Investigating activity patterns of wild mammals is important for proper conservation and management (Bennie et al. 2014). Most mammalian species are active at night (nocturnal), but many others are diurnal, crepuscular, or cathemeral (Bennie et al. 2014). However, within species, activity patterns can vary depending on the animal's environment. Climatic zones partially determine the activity patterns of mammals (Ware et al. 2012; Bennie et al. 2014). For example, bats (Chiroptera) and rodents (Rodentia) in Venezuela (located in the tropics) are nocturnal, whereas bats and rodents on the Tibetan Plateau (located in the Alpine tundra zone) are diurnal due to low nighttime temperatures and extremely high nocturnal energy costs (Bennie et al. 2014). In addition, the presence of humans and other animals affects the activity patterns of mammals. Human presence can cause a strong sense of fear in wild animals, and they may adjust their activities to avoid human contact (Gaynor et al. 2018). Their impact is often minuscule in small and medium-sized mammals (Zapata-Ríos and Branch 2016; Oberosler et al. 2017), but substantial in large carnivores (Ordiz et al. 2017) because large carnivores require a larger area for survival and are more likely to encounter human activity than small animals (Gaynor et al. 2018). North American wolves (Canis lupus) are diurnal, whereas European wolves are mainly active during the night and twilight (Mech 1992; Ciucci et al. 1997; Theuerkauf et al. 2007). The persecution of large carnivores by humans has been going on for centuries in Europe, but more recently, more intensely and effectively in North America (Frank and Woodroffe 2001). It is possible that the history of persecution by humans has influenced the activity patterns of large carnivores. Additionally, regional differences in food resource availabilities affect the activity patterns of large carnivores (Klinka and Reimchen 2009). The predation patterns of large carnivores change in response to changes in prey catchability (Metz et al. 2012). Large carnivores of the same species may have different activity patterns in different regions depending on the ease with which prey can be captured.

Brown bears (Ursus arctos) are large carnivores that are widely distributed in the Northern Hemisphere (Penteriani and Melletti 2020). Brown bears exhibit various activity patterns in different areas. Some studies have reported the activity patterns of brown bears as nocturnal (Rauer and Gutleb 1997; Kaczensky et al. 2006), crepuscular (Ordiz et al. 2014, 2017), and diurnal (MacHutchon et al. 1998; Klinka and Reimchen 2002). These differences in activity patterns are suggested to be a result of human activities. In Sweden, brown bears were more active at night than at day in the northeast, where roads were more dense; however, in the northwest, where there were fewer roads, they were more active during the day than at night (Ordiz et al. 2014). Klinka and Reimchen (2002) showed that brown bears were predominantly diurnal in areas with low human presence in British Columbia; however, in areas with high human presence, bears shifted to nocturnal. It has been reported in Alaska that in situations where there is little aggressive behavior by humans, such as hunting, and abundant food is available, brown bears become accustomed to humans and more active during the daytime, even when human activity is high (Smith et al. 2005). Kawamura et al. (2022) showed monthly activity patterns of brown bears in the Shiretoko Peninsula, Hokkaido, Japan, by sex-age class using camera traps. The study compared activity patterns inside and outside the National Park (NP, hereafter), suggesting that daytime activity is higher inside the NP than outside, an area with less human-caused mortality and more harmless encounters with humans, and that adult brown bears increase their nocturnal behavior in September and October.

Additionally, social relationships among individuals in a population can influence the activity patterns of brown bears. In an intraspecific brown bear society, individuals may exhibit different activity patterns for each sex-age class due to the hierarchy among classes, sexually selected infanticide by adult males, and mating behavior during the mating season (Steyaert et al. 2013). Klinka and Reimchen (2002) explained that the diurnal feeding activity of females with cubs and subadult bears targeting spawning salmon was high primarily to avoid larger nocturnal adult males. Kaczensky et al. (2006) hypothesized that subadult brown bears are more aversive toward other socially dominant sex-aged groups (especially adult males) than toward humans because of their limited experience with negative encounters with humans, and therefore become more active during the day when other sex-aged groups are less active but humans are more active. Parres et al. (2020) demonstrated the seasonal activity patterns of brown bears in the Pyrenees Mountains for each sex-age class using camera traps. In that study, adult males were highly active during the night and twilight, whereas females with cubs and subadults were active during the daytime. These examples indicate that even among bears living in the same area, individuals of the more vulnerable class adjust their activity patterns and try to reduce encounters with dominant individuals.

In the Akan–Shiranuka region, located in eastern Hokkaido, brown bears are distributed throughout the forest area, and human–bear conflicts such as crop depredation occurs around the forest, and consequently, nuisance control kills are carried out from May to September every year (Sato et al. 2011, 2014). During the hunting season in autumn, the hunting of sika deer (Cervus nippon) and brown bear is permitted throughout the area except for some small wildlife refuges and special protection zones in the Akan–Mashu NP. Therefore, brown bears in this area are at risk of human-caused mortality throughout the year, especially during hunting season when hunting pressure in the forest increases. Although diel activity patterns in brown bears in Hokkaido have been reported in Shiretoko NP, which is designated as a World Natural Heritage site (Kawamura et al. 2022), the information is still insufficient to assess the effects of human activities because the previous study was conducted for a short duration and the survey area was mostly occupied by protected areas. Diel activity patterns in brown bears among different sex-age classes outside protected areas are expected to be affected by human activities and the intraspecific relationships, and vary depending on the season. We tested a hypothesis: the activity patterns reflect social relationships among sex-age classes of brown bears, with inferior classes avoiding dominant classes. The objective of this study was to clarify the seasonal changes in diurnal activity by sex-age class based on long-term observation data. For sex-age classes, we also examined the effects of social relationships by dividing adult females into two groups: females with cubs of the year and solitary females, from the perspective of risk of infanticide.

Fig. 1.

The Location of the Akan–Shiranuka area where the field survey was conducted. Black points indicate the location of camera traps. The map was drawn by QGIS ( http://www.qgis.org). The vegetation map layer was drawn based on the vegetation survey report of the 5th Basic Survey for Nature Conservation (Biodiversity Center of Japan, Ministry of the Environment,  https://www.biodic.go.jp/dload/mesh_vg.html).

Materials and methods

Overlap

The overlap coefficients were estimated using the overlap package (Meredith and Ridout 2018) to analyze the degree of temporal overlap between the age-sex classes. The overlapping coefficient (Δ) is an estimator that determines the area under the curve formed by taking the minimum of the two density functions at each time point and ranges from 0 (no overlap) to 1 (complete overlap) (Linkie and Ridout 2011). The overlapping coefficient was used to identify the avoidance level between the individuals, following a previous evaluation (Monterroso et al. 2014). The following ranges of avoidance were considered: (1) low avoidance (Δ ≥ 0.67), (2) medium avoidance (0.56 ≤ Δ ≤ 0.66), and (3) high avoidance (Δ ≤ 0.55). The daily and seasonal differences in time patterns between the coexisting pairs of sex-age-specific bear groups were evaluated using a multiple comparison Mardia–Watson–Wheeler test (MWW) for testing whether the 24-hour period circulation observations differ significantly or not (Batschelet 1981). All statistical analyses were conducted in R version 3.4.3 (R Core Team 2021).

Table 1.

Seasonal variation in the number of events where brown bears were captured by camera traps by sex-age class in the Akan–Shiranuka region, 2016–2018

Results

A total of 2479 events were recorded between 2016 and 2018. The sex-age class was determined in 2183 events, where 832, 322, 418, and 611 events showed the presence of AMs, SFs, FCs, and SAs, respectively (Table 1). In spring, early summer, late summer, and autumn, 318 (14.5%), 578 (26.4%), 714 (32.7%), and 573 (26.2%) events were recorded, respectively. Of all events, 231 events in which body length could not be measured and 65 events in which body length ranged from 1.3 to 1.6 m but external reproductive organs could not be identified were excluded from the analysis as individuals of unknown sex-age class (13.5% of total). The number of events for the AMs was high throughout the year (832 events; 38.1%), and the number of events tended to be especially high in early summer (266 out of 832 events; 31.9%), which is the mating season (from May to July: Tsubota et al. 1985; Ishikawa et al. 2003). SFs and SAs had the highest number of events in late summer (130 out of 322 events, 40.3% and 234 out of 611 events, 38.3%, respectively) and FCs had the highest number of events in autumn (175 out of 418 events; 41.8%) (Table 1).

Fig. 2.

The seasonal variation in the activity patterns and activity peak analysis of adult male brown bears during a) spring, b) early summer, c) late summer, and d) autumn. Dark filled area indicates the core activity range, light filled area indicates the activity range, and the diagonally striped area indicates twilight. AR, activity range; CAR, core activity range; 95% CI, 95% confidence interval of AR.

Activity patterns

The twilight period is from 03:30 to 05:30 and from 17:30 to 19:30 in spring, from 03:00 to 05:00 and from 18:12 to 19:12 in early summer, from 03:54 to 05:54 and from 17:06 to 19:06 in late summer, and from 05:06 to 07:06 and from 15:30 to 17:30 in autumn. Brown bears of all sex-age classes in the study area generally tended to be nocturnal and were active during crepuscular or vesperal periods, except for the SAs (Figs. 2–5).

The estimated daily activity range (95% kernel) of the AMs was 21.15 h in spring and lasted roughly from 13:18 to 10:27 (Fig. 2a). The core activity range (50% kernel) was 7.06 h from 16:41 to 23:45. Early summer showed two peaks, from 3:30 to 6:47 and from 16:40 to 21:30 (activity range: 21.19 h), with the highest combined core activity range of 8.12 h (Fig. 2b). Late summer showed the same trend as spring but with a decrease in activity range and core activity range (activity range: 17.88 h; core activity range: 5.75 h) (Fig. 2c). The activity range was the lowest in autumn, with an estimated activity range of 14.81 h from 15:04 to 5:53 (core activity range: 6.35 h) caused by a considerable decrease in daytime activity (Fig. 2d).

The estimated daily activity range (95% kernel) of the SFs was 20.53 h in spring and lasted roughly from 2:30 to 9:10 and from 11:19 to 1:09 (Fig. 3a). The core activity range (50% kernel) was 7.16 h from 5:28 to 5:44 and from 16:20 to 23:14. In early summer, the estimated daily activity range was 19.83 h from 2:33 to 10:13 and from 10:36 to 22:46 (core activity range: 7.25 h, Fig. 3b). In late summer, the estimate showed the highest value of 21.72 h, from 12:16 to 9:59 (Fig. 3c). The core activity range (50% kernel) was 6.72 h from 15:22 to 22:06. The activity range was the lowest in autumn, with an estimated activity range of 18.26 h from 13:45 to 6:00 and from 7:45 to 9:43 (core activity range: 5.80 h) caused by a considerable decrease in daytime activity (Fig. 3d).

Compared to the AMs and SFs, the FCs were less active in spring and early summer (activity range: spring: 19.73 h and early summer: 19.77 h; core activity range spring: 5.55 h and early summer: 7.17 h) and inactive at night (Fig. 4a and b). In autumn, the estimated daily activity range was 19.94 h from 12:08 to 8:04. The estimated daily activity range in autumn was the same as that in other seasons, whereas the diurnal activity decreased, indicating that the FCs shifted their activity to nocturnal as the AMs and SFs (Fig. 4d).

Fig. 3.

The seasonal variation in the activity patterns and activity peak analysis of solitary adult female brown bears during a) spring, b) early summer, c) late summer, and d) autumn. Dark filled area indicates the core activity range, light filled area indicates the activity range, and the diagonally striped area indicates twilight. AR, activity range; CAR, core activity range; 95% CI, 95% confidence interval of AR.

SAs showed diurnal activity patterns from dawn to dusk from spring to late summer (spring: 18.42 h, early summer: 19.58 h, and late summer: 20.76 h) (Fig. 5a–c). They showed the highest estimated activity range of 22.70 h in autumn, with reduced diurnal activity and a slight shift to nocturnal activity (Fig. 5d).

Overlap

According to the results of the MMW test, the activity patterns of AMs were significantly different from those of all sex-age classes in all seasons, except for SFs in spring and early summer (Tables 2 and 3). There were no significant differences in activity patterns between SFs and FCs in any season. The SFs had low avoidance for AMs (Δ = 0.82) in spring, and MWW test showed that differences in activity range between SFs and AMs were not significant (W = 0.37 n = 227, and P = 0.83). In early summer, they showed high avoidance (Δ = 0.47) and MWW test showed that differences in activity range between SFs and AMs were not significant (W = 4.98, n = 333, and P = 0.08). In late summer, they showed medium avoidance (Δ = 0.56), which was significantly different (W = 11.71, n = 333, and P = 2.00 × 10^–3), and in autumn, they showed high avoidance (Δ = 0.48), which was significantly different (W = 8.90, n = 260, and P = 0.01).

The FCs highly avoided the AMs in spring, early summer, and late summer, with significant differences in activity range (spring: Δ = 0.51, W = 10.76, n = 245, and P = 5.00 × 10^–3; early summer: Δ = 0.47, W = 8.60, n = 315, and P = 0.01; late summer: Δ = 0.51, W = 23.32, n = 350, and P = 8.65 × 10^–6). In autumn, they showed medium avoidance toward AMs and their activity range were significantly different (Δ = 0.64, W = 11.18, n = 340, and P = 4.00 × 10^–3).

The activity range of SAs was significantly different from that of AMs in all seasons (spring: Δ = 0.57, W = 23.46, n = 242, and P = 08.05 × 10^–6; early summer: Δ = 0.68, W = 25.49, n = 462, and P = 2.92 × 10^–6; late summer: Δ = 0.51, W = 98.965, n = 437, and P = 2.20 × 10^–16, autumn: Δ = 0.54, W = 35.44, n = 350, and P = 2.02 × 10^–8).

Fig. 4.

The seasonal variation in the activity patterns and activity peak analysis of females with cubs during a) spring, b) early summer, c) late summer, and d) autumn. Dark filled area indicates the core activity range, light filled area indicates the activity range, and the diagonally striped area indicates twilight. AR, activity range; CAR, core activity range; 95% CI, 95% confidence interval of AR.

SAs and SFs showed low avoidance in spring (Δ = 0.53) and avoidance or medium avoidance in the other seasons (early summer: Δ = 0.45, late summer: Δ = 0.78, and autumn: Δ = 0.57). Significant differences in activity range were shown in all seasons except early summer (spring: W = 7.50, n = 73, and P = 0.03; early summer: W = 5.04, n = 263, and P = 0.15; late summer: W = 28.620 n = 364, and P = 2.27 × 10^–6, fall: W = 9.91, n = 233, and P = 2.82 × 10^–6).

SAs and FCs showed medium avoidance in spring (Δ = 0.56) and low avoidance in the other seasons (early summer: Δ = 0.72, late summer: Δ = 0.71, and autumn: Δ = 0.81). Significant differences in activity range were shown in all seasons except early summer (spring: W = 7.25, n = 91, and P = 0.04; early summer: W = 0.83, n = 245, and P = 0.18; late summer: W = 19.79, n = 381, and P = 2.67 × 10^–9, fall: W = 14.67, n = 312, and P = 2.54 × 10^–7).

Discussion

Brown bears of all sex-age classes, except the SAs, in the Akan–Shiranuka region were mainly crepuscular or vesperal. However, the core activity period differed by sex-age class and season. AMs and SFs showed similar activity patterns of crepuscular or vesperal throughout the seasons, although SFs were slightly more active during the day (Figs. 2 and 3). Especially in the mating season of spring and early summer, the differences in activity patterns between AMs and SFs were insignificant and the overlap coefficients were high, indicating that they selected and used the same time of the day (Table 2). Proximity in the activity patterns of AMs and SFs would increase their encounter frequencies and consequently, increase their reproductive opportunities (Sahlén et al. 2015). Additionally, FCs showed an activity peak that occurs at approximately the same time as AMs and SFs; however, in spring and early summer, FCs were more active during the day compared to the AMs and SFs. Furthermore, the overlap test results showed that FCs highly avoided AMs, indicating a different activity pattern than the other two classes (SFs and SAs) (Table 2, Fig. 4).

Since SFs did not avoid the core activity period of AMs during mating season, it was possible that they were more likely to encounter AMs for mating during the mating season (Table 2). On the other hand, FCs were active at a different time of the day from AMs during the mating season, probably to avoid infanticide. These results are consistent with those reported in previous studies (Ordiz et al. 2012; Steyaert et al. 2013; Parres et al. 2020). FCs have been suggested to temporally avoid AMs to reduce the risk of sexually selected infanticide by the AMs.

Fig. 5.

The seasonal variation in the activity patterns and activity peak analysis of subadult brown bears during a) spring, b) early summer, c) late summer, and d) autumn. Dark filled area indicates the core activity range, light filled area indicates the activity range, and the diagonally striped area indicates twilight. AR, activity range; CAR, core activity range; 95% CI, 95% confidence interval of AR.

Table 2.

Number of independent detections (n), statistical value of Mardia–Watson–Wheeler test (W), and coefficient of overlap between sex-age classes (Δ) among the activity patterns of coexistent groups of sex-age class of brown bears in the Akan–Shiranuka region, 2016–2018

Table 3.

Visual summary of Table 2. Significant differences in each seasonal (Sp: spring, Es: early summer, Ls: late summer, and Au: autumn) by sex-age classes (AMs: Adult males, SFs: Solitary females, FCs: Females with cubs, and SAs: Subadults) are indicated by the size of the symbol (n.s.: no significant difference, medium: P < 0.05, significant, large: P < 0.01, highly significant), and avoidance is indicated by the color and shape (low avoidance: open square, medium avoidance: light filled triangle, high avoidance: dark filled circle)

Compared with the other sex-age classes, SAs showed a different diurnal activity pattern with fewer night activities and more daytime activities (Fig. 5). American black bears U. americanus exhibit extremely different activity patterns from sympatric brown bears because smaller black bears tend to avoid encounters with larger brown bears (Schwartz et al. 2010). In other study areas, SAs of brown bears are more active during the day to avoid aggressive or competitive AMs (Kaczensky et al. 2006). Additionally, SAs are less vigilant about humans and tend to be less susceptible to human presence (Aumiller and Matt 1994). SAs are often disproportionately found in areas close to human activity and development (MacHutchon et al. 1998; Elfström et al. 2012). In North America SAs are more active during the day and frequently tend to expose themselves to humans than adult bears in the process of dispersal, for example (MacHutchon et al. 1998). These results seem to be due to the intra- or inter-specific competition and the naïve behavior (reflecting lack of experience) of SAs. Thus, SAs, which are typically vulnerable individuals, tend to avoid the risk of encountering AMs, which are dominant, and may shift to being active during daytime. SA bears were still active during the daytime in autumn when all sex-age classes shifted to nocturnal, although the proportion of activity was smaller than that in other seasons possibly because they selected instances when human activity was high to avoid socially dominant AMs. Since SAs are diurnal, they are at an increased risk of human encounters. However, more diurnal activity pattern may increase the opportunities for SAs to acquire foraging resources. Prey species exposed to predation or hunting would prioritize avoiding humans when predator ratios are low (Eriksen et al. 2011). The high density of AMs in our study area (36% of the total), as observed from the capture rates using camera traps, likely caused FCs, and especially SAs, to prioritize avoiding AMs, leading to increased daytime activities.

The shift to the nocturnal pattern in autumn from the crepuscular or diurnal pattern in late summer was confirmed in this study for all sex-age classes (Figs. 2–5). Similar changes in activity patterns during the hunting season were reported for Scandinavian brown bears (Ordiz et al. 2012; Frank et al. 2017). Additionally, regulations that limit hunting to daytime may cause brown bears to be more nocturnal (Frank et al. 2017). Kawamura et al. (2022) compared the activity patterns of brown bears outside the NP, where human-caused mortality occurs, and inside the NP, where the occurrence is rare. Brown bears are more active during the daytime in the NP, suggesting that vigilance toward humans influences their activity patterns. In our study area, the hunting season begins in October. Hunting is permitted during the daytime, and the main target for hunters is sika deer. Therefore, bears possibly change their activity patterns to being nocturnal in autumn so that they can avoid the risk of encounters with hunters during the day in the hunting season.

Autumn is also a period of hyperphagia when bears need to acquire more foraging resources (Friebe et al. 2001). Scandinavian brown bears increase daytime activity to forage for berries in summer and early autumn (Ordiz et al. 2017). Foraging efficiency is the highest during daylight hours when color and detail discrimination at close range is possible (Larivière et al. 1994). Brown bears in the study area also need to increase their foraging to accumulate nutrients before hibernation, indicating that their being active during the day and night would be beneficial in autumn. However, human influence may have caused them to shift their foraging activity to nighttime. Particularly, AMs are significantly less active during the daytime in autumn, and foraging opportunities may thus be reduced. A comparison of the diurnal activity patterns of SFs and FCs during the hunting season shows that the FCs were more active during the day than SFs (SFs: from 13:45 to 6:00 total 18.26 h; FCs: from 12:08 to 8:04 total 19.94 h) (Figs. 3 and 4). As mentioned in the previous paragraph, brown bears in Scandinavia also become nocturnal during hunting season as in our study area, but the change in activity pattern in FCs is smaller than that in SFs, possibly because FCs were legally protected from hunting and had to meet the increased energy requirements to care for their cubs. Although daytime activity declines in the autumn likely due to hunting pressure in our study area where FCs were not protected from hunting, FCs are more likely to be active during the day than SFs because of maternal care.

Recently, brown bears have been appearing in urban areas during the daytime, and forest encounters have been increasing in some areas of Hokkaido (Sato 2017; Hokkaido Prefectural Government 2022). Species with behavioral plasticity are more likely to adapt to human presence and may be more resistant to extinction (Beckmann and Berger 2003). One of the most important prerequisites for brown bears and humans coexisting is the maintenance of the nocturnal behavior of brown bears near human-dominated landscapes (Kaczensky et al. 2006). Although brown bears are broadly viewed as a species that varies the time and location of their activities while responding well to human activity, a closer analysis by sex-age class suggests that the socially inferior classes of brown bears, such as FCs and SAs, tend to be more active during times when humans are active.

Here, we identified the seasonal behavioral patterns of each sex-age class of brown bears. This study showed that FCs and SAs were more active during times when humans were more active. For humans to coexist with large wildlife, it is essential that they should be segregated geographically and temporally. This study would help provide clues to the causes of brown bears' urban appearance and increased conflicts with humans (Sato 2017; Hokkaido Prefectural Government 2022), thereby effectively protecting and managing brown bears. Activity patterns of all sex-age classes of brown bears in the Shiretoko NP, which is also located in the eastern part of Hokkaido, were more active during the daytime than in our study area (Kawamura et al. 2022). This difference is thought to be related to the fact that brown bears are less wary of humans in Shiretoko NP, where the use of human-derived resources such as crops is not observed, and the frequency of harmless encounters with humans is higher (Sato et al. 2004; Shirane et al. 2021; Kawamura et al. 2022). The diel activity patterns shown in our study might be applicable to brown bears outside of protected areas over a wide area of Hokkaido. Future research on behavioral patterns in various areas will provide a deeper understanding of the factors that influence the behavioral patterns of brown bears and will allow for the prediction of behavioral patterns based on local current conditions, which will be useful for brown bear conservation management.