AT M O S P H E R I C S C I E N C E Considering intentional stratospheric dehydration for climate benefits Joshua P. Schwarz1*, Ru- han Gao1, Troy D. Thornberry1, Andrew W. Rollins1, Karen H. Rosenlof1, S Robert W. Portmann1, ThaoPaul Bui2, Eric J. Jensen1,3, Eric A. Ray1,3 We introduce a climate intervention strategy focused on decreasing water vapor (WV) concentrations near the tropopause and in the stratosphere to increase outbound longwave radiation. The mechanism is the targeted injection of ice- ucleating particles (INP) in air supersaturated with respect to ice at high altitudes in the tropical n entryway to the stratosphere. Ice formation in this region is a critical control of stratospheric WV. Recent airborne in situ data indicate that targeting only a small fraction of air parcels in the region would be sufficient to achieve substantial removal of water. This “intentional stratospheric dehydration” (ISD) strategy would not counteract a large fraction of the forcing from carbon dioxide but may contribute to a portfolio of climate interventions by acting with different time and length scales of impact and risk than other interventions that are already under consideration. We outline the idea, its plausibility, technical hurdles, and side effects to be considered. INTRODUCTION 1 NOAA Chemical Sciences Laboratory, Boulder, CO, USA. 2NASA Ames Research Center, Moffett Field, CA, USA. 3Cooperative Institute for Research in Environmental Sciences, Boulder, CO, USA. *Corresponding author. Email: joshua.p.schwarz@noaa.gov Schwarz et al., Sci. Adv. 10, eadk0593 (2024) 28 February 2024 considered before, via modifying clouds and convection (11, 12). Here, we focus narrowly on the TTL pathway to the stratosphere. Figure 1 provides a conceptual representation of the idea. The figure shows an idealized case for an air parcel with few INP slowly rising into the stratosphere while experiencing horizontal advection and temperature fluctuations. As temperatures drop, relative humidity increases and the WV increasingly tends toward condensation into ice. Conversely, higher temperatures result in lower RHi for fixed WV content. The figure considers an air mass that fails to reach high enough RHi to result in homogeneous ice nucleation en route to the stratosphere and hence has fixed WV content. By adding INP, ice would be formed to remove the WV generating the supersaturation with respect to ice. After the resulting ice falls into lower altitudes, the WV content of the parcel would stay constant. The INP injection would also result in increased short-ived cloud presence due to the l creation of ice crystals (while reducing possibilities for homogeneously nucleated ice clouds to form). “Natural experiments” of injections of INP have already been evaluated: Ueyama et al. (13) showed that ice crystals (which act as efficient INP) “injected” into the TTL from the anvils of large convective events generate a net dehydration of the air, even as they themselves represent injections of water. Laboratory experiments also support the intervention concept (14). This approach to intentional stratospheric dehydration (ISD) has many similarities to cirrus cloud thinning (12, 15) (CCT), which has been considered for over a decade. In CCT, INP injection into air masses that might otherwise undergo homogenous freezing in commercial air lanes, primarily in the northern midlatitudes, is considered. CCT generally is envisioned at lower altitudes than ISD and is primarily focused on influencing cirrus. ISD and CCT also share some technical aspects with cloud seeding approaches for weather modification. However, the primary goals of the various approaches, the desired timescale of the impact of injections of INP, and the geographic scale of impacts are distinct. RESULTS The conceptual scheme outlined above leaves many questions open. In this section, we present model results exploring INP injection in the TTL, measurements of WV supersaturations relevant to assessing the amount of water susceptible for removal and the scale of the 1 of 7 Downloaded from https://www.science.org on June 25, 2024 Stratospheric water vapor (WV) is recognized as a contributor to the greenhouse gas effect, with both clear annual cycles and decadal trends with radiative significance (1). In our current climate, reductions in stratospheric water help reduce the radiative forcing (RF) of the Earth system (2). The dominant pathway of WV to the stratosphere is through the tropical tropopause layer (TTL) (3). Here, most stratospheric water enters the overworld after air parcels are “freeze- ried” at the lowest d temperatures experienced en route from lower, warmer, wetter altitudes (4, 5). This dehydration often occurs at the so- alled “western Pacific c cold point” (WCP) (3). Once in the stratosphere, air spends approximately 1 to 4 years circulating to the poles (highly dependent on the specific transport pathway) before returning to the troposphere. WV’s most important radiative effects occur near the tropopause. The effectiveness of the TTL pathway for removing WV depends critically on both the lowest temperatures experienced by individual air parcels rising into the region and on the presence of ice- ucleating n particles (INP) in those parcels (6–8), as follows. The colder the air is, for a given WV concentration, the higher the resulting relative humidity with respect to ice (RHi), which scales with the vapor’s tendency to condense into ice. At a given temperature, RHi and WV are linearly related: doubling WV doubles RHi. Above 100% RHi, the lowest energy state of water is ice. However, in the absence of existing ice or INP, RHi must reach extremely high levels (~200%) before WV can “homogeneously” generate ice under TTL conditions (9); if such high RHi is not reached and INP are not present, then all the WV in the parcel is carried into the stratosphere. On the other hand, if a sufficient concentration of ice particles (>100 liter−l) form on INP, then they can fall into lower altitudes, thereby removing WV responsible for RHi in excess of about 100% from the air. We know that the abundance of natural INP in this region is often insufficient to prevent considerable water, approximately a quarter of all the water transported via this route, from entering the stratosphere (6, 7, 10). Intentional reduction of the flow of WV into the stratosphere via INP injection may provide beneficial climate impacts, as has been Copyright © 2024 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. Distributed under a Creative Commons Attribution NonCommercial License 4.0 (CC BY-NC).
A −40 −60 Altitude (km) Latitude (°) S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e 20 -250 -200 -150 -100 -50 0 50 Longitude (°) B 18 16 200 190.5 150 Reduced RHi C 190.0 100 189.5 189.0 Water vapor (synthetic) , , Low INP High INP 50 INP addition 0 Depleted water vapor D −20 −15 −10 −5 0 Days of transport (day) Fig. 1. Schematic of the intentional stratospheric dehydration (ISD) concept to reduce water vapor transport across the tropical tropopause and into the stratosphere. (A) A 22-day back trajectory for an air parcel ending at the solid circle and starting at the open circle; this is for an air mass that rose into the stratosphere near the WCP, as modeled by HYSPLIT and adjusted for clarity. (B) HYSPLIT altitude associated with the parcel. (C) Synthetic data to conceptually represent the temperature and relative humidity with respect to ice (RHi) fluctuations experienced by an air parcel with few INP that does not experience high enough RHi to initiate homogeneous nucleation of ice. An injection of INP either when the air mass exhibits a clear supersaturation or in advance of the supersaturation condition removes the WV generating the supersaturation with respect to ice at the low temperature of the TTL by forming ice that falls out of the air mass [shown in (D)]. After injection, the air mass continues its journey to the stratosphere with additional temperature fluctuations. Once in the stratosphere, temperature increases quickly with altitude, so relative humidity quickly decreases with altitude above the tropopause. The shaded areas represent the change in relative humidity and in WV content in the air mass due to the loss of ice generated by injection of INP, with the high-INP RHi remaining a fixed fraction of the low-INP RHi case after the ice has fallen from the air mass. Homogeneous nucleation occurs at RHi of ~200% at the relevant temperatures (9). challenge of accessing it, the outline of the challenges of seeding large air masses with INP, and some side effects of INP injection. distribution. The simulated cloud properties and WV fields have been shown to compare favorably with observations (7, 16). To crudely simulate the impacts of supplemental INP, we treated all aerosol parModeling INP injection in the TTL ticles (rather than a tiny fraction) as effective INP throughout the The basic concept shown in Fig. 1 is effective at reducing stratospher- model. We assumed a threshold ice supersaturation of 10% for actiic WV in principle. We ran a large ensemble of Lagrangian one- vation of the INP and hence did not remove all supersaturations. We dimensional (1D) microphysical model simulations driven by back compared cases with and without the addition of the supplemental trajectories and analysis temperatures in the TTL. The approach for INP across the whole TTL and averaged WV fields over the last simulating natural TTL cirrus combined 60- ay diabatic back tra- week of the simulation. The model showed a 10% reduction in the d jectories with 1D simulations of ice crystal nucleation, deposition tropical- verage 100- Pa MLS kernel–averaged WV concentration a h growth, sublimation, and sedimentation (7, 16). Briefly, back trajec- with INP injection. WV at the tropopause was reduced in an absotories were launched throughout the tropics from a 2° longitude × 2° lute sense by 0.3 parts per million (ppm; from 2.5 to 2.2 ppm). An latitude grid at 372 K potential temperature over ±20° latitude and undesirable side effect of the water removal shown by the microphysall longitude. Curtains of temperature versus height and time were ical model, as expected a priori, was that the TTL cloud frequency extracted from the ERA- nterim analysis fields. To improve realism, increased: tropical- verage 16-to 18- m cloud fraction increased I a k the model superimposed high- requency temperature fluctuations from 0.11 to 0.19. Note that the net impacts on both cloud frequency f on these fields to represent wave structures as parameterized by Jensen and WV removal would be reduced by INP injection targeted only and Pfister (17). These temperature curtains were used to drive ice at those air masses destined to enter the stratosphere with substancloud simulations. The cloud model tracked the growth, sublima- tial supersaturation rather than over the entire TTL region, as modtion, and sedimentation of thousands of individual ice crystals to eled here. We expect that targeting air masses expected to experience represent clouds. This Lagrangian approach avoids numerical diffu- supersaturation only shortly before entry into the stratosphere sion artifacts. The ensemble of simulations provides statistics of TTL would result in a better ratio of WV reduction and related cooling cirrus occurrence and microphysical properties, as well as the WV to cloud frequency increase with potential counter effects. More Schwarz et al., Sci. Adv. 10, eadk0593 (2024) 28 February 2024 2 of 7 Downloaded from https://www.science.org on June 25, 2024 188.5 6.0 5.5 5.0 4.5 4.0 3.5 RHi (%) (synthetic) Temperature (synthetic) 191.0
sophisticated model approaches, for example, evaluating targeted injections rather than widespread inclusion of IN in the TTL, perhaps by analyzing only back trajectories from the stratosphere that imply excess water transport, could reduce the overestimates arising from our first- rder evaluation of cirrus cloud frequency changes. o This issue is further explored below. This scale of stratospheric water depletion maps to an RF of −0.07 to −0.09 W m−2 based on previous analyses of the RF of WV in the stratosphere in which a 1- pmv stratospheric WV increase p over 20 years was estimated to produce an adjusted RF of 0.2 to 0.3 W m−2, slightly less than the ~0.36 W m−2 attributed to CO2 from 1980 to 1996 (1, 2). Most of the forcing occurs near the tropical tropopause (2). We do not estimate the impact of the additional cloud fraction under targeted INP injection, but we expect it to be small (see Discussion below) and represent forcing over shorter timescales than that due to the WV reduction. Furthermore, targeting only supersaturated air masses raises the possibility that portions of the mechanism of CCT (reduction of homogeneously nucleated cirrus and a shift of such cirrus to lower altitudes/higher temperatures) may prove to be a beneficial side effect of ISD. Fig. 2. Distribution of ice saturation ratios in the clear skies of the tropics. Data from the Airborne Tropical TRopopause EXperiment (ATTREX) campaign. As air rises into lower temperature altitudes with fixed WV content, relative humidity with respect to ice (RHi) increases, with increasing fractions of air showing supersaturations. The dashed line shows the relative humidity required for homogeneous nucleation from Schneider et al. (9). Schwarz et al., Sci. Adv. 10, eadk0593 (2024) 28 February 2024 3 of 7 Downloaded from https://www.science.org on June 25, 2024 with a higher likelihood for higher RHi. At the lowest temperatures encountered, RHi values were roughly centered on 100%, with half the air having some supersaturation. We focus on the coldest air masses sampled at and below 186 K, where the ultimate dehydration of rising air was likely, as a proxy for the multiple routes of air parcels into the stratosphere; these made up 15% of the air in the 17-to 18- m altitude range sampled in ATTREX. By integrating observed k WV associated with RHi of at least 120%, we find that 7% of the total water in the rising air masses could be removed by bringing those parcels down to saturation. This estimate of water susceptible to removal approaches the 10% reduction predicted by the model treatment for INP injection in TTL air generally. We use the ATTREX data as a proxy to estimate the granularity of supersaturated regions within the WCP. If the highly supersaturated 1% of the air in the region of interest was near- niformly dispersed u in the WCP, then there would be almost no reduction in effort needed to target it for INP injection relative to seeding the entire WCP. We searched the ATTREX data to identify the duration of individual intercepts of highly supersaturated air (>120% RHi). ATTREX does not provide the ideal dataset with which to do this, as the aircraft was near- ontinuously profiling vertically between ~14.5-and 19- m alc k titude. However, we are not aware of a dataset that would be more Implications for ISD drawn from in situ observations of TTL appropriate for a first- rder evaluation of this issue. We assume that o water vapor In situ observations support the idea that targeted INP injections the duration of each intercept maps the diameter of a cylinder of could substantially reduce WV transport from the TTL to the strato- supersaturated air with a fixed depth of 1000 m. The depth assumpsphere without generating large cloud fractions. Figure 2 shows tion is based on wave structure in the TTL (19). The data show that clear- ky frequency distributions of ice- aturation ratios measured supersaturated regions are quite coherent, likely due to the length s s in the tropics during the NASA Airborne Tropical TRopopause EX- scales of the wave activities that generate them. Figure 3 shows the observed durations (sorted from shortest to periment (ATTREX) campaign (18). In this campaign, a high- altitude unmanned aerial vehicle (UAV), the NASA Global Hawk, longest with matching event index) and the cumulative volume of carrying instrumentation to quantify WV and ambient tempera- these air parcels. The figure shows that 550 regions of supersaturatures in situ, explored the TTL. Details are provided in Materials tion with respect to ice (RHi > 120%) were observed in the dataset, and Methods. Just as in the simplified conception of the intervention yet just 1% of them accounted for fully half the total saturated vol(Fig. 1), the data in Fig. 2 show that lower temperatures are associated ume. Ten percent of the observed supersaturated regions contained
Fig. 3. Exploration of the length scales of supersaturated air masses in the tropical tropopause layer. ATTREX observations of air parcels with >120% relative humidity with respect to ice were sorted and indexed according to their duration in the measurement record. Shown are cumulative volume (assuming a fixed depth for each observed region), and the duration in seconds of each transect. This dataset included air sampled below the altitudes that are more likely representative of final dehydration conditions. 1000 INP per liter would grow to ~8- m diameter in less than an μ hour, and then fall at about an order of magnitude faster than the general uplift speed in the TTL. The assumptions about necessary INP concentration and composition lead to an estimate that between 2 kg (for the 10-nm INP) and 2 tons of material (for the 100-nm particles) would need to be injected each week. The challenge of lifting 2 tons of material per week to 17- m altitude is readily k achievable with present- ay technology. Available high- ltitude aird a craft capable of reaching the necessary altitude range include WB- 57Fs, each with an ~4-ton useful payload, U-2s with ~2-ton payload, h and Global Hawk UAVs, with ~1 1/2–ton useful payload (and 24- our endurance). Schwarz et al., Sci. Adv. 10, eadk0593 (2024) 28 February 2024 4 of 7 Downloaded from https://www.science.org on June 25, 2024 The challenges of dispersing INP into specific large air masses The challenge of dispersing INP into specific large volumes of supersaturated air in the conditions of the TTL, potentially with or without substantial wind shear and/or turbulence, is considerable. The literature does not yet reflect substantial progress in evaluating the technical aspects of particle injections accounting for mixing, ice nearly all the water susceptible for removal. This is a substantial nucleation, and sedimentation under these conditions. At present, amount of water potentially available for removal from small, fairly 1- ay temperature forecasts coupled with satellite-nferred TTL d i contiguous fractions of air. This finding is supported by previous temperatures (23) are sufficient to locate the coldest air masses in the research in cirrus regimes (20), which indicated that larger ice- region; this was demonstrated during the ATTREX campaign. Misupersaturated regions were associated with higher WV content. crowave temperature profiler instruments (24), such as have flown These results indicate that nontargeted widespread INP injection (as extensively on high- ltitude aircraft including the ER- (a research a 2 in the model treatment) is not necessary to effect substantial water aircraft of similar design to the U-2), the B-57, and the Global Hawk, removal. It follows that cirrus generation would be limited to ~<1% provide near- ircraft temperature measurements over a couple of a of the air in the WCP region only (from the initial injections, sec- kilometers above, below, and in front of the plane, and allow some ondary impacts are not considered in this estimate). real- ime identification of temperature boundaries. WV concentrat tion forecasts are much lower quality. Improved forecasts could beEstimating the amount of INP necessary to effect WV come available with increased radiosonde launches in the region, reduction in the WCP TTL improved (or even maintained) satellite assets, and potential inteEven a “small fraction” of the atmosphere could be too large to mean- gration of dedicated in situ temperature/pressure measurements, for ingfully influence with INP injections. We assume a target region for example, from weather balloons. On this basis, we continue to exthe final dehydration of the volume of a 1000-m-thick band of air near plore the challenge of dispersing INP, accepting that targeting relevant the top of the TTL (~17-to 18- m altitude) over a region the size of air masses is not unreasonable, yet note that forecasts of supersatuk Australia (7.7 × 106 km2). The distribution of the final dehydration rations allowing seeding of air multiple days in advance of the occurpoints in the western Pacific covers approximately this area (3). In this rence of supersaturations would ease targeting, extend the timescales region, a typical uplift speed of 0.0015 m s−1 (21) and typical wind for introducing INP, and provide longer timescales for INP to mix speeds <10 m s−1 suggest that this air would exit the region in ap- into air parcels. proximately 1 week. For this first estimate of the challenge of seeding Figure 3 indicates that the supersaturated regions containing 90% air in the TTL for ISD, we assume 1 week to seed 1% of this size re- of the susceptible WV were traversed in ~4 min to half an hour at gion, as suggested by the observational evidence above. Global Hawk speed (~180 m/s). This corresponds to lengths of ~40 The mass of INP material necessary to seed such a volume de- to 300 km across. We use contrail size estimates [~0.5 km × 0.5 km pends on the type of INP and on the assumption that they could be (25)] from Boeing 737 aircraft 0.5 hours after emission as a starting dispersed without large- cale coagulation or deactivation. We con- estimate for the cross- ectional area that could be seeded with INP s s sider bismuth triiodide (BiI3; bulk density of ~6 g cm−3), in particles or contrail ice particles with a single aircraft pass. Assuming uniform of either 10 or 100 nm as previously proposed for CCT (15). The dispersion of INP within this cross- ection, seeding the individual s minimum size estimated for INP to be effective for vapor deposition supersaturated volumes suggested by the measurements would take nucleation in the conditions of the TTL is 10 nm in diameter (22). ~4 to 150 hours for a single plane. This assumes full seeding of the One hundred–nanometer diameter represents a more conservative vertical extent of the supersaturated air masses, without the expectaestimate of the size and would likely ease technical issues associated tion that upper- evel INP would then generate crystals that would l with coagulation with background aerosol. Larger INP, such as sediment down to remove lower-evel supersaturations. This estil those associated with mineral dust sources, would pose increasingly mate neglects transit times for the aircraft, which would likely delarge engineering challenges to application. We assume that 1000 INP pend critically on forecasting skills. Considering transit needs, the per liter concentrations would be sufficient to reduce excess WV potential value of the WB- 7 for injection relative to the longer- 5 based on the simulations in (6) and will generate large enough ice duration platforms (U- and Global Hawk) is lessened. Seeding 2 crystals for high sedimentation rates. For 2 ppm of excess water, challenges using different injection platforms such as small UAVs,
balloons, or high- ltitude dirigibles would need to be considered in a the context of the specifics of INP dispersal technology and platform capabilities. DISCUSSION We have provided a high-evel overview of the ISD idea and prol vided early-stage assessments of its potential based on modeling and evaluation of a relevant in situ dataset. Conclusions from the observational data, combined with an estimate that the WCP region dehydrates ~40% of all modeled air parcels rising into the stratosphere (3), suggest an ~0.1- pm reduction in stratospheric WV associated p with a few tens of milliwatts per square meter cooling possible from targeting 1% of the WCP with INP injection. It is clear that this Schwarz et al., Sci. Adv. 10, eadk0593 (2024) 28 February 2024 5 of 7 Downloaded from https://www.science.org on June 25, 2024 Side effects of ISD via INP injection in the TTL There are a number of apparent side effects from INP injection in the TTL that should be considered. First among them are the effects of additional clouds generated by the injections, both at the altitude of interest and, potentially, at lower altitudes after ice crystals potentially containing INP sediment to warmer temperatures. Although it is possible that the radiative effects of these clouds will be counteracted by the reduced presence of homogenously nucleated clouds in the TTL, a full and confident accounting will likely require both testing and advances in modeling cirrus in the region. Additional uncertainties about the radiative impacts of additional ice clouds in specific locations arise from dependencies on pre-existing (or, in the case of ISD, potentially additional) clouds in the column (26, 27). In this context, we note that model evaluation of CCT, with similar issues, has already reached the point where model results must wait on improved physical understanding and representations to provide clarity on these effects (15). Changing stratospheric water could have implications for feedback to the large- cale global circulation (28). The rate and height of s the Brewer-Dobson circulation have been predicted by models to be affected by greenhouse gas concentrations (29), so ISD could counteract some possible changes. However, as ISD has the potential to modify stratospheric WV only at levels below natural variability, it seems unlikely that it would drive a large- cale change in circulation. s By influencing the chemistry of ozone- epleting substances, d stratospheric water contributes to the destruction of ozone in this region (30, 31). Although ISD could only generate a systematic shift in stratospheric WV of lesser amplitude than natural variability, it may affect the recovery of stratospheric ozone (in a positive, yet negligible, sense for the current climate). In a future climate with different concentrations of ozone and ozone- epleting substances, the d effects of ISD on ozone could be different. As there are multiple pathways of water to the stratosphere that could change in the future, we can anticipate that long-term stratospheric water trends will overshadow any absolute effects of ISD. Nevertheless, ISD may provide a consistent absolute reduction in WV with corresponding radiative implications. Depending on the technological approach considered to inject INP, direct emissions associated with the effort could vary tremendously but are likely to be absolutely negligible relative to global emissions. We assume that for an ~10-aircraft approach as discussed below, the amount of fuel burned to effect INP injection, with resulting CO2, INP, and WV emissions, would be a negligible fraction of global aviation- elated fuel burn. r intervention strategy would not negate a large fraction of the forcing generated by CO2 over the long term. However, it may be valuable as an element within a larger portfolio of climate intervention strategies each with differing impacts and positive and negative consequences, and as a case study to guide further study of stratospheric WV controls and impacts. The basic technical challenges of ISD would require the development of injectors, an improved understanding of INP ice activation, cloud evolution, and WV transport and forecasting in the TTL, and the use of either existing aircraft or newly developed and developing alternative platforms such as long-duration stratospheric/upper tropospheric balloons (32), dirigibles (33), or solar- owered UAVs p (34). If only kilograms of INP material were needed on a weekly basis (the low end of our estimate), then the major challenges would be accessing large air masses and understanding the optimal dispersion of the INP. At the high end of our estimate for an INP mass, of 2 tons/week, the use of existing aircraft such as the Global Hawk appears to be a starting point for consideration. In this case, we estimate a total flight time requirement per week of ~800-hour flight, excluding transit time, and requiring nonstop flight from five aircraft. One- alf of this duty cycle would be required for 10 dedicated h aircraft. Note that current forecasting skills are insufficient to optimize flight times and aircraft routing. Negative side effects of INP injection in the TTL appear limited if their effects are highly concentrated in target regions, yet still deserve further study; first, among them are the secondary impacts of ice falling into lower air masses, acting directly there or evaporating, releasing IN and water, potentially to return to the upper TTL and enter the stratosphere. We have not evaluated these possibilities. Note that there may be technological solutions to such secondary issues. For example, molecular INPs active low in the troposphere under very different conditions than the TTL have been shown to be deactivated by acid and oxidation (35). This raises the possibility of synthetic INP that would only maintain effectiveness over a short time. Alternatively, the supersaturated air masses in focus here contain excess water; an engineering solution in which ice crystals are generated in situ and dispersed could lead to eventual evaporation at lower altitudes without a corresponding increase in INP. For example, aircraft pollution may be generated that is insufficiently ice- active to nucleate in the TTL outside of the extreme conditions of the emitted hot/wet exhaust (36, 37). One of the issues commonly raised in the context of considering climate interventions is whether testing of the intervention is possible. In the case of ISD, INP injections would be expected to be accompanied by the formation of ice clouds of sufficient optical depth for detection by lidar. The statistical evaluation of measured WV content of seeded and unseeded air masses, for example, on timelines such that supersaturations would be expected to be diminished in both cases, could address the basic effectiveness of injection. Note that natural variability would likely disguise net long- erm impacts t in the stratosphere. The ISD idea highlights the value of an improved understanding of the controls of WV transport into the stratosphere. This includes continued and improved tracking of TTL WV and temperature distributions and trends, and further research into the behavior of INP in very cold low- ater mixing ratio conditions. Consideration of w ISD within as wide a range of climate interventions as possible could be useful in assessing the likelihood (or lack of it) of any alternatives to the reduction of CO2 emissions to alleviate climate stresses.
MATERIALS AND METHODS REFERENCES AND NOTES 1. P. M. de F. Forster, K. P. Shine, Assessing the climate impact of trends in stratospheric water vapor. Geophys. Res. Lett. 29, 10- –10–4 (2002). 1 2. S. Solomon, K. H. Rosenlof, R. W. Portmann, J. S. Daniel, S. M. Davis, T. J. Sanford, G.- . Plattner, Contributions of stratospheric water vapor to decadal changes in the rate K of global warming. Science 327, 1219–1223 (2010). 3. M. R. Schoeberl, A. E. Dessler, Dehydration of the stratosphere. Atmos. Chem. Phys. 11, 8433–8446 (2011). 4. A. W. Brewer, Evidence for a world circulation provided by the measurements of helium and water vapour distribution in the stratosphere. Q. J. Roy. Meteorol. Soc. 75, 351–363 (1949). 5. R. E. Newell, S. Gould- tewart, A stratospheric fountain? J. Atmos. Sci. 38, 2789–2796 S (1981). 6. E. J. Jensen, G. Diskin, R. P. Lawson, S. Lance, T. P. Bui, D. Hlavka, M. McGill, L. Pfister, O. B. Toon, R. Gaog, Ice nucleation and dehydration in the tropical tropopause layer. Proc. Natl. Acad. Sci. U.S.A. 110, 2041–2046 (2013). 7. R. Ueyama, E. J. Jensen, L. Pfister, J.-E. Kim, Dynamical, convective, and microphysical control on wintertime distributions of water vapor and clouds in the tropical tropopause layer. J. Geophys. Res. Atmos. 120, 10483–10500 (2015). 8. W. J. Randel, E. J. Jensen, Physical processes in the tropical tropopause layer and their roles in a changing climate. Nat. Geosci. 6, 169–176 (2013). 9. J. Schneider, K. Höhler, R. Wagner, H. Saathoff, M. Schnaiter, T. Schorr, I. Steinke, S. Benz, M. Baumgartner, C. Rolf, M. Krämer, T. Leisner, O. Möhler, High homogeneous freezing onsets of sulfuric acid aerosol at cirrus temperatures. Atmos. Chem. Phys. 21, 14403–14425 (2021). 10. A. W. Rollins, T. D. Thornberry, R. S. Gao, S. Woods, R. P. Lawson, T. P. Bui, E. J. Jensen, D. W. Fahey, Observational constraints on the efficiency of dehydration mechanisms in the tropical tropopause layer. Geophys. Res. Lett. 43, 2912–2918 (2016). 11. B. Chen, Y. Yin, Can we modify stratospheric water vapor by deliberate cloud seeding? J. Geophys. Res. 119, 1406–1418 (2014). Schwarz et al., Sci. Adv. 10, eadk0593 (2024) 28 February 2024 12. D. L. Mitchell, W. Finnegan, Modification of cirrus clouds to reduce global warming. Environ. Res. Lett. 4, 045102 (2009). 13. R. Ueyama, M. Schoeberl, E. Jensen, L. Pfister, M. Park, J.- . Ryoo, Convective impact on M the global lower stratospheric water vapor budget. J. Geophys. Res. Atmos. 128, e2022JD037135 (2023). 14. J. Skrotzki, P. Connolly, M. Schnaiter, H. Saathoff, O. Möhler, R. Wagner, M. Niemand, V. Ebert, T. Leisner, The accommodation coefficient of water molecules on ice – cirrus cloud studies at the AIDA simulation chamber. Atmos. Chem. Phys. 13, 4451–4466 (2013). 15. C. Tully, D. Neubauer, D. Villanueva, U. Lohmann, Does prognostic seeding along flight tracks produce the desired effects of cirrus cloud thinning? Atmos. Chem. Phys. 23, 7673–7698 (2023). 16. R. Ueyama, E. J. Jensen, L. Pfister, Convective Influence on the humidity and clouds in the tropical tropopause layer during boreal summer. J. Geophys. Res. Atmos. 123, 7576–7593 (2018). 17. E. Jensen, L. Pfister, Transport and freeze- rying in the tropical tropopause layer. J. d Geophys. Res. Atmos. 109, (2004). 18. E. J. Jensen, L. Pfister, D. E. Jordan, T. V. Bui, R. Ueyama, H. B. Singh, T. D. Thornberry, A. W. Rollins, R.- . Gao, D. W. Fahey, K. H. Rosenlof, J. W. Elkins, G. S. Diskin, J. P. DiGangi, S R. P. Lawson, S. Woods, E. L. Atlas, M. A. N. Rodriguez, S. C. Wofsy, J. Pittman, C. G. Bardeen, O. B. Toon, B. C. Kindel, P. A. Newman, M. J. McGill, D. L. Hlavka, L. R. Lait, M. R. Schoeberl, J. W. Bergman, H. B. Selkirk, M. J. Alexander, J.-E. Kim, B. H. Lim, J. Stutz, K. Pfeilsticker, The NASA Airborne Tropical TRopopause EXperiment: High- ltitude aircraft measurements in a the tropical western pacific. Bull. Am. Meteorol. Soc. 98, 129–143 (2017). 19. J. E. Kim, M. J. Alexander, T. P. Bui, J. M. Dean-Day, R. P. Lawson, S. Woods, D. Hlavka, L. Pfister, E. J. Jensen, Ubiquitous influence of waves on tropical high cirrus clouds. Geophys. Res. Lett. 43, 5895–5901 (2016). 20. M. Diao, M. A. Zondlo, A. J. Heymsfield, L. M. Avallone, M. E. Paige, S. P. Beaton, T. Campos, D. C. Rogers, Cloud- cale ice- upersaturated regions spatially correlate with high water s s vapor heterogeneities. Atmos. Chem. Phys. 14, 2639–2656 (2014). 21. S. Park, R. Jiménez, B. C. Daube, L. Pflster, T. J. Conway, E. W. Gottlieb, V. Y. Chow, D. J. Curran, D. M. Matross, A. Bright, E. L. Atlas, T. P. Bui, R. S. Gao, C. H. Twohy, S. C. Wofsy, The CO2 tracer clock for the tropical tropopause layer. Atmos. Chem. Phys. 7, 3989–4000 (2007). 22. C. Marcolli, Technical note: Fundamental aspects of ice nucleation via pore condensation and freezing including Laplace pressure and growth into macroscopic ice. Atmos. Chem. Phys. 20, 3209–3230 (2020). 23. E. E. Borbas, W. P. Menzel, E. Weisz, D. Devenyi, Deriving atmospheric temperature of the tropopause region- pper troposphere by combining information from GPS radio u occultation refractivity and high- pectral- esolution infrared radiance measurements. J. s r Appl. Meteorol. Climatol. 47, 2300–2310 (2008). 24. M. Heckl, A. Fix, M. Jirousek, F. Schreier, J. Xu, M. Rapp, Measurement characteristics of an airborne microwave temperature profiler (MTP). Atmos. Meas. Tech. 14, 1689–1713 (2021). 25. V. Freudenthaler, F. Homburg, H. Jäger, Contrail observations by ground- ased scanning b lidar: Cross- ectional growth. Geophys. Res. Lett. 22, 3501–3504 (1995). s 26. X. Tan, Y. Huang, M. Diao, A. Bansemer, M. A. Zondlo, J. P. DiGangi, R. Volkamer, Y. Hu, An assessment of the radiative effects of ice supersaturation based on in situ observations. Geophys. Res. Lett. 43, 11039–11047 (2016). 27. E. Johansson, A. Devasthale, A. M. L. Ekman, M. Tjernström, T. L’Ecuyer, How does cloud overlap affect the radiative heating in the tropical upper troposphere/lower stratosphere? Geophys. Res. Lett. 46, 5623–5631 (2019). 28. E. Charlesworth, F. Plöger, T. Birner, R. Baikhadzhaev, M. Abalos, N. L. Abraham, H. Akiyoshi, S. Bekki, F. Dennison, P. Jöckel, J. Keeble, D. Kinnison, O. Morgenstern, D. Plummer, E. Rozanov, S. Strode, G. Zeng, T. Egorova, M. Riese, Stratospheric water vapor affecting atmospheric circulation. Nat. Commun. 14, 3925 (2023). 29. S. Oberländer-Hayn, E. P. Gerber, J. Abalichin, H. Akiyoshi, A. Kerschbaumer, A. Kubin, M. Kunze, U. Langematz, S. Meul, M. Michou, O. Morgenstern, L. D. Oman, Is the Brewer-Dobson circulation increasing or moving upward? Geophys. Res. Lett. 43, 1772–1779 (2016). 30. D. B. Kirk-Davidoff, E. J. Hintsa, J. G. Anderson, D. W. Keith, The effect of climate change on ozone depletion through changes in stratospheric water vapour. Nature 402, 399–401 (1999). 31. V. L. Dvortsov, S. Solomon, Response of the stratospheric temperatures and ozone to past and future increases in stratospheric humidity. J. Geophys. Res. Atmos. 106, 7505–7514 (2001). 32. M. Wall, World view’s “Stratollite” balloon stays aloft for record 32 days; www.space.com/ world-view-stratollite-balloon-flight-record.html (2019). 33. M. V. Frandsen, D. Kim, Stratospheric platforms to improve life on our planet; www.sceye. com/ (2023). 34. A. Morrison, T. Caverley, PHASA- 5® completes first successful stratospheric flight; www. 3 baesystems.com/en/article/phasa-35--completes-first-successful-stratospheric-flight (2023). 35. B. G. Pummer, H. Bauer, J. Bernardi, S. Bleicher, H. Grothe, Suspendable macromolecules are responsible for ice nucleation activity of birch and conifer pollen. Atmos. Chem. Phys. 12, 2541–2550 (2012). 6 of 7 Downloaded from https://www.science.org on June 25, 2024 In situ observations of water vapor and temperature used to calculate ice supersaturation ratios Ice supersaturation observations were generated from observations of WV and in situ temperature obtained on the NOAA/NASA Global Hawk UAV during ATTREX. WV mixing ratio was measured with the NOAA UT/LS hydrometer instrument (38) at 1- intervals with s an absolute uncertainty of 0.2 ppm. Ambient temperatures were measured with the NASA Meteorological Measurement System (39) with an absolute uncertainty of 0.3 K. For the purposes of this manuscript, these uncertainties are negligible. The ice supersaturation ratio was derived from the temperature and WV concentration measurements based on the relationships in (40). The data were collected over 199 hours of sampling between ~20°N and 10°S latitudes and ~90°W to 230°W longitude, between 14-and 19- m altitude k above sea level. A total of 105 hours of sampling occurred from Guam and stayed north of Papua New Guinea and the Solomon Islands. These measurements were collected in 2013 and 2014. We use clear-sky measurements for our statistics; however, in the 17-to 18-km altitude range of interest, only 5% of the observations occurred in detectible (ice water content > 0.05 ppm) clouds. As aircraft operational considerations require avoidance of clouds when possible, the in situ in- loud data here are necessarily skewed. Recent studies (41, c 42) indicate that in- loud supersaturations relevant to the TTL are c likely transient on timescales shorter than relevant here, but inspection of in-loud processes and resulting WV transport into the c stratosphere, especially for extremely low ice concentration clouds, may further extend the promise of ISD. The Global Hawk was limited in its ability to target the coldest air masses (which would be associated with the highest supersaturations of WV). This limitation arose because of operational limits on the temperatures the aircraft can fly through. This limitation likely slightly biases the observed frequency distributions of ice- upersaturated air s to slightly lower frequencies of supersaturations.
36. R. H. Moore, K. L. Thornhill, B. Weinzierl, D. Sauer, E. D’Ascoli, J. Kim, M. Lichtenstern, M. Scheibe, B. Beaton, A. J. Beyersdorf, J. Barrick, D. Bulzan, C. A. Corr, E. Crosbie, T. Jurkat, R. Martin, D. Riddick, M. Shook, G. Slover, C. Voigt, R. White, E. Winstead, R. Yasky, L. D. Ziemba, A. Brown, H. Schlager, B. E. Anderson, Biofuel blending reduces particle emissions from aircraft engines at cruise conditions. Nature 543, 411–415 (2017). 37. D. J. Cziczo, K. D. Froyd, C. Hoose, E. J. Jensen, M. Diao, M. A. Zondlo, J. B. Smith, C. H. Twohy, D. M. Murphy, Clarifying the dominant sources and mechanisms of cirrus cloud formation. Science 340, 1320–1324 (2013). 38. T. D. Thornberry, A. W. Rollins, R. S. Gao, L. A. Watts, S. J. Ciciora, R. J. McLaughlin, D. W. Fahey, A two- hannel, tunable diode laser- ased hygrometer for measurement of c b water vapor and cirrus cloud ice water content in the upper troposphere and lower stratosphere. Atmos. Meas. Tech. 8, 211–224 (2015). 39. S. G. Scott, T. P. Bui, K. Roland-Chan, S. W. Bowen, The meteorological measurement system on the nasa er- aircraft. J. Atmos. Oceanic Tech. 7, 525–540 (1990). 2 40. D. M. Murphy, T. Koop, Review of the vapour pressures of ice and supercooled water for atmospheric applications. Q. J. Roy. Meteorol. Soc. 131, 1539–1565 (2005). 41. B. Kärcher, E. J. Jensen, G. F. Pokrifka, J. Y. Harrington, Ice supersaturation variability in cirrus clouds: Role of vertical wind speeds and deposition coefficients. J. Geophys. Res. Atmos. 128, e2023JD039324 (2023). 42. E. J. Jensen, T. D. Thornberry, A. W. Rollins, R. Ueyama, L. Pfister, T. Bui, G. S. Diskin, J. P. Digangi, E. Hintsa, R. S. Gao, S. Woods, R. P. Lawson, J. Pittman, Physical processes controlling the spatial distributions of relative humidity in the tropical tropopause layer over the Pacific. J. Geophys. Res. 122, 6094–6107 (2017). Acknowledgments: We thank the pilots and crew of the NASA Global Hawk for supporting the ATTREX campaign. Funding: ATTREX was supported by the NASA Upper Atmospheric Composition program. Some ATTREX activities were supported via NASA (grant #NNH12ZDA001N- UACO, to R.- .G.). Author contributions: Conceptualization: J.P.S., R.- .G., E.J.J., E.A.R., A.W.R., T.D.T., S S K.H.R., and R.W.P. Formal analysis: J.P.S., T.B., T.D.T., and R.W.P. Investigation: J.P.S., T.D.T., A.W.R., and T.B. Methodology: J.P.S., R.- .G., E.J.J., A.W.R., and T.D.T. Visualization: J.P.S., R.- .G., T.D.T., and K.H.R. S S Writing—original draft: J.P.S. and R.W.P. Writing—review and editing: J.P.S., E.J.J., R.- .G., K.H.R., T.B., S A.W.R., and T.D.T. Data curation: T.B. and T.D.T. Validation: J.P.S., T.B., and T.D.T. Resources: E.A.R. and T.D.T. Supervision: J.P.S. and K.H.R. Project Administration: J.P.S. Competing interests: The authors declare that they have no competing interests. Data and materials availability: ATTREX data are available at: https://asdc.larc.nasa.gov/project/ATTREX, under citation: NASA/LARC/SD/ASDC, 2015. ATTREX-Aircraft_insitu_TraceGas_Measurements, https://doi.org/10.5067/ASDC_DAAC/ ATTREX/0003. All data needed to evaluate the conclusions in the paper are present in the paper. Submitted 7 September 2023 Accepted 25 January 2024 Published 28 February 2024 10.1126/sciadv.adk0593 Downloaded from https://www.science.org on June 25, 2024 Schwarz et al., Sci. Adv. 10, eadk0593 (2024) 28 February 2024 7 of 7
Fleepit Digital © 2021