A mitigation strategy for commercial aviation impact on NOx‐related O3 change

An operational mitigation strategy for commercial aircraft impact on atmospheric composition, referred to as the turboprop replacement strategy (TRS), is described in this paper. The global air traffic between 2005 and 2011 was modeled with the TRS in which turbofan powered aircraft were replaced with nine chosen turboprop powered aircraft on all routes up to 1700 nautical miles (NM) in range. The results of this TRS double the global number of departures, as well as global mission distance, while global mission time grows by nearly a factor of 3. However, the global mission fuel and the emissions of aviation CO2, H2O, and SOx remain approximately unchanged, and the total global aviation CO, hydrocarbons (HC), and NOx emissions are reduced by 79%, 21%, and 11% on average between 2005 and 2011. The TRS lowers the global mean cruise altitude of flights up to 1700 NM by ~2.7 km which leads to a significant decrease in global mission fuel burn, mission time, distance flown, and the aircraft emissions of CO2, CO, H2O, NOx, SOx, and HC above 9.2 km. The replacement of turbofans with turboprops in regional fleets on a global scale leads to an overall reduction in levels of tropospheric O3 at the current estimated mean cruise altitude near the tropopause where the radiative forcing of O3 is strongest. Further, the replacement strategy results in a reduction of ground‐level aviation CO and NOx emissions by 33 and 29%, respectively, between 2005 and 2011.


Introduction
Short-haul flight (intracontinental and domestic missions) accounted for 92% of all global departures recorded, and over a half of the estimated total global emissions of CO 2 and NO x between 2005 and 2011 [Wasiuk, 2014]. The short-haul flight generally climbs out to the desired cruise altitude, spends a relatively small proportion of the total mission time in cruise, and descends [Erzberger et al., 1975]. However, long-haul missions (mostly intercontinental) require the aircraft to carry large volumes of fuel on board contributing to an increased aircraft weight, which in turn leads to higher fuel burn and emissions [Wasiuk et al., 2015]. In terms of fuel consumption, climb is the most expensive part of any flight. Flight is always more efficient in cruise as, by design, the cruise altitude is the altitude at which aircraft burn the least amount of fuel. Propulsion for most commercial aviation aircraft is provided by turbofan or turboprop engines. Both are air-breathing engines in which the turbine produces work to run the compressors: in a turbofan, the turbine also rotates the fan-housed upstream of the hot section of the engine, and this fan produces most of the thrust; in a turboprop, the turbine drives a propeller which generates the majority of the thrust. Turboprop aircraft are better suited than turbofan for short-haul missions as they spend more time in cruise and are noted for their low fuel consumption [Babikian et al., 2002]. But the efficient technology (e.g., turboprop) was abandoned in favor of the less efficient and less environmentally friendly technology (e.g., turbofan) in the 1980s because of the low fuel cost, the advantage of higher flight speeds, and hence utilization. Turbofans are also chosen in favor of turboprops due to a "level of passenger service in the form of comfort and perceived safety" [Ryerson and Hansen, 2010]. The findings in Ryerson and Hansen [2010] indicated that rising fuel prices could reverse this trend in the short-haul markets, and they concluded that high fuel costs could potentially overshadow the importance of passenger convenience. There is a dramatic drop in fuel cost over recent years; however, the environmental cost could tip the advantage in favor of the turboprop. This can be coupled with the findings on the availability of fuel by Nygren [2008] which suggest that aviation fuel production is predicted to decrease by several percent a year after the crude oil production peak is reached. The growth in demand for air traffic, coupled with falling availability of aviation fuel, "envisages a substantial lack of jet fuel by the year 2026." turbofans, which on a per seat basis are less fuel efficient [Ryerson and Hansen, 2010]. This represents in a significant move away from a more environmentally beneficial option. If turboprop aircraft can be utilized on short-haul missions instead of turbofans, there is potential for increasing fuel efficiency of regional air traffic on a regional as well as global scale. Such a substitution across regional fleets on a global scale would cause the global mean cruising altitudes to be lowered.
The aircraft emissions of NO x , CO, hydrocarbons (HC), and particles are of concern in terms of local air quality [Yu et al., 2004;Peace et al., 2006], and the emissions of CO 2 , H 2 O, NO x , SO x and particles are of concern in terms of global climate change [Rogers et al., 2002;Köhler et al., 2013;Gilmore et al., 2013;Skowron et al., 2015]. Aircraft emissions of CO 2 and H 2 O released at commercial cruise altitudes can contribute directly to climate change by increasing the levels of greenhouse gases in the upper troposphere and lower stratosphere. Aircraft NO x emissions have an indirect effect on our climate via tropospheric O 3 production and through removal of CH 4 [Olsen et al., 2013;Wasiuk et al., 2016a]. Aircraft SO x emissions can be oxidized in the atmosphere to form sulfate aerosol particles, which also contribute to climate change (notably through cooling) [Pitari et al., 2002]. Contrails from aircraft engines increase cloud cover directly or indirectly causing a positive mean radiative forcing at the top of the troposphere [Schumann, 2002;Burkhardt and Kärcher, 2011].
Utilizing a turboprop instead of a turbofan on a short-haul mission reduces CO 2 and NO x emissions because of the lower fuel burn [Babikian et al., 2002]; thus, there is a potential for environmental and atmospheric benefit in substituting turbofans with turboprops on short-haul missions. An 8% of all global departures between 2005 and 2011 were estimated as long-haul departures, but this 8% accounted for nearly half of the total global mission fuel burn, CO 2 and NO x emissions in this time period [Wasiuk, 2014]. Moreover, between 2005 and 2011 long-haul departures increased continuously, by~40% in total [Wasiuk, 2014]. In the light of these findings, a substitution of turbofans with turboprops on long-haul missions could potentially generate a significant saving in terms of the fuel burn, CO 2 , and NO x emissions. Due to the lack of a design specification for a possible future long-range turboprop aircraft, we investigated a scenario in this study where all turbofans were replaced only on short-haul routes with existing turboprop aircraft.
In our recent study [Wasiuk et al., 2015[Wasiuk et al., , 2016b, Aircraft Performance Model Implementation (APMI) software containing a database of global aircraft movements, a model of aircraft performance for all phases of flight and an aircraft emissions estimation method was used to create a 4-D Aircraft Fuel Burn and Emissions Inventory for the time period of 2005 to 2011. In this study, the replacement of nine selected turbofan aircraft by turboprops for flights up to 1700 nautical miles (NM) was made to build an amended 4-D Aircraft Fuel Burn and Emissions Inventory, referred to as the turboprop replacement strategy (TRS) inventory for short. Fuel burn and emissions of flights over 1700 NM were unchanged in the TRS inventory. The resulting emissions were redistributed and the global seasonal TRS 3-D NO x Emissions Distribution Fields were created. These NO x fields were used as an input into the 3-D global Lagrangian chemical transport model (CTM); STOCHEM-CRI and sensitivity simulations (SS) were performed for each year between 2005 and 2011 to investigate the impact of TRS relative to the reference inventory on tropospheric composition. In this study, the estimates of the mission fuel burn, mission time, mission distance, and the emissions of CO 2 , CO, H 2 O, HC, NO x , and SO x due to the replacement of turbofans on short-haul missions by a current make of turboprop airliner were compared with those due to the original fleet. The global burden and tropospheric distribution of NO x and O 3 under the TRS is also compared with that due to the original fleet.

Four-Dimensional TRS Aircraft Fuel Burn and Emissions Inventory
The air traffic movement statistics database from 2005 to 2011 was mined from the global airline schedules data, CAPSTATS (http://www.capstats.com/). The Aircraft Performance Model Implementation (APMI) described in Wasiuk et al. [2015] was used to generate a 4-D TRS Aircraft Fuel Burn and Emissions Inventory from the aircraft database. In the TRS inventory, all unique route and aircraft combinations with a mission distance ≤ 1700 NM (3148 km) were extracted for the time period of 2005 to 2011 and classified according to the aircraft type (turbofan/turboprop) used on the routes as summarized in Figure 1. Each unique route with a mission distance ≤ 1700 NM serviced with a turbofan was considered for replacement. Nine turboprops from an aircraft database were selected to be used as the replacement aircraft. International Civil Aviation Organization  (Table 1). Only turboprop aircraft with a passenger capacity greater than 40 were selected in order to minimize the number of departures necessary to carry the original volume of passengers. All unique routes and turbofan aircraft combinations selected for replacement were parsed and all unique input parameter triples (the ICAO aircraft code, the mission type, and the mission distance) were extracted. Mission type and mission distance duplicates were removed, and each unique pair was assigned a turboprop from the turboprop aircraft available for replacement.
The new parameter triples in TRS were used as input into the APMI which assigned mission fuel burn and the emissions of CO 2 , CO, H 2 O, HC, NO x , and SO x to a simulated flight trajectory. The output from the APMI was used to update all the entries in the reference case (RC) inventory (the inventory with the original fleet composition) that were selected for replacement. A multiplication factor was used to calculate the number of turboprop departures required on each route selected for replacement in order to transport approximately the same volume of passengers as with the turbofans. The original number of departures associated with a unique route and aircraft combination selected for replacement was adjusted according to   (1) where departures tp is the adjusted number of departures, capacity tf is the passenger capacity of the turbofan used on route, capacity tp is the capacity of the turboprop selected as the replacement aircraft, and departures tf is the original number of departures made on route with the turbofan.
The total global aircraft NO x emissions in the TRS inventory were distributed on a 3-D 5°latitude × 5°longitude grid resolution [Wasiuk et al., 2016b]. The vertical grid resolution followed the pressure levels and approximate height bands which are based on the vertical model resolution of the 3-D STOCHEM-CRI chemistry transport model [Collins et al., 1997;Utembe et al., 2010]. The nine height bands across the 5°× 5°grid result in the physical transport system in the STOCHEM model comprising 50,000 constant mass air parcels, the centroids of which were advected, on a three hour time step, through the model [Collins et al., 1997;Derwent et al., 2008]. The advection of the parcels was on Lagrangian trajectories, using meteorological data provided by the UK Met Office. The chemical processes that occurred within the air parcel, together with emission, deposition, mixing, and removal processes were uncoupled from transport processes to enable local determination of the chemistry time step [Utembe et al., 2010]. In previous studies [Stevenson and Derwent, 2009;Stevenson et al., 2004], STOCHEM was successfully used for measuring the impact of aviation NO x emissions on climate. In this study, the new chemical mechanism is added in STOCHEM which is the Common Representative Intermediates mechanism version 2 and reduction 5 (CRI v2-R5), referred to as "STOCHEM-CRI." The detailed description of the CRI v2-R5 mechanism is given by Jenkin et al. [2008], Watson et al. [2008], and Utembe et al. [2009Utembe et al. [ , 2010.
The emission totals of 27 species including CO, NO x , and nonmethane hydrocarbons employed in STOCHEM-CRI were adapted from the Precursor of Ozone and their Effects in the Troposphere inventory [Granier et al., 2005]. Emission totals for CH 4 were taken from the inverse model study of Mikaloff-Fletcher et al. [2004], except for the ocean emissions which were taken from Houweling et al. [2000]. More details about the global emission data used in STOCHEM-CRI can be found in Khan et al. [2014] and Wasiuk et al. [2016a].
The adjusted aircraft NO x emissions from 2005 to 2011 estimated by the TRS 4-D Aircraft Fuel Burn and Emissions inventory were normalized to give global yearly aviation NO x emissions as an input to STOCHEM-CRI. A set of seven sensitivity simulations based on the volume and distribution derived from air traffic movements recorded between 2005 and 2011 were performed in which a detailed 3-D spatial distribution of the global annual aviation NO x emissions used as input into the CTM was modified to reflect the changes resulting from the replacement of turbofans with turboprops in regional fleets on a global scale. All simulations were run with meteorology from 1998 for a period of 24 months with the initial 12 months being discarded as a spin-up year. Three sets of results were obtained from the study: a TRS Inventory, global seasonal TRS 3-D NO x Emissions Distribution Fields, and TRS STOCHEM-CRI sensitivity simulation (SS) results. These are presented and compared with the reference case (RC) inventory. (roughly half that of the average for the shorter-range turbofan aircraft they are replacing). A practical strategy would entail the design and operation of larger-capacity turboprop aircraft so that number of departures would not double: hence, air traffic management problems would not be incurred; fuel burn and operating costs would be lower than the theoretical approach. From the perspective of emissions, too, the doubling of departures represents a conservative scenario.

Results and
As the emissions of CO 2 , H 2 O, and SO x are directly proportional to the amount of the fuel burned [Penner et al., 1999], the changes in the quantity of these emissions under the TRS mirror that of the mission fuel burn. The total globalaviationemissionsofCO,HC,andNO x ontheotherhandalldecrease,HCby0. More importantly, a significant saving in terms of NO x , CO, and HC emissions was achieved. Turboprops which were at most half the capacity of an average turbofan assuming a 75.5% load factor were used [Wasiuk et al., 2015], but turboprops with a much higher passenger capacity would reduce the number of departures needed. Most flights rarely fly at 100% capacity and thus substituting a turboprop with a slightly smaller capacity can increase load factor which would be an improvement in terms of efficiency and potentially lead to even greater environmental and atmospheric benefits.

Global Geographical Distribution Changes of Aviation NO x Emissions Under the TRS
The average (2005-2011) spatial distribution change of the total global aircraft NO x emissions after swapping turbofan aircraft with turboprop aircraft for short-haul missions (Figure 2) highlights the affected areas of flight activity and major global flight paths between 5.6 and 16.2 km altitude. The allocation of the total global aircraft NO x emissions between the northern hemisphere (NH) and southern hemisphere (SH) under the TRS is virtually unchanged, with~90% of the total global aircraft NO x emissions released in the NH. The distribution of the total global NO x emissions across the latitudes in the NH changes negligibly as well. The replacement of turbofans with turboprops in regional fleets on a global scale leads to changes in the global geographical distribution of the total global aviation NO x emissions between 5.6 and 7.2 km (see Figure 2a) when compared with the global geographical distribution of the total global aviation NO x emissions from the unmodified fleet. The changes between 5.6 and 7.2 km appear to be more or less randomly distributed. Between 7.2 and 9.2 km, the replacement leads to a significant increase in the level of aviation NO x emissions by up to 500%. These are concentrated in specific regions of the world as seen in Figure 2b. Above 9.2 km, the replacement leads to a decrease in the level of NO x emissions only with a maximum of 100%. A regional pattern in these changes is discernible which are shown in Figures 2c and 2d.

Global Budget of Tropospheric O 3
The substitution of turbofans by turboprops in regional fleets on a global scale decreases the global annual mean burden of NO x by 4.0 Gg (0.8%) which resulted in a decrease of global annual tropospheric burden of PAN by 1.7 Gg (0.3%). The global annual mean tropospheric burden of O 3 decreased by 0.9 Tg (0.3%) on average between 2005 and 2011, and as a consequence of decreased O 3 levels, the global annual mean tropospheric burden of OH decreases by 0.002 Gg (0.8%). Table 3 shows the global O 3 budget under the TRS in which both total production and loss decrease slightly by 11.6 Tg/yr (0.1%) and 11.1 Tg/yr (0.1%), respectively. Two dominant chemical production channels, HO 2 + NO and CH 3 O 2 + NO,  Journal of Geophysical Research: Atmospheres 10.1002/2016JD025051 decreaseby9.6 Tg/yr(0.2%)and1.7 Tg/yr(0.1%)respectively,whiletheotherchannelsremainunchanged.There is a decrease in all channels that make up the chemical loss, the most being OH + O 3 at 3.5 Tg/yr (0.5%). Net O 3 productiondecreasesby0.5 Tg/yr(0.6%)onaverageundertheTRS.IntheNO x -saturatedregime(taxiout,takeoff, approach, landing, and taxi-inphases when VOC/NO x ratio is small), the decreased NO x atground level under TRS (Figure 3f) leads to no significant change in O 3 mixing ratios. However, in the NO x sensitive regime (ascent, cruise, and descent phases when VOC/NO x ratio is high), the significant decrease in NO x at 9.2-16.2 km (Figure 3f) reduces O 3 mixing ratios notably. There is an overall reduction in levels of tropospheric O 3 after substitution of turbofans by turboprops on a global scale in regional fleets on short-haul missions which would be an improvement in terms of environmental pollution. Figure 4 shows the global zonal distribution of the average percent changes in NO x and O 3 mixing ratios under the TRS. Following the replacement of turbofans by turboprops in regional fleets on a global scale and consequently lowering the estimated global mean cruise altitude, the concentration of NO x is generally suppressed by 5-10% above 9.2 km in the NH. In particular, there are two areas of greatest negative changes: one of up to 10% between 30°S and 40°S in the SH and one of up to 25% between 25°N and 55°N in the NH at 11.8-16.2 km as seen in Figure 4. Brazil and southern coast of Australia emerge as new NO x mixing ratio hot spots during 2005-2011 [Wasiuk et al., 2016a]; the TRS replacement strategy reduces the NO x change in Brazil and Australia mostly which reflects the negative NO x change in the SH. Between 7.2 and 9.2 km NO x mixing ratios are increased by up to 8% which can be attributed to the lowering of the global mean cruise altitude under the TRS.  Following the reduction in NO x concentrations due to the substitution of turbofans by turboprops in regional fleets on a global scale, overall O 3 mixing ratios decrease and most notably between 10°N and 90°N starting at 9.2 km ( Figure 4b). The greatest decrease (≥2%) takes place at 11.8-16.2 km, between 25°N and 50°N. When averaged zonally, no increase in O 3 mixing ratios corresponding to the increases in NO x concentrations is seen between 7.2 and 9.2 km.

Global Geographical NO x and O 3 Distribution
The global geographical distribution and magnitude of the NO x changes under the TRS at 11.8-16.2 km and 7.2-9.2 km are shown in Figure 5. Globally, the reduction in NO x mixing ratios at 11.8-16.2 km is found to be 8% and regionally up to 50% reduction has been found over the North American east coast (Figure 5a). Figure 5b shows the positive changes in the geographical distribution of NO x at 7.2-9.2 km where the global mean change (2%) and the regional change (up to 40% in central mainland Europe, and the east coast of North America) have been observed.
The decreased aircraft NO x emissions at 11.8-16.2 km (NO x sensitive region) under the TRS leads to a decrease in global mean O 3 mixing ratios by 0.8% and regional O 3 mixing ratios by up to 2.5% (2 ppb) over the North American east coast (Figures 6a and 6c), but the aircraft emit an increased amount of NO x at 7.2-9.2 km (NO x sensitive region) under TRS where the additional NO x leads to a decrease in global mean O 3 mixing ratios by 0.3% and an increase in regional O 3 mixing ratios by up to 0.2% (2 ppb) North Atlantic Ocean, southern Europe, and south East Asia (Figures 6b and 6d). The increases in the concentration of O 3 are found between 7.2 and 9.2 km due to the lowering of the estimated global mean cruise altitude and the displacement of the total global aviation NO x emissions to a lower altitude materialize away from the continents and over the oceans. A belt of decreased O 3 levels at 0-30°N stretches from the east to the west ( Figure 6b); hence, despite increased NO x levels between 7.2 and 9.2 km, there are negative changes in O 3 levels (1%) in this region. In terms of the impact on the global climate, the turboprop replacement strategy resulted in an average overall   [Brasseur et al., 1998]. This is likely to be in a regime closer to the net O 3 production compensation point (the point at which the net O 3 production is equal to zero), which is preferable in terms of the atmospheric impact. The findings in the study are in qualitative agreement with the results of Grewe et al. [2002] and Søvde et al. [2014] where they found reduced levels of O 3 after lowering the cruise altitude by 1 km and 0.6 km, respectively. However the magnitudes of O 3 changes cannot be compared because of the different methodology used and the smaller altitude decrease considered in the Grewe et al. [2002] and Søvde et al.
[2014] studies. We have not assessed the potential of TRS strategy on the long-term effects of NO x emissions, effects of CO 2 emissions, changes in contrail coverage, or aircraft-produced aerosol emissions. However, considering the relationship of the radiative forcing with the cruise altitude [Søvde et al., 2014], it can be concluded that the TRS strategy reduces the global O 3 levels by up to 2 ppb (Figure 6c) which subsequently reduces the radiative forcing (RF) by 2-3 mW m À2 . The total RF from aviation (in 2005, excluding cirrus effects) is 55 mW m À2 [Lee et al., 2009], i.e., the mitigation strategy may reduce total aviation RF by~5%. The manipulation of the aviation NO x emissions between 2005 and 2011 due to TRS has the effect of decreasing the tropospheric burden of O 3 by 0.9 Tg. The net O 3 production decreases by 0.5 Tg (0.6%) between 2005 and 2011. TRS leads to a substantial decrease of NO x mixing ratios at the current estimated mean cruise altitude near the tropopause. The displacement of the total annual global aviation NO x emissions to the lower altitude between 7.2 and 9.2 km leads to an overall decrease of O 3 at that altitude. The findings from this study imply that the substitution of turbofans with turboprops on a global scale in regional fleets on short haul missions would lead to an overall reduction in levels of tropospheric O 3 in a region of the atmosphere where it is most harmful in terms of the radiative forcing of climate change.

Conclusion
A reduction in surface NO x , CO, and HC local to airports is a matter of much current debate, but it should be noted that studies have shown that there is a direct relationship between such compounds and negative health outcomes and that such exposure can have long-term impacts [Janke et al., 2009;Hansell et al., 2015], and so any reduction in primary emissions would be welcome. The global model used here will not be able to resolve changes in O 3 on city wide and regional scales and so O 3 may increase during outflow as the airport is likely to be in a NO x -saturated, VOC-limited environment, but total O 3 production will scale approximately with NO x away from urban areas and so any reduction in NO x is likely to reduce surface O 3 regionally with multiple health benefits for animals, plants, and building structures [e.g., Jenkin and Clemitshaw, 2000].