There are currently two national nature reserves in Xishuangbanna Dai Autonomous Prefecture, namely the Xishuangbanna National Nature Reserve and the Napan River Basin National Nature Reserve. Among them, Xishuangbanna National Nature Reserve consists of five sub-reserves, namely Mengyang, Menglun, Shangyong, Mengla and Mangao, which are geographically close to each other but not connected, with geographical coordinates from 21°10’N to 22°24’N and 100°16′E to 101°50′E, with a total area of 242,510 hm², accounting for 12.9% of the state’s area. In Xishuangbanna, Asian elephants are mainly distributed in and around the three sub-protected areas of Mengyang, Mengla and Shangyong in Xishuangbanna National Nature Reserve. Although there is Asian elephant habitat in the Napanhe Reserve, no Asian elephants are currently active in the area, and because it is separated from the Mengyang Reserve by the Lancang River, elephant herds have not yet entered the area (Lin et al., 2015). Therefore, three sub-protected areas, Mengyang, Mengla and Shangyong, were selected as the study area to investigate the human activities around them (Figure 1).

Based on the conclusion that the range of Asian elephant activities may be beyond the existing PA, this paper proposes to establish a 10 km buffer zone for the protected area (PAB) to carry out the study. Rowell (2014) showed that wild Asian and African elephants are estimated to move an average of 5–10 km per day, and the elephants they examined walked an average of 9.05 ± 0.61 km per day. Holdgate et al. (2016) used foot rings equipped with GPS data loggers to measure the daily outdoor walking distance of African and Asian elephants kept in zoos, and concluded that the elephants walked an average of 5.3 km per day conclusion. Zhou and Zeng (2022) monitored Asian elephants entering Menglun Nature Reserve and concluded that Asian elephants had a daily activity radius of 2–15 km. We finally chose a buffer zone of 10 km for the study, and some experts (Li, 2021) also pointed out that the average habitat extent of Asian elephants in natural forests is 10 km².

This study mainly used the annual mean products of the NTL dataset from 2012–2021 to characterize human activity space, and the global land cover dataset released by J. Yang and Huang in 2012 and 2021 (Yang and Huang, 2021) to explore the present-day Asian elephant habitat status. In addition, this paper also used the boundary data of China Nature Reserves in June 2020 from the World Database on Protected Areas (WDPA).¹

Human activities around Asian elephant habitats have been increasing in the last decade. Despite some interannual fluctuations, there is a clear trend of increasing SNL (Figure 2A) and ANL (Figure 2B) from 2012 to 2021 both within the PA and PAB. SNL represents the variation of total night-time light in the area. In Figure 2A, even though the SNL of the PA is far lower than PAB, the SNL of the PA would increase when the SNL of PAB increases, indicating that the increase and decrease changes of PA and PAB are consistent, e.g., the SNL in both areas are increasing from 2012 to 2014, and the SNL increased first and then decreased around 2016. However, there are also inconsistencies. In 2020, the SNL of PAB decreased compared with the previous year, while the PA increased. ANL indicates the percentage of bright pixels appearing in the region, and the changes of PA and PAB in Figure 2B are more consistent, and both have a significant increasing trend (R²>0.82). When the bright pixels of the PAB decreases, it is also decreasing in the PA, such as in 2015, 2018, and 2020.

Consistency of NTL changes inside and outside the PA is increasing, the average and maximum NTL (ANLv and MNL) in PA and PAB are converging. The interannual fluctuations of PAB’s ANLv are large and have been slowly decreasing in recent years. Unexpectedly, overall ANLv within PA is trending upward and two lines representing different regions have a tendency to approach during Figure 2C. From the regional maximum value (Figure 2D), the MNL of the PAB region also shows a decreasing trend year by year 27.92 in 2021 is much lower than 45.79 in 2012, a decrease of 39.03%. However, the MNL within PA has climbed rapidly since 2019 and gradually approaching to the MNL of PAB. Even though Figures 2C,D show a decreasing trend of both the mean and maximum values in the PAB, the SNL in Figure 2A still shows an increasing trend, indicating that the number of illuminated pixels within the PAB has increased and the range of human activities is expanding. Figure 2B also shows such a feature.

The level of human activity in PA is greater than that in PAB. During Figure 2E, the INL of both PA and PAB fluctuate considerably, with positive and negative values, indicating that NTL has been unstable over the last decade. Nevertheless, the overall values of INL in the PA are larger than those in the PAB, indicating that the increasing trend of human activities in PA is more obvious than PAB.

In a word, the NTL characteristics of Xishuangbanna Reserve and its surrounding areas show an obvious trend of growth in both value and range, and individual areas such as MNL and ANLv are approaching, indicating that the boundary effect of the reserve may be weakening due to human activities. The SNL, ANL, MNL, and ANLv of PA had an increasing trend, and the SNL of PAB is increasing, ANLv decreasing but the percentage of SNL expands, indicating that the area and range of activities are expanding. And the INL changes because the interannual fluctuation of SNL is also influenced by the number and value of pixels.

From the spatial pattern of Xishuangbanna, the brightest part of the NTL in its middle and close to the PA, and the NTL in 2021 is brighter and broader than in 2012. In 2012, NTL is mainly concentrated in the central part, and some of the southern part of the PA had faint lights (Figure 3A). In 2021, more areas around the PA were lit up, the original lighted area in the south was brighter, and even appeared in the interior and west side (Figure 3B). In addition, in 2012 and 2021, the NTL around PA located in the central region were the brightest, and PA in the western and southern regions will also be illuminated.

The NTLs in the PAB area show a central clustering trend. Figures 3C,D are the SNL and MNL of NTL around the PA from 2012 to 2021 based on pixel scale statistics. Their spatial distribution is relatively similar. With faint light sources around the three PA, and some of them even inside the PA, such as Mengyang. The SNL and MNL of NTL in the Mengyang are relatively large compared to the other, and there are clusters of NTL within PAB in both the Mengyang and Mengla areas. The southern part of PA has NTL and aggregation phenomenon, especially the NTL of central PA aggregation phenomenon is more obvious, showing an obvious bicentric trend, and the western PA shows a weak bicentric trend.

The scope of human activities outside the PA is gradually expanding, and the impact on the PA is gradually enhancing. The statistics obtained 86.21% of the areas with increasing trend and 13.78% with decreasing trend of NTL in the PAB area from 2012 to 2021 (Figure 4A, p < 0.1). The average NTL per pixel of PAB increased by 1.24 in the last decade, which is equivalent to 2.23 times of the regional mean NTL of 0.57 in PA in 2012. Calculation of NTL trend changes in the PAB over the past 10 years found that the NTL gathering center in the central PAB has a weakening trend, but the NTL around the center has an increasing trend, indicating that human activities have a trend of spreading to the periphery (Mengyang). The NTL aggregation center in Mengla has a higher increasing trend, but the NTL of the part intersected with PA has a decreasing trend, and the southern PA is currently at some distance from the NTL.

Human activities in PA and PAB have a tendency to move southeastward and gradually expand in extent. The long semi-axis of the SDE indicates the direction of the center of gravity, which points to the southeast in 2012, 2015, 2018 and 2021 (Figure 4B). The center of the ellipse shifted along the southeast for two consecutive periods, and the period during 2018–2021 was more southerly compared with the period before 2018. It can be inferred that the center of gravity of the NTL distribution in the PAB shifted roughly to the southeast during 2012–2021, indicating a trend of human activities to the southeast.

The areas with greater NTL brightness in Xishuangbanna appear in the southern part of the PA and inside the PAB, and most of the NTL showed an increasing trend. From the SDE, it is clear that there is a tendency for human activities to move in a southeasterly direction, and basing on conclusion of the northward movement of Asian elephants, it is likely that this movement is a “purposeful” trip, which shows that PA is increasingly affected by human activities (Jiang et al., 2023). In the future, monitoring and planning of habitats based on Asian elephant activities should be necessary.

The main LC type in PA and its surrounding areas is predominantly forest, and reduced forest have been converted to cropland and building in the last decade. Figures 5A,B are the LC classification maps of PAB in 2012 and 2021, respectively. The cropland in PA has increased significantly, and there are a large number of constructions in the south, and the scope has expanded, indicating that human activities are close to PA. Figure 5C shows the statistical analysis of LC in 2012 and 2021, where the forest of PA and PAB decreased by 1.61 and 2.30%, respectively, in 10 years, in which 99.37% of the decreased vegetation in PA area converted into cropland, and 94.78 and 5.22% of the decreased vegetation in PAB converted into farmland and construction, respectively.

In the past decade, all LC types showed more increasing than decreasing trends in the areas with significant NTL change trends. Nearly 50% of the forest showed an increasing trend of NTL and nearly 10% showed a decreasing trend. Combining with the areas in Figure 4A where NTL showed different trends, we counted the NTL trends of different LC under the same location and plotted Figure 5D. The areas where NTL has a significant trend of change, forest and cropland more dominant, and the areas where forest shows different trend are decreasing and farmland is increasing. The increasing trend of impervious from 2012 to 2021 is not obvious, but the increasing trend presented by NTL is significant, from 0.43 to 2.56%.

The tree cover in the PAB area was higher than 90%, but nearly half of the vegetation cover showed a significant trend of increasing NTL, indicating that the Asian elephant habitat may still have a great risk of human disturbance in the future. However, the data of NPP—VIIRS started from 2012, it is impossible to explore the current status of NTL in the study area by using the same data, which is not conducive to the in-depth analysis of long-term change trajectories. And there are few existing LC products with high spatial resolution, and if multiple datasets are used, some information may be lost due to the uniformity of scale, so if it can be supplemented with more high-precision LC data, the understanding of Asian elephant habitat status in Xishuangbanna region will be more comprehensive.