Study area and animals
This study was conducted on the Shiretoko Peninsula (43°50′–44°20′ N, 144°45′–145°20′ E; Fig. 1), eastern Hokkaido, Japan. We conducted a monitoring survey of adult female bears inhabiting a region extending from the Horobetsu-Iwaobetsu area (44°5′ N, 145°1′ E; Fig. 1) to the Rusha area (44°11′ N, 145°11′ E), and divided the bears into three groups according to their home ranges and levels of human habituation. The home ranges were determined by the core areas revealed by radiotracking data for captured bears, and by the locations of DNA sampling (collection of hair, feces, and skin samples) for the remaining bears. The level of human habituation was determined based on the bears’ behavioral responses to humans, as follows: very high, no aversive reaction when approached by humans within 20 m; high, no aversive reaction to humans > 20 m away who move to within 20 m; moderate, movement away when approached within 50 m; and low, rare appearance before humans and rapid movement away upon awareness of humans. For bears inhabiting the area patrolled daily by park managers (i.e., the area excluding the Rusha area, as described below), this classification was done by two park managers with long experience with those bears from 1997 to 2018, based on patrol records that described the response of the bears when they approached to chase the bears away. We reviewed multiple randomly selected records for each bear and determined the habituation level based on the response most frequently observed. Alternatively, for bears inhabiting the Rusha area, outside the area patrolled for management purposes, two persons (one in common with the former case) determined the level, based on the response when they approached the bears multiple times during the long-term survey conducted between 2006 and 2018 (detailed in our previous reports31). All bears whose behavioral responses were not available, e.g., bears identified by genetic sampling that did not appear in front of humans, were categorized as low habituation level.
The first group comprised bears living around the front portion of the national park, including the Horobetsu-Iwaobetsu area, Shiretoko Five Lakes (44°7′ N, 145°5′ E), and the Utoro Plateau (44°3′ N, 144°59′ E), which is adjacent to the town of Utoro (44°4′ N, 144°59′ E; Fig. 1), a gateway community. Group 1 comprised 23 females, 14 of which were captured and released with ear tags for radiotracking in the Horobetsu-Iwaobetsu area between 1998 and 2013 and 9 of which were determined by genetic sampling to have been alive and present in this area from 1999 to 2018 [2 with distinctive characteristics (e.g., chest marking) were periodically identified in this area between 2013 and 2018]. This area is the main place where park visitors experience activities in nature. In the last two decades, bear sightings have been increasingly common in this area. When a bear appears near a road or development, park managers rush to the site and chase the bear away by shooting it with non-lethal rubber slugs and cracker shells. Thus, some bears are moderately habituated, but are not allowed to act freely during daylight hours. The second group comprised bears living between the Idashubetsu River (44°9′ N, 145°6′ E; Fig. 1) and the Rusha area, about 12.4–21.8 km from Utoro. Group 2 comprised 26 females, 4 of which were captured and released with ear tags for radiotracking in this area between 1993 and 2015 and 22 of which were determined by genetic sampling to have been alive and present in this area from 2000 to 2015 (13 were periodically identified in the area between 2013 and 2018). Kamuiwakka Falls (44°9′ N, 145°8′ E; Fig. 1) is the final tourist destination in this region; visitors are not allowed to pass through from Kamuiwakka Falls to the Rusha area. The third group comprised human-habituated bears living around the Rusha area, a special wildlife protection area with no public access without permission, and no human residence except for one fishermen’s settlement. Because the fishermen have not excluded bears from the settlement area in the last few decades, the bears have become habituated to the existence of humans and ignore people, which enables direct observation at close range (Fig. S1). Group 3 comprised 13 very highly habituated bears that had been visually identified and frequently observed in this area from the late 1990s to 201828,31,32,33. Among them, six bears were captured and released with ear tags for radiotracking between 2013 and 201831. Non- and less-habituated bears (low–high habituation levels) observed and sampled in the Rusha area were allocated to Group 2. All collared bears were captured live in accordance with the Guidelines for Animal Care and Use of Hokkaido University and all procedures were approved by the Animal Care and Use Committee of the Graduate School of Veterinary Medicine, Hokkaido University (Permit Number: 1152 and 15009, 17005, 18-0083). The protocols for capture received annual approval from the Ministry of the Environment, Japan, and the Hokkaido Government through research permit applications.
Sample collection
During 1998–2018, we collected genetic samples from bears throughout the peninsula, including the towns of Shari, Rausu, and Shibetsu (Fig. 1), using multiple methods. Most samples were from bears that were dead due to nuisance control or hunting, or that were captured for research purposes. Blood, hair, liver, and/or muscle samples were obtained. Hunting is limited from October to January and is not allowed inside the national park. When a bear is killed, all hunters are required to report the reason why it was killed (i.e., nuisance control or hunting), date, location, sex, estimated age (based on body size), body mass, and size (if available), and to provide the genetic samples described above. Due to the strong relationship between park managers and hunters in the peninsula, poaching or hunting without a report are very unlikely to have occurred over the past two decades. In the current analysis, nuisance killing and hunting were included as human-caused mortality because, in most situations, hunting occurs near human settlements, especially around agricultural land, to mitigate the human–bear conflict in autumn, which makes it reasonable to consider that the likelihood of mortality by hunting is also influenced by the bears’ behavioral traits, including habituation level. We also obtained: (1) hair samples collected from hair traps, including fence traps and tree-rub traps placed in several locations (including the Horobetsu-Iwaobetsu and Rusha areas) during 2010–2018; (2) skin tissues collected by biopsy dart sampling during 2011–2018; and (3) fecal samples collected during 2009–2018. The sampling methods have been detailed in our previous reports30,31,32,34. For bears captured or killed between 1998 and 2018, age was estimated by counting the dental cementum annuli35.
Human-caused mortality of the offspring of adult females
To determine how often offspring of adult females were killed due to human activities after separation from their mothers at 1.5–2.5 years of age31, the number of offspring (aged ≥ 1 year) produced by those females during the follow-up period must be known. However, we were unable to determine the exact number of offspring because the females, except for habituated bears in the Rusha area, were not monitored annually by direct observation. Therefore, the number of offspring produced by each female was estimated using the following formula: bear-years (follow-up years at ages ≥ 5 years) × reproductive rate (young born/year/reproductive adult female aged ≥ 5 years) × cub survival rate (from approximately 0.5 to 1.5 years of age). We assumed an equal sex ratio at birth. The latter two parameters were investigated in a long-term, individual-based monitoring survey in the Rusha area; values obtained were 0.74 and 0.63, respectively31. The bear-years were calculated from the year of first identification at ≥ 5 years of age (the earliest age of first reproduction in the Rusha area31) to the year of last identification. The initiation year was determined based on: (1) direct observation of a bear whose birth year was known, (2) dental age estimation at the time of capture or death, or (3) reproductive history (i.e., the age at first identification with cubs was assumed to be ≥ 5 years). The latter parameter was determined by not only direct observation of a female with cubs, but also indirect evidence of birth experience based on parentage analysis (described below). For example, if a female was assigned to the mother of 3-year-old bears (estimated by dental examination) killed in 2012, the mother was assumed to be ≥ 5 years old in 2009. The completion year was determined based on: (1) the year of last observation or death or (2) the reproductive history, revealed by parentage analysis. In the latter case, for example, if a female was assigned to the mother of 3-year-old bears in 2012, we assumed that the female survived at least until separation of the offspring (i.e., 2010), considering that a cub is not likely to survive its first year without the mother. Because genetic analysis of killed bears was initiated in 1998, the earliest year for the start of bear-year calculation was set at 1997, considering that offspring born in 1997 were potentially killed at 1 year of age in 1998. The latest year for bear-year calculation was set at 2015, as most human-caused mortality of offspring occurs at 1–3 years of age; thus, dead and alive bears born in 2015 can be evaluated in 2016–2018. The number of offspring (aged ≥ 1 year) killed by humans after separation from the mother was calculated by parentage analysis. Bears killed with their mothers were not included in the calculation. The age of offspring was estimated by counting the dental cementum annuli, for bears without birth records based on direct observation. Offspring sex was confirmed by genetic methods (described below) and external characteristics.
DNA extraction and genotyping
DNA was extracted using the DNeasy Blood & Tissue Mini Kit (Qiagen Inc., Tokyo, Japan) for blood and tissue samples, the DNA Extractor FM Kit (Wako, Osaka, Japan) or Isohair Easy (Nippon Gene, Inc., Tokyo, Japan) for hair samples, and the QIAamp DNA Stool Mini Kit (Qiagen Inc.) for feces samples, according to the manufacturers’ protocols. Twenty-one microsatellite markers and one sex marker, amelogenin, were analyzed by multiplex PCR assay under conditions described previously (primers are listed in31). Allele size was determined using an ABI PRISM 310 genetic analyzer (Life Technologies Japan Ltd.). In addition, a mitochondrial DNA haplotype analysis targeting the control region was conducted to select candidate mothers of each individual in the parentage analysis (for details, see30).
Parentage analysis
The parentage analysis was performed using a likelihood-based approach with CERVUS version 3.0.736. The simulation parameters were 10,000 cycles, 150 candidate mothers and fathers per offspring, 40% of candidate parents sampled, and 1% of loci mistyped32. In the first step of the CERVUS analysis, we assigned a parent pair. The confidence level was set at 80%, and no mismatching was allowed in a parents–offspring combination (i.e., mother–father–offspring trio). One mismatch was allowed in a parents–offspring combination obtained at a ≥ 95% confidence level when the same mother and father were selected as the most likely parents (≤ 1 mismatch per pair) in maternity and paternity assignment analyses, respectively. If a parent pair could not be assigned due to a low (< 80%) confidence level or the presence of ≥ 1 mismatching loci, we assigned maternity as a second step. The confidence level was set at 80%, and no mismatching was allowed in a mother–offspring combination.
Statistical analysis
Statistical analyses were performed with IBM SPSS Advanced Statistics ver. 23 (IBM Corp., Chicago, IL, USA). Due to several limitations in the dataset (including the limited sample size, wide variations of bear-years among the subject mothers, and biased geographical distribution of human habituated mothers), it was inappropriate to apply an individual-based approach, e.g., generalized liner model, to examine how multiple factors, including sex, proximity of the birthplace to areas of human activity, age, and human habituation of the mother, affect the likelihood of human-caused death of offspring. Instead, we applied group-based comparisons to test each hypothesis described in the Introduction. For all but the fourth hypothesis (age × birthplace hypothesis), we used Fisher’s exact test. For example, the first hypothesis (“sex-biased mortality hypothesis”) was tested by comparing the frequency of offspring mortality between males and females. The correspondences between each hypothesis and comparison pair are shown in Table 3. The method of Holm37 was used to adjust the P values obtained from multiple comparisons. In our analysis, the potential number of offspring produced by targeted mothers was estimated by bear-years, so the likelihood of human-caused death might be affected by this estimate, i.e., relatively few bear-years (small denominator) potentially increase the morality rate, especially in bears whose birth-years were determined by the genetic identification of the offspring, and not by direct observation or dental examinations. Therefore, the number of bear-years was compared among groups by one-way analysis of variance (ANOVA), followed by the Tukey–Kramer test to examine these possibilities. For the fourth hypothesis (age × birthplace hypothesis), group differences in ages of death for the offspring were analyzed by one-way ANOVA, followed by the Tukey–Kramer test. The homogeneity of variances in the data for bear-years and for ages of death was verified with Levene’s test. All comparisons were considered significant at P < 0.05. All values are presented as means ± standard deviations.
Source: Ecology - nature.com
