key: cord-1015294-2pjftgac
authors: Zhang, Linfeng; Yang, Hangjun; Wang, Kun; Zhan, Yi; Bian, Lei
title: Measuring imported case risk of COVID-19 from inbound international flights ––– A case study on China
date: 2020-08-28
journal: J Air Transp Manag
DOI: 10.1016/j.jairtraman.2020.101918
sha: 8e19ee79a944137ee36e2d7ba43e4b5a280628d9
doc_id: 1015294
cord_uid: 2pjftgac

With COVID-19 spreading around the world, many countries are exposed to the imported case risk from inbound international flights. Several governments issued restrictions on inbound flights to mitigate such risk. But with the pandemic controlled in many countries, some decide to reopen the economy by relaxing the international air travel bans. As the virus has still been prevailing in many regions, this relaxation raises the alarm to import overseas cases and results in the revival of local pandemic. This study proposes a risk index to measure one country's imported case risk from inbound international flights. The index combines both daily dynamic international air connectivity data and the updating global COVID-19 data. It can measure the risk at the country, province and even specific route level. The proposed index was applied to China, which is the first country to experience and control COVID-19 pandemic but later becoming exposed to high imported case risk after the pandemic centers switched to Europe and the US afterward. The calculated risk indexes for each Chinese province or region show both spatial and temporal patterns from January to April 2020. It is found that China's strict restriction on inbound flights since March 26 was very effective to reduce the imported case risk by half than doing nothing. But the overall index level kept rising because of the deteriorating pandemic around the world. Hong Kong and Taiwan are the regions facing the highest imported case risk due to their superior international air connectivity and looser restriction on inbound flights. Shandong Province had the highest risk in February and early March due to its well-developed air connectivity with South Korea and Japan when the pandemic peaked in these two countries. Since mid-March, the imported case risk from Europe and the US dramatically increased. Last, we discuss policy implications for the relevant stakeholders to use our index to dynamically adjust the international air travel restrictions. This risk index can also be applied to other contexts and countries to relax restrictions on particular low-risk routes while still restricting the high-risk ones. This would balance the essential air travels need and the requirement to minimize the imported case risk.

With the improvement of transportation infrastructure and the reduction in travel cost around the world, international travels, especially by air, have dramatically increased. In 2018, the global air passenger travel has reached 8.3 trillion revenue passenger-kilometers (RPKs) (IATA, 2019) . The global RPKs have kept an average annual 6% growth since 2010. Thanks to the globalization and increasing liberalization of bilateral air service agreements (ASAs), the international airline network and air passenger have been expanded dramatically (Winston and Yan, 2015; Oum et al., 2019) . China and other Asia Pacific regions lead the air travel increase in the world (Wang et al., 2018) . China has been the world's second largest airline market since 2005 (Fu et al., 2015) . In 2018, the market served 1.27 billion passenger travels, with 60 airlines, 3,639 transport aircraft, 4,945 regular flight routes, 230 domestic cities with regular flights (excluding Hong Kong, Macao, and Taiwan), operating flights to 165 cities in 65 countries (Civil Aviation Administration of China (CAAC), 2018). Such domestic and international air connectivity allows people to travel around the world easily, stimulating trade and people-to-people exchanges. But it may also facilitate infectious disease to spread rapidly around the world. International air travel has acted as important media to contribute to fast spread of several pandemics in the past, for example the SARS in 2013 and H1N1 in 2009 (Wilder-Smith et al., 2013; Browne et al., 2016; Cai, 2019) .

The first COVID-19 case was recorded in Wuhan China on December 9, 2019 (Huang et al., contain COVID-19 spread domestically and globally (Oum and Wang, 2020) . For example, China has quickly locked down Wuhan and the surrounding regions since January 23, suspending all intra-city public transit, and blocked all inter-city and international travels between Wuhan and other regions. Other Chinese cities also responded actively, banning inter-provincial travels and cutting inbound flights. With many earliest overseas confirmed cases reporting Wuhan travel history, China and other countries greatly reduced international flights to China and imposed strict inspections on Chinese flights. Many airlines also voluntarily canceled most of their flights with China even before their governments formally responded to mitigate virus spread risk, including British Airways, Lufthansa, Air Canada, etc. 1 Thanks to the government's strict quarantine and case tracing measures, the pandemic in China has been controlled since early March. The hospitalized cases have dropped from 58,090 in mid-February to less than 20,000 at the end of March. On April 8, 2020, Wuhan reopened after more than 10 weeks in lockdown. Air and train services have resumed.

However, COVID-19 has quickly spread globally, leading to a serious pandemic in Europe and then North America. The pandemic centers are thus switched from China to these two regions, which in turn makes China to be exposed to high imported case risk via its international air connectivity. In particular, China reported more than 300 imported cases from other countries, with about 100 cases confirmed at Beijing airport, where the largest amount of direct overseas flights operated. Thus, to minimize such imported case risk, since March 29, CAAC has promulgated the "one-airline, one-country, one-flight-per-week policy" (CAAC, 2020) . That is, only one airline from China or the foreign country is allowed to operate the route between two countries; One route can only have only one flight per week;

China only maintains direct flights to one destination of each foreign country. In addition, CAAC launched a comprehensive inspection and quarantine procedure for the inbound airline passengers. Now, many countries consider reopening the economy and relaxing the international travel 1 Please see the following links for information: https://www.businessinsider.com/airlines-canceling-changing-flights-to-china-amid-coronavirus-fears-2020-1 J o u r n a l P r e -p r o o f ban. For example, more than forty states in the US decided to open up in May and several European countries designed multiple-phase reopening plans, which have been implemented since early May. However, the pandemic has not been well controlled in many countries, such that the aggressive relaxation of international air travels restrictions would inevitably bring about high imported case risk. This could result in a revival of pandemic in many places where the COVID-19 spread has been initially controlled. Thus, before fully removing the international air travel ban, it is important to comprehensively evaluate the imported case risk.

This study thus aims to propose an index to evaluate the imported case risk of COVID-19 through international air travel. Specifically, we use the Chinese data to construct such an index in order to measure the imported case risk faced by each Chinese province. The index explicitly accounts for the international air connectivity between China and the foreign countries, and also the detailed pandemic data in each foreign country. The risk index is calculated for each Chinese province on a daily basis from January 1 to April 16, 2020. Both the spatial and temporal patterns of the imported case risk can be examined in detail.

Moreover, we can also evaluate how such risk is affected by the government's restriction on international air travel and the dynamic change in global pandemic development in each foreign country.

Such a risk index also has important policy implications. The policy maker can apply this index to dynamically monitor and adjust its air travel restrictions with each foreign country. For example, they can consider relaxing the air travel restriction and resuming air connectivity with those low imported risk regions, but probably further tightening or lengthening the duration of the restrictions with those high imported risk regions. With more and more countries succeeding in controlling the domestic pandemic, such imported case risk index can be applied in more contexts to provide in-advance alert and also suggest a more effective strategy to gradually lift international air travel bans, while minimizing the imported case risk and resultant pandemic revival.

The paper is organized as follows. Section 2 reviews the related recent studies on the J o u r n a l P r e -p r o o f interrelationship between international air travel and COVID-19 global spread. In Section 3, we combine the real-time data of global flights with the dynamic global COVID-19 pandemic data to construct an index measuring imported case risk for each country with which China has air connection. Section 4 discusses the results of the index calculation and tries to distinguish the patterns from temporal, spatial, source, and route distribution perspectives, respectively.

The policy implications are also discussed in this section. Section 5 concludes this study.

Since China's COVID-19 pandemic started in January, many scholars around the world modeled and predicted its transmission path and growth trajectory. In particular, people flow plays an essential role in the fast spread of COVID-19. For example, Tang et al. (2020) calibrated a random transmission model of COVID-19 to predict the virus spread within China by using the people outflow from Wuhan to different parts of China by different transport modes, such as airlines, rail, and coaches. Lau et al. (2020) evaluated the effect lockdown of Wuhan and the air travel ban to contain China's domestic pandemic. They found the air travel ban has greatly helped slow the COVID-19 spread in China. Zhang et al. (2020) compared the effect of different transport modes on COVID-19 spread in China. They found that air transport is more closely associated with the virus spread than high-speed train.

Later on, more attention has been paid on the global spread via international travels, especially air travel. Gilbert et al. (2020) J o u r n a l P r e -p r o o f

This section develops the index to measure the imported case risk through international air travel for one region. We apply this index onto different Chinese regions, including 31

provinces, Hong Kong, Macao, and Taiwan regions (hereafter, "province" is used to refer to each Chinese region). Specifically, the imported case risk from one particular foreign country should be positively correlated to the real-time bilateral air connectivity and also the pandemic situation of this foreign country. Our index thus explicitly accounts for both the air connectivity and also the severity of pandemic of the foreign country. Table 1 collates the notation glossary used in our calculation methods. 

In order to estimate the new cases on day , namely , we use the confirmed new cases in previous days. This is achieved by solving the following ordinary differential dynamical system.

To solve for the above ordinary differential dynamical system, we need to give an explicit and multi-step discrete numerical scheme. We select a two-step eccentric scheme, called

Admas-Bashforth scheme (Beeman, 1976) , This method is widely used in disease spread, physics, and biology research to make correction of the time-series changes for one variable (Xiao et al., 2009; Khan et al., 2019; Kumar et al., 2020) .

where ∆ is the time difference, which is defined as one day in our model. This leads to the following expression of as Eq. (4).

where and '( are the numbers of new confirmed cases on days and − 1 of the region .

As most of the countries in the world do not report the imported cases or they are unable to trace such cases accurately, we adopt the following approach to estimate the number of imported cases of the region , which is , by Eq. (5). 

Then for each Chinese province indexed by ., we are able to define its imported case risk on day as follows,

where , is the air connectivity between Chinese province . and foreign country on day .

The proposed imported case risk index involves several assumptions to simplify calculation and to generate clearer implications. First, to measure the share of infected population in foreign regions, namely the variable , we treat most of the foreign countries as a whole, not distinguishing the different provinces/states/cities (airports) within that country . This is mainly due to the unavailability of the detailed pandemic data at the more specific province/state/city level for most countries, including Russia, Canada and the European countries. But, as will be discussed later, the US has maintained very detailed pandemic data available at the state level, such that we are able to evaluate the imported case risk from each individual US state. With daily more than fifty thousand new confirmed cases reported each day since July, the US has accumulated more than 3.5 million cases, suffering the most severe pandemic in the world. The pandemic conditions also vary significantly among different states over time. governments and airlines impose strict rules to control the load factor of the international flights, and a widely adopted policy is to leave one empty seat between two passengers.

Therefore, using the scheduled seats might exaggerate the air connectivity. But as almost all the airlines adopt the same principle, it would not change the relative level of imported risk from different countries when compared among regions. Third, the index accounts for the first entry point of the international flights. Since late March, China diverted many international flights from Beijing, Shanghai and Guangzhou to other secondary airports in

Xi'an, Shenyang, Shijiazhuang, Tianjin, etc. This is to alleviate the huge workload of the gateway airports after strict inspection and quarantine policies have been implemented upon the inbound flights and passengers. 3 Last, our index does not consider the effect of strict quarantine and inspection adopted on inbound flights. The index only measures the potential imported case risk brought by the international air connectivity from foreign countries with 3 All inbound airline passengers were asked to conduct nucleic acid testing and undergo 14-day quarantine since the pandemic. Instead, the index can help policy makers adjust the strictness to inspect the inbound flights for provinces with different risk levels.

The absolute value of the index does not have a specific meaning, while it is intuitive that a higher value of indicates a higher imported case risk for province . in China. As an alternative, we normalize into a relative risk index as below, 

Our data involves both the dynamic real-time international flight information and the 

This section presents our calculated imported case risk indexes for China and its different provinces. The evolution of risk patterns is also discussed with the policy implications suggested. Specifically, this section includes the following four parts. First, we examine the evolution of China's overall imported case risk. Second, the imported case risk for each province is demonstrated and ranked. Third, we further identify specific foreign countries that impose the highest imported case risk for each Chinese province. Finally, we investigate the imported case risk on different major Chinese international routes. This would provide more specific policy implications for the Chinese government to dynamically adjust the international travel restrictions for individual routes or airlines.

First, China's overall imported case risk can be obtained by aggregating the risk index of each Chinese province (i.e., Eq. (8)). This overall risk index is shown in Figure 2 6 Shanghai is selected as the baseline as it is the most economically developed city in China, and it is also the most connected Chinese city with the world by air and maritime transport. Shanghai also has the fastest and most frequent update on daily COVID-19 pandemic development in China. The accuracy and reliability of its released data could be the highest in China. In addition, February 2 is selected is because the first confirmed case in Shanghai was reported on January 20 (Yao, 2020) . The latent period for COVID-19 can be as long as two-week (14 days). Thus, the transmission and outbreak caused by the first reported case can start as late as February 2. We therefore use February 2 as the remarking date for Shanghai to face the real challenge of COVID-19 outbreak. 

In this section, we further identify the major foreign countries (for the US, the scope is the individual state) which impose the imported case risk to each Chinese province. Such analysis could provide more direct implications for policy makers to tailor its flight restrictions targeted to different foreign countries. As China's imported case risk was low in February, our analysis is thus conducted for March and April. Table 5 exhibits the main foreign countries or US states that impose the imported case risk for the five riskiest Chinese provinces in March and April.

In early and mid-March, China was mainly exposed to the imported case risk from several Asian As shown in our previous section, the risk indexes of Hong Kong, Beijing, and Shanghai are highly correlated to the pandemic development in Europe and the US. Figure 8 shows the evolution of daily confirmed cases of major pandemic countries, including South Korea, Japan, Singapore, Spain, the US, Italy and Russia. It can be seen that the number of confirmed cases in each foreign country explains their imported case risk imposed on Chinese different provinces. Figure 8 . Evolution of daily confirmed cases for major countries imposing the risk of imported cases for China

In April, China still faced significant imported case risk from Europe and the US, despite its significant cut of inbound flights. Especially for Beijing and Shanghai, they were still exposed to very high risk of imported cases. This is mainly because mainland China has not cut its flights with Hong Kong. A significant amount of the US and European passengers can still travel to mainland China by transferring from Hong Kong, while Hong Kong did not implement as strict flight cut and inspection policies as adopted in mainland China.

Most of the countries have also adopted stricter international air travel bans and cut the international air connectivity. This indeed helped China to control the imported case risk as well. 

This subsection focuses on more micro-level risk by examining individual routes between China and foreign countries or different states of the US. First, as shown in shown in Table 7 . The top-ranked routes in China not only concentrated in Taipei and Beijing, but also involved some secondary Chinese airports such as Yanji, Xi'an, and Shenyang. This is because these cities have been designed as the first entry points for the inbound international flights, which were originally scheduled to land in Beijing, Shanghai, and Guangzhou. reporting an increasing number of imported cases in April.

Such a route-level risk index is more useful for the relevant stakeholders to adjust international air travel restrictions, which can be specific to an individual Chinese airport and a particular foreign country. Because our proposed approach is based on the real-time data, the risk index can be calculated and updated day by day. It would help policy makers, airlines, customs, border inspections, epidemic prevention departments and the hospitals to prepare in advance. It is useful to judge whether the future flight plan can be adjusted according to different risk levels. This provides a theoretical and quantitative basis for timely flight plan adjustment. Especially, for the low-risk routes, the air travel ban can be lifted to resume the normal operation as soon as possible. This is important to mitigate the negative economic and social impact of the COVID-19

pandemic. 

International air travel has contributed to the global spread of COVID-19. Although many countries have controlled local pandemic or seen the curve of newly reported cases flattened, there has been no sign for the virus to disappear in the near future. Given the fast spread of COVID-19 around the world, it is still essential and even urgent for each country to stipulate effective measures to control the imported case risk from inbound international flights. A too soon relaxation of international air travel restriction is likely to result in new imported cases from overseas and lead to local pandemic revival. Moreover, it is necessary to accurately measure the imported case risk of inbound flights from different foreign countries and even specific routes. This is useful to tailor the restriction rule based on different risk levels. This helps mitigate the imported case risk, while minimizing the impact on the essential international air travels. This paper proposed a risk index to measure the risk of imported cases by inbound international flights. The index combines both international air connectivity and COVID-19 pandemic data.

The method was applied to calculate risk index for each province in China. With the COVID-19 pandemic first emerged in Wuhan in January 2020, China has adopted a series of powerful measures, including the lockdown of Wuhan and other cities. It has been the first country to effectively contain the pandemic and reopen the domestic economy. However, the country was then exposed to very high risk of imported cases, when COVID-19 quickly spread globally.

Our calculated risk indexes for China have shown clear spatial and temporal patterns. First, there is clear evidence that China's inbound flight restriction since late March has been quite effective to control the imported case risk. Our counterfactual analysis showed that the overall risk index level has been reduced by half thanks to the strict restriction. Second, China's imported case risk has kept growing from January to April, despite the cut of international air connectivity by China and many other countries. This is mainly because COVID-19 was spreading extremely fast around the world, which dominated the air connectivity reduction to raise the imported case risk.

This thus calls for continuous caution for policy makers to restrict international air travel from high risk regions. Third, in February and early March, South Korea and some other Asian J o u r n a l P r e -p r o o f countries imposed higher imported case risk to China. Later, Europe and the US became the pandemic centers since mid-March. This made China face more imported case risk from the inbound flights from these two regions. Fourth, it is noted that Hong Kong did not strictly restrict its international air connectivity, while mainland China has still kept most of the flights with Hong Kong. This thus indirectly made mainland China exposed to higher imported case risk from those overseas passengers who can transfer from Hong Kong to enter mainland China. Last, we also examined the risk index for each province and identified the major foreign countries that imposed high risk. Some route-level risk indexes have also been scrutinized. The proposed index can provide useful references for policy makers, airlines and other relevant stakeholders to adjust the international air travel restrictions in advance.

This study, however, is also subject to some limitations, which open avenues for future studies.

First, for the large foreign countries, we focus on the US and distinguish the imported case risk from different US states. While for other large countries, such as US and Russia, we are not able to examine at the province level, simply because of the data unavailability. But the research method and framework can be directly applied to measure the province-specific imported case risk for these countries, once more micro-level data become available. Second, as most of the connecting flights and transfer passengers have been banned for international air service, our imported case risk now focuses only on the direct international flights. When the restrictions are gradually lifted, the connecting flights would rebound and transfer passengers are allowed to travel, the existing method can be easily adapted to account for the imported case risk of the connecting flights. As long as having the detailed itinerary information of the passengers or at least knowing the share of origin and transfer passengers per flight, we should be able to distinguish how much the imported case risk from the origin and intermediate countries of the particular flight. These are all very meaningful extensions with valuable policy implications in the near future, but are out of the scope of this paper.

J o u r n a l P r e -p r o o f

The SARS-CoV-2 epidemic outbreak: a review of plausible scenarios of containment and mitigation for Mexico

Some multistep methods for use in molecular dynamics calculations

Risk assessment of novel coronavirus COVID-19 outbreaks outside

The roles of transportation and transportation hubs in the propagation of influenza and coronaviruses: a systematic review

Roles of Different Transport Modes in the Spatial Spread of the 2009 Influenza A (H1N1) Pandemic in Mainland China

The effect of travel restrictions on the spread of the 2019 novel coronavirus (COVID-19) outbreak

An interactive web-based dashboard to track COVID-19 in real time

Low cost carrier competition and route entry in an emerging but regulated aviation market-The case of China

Preparedness and vulnerability of African countries against importations of COVID-19: a modelling study

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

World Air Transport Statistics

COVID-19 Dashboard by the

Fractional order ecological system for complexities of interacting species with harvesting threshold in imprecise environment

A study of fractional Lotka Volterra population model using Haar wavelet and Adams Bashforth Moulton methods

The positive impact of lockdown in Wuhan on containing the COVID-19 outbreak in China

Prospects for air policy liberalization in China as a Result of China-ASEAN Open Skies: Changing role of Chinese mega carriers in global scene and anticipated Low Cost Carrier competition

Identification of critical airports for controlling global infectious disease outbreaks: Stress-tests focusing in Europe

Measuring the effects of open skies agreements on bilateral passenger flow and service export and import trades

Socially optimal lockdown and travel restrictions for fighting communicable virus including COVID-19

Stochastic discrete epidemic modeling of COVID-19 transmission in the Province of Shaanxi incorporating public health intervention and case importation

Estimation of COVID-19 outbreak size in Italy based on international case exportations

Key determinants of airline pricing and air travel demand in China and India: Policy, ownership, and LCC competition

Low risk of transmission of severe acute respiratory syndrome on airplanes: the Singapore experience

Open Skies: estimating travelers' benefits from free trade in airline services

A multi-step heart rate prediction method based on physical activity using Adams-Bashforth technique

Connectivity at Chinese airports: The evolution and drivers

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

We would like to thank the anonymous reviewer for providing us with many excellent comments and suggestions, which lead to significant improvement of this paper. We also acknowledge gratefully the financial support from National Social Science Foundation of China (18ZDA071) and National Natural Science Foundation of China (71901065).

A method is proposed to calculate the imported case risk of COVID-19 by international air transport.We calculated the imported case risk imposed by foreign countries on Chinese provinces.The imported case risk index accounts for both foreign country's pandemic condition and the air connectivity with China.The results of imported case risk index are discussed with policy implication suggested.J o u r n a l P r e -p r o o f