4.3.3. Model Results for Supersonic Aircraft Emissions

4.3.3.1. NOy and H2O Enhancement

Supersonic aircraft emissions of NOx and H2O were incorporated within six 2-D models and three 3-D models. These supersonic aircraft emissions are primarily deposited within the Northern Hemisphere LS. Figure 4-6 shows calculated perturbations in NOy and H2O from HSCT emissions in June 2015 with EI(NOx)=5 (scenario S1c-D). Maximum perturbations occur in the Northern Hemisphere LS and range from 0.6 to 1.0 ppbv for NOy and from 0.4 to 0.7 ppmv for H2O. Based on the calculated perturbation in NOy and H2O, one can conclude that the transport fields of these multi-dimensional assessment models are significantly different. For example, the GSFC model shows a significantly larger calculated perturbation in Northern Hemisphere lower stratospheric NOy than the other models (20-40% higher peak NOy abundance). This model also transports more NOy and H2O higher into the middle stratosphere (MS) and into the Southern Hemisphere than the other models. By contrast, the AER model tends to isolate supersonic emission increases in NOy and H2O within the Northern Hemisphere LS. The THINAIR and LLNL models calculate a NOy perturbation from supersonic aircraft in the Northern Hemisphere LS that approaches that of the GSFC model. However, these models transport much less aircraft-enhanced NOy and H2O to the Southern Hemisphere. CSIRO model NOy and H2O perturbations are similar to those of LLNL, except that the maximum NOy abundance increase in the Northern Hemisphere LS is significantly less. The SLIMCAT model shows greater NOy and H2O enhancements in the Southern Hemisphere than the GSFC model but smaller Northern Hemisphere perturbations. Aircraft-induced NOy enhancements in the middle stratosphere are greatest in the THINAIR, GSFC, and SLIMCAT models for Northern Hemisphere and tropical regions. Overall, the AER, CSIRO, LLNL, and THINAIR models show the greatest Northern Hemisphere isolation of supersonic aircraft effluent among the participating assessment models. The UNIVAQ, LARC, and SCTM1 models derive a maximum Northern Hemisphere LS NOy abundance increase similar to CSIRO. However, these models derive relatively small horizontal gradients in the LS horizontal spread in NOy and H2O, extending from Northern Hemisphere mid-latitudes to Southern Hemisphere high latitudes, suggesting a weak tropical/mid-latitude barrier to mixing. The UNIVAQ, SLIMCAT, SCTM1, and LARC models generate Southern Hemisphere LS NOy and H2O magnitudes that are similar to those of the GSFC model. To first order, the spread in model-derived changes in total O3 for the ensemble of scenarios shown in Tables 4-11 and 4-12 is a direct result of large differences between supersonic aircraft-induced perturbations in the NOy and H2O fields.

4.3.3.2. Profile and Column Ozone Change

The percentage change in profile O3 for June is shown in Figure 4-6c for scenario S1c-D. Above 25 km, in all models, local O3 abundance is reduced as a result of including supersonic aircraft emissions of NOx and H2O. This result is not surprising because all assessment models represent the NOx (middle stratosphere) and HOx (upper stratosphere) chemical families as prime contributors to odd-oxygen loss above 25 km, in approximately the same relative importance. Therefore, sensitivity to NOx and H2O emissions from supersonic aircraft exhaust is similar in these regions and roughly proportional to calculated perturbations in NOy and H2O. Below 25 km, the relative odd-oxygen loss partitioning between NOx, HOx, and ClOx-BrOx chemical families is highly model-dependent. This situation is evident in the calculated local O3 change. Models that are more HOx dominant in the LS tend to show a more positive response to added aircraft NOx. The opposite is true for models that are more NOx dominant. These variations are a result of different methodologies of transport and chemistry incorporated in the individual models.

The percentage change in total column O3 is shown in Figure 4-6d for scenario S1c-D. Six of the models (AER, CSIRO, GSFC, LLNL, SCTM1, and UNIVAQ) indicate smaller depletions in the tropics and larger depletions at mid- and high latitudes. The other three models (LARC, SLIMCAT, and THINAIR) show more complicated features. The LARC model predicts increases in Northern Hemisphere mid- and high latitudes and decreases in the tropics. The SLIMCAT model predicts decreases of more than 0.6% in the tropics, with less depletion predicted at Northern Hemisphere mid- and high latitudes and Southern Hemisphere mid-latitudes. The THINAIR model generally predicts larger depletions at lower latitudes. Decreases in Northern Hemisphere polar latitudes of greater than 1% are shown in the AER model results. The GSFC and SLIMCAT models predict decreases in the Southern Hemisphere polar spring of greater than 1%.

4.3.3.3. Column Ozone Sensitivity to Supersonic Emission of NOx and H2O

Sensitivity studies of supersonic engine emissions of H2O and NOx were carried out. These studies investigated individual and combined impacts on O3 from H2O and NOx emissions within an atmosphere that was volcanically clean (SA0). In Figure 4-7a, the change in annual average Northern Hemisphere total column O3 as a function of EI(NOx) is shown for a fleet of 500 supersonic aircraft in the year 2015 (Scenarios S1b-e). The reference atmosphere included emissions from a representative 2015 fleet of subsonic aircraft. Under the conditions of these scenarios, the models predict that water emissions [corresponding to EI(NOx)=0] have the most significant impact. For EI(NOx)=0, the participating models all derive a total O3 depletion in the 0.2-0.6% range.

For low emission indices of NOx [EI(NOx)=5 to 10], the predicted decrease in O3 is dominated by emitted H2O; in most models, the addition of NOx emissions leads to less predicted O3 depletion. Because O3 loss in the lower stratosphere tends to be dominated by HOx constituents and supersonic aircraft emissions are most important in the lower stratosphere (see Figure 4-6), the predicted importance of H2O emissions is understandable.

Several models show relatively little sensitivity to NOx emissions over the EI(NOx)=5 to 15 range. Including NOx emission at an EI(NOx)=5 level, S1c had little effect on total column O3 change for the LLNL, GSFC, and CSIRO models [relative to EI (NOx)=0]. For the AER, SLIMCAT, and UNIVAQ models, added NOx buffers the larger total O3 depletion derived when H2O is the only emitted constituent. When NOx emission is included, six of the models (CSIRO, GSFC, LLNL, SCTM1, SLIMCAT, and THINAIR) calculate a larger O3 depletion as EI(NOx) increases from 5 to 15 (S1c-e). The AER model showed little sensitivity to EI(NOx) in this range. Within the same range, the LARC and UNIVAQ models showed a positive O3 change slope. In general, the model-derived O3 change is in better agreement among participating models at EI(NOx)=5, with a spread among models from approximately 0% total column O3 depletion (UNIVAQ) down to -0.4% (GSFC and SLIMCAT). At EI(NOx)=15, the overall spread in model-derived total column O3 change is much larger (+0.4% to -0.9%).

There is also a large spread in interhemispheric gradients in model-derived total column O3 change (see Tables 4-11 and 4-12). For example, when the emission scenario is S1c [EI(NOx)=5 with 500 supersonic aircraft], the AER, SCTM1, and CSIRO models derived a larger depletion in total column O3 in the Northern Hemisphere than in the Southern Hemisphere. The UNIVAQ model had very little change in column O3 in either hemisphere. The LLNL and THINAIR models derived a similar change in total column O3 in both hemispheres. However, the ratio of Southern Hemisphere to Northern Hemisphere change in total column O3 is greater than 1.5 in the GSFC, LARC, and SLIMCAT models. This larger depletion in the Southern Hemisphere correlates with increased H2O and NOy being transported to the Southern Hemisphere and the sensitivity of cold aerosol processes implemented within the GSFC, LARC, and SLIMCAT models.

4.3.3.4. Ambient Surface Area Density Sensitivity

Volcanic activity within the past several decades has greatly modulated ambient SAD. Two major volcanic eruptions-El Chich�n (April 1982) and Mt. Pinatubo (June 1991)-are responsible for most of the observed SAD change. Satellite instruments (SAM, SAGE I, SAGE II, and SME; e.g., Thomason et al., 1997) have shown that average SAD in the lower stratosphere between 1979 and 1995 was greater than that during volcanically clean periods (SA0; WMO, 1992) by a factor of 8 in the equatorial region and a factor of 2 to 4 at higher latitudes. These enhancements decay with e-folding time scales on the order of 1 to 2 years (Yue et al., 1991; Thomason et al., 1997).

For this assessment, we examined the impact of supersonic aircraft at a SAD of four times the volcanically clean condition (4xSA0). Results from these higher SAD sensitivity studies are shown in Figure 4-7b. Unlike the volcanically clean ambient atmosphere (Figure 4-7a), when the EI(NOx) varies between 0 and 15 (S3a-d), all participating models derived a positive slope to the Northern Hemispherical change in total column O3. All models show a strong interference of the H2O only [EI(NOx)=0] total O3 depletion when NOx emissions are included [EI(NOx)=5]. This response is consistent with the idea that at higher ambient SAD, NOx catalytic processes are less important, mainly as a result of conversion of active NOx radicals to nitric acid by heterogeneous reactions on sulfate aerosols. As supersonic aircraft NOx is enhanced, the additional NOx interferes with ClOx, BrOx, and HOx odd-oxygen loss processes, reducing the net total odd-oxygen loss.

It should be noted that the model-derived impact on total column O3 from increasing ambient SAD by natural causes (approximately 2-3% decreases in Northern Hemisphere mid-latitude total column O3 change; Solomon et al., 1996) exceeds the model-derived impact from supersonic aircraft (typically less than 1% Northern Hemisphere mid-latitude total column O3 change).

4.3.3.5. Sulfate Aerosols and Polar Stratospheric Cloud Sensitivities

Heterogeneous reactions are important in defining the O3 removal rate in the stratosphere. These reactions occur on sulfate aerosols and PSC particles, though sulfate aerosols have the greater influence because of their global distribution. New sulfate aerosols are produced by aircraft emissions; sulfate particles are generated by plume processing of part of the emitted SO2/SO3 (Fahey et al., 1995; Brown et al., 1996; Danilin et al., 1997; K�rcher, 1997; K�rcher and Fahey, 1997). The remaining gas-phase SO2 is then transported on the atmospheric large/global scale, where it is oxidized to sulfuric acid and condenses onto preexisting sulfate aerosol particles (Bekki and Pyle, 1992; Pitari et al., 1993; Weisenstein et al., 1998). Aircraft also emit H2O and NOx and therefore have the potential to increase the saturation ratios of water vapor and HNO3, so that the size distribution, mass, and SAD of PSCs may be perturbed (Peter et al., 1991; Pitari et al., 1993; see Section 3.3.6).

4.3.3.5.1. Sulfate aerosol sensitivity-SO2 gas-to-particle conversion in the supersonic aircraft plume

Supersonic aircraft consuming sulfur-containing fuel will have some effect on sulfate aerosol amounts in the stratosphere (see Section 3.7.3). In this assessment, the increased SAD from sulfur conversion to particles in the plume had a significant impact on stratospheric ozone. Figure 4-8 shows the percentage change in annual average Northern Hemisphere total column O3 correlated to the percentage increase in SAD for given SO2 gas-to-particle conversion assumptions. The percentage SAD increase is derived by averaging the SAD change relative to the volcanically clean atmosphere (SA0) within the region between 14-21 km and 33-90�N. The SAD percentage increases for 0, 10, 50, and 100% SO2 gas-to-particle conversion assumptions in the plume are 27, 38, 82, and 111%, respectively. The 0% SAD increase on the abscissa is taken from supersonic scenario S1c; it represents a perturbation without any consideration of aircraft sulfur emissions. The AER model was the only model to compute the 0% SO2 gas-to-particle conversion impact on total O3 change. Here, the additional aircraft-emitted SO2 gas increased the ambient SAD by 27%, altering the total O3 decrease for the AER model from 0.3 (SA0) to 0.6% (SA7). When particle conversion was assumed, all participating models derived relatively large decreases in total O3. In fact, AER and GSFC model results for the Northern Hemisphere showed O3 depletions of greater than 1% when a 50% SO2 gas-to-particle conversion efficiency was assumed.

4.3.3.5.2. PSC sensitivity

The change in SAD of PSCs from HSCT aircraft emissions has been calculated by the AER, GSFC and UNIVAQ 2-D models. The surface area increase is a consequence of the stratospheric injection of H2O and NOy from aircraft. This process has the effect of increasing the saturation ratios of both water vapor and nitric acid, producing an enhancement of PSC particle mass and SAD. Changes in H2O and NOy from HSCT aircraft at 18 km and 60�N and 60�S are listed in Tables 4-13a and 4-13b. The average additional SAD of PSC1+PSC2 aerosols in the 12-24 km altitude layer and poleward of 60�N during December-January-February (DJF) or poleward of 60�S during June-July-August (JJA) are shown in Tables 4-13c and 4-13d. These results should be interpreted with caution because PSC distributions are highly localized.

All three models assume PSC1 to be solid NAT particles, although the treatment for PSC1 and PSC2 formation varies from model to model. AER and GSFC use a thermodynamic equilibrium between gas and condensed phase for HNO3 and H2O (Hanson and Mauersberger, 1988), with an imposed size distribution. Supersaturation ratios of 10 and 1.4 are required before PSC1 and PSC2 form in the GSFC model, thus lowering the temperature threshold for particle nucleation by about 3 and 2 K, respectively. Denitrification and dehydration are included in the models by calculating HNO3 and H2O loss terms from sedimentation. The UNIVAQ model makes the calculation by integrating on the NAT/ice particle size distribution; the AER and GSFC models assume that all condensed HNO3 and H2O form NAT and ice particles of a prescribed size distribution. No supersaturation is assumed in the AER model. In these two models, perturbations of sulfate aerosols from aircraft emission of SO2 leave unaffected the PSC. The models do allow H2O and HNO3 changes from aircraft to affect PSC. The UNIVAQ model allows interaction between sulfate aerosols and PSC by adopting a microphysical code for both types of particles (Pitari et al., 1993). As the atmosphere cools, PSC1 particles are first formed by nucleating on preexisting sulfate aerosols; below the frost point, PSC1 become coated with ice, thus forming PSC2 particles. With this scheme, a simultaneous presence of sulfate, PSC1, and PSC2 particles is possible, and size distributions are calculated for all particle types.

The spread in absolute change of PSC SADs shown in Tables 4-13c and 4-13d is not surprising because of the different schemes for PSC formation adopted in the models. In particular, the absence of supersaturation assumptions in the AER model consistently produces larger PSC SAD. The spread in relative changes is closely related to different relative increases of H2O and NOy. About 50-60% (70-90%) of the calculated PSC perturbation in the Northern Hemisphere (Southern Hemisphere) is caused by HSCT H2O emissions when EI(NOx)=5. The more important role of H2O emissions in the Antarctic winter is a consequence of persistent PSC2 particles there; in the Arctic winter, PSC1 particles contribute to almost the entire available PSC SAD.

4.3.3.5.3. Column ozone sensitivity to PSCs

Representation of heterogeneous processes on cold aerosols (PSCs) is not consistent among the participating models (Sections 4.3.1.3 and 4.3.3.5.). These differences are manifest when increased abundances of NOY and H2O from supersonic aircraft are considered. As mentioned previously, the transport fields of participating assessment models are significantly different; therefore, the amount of supersonic H2O and NOy dispersion between hemispheres and vertically within the same hemisphere is also different. Three models (AER, GSFC, and UNIVAQ) investigated the impact on total O3 when cold aerosol representation was not included (S4b). These models removed their representation of SAD, dehydration, and denitrification that would be associated with cold aerosol processes. The results were mixed. When cold aerosol processes were included in the model chemistry formulation, the model-derived percentage change in Northern Hemisphere total column O3 increased by factors of 1.1 and 1.5 for the GSFC and AER models, respectively (S1c vs. S4b). However, in the Southern Hemisphere the GSFC model-derived depletion in total O3 increased by 2.5 when cold chemistry processes were included. There was minimal change in Southern Hemisphere average O3 depletion in the AER model. When the EI(NOx) is set to zero (S4a), a similar conclusion is also drawn (compare to S1b) by the AER and GSFC models. These results are consistent with the dispersion of supersonic H2O and NOy between the AER and GSFC models. The UNIVAQ model did show sensitivity to supersonic emissions in the total O3 hemispheric averages; because the absolute impact was small and centered around zero, however, it is difficult to draw any significant conclusions.

4.3.3.6. Ambient Chlorine Sensitivity

Total radical chlorine (Cly) abundance in the atmosphere is predicted to decrease in the future as a result of international agreements controlling production of anthropogenic organic source gases such as chlorofluorocarbons (CFCs) and hydrochlorofluorocarbons (HCFCs) (WMO 1995, 1999). Therefore, several sensitivity studies were performed to investigate the impact of ambient Cly levels on a given supersonic aircraft perturbation. Here, atmospheric chlorine amounts were varied by changing halocarbon boundary conditions. Four different ground halocarbon amounts were considered, which resulted in reactive Cly levels of 1.0, 2.0, 3.0, and 4.0 ppbv in the upper stratosphere. For reference, the 3.0 and 2.0 ppbv Cly levels are representative of predicted 2015 and 2050 atmospheric conditions, respectively. In all Cly sensitivity studies, only EI(NOx)=5 and a fleet of 500 supersonic aircraft were considered. Sensitivity studies were conducted under volcanically clean conditions without aircraft sulfur emissions (SA0) and with sulfur emissions assuming 50% SO2 gas-to-particle conversion in the plume (SA1). Without aircraft sulfur emissions, there was little discernible sensitivity to background Cly abundance (S1c, S5a, S6a, and S9a). With sulfur emissions included, however, total ozone depletion for all three participating models was smallest when Cly abundance was 1.0 ppbv. As Cly increased in abundance, the model-derived ozone depletion increased, maximizing and leveling off between 3.0 and 4.0 ppbv Cly (S1f, S5b, S6b, and S9c). Among the participating models, AER showed the most sensitivity to Cly abundance, followed by GSFC and UNIVAQ. It should be noted that the Cly=2 ppbv scenario (S9c) had N2O and CH4 boundary conditions consistent with the year 2050. The other three Cly sensitivity scenarios (S1f, S5b, and S6b) had N2O and CH4 abundances consistent with the year 2015. The AER model did investigate Cly sensitivity to N2O and CH4 boundary conditions by repeating the Cly=2.0 ppbv scenario with N2O and CH4 values representative of 2015 conditions. The model-derived change in NH total O3 was -0.6% (instead of -0.7% under 2050 CH4 and N2O conditions).

4.3.3.7. Fleet Design

4.3.3.7.1. Supersonic cruise altitude sensitivity

Figure 4-9 (S1c, S2a, and S2b) shows the effect of lowering or raising the supersonic cruise altitude by 2 km relative to the baseline flight cruise altitude band of 17-20 km for an assumed fleet of 500 EI(NOx)=5 aircraft. In most altitude sensitivity studies, ambient SAD was considered to be volcanically clean, and sulfur emission by HSCT aircraft was not included. The model-derived change in annual average Northern Hemisphere total column O3 had a large dependence on emission altitude in participating assessment models. This cruise altitude impact on O3 highlights the fact that as NOx is emitted lower in the stratosphere, the residence time of supersonic aircraft emitted NOx is significantly shorter. In addition, at a lower cruise altitude the chemical regime is quite different, and past modeling studies have shown that a crossover point exists where an additional NOx molecule will produce O3 based on the well-known CH4-smog mechanism (Johnston and Quitevis, 1975). When the cruise altitude is shifted higher, emissions have a longer stratospheric lifetime and can be transported more readily to regions dominated by NOx destruction of O3.

Each of these models showed less total column O3 depletion when the cruise altitude was shifted down by 2 km. Three models also investigated sensitivity to higher values of EI(NOx) (10 and 15) at the lower cruise altitude (S2c-d). The model-derived total column O3 change generally increased in the direction of higher EI(NOx). These results are consistent with the 1995 NASA supersonic assessment (Stolarski et al., 1995).

Two models (AER and UNIVAQ) did look at sensitivity to cruise altitude with increased SAD from sulfur emissions as 50% particles (S7c). In this scenario, an enhanced SAD was not supplied but derived within the 2-D model, coupled with a sulfate microphysical model. The model-derived annual average Northern Hemisphere depletion in total column O3 was reduced from -1.0 (S7a) to -0.6% (S7c) in the AER model and from -0.5 (S7a) to -0.2% (S7c) in the UNIVAQ model when the supersonic cruise altitude was lowered by 2 km. The opposite impact was derived by the AER model when the cruise altitude was raised 2 km. Here, the model-derived annual average Northern Hemisphere depletion in total O3 increased from -1.0 (S7a) to -1.4 % (S7b) in the AER model and from -0.5 to -0.7% in the UNIVAQ model.

4.3.3.7.2. Supersonic fleet size sensitivity

In this assessment, fleet sizes of 500 and 1,000 supersonic aircraft were considered at EI(NOx) values of 5 and 10. The increase in fleet size from 500 to 1,000 aircraft included additional routes as well as additional aircraft; as a consequence, the geographical distribution of emissions changes with proportionally more flights in the tropics (Baughcum and Henderson, 1998). In addition, by increasing the fleet size, the amount of H2O emitted is also increased proportionally. For most of the cases considered, doubling the fleet size had a nearly linear effect on the Northern Hemisphere column O3 reduction. The exception is the UNIVAQ model, which showed almost no change with increasing fleet size. For the 2050 atmosphere with aircraft sulfur emission as 50% particles, increasing the fleet size from 500 to 1,000 aircraft has a less than linear effect, especially with the AER model.

IPCC Homepage