key: cord-1004784-9ukew54e
authors: Xue, Dabin; Liu, Zhizhao; Wang, Bing; Yang, Jian
title: Impacts of COVID-19 on aircraft usage and fuel consumption: A case study on four Chinese international airports
date: 2021-07-01
journal: J Air Transp Manag
DOI: 10.1016/j.jairtraman.2021.102106
sha: 46004b3d65ef428172dd7d6d955f7c3675cf69f0
doc_id: 1004784
cord_uid: 9ukew54e

COVID-19 pandemic starting in early 2020 has greatly impacted human and industrial activities. Air transport in China shrank abruptly in February 2020, following a year-long gradual recovery. The airline companies reacted to this unprecedented event by dramatically reducing the flight volume and rearranging the aircraft types. As the first major economy that successfully controls the spread of COVID-19, China can provide a unique opportunity to quantify the medium-long impacts on the air transport industry. To quantify the corresponding changes and to elucidate the effects of COVID-19 in the wake of two major outbreaks centered in Wuhan and Beijing, we analyze twelve flight routes formed by four selected airports, using the Automatic Dependent Surveillance-Broadcast (ADS-B) data in 2019 and 2020. Our results show that the total flight volume in 2020 reduced to 67.8% of 2019 in China. The recovering time of flight volume was about 2–6 months, dependent on the severity. In order to unwind the severe challenge, airlines mainly relied on aircraft B738 and A321 between February and June in 2020 because the fuel consumption per seat of these two aircraft types is the lowest. Besides, fuel consumption and aircraft emissions are calculated according to the Base of Aircraft Data (BADA) and the International Civil Aviation Organization's Engine Emissions Databank (ICAO's EEDB). At the end of 2020, the ratios of daily fuel consumption and aircraft emissions of 2020 to 2019 rebounded to about 0.875, suggesting the domestic commercial flights were nearly fully recovered. Our results may provide practical guidance and meaningful expectation for commercial aircraft management for other countries.

The outbreak of the Coronavirus pandemic 2019 has an unprecedented effect on public life worldwide , as the virus can infect both humans and a wide range of animals (Velavan and Meyer, 2020) . A global response is imperative to reduce the spread of the COVID-19 pandemic and slow down the number of new cases. Many countries raised their infectious disease response activities to the highest alert levels (Chang et al., 2020) and adopted a combination of containment and mitigation activities, including school shutdown, travel restrictions, industries, and construction activities suspension . China, the United States, and several other countries have instituted temporary travel restrictions to slow down COVID-19 spread (Fauci et al., 2020) , while most countries closed borders too late to contain the spread of COVID-19 (Sun et al., 2021a) .

As presented by Zhang et al. (2020b) , aviation is one of the most significant initial contributors to the COVID-19 spread. Many countries then reduced air transport associated with pandemic centers. As a result, there was a sharp decrease in the number of domestic and international passenger flights. International markets suffered a more severe depression than domestic markets (Sun et al., 2020) , except air logistics airlines suffering less depression (Li, 2020) , and ultra-long-haul flights outperforming other business models (Bauer et al., 2020) . From the perspective connectivity with other airports, the Southern hemisphere undergoes a more severe drop than the Northern part (Sun et al., 2021b) . According to the International Civil Aviation Organization, by the end of March 2020, more than 20 commercial airlines had stopped flights entirely, and about 12 airlines stopped all international flights all over the globe. These aviation losses caused a damaging decline in the World's GDP by 0.02%-0.12% in the first quarter of 2020 (Iacus et al., 2020) . Immediate policy designs are necessary to alleviate the impact of COVID-19 on the airline industry around the globe (Maneenop and Kotcharin, 2020) . For example, governments have prioritized maintaining air transport connectivity (Abate et al., 2020) , and airlines have adopted the typical crisis response strategies of retrenchment, persevering, innovating, and exit (Albers and Rundshagen, 2020) . Due to the preliminary control of COVID-19 in some countries, international air travel bans are relaxed to recover the economy (Zhang et al., 2020a) , which may lead to a likelihood of the aviation business rebounding at a slower pace with V-shape and U-shape recovery (Serrano and Kazda, 2020) .

Significantly disrupted aviation caused a subsequent reduction in fuel consumption and aircraft emissions. Fuel consumption influences airlines' profit and aircraft emissions can affect the environment and climate (Xue et al., 2020) . For example, the contribution of CO 2 is estimated to represent 36%-51% of the total aviation radiative forcing of climate, including short-term climate forcers (Kärcher, 2018; Terrenoire, 2019) . In addition to CO 2 , NO x emission from fuel combustion is estimated to account for about 65% of the global total NO x emission (Bauwens et al., 2020) . NO x also induces some extra consequences for climate and human health (Atkinson et al., 2018) . Thus, accurate fuel burn estimation models are necessary to evaluate the amount of fuel consumption and CO 2 emissions (Seymour et al., 2020) . The Aircraft Performance Model Implementation software is developed to simulate the global commercial flight fuel burn and emissions during 2006-2011 (Wasiuk et al., 2015) . Fuel consumption and CO 2 emission of a specific route are estimated by cluster analysis on historical trajectory data (Pagoni and Psaraki-Kalouptsidi, 2017) . To better calculate aircraft fuel burn during ground operations, Kim and Baik (2020) used aircraft trajectory data acquired from an airport surface surveillance system. Different from the methods mentioned above, our study uses the ADS-B (Automatic Dependent Surveillance-Broadcast) data to calculate the fuel consumption and aircraft emissions based on the International Civil Aviation Organization's Engine Emissions Databank (ICAO's EEDB) and Base of Aircraft Data (BADA), because ADS-B data are easily accessible.

Although the immediate impact of COVID-19 on aviation has been actively studied (Naboush and Alnimer, 2020; Serrano and Kazda, 2020) , the medium-term and long-term impacts are still unclear, as air travel in most of the major countries is still struggling to rebound to the level before the pandemic. However, China's successful control of COVID-19 provides us with a rare opportunity to quantify the impacts, as its domestic air demand has largely recovered by the end of 2020. In this study, we analyzed the two-year ADS-B data and calculated flight volume, aircraft usage, fuel consumption, and aircraft emissions among routes between four selected major Chinese international airports, Beijing (ZBAA), Shanghai (ZSSS), Guangzhou (ZGGG), and Wuhan (ZHHH). Since two of the cities were epicenters of the major outbreaks in the early and middle of 2020, our study allows us to elucidate both the abrupt decline and the year-long recovery of domestic aviation. Therefore, our study may shed some light on how the air travel activity was impacted both in different countries and at different pandemic stages.

The paper is organized as follows. Section 2 describes the dataset and the methodology. Section 3 shows the results. Section 4 discusses the impacts and concludes the paper.

Current strategic plans for Air Traffic Management envisage a transition from radar control to Communications, Navigation, and Surveillance/Air Traffic Management (CNS/ATM). Automatic Dependent Surveillance-Broadcast (ADS-B) is a surveillance technology. As an indispensable part of the Next Generation Air Transportation System and the Single European Sky ATM Research project (Gugliotta, 2009 ), ADS-B works based on the Global Navigation Satellite System (GNSS) (Valovage, 2007) . Fig. 1 shows the schematic diagram of the ADS-B working principle. The aircraft are positioned by the GNSS, and onboard ADS-B OUT system broadcasts the real-time positions and other information (callsign, ground speed, etc.). Other aircraft with ADS-B IN system and ground ADS-B receivers can receive these ADS-B signals, which can improve flight situational awareness and ATM. ADS-B data shown in Table 1 can be recorded and accessible from the Flightradar24 (https://www.flightradar24.com).

The data analysis process is presented in Fig. 2 . Data cleaning has to be performed before analyzing the ADS-B data. The cleaning contains three parts. (1) At the time t i , the aircraft flight height H i should satisfy

where H max denoted the maximum flight height depending on the aircraft performance, and H min denotes the minimum flight height constrained by the airport elevation. (2) The true airspeed TAS i should satisfy TAS i ∈ [TAS min , TAS max ], where TAS min and TAS max are the minimum and maximum operation speeds, respectively. These two values are constrained by the flight envelope, which is available from the Base of Aircraft Data (BADA).

where RC max and RD max are the maximum climb rate and the maximum descent rate. These two values are both positive and related to aircraft performance. The method for calculating R i is based on two adjacent time-series ADS-B data at the time t i and t i+1 :

Combining the flight plans and aircraft types, we can calculate the flight volume in different flight routes and aircraft usage.

The flight operation contains two cycles, Landing and Take-off (LTO) and Climb/Cruise/Descent (CCD), with the boundary of altitude 3000 ft above the airport elevation (Society of Automotive Engineers, 2009). Fig. 3 presents comprehensive explanations of flight phases in detail. In the departure step, the aircraft operational sequence is engine start, taxi to the runway, holding on the ground (sometimes), take-off roll to lift-off, initial climb to power cutback, acceleration, and clean-up and en-route climb. In the arrival step, the sequence is the final approach and flap extension, flare, touchdown and landing roll, taxi from the runway to parking stand/gate, and engine shutdown.

Due to the ADS-B data availability, the exact time of approach, taxi and ground idle, take-off, and climb out in the LTO cycle cannot be determined precisely. Therefore, we refer to the Airport Air Quality Manual (AAQM) (International Civil Aviation Organization (ICAO), 2011) to estimate the Time-In-Mode (TIM) and thrust setting as shown in Table 2 . Four operating phases (approach, taxi and ground idle, take-off, and climb out) are numbered with i = 1, 2, 3, 4, respectively.

The fuel consumption in LTO (F C LTO in kg) can be calculated by the below formula:

where N e denoted the number of engines of each aircraft type; TIM i and F i (in kg/min) represent the operation time and fuel flow in the i th mode as shown in 

According to the BADA, the fuel consumption during the CCD cycle can be estimated based on the flight phase FP i and the aerodynamic configuration AC i . Here FP i is categorized using the flight height change rate R i .

where RC min and RD min are both positive, denoting the minimum climbing rate and the minimum descending rate, respectively. The subscription i denotes the quantity at a given time t i . AC i is determined using the following,

The fuel consumption in CCD (F C CCD in kg) can be calculated by the following formula using the time-series ADS-B data:

The nominal fuel flow, F i (in kg/min) can then be calculated based on : where:

Here C f1 (in kg/(min⋅kN)), C f2 (in knots), C f3 (in kg/min), C f4 (in feet), and C fcr (dimensionless) are fuel flow coefficients, available from BADA. TAS i (in m/s) is the true airspeed, available from ADS-B data. The thrust T i (unit in Newtons) can be calculated by:

Here m i (in kg) is aircraft weight; g is the gravitational acceleration (9.8 m/s 2 ); the change rate C i can be calculated as:

and the aircraft drag D i (in Newtons) is obtained as followed:

where ρ i is the atmosphere density (in kg/m 3 ) related to flight altitude; S (in m 2 ) is the total area of aircraft wings; C D0 and C D2 are related to aircraft types and flight phase FP i , available from BADA (Nuic, 2010) .

The lift coefficient C L,i at a given t i is calculated as

CO 2 , SO 2 , H 2 O (g), NO x , and CO are the primary pollutants, which are labeled with j = 1, 2, 3, 4, 5, 6. According to the International Civil Aviation Organization's Engine Emissions Databank (ICAO's EEDB), aircraft emission for a given pollutant E j (in gram) is the product of fuel consumption F C (in kg) and the engine emission indices (EI j ) (in g/kg), which are available from (Society of Automotive Engineers, 2009),

Fig. 4 provides the number of daily new cases of COVID-19 in Beijing, Shanghai, Guangdong, and Hubei provinces. The twelve flight routes are formed by Beijing (ZBAA), Shanghai (ZSSS), Guangzhou (ZGGG, the provincial capital of Guangdong), and Wuhan (ZHHH, the provincial capital of Hubei) airports. Three essential features in the graph are worth pointing out. First, the number here includes domestic cases that were believed to be infected within China and imported cases from abroad, confirmed at the immigration ports, and then quarantined and treated immediately. Therefore, only the domestic cases would induce alert to the air travel between the airports we are interested in. Second, the spikes in Shanghai in early April (http://news.sina.com.cn /o/2020-04-12/doc-iirczymi5796291.shtml) and from August to December, and the spikes in Guangdong in early June and from August to December are due to the imported new cases. Third, the massive spike in middle February in Hubei is due to a correction of previous reports when the new criteria were applied. Therefore, the two significant outbreaks that severely affected domestic air travel are the one that occurred in Wuhan Hubei from January 23 to early March and the one that occurred in Beijing from middle June to early July. As the epicenter, the complete lockdown in Wuhan lasted 76 days until April 8, during which no commercial flights were allowed (http://www.xinhuanet. com/english/2020-04/08/c_138958718.htm).

COVID-19 pandemic breakout in 2020 resulted in significantly fewer trips compared to 2019. Fig. 5 illustrates the distribution of the total number of flights for each flight route. For simplification, (ZBAA, ZSSS) denotes flight routes from ZBAA to ZSSS and from ZSSS to ZBAA combined. In 2019, flight volume in (ZBAA, ZSSS) was the highest, with flight volume 24453, followed by (ZGGG, ZSSS) 21819. Due to the impact of COVID-19, there was a tremendous decrease in flight volume in 2020 compared to 2019. Flight volume in (ZGGG, ZSSS) became the highest, with 19187 accounted. Herein, the flight volume ratio is introduced to show the variation from 2019 to 2020. It is clear that flight volume ratios in flight routes connecting to Wuhan (ZHHH) are the smallest, especially in (ZBAA, ZHHH) (2731/8234 = 0.332) and (ZHHH, ZSSS) (2250/4830 = 0.466). This is because Wuhan was the epidemic center, and the Wuhan international airport was shut down for all commercial flights for straight 76 days. The total flight volume in 2020 (56430) was only 67.8% of that in 2019 (83189). Fig. 6 compares the flight volume for selected (a) domestic flight routes in China, (b) domestic flight routes in other major countries, and (c) international flight routes. It is evident that by the end of 2020, the commercial flights between major cities in China have largely recovered to the volume before the pandemic due to the successful control of COVID-19. However, the number of domestic flights within the US, Germany, Australia, and England was vastly reduced throughout 2020, except for two occasions between Sydney and Melbourne. The international flight volume also experienced the same trend, even between China and other countries. Apparently, the much slower decline of COVID-19 cases in other major countries than China was to blame.

This section shows the airline companies' response to fewer passengers from the perspective of aircraft usage. Because fuel consumption differs in terms of aircraft types and affects airlines' financial performance, accommodation of different aircraft types became a critical way to unwind the declined demand. There were seven aircraft types (A320, A321, A332, A333, B738, B77W, B788) widely used for the selected flight routes. Fig. 7 (a) shows B738 and A333, two types of the most frequently used aircraft, accounted for the majority of the decrease. Interestingly, there was a slight increase in flight volume in A321 and A332. Fig. 7 (b) and Fig. 7 (d) show a comparison in the percentage of aircraft types between 2019 and 2020. The percentage of B738 accounted for the majority part in 2019 (43.6%) and 2020 (38.4%). Percentages of A320, A321, and A332 experienced an increasing trend, especially A321 from 6.4% to 11.7%. Fig. 7 (c) shows the monthly percentage of usage of different aircraft types. A320 and A321 were used relatively widely in Feb-May 2020, with a remarkable increase in market share. Simultaneously, the percentages of A332, A333, B77W, and B788 all declined. The passenger capacities of different aircraft types are shown in Fig. 7 (f) . The passenger capacities of A320, A321, and B738 are 150, 185, and 180, fewer than other aircraft types. Fig. 7 (e) shows fuel consumption per seat in different flight routes in different aircraft types. The variation between the different routes is due to the travel distance. However, we can easily notice that, for the same route, B738 is the most fuel-saving for each seat than other aircraft types, followed by A320 and A321. This is why the usage of B738 only suffered a brief decline in volume in February, as shown in Fig. 7 (c) , while the usage of A320 and A321 in percentage even jumped to its highest from February to April. Therefore, we can conclude that airlines adopted tactical adjustments by using relatively economical aircraft such as A320, A321, and A738 to reduce cost because fewer passengers traveled by airplane. The percentages of A333, B77W, and B788 decreased because these three aircraft are much larger and heavier than other aircraft types. Such transformation in market share among different aircraft types in China lasted until August when the fleet configuration before the pandemic was restored.

To compare the difference in fuel consumption and aircraft emissions between 2019 and 2020, we investigate the changes in four quarters, shown in Table 3 . The ratio of 2020 to 2019 is ~0.38 in Q1 and Q2, and ~0.85 in Q3 and Q4. CO 2 , SO x , and H 2 O emission ratios of 2020 to 2019 are the same as fuel consumption because emissions of CO 2 , SO x , and H 2 O are proportional to fuel consumption (Lai et al., 2020) . However, NO x and CO emissions ratios are slightly different because engine thrust level affects NO x and CO emissions. The high temperature in the internal engine accelerates NO x formation, so less NO x is generated during the taxi-in and taxi-out mode with a low engine thrust level of 7% . In contrast, CO emissions in taxi-out and taxi-in modes are much higher than other modes of LTO because lower thrust level results in incomplete fuel burning, leading to more significant CO emissions. It can be summarized from Table 3 that fuel consumption and aircraft emissions reduced significantly (over 60%) in the first half-year 2020 and gradually returned to more than 80% in the second half-year 2020.

We also compare the change in daily fuel consumption to quantify the short-term and medium-short impact of COVID-19. Fig. 8 shows a sharp decrease in fuel consumption starting January 23, 2020 (indicated by the purple dot line), when the lockdown in Wuhan was announced. During the lockdown, daily fuel consumption values in the flight routes connecting to Wuhan in 2020 were zero. Fuel consumption started a noticeable increase after the Labor Day holiday on May 1, 2020. However, another travel restriction starting June 14 (indicated by the orange vertical dot line) was due to the secondary COVID-19 outbreak in Beijing because some COVID-19 cases caused by salmon were confirmed. Therefore, we can see a subsequent sharp decrease in fuel consumption in flight routes connecting to Beijing. At the end of 2020, the daily fuel consumption was nearly equal to that at the beginning of 2020. Likewise, aircraft emissions were also in the same trends because of proportional relation to fuel consumption.

The COVID-19 pandemic has caused a global lockdown and significantly reduced flight volume since January 23, 2020. In this study, four representative airports (Beijing, Shanghai, Guangzhou, and Wuhan) in China are selected because of their political, economic importance and the centre stage during the pandemic. We focused on twelve domestic flight routes formed by four airports to investigate the short-term and medium-term impact of COVID-19 on flight volume, aircraft usage, fuel consumption, and aircraft emissions from commercial flights.

Detailed analysis shows several important results. The total flight volume ratio of 2020 to 2019 is 67.8% for the selected Chinese domestic routes. The fuel consumption and aircraft emissions in the first half year of 2020 were reduced to ~0.38 of that in 2019; the ratio rebounded to ~0.85 in the second half of 2020. This is a straightforward conclusion as a result of decreased air travel activity associated with COVID-19. The findings in this research can help the aviation authorities with the assessment of aircraft emission impact and environmental impact for the green and sustainable air transport system in the future. The international travels connecting major cities in China to other countries were still severely impacted. Therefore, the long-term recovery for the air industry in the domestic market is leading the international market. Comparison analysis also shows that domestic flights in other major countries (the US, Germany, Australia, England) were also significantly lagged behind China, suggesting the COVID-19 control as the critical factor. Thanks to the successful control of COVID-19, at the end of 2020, the commercial air travel activity in these routes has mostly fully recovered. In the domestic market, the initial recovery of flight volume between non-epicenter cities (Guangzhou and Shanghai) took as long as six months from middle February to July (Figs. 6a and 8b). Different from the initial recovery, we believe that the reopening of the epicenter Wuhan on April 8, 2020, injected significant confidence in the safety of air travel. Therefore, much faster recovery rates are found in May, one month after the reopening, and in July another month after the outbreak in Beijing. This suggests that the sustained low number of new COVID-19 cases would lead to a fairly fast recovery of air travel on a time scale of ~ 1-2 months. The recovery time may differ in different cities and countries due to government measures, people awareness, and medical conditions. However, for a certain city or country, there may be some characteristics or features, which are related to the resilience of air transportation systems. Preference for smaller aircraft is a natural choice of airline companies to reduce financial damage. Our calculation shows that aircraft A320, A321, and B738 were more widely used for flight plans compared with other aircraft types because these models have the lowest fuel consumption per seat. The empirical situation in China can somehow provide a true and promising benchmarking for future commercial flight management. We hope other countries can learn the successful experience from China and recover flight operation as soon as possible. During the recovery phase, airlines could reduce some unnecessary costs and make good use of the existing aircraft such as converting passenger aircraft to cargo aircraft.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. 

Government support to airlines in the aftermath of the COVID-19 pandemic

European airlines' strategic responses to the COVID-19 pandemic

Long-term concentrations of nitrogen dioxide and mortality: a meta-analysis of cohort studies

Ultra Long-Haul: an emerging business model accelerated by COVID-19

Impact of coronavirus outbreak on NO2 pollution assessed using TROPOMI and OMI observations

Puzzling haze events in China during the coronavirus (COVID-19) shutdown

Covid-19-navigating the Uncharted

An air-traffic upgrade to improve travel by plane

The Silver Lining of COVID-19: Estimation of Short-term Health Impacts Due to Lockdown in the Yangtze River Delta Region

Estimating and projecting air passenger traffic during the COVID-19 coronavirus outbreak and its socio-economic impact

Formation and radiative forcing of contrail cirrus

Empirical Method for Estimating Aircraft Fuel Consumption in Ground Operations

Investigating the evolving context of an unstable approach in aviation from mental model disconnects with an agentbased model

Aircraft emission inventory and characteristics of the airport cluster in the Guangdong-Hong Kong-Macao greater bay area

A SWOT analysis of China's air cargo sector in the context of COVID-19 pandemic

The impacts of COVID-19 on the global airline industry: an event study approach

Air carrier's liability for the safety of passengers during COVID-19 pandemic

User manual for the Base of aircraft data (BADA) revision 3.10. Atmosphere

Calculation of aircraft fuel consumption and CO2 emissions based on path profile estimation by clustering and registration

The future of airport post COVID-19

Fuel Estimation in Air Transportation: modeling global fuel consumption for commercial aviation

On the degree of synchronization between air transport connectivity and COVID-19 cases at worldwide level

COVID-19 pandemic and air transportation: successfully navigating the paper hurricane

How did COVID-19 impact air transportation? A first peek through the lens of complex networks

The contribution of carbon dioxide emissions from the aviation sector to future climate change

Enhanced ADS-B research

The COVID-19 epidemic

A method for computing flight operation fuel burn and emissions based on ADS-B trajectories

An aircraft performance model implementation for the estimation of global and regional commercial aviation fuel burn and emissions

Multi-objective flight altitude decision considering contrails, fuel consumption and flight time

Measuring imported case risk of COVID-19 from inbound international flights -a case study on

Exploring the roles of high-speed train, air and coach services in the spread of COVID-19 in China