Skip to main content

Reducing Emissions from Aviation by Changing Trajectories for the benefit of Climate

Final Report Summary - REACT4C (Reducing Emissions from Aviation by Changing Trajectories for the benefit of Climate)

Executive Summary:
The European project REACT4C (FP7) investigated the potential of climate-optimised flight routing as a measure of reducing the atmospheric impact of aviation. The novelty in REACT4C is a modelling chain for an optimisation of aircraft trajectories with respect to their climate impact, which is depending on the actual weather situation, taking into account the weather-dependent climate impact of aviation emissions (CO2 and non-CO2, such as NOx, H2O and contrail cirrus) released during individual flights.
We have performed a weather classification for the North-Atlantic region, computed 4-dimensional climate cost functions for various species and using different metrics, combined them with the traditional operational cost functions airlines use to optimise their routing. Most efficient reduction of climate impacts is calculated for given changes in operating costs (so-called Pareto fronts). This novel methodology turns out to more effectively reduce the air traffic’s climate impact than simply flying higher or lower, as we showed using results from a variety of participating chemistry-transport models. The collaborative project started in January 2010 and has been successfully completed in April 2014.

Key findings of the project are:
• Climate-friendly routing can be performed using weather classes. The weather situations over the North Atlantic at cruise altitudes were categorised into a small set of distinct frequently-occurring patterns. Strength and orientation of the jet stream; the flight routing and durations, for both eastbound and westbound flights, depend on these patterns.
• These weather patterns were used to estimate the CO2 and non-CO2 climate impacts of aviation, and formed the basis of for calculating the routes that had minimal climate impact, i.e. climate-optimised trajectories that depend on the actual weather. Climate cost functions are the basis of this optimisation which describe the atmospheric sensitivity to aviation emissions with respect to climate, depending on the geographic position, the flight altitude and time.
• By this weather dependent flight trajectory optimisation it is possible to significantly reduce the overall climate impact of aviation considering CO2 and non-CO2 impacts at only moderate cost increases. It is even possible to convert costs into gains for airlines by combining our method with an emission trading system and turning the these climate optimised trajectories into non-regret measures.
• A novel analysis of airflows in North-Atlantic weather systems examined the typical duration of regions of ice supersaturated air, when following a parcel of air along these air flows. It was shown that individual parcels normally remain supersaturated for less than 6 hours. The weather conditions in which longer-lived areas of super-saturation can occur were identified, which is crucial for contrail formation by air traffic.
• A higher temporal and spatial resolution of meteorological fields is needed in order to sufficiently resolve contrail cirrus. It is still AIC (aviation induced cloudiness), followed by sulphur aerosols whose climate cost functions have the largest uncertainties. One needs to use the full set of climate cost functions (all effects) in order to minimise the total climate impact of aviation.
• Uncertainty studies have been performed and it was shown that while absolute values may change when uncertain parameters are varied, the overall results of the study were robust. Simplified mitigation procedures (flying higher and flying lower) are less cost effective than weather dependent routing.
• Our method is flexible such that new results and findings on species’ impacts on climate can easily be incorporated, resulting in additional components of the climate cost functions. This conceptual approach can be used for further mitigation assessment studies. A road map for implementation of our method was assessed for moving towards conceptual use in any flight planning system on a long-term time horizon.

Project Context and Objectives:
Impact of aviation emissions on the atmosphere
Air transport is a very important economic factor in the era of globalisation, both for passenger and freight transport. In fact, its importance is growing relative to other factors since many years, which is reflected in past and ongoing above-average growth rates of about 5% per year. However, the dark side of this success-story is that aviation contributes to climate change and its relative share (compared to other economic and traffic sectors) increases, despite substantial improvements in fuel efficiency. Aviation emits gases that directly contribute to the greenhouse effect (e.g. carbon dioxide CO2 and water vapour H2O) and gases (e.g. nitrogen oxides, NOx) that alter concentrations of other greenhouse gases (ozone O3 and methane CH4) by participation in air-chemical reactions or lead to formation of contrails by the emission of water vapour. Furthermore, aircraft engines emit gases (sulphuric oxides) that can be converted into aerosol droplets and solid particles (soot). These support the formation of contrails and have the potential to affect natural cloud formation processes as well. On average, contrails contribute to global warming. Overall, total aviation emissions have a warming effect on climate (see IPCC 1999, Sausen et al. 2005, Lee et al. 2009, 2010, and the series of ACCRI papers in the April 2010 issue of Bull. Amer. Meteorol. Soc. for more details).

Principles of mitigation concepts
There are two principle ways to a less climate-adverse aviation system, (1) to reduce the absolute amount of emissions (per year) and (2) to reduce the specific impact of a unit emission on the atmosphere. World’s aviation industry follows the first way already since decades. Aircraft and engines have always been improved, new and lighter materials have been introduced in the manufacturing process, aerodynamics has been developed (e.g. shark skin, winglets) and new developments are underway (e.g. laminar wing), and engine efficiencies have been increased, leading to an emission reduction. These measures allowed reducing costs by saving fuel and thus fuel consumption per passenger kilometre has been reduced by a factor of about 40% since the introduction of jet engines (e.g. Lee et al. 2009). However, the demand for aviation has increased as well during this period and the absolute amount of fuel burnt by aviation has increased over the years. Accordingly, aviation emissions have increased as well in total amounts. Although further technological developments are still possible and will probably lead to a further reduction of specific fuel consumption and specific emissions (Green 2005), it will not outweigh the strong increase of demand and thus technical solutions of fuel saving will not suffice to make aviation sustainable.

Thus, in order to lessen aviation impacts on climate we need additional efforts. The second way, the specific impact of emissions (see above), is possible whenever the emission location affects the emissions’ impact. This holds for all emissions except carbon dioxide CO2. CO2 has a very long residence time in the atmosphere of decades to centuries and it does not matter when and where a unit mass of CO2 is emitted because after a sufficiently long residence time it will be well mixed throughout the atmosphere. On the contrary, the climate impact of all other types of emissions (non-CO2) depends on the actual state of the atmosphere and thus on location and time of emission. This implies that there are atmospheric regions where an emission has a relatively large effect on climate and others where its effect is relatively small. If we knew under which circumstances an emission has only a small effect or if we could calculate it, we could exploit this knowledge for a climate-friendly aircraft routing.

Spatial and temporal variation of aviation impact in the atmosphere
Unfortunately the regions where aviation emissions have large and small impact are not fix; they change with the changing weather. However, certain characteristics of the atmosphere are either independent of the weather or they change only on very long (climatological) time-scales. If it would suffice to exploit these characteristics of the atmosphere for climate-friendly routing, the result would be a set of more or less simple rules and the computational effort to find these rules would be relatively low. This way has been explored in the past with little success. As an example of such a simple solution we mention the strategy to avoid formation of contrails by flying at a sufficiently low level in the troposphere where temperatures are too high for contrail formation.

Principles of operational mitigation strategies
In order to explore mitigation strategies various approaches have been investigated in the past: as a matter of principle flying higher or lower (Sausen et al. 1998; Grewe et al. 2002; Williams et al. 2002; Fichter et al. 2005; Gauss et al. 2006; Köhler et al. 2008; Rädel and Shine 2008), flying around regions where contrail can form vertically or laterally (Mannstein et al. 2005; Gierens et al. 2008; Campbell et al. 2009; Sridhar et al. 2011), climate optimized aircraft design (Koch et al. 2011; Schwartz Dallara et al. 2011), and weather-dependent route optimization for minimum contrail climate impact (Schumann et al. 2011) and for overall climate impact (Matthes 2011).

These studies indicate significant mitigation potential for reducing climate impact by avoiding regions of the atmosphere which are particularly sensitive to aviation impact. Obviously, taking into account the specific synoptic weather situation, hence optimizing weather-dependent, an even larger climate impact reduction potential is anticipated and available airspace can be most efficiently used (Matthes et al. 2012).

The idea of avoiding such regions which are more sensitive to aviation emissions with regards to climate impact can be nicely illustrated when considering contrail avoidance. Contrail persistence is only possible if they are formed in ice-saturated or supersaturated air, so-called ice supersaturated regions (ISSR). Such regions are preferentially connected to certain synoptic features (Gierens and Brinkop 2012, Irvine et al. 2012). ISSRs can be predicted in weather forecast models (Tompkins et al. 2007) and thus they can be taken into account for contrail-avoiding flight routing.

Contrails are a relatively large but not the only non-CO2 contribution to climate change. Hence, when aiming to minimize overall climate impact of aviation, it is required to consider simultaneously as well other non-CO2 emissions, as nitrogen oxides (NOx), particulate matter (soot) and water vapour (H2O), which has been explored by weather-dependent flight trajectory optimisation in the REACT4C project. As a matter of course, suggested measures need to be compliant with existing procedures in air traffic control (ATC) and air traffic management (ATM).

Project objectives
In order to reduce aviation's emissions and reduce its climate impact, the project REACT4C addressed those inefficiencies which exist in the aviation system with respect to fuel consumption and emissions by investigating the potential of alternative flight routing for lessening the atmospheric impact of aviation.

Hence the main objectives of REACT4C were:

• to explore the feasibility of adopting flight altitudes and flight routes that lead to reduced fuel consumption and emissions, and lessen the environmental impact;
• to estimate the overall global effect of such ATM measures in terms of climate change.

The objective of REACT4C was to demonstrate that environmentally friendly flight routing is feasible, but does not address the operational implementation of such advanced ATM procedures. The latter would require much more time than is available during the present project. However, REACT4C delivers substantial scientific foundation and operational specification for novel ATM procedures, which might be explored in a later phase of the SESAR JU. Analogously, REACT4C delivers fundamental concepts of aircraft that are better suited for environmentally flight routing, which will have the potential to enter the Clean Sky JTI in a later phase. In this way we were able to achieve objectives that were ambitious but realistic for a four-year medium-scale collaborative project.

Concept
The underlying concept is to upgrade conventional flight planning tools to incorporate climate-optimization flight planning tools. On the basis of a comprehensive Lagrangian climate-chemistry model we establish 4D climate cost functions for selected example days, which determine the incremental climate change for unit emissions as functions of location (latitude, longitude), altitude and time. The integral along a given trajectory of an aircraft's emissions, weighted by the cost functions, gives the incremental impact of the considered flight. The environmentally optimal trajectory is found in the flight planning tool by minimizing this integral. A quantification of the associated reduction of fuel consumption and emissions and improvement of climate impact is obtained as a side product of the optimization procedure. The calculated environmental benefit is evaluated using atmosphere models that are independent of the optimisation procedure; at the same time the uncertainty associated with the environmental benefit is estimated. Additionally, we explore potential aircraft design that is better suited for environmentally friendly flight routing than present day aircraft.

Project Results:
Methods developed within REACT4C
The REACT4C concept required the development of a novel methodology. Here we sketch the individual steps that have been taken to realise the REACT4C concept. The details are described in open access journal papers. The steps are a first development of a weather classification for the North-Atlantic, focusing on air traffic aspects, which is described in more detail in Section 3.1.1. Second, the calculation of climate-cost functions for each of these weather classes (Section 3.1.2) and third the air traffic flow optimisation on the basis of these climate-cost functions (Section 3.1.3).
We have developed our methods for the North Atlantic region since this is a region with a lot of air traffic each day between Europe and North America. The traffic density is large but not too large such that it is still possible to optimise flight routes without compromises to safety and flight distance so that route changes are feasible. Broadly the same methodology could be applied to other regions although constraints will be different (for example, flights within Europe or North America operate in a more crowded air space, and flights within or through the tropics operate in a quite different meteorological regime).

Classification of meteorological situation
The calculation of climate cost functions is a very time-consuming process that cannot be done anew for every day’s specific weather situation, for instance during a weather forecast numerical simulation. Meteorological experience, however, says that weather situations can be classified (the German language has the word “Wetterlage” to denote this experience; appropriate English translations are “archetypical meteorological situation”, “archetype”, or similar expressions). Weather situations belonging to the same archetype are similar in certain respects, and this similarity can be exploited to simplify the computation of climate cost functions.
We used ECMWF data to classify the weather situations and then look for analogues in the climate model, EMAC. Irvine et al. (2013) used "quasi-observational" data (i.e. the meteorological analyses from the forecast centres) and applied atmospheric circulation indices as the basis for meteorological classification of weather patterns, e.g. for the North Atlantic specifically NAO (North Atlantic Oscillation) and EA (Eastern Atlantic) oscillations. This allowed the characterisation of prevailing weather patterns and their probability. For each selected weather pattern a representative (typical) real case serves as period for detailed analysis. Weather types were classified and named according to the prevailing jet characteristic, e.g. strong zonal jet or strong tilted jet. These weather types have a strong influence on actually selected flight trajectories. Thus, they effectively classify not only the weather situation but the selection of flight routes as well so that climate cost functions and climate-friendly flight routes for different weather archetypes differ sufficiently to justify the computation effort of a climate-cost function for each weather type. It was then possible to identify days in the climate model simulations that corresponded closely to the observed weather patterns, and it was these climate-model analogues that were used for the detailed climate cost function calculations.

Concept of climate cost functions
Climate cost functions measure the effect of a certain unit emission on climate change. For each emitted species, such as CO2, NOx, and H2O, and flown distance for the effects from contrails, they comprise 4-dimensional scalar fields, where the four dimensions are time and location of the emission. The climate effect can be measured with various metrics (absolute global warming potential AGWP, absolute global temperature potential AGTP, etc.; for definitions and an overview see Fuglestvedt et al. 2010) so that the actual value of a climate cost function depends on the choice of metric (this question will not be discussed here; the reader may consult Grewe et al., 2014a for the REACT4C specific question and Fuglestvedt et al., 2010 for a general introduction to climate metrics). Usually it is easier to deal with non-negative cost functions since otherwise inappropriate incentives may occur for flying (and indeed "lingering") in areas where aircraft emissions cause a negative forcing and therefore tend to cause a climate cooling. In fact, both NOx and contrail climate cost functions are negative in some regions. To prevent inappropriate aircraft trajectories, a brute force approach was chosen within REACT4C, where a large number of possible trajectories were defined and the climate impact of each is calculated based on the climate cost functions. For an implementation of the REACT4C rerouting strategy in the air traffic system, target functions of different nature, in our case climate cost functions and the traditional ones like fuel costs, crew costs, punctuality, etc, have to be converted to a common scale (e.g. money). Choices of conversion factors are political or economic choices and are outside the scope of the project. Here we focus on understanding the tradeoffs between individual cost factors and perform a multi-objective optimisation to understand the dependencies between climate impacts and cash operating costs. This procedure allows for an improved understanding of importance of individual emissions (CO2, NOx, contrails) for a climate optimised re-routing.
As said above, we select from the weather archetypes representative weather situations. These are not in any way idealised, as each weather situation has been selected on the basis of it being frequently occurring. For these representative situations we calculate the climate cost functions in the comprehensive global climate-chemistry-model EMAC. For the computation of the cost functions we follow the emitted species along air parcel trajectories through the atmosphere using a Lagrangian approach. We calculate their microphysical (e.g. cloud effects) and chemical effects (e.g. ozone production) and eventually the corresponding influence on the flow of radiation energy, that is, the global-mean radiative forcing. The integral of radiative forcing along the trajectory is the basic quantity from which the value of the cost function at the time and location of the emission can be determined, depending on the choice of a metric. Separate cost functions are calculated for nitrogen oxide emission, water vapour, soot, and carbon dioxide. Details on the calculation of climate cost functions are given in Grewe et al. (2014).

Environmentally friendly flight planning
As the next methodological step, these climate cost functions are combined with traditional cost functions and the combination is used in the flight planning tool from air traffic management (ATM). Within our setup we combine climate cost functions with waypoint based tools, which have been enabled to calculate climate impact contributions for individual flight segments. For a given set of city pairs and for a specific weather pattern the air traffic flow tool (SAAM) is used to analyse a large set of available options of waypoint combinations and intersections. Both flight level variations and horizontal variation of the aircraft trajectory are taken into account. We calculate fuel consumption, operating costs and climate impact of each individual option simultaneously. It is then possible to identify both the minimal cost trajectory and the minimal climate impact trajectory by means of the same flight planning tool.
The simulated air traffic consists of roughly 800 flights between Europe and the US. Any single flight path from A to B is varied horizontally and vertically such that we can choose between 105 possible aircraft trajectories, from which we determine the one with lowest operational costs, the one with the lowest climate impact, and any other combination. We do this for every flight, thereby avoiding too close encounters, such that for 800 flights we end up with roughly 105800 possibilities to organise the trans-Atlantic air traffic flow. Every combination has a certain total operational cost and a certain overall climate impact. From all combinations, we identify those, for which we cannot improve costs and climate simultaneously. This line describes the trade-off between costs and climate impact and is called a Pareto front. That is, the Pareto front represents the best effect on climate that we can get for a certain increase in costs. The two endpoints of the Pareto front are composed of (1) all flights with least individual operational costs (representing todays routing) and (2) all flights with least individual climate effect. The latter endpoint can be used to determine the maximum gain that is possible in such a system (see below). The Pareto fronts are the main tool to visualise the potential benefit of climate-friendly flight routing. Their shape depends on the weather archetype, flight direction (eastbound vs. westbound), and in particular on the chosen metric and its associated time horizon.
As final step, optimised individual trajectories need to be combined with an emission inventory, which is to be evaluated by means of chemistry-transport models with respect to its overall climate impact. This will determine the overall mitigation gain of the optimised trajectories and it eventually establishes the robustness of our modelling chain for climate-optimised flight planning.

A one-day case study of climate-optimized air traffic
Here we present a case study, which demonstrates for one specific winter weather situation, the principle of the REACT4C re-routing strategy. Details of the modelling approach can be found in Grewe et al. (2014a), and details on the application, the one day case study, are published in Grewe et al. (2014b).
We focus on one specific weather situation for winter, and analyse the air traffic routing options for this weather situation for one day. The weather pattern at the main cruise levels (200 hPa-300 hPa) is shown in Fig. 3.1. This weather situation has a trough over the north Atlantic, with a high pressure ridge over Europe (left). It is characterised by a strong zonally-oriented jet stream with wind speeds exceeding 65 m/s in the core of the jet stream around 35°N (right). This pattern is a representative of winter weather type 1 defined by Irvine et al. (2013), and would typically occur on 17 days each winter.

(Figure 3.1 would appear here, see appendix)

For this day, real air traffic data are used. This traffic flow includes 391 and 394 flights with 28 and 30 different types of aircraft for eastbound and westbound routes. The transport volume and fleet mixture are similar in either direction, though not identical. The flights from Europe to the U.S. start in the (local) morning and the flights from the U.S. start in the (local) afternoon and evening arriving in Europe on the next day. Therefore, this one day air traffic actually spans a little bit more than 24 h.

Climate-cost functions
Fig. 3.2 shows the climate cost functions for contrails, ozone, methane and total NOx effect (=sum of ozone, methane, and primary mode ozone) at 200 hPa for 12 UTC. Meteorological fields from Fig. 3.1 are overlaid, e.g. the location of the low pressure system (thick black line) and the terminator (violet line). Wind speeds are indicated by the blue dashed contours, which makes it possible to locate the position of the jet stream. Contrails (Fig. 3.2a) show very different evolution and impacts. In the north-eastern part, e.g. over Greenland and East of Greenland, contrails form and are transported northward with a lifetime of around 2 h. They are mainly occurring in darkness and lead to a warming. By contrast in the region of the Gulf of Saint Lawrence the contrails show a larger optical thickness and remain at very low solar zenith angles, which lead to a cooling effect, since the negative solar forcing dominates over the positive longwave forcing.

(Figure 3.2 would appear here, see appendix)

The climate impact of a NOx emission (Fig. 3.2d) results from an increase in ozone and a warming effect (Fig. 3.2b) and a decrease in methane and an associated reduced warming effect (Fig. 3.2c). The ozone impact (Fig. 3.2b) is larger in the area of the jet stream and shows a minimum in the location of the low pressure system. Air masses, which are transported towards higher latitudes, e.g. originating from around 30°W and 60°N, are also transported to higher altitudes. This implies a longer atmospheric lifetime of the emitted species NOx and H2O, but also a smaller production of ozone, since in that region (lower stratosphere) and time (winter) the chemical reaction rates are low. At lower latitudes (e.g. at 75°W and 60°N), the emitted species are transported to the tropics and experience a large ozone production, though for a short time period, since the nitrogen compounds have a lower residence time, caused by washout. Still the higher ozone production dominates and leads to strong warming effects. The range in the ozone induced warming is one order of magnitude in the displayed area (Fig. 3.2b).
The total NOx impact on global-mean temperature is positive for emissions at lower latitudes and negative for emissions at at higher latitudes (Fig. 3.2d). This is in agreement with earlier studies, which showed stronger ozone induced warming at lower latitudes and roughly a balance between ozone warming and a reduction in methane reduced warming effects at higher latitudes (Grewe and Stenke, 2008, Köhler et al. 2013). The results show a broad variability in the temporal evolution of the atmospheric masses of H2O, NOx, ozone and methane, caused by the pulse emission at the climate cost function grid box, which are in agreement with earlier studies (for more details see Grewe et al., 2014a).

Air traffic optimisation in the North Atlantic Flight Corridor
As an illustration, Fig. 3.3 shows various flight options for one city pair connection (Washington DC to Vienna). The flight on that day (light brown) clearly follows the jet stream (arrows). We have performed an optimisation of the whole transatlantic air traffic with respect to its short-term climate impacts as measured with AGWP20 and with respect to costs. We obtain a slightly different cost optimised aircraft trajectory for the above-mentioned city pair (blue), which more closely follows the jet stream. However, when the traffic is cost-optimised and includes conflict avoidance, as in reality, then the real route (light brown) and the cost-optimal (cheapest) route within conflict-free traffic (dark brown) are much closer. The difference in distance and time is around 1% and in fuel about 3 %.

(Figure 3.3 would appear here, see appendix)

For this city pair, the climate optimal routes (with and without conflict avoidance) with respect to short-term climate impact (see above) are located further north and at lower flight altitudes (FL330 and FL310). The conflict-free, climate-optimised route (green) avoids more contrails and leads to a decrease in contrail AGWP20 by 16% and a decrease in NOx AGWP20 by 4% with an increase in fuel consumption by 14 %, which is related to an increase in the CO2 induced AGWP20 by 1 %.

We have optimised the trans-Atlantic one-day air traffic with respect to different objectives: economic costs, short-term climate impacts (F-ATR20 and P-AGWP20) and long-term climate impacts (P-AGWP100). The results, shown in Fig. 3.4 are separated for westbound (blue shaded) and eastbound (red shaded) flights, since the impact of meteorology on routing, largely differs depending on the flight direction, as tail and head winds play a large role (e.g. Irvine et al., 2013). We define a reference point which is that air traffic routing with the minimum cost and discuss the economic costs and climate changes from traffic changes in relation to this reference point. The climate optimised air traffic leads to a maximum reduction of the climate impact in the range of around 25%-60% associated with an increase in economic costs around 15%.

(Figure 3.4 would appear here, see appendix)

We find a clear difference between eastbound and westbound flights. Eastbound routes benefit from the tail winds of the jet stream. Each re-routing option for eastbound routes which leaves the jet stream has a significant increase in fuel consumption. In contrast, westbound flights and any re-routing option both avoid head winds. Hence the difference in fuel consumption between each westbound re-routing option and the reference (minimum cost) flight is less compared to the eastbound flights. In addition, any increase in fuel consumption also implies an increase in NOx emissions. Therefore, for this specific meteorological situation, the westbound air traffic has more re-routing possibilities avoiding
non-CO2 warming effects with only small increases in fuel consumption and compensating CO2 induced warming, compared to eastbound air traffic. For each metric and flight direction, the Pareto front (optimal relation between climate change and costs) is included in Fig. 3.4. Starting from the economic best flights (lower right), we change successively re-routing flight options, starting with the most promising, i.e. largest climate reductions at lowest cost increase. Already large reductions in the climate impact of up to 25% can be achieved by only small increases in economic costs of less than 0.5% (red solid line).

Mitigation gain and corresponding costs due to trajectory optimisation
The maximum mitigation gain (i.e. the maximum reduction of climate effects) in a certain situation is represented by the upper left endpoint of the Pareto fronts shown in the previous sections (Fig. 3.4). The question is now, how does the picture look like, if we consider longer periods of time instead of single days or single situations, for instance a complete season. In order to achieve this we need to know how often certain weather archetypes occur during a long period. We have done this for summer and winter seasons using the 5 winter and 3 summer weather types determined in the project. Using ERA-interim reanalysis data from 1989 to 2010 we find that the relative frequency of these weather types (pattern probabilities) changes considerably from year to year. For the present estimate we use the mean fractions over all considered years. The maximum gains for all weather patterns are then averaged using the pattern probabilities as weights, resulting in the average maximum gain over the respective season. The results for various chosen metrics are chosen in Fig. 3.5.

(Figure 3.5 would appear here, see appendix)

In Fig. 3.5 the average maximum gain over a winter or summer season is given by the respective centres of the crossing error bars (large symbols). We see that the actual value depends strongly on the chosen metric (symbols) with the highest gains found for AGWP100 and the lowest for ATR20. Also the flight direction has a strong influence. The potential gains are larger for westbound flights (red) than for eastbound flights because leaving the jet stream on eastbound flights for a reduction of climate impact incurs relatively large increases of fuel demand, thus diminishing the climate gain, and increase of operational costs. The inspection of results for individual weather situations shows that the contrast between the climate impact of westbound and eastbound air traffic is large for zonally jets (see example in Sec. 3.2) but almost disappears in situations where the jet meanders highly.
The figure shows a quite large potential to minimise climate impacts of up to 40% on average (for the AGWP100 metric). The error bars shown result from the year-to-year variation of the weather pattern probabilities which affects both the climate gain and the resulting costs.
In this sense, the error bars do not represent an error at all, only a natural variability of the atmosphere. Actual uncertainties which arise from uncertainties in the cost-function computation are not included in the figure. Such uncertainties originate from, inter alia, using a representative day for a weather pattern neglecting the day-to-day variation, from the coarse grid of emission locations and time points, from “noise” in the climate impact of emissions (i.e. emissions released from nearby locations follow different trajectories and have a different climate impact), and from uncertainties in the representation of atmospheric processes in the CTM used for the calculation of the cost functions.
Such uncertainties were addressed partially in Grewe et al. (2014b), e.g. the impact of the resolution and the impact of a chaotic dispersion on the results, and those were found not to critically affect the pattern of the climate-cost functions, e.g. for the impact of NOx emissions. Since uncertainties play a critical role, we tested the effect of uncertainties in the climate cost functions on the obtained Pareto-front in sensitivity studies by parametric changes in the climate-cost function. The absolute numbers in the climate impact reductions changed, but the main results, such as large climate reductions are feasible at low cost changes were unaffected.
The large increases in economic costs, compared to general return of investments of airline companies that would be incurred by choosing the maximum mitigation option make this choice currently unlikely to be acceptable to the aviation industry or legislators. More likely is that a mitigation option would be chosen that corresponds to an acceptable increase in economic cost. For this approach, for each weather pattern, we choose the mitigation option corresponding to a certain increase in cash operating costs, and calculate the resulting seasonal-mean reduction in climate impact. We can also calculate the range in climate impact reduction resulting from the natural variation in the frequency of weather patterns each winter or summer. We have calculated this for a 0.25%, 0.5%, 1.0%, 2.0% and 4.0% increase in COC (Figures 5.2). For AGWP20, a 0.5% increase in cash operating costs can achieve a reduction in climate impact of roughly 8% (eastbound) and 12% (westbound); this demonstrates that a substantial fraction of the maximum possible reduction in climate impact may be achieved through relatively small increases in cash operating costs. It is important to note that the reduction in climate impact achieved for a given increase in economic cost is sensitive to the choice of metric, particularly for higher values of economic cost increase.

(Figure 3.6 would appear here, see appendix)

The implementation of climate-friendly flight routing could be fostered by economic instruments, such as the emission trading system. If there were an emission trading system that includes not only CO2 but the other non-CO2 impacts as well, the calculation of their CO2 equivalents for these impacts would follow the conventional procedure, e.g. as currently imposed for emissions of methane or nitrous oxides. The conversion factors to establish this equivalence can be directly computed from the cost functions (i.e. they depend on the metric). We have assumed a cost of 5€ per tonne of equivalent CO2 emitted (representative of current day). This will give incentives for flying climate-friendly, especially when metrics with short time scales are addressed, since then the avoidance of other climate impacts can easily outweigh the increase in CO2 emission resulting from deviations from the most fuel saving flight path. Figure 3.7 shows that quite substantial reductions in the climate impact can be achieved if an airline chooses to fly such that the cost optima (minima of the curves) are met. In our example (Fig. 3.7) the cost reduction by avoiding buying emission certificates outweighs the re-routing induced increase in cash operating cost. For metrics with short time horizons, i.e. with a focus on avoiding climate impacts in the near future, airlines could even benefit from the trading in the order of 1% of the cash operating costs. For metrics with long time horizons, i.e. with a focus on avoiding climate impacts in the far future, such a benefit can be achieved for moderate re-routing options. Both situations are hence considered as non-regret measures for the airline companies.

(Figure 3.7 would appear here, see appendix)

Mitigation gain from simply flying lower or higher
Described here are some of the results from WP 5. This is important for a comparison with the results detailed in section 3.3.1. The question to be addressed is: Is it necessary to spend the large effort computing the cost functions or can a similar gain be achieved by such simple measures as flying higher or lower?
To explore the uncertainty of the procedure explained above, a multimodel-estimate with an updated aviation emission inventory based on the CAEP/8 movement data for the year 2006 was prepared within REACT4C. Two corresponding mitigation scenarios were compiled: one shifting aircraft cruise altitudes upwards, and a second one shifting them downwards by 2000 ft. The updated aviation climate impact estimates are calculated by means of a multi-model assessment.
Using all RF estimates assessed during REACT4C, along with metric calculations based on current literature values, we have calculated the impacts of different components: NOx (O3, CH4, H2O strat.), H2O direct (from aircraft engines), sulphate particles, soot particles, soot-cirrus, and contrail cirrus. When all aircraft as a matter of principle fly 2000 feet higher if their performance allows, the climate impact of individual aviation emissions increases except for sulphate and CO2. When flying 2000 feet lower, the opposite is the case, and all individual effects decrease except sulphate and CO2.

Considering a 1-year aircraft pulse emission, contrail cirrus is the largest contributor to temperature change within the next 20 years, before CO2 takes over as the most important contributor. Rather assuming sustained emissions (and no emissions prior to year 1), contrail cirrus keep its role as the largest contributor to temperature change for 100 years, before CO2 becomes more important. Changing cruise altitude changes the radiative forcings and our results suggest that in a sustained emission scenario, contrail cirrus have a larger impact on temperature than does the change in CO2 due to fuel changes. Also changes originating from NOx are larger than the CO2 changes.

However, it turns out that the climate gain that can be achieved by simply flying lower is significantly smaller than what can be achieved with our smarter novel method, while increases in fuel consumption and flight times are similar. We conclude here, that a computation of the cost functions is a worthwhile effort because much higher rewards can be achieved in terms of climate impact reductions when the weather situation is taken into account compared with simply flying lower.

Recommendations
A pre-requisite for climate-optimization is the availability of high quality weather forecast data, for air traffic management. Our experience from many model runs performed to compute climate cost functions can be condensed into simple rules for routing in the North Atlantic as follows: For east-bound traffic, when flying inside the jet stream, the climate cost functions are generally higher, which increases climate impact of emissions. Hence avoiding the jet stream would allow reducing climate impact of non-CO2 emissions.
When comparing east-bound and west-bound traffic, we identify a higher mitigation potential for west-bound traffic in a strong jet (zonal) situation, as east bound traffic receives a high penalty when leaving major jet maximum, while more air space is available for optimisation when aiming to avoid the jet stream.
Simultaneous consideration of individual effects is a pre-requisite for identifying climate optimal solutions. It is beneficial for both climate and airlines if an ETS is introduced for equivalent CO2 and based on a metric with short time horizon. Aircraft design adapted to climate-friendly routing can lead to even further reductions in climate-impact.

Potential Impact:
REACT4C completed successfully a feasibility study on mitigation strategies relying on weather dependent climate optimisation of aircraft trajectories and provided quantitative estimates of the associated mitigation potential. Implementation of the climate-friendly flight routing concepts developed within REACT4C has strategic impact on various stakeholders. The European research agenda is driven by ACARE council.

The ACARE goals comprise a list of emission reductions that individually would lessen aviation's impacts on the atmosphere. The compliance of the REACT4C project with the ACARE goals is addressed in the individual paragraphs below
• Reduction of fuel consumption and hence CO2 emissions
• Reduction of NOx emissions
• European added value (sustainable growth, future economic development, environment)

Strategic impact on research agenda on aviation and environment
Implementation of the climate-friendly flight routing concepts developed within REACT4C would help to achieve the ACARE goals indirectly, which means that the climate impact of emissions will be lessened without necessarily overall lower emissions as foreseen in the ACARE goals. This is possible by the explicit consideration of non-CO2 effects whose climate impact depend on time and location of emission.
Project partner AIRBUS considered new aircraft designs that are better suited to novel concepts of flight routing. New design might even lead to fuel savings and thus part of the ACARE goals could be directly addressed. Furthermore, such concepts can be taken up within the Clean Sky JTI. We could show within the project that or concepts are feasible and compliant to current ATM and ATC rules and regulations. This has been assured by project partner EUROCONTROL. Experiences from the project are thus expected to be taken up by the SESAR JU.Thus our project has impacts on both large European efforts in aeronautics, Clean Sky and SESAR, leading to an added value to either project.

European added value (sustainable growth, future economic development, environment)
In light of the projected rapid growth of the transport volume of aviation, the achievement of the ACARE goals will not be sufficient to stop the growth of the contribution of aviation to cli-mate change, if absolute figures are considered. The situation looks even worse if a relative contribution of aviation to the total anthropogenic impact is considered as, at least in the European Union, the equivalent CO2 emissions of the non-transport sectors decrease.
Environmentally friendly flight planning, as developed in the REACT4C project, offers a chance to further reduce the impact of aviation on the environment/climate for a given transport volume.
An important result was that large reductions of climate impact are principally possible, but the maximum that could in principle be achieved will incur unacceptable rise of operational costs. However, still quite substantial reductions of climate impacts are achievable at moderate increases of operational costs of the order 1%; details depend on the weather situation and the chosen metrics. The situation looks even more promising if an emission trading system including non-CO2 emissions would apply, as this would give incentives to airlines to fly climate-friendly. This possibility to gain money can increase airlines’ willingness to accept an ETS with reasonable prices.
Within the process of extending the EU Emission Trading System (ETS) to international aviation, the inclusion of the non-CO2 effects is controversially discussed. However, this is the key to convert costs into gains for the airlines. The method we used was based on the ratios of cost function components, which provides to our view a fair pricing for non-CO2 effects. This method could be used as well in a formulation of the ETS; it offers an elegant and fair method to include non-CO2 effects in the EU ETS.

The REACT4C project developed a method to quantify all non-CO2 effects on climate and at the same time a concept by which these effects are reduced through improved ATM and aircraft design. Hence, environmentally friendly flight planning supports a sustainable growth of aviation (with the related impacts on economy and environment).
While technical improvements in airframe and engines only slowly become effective because of the long development times, fleet rollover and long in-service times of aircraft, reducing aviation's environmental impact by means of ATM can be realised more quickly. It even might already substantially contribute the EU target of reducing the EU contribution to climate change by 20% in 2020 relative to 1990.

Dissemination of foreground - advantages for Europe
Results of REACT4C have been published in peer reviewed journals, reports, popular science magazines and on the internet (e.g. the project website http://www.react4c.eu) and have been presented at international conferences. The EC and co-operating partners received information at the earliest possible stage. In this way, which is via the European Commission, the Network European Environmental Advisory Committee (EEAC) could directly and indirectly be informed on project results. In particular, initial results from REACT4C have been available early enough for any upcoming IPCC Initiative to update the Special Report on Aviation. Furthermore, policymakers, stakeholders in the aviation sector and the general public have been informed via the internet, brochures and targeted workshops.
Within REACT4C a large number of multiplicators are involved, who are ICAO/CAEP members, participants of the Clean Sky JTI and the SESAR JU, IPCC climate experts, coordinators of other European projects involved in the Consortium. This assures efficient dissemination of project results.
Stakeholders in REACT4C range from airlines, aircraft manufacturers, engine manufacturers, service agencies, but also include regulators, policy makers and the general public. Adequate means of dissemination to these groups are presented below.

Dissemination to aeronautics stakeholders
As REACT4C has provided strategic important information on efficient flying with respect to fuel consumption, emissions and the climate impact of aviation, stakeholders have been involved in the project. For this reason an REACT4C Expert Panel has been established under the chairmanship of Dr. Herbert Pümpel, former Chief of the Aeronautical Meteorology Unit of WMO. This Expert Panel has been involved in this collaborative project from the beginning. Their contextual information on the challenges for ATM and other potential obstacles has been taken into account by the project. The meetings of the Expert Panel have been flanked by technical seminars specifically designed to the needs of stakeholders, and technical papers (e.g. on environmental cost functions), thematic papers, and position papers (e.g. on recommendations to include non-CO2 effects in ETS). Finally, in the fourth year focus has been given on identifying possible obstacles for implementing our newly developed procedure, and recommendations how to over-come them, have been provided. A public outreach event as a REACT4C stakeholder seminar was performed in the fourth year.

Community societal objectives
An objective of the European Union is to provide future mobility to the European and world population. Such mobility requires sustainable paths for future development. The ACARE targets were set out in order to contribute towards this sustainable development path. This requires eliminating inefficiencies within the air transport system. In air traffic management still such environmental inefficiencies do exist and provide the potential for improving environmental performance of aviation.
Another objective of the European Union is to reduce the risks arising from climate change to the European and world population. This requires the development of strategies for reducing climate-impacting emissions and for adapting to a changing climate. In both cases, quantitative knowledge of the various factors contributing to climate change is required. Future commitment periods to the Kyoto Protocol will potentially require further reductions of emissions and the inclusion of additional species. While the Global Warming Potential (GWP) provides a measure for comparing the six gases or groups of gases from the Kyoto basket, it fails when comparing (or trading) CO2 with precursors of shorter-lived species like ozone or with aerosols and contrails. For instance, Lee and Sausen (2000) pointed out the difficulties of including international aviation into the emission trading concept of the Kyoto Protocol.
REACT4C has explored possibilities to optimise flight trajectories under environmental aspects. The tool developed allows planning trajectories with less climate impact and its availability helps European policymakers in formulating policies for reducing emissions and evaluating mitigation measures. The improved information on the climate impact of emissions that has been achieved in the project allows industry to incorporate, with greater confidence, environmental considerations into both their design and development work and flight planning.

Reinforcing European Competitiveness
The European Commission has the possibility to exploit the new information provided by the project in negotiations on international treaties. As for the topic "Impact of Aviation on Climate", where Europe is in a leading position, REACT4C results and the new possibilities increase the competence, visibility and reputation of European research. The early access to new information has strengthened the position of the European Commission and that of present European representatives in ICAO. REACT4C. The European influence in IPCC fora is expected to increase. Additionally, the project results help improving the competitiveness of European industry through the provision of the most up to date information (link to Clean Sky).

General dissemination activities
Although the project disseminated results and achievements over the full duration, we concentrate here on activities since January 2013. Results and achievements of REACT4C were presented at several international scientific conferences, in Europe and USA, e.g. as a highlight during the ARAM conference, 7-10 Jan 2013 (Austin, USA), ATM Seminar, June 2013 (Chicago, USA), DACH Conference, Sep 2013 (Innsbruck, Austria), EMS in October 2013 (Reading, UK) and during the ECATS Conference, 18-21 Nov 2013 (Berlin, Germany).

For the consolidation of results, discussion on achievements, implementation and perspectives the final meeting took place from 11-12 Nov 2013 in Manchester. Usage of novel data developed via our REACT4C data protocol, making it available also for future research within the scientific community, was seen as important element for future implementation. A critical review of dissemination of results was undertaken, and consequently focus on publication of results was decided.

A dedicated REACT4C stakeholder event took place during the 1st ECATS Conference on “Technical challenges for aviation in a changing environment” in Berlin, Germany. The project office organized the meeting and acted as local host during the event. The event was attended by more than 50 researchers. Major results and achievements were presented by the Consortium by a series of presentations followed by a panel discussion. Constructive exchange on ideas for evaluation and further exploitation took place. Main message from the event was that alternative flight trajectories offer a substantial benefit for climate impact. Stakeholders were in particular interested in how to develop a roadmap for implementation, as well as remaining uncertainties.

Regular exchange with on-going major aviation programs continued. In September 2013 specific ideas on research directions regarding trajectory optimisation were discussed with partners of Clean Sky. Exchange with SESAR Environmental officer Celia Rodriguez continued leading to a key note speech during the REACT4C stakeholder event (Nov 2013). Concept, results, data products and achievements were presented during various European Scientific meetings, e.g. TEAM Play Meeting (Brussels, February 2013), Turbine Symposium (Delft, May 2013), NASA-DLR Cooperation, WeCare Annual Meeting (January 2014), and Forum-AE Climate Impact workshop (April 2014).

Communication on thematic expertise with stakeholders was performed by relevant publications, conference presentations, workshop and technical meeting presentations, and bilateral contacts, e.g. to Eurocontrol tactical research unit (Maastricht), engine manufacturers. Main focus is given on improved understanding of aviation climate impacts and mechanisms and perspectives for implementation in ATM. Lessons learned from early project results are already discussed with experts and stakeholders of all fields of expertise.

Training and education
REACT4C project partners contributed to the education of PhD students and training of young researchers by letting them work in the project’s tasks, discussing at scientific conferences and providing guidance on research. These students and researchers carry on their new abilities and the knowledge they gained to new stations in their career, in science, industry and administration.

List of Websites:
http://www.react4c.eu
Coordinator REACT4C: Dr. Sigrun Matthes (sigrun.matthes@dlr.de)
DLR Institute of Atmospheric Physics, Oberpfaffenhofeh, 82 234 Wessling, Germany