The Impact on the Ozone Layer of a Potential Fleet of Civil Hypersonic Aircraft

The aeronautical community is currently researching technology that might lead to commercial hypersonic aircraft that would cruise at Mach 5–8 in the middle or upper stratosphere and would transfer passengers from London to New York or from Los Angeles to Tokyo in just a couple of hours. Depending on the engine technology to be adopted, these aircraft will potentially release substantial amounts of water vapor and nitrogen oxides around 30–40 km altitude. We show here that the operation of a large fleet of such aircraft could potentially deplete considerable amounts of ozone in the stratosphere, which would lead to a substantial increase in biologically damaging ultraviolet radiation reaching the Earth's surface. The calculations are based on a specific emission scenario, which carries large uncertainties but can easily be scaled to account for the type of aircraft engine to be eventually adopted, improved technology to be expected, and the size and operation conditions of the future aircraft fleet. Plain Language Summary Commercial hypersonic aircraft, if developed in the future, will be flying at Mach 5 to 8 in the middle to upper stratosphere (30 to 40 km altitude) to carry passengers in a couple of hours from London to New York or from Los Angeles to Tokyo. Depending on the adopted technology and fleet size, the powerful engines of such airplanes may release substantial amounts of water and nitrogen oxides in the stratosphere, which could potentially damage the protecting ozone layer and hence increase the level of biologically damaging ultraviolet radiation reaching the Earth surface. The paper uses an advanced global atmospheric model to assess the impact of a potential fleet of hypersonic aircraft.


Introduction
The aeronautical community is currently researching technology which might lead to the development of a fleet of commercial hypersonic transport (HST) at some time in the future. This new generation of aircraft with a 5,000 nautical mile range (about 9,000 km) would allow passengers to cross the Atlantic or the Pacific Ocean in just 2 to 3 hr at a speed of Mach 5-8. HST are expected to operate at or above 30 km altitude (100,000 feet), thus within or slightly above the protective ozone layer. The potential environmental impacts of such a fleet could be substantial and must therefore be assessed. Here, we focus on the ozone depletion that could result from the operation of an HST fleet.
In the 1970s, there was concern about the potential impact on stratospheric ozone (O 3 ) of a fleet of supersonic transport aircraft (SSTs flying typically at Mach 2-2.5) that was planned to operate in the decades ahead at an altitude of 15-20 km (Johnston, 1971). At the request of Congress in the United States, a crash program, the Climatic Impact Assessment Program (CIAP), was initiated by the Department of Transportation to provide a quantitative estimate of the climatic impact of these planned aircraft. Similar studies were conducted in other countries, including the United Kingdom, France, and Australia. The U.S. assessment (Grobecker et al., 1975) concluded that the emission of nitrogen oxides by a hypothetical fleet of 500 SSTs would not significantly affect the Earth's climate but would damage the ozone layer (Crutzen, 1972;Dameris et al., 1998;Hesstvedt, 1974;Johnston et al., 1989;McElroy et al., 1974;Tie et al., 1994).
The impact of current commercial subsonic aircraft, cruising in the upper troposphere (and in the lower stratosphere for flights occurring in polar regions), was assessed more recently (Brasseur et al., 1996(Brasseur et al., , 1998(Brasseur et al., , 2016Lee et al., 2009;Penner et al., 1999;Schumann, 1994). Aviation CO 2 represents 2% to 3% of the total annual CO 2 emissions. Nitrogen oxides released by jet engines generate additional ozone in the upper troposphere and therefore enhance radiative forcing. Condensation trails produced by the aircraft and the induced cirrus clouds also contribute to significant radiative perturbations (Penner et al., 1999). The overall radiative forcing by current commercial aviation is estimated to be of the order of 0.08 W/m 2 (Intergovernmental Panel for Climate Change, 2007).
The concept of commercial hypersonic aircraft is still under discussion, and no scenario about future aircraft emissions can be currently retained as representative of the future since these emissions depend on the type of engines to be adopted, the combustion system in such engines, the size of the future fleet, the conditions of operation including flight altitudes, and preferred routes of the aircraft. This study should therefore be viewed as an attempt to quantify the sensitivity of the stratospheric ozone concentration and of the ozone column abundance to localized emissions of nitrogen oxides and water vapor in the middle and upper stratosphere and hence to identify a possible threat to the biosphere (including humans) from the potential operation in the future of a fleet of hypersonic aircraft.

Emissions by Hypersonic Aircraft
Hypersonic aircraft are expected to operate at Mach 5-8 in an altitude range of 30-40 km (Yanes, 2020). The engine technology to be adopted remains a matter of debate, and intense research is conducted to investigate the advantages and disadvantages of different alternatives. One option under consideration is the adoption of air-breathing ramjet or scramjet engines. These engines use the fast motion of the aircraft to compress the incoming air rather than a built-in compressor as used in classic turbojet airliners. Acceleration from takeoff to supersonic speeds must be achieved by another type of engine, for example, a turbojet engine burning hydrocarbon fuel or a rocket engine fueled by stored liquid oxygen and hydrogen. Scramjets, considered for aircraft operation at speeds higher than Mach 5, represent a modified version of the ramjets in which the air flows through the engine at supersonic speed, which improves the conversion of thermal energy into thrust. In one specific HST concept (Bowcutt, 2020), a turbofan jet engine, used to provide low-speed thrust during the initial part of the flight, is combined with an air-breathing ramjet to operate the vehicle at a hypersonic speed of Mach 5.
Other options are being considered. For example, the LAPCAT project (Steelant, 2008(Steelant, , 2011 was established by the European Commission to develop different hypersonic vehicle concepts, enabling an antipodal flight time to about 4 hr. One of the proposed configurations (LAPCAT-A2) was based on precooled turbo-ramjet fueled by hydrogen. The Synergetic Air Breathing Rocket Engine (SABRE), under consideration in the United Kingdom (Szondy, 2019), combines in a single system a conventional air-breathing jet engine (to be used for speed of less than approximately Mach 5) and a rocket engine (capable of accelerating the aircraft to speeds as high as Mach 25). The European STRATOFLY 300-passenger airliner was conceived to reach Mach 8 (European Commission, 2018) and would be equipped with two types of ramjet engines, one to be used for the first part of the flight up to the speed of Mach 4.5 and the second for flight operations from Mach 4.5 to Mach 8. A single cryogenically precooled air-breathing turbojet engine would operate from take-off to cruise speed of Mach 5 as part of the HST under consideration by the Japan Aerospace Exploration Agency (JAXA) (http://www.aero.jaxa.jp/eng/ research/frontier/hst/). Air-breathing engines have a high fuel efficiency compared to rockets in terms of propulsion performance, but combustion produces carbon dioxide, nitrogen oxides, and water vapor that are released in the atmosphere.
The combustion products to be released in the atmosphere by the proposed engines are not well characterized and quantified because they depend on the technology to be adopted. Early estimates made in the 1970s for the ramjet engine (Letzberg, 1973;Petersen & Waters, 1972) during the CIAP Program suggested that, due to the high temperatures in the combustion chamber, the emission index (EI) for nitrogen oxides (NOx) should be an order of magnitude greater than that of classic turbojets. The CIAP study adopted for the air-breathing engines of Concorde and of a hypothetical American SST an EI of 18 grams of NO 2 emitted per kg of fuel burned. Letzberg (1973) estimated that the EI for a ramjet would vary from 40 to 300 g NO 2 per kg of fuel for Mach numbers ranging from 5 to 12. That report also indicated that the amount of nitrogen oxides produced in the HST flow field by the shock wave occurring around the nose of the aircraft would be an order of magnitude lower than the emission by a ramjet engine, thus of the same order as the emission by a classic turbojet. Brooks et al. (1993) indicate that, for Mach numbers between 1 to 10, the effect of NO output produced by the high-temperature boundary layer of the HST can be neglected compared to the emission by the engines. At Mach 16, however, the output of NO from the boundary layer of the aircraft would become roughly equal to the emission from a scramjet engine, and thus considerably higher than the emission of a classic air-breathing turbojet. Letzberg (1973) also provides the emission for hydrogenated species per kg of fuel burnt: 500 to 900 g of H 2 O, 19 to 300 g of H 2 , and 38 to 333 g of OH, depending on the speed of the aircraft. Chemical kinetics computations (Grobecker et al., 1975) show that, due to the small value of the EI (OH)/EI (NOx) ratio, the conversion of nitrogen oxides to nitric acid is small. It can be therefore assumed that nitrogen compounds resulting from the combustion are emitted as nitric oxide. Finally, one should emphasize that the emission indices quoted in these early studies are very uncertain and can only be used as guides for future studies; they may represent overestimates if new engines benefit from advanced technology and optimized modes of operation.
Because of the large uncertainties existing on the number of aircraft to be deployed, the engine technology to be adopted, and the emission indices to be achieved in modern combustion chambers, we present here sensitivity studies that illustrate the potential impact of a hypothetical fleet of future hypersonic aircraft on the chemical composition of the atmosphere and specifically on the ozone abundance. The results produced for specific emissions can be easily scaled and adapted to represent the conditions that could prevail in the future. The model simulations presented in this paper should therefore be considered as providing a guide for the developers of hypersonic aircraft engines. In the present study, we only consider the key drivers of ozone change, specifically the emissions by aircraft engines of nitrogen oxides and water vapor. The contribution of molecular hydrogen is implicitly included in the water vapor emissions since both compounds, when oxidized, initiate the formation of the hydroxyl radical that catalytically destroys the ozone molecule.
For our model simulations and due to the lack of reliable data, we assume that the geographical distribution of the HST routes and the HST emissions are identical to the scenario adopted for a fleet of 500 stratospheric SSTs Kawa et al., 1999), operating at Mach 2.4 at 17-20 km altitude, but we assume that the injection of NOx and water vapor by the proposed HST takes place around 30 km (Case 1) or around 40 km altitude (Case 2) in a thin layer of about 2 km. We assume ( Table 1) that 129 million kg of fuel is burned each day by the HST fleet with emission indices for water vapor and nitrogen oxides of 1,237 g H 2 O/kg fuel burned and of 20 g NO 2 /kg fuel burned, respectively. The corresponding total HST emissions for water vapor and nitrogen oxides are therefore equal to 58.3 Tg (H 2 O)/year and 0.94 Tg (NO 2 )/year, respectively. Note that in the adopted scenarios, about 84% of the HST emissions are taking place in the Northern Hemisphere. The total Northern Hemisphere emissions are therefore 48.8 Tg (H 2 O)/year and 0.79 Tg (NO 2 )/year. For comparison with natural processes as simulated by our model, which is described in section 3, the total natural atmospheric production of water vapor by methane oxidation and of nitrogen oxides by nitrous oxide oxidation is 60 Tg (H 2 O)/year and 2.8 Tg (NO 2 )/year, respectively. All these values represent orders of magnitude; in the case of water specifically, it is assumed here that the oxidation of background methane leads to two molecules of water. This number is, however, only approximate since it does not account for the part of methane diverted to the production of molecular hydrogen primarily in the mesosphere (Le Texier & Solomon, 1988) and for other chemical processes that reduce the water production yield to about 1.7 in the lower stratosphere (Frank et al., 2018).

Earth's Future
In our simulations, the release of water and nitrogen oxides is assumed to take place primarily above the oceans, where most of the hypersonic flights will take place. Figure 1 shows that, with the adopted emissions, the stratospheric injection of water vapor and nitrogen oxides by the fleet of aircraft is larger than the natural atmospheric production of these species at the planned flight levels. The figure also highlights two prominent corridors in which high HST traffic is expected to occur: the northern Atlantic and Pacific Oceans.

Model Description
We derive the impact of these emissions on the global atmosphere by using the Community Earth System Model version 1 (CESM1), Whole Atmosphere Community Climate Model (WACCM) for the 2050 period. WACCM is a coupled chemistry-climate model ranging from the Earth's surface to the lower thermosphere Marsh et al., 2013). WACCM is a vertical extension of the Community Atmosphere

Earth's Future
Model version 4 (CAM4) and includes all of the physical parameterizations of CAM4 (Neale et al., 2013) and a finite-volume dynamical core (Lin, 2004) for the tracer advection. The horizontal resolution is 1.9°l atitude × 2.5°longitude. The vertical resolution in the lower stratosphere ranges from 1.2 km near the tropopause to about 2 km near the stratopause; in the mesosphere and thermosphere, the vertical resolution is~3 km. Simulations used in the work are based on the International Global Atmospheric Chemistry/ Stratosphere-troposphere Processes And their Role in Climate (IGAC/SPARC) Chemistry Climate Model Initiative (CCMI) (Morgenstern et al., 2017). Improvements in CESM1 (WACCM) for CCMI include a modification to the orographic gravity wave forcing that reduced the cold bias in Antarctic polar temperatures (Calvo et al., 2017;Garcia et al., 2007) and updates to the stratospheric heterogeneous chemistry which improved the representation of polar ozone depletion (Solomon et al., 2015;Wegner et al., 2013). The chemistry used in this version of WACCM is representative of the troposphere through the lower mesosphere Tilmes et al., 2016). The species included within this mechanism are contained within the Ox, NOx, HOx, ClOx, and BrOx chemical families, along with CH 4 and its degradation products. In addition, 20 primary nonmethane hydrocarbons and related oxygenated organic compounds are represented along with their surface emissions. There is a total of 183 species and 472 chemical reactions; this includes 17 heterogeneous reactions on multiple aerosol types (i.e., sulfate, nitric acid trihydrate, and water-ice). In this work, the baseline scenario follows the CCMI REF-C2 definition (Morgenstern et al., 2017), where the model is coupled to an interactive ocean. This scenario includes forcing of greenhouse gases (CH 4 , N 2 O, and CO 2 ), organic halogens, volcanic surface area density and heating, and 11-year solar cycle variability. The Quasi-Biennial Oscillation (QBO) is nudged to observed monthly mean tropical winds from the 86 to 4 hPa levels (Matthes et al., 2010). For the simulations shown here, the model is branched off of an existing REF-C2 simulation in year 2040 and run through year 2050. High-frequency meteorological fields are output instantaneously every 3 hr. In addition, sea surface temperatures and sea ice from this branched run are derived. With this information, the reference period and HST simulations are run in a specified dynamics (SD) mode (Kunz et al., 2011;Lamarque et al., 2012). Here, temperature, zonal and meridional winds, and surface pressure are used to drive the physical parameterization that control boundary layer exchanges, advective and convective transport, and the hydrological cycle. These meteorological fields are nudged from the surface to 125 km, with a 5-hr relaxation time constant. The goal here is to have all scenarios have identical temperature and transport distributions from the surface through the mesosphere. Table 1 shows the scenarios used in this work. All simulations considered in the present study were carried out for a period of 11 years; the results presented below represent the annual average of the last year of model integration.

The Response of Ozone to the HST Injection of Water Vapor and Nitrogen Oxides
It is well known that the destruction of odd oxygen (Ox = O 3 + O + O( 1 D) + other terms; Brasseur & Solomon, 2005) is catalyzed by nitrogen oxides in the stratosphere (Crutzen, 1970(Crutzen, , 1979 and by hydroxyl radicals in the upper stratosphere and mesosphere (Bates & Nicolet, 1950). The simultaneous release of both types of compounds could lead to chemical interactions that could mitigate the effects of the individual emissions on ozone destruction. For example, the injection of water vapor contributes to the conversion into nitric acid of a fraction of aircraft generated reactive NOx. Reactive ozone-depleting chlorine (e.g., ClO) produced by the photochemical degradation of industrially manufactured chlorofluorocarbons is partly destroyed by the addition of nitrogen oxides released by the HST aircraft engines. However, this effect will gradually become less pronounced as the atmospheric levels of anthropogenic chlorine are expected to decrease in the coming decades. In this study, we have adopted a level of total stratospheric chlorine of 2.3 ppbv, the concentration expected to be representative for the 2050 time period. This is a decrease of 35% from the peak near year 2000.
We illustrate the response of the atmospheric composition to the injection by a potential HST fleet by displaying the change in the zonal and annual mean concentration (expressed in percent) of water vapor, nitrogen oxides, and ozone. Figure 2 shows the percentage increase in the water vapor concentration of the stratosphere resulting from the operation of the HST fleet. This scenario assumes that the cruise altitude of the aircraft is 30 km. The largest perturbation (larger than 30% or~3.5 ppmv) in H 2 O takes place between 30°and 70°N, although the noticeable effect of the water emissions reaches the North Pole, the equator, and When the altitude of the injection (with the same emission) is located at 40 km altitude, the maximum perturbation for H 2 O reaches 140% (10 ppmv). The atmospheric region affected by the emission of water extends from 20 to 60 km in the northern midlatitudes and from 35 to 60 km in the tropics. It also affects most of the Southern Hemisphere with some small effect (20%) between 30 and 70 km. Interestingly, the predicted increase of water vapor reaches 5-15% in the upper mesosphere/lower thermosphere.
The perturbations in total odd nitrogen (noted NOy and defined as the sum of N + NO + NO 2 + NO 3 + HO 2 NO 2 + HNO 3 + 2N 2 O 5 + ClONO 2 + minor contributions from other nitrogen species) are also shown in Figure 2 for aircraft emissions occurring at the cruise altitudes of 30 and 40 km, respectively. In the first case, the relative perturbation is largest (50-90%) between 15 and 35 km at middle and high latitudes and from 20 to 40 km in the tropics (40-60%). In the second case, the altitude with the largest relative perturbations ranges from 30 to 45 km (up to 300% increase) and from 35 to 60 km at midlatitudes and in the tropics, respectively. In each case, a tongue of elevated NOy extends in the Southern Hemisphere to about 70°S.
It is interesting to note that only a relatively small fraction of the material emitted by the aircraft's engines in the Northern Hemisphere is transported and dispersed south of the equator. Interhemispheric exchanges occur with a time constant of about 1 to 1.5 years (Bowman & Cohen, 1997;Geller et al., 1997) due to the strong upwelling that takes place in the equatorial region (Plumb & Eluszkiewicz, 1999;Randel et al., 2002). Tracers injected into the lower or middle stratosphere are therefore transported to higher levels before being transported across the equator. The lifetime of water vapor is larger than 1 year below 40 km altitude and larger than 10 years below 30 km. The transformation of water molecules released in the atmosphere by hypersonic aircraft into ozone-depleting reactive hydrogen radicals (H, OH, and HO 2 ) is therefore slow in the stratosphere but becomes efficient in the mesosphere and lower thermosphere for the water molecules that are transported up to these altitudes. At 80 km, the lifetime of H 2 O is only of the order of 10 days, and large amounts of hydrogen radicals are formed. As shown by Figure 4, the perturbation in HOy (defined as the sum H + OH + HO 2 + 2H 2 O 2 + 2 H 2 + 2H 2 O) generated by the aircraft emissions is most pronounced near the location of the emissions at 30 km in the Northern Hemisphere; a limited intrusion toward the Southern Hemisphere occurs, however, in the upper stratosphere and lower mesosphere. Figure S1 in the supporting information shows the 40 km injection altitude result. When we contrast the behavior of HOx defined as the sum of all hydrogen species without water

10.1029/2020EF001626
Earth's Future vapor and molecular hydrogen included (HOx = HOy − 2 H 2 -2 H 2 O) with the behavior of the total HOy, we note that the increase in HOx is largest (of the order of 1 ppm) above 80 km altitude, but it is very small in the stratosphere. In fact, as shown by the two lower panels in Figure 4, the change in the HOx concentration in the lower stratosphere is positive in the case of an H 2 O-only injection. It is negative, however, in the case of a combined (H 2 O, NOx) emission since the nitrogen oxides released by the aircraft tend to convert HOx into nitric acid. In summary, water vapor injected by aircraft should not destroy substantial amounts of ozone in the stratosphere, but the material injected transported upward should destroy a significant quantity of ozone and atomic oxygen in the lower thermosphere with very small effects on the ozone column abundance.
In the case of NOy (Figure 5), the situation is somewhat similar to that of water vapor but with some quantitative differences. Odd nitrogen is destroyed by the photolysis of NO in the upper stratosphere, mesosphere, and lower thermosphere (producing N and O), followed by the recombination of N with NO to form molecular nitrogen (N 2 ). The lifetime of NOy of more than 10 years in the lower stratosphere is reduced to less than 3 months above 40 km, a value that is smaller than the interhemispheric transport time scale. In this case, the transport of NOy across the tropical region to the Southern Hemisphere is therefore somewhat hampered by the existence of NOy loss process.
The six panels in Figure 6 show the change in zonal and annual mean ozone concentration resulting from the emissions at 30 and 40 km of water vapor, nitrogen oxides, and both combined. When only water vapor emissions are considered, the largest percent ozone reduction is found around 80 km, where ozone is very sensitive to the hydroxyl and hydroperoxyl radicals. The maximum depletion is 9% when the emissions are occurring at 30 km and 20% when they are located at 40 km. For an injection of water at 30 km, the ozone reduction is typically 2% at the stratopause (50 km) and 5% around 60-70 km altitude. A slight ozone increase of 1% calculated near 30 km latitude results from the self-healing effect of the ozone layer (Solomon et al., 1985). In the case of pure NOx emissions, the ozone reduction is larger than 10% between 20 and 35 km altitude (between 30 and 40 km in the tropics) for a 30 km injection and between 20 and 45 km (between 35 and 45 km in the tropics) for an injection at 40 km. The peak ozone reduction is 30% and 37%, respectively.
When considering the combined effects of NOx and H 2 O emissions, we calculate a maximum reduction in the zonal annual mean ozone concentration of the order of 25% (1.5 ppmv) and of 35% (2.5 ppmv) for flight altitudes of 30 and 40 km, respectively.

Earth's Future
The change in the ozone column for individual and combined injections of water vapor and nitrogen oxides at 30 and 40 km is shown in Figure 7 as a function of latitude and season. As expected, the largest depletion of ozone takes place in the Northern Hemisphere, where most aircraft operate. At the high northern latitudes, the reduction in the ozone column is of the order of 8-10% (25 Dobson units) for a 30 km injection  and close to 5-8% (20 Dobson units) for a 40 km injection. Most of the ozone reduction is due to the release in the atmosphere of NOx. The maximum percentage reduction (10-12%) occurs in the polar region in spring and fall. The maximum reduction at high latitudes in the Southern Hemisphere is of the order of 2%. Again, these quantitative values are directly linked to the adopted emission scenarios.
In order to provide values with little dependence on the adopted magnitude of emissions by HST, we present in Tables 2 and 3 the global and Northern Hemispheric average ozone reduction relative to an annual injection per Tg of NOx and per Tg of H 2 O, respectively. Note that these ozone changes refer to emissions integrated over the entire Earth and the Northern Hemisphere, respectively. All presented values are annual average, zonal mean, for year 2050.
These tables show that, per Tg/year of injected substance, the ozone column is considerably more sensitive to nitrogen oxide than water vapor. The ozone reduction in response to water vapor takes place primarily in the upper atmosphere with little effect on the ozone column. Further, the ozone sensitivity to nitrogen oxides is larger when the flight altitude is located at 30 rather than 40 km. We note, in particular, that per Tg/year of injected NO 2 at 30 km injection, the average ozone depletion calculated in the Northern Hemisphere is as large as 24.5 DU or 8%. For an injection at 40 km, this reduction is 14.5 DU or 4.7%.
Reductions can be locally larger, specifically at high latitudes.
Finally, in order to assess the perturbation in the chemical budget of odd oxygen as a function of altitude, we show in Figure 8 the Northern Hemisphere annual averaged change in the different contributions to the Ox loss rate function of altitude. The top panels depict the loss rates for a 2050 reference atmosphere, in percent of total Ox loss and in absolute value. They show that at the cruise altitude planned for the HST, more than 50% of the background ozone loss results from the catalytic destruction by NOx. The contribution by water vapor becomes dominant only above 45 km altitude and remains such in the upper atmosphere. Figure 8 also shows that the injection by aircraft of water or nitrogen oxides modifies the Ox loss rate by other reactive species, which highlights the existence of chemical interactions between the different ozone catalytic destruction cycles. This is especially evident for the 40 km injections (i.e., H 2 O only, NOx only, and combined). Here the H 2 O only simulation increases HOx (see Figure S1) and increases the HOx family chemical loss (bottom row, left panel of Figure 8). The increase in HOx decreases the NOx family chemical loss by the reaction of OH + NO 2 forming nitric acid (HNO 3 ). This process is converting the more Earth's Future catalytically active NO 2 molecule into a longer-lived molecule (HNO 3 ) and therefore reducing the overall NOx family odd oxygen loss. When NOx only is emitted, the Ox, ClOx, BrOx, and HOx odd oxygen loss families all decrease relative to the H 2 O only simulation. Again, this is due to interfamily coupling reactions removing active odd oxygen loss species from these families. However, the overall all odd oxygen loss increases due to the NOx family, and substantial ozone depletion is derived (Figures 6 and 7).

Chemical-Dynamical Interactions
The results described in the previous sections are based on model simulations in which the dynamical fields (winds and temperature) are specified as model input and are therefore insensitive to changes in the chemical composition of the atmosphere and specifically to the HST injection. Deriving the full dynamical response to changes in stratospheric ozone concentrations and specifically the response of the atmospheric circulation to the HST operations requires comprehensive Earth system model simulations in which the dynamics of the atmosphere is simulated consistently with changes in chemical and related radiative processes ("online model"). Due to the internal dynamical variability of the atmosphere, an ensemble of several (typically 10 or more) model realizations with slightly different initial conditions should be produced for the background atmosphere as well as for the two perturbed cases. Such computationally expensive ensemble could be performed as part of a more extensive modeling assessment exercise. Here, in order to highlight the potential importance of temperature and circulation feedbacks, we perform two additional simulations in which chemical-dynamical interactions are taken into account in a reference case and in a perturbed case (with a 30 km

10.1029/2020EF001626
Earth's Future injection by the HST fleet as considered in the previous sections), respectively. Although only one realization of the model dynamics is represented for each case, these two free-running (FR) calculations provide some information on the magnitude of interannual variability, the perturbation in the temperature structure with the related slow-down of the catalytic ozone destruction, the change in the strength of the Brewer-Dobson circulation, and with the intrusion of water vapor through the cold trap of the tropical tropopause.
The differences between the time variation of ozone derived by specified dynamics and calculated dynamics are substantial. For example, in the Arctic at 78°N, the relative difference in the ozone column abundance between the perturbed and background cases during the last 5 years of the model integration varies from −7% to −11% in the SD case and from +8% and −23% in the FR case (the difference between two single realizations). When averaged over a period of 5 years (model years 2047 to 2051) to reduce the effect of interannual variability, the changes in the zonally averaged ozone concentration and temperature are those represented in Figure 9. First, as expected, in the FR case, we note a cooling of the Northern Hemisphere upper stratosphere that reaches 3 K at 30 km altitude. This temperature decrease results from a reduction in the absorption of solar ultraviolet radiation by ozone. The mean ozone depletion calculated in the FR simulation is smaller than in the SD case (maximum ozone reduction of 18% versus 28%) due to a large extent to the positive relationship between photochemical ozone destruction and temperature.
Large differences between the specified dynamics and free-running cases are also noticeable in the vicinity of the tropopause where large vertical ozone gradients and a long chemical lifetime makes ozone particularly sensitive to dynamical variability. Stone et al. (2018) showed from an ensemble of model simulations that the response of ozone to external perturbations is very variable in the lower stratosphere, particularly in the tropics. In the realization shown in Figure 9, the ozone concentration increases by up 5% in the vicinity of the tropopause. This value could be very different in other model realizations.

Conclusions
As shown by the model simulations presented here, a fleet of hypersonic aircraft that is being considered for future intercontinental travel has the potential to substantially reduce the atmospheric column abundance if sufficiently large amounts of nitrogen oxides are released in the atmosphere. The impact of water vapor on the ozone column is found to be considerably smaller. The study shows that, for a given amount of emissions, the impact on ozone is larger if the flight altitude is located at 30 km (closer to the ozone maximum) rather than 40 km altitude. As expected, the effects are more pronounced in the Northern Hemisphere, where the largest fraction of the fleet will be operated. However, due to the long-range transport aircraft effluents and probable local emissions, the potential effects on ozone in the Southern Hemisphere cannot be ignored. The calculated ozone depletion is largest in the polar region of the Northern Hemisphere and is most pronounced in spring (March and April) and fall (October and November). In the Southern Hemisphere, its largest effect occurs in Antarctica during the period of the ozone hole formation, but the amplitude is considerably lower than the ozone reduction associated with the development of hole.

Earth's Future
The projections derived with a free-running chemistry-climate model show qualitatively the same results as the simulations performed with fixed dynamics. However, the maximum ozone depletion resulting from the injection of engine effluents at 30 km altitude is reduced by about a third due to the chemistry/temperature feedback in this region of the stratosphere. Further, the ozone response in the vicinity of the tropopause appears to be strongly affected by dynamical variability. A full analysis of the mean ozone change near the tropopause should be performed on the basis of an ensemble of a large number of model realizations.
The impact on the atmosphere of a potential future fleet of hypersonic aircraft will depend on the fleet size, the type of operations, the fuel consumption by each aircraft, the type of fuel to be used, and the emission indices for the different effluents released by the engines. If HST engines release nitrogen oxides in substantial quantity, the average level of biologically harmful ultraviolet radiation (UV-B) radiation could increase and produce detrimental effects on the biosphere; in particular, it could enhance the number of skin cancers affecting humans.