for best- and worst-case leak rates based on 1 kg of either green or blue hydrogen deployed.

For estimating absolute temperature responses to future hydrogen leakage, we consider three levels of leakage (1 %, 5 %, and 10 %) and several levels of hydrogen demand from the present-day level (around 100 Tg yr-1) to a theoretical maximum projected for mid-century (around 3000 Tg yr-1). Depending on the scenario and source, projections for future hydrogen demand range from 100 to 210 Tg by 2030 and from 130 to 1370 by 2050 (Table 3). Of 21 published estimates for hydrogen demand in 2050, the average is 590 Tg (median is 570 Tg). The theoretical maximum of using hydrogen to supply the entire final energy demand in 2050 is determined based on the estimates of hydrogen demand as a percent of the final energy demand provided by Hydrogen Council (2017) and BloombergNEF (2020a), 3055 and 2900 Mt, respectively, that are each for scenarios of a decarbonized world.

To estimate the potential climate concern with respect to hydrogen technologies, we compare the net climate impacts over time from green and blue hydrogen relative to their fossil fuel counterparts based on the anticipated avoided greenhouse gas emissions from the consumption of 1 kg of hydrogen continuously each year. We consider emissions of both carbon dioxide and methane. We do not include hydrogen emissions that would be avoided from the cessation of the combustion of fossil fuels nor other co-emitted climate pollutants such as particulates, sulfur dioxide, and nitrogen oxides that contain a mix of warming and cooling forcers.

While the carbon dioxide and methane emissions avoided from the deployment of 1 kg of hydrogen will ultimately depend on the specific technology, as a first-order approximation we explore the impacts from a generic case in which a variety of fossil fuel technologies are replaced. We use estimates from the Hydrogen Council (2017) that quantify avoided carbon dioxide emissions from a scenario of supplying 18 % of the final energy demand in 2050 with hydrogen applications. They estimate that a consumption of 550 Mt of hydrogen (roughly the same amount as the average of the 21 projections published in the literature for year 2050; Table 3) can avoid 6 Gt of carbon dioxide emissions annually. In their analysis, fossil-fuel-powered end-use applications that are decarbonized by hydrogen alternatives include segments of transport, industry energy use, building power and heating, and as an industry feedstock. For transport, their vision includes hydrogen-powering hundreds of millions of cars, trucks, buses, passenger ships, and locomotives, with hydrogen-based fuels powering a share of airplanes and freight ships. For heat and power for buildings and industry, hydrogen could provide around 10 % of the heat and power required for global households and industry sectors. Of the avoided 6 Gt of CO2 annually from this level of hydrogen deployment, around half is from hydrogen applications in the transport sector, and one-third is from industry energy and feedstocks. Using the scenario and calculations from Hydrogen Council (2017) provides a central estimate of 11 kg CO2 avoided per 1 kg H2 consumed. While this estimate is for the year 2050, in the absence of better estimates, we assume that it can generally apply to earlier decades as well. However, to test the sensitivity of our results to different levels of avoided CO2 (which arguably is of further importance for specific technologies as opposed to different years), we consider three different levels of avoided carbon dioxide emissions (5, 10, and 15 kg).

Furthermore, given that the Hydrogen Council (2017) analysis does not provide avoided methane emissions associated with their hydrogen economy vision, additional assumptions need to be made to include their impact on the net radiative effect of fossil fuel applications vs. their hydrogen alternatives. First, the methane avoided will depend on the specific fossil fuel (e.g., coal, oil, or gas) used in the displaced fossil fuel technologies. For example, a natural-gas-driven technology will likely emit more methane than a coal-driven technology due to emissions associated with natural gas production and distribution. However, a natural-gas-driven technology will also likely emit less CO2 than a coal-driven one because burning natural gas emits less CO2 than coal. Therefore, for each level of avoided carbon dioxide emissions in our sensitivity analysis, we also calculate the resulting radiative impact from these emissions if the CO2 is generated from burning natural gas (i.e., considerable methane emissions). Burning 1 kg of natural gas emits 2.75 kg of CO2 if the natural gas is almost entirely methane, and we consider methane leakage rates from 1 % to 3 %, as discussed earlier. The resulting emissions of methane are shown in Table 4.

Global warming potential has become the most familiar metric for grasping the importance of a climate forcer as an agent of climate change. Hydrogen's GWP has been reported for decades, although only with respect to its tropospheric effects and for a 100-year time horizon (thereby including numerous decades when hydrogen is not influencing the atmosphere) (Derwent et al., 2001, 2006, 2020; Derwent, 2018). This has led to an undervaluing of its impact. Recent research has reported hydrogen's GWP for both tropospheric and stratospheric effects and for both 20- and 100-year timeframes, revealing that hydrogen's 100-year GWP is twice as high as previously thought, and its 20-year GWP is 3 times higher than its 100-year GWP (Warwick et al., 2022). Fig. 3a extends this work to calculate hydrogen's GWP over time.

Hydrogen's maximum GWP occurs around 7 years after the initial pulse of emissions, with a range of 25 to 60 based on uncertainties, and a central estimate of 40. This is around 8 times higher than the most well-known GWP for hydrogen (Derwent et al., 2020). Hydrogen's GWP initially increases before it declines again, as it takes several years for methane's atmospheric lifetime to increase in response to less OH being available from the reaction with hydrogen. For time horizons of 10 to 100 years, GWP decreases, which is expected when the warming effects of a pulse of emissions from a short-term forcer are compared to those of a long-term forcer; the CO2 is still in the atmosphere 100 years later, whereas the short-term forcer's impacts are long gone – meaning that the relative potency of the short-term forcer declines. In fact, the factor of 3 difference between hydrogen's GWP-20 (central estimate of 33) and GWP-100 (central estimate of 11) is similar in ratio to that from methane (80 and 30, respectively; Forster et al., 2021).

In Fig. 3b, we use an identical GWP calculation except that a constant emission rate is considered, rather than pulse emissions. The constant-emission-rate approach is a more realistic representation of hydrogen leakage in a hydrogen economy, as opposed to a one-time pulse of emissions, and is also more sensible in that one is calculating hydrogen's warming effects compared with carbon dioxide for cases where they are both impacting the atmosphere in each time horizon.

When continuous equal emissions of both hydrogen and carbon dioxide are considered, as opposed to just one pulse at time = 0, the potency of hydrogen relative to carbon dioxide can be 50 % higher than that of the pulse approach. However, this is not uniform across all timescales. In fact, before 10 years, the pulse approach (GWP) yields higher potency values than the constant-emission-rate approach. This is because the carbon dioxide impact builds up more quickly in the near term for constant emissions compared with the hydrogen impact, as the hydrogen impact takes several years to reach its full impact. However, as time goes on, the replenishing effect of constant hydrogen emissions (as opposed to decaying impacts) dominates and leads to a greater relative potency compared with the pulse approach. For hydrogen's GWP-20, constant emissions lead to around a 15 % increase in hydrogen's potency; this increases to 50 % by a time horizon of around 70 years and to almost 60 % by a time horizon of 100 years.

The results of our analysis of the climate impacts of hydrogen and methane emissions are shown in Fig. 4. If there were zero climate forcer emissions from the hydrogen applications, the result would be a -100 % change in cumulative radiative forcing; moreover, if there was no replacement, the result would be 0 %. If the climate forcer emissions from hydrogen alternatives yield more (less) warming than their fossil fuel counterparts over a particular time period, it would amount to a positive (negative) percent change in cumulative radiative forcing.

Overall, any amount of hydrogen leakage will diminish the climate benefits of avoided carbon dioxide emissions to some degree, but there are vastly different outcomes – favorable and unfavorable – depending on the production method, total emissions, and time horizon. For example, the worst case for blue hydrogen (10 % hydrogen leakage and 3 % methane leakage) could initially be worse for the climate than the CO2 emissions from the corresponding fossil fuel technologies, yielding up to 60 % more warming over the first 10 years and taking around 50 years before the benefits of the technology switch are realized. On the other hand, the best case for green hydrogen (1 % hydrogen leaks) could yield a near elimination of the climate impact compared with fossil fuel's CO2 emissions. Recall, however, that we do not include greenhouse gas emissions associated with installing the infrastructure that will be needed to support the growing demand for hydrogen and its applications.

The importance of the clean hydrogen production method – i.e., green (renewable electricity with water) or blue (steam methane reforming with CCUS) – in determining the magnitude of climate benefits is clear (Fig. 4). While hydrogen emissions can yield climate impacts for green hydrogen that are far from climate neutral over all timescales, the cumulative radiative impact is still less than that from fossil fuels; this signifies a decrease in warming from using green hydrogen alternatives. On the other hand, blue hydrogen can be better or worse for the climate depending on the leakage rate and time horizon. For example, over a 10-year time period, worst-case blue hydrogen emissions could increase the warming impact from fossil fuels by 40 % [25, 60], whereas worst-case green hydrogen emissions could decrease warming by 60 % [43, 76]. For best-case leak rates for both, blue hydrogen could still only reduce the warming impact from fossil fuels by 65 % over the first 10 years, whereas green hydrogen could reduce the impact by more than 95 %. For a 100-year time horizon, the story is similar, with worst-case leak rates yielding a doubling of the climate impact of blue hydrogen compared with green hydrogen. In fact, the worst-case green hydrogen benefits are roughly the same as the best-case blue hydrogen benefits across all timescales (such as a ∼ 65 % decrease in the warming impact from fossil fuel CO2 emissions over a 10-year period and an 85 % decrease over a 100-year period). Given that the hydrogen emissions are the same in both the blue and green cases, the difference is due entirely to the warming effects of methane emissions from the natural gas supply chain.

While the production method matters greatly, so does the level of emissions. For example, how beneficial green hydrogen is to the climate in both the near and long term will depend strongly on the level of leakage, with benefits ranging from more than a 95 % reduction in climate impacts from fossil fuel technologies to only 65 % over the first 10 years for total leakage rates of 1 % and 10 %, respectively. Even in the long term (100-year time horizon), green hydrogen may only reduce climate impacts by 85 % if there is high leakage. The impact of leakage levels is also apparent for blue hydrogen, where high leak rates for both hydrogen and methane could lead to an increase in warming relative to their fossil fuel counterparts for decades, but the low leak rates for both could cut climate impacts by more than half within 10 years. In the longer term (over 100 years), both the worst- and best-case leak rates for blue hydrogen would likely yield reductions in the climate impacts; however, the magnitude of the benefits ranges from a 45 % to a 85 % reduction, respectively. These results show the importance of the emission rate in determining the climate benefits (and potential disbenefits) of replacing fossil fuel technologies with hydrogen alternatives.

Whereas most assessments of the climate benefits of alternative technologies inherently focus on the long-term impacts due to use of the GWP-100 metric, our analysis shows how different the picture looks when considering time horizons from 10 to 100 years. This is because, unlike carbon dioxide, the warming effects of hydrogen (and methane) are short-lived and do not accumulate over time. Therefore, the benefits of hydrogen applications grow larger over time due to the prevention of the build-up of carbon dioxide in the atmosphere. If only a long-term perspective is pursued when evaluating hydrogen applications, the results will not convey the much larger relative climate impacts over shorter time horizons. For example, for the first few decades, the worst-case green hydrogen scenario may only halve the warming impacts of the fossil fuel applications that it is replacing, but the warming impacts could be reduced by three-quarters over 100 years. For blue hydrogen, the temporal significance is even more stark due to the combination of emissions of two short-term forcers. For example, the worst-case blue hydrogen alternatives could increase warming relative to fossil fuel technologies for the first several decades, but they would cut the warming impact by nearly half over 100 years. Therefore, depending on the time horizon that is considered in the analysis, one could receive very different insights into the climate benefits of the decarbonization potential of hydrogen.

This is even more acute if the GWP metric with a pulse approach is used, as opposed to a constant emission rate. While we consider constant emissions in our analysis, Fig. 4 shows the corresponding result if a pulse approach is used (see the “x” and “o” markers). While the pulse approach reasonably captures the near-term impacts of hydrogen applications relative to those of fossil fuels, it diverges over time and ultimately undervalues the cumulative radiative forcing. For example, the worst-case blue hydrogen alternative could yield a decrease in warming of only 45 % even after 100 years of replacing fossil fuel technologies, but GWP-100 suggests a decrease in warming of 65 %. Moreover, if GWP-100 is used exclusively and taken to represent hydrogen's impacts over any timescale (as it often is), the near- and middle-term impacts of hydrogen (and methane) leakage will be overlooked entirely – which, in some cases, means assuming a benefit to the climate when it is actually a disbenefit for decades.

In the above, we considered a generic case for avoiding carbon dioxide emissions from fossil fuel technologies. However, the perceived climate benefits of hydrogen alternatives will depend on the amount of CO2 avoided, which will vary depending on the technology that is replaced. Therefore, to test the sensitivity of our results to the amount of CO2 avoided, we consider avoided emissions of 5, 10, and 15 kg per 1 kg of hydrogen deployed (compared with our central estimate of 11 kg) and compare the relative climate impacts of the hydrogen applications over a 20-year time horizon (solid bars in Fig. 5). We find that blue hydrogen could yield more than a 150 % increase in warming over the first 20 years if avoided emissions of CO2 are on the lower end and if leak rates from hydrogen and methane are at the upper end, whereas green hydrogen may only reduce warming by 20 %. However, if avoided emissions of CO2 are on the higher end, both the worst-case blue and green hydrogen alternatives would yield climate benefits, reducing warming by 10 % and 75 %, respectively.

Given that methane emissions may also be avoided by replacing fossil fuel technologies, we extend the analysis to consider a case where the fossil fuel that was burned to produce the CO2 was natural gas (hatched bars in Fig. 5), using the same best- and worst-case methane leak rates as in the hydrogen applications. We find that the avoided methane emissions may play a significant role in increasing the near-term benefits of hydrogen applications, but there is a strong dependence on the corresponding CO2 emissions that are avoided. For example, while the worst-case blue hydrogen alternative with the lower-end avoided CO2 values would still be worse for the climate over the first 20 years, even when including avoided methane, the central estimate of avoided CO2 would switch from worse for the climate to better for the climate. For the worst-case green hydrogen alternative, climate benefits would double for all levels of avoided CO2 when including avoided methane emissions. However, given that natural gas emits less CO2 when burned than coal, it is likely that CO2 emissions would be lower if methane emissions were higher, as opposed to both being on the higher end. Therefore, a case-by-case study with reported data on both carbon dioxide and methane emissions from fossil fuel technologies is warranted to fully understand the impact of avoided methane emissions.

We find that the present-day hydrogen demand (around 100 Tg) may cause at most 0.01 ∘C warming for all levels of hydrogen emissions. For 2030 projections, five estimates based on different scenarios and sources suggest an average hydrogen demand of 150 Tg (see Table 3), which could double the 100 Tg impact for upper-end leak rates (10 %) and uncertainties (0.02 ∘C). For 2050 projections, 21 different estimates suggest a range in demand from 130 to 1370 Tg (Table 3), with an average of 590 Tg. For the worst-case hydrogen leak rates (10 %), these levels of demand could yield anywhere from 0.01 ∘C to 0.1 ± 0.05 ∘C warming. On the other hand, if total hydrogen emissions are kept minimal (1 %), temperature responses could be less than 0.02 ∘C including uncertainties. For context, a 590 Tg hydrogen demand could supply around 20 % of the final global energy demand in 2050 under a 2 ∘C scenario (Hydrogen Council, 2017; BloombergNEF, 2020a).

Figure 6 shows the long-term temperature responses to various hydrogen demand levels, up to a theoretical maximum of 3000 Tg estimated for 2050 (which would correspond to using hydrogen for the total final energy demand in a 2 ∘C decarbonization scenario). Using hydrogen for the total final energy demand in 2050 could lead to greater than 0.1 ∘C warming with a 5 % leak rate and up to 0.4 ∘C warming with 10 % leak rates and uncertainties in hydrogen's radiative effects.

However, this level of hydrogen demand is not realistic. Of the available projections in the literature for hydrogen demand in 2050, 4 suggest values between 100 and 199 Tg, 3 suggest values between 200 and 499 Tg, 11 suggest values between 500 and 999 Tg, and 3 suggest values between 1000 and 1999 Tg (Table 3). None project hydrogen demands below 100 or above 2000. Sustained hydrogen demands around 800 Tg or greater (which could account for around a quarter of the final energy demand in 2050) could contribute at least 0.1 ∘C of warming if leak rates and uncertainties are at the upper end. For context, this amount of warming could offset the avoided warming in 2050 from deploying all cost-effective options to mitigate methane emissions globally over the next decade, which otherwise could have slowed down global-mean warming rates by up to 15 % (Ocko et al., 2021), or the avoided warming anticipated from the phasing out of hydrofluorocarbons (Xu et al., 2013). This amount of warming (∼0.1 ∘C) is also equal to the amount of warming projected in 2100 from carbon dioxide emissions from international shipping and aviation combined in the absence of climate action (Ivanovich et al., 2019). However, if leakage does not exceed 1 %, the temperature response could be an order of magnitude smaller.