Seasonal recreational closures based on bear foods

Habitat quality can influence the distribution of bears on the landscape and, therefore, the chances of bear–human encounters and the frequency of bear-caused human injuries (Herrero 2002). Grizzly bears select foods that are relatively high in nutrients, are easily digested, and are concentrated spatially on the landscape to make for efficient foraging (Gunther et al. 2014a). We identified 22 seasonal bear foods present in Hayden Valley and the average peak periods of their use from literature reviews and field experience. We then compared the average peak period of use for each food to the dates of past grizzly bear–inflicted human injuries that occurred in Hayden Valley from 1970 to 2017, to estimate how many incidents may have been prevented if seasonal recreational closures had been in place during the peak period each food is consumed by bears (Table 1;  Text S2 (ursu-31-01-04_s01.docx)).

Table 1.

Common grizzly bear (Ursus arctos) foods and peak season of their use in Hayden Valley, Yellowstone National Park, Wyoming, USA, 1970–2017.

Recreational closures and restrictions based on bear activity

Recreational closures based on seasonal activity. In addition to seasonal foods, the presence of rub trees, scent-marking areas, ponds for thermoregulation, estrous females, day-bedding sites, denning sites, mineral feeding sites, travel routes, and other factors also influence the distribution of grizzly bears on the landscape, and therefore, the probability of encounters with backcountry recreationalists. We analyzed 163 very high frequency (VHF) telemetry locations from 7 male and 267 locations from 7 female radiocollared grizzly bears that frequented Hayden Valley during spring, summer, and autumn of 1986–2017, to determine the peak periods during which male and female grizzly bears were active in the valley. We obtained these VHF locations via fixed-wing aircraft approximately once per week during early to mid-morning hours. We calculated the proportion of these locations that occurred in Hayden Valley each week from den emergence through den entry.

We compared the peak period of activity for male, female, and male and female grizzlies combined in Hayden Valley with the dates of all bear-inflicted human injuries in the valley during 1970–2017. From this comparison, we estimated how many bear attacks may have been prevented if seasonal recreational-activity closures corresponding to the peak periods during which male and female grizzly bears were active in the valley had been in place the years the incidents occurred.

No grizzly bears were captured or radiocollared specifically for this study. All radiotelemetry data used for this study were collected by the Interagency Grizzly Bear Study Team as part of other research projects. Grizzly bear capture and handling procedures were reviewed and approved by the Animal Care and Use Committee of the U.S. Geological Survey (IACUC #2018-1); procedures conformed to the Animal Welfare Act, and to the U.S. government principles for use and care of vertebrate animals in testing, research, and training. Captures were conducted under U.S. Fish and Wildlife Service Endangered Species Permit [Section (i) C and D of the grizzly bear 4(d) rule, 50 CFR17.40 (b)], with additional state research permits for Wyoming, Montana, and Idaho, and National Park Service research permits for Yellowstone and Grand Teton National Parks.

Time-of-day recreational restriction based on grizzly bear diel activity. Grizzly bear diel-behavior patterns may also influence the frequency of human–bear encounters (Penteriani et al. 2017). Grizzly bears in YNP are most active during crepuscular and nocturnal time periods (Schleyer 1983, Harting 1985, Gunther 1990). A time-of-day restriction (e.g., recreational activity allowed only from 0900 to 1900 hours that prohibits recreational activity during nocturnal and crepuscular periods when grizzly bears are most active) could potentially reduce the temporal overlap between human recreational activities and bear foraging and traveling activity, thereby reducing the chances of surprise encounters and subsequent defensive bear aggression. A time-of-day restriction would also contribute toward making human activity in Hayden Valley more predictable to bears, thereby potentially reducing the chances bears would react defensively during surprise encounters. We investigated the potential impact that prohibiting recreational activity in Hayden Valley during crepuscular or nocturnal hours (1900 to 0900 hr) would have had on the frequency of bear-caused human injuries. We analyzed timing of bear attacks in the valley from 1970 to 2017 to determine how many occurred during nocturnal and crepuscular periods versus diurnal periods. We then calculated the proportion of bear-inflicted human injuries that may have been prevented by prohibiting hiking during nocturnal and crepuscular time periods.

Recreational restrictions based on human behavior

Certain human behaviors may influence the probability of encountering bears, as well as outcomes of those encounters (Herrero 2002, Penteriani et al. 2017). Herrero (2002) analyzed grizzly bear–inflicted human injuries from throughout North America and identified many similarities between the circumstances of, and human behavior in, human–grizzly bear encounters in backcountry areas. Herrero (1985, 2002) and Gunther and Hoekstra (1998) reported that party size and whether or not the hiker(s) were traveling on maintained trails or hiking cross-country off-trail both appeared to influence the frequency of encounters and bears reactions to them. Physical contact by bears was more likely to occur during encounters with small party sizes (≤3 people) and during off-trail encounters (Herrero 1985, 2002; Gunther and Hoekstra 1998). In addition, bear spray has proven effective at stopping aggressive behavior in bears during encounters and reducing the length and severity of bear attacks when they occur (Herrero and Higgins 1998, Smith et al. 2008). However, many backcountry hikers in YNP choose not to carry bear spray (Gunther et al. 2015). We developed visitor use restrictions that potentially would prevent the most common situations where human behaviors result in injuries by grizzly bears, including 1) requiring a minimum party size of ≥3 people for hiking in Hayden Valley, 2) prohibiting off-trail travel in the valley, and 3) requiring all hikers in the valley to carry bear spray. We then retrospectively applied these restrictions to Hayden Valley for the 1970–2017 time period to determine how many past injuries might have been prevented had these restrictions been in place. We calculated estimates of the potential effectiveness of the human-behavior-based recreational restrictions as follows:

Group size restriction. We assumed the proportion of bear-caused human injuries involving groups of ≥3 people in Hayden Valley would be approximately equal to the proportion of all hikers injured by bears in YNP during 1970–2017 that were traveling in groups of ≥3 people. Over the past 48 years, only 4% (n = 2) of the backcountry recreationalists injured by grizzlies in YNP were traveling in groups of ≥3 people when attacked. Most (n = 42, 95%) bear-inflicted human injures involved people hiking alone (n = 21, 48%) or with just 1 (n = 21, 48%) partner. Therefore, we assumed only 4% of those people hiking in groups of ≥3 people in Hayden Valley would be injured by grizzly bears.

Off-trail travel restriction. To estimate the efficacy of requiring hikers in Hayden Valley to stay on designated trails, we assumed that prohibiting off-trail travel would have prevented all of the past grizzly bear–caused human injuries in Hayden Valley that occurred while people were traveling cross country off-trail.

Requirement to carry bear spray. When deployed, bear spray is about 92% effective at stopping aggressive behavior in bears (Smith et al. 2008). However, not all people carrying bear spray are able to successfully deploy it during close-range aggressive encounters with grizzly bears. During 1970–2017, 30% of the people injured by bears in backcountry areas of YNP were carrying bear spray but were not able to deploy it, either because they did not have time to retrieve it, or they failed to get the safety tab off in time. To estimate the efficacy of implementing a requirement to carry bear spray, we assumed that the proportion of bear attacks stopped or defused by bear spray in Hayden Valley would be similar to the proportion of all hikers injured by bears in YNP for the years 1970–2017 that were carrying bear spray and had time to deploy it when attacked by a bear (70%), multiplied by the known proportion of bear spray deployments that are successful at stopping aggressive behavior in bears (92%; Smith et al. 2008). Therefore, we estimated bear spray would be effective at stopping or reducing the severity of attacks 64% of the time (0.70 proportion able to deploy bear spray × 0.92 proportion effective when deployed = 0.64 efficacy in stopping or reducing severity of attacks).

Characteristics of grizzly bear attacks in Hayden Valley, 1970–2017

Ten people were injured by grizzly bears in backcountry areas of Hayden Valley during 1970–2017, with 3 dying from their injuries (Table 2). The bodies of 2 of the 3 fatalities were partially consumed by bears. All 10 injuries occurred from late May through early October and between 1030 and 2100 hours. Eight of the injured people were day-hiking, 1 was backpacking, and 1 was day-riding but was on foot leading a mule at the time of attack. Four of the people were traveling alone and 6 were traveling with 1 other person. None of the injured people were in a group of ≥3. Seven of the injured people were traveling cross-country off of the designated trail and 3 were traveling on the designated trail. Eight of the 10 incidents, including 1 fatality, involved defensive aggression by grizzlies following surprise encounters. In 2 incidents (both fatalities with bear consumption of the body) the cause of attack was not known. In 1 of these 2 incidents, evidence suggests a photographer likely approached the grizzly bear for a picture prior to the attack. The circumstances surrounding another fatality are not known; however, there was evidence of a bison carcass nearby. Female grizzly bears (4 accompanied by young) were known to be involved in 8 of the10 incidents. Three of the 10 incidents were known to have occurred near bison carcasses. However, the presence of cubs and ungulate carcasses may have been underreported because cubs and bear-cached carcasses near encounter sites may not always have been visible to the injured hikers.

Table 2.

Characteristics of grizzly bear (Ursus arctos) –inflicted human injuries in Hayden Valley, Yellowstone National Park, USA, 1970–2017.

Efficacy of seasonal recreational closures based on bear foods

We analyzed the efficacy of closing Hayden Valley to recreational use during 22 different time periods that coincided with the peak seasons during which grizzlies consume a variety of foods found in the valley ( Text S1 (ursu-31-01-04_s01.docx)). Applying seasonal closures that coincided with the peak period during which grizzlies consume specific foods in Hayden Valley potentially would have prevented from as many as 6 (60%) to as few as none (0%) of the bear-caused human injuries that occurred in the valley during 1970–2017 (Table 3). The food-based closure that potentially would have prevented the most bear-caused injuries occurred during mid to late summer and coincided with the peak season during which bears scavenge bison carcasses that died from injuries sustained during the bison breeding season. Hayden Valley and Lamar Valley are the 2 primary bison breeding areas in YNP (White et al. 2015). However, Hayden Valley has much higher grizzly bear density index values (values of 19 and 21) than does the Lamar Valley (11), and more bear activity is generally observed on bison carcasses in Hayden Valley. Although the bison rut occurs from mid-July through mid-August (Meagher 1973, White et al. 2015), some adult males do not die from rut-related injuries and/or infection until early to mid-September. Bison carcasses can attract and hold large numbers grizzly bears for several days to several weeks. Telemetry data from radio-instrumented grizzly bears have shown direct straight-line movements toward carcasses from as far as 15 km away (van Manen et al. 2017). Twenty grizzly bears were observed in the immediate vicinity of a bison carcass in Hayden Valley during an aerial telemetry flight on 3 August 2007 (Fig. 3), and Craighead et al. (1995) reported a ground observation of 23 grizzly bears congregating around a single bison carcass in YNP. The number and concentration of bears attracted to bison carcasses increases the likelihood of surprise encounters between hikers and bears in the vicinity of carcasses. In addition, bison have large body size and, thus, their carcasses generally have edible biomass that attracts and holds bears for approximately 2 weeks before complete depletion by scavengers (Green 1994), further increasing the probability of human–bear encounters. Bears will often defend bison carcasses from conspecifics, wolves, and humans because of the high caloric value of the carcasses. Closing Hayden Valley during the peak period of rut-related bison mortality could reduce the chances of hikers encountering bears at bison carcasses. Had a seasonal closure been in effect in the Hayden Valley during the peak period of rut-related bison mortality (15 Jul–15 Sep), it may have prevented 6 of the 10 (60%) grizzly bear–inflicted human injuries that occurred in Hayden Valley during 1970–2017.

Table 3.

Potential reduction in grizzly bear (Ursus arctos) –inflicted human injuries in the Hayden Valley area of Yellowstone National Park, USA, 1970–2017, based on a retrospective evaluation of different types of grizzly food-based and activity-based seasonal recreational closures and human behavior-based recreational restrictions.

Fig. 3.

Photo of 17 of 20 grizzly bears (Ursus arctos) counted congregating around a bison (Bison bison) carcass in Hayden Valley, Yellowstone National Park, USA, on 3 August 2007. Three additional grizzly bears were observed just outside of the photo frame. (Photo by Interagency Grizzly Bear Study Team).

Although bears consume several foods during spring and early summer that are concentrated sources of calories and are known to attract and hold bears at specific locations (e.g., wolf-killed and winter-killed non-neonate bison and elk), none of the spring closures (Mar–May) would have been very effective at preventing the bear encounters that resulted in human injuries because few people hike in Hayden Valley during that time of year. The road through Hayden Valley does not open to the public until early May. Even after the road opens, the depth of accumulated snow and wet muddy conditions associated with spring snow melt limit visitor activity until later in June.

Closure of Hayden Valley during early summer (Jun–mid-Jul) would also be relatively ineffective at preventing encounters with grizzly bears that result in human injury. Except for elk calves, most foods consumed by bears in the valley during early summer are of lower caloric value, generally more dispersed across the landscape, and, therefore, not subject to directed grizzly feeding activity and less likely to be defended by bears. Although elk calves are a concentrated source of calories, they are generally consumed in just 1 or 2 feeding bouts immediately after predation (Gunther and Renkin 1990). Predation by grizzly bears on neonate elk does not attract and hold multiple bears for long periods of time; therefore, it is less likely to lead to human–bear encounters than larger carcasses such as bison.

Closures coinciding with the peak periods during which bears eat autumn foods would also not be very effective at preventing grizzly attacks. By late September and October, visitation to YNP (Fig. 2) and the number of people hiking in Hayden Valley decreases substantially, reducing the odds of bear–human encounters and, therefore, bear-inflicted human injuries.

Most of the food-based seasonal closures would have been ineffective at preventing bear-caused human injuries. The dates of our food-based closures were based on the average peak period during which bears consume different foods. However, ungulate migrations in spring and plant phenological development can vary by several weeks to a month or more depending on winter snow accumulation, spring snow melt, and spring and summer temperatures and precipitation. Therefore, the average dates of our food-based closures may not have aligned well with the actual periods during which bears sought specific foods during any given year.

Efficacy of recreational closures based on bear activity

Recreational closure based on seasonal activity. For the 7 radiocollared male grizzly bears that frequented Hayden Valley, 43 of 163 (26%) telemetry locations recorded were inside and 120 outside of the valley. Of the 43 telemetry locations inside Hayden Valley, over half (56%, n = 24) occurred during the 62-day period between 1 July and 31 August. For the 7 radiocollared female grizzly bears that frequented Hayden Valley, 90 of 267 (34%) telemetry locations recorded were in the valley. Of these 90 locations, over half (58%, n = 52) occurred during the 62-day period between 15 June and 15 August. Applying a seasonal closure during the period in which male grizzly bears are most active in Hayden Valley (1 Jul–31 Aug) may have prevented 5 of the 10 (50%) human injuries that occurred during 1970–2017 (Table 3). Had a seasonal closure been in effect in Hayden Valley from 1970 to 2017 during the peak period of female grizzly bear activity in the valley (15 Jun–15 Aug), 4 of the 10 (40%) grizzly bear–caused human injuries that occurred in the valley may have been prevented. Applying a seasonal closure that combines the peak period during which both male and female bears are active (15 Jun–31 Aug) may have prevented 7 of the 10 (70%) human injuries.

Time-of-day recreational restriction based on diel activity. Only 1 of 10 people injured by grizzlies in Hayden Valley from 1970 to 2017 was injured during nocturnal or crepuscular time periods. Nine of the 10 (90%) people were injured during diurnal time periods. Therefore, time-of-day restrictions closing Hayden Valley to recreational activity during nocturnal or crepuscular periods and allowing recreational use only during diurnal periods (0900–1900 hr) may have prevented only 1 (10%) of the bear-caused injuries in Hayden Valley (Table 3). The ineffectiveness of a time-of-day restriction in reducing bear attacks is likely due to 2 factors. First, most bear-inflicted human injuries in YNP occur during the time periods when people are most active (1000–1700 hr) and, thus, have the greatest chance of surprising bears. Few backcountry recreationalists are active during nocturnal and crepuscular time periods; therefore, there is little chance of hikers being injured during nocturnal and crepuscular time periods. In addition, female grizzly bears with cubs are more diurnal than other grizzly bears (Gunther 1990). Therefore, prohibiting recreational activity during nocturnal and crepuscular time periods would be most effective at reducing the chances of surprising adult male grizzly bears and female bears during years when they do not have cubs, but would be less effective at reducing the chances of encounters with females accompanied by cubs. Female grizzly bears with cubs are responsible for most human injuries involving defensive reactions during surprise encounters (Herrero 1970, Gunther and Hoekstra 1998). The presence of cubs may increase a bear's individual space requirements and the likelihood the bear will elect to fight if other bears or hikers enter that space (Craighead et al. 1995, Smith 2006). In addition, overnight camping is prohibited in Hayden Valley (National Park Service 1983), so there is very little chance of bear attacks during nocturnal and crepuscular time periods. A regulation limiting hiking in Hayden Valley to diurnal time periods would not be very effective at reducing the frequency of grizzly bear–caused human injuries in Hayden Valley.

Potential efficacy of recreational restrictions based on human behavior

Group size restriction. There is evidence that hiking in large groups reduces the risks of bear attack (Herrero 1985, Gunther and Hoekstra 1998). Large groups tend to make more noise that alerts bears to the group's presence at a distance, and large parties are also more intimidating to grizzly bears than smaller groups (Mc-Clung 2001, Herrero 2002). In addition, larger groups generally have better vigilance (the ‘many eyes’ effect) than smaller groups (Penteriani et al. 2017), thereby reducing the chances of surprise encounters. Herrero (2002) analyzed 165 grizzly bear–inflicted human injuries from North America and found no encounters involving parties of ≥6 people that resulted in human injuries. In YNP, groups of ≥3 people are rarely injured by grizzly bears (Gunther and Hoekstra 1998), and nobody in a group of ≥5 people has been injured in the past 47 years (1970–2017, YNP Bear Management Office records). Park-wide, solo hikers account for 48% of all bear-inflicted human injuries, hiking parties of 2 people account for 48% of all injuries, and groups of ≥3 people account for only 4% of all injuries (YNP Bear Management Office records).

All 10 people injured by bears in Hayden Valley from 1970 to 2017 were traveling alone or with just one partner. Therefore, retroactively applying a party size restriction requiring groups of ≥3 people to hike in Hayden Valley, and assuming 4% of those people would still be injured during encounters with bears (10 people × 0.04 = ∼1 person injured), the party size restriction may have prevented approximately 9.5 of the 10 (95%) bear-caused human injuries that occurred in Hayden Valley during 1970–2017 (Table 3).

Prohibition of off-trail travel. The danger of surprise encounters with grizzly bears decreases if people are alert and bears know where to expect people (Herrero 2002). With the current popularity of wildlife photography and advances in ultra-light backcountry gear, cross-country travel is becoming increasingly popular, making human activities less predictable to grizzlies and surprise encounters more probable. A restriction prohibiting off-trail travel would make human activity in Hayden Valley more spatially predictable to bears, possibly reducing the chances that bears will react defensively during surprise encounters. Bears may associate trails with people and react less aggressively toward people on trails where they expect to encounter them (Gunther and Hoekstra 1998). Bears appear to react more aggressively to encounters with people in off-trail areas where they are not expecting people (Gunther and Hoekstra 1998). Herrero and Fleck (1990) reported an increasing trend for grizzly bear–inflicted human injuries in North America to occur in off-trail areas. Gunther and Hoekstra (1998) reported that approximately half of the grizzly bear–inflicted human injuries in backcountry areas of YNP occurred while people were hiking off-trail. Yellowstone National Park has >1,600 km of maintained designated hiking trails. It is difficult to accurately compare the proportion of injuries that occur while people are travelling on trails versus off of trails because YNP does not keep records of how many recreationists travel on-trail versus off-trail. Maintained trails make travel and navigation easier and more efficient, so most backcountry recreationists in YNP likely spend the majority of their time on the designated trail system. Coleman (2012) reported that 67% of the backcountry recreationalists he surveyed in YNP remained on designated trails for the entirety of their hikes. In YNP during 1970–2017, 21 of the 44 (48%) people injured by grizzly bears in backcountry areas were traveling off-trail and 21 (48%) were traveling on designated trails (YNP Bear Management Office records). Two incidents (4%) occurred in backcountry campsites.

Seven of the 10 people injured by grizzly bears in Hayden Valley during 1970–2017 were traveling cross-country off-trail when the incidents occurred. Therefore, retroactively applying a prohibition of off-trail travel may have prevented 7 of the 10 (70%) bear-caused human injuries that occurred in Hayden Valley (Table 3).

Requirement to carry bear spray. Bear spray has proven effective at stopping aggressive behavior by bears during surprise encounters (Herrero and Higgins 1998, Smith et al. 2008) and is often recommended as a defensive tool to stop or defuse bear attacks (Brown 1996, Lapinski 1998, Herrero 2002, Rubbert 2006, Smith 2006, McCloskey 2009, Snow 2016). Although the first documented use of bear spray in YNP occurred in 1984, bear spray was not readily used by employees or recreationists in the park until the early 1990s. In the early to mid-1980s, bear spray was still primarily in the test phase (Hunt 1984, Rogers 1984), and in the late 1980s many people were still skeptical of its efficacy. Only 10 of the 46 (22%) people attacked by bears in backcountry areas of YNP during 1970–2017 were carrying bear spray. Seven of the 10 (70%) people had time to deploy it and all received only minor injuries. Five of these 7 did not deploy their bear spray until after the bear had already made contact. Hypothetically, if all 10 people injured by grizzly bears in Hayden Valley over the past 48 years (1970–2017) had been carrying bear spray, approximately 70% of them would likely have been able to deploy it. With a 92% efficacy rate (Smith et al. 2008), approximately 6–7 of them would likely have been able to stop or reduce the duration and severity of the interaction (10 previous injuries × 0.70 able to deploy bear spray × 0.92 success at stopping attack when bear spray deployed = 6.4 attacks stopped or of reduced severity; Table 3). Under the current paradigm, where more effort is being made to teach visitors about the importance of bear spray and how to use it effectively (Gunther et al. 2014b), we might expect that an even greater proportion of backcountry recreationists would be able to successfully deploy their bear spray to stop aggressive interactions with bears. A regulation requiring backcountry recreationists to carry bear spray has the potential to significantly reduce the number and severity of human injuries from grizzly bears in Hayden Valley.