In this study, we used precise tracking data to elucidate the migration patterns of the eastern great bustard and found clear sex-dependent differences. Males started migration later but arrived earlier than females in spring; males migrated for half as long a distance as females and spent one-third times of females in transit. Additionally, we found that the eastern great bustard had high site fidelity outside of migration season but that their habitats are not well protected.
We found male great bustard arrived at breeding sites earlier, which may be explained by the higher intrasexual competition for a mating that occurs among males. Generally, males end their migration earlier than females to acquire better breeding territories or maximize their mating opportunities [4, 10, 16, 26, 43, 44], as also known as the arrival-time hypothesis . This mechanism has been convincingly shown to be at play for O. t. tarda, the western subspecies [2, 26]. We also found that males left wintering sites later than female. To our knowledge, this phenomenon hasn’t been reported in other protandrous species before, because males generally initiate migration earlier or at the same time as females [4, 10, 43]. We speculate that this phenomenon occurs because of the shorter migration distance of the males, and the long stopovers time of the females during the spring migration. Considering the sex-based differences in migration distance, the early departure of the females may be to compensate for the longer transit.
Great bustards showed sex-dependence in their migration distances: males migrated nearly 1000 km per route, which is about half the distance of females. Their wintering site was latitudinally separated by over 6° (Fig. 1), with the males wintering farther north. This pattern is also found in other species such as Dark-Eyed Juncos (Junco Hyemalis) , Yellow-bellied Sapsucker (Sphyrapicus varius) and other birds . Males wintered closer to their breeding sites, which contributed to bustard’s ability to return to the breeding site in a shorter amount of time, no doubt helping them cope with the intrasexual competition . Besides, high sexual size dimorphism influences their tolerance to environmental conditions, such as cold weather. Ketterson (1976) found that larger birds can remain in harsher environments closer to the breeding grounds, whereas smaller individuals are constrained by winter chills and need to migrate farther. In our study, male O. t. dybowskii weighed 12.24 ± 3.14 kg on average, which is approximately 2.9 times that of the female (Table 1). The maximum weight we recorded was 17.8 kg, which is close to the maximum weight ever recorded for European great bustards [21, 46, 47].
Further, males may winter in Mongolia for some years, depending on weather conditions and snow cover. For example, Mongolian researchers have observed four over-wintering male great bustards near the croplands in northern central Mongolia, on January 15th, 2005 (Batsaikhan, N. personal communication) and 17 great bustards in Western Mongolia, on January 31th, 2004 (Batdemberel, D. personal communication). Natsagdorj  also notes great bustards as being present in Mongolia in winter. These observations likely underscore the physiological limitations of body size and the reason for the different migration distances of the two sexes.
The migration distance of the great bustard also varies in subspecies. We found males migrated over 10 times longer than males in Spain . The male and female great bustards in this study moved up to 540 km and 594 km in one day, respectively, which was longer than the maximum distance per day reported for the western subspecies recorded in Russia (maximum 325 km per day) . Kessler et al.  considered that migratory distances increase longitudinally from west to east across the range of this species because of the harsh environmental conditions in Asia. Our results were in line with this suspicion.
Besides, we found females spent more days in spring than autumn during migration, which may be caused by the long stopover duration in spring . Temperature and wind support mainly trigger the migration movement of the great bustard [50, 51]. In northern China and Mongolia, weather condition remains harsh and unstable in March. Thus, females may wait at stopovers until the environmental conditions become more comfortable [50, 52]. Further, female great bustards incubate eggs and raise chicks singly in summer , so a long stopover duration provided enough time for energy replenishment for egg production in the breeding season .
We also found great bustards exhibited occasional nocturnal movement during migration. Great bustard usually roosts when darkness falls, but occasionally conducts nocturnal activity in spring or winter . However, few studies described the nocturnal migratory behaviours of great bustards to our knowledge . Diurnal species have flexibility in circadian timing strategies during migration  when they are crossing the inhospitable terrain or with comfortable atmospheric conditions . Besides, feeding is the most time-consuming activity for great bustards . Thus, another possibility for the nocturnal migration is that migratory flights at night do not interfere with foraging during the days [57, 58].
Habitat fidelity and distribution
We found that all four individuals showed strong site fidelity (over 50%) during both wintering and breeding seasons. Understandably, both females and males gathered in the breeding season, given the breeding displays and mating behaviour of the species. Males first aggregate, showing ground-displaying behaviour in a high visibility area to attract females , and then form dispersed leks . This high lek site fidelity was also found in the western subspecies [60, 61].
Our tracking data revealed that eastern great bustards have three different wintering sites including Xinxiang, Henan, Xilingol, Inner Mongolia and a previously unknown wintering site in Datong, Shanxi; two breeding sites in Bayandun and Dashbalbar in Dornod province; and one newly found post-breeding sites in Choibalsan Soum in Dornod Province (Fig. 1 and Fig. 5). The stopover sites during the migration for each individual were random and scattered, similar to what has been observed for the population in central Mongolia . Newly migration stopover sites were found in Choibalsan and Bulgan Soums in Dornod Province in eastern Mongolia, and Tianjin, Zhangjiakou, Hebei and Xilingol, in Inner Mongolia in China. Given that this was determined from tracking just six individuals, we strongly suspect that the distribution of eastern great bustards may be more widespread than was previously believed and that they are diffusely scattered throughout the east of Asia.
Existing reserve systems are a cornerstone of efforts to protect biodiversity . Our results have revealed that only one breeding site and one wintering site are inside a protected area and that there is generally a low overlap between their home ranges and PAs, with large numbers of stopover habitats occurring far outside existing nature reserves (Fig. 6). During winter and migration season, great bustards spend less than 5% of their overall time in the protected areas. Half of the great bustards we tracked died in their wintering sites or during migration, indicating the protection of the great bustard's habitat is inadequate and this subspecies is at risk of disappearing. In addition, it may be difficult for this species to adapt to a changing environment given its high site fidelity More research is urgently required to assess habitat quality for great bustards and develop efficient plans to provide suitable habitat for the species.
Great bustards are threatened by predators and sensitive to human disturbance , suffering rapid population reductions across most of their range owing to hunting, powerline collisions or degradation and fragmentation of their habitat . The deaths we observed demonstrate the high mortality risk faced by this species. Bustard NO.1, the only female, died during the autumn migration on November 27, 2020 in Xilingol, Inner Mongolia because of a collision with an agricultural fence. Individual NO.5 died on March 3, 2019, in his wintering site because of a collision with agricultural sprinkler irrigation equipment. Bustard NO.6 died at midnight on March 29, 2020 in Xilingol due to a collision with a power line. This was the largest individual in our study (NO.6, 17.80 kg; see Additional file 1: Figure S1) and since dying, no young male has yet taken over his lek site. The great bustard is thought to be a “poor” flier and has experienced record numbers of powerline collision casualties in recent years . We suggested that conspicuous marking should be considered where birds are highly concentrated to reduce those collisions [64, 65], and international cooperation is also needed to protect this endangered species effectively.
Some constraints exist for our study. First, we tracked six adult bustards with one female, which may limit our analysis. However, we found similar trends after adding migratory data of three females from Kessler et al. . Second, environmental factors  or refuelling patterns  may also affect animals’ migration rhythm. We cannot exclude the effect of the above factors on sex differences in these subspecies. Besides, the great bustard is distributed widely in Mongolia, and the migration routes vary among populations. However, we only tracked one of those populations. Thus, future studies will need to be conducted in more areas to completely reveal the migration patterns of great bustards and provide a more generalisable insight for comprehensive conservation.