As it is difficult to obtain detailed data on the number of tourists and tourism revenue, the core data for this study on the measurement of tourism resilience is the operating revenue of star hotels in each quarter of each province, which comes from the quarterly National Star Hotel statistical report issued by the Ministry of Culture and Tourism of the People’s Republic of China. It should be noted that the operating income of star-rated hotels can only reflect the situation of the tourism industry to a certain extent as it only reflects the income brought by tourists’ accommodation, but cannot represent the tourism activities that occur without the accommodation demand of star hotels. The data collected covers the period from the first quarter of 2018 to the second quarter of 2021, and includes the data of 31 provincial administrative regions in China (hereinafter referred to as “provinces and regions”) excluding Hong Kong Special Administrative Region, Macao Special Administrative Region and Taiwan. After preliminary sorting, the following characteristics were found.

In order to more clearly show the overall changes in the revenue of star hotels in various provinces and regions across the country before and after the epidemic, only the trends of the top three, the middle three, the last three provinces and regions in the ranking of pre-epidemic star hotel revenue, and the national average of star hotel revenue are displayed (Fig. 1). In the first wave of COVID-19, the revenue of the nationwide star-rated hotels declined, and then it rose gradually. Under the impact of the second wave of the epidemic, the decline of operating income decreased, showing that it had a certain toughness. Before the outbreak of the epidemic, the operating revenue of national star hotels was relatively stable, and the national total operating revenue remained at about 49 thousand million yuan in each quarter of 2018 and 2019. In each year, the total operating revenue of national star hotels shows an overall upward trend over time. In the fourth quarter of 2018, it increased by 7.9% compared with the first quarter, and then resumed growth after decreasing by 17.3% in the first quarter of 2019. The overall operating revenue in 2019 increased by 13.4%. Due to the strong impact of COVID-19 in early 2020, the total revenue of the star hotels in the first quarter dropped by 63.9%, and the vulnerability of the national tourism industry under the epidemic situation was highlighted. Subsequently, the total operating revenue continued to rise. In the fourth quarter of 2020, the total revenue increased by 107.1% compared with the first quarter. The tourism industry gradually recovered and developed well, but it was still at a lower level than before the epidemic. In the first quarter of 2021, the total operating revenue of national star hotels decreased compared with the previous quarter, but increased by 52% compared with the first quarter of 2020. In the second quarter of 2020, the total operating revenue still maintained an upward trend, with a year-on-year increase of 71% in the second quarter of 2020. Therefore, although the epidemic has caused a serious blow to the domestic tourism industry, it still maintains a trend of continuous growth of tourism revenue over time, and the overall level shows an increasing trend year by year. This change not only includes the internal trend characteristics of the tourism industry changing with the quarters, but also highlights the internal motivation of resilience to promote the tourism industry to gradually return to normal, which is worthy of further discussion.

Referring to the resilience measurement model of Zhou et al. (2019), the “time axis integral of the change range of operating income of star hotels”, which integrates the time change and time accumulation, is defined as the “resilience characterization quantity”. Based on this, the tourism resilience measurement model is constructed, and its basic expression is:

where R is the resilience index, S' represents the time axis integral of the predicted trajectory, ∆S represents the time axis integral of the difference between the actual trajectory and the predicted trajectory, t is the time, f'(t) is the actual trajectory of the epidemic impact, f (t) is the predicted trajectory without the epidemic situation, and t1 and t2 are time points, as shown in Fig. 2.

Formula (1) is the general formula of the resilience measurement, and different time periods correspond to resilience indexes with different connotations. The tourism resilience of the whole investigation period from the first quarter of 2020 to the second quarter of 2021 is defined as the comprehensive tourism toughness index. The tourism toughness from the beginning of a wave to a specific time point is defined as the time section resilience index (noting that the first wave was in 2020 and the second wave was in the first two quarters of 2021). Generally, the larger the R value, the smaller the ΔS value and the stronger the tourism resilience, and vice versa. If R=1, this indicates that the epidemic had no significant impact on tourism; if R>1, the revenue of star hotels is higher than predicted, which may be due to major technological progress, improvement of management efficiency, the launch of innovative tourism products and other reasons which establish a new growth path. For ease of expression, the “resilience index” mentioned in this study refers to the resilience index of a certain time section unless otherwise specified.

Generally, the national tourism resilience gradually increased over time in 2020. After a slight decline in the first quarter of 2021, it resumed an upward trend in the second quarter. The resilience index increased from 0.39 to 0.67, which shows that the fundamental improvement of the epidemic situation has helped tourism to resume. After the first round of the epidemic hit the country, the introduction and implementation of various prevention and control policies greatly affected the flow of personnel, but at the same time they fundamentally reversed the epidemic situation, and the tourism industry as a whole was slowly recovering. At the beginning of 2021, the tourism industry was hit again due to the second wave of the large-scale winter epidemic, but the resilience index only dropped slightly, almost the same as in the fourth quarter of 2020, and it quickly increased to a new high in the second quarter. The seven provinces and regions of Zhejiang, Jiangsu, Henan, Yunnan, Hainan, Jiangxi and Jilin, are completely consistent with the overall trend of tourism resilience, accounting for only 21.88% of the total sample.

The tourism resilience trends of most other provinces are quite different, and the “partial outbreak” of the epidemic and corresponding control policies are important reasons why. For example, as is shown in Fig. 3, the resilience index of 14 provinces (municipalities directly under the Central Government) declined in the second quarter of 2020, of which Hubei had the largest decline, exceeding 0.1 (the declines in the other regions were below 0.1). As the first outbreak location of the epidemic, Wuhan of Hubei took on strict lockdown measures which were not fully lifted until April 2020. After the measures were lifted, its tourist attraction still declined due to the impact of the epidemic. In addition, Hubei Province did not lower its response level of public health emergencies from level 1 to level 2 until June 2020, with all existing confirmed and suspected cases being cleared to zero. The severe impact of the epidemic and the long period of prevention and control are the main reasons for the significant decline in tourism resilience in Hubei. In the third quarter of 2020, only Xinjiang’s tourism resilience declined, mainly due to the outbreak of the epidemic in Xinjiang from July to August of that year and the implementation of extremely strict prevention and control measures. Major cities such as Urumqi were almost in a fully “closed” state, and tourism revenue declined more severely. In the first quarter of 2021, the resilience of tourism in 11 provinces and regions across the country declined. Among them, the Beijing Tourism Resilience Index had a largest drop, by 0.19. Under the uncertain situation that the first case of the British mutant virus was discovered in January 2021, as the capital, Beijing had tightened the control of the epidemic during the “two festivals” in winter, New Year’s Day and the Spring Festival, and advocated that travel agencies and online travel companies should not organize travel teams to enter Beijing, and that scenic spots in Beijing strictly control the number of reservations, which resulted in a significant decline in tourism resilience in the second quarter of 2021. The tourism resilience of most other provinces and regions maintained an upward trend in the first quarter of 2021. The increase in tourism resilience in Gansu exceeded 0.2 and the resilience level exceeded 1, ranking first among all provinces. Gansu was a relatively low-endemic area, and products such as “Gansu Lunar New Year” tour group and independent travel were launched during the Spring Festival, making Gansu one of the most popular tourist destinations for the Spring Festival, and the revenue of star hotels even exceeded the pre-epidemic level. In the second quarter of 2021, only the four provinces (municipalities directly under the Central Government) of Tianjin, Gansu, Qinghai, and Tibet experienced a slight decline in tourism resilience (none of which declined by more than 0.04), and the country generally maintained an upward trend.

The spatial pattern of comprehensive tourism resilience during the study period is shown in Fig. 4. In terms of specific provinces and regions, the provinces with comprehensive tourism resilience over 0.75 during the study period included Gansu, Hainan, Guizhou, Hebei, and Shandong. Many factors allowed them to have a high level of tourism resilience. Among them, Gansu, Hainan, and Guizhou were provinces with relatively light epidemics and obvious advantages in tourism characteristics; while Hebei relies on the strong tourism consumption demand of Beijing and Tianjin, and Shandong relies on the recovered customer source from it and neighboring provinces such as Jiangsu, so their tourist industries showed strong resilience as a whole. The provinces with low comprehensive tourism resilience were Tibet, Ningxia, Shanxi, and Beijing, with resilience values below 0.5. The reasons for the low level of tourism resilience in each of these provinces were different. Among them, Ningxia, Tibet, and Shanxi are all provinces with poor tourism development levels. The weak tourism transportation infrastructure construction and related operation and maintenance, the unreasonable proportion of star hotels’ operating income, the lack of tourism management professionals, the insufficient development of tourism resources, the incomplete supporting facilities, the lack of publicity and many other issues significantly magnified the adverse impact of the national epidemic on their tourism industries, so they became the provinces with the lowest level of comprehensive tourism resilience. As the capital, the main reason for the low tourism resilience of Beijing was that its epidemic prevention and control measures were always extremely strict. Since the outbreak of the “epidemic of Xinfadi” in June 2020, the prevention and control of personnel movement has been strictly tightened, and the Civil Aviation Administration issued a “fuse instruction”. The mobility of people inside and outside the city has been severely weakened, and the tourism industry has also suffered a relatively large blow. In short, due to many considerations and factors such as transportation convenience and tourism experience quality, passengers would be less likely to choose the above-mentioned provinces as their travel destinations when there were fewer travel opportunities.

The comprehensive tourism resilience values of most other regions were between 0.5-0.75, showing the overall pattern of “weak north and strong south”, but Guangdong, Guangxi, Hunan, Hubei, Chongqing and Shanghai were less resilient than other southern regions, showing certain particularity. The three northeastern provinces, Inner Mongolia and Xinjiang are located in the northern border regions of China, which were facing certain risks of imported cases during the epidemic. For example, the main epidemic risk threat of China in April 2020 was the imported cases from Russia. Heilongjiang Province accounts for the largest number of imported cases which caused a locally-associated epidemic. Also, there were cases imported from Russia into Inner Mongolia in November 2020, and an epidemic of cold-chain food imports in Tianjin during the same period. Relevant prevention and control measures would severely hinder the inflow of tourists and affect the development of the local tourism industry. In addition, the tourism foundations of most northern regions are relatively weaker than those of the south, especially due to the geographical location, and the tourism industry of northern cities in China is under greater pressure in autumn and winter. Coupled with the repeated anticipation of the winter epidemic, people’s willingness to choose northern cities as tourist destinations had become weaker, which to a certain extent inhibited the recovery of the tourism industry in northern China. Therefore, Beijing was more severely affected, while Shandong and Qinghai also showed a consistent trend of revenue decline in the fourth quarter of 2020. The “Matthew Effect” of the tourism industry under the impact of emergencies was manifested. Among the southern provinces, Hubei Province has been most affected by the epidemic, and its tourism industry has also suffered a greater impact. At the same time, it has affected neighboring areas such as Chongqing and Hunan since it is the main source of tourists. The average revenues of star hotels before the epidemic in Guangdong and Shanghai were ranked second and third. The scale of tourism was relatively large, but the resilience performance under the epidemic was mediocre. Similarly, the reason is that these two places are important transportation hubs at home and abroad, and the mobility of people is extremely high under normal circumstances. Under the control of the epidemic, there was a large loss of foreign tourists, and the high risk of cases imported from overseas further affected the recovery of tourism in these areas. The main source of tourists from other provinces in Guangxi is Guangdong, which also affected its tourism resilience to a certain extent.

Based on the trend of national tourism resilience changes, it is easy to see that tourism resilience was at its lowest level in the first two quarters of 2020 and the improvement was slow. Tourism resilience recovered rapidly in the third and fourth quarters. After a small shock in the first half of 2021, it continued to maintain a growth trend. The degree of recovery varies across provinces (municipalities directly under the Central Government) over time. On this basis, the resilience changes during the investigation period could be divided into three stages.

The first stage (the first quarter to the second quarter of 2020, shown in Fig. 5): Hard resistance. All parts of the country have been severely affected by the epidemic, the first-level responses to major public health emergencies have been launched one after another, and tourism resilience has successively reached the lowest level (except in Beijing). In the first half of 2020, the spread of the epidemic caused the vulnerability of the tourism industry to be revealed, which caused a significant decline in tourism resilience across the country. In the first and second quarters, the average tourism resilience of each province (municipalities directly under the Central Government) was 0.37 and 0.39, respectively, that is, the national tourism market had shrunk by nearly two-thirds compared to before the epidemic. Ningxia, Tibet, Shanxi, Inner Mongolia, Hunan, Sichuan, and Chongqing had minimum resilience values of less than 0.3, and the southwestern and southeastern coastal areas were also seriously affected. Among them, the local epidemics in Tibet and Ningxia provinces were relatively mild (as of October 2021, Tibet had confirmed a total of one case, and Ningxia had confirmed a total of 77 cases), but the two provinces were at an extremely low level of tourism resilience throughout the study period. This is because tourism in Tibet is restricted by terrain, weather, and road conditions, and the convenience of round-trip travel is lower than that of other regions. In addition, the traffic interruption caused by the epidemic and the extremely strict control policies have further reduced the willingness and possibility of tourists to travel to Tibet. In addition, the average revenues of star hotels in Ningxia and Tibet before the epidemic were the two lowest in the country, and the tourism foundation was extremely weak. During the epidemic, tourists were less concerned about such tourist destinations given the limited choices of tourist destinations. Similarly, the average revenue of star hotels in Inner Mongolia and Shanxi before the epidemic ranked 9th and 10th at the bottom of the country, and their tourism attractiveness was low. The spread of the epidemic then magnified the tourism shortcomings of these two places. Except for Beijing, Shanghai and Guangzhou, the main source of tourists from other provinces in Hunan, Sichuan, and Chongqing is Hubei. Hubei had a serious epidemic, and Beijing, Shanghai and Guangzhou implemented strong control policies, so the loss of tourist sources directly caused the tourism industry in these provinces to be greatly hit in the early stage of the epidemic. The southeastern coastal region and the southwestern region have concentrated on China’s richest tourism resources. As major tourism provinces accounted for a larger share of the national tourism industry, the impact of the epidemic was even more pronounced at the beginning of the epidemic.

The second stage (the third quarter to the fourth quarter of 2020, shown in Fig. 6): Accelerated recovery. Most parts of the country had passed the peak of the first epidemic. As Beijing-Tianjin-Hebei, Hubei, etc. respectively lowered the response level of their public health emergencies from Level 1 to Level 2, and the Mid-Autumn National Day holiday approached, major tourism provinces promoted the obvious improvement of tourism resilience in the surrounding areas. In the third and fourth quarters of 2020, the average tourism resilience of each province (municipalities directly under the Central Government) was 0.5 and 0.6, respectively, which was a substantial increase compared to the first half of the year. With the improvement of the epidemic situation, various industries gradually resumed work and production. In the third quarter of 2020, economic growth across the country turned from negative to positive, and tourism consumption demand increased accordingly. In addition to the National Day and Mid-Autumn Festival holiday, the national civil aviation transported about 1.67 million passengers on 1 October, which was basically the same as in the previous year. During the National Day and Mid-Autumn Festival, 637 million domestic tourists were received throughout the country, which strongly promoted the restoration of tourism resilience across the country. In this stage, Hebei Province, the southwest region represented by Sichuan and its neighboring Guizhou, Yunnan, etc., the Yangtze River Delta region represented by Jiangsu and Zhejiang, and its neighboring eastern provinces such as Shandong took the lead in tourism resilience recovery and upgrading to a higher level. Among them, Hebei resumed inter-provincial team tours in July, and its largest tourist sources from other provinces-Beijing and Tianjin- rapidly released tourism consumption demand, causing its resilience to increase rapidly. The average revenues of star hotels before the epidemic in Zhejiang, Jiangsu, Shandong, and Sichuan ranked 7, 8, 9, and 11, respectively, and their tourism resources are relatively rich. During the 11th Golden Week, Shanghai, Guangzhou and Beijing remained the most popular travel destinations, while Hangzhou, Chengdu, Chongqing, Shenzhen, Xi’an, Wuhan and Kunming ranked fourth to tenth. With the improvement of the epidemic situation, the major tourism provinces had more tourism resources to use and release, and a more complete tourism industry chain also enabled them to have a more significant recovery capacity. Moreover, the tourism consumption demand of the people in the above-mentioned areas was relatively strong, so they were important tourist sources. The demand for intra-provincial travel and inter-provincial peripheral travel in such areas had been released, and tourism resilience had been intensively rising in the regional space. During this period, sporadic areas such as Inner Mongolia and Sichuan occasionally experienced localized epidemics, but they were quickly contained. For example, there were imported cases from Russia in Inner Mongolia in November, and no new local cases in December; while the cases in Chengdu, Sichuan Province, were contained in about 10 days, and the medium-risk areas were cleared in about 20 days. Therefore, they did not have a major impact on the overall recovery of tourism resilience.

The third stage (the first quarter to the second quarter of 2021, shown in Fig. 7): Increasing with fluctuation. Compared with the first and second quarters of 2020, the overall level in this stage had improved significantly, but the differences in tourism resilience between provinces increased, and the characteristic tourism provinces performed well. Affected by the second wave of the winter epidemic and the tourism industry cycle, the overall resilience in the first quarter of 2021 decreased from 0.62 in the fourth quarter of 2020 to 0.59, but it was much higher than the 0.39 level of resilience under the impact of the first wave of the epidemic in the first quarter of 2020. Although local epidemics broke out in many provinces (municipalities directly under the Central Government) in January 2021, zero new additions in local areas were realized in early February, and all medium- and high-risk areas across the country were cleared, so the epidemic was effectively controlled. During the Spring Festival, about 80% of A-level tourist attractions nationwide were open normally, which further restored the tourism industry in most provinces and regions across the country. In particular, the tourism resilience of Gansu and Guizhou successively exceeded 1, showing a bit of “resilience improvement”. The reason for this improvement lies in two aspects. Firstly, the severity of the epidemic in these two provinces was at a relatively low level across the country, creating the possibility of short-distance “intra-provincial travel” during the epidemic, and offsetting the adverse impact of the epidemic on the tourism industry to a certain extent. Secondly, in the later period of the epidemic, the distinctive tourism resources of these two provinces once again became the favored choice of tourists. Among them, Gansu’s tourist growth rate after the resumption of inter-provincial tourism was close to 280%, ranking first in the country; while Guizhou successively introduced “Wine Tourism” “Integration” and other specialty tourism products, making it a more desirable travel destination for tourists. In addition, the resilience of Xinjiang, Shaanxi, Inner Mongolia and other northwestern provinces and regions exceeded 0.75 in the second quarter of 2021. In the post-epidemic era, the non-crowded and distinctive tourism resources make this type of area more attractive, and their degrees of tourism recovery rank among the top in the country. Similarly, Hainan’s tourism industry has performed well. Sanya, an important tourist city in the province, has played a significant and positive role in the competitive advantage of “Chinese Hawaii”. Coupled with the accelerated layout of duty-free shops and the pulling effect of the China International Consumer Goods Fair, Hainan has shown extremely strong resilience under the impact of the epidemic. However, the three provinces of Tibet, Qinghai, and Ningxia are at an extremely low level of the resilience index due to various reasons, such as their weak tourism foundation. By the end of the study period, most provinces had not recovered to their pre-epidemic levels, and the import of cases from overseas and local epidemic outbreaks had a negative impact on some areas.

Although the overall tourism resilience to the epidemic has improved significantly after one and a half years of recovery, only one province (Guizhou) has a resilience index exceeding 1 in the second quarter of 2021, and none of the other provinces have returned to their pre-epidemic levels. Heilongjiang, Jilin, Tianjin, Guangdong, Shanghai and other provinces were more prominently affected by overseas case imports and local outbreaks at this stage. For example, a local diagnosis in Heilongjiang Province occurred in December 2020, and an asymptomatic infection occurred in January 2021, and then the province entered an emergency state, which reduced its resilience in the first quarter. The super-spreading incident in Tonghua, Jilin in January 2021, coupled with the relatively inadequate northeastern regional tourism attractiveness, made recovery relatively difficult. Another case was detected in Ruili, Yunnan in March 2021, and then traffic control was implemented as “no entry, no exit”. This continued to relax until May Day. Normalized epidemic prevention management was implemented, and the resilience index increased in the second quarter. Guangdong has always been troubled by the import of cases from overseas. There was one case of overseas import in December 2020, the first British mutant virus in January 2021, the first Nigeria mutant virus imported from abroad in March, the Indian mutant strain in May, and as of the end of June, all four medium-risk regions across the country were in Guangdong (one in Foshan, two in Dongguan, and one in Shenzhen). As a hub area connecting home and abroad, the risk of overseas imports has largely affected the recovery of its tourism industry. Therefore, the resilience performance of the above-mentioned provinces at the end of the study period is not as good as those of some northwestern regions and characteristic tourist regions.

The lowest resilience index value of each province tends to be below 0.5, and the tourism industry is generally weak. The worst level of resilience performance under the epidemic reflects the resistance of the tourism industry in each province to the epidemic, and the higher the lowest value, the stronger the resistance of the tourism industry to the epidemic. The spatial pattern of the lowest value for the resilience index in each province during the study period is shown in Fig. 8a, and Tibet, Sichuan, Chongqing, Hunan, Ningxia, Shanxi, Beijing, and Inner Mongolia are the provinces with poor resistance. The lowest value of the resilience index is less than 0.3, so the epidemic had a greater impact on the local tourism industry. The minimum resilience index values of Xinjiang, Gansu, Henan, Shandong, Hebei, Tianjin, and Jiangxi are relatively high, reaching above 0.4. Therefore, the tourism industries in these provinces were relatively weakly affected by the epidemic, but they are still at a very low level. The lowest resilience indices of the provinces are not very different, and they tend to be below 0.5, indicating that under the impact of the epidemic, the vulnerability of the tourism industry is universal.

The recovery ability of the tourism industry varies greatly among provinces, with Gansu and Guizhou showing increased resilience, while the developed areas and areas surrounding Hubei province were subject to stricter epidemic prevention and control measures, and they have moderate resilience and recovery ability. The optimal level of resilience performance during the epidemic reflects the recovery ability of the tourism industry in the various provinces, that is, the higher the maximum resilience index, the stronger the recovery ability. The spatial pattern of the highest value of the resilience index in each province during the study period is shown in Fig. 8b. The resilience values of Gansu and Guizhou exceeded 1, showing a significant improvement in resilience; while Xinjiang, Sichuan, Shaanxi, Hebei, Shandong, Fujian, and Inner Mongolia also showed strong recovery ability, with their highest resilience index values reaching above 0.75; and the developed regions, Hubei and some of its surrounding areas and the border provinces performed at a mediocre level. According to calculations by Zhang et al. (2020), the four developed regions of Shanghai, Beijing, Jiangsu, and Guangdong ranked first in the country in terms of their epidemic prevention and control capabilities; while Henan, Hunan, Chongqing, Anhui, Shanxi and other neighboring areas in Hubei also had excellent epidemic prevention and control capabilities. This also means that more stringent control measures would present a multi-faceted barrier to travel within the province and cross-provincial travel from outside the province. However, the highest index values of Tibet, Ningxia, and Shanxi are less than 0.4, and the highest value of Tibet’s resilience index is only 0.26, showing that the effect of tourism recovery was minimal.