Zoo Closure Does Not Affect Behavior and Activity Patterns of Palawan Binturong ( Arctictis binturong whitei )

: Exploring the interaction between humans and animals has become increasingly important in the evaluation of well-being for species housed in zoos and aquaria. The COVID-19 pandemic saw the global closure of zoos and aquaria to visitors. Chester Zoo, U.K., was no exception, with the charity shutting its doors for the longest period in its 90-year history. Whilst access to site was strictly limited to essential animal care staff, recent investment in networked infrared CCTV camera systems allowed some species to be monitored remotely during this extraordinary period of zoo closure. Here, we used this equipment to investigate whether zoo closure inﬂuenced activity patterns and behavior of two adult Palawan binturong, Arctictis binturong whitei . The cameras facilitated behavioral monitoring over 24 h enabling the collection of a full activity budget, which revealed a natural crepuscular activity pattern. Overall, visitor presence was found to have a neutral effect on this species, with no signiﬁcant difference observed in time spent engaging in den use, vigilance or travel behaviors during zoo open and zoo closed conditions. A neutral visitor effect was found when evaluating behavior over a 24 h period and during hours which the zoo would normally be open to visitors (10:00–16:30). This research presents new information on this elusive and understudied species in captivity, and promotes investment in monitoring equipment which enables more comprehensive behavioral sampling than traditional visitor-effect methods.


Introduction
Understanding the association between animals and humans has become increasingly important in the evaluation of animal well-being in zoos and aquaria. The newly updated Five Domains model for animal welfare assessment highlights the significance of this relationship, by integrating 'animal-visitor' interactions into their assessment approach. In this amendment, Mellor et al. [1] highlights the significance of human-animal relationships, stating that human behavior can elicit both positive, neutral and negative behavioral responses in animals [1,2].
A large body of research has accumulated over recent decades on the impact of visitor presence on zoo animal welfare [3][4][5]. As with many zoo studies, there is a taxonomic bias, with the largest number of visitor-effect studies focusing on non-human primates [6,7]. Although this bias is being slowly addressed, very few studies to date have focused on the impact visitors may have on crepuscular or nocturnal species [8]. This may be due to the assumption that these species might be less impacted by visitors as they spend most time active when visitors are not present [9]. However, this assumption should be treated with caution, as visitor presence might disturb vital inactive behaviors, such as rest and sleep which can have a detrimental impact animal wellbeing [10,11]. Furthermore, documenting the impact of visitors on nocturnal or crepuscular species presents a challenge as institutions do not always have the ability to observe them during their natural period of activity. As animal care staff are largely present during daylight hours, monitoring opportunities are often limited to periods in which species, such as binturong, are largely inactive, causing the potential for welfare or behavioral issues to be unobserved [12]. Investment in infrared camera equipment with remote access options can help address these monitoring issues [13].
In 2020, the outbreak of COVID-19 forced zoos around the globe to close their doors to visitors for an extended period. Despite these circumstances, the abrupt and prolonged closure of zoos and aquaria worldwide provided a unique opportunity to investigate the effect that visitors may have on zoo-housed species [14][15][16]. For example, during this period, visibility of amphibian species was suggested to be influenced [17] whilst differences in behavior were attributed to seasonal rather than visitor effects for two species of flamingo (Phoenicopterus roseus and Phoenicopterus chilensis) [18]. Enclosure use alongside behavior was also explored in published literature in relation to visitor presence, with meerkats (Suricata suricatta) using more of their exhibit areas during periods of zoo closure, whilst space use for African penguins (Spheniscus demersus) was unaffected [19]. Additionally, space utilization combined with fecal glucocorticoid concentrations revealed an intra-individual difference in response to zoo closure for banteng (Bos javanicus), with a temporary elevation in glucocorticoid concentrations for two of the five animals studied and an increase in time spent close to visitor areas when zoo-visitors returned after closure [16]. Studies also varied in length, with some outlining a shorter pre to post closure comparison [20] whilst others looked more longitudinally, assessing seasonal differences in behavior alongside changes in visitor density [21].
This study documents activity patterns and behavior of two binturong during the closure of Chester Zoo, UK in response to the COVID-19 pandemic. The Palawan binturong (Arctictis binturong whitei) is a threatened frugivorous carnivore species, native to the forests of South-East Asia [22]. Fondly referred to as the 'bear cat', binturong are held in zoos due to their distinctive morphology and threatened conservation status. However, their naturally crepuscular activity pattern [23] means they are often resting during visitor opening hours. Using networked infrared CCTV cameras, we adopted a 24 h approach in monitoring of this under-studied species allowing for a more detailed insight into activity patterns and behavior in response to the absence of zoo visitors. Based on species natural history and previous published work [20,24,25], we make the following predictions: that binturong activity over 24 h will follow a crepuscular pattern (P1) and that these activity patterns are mostly likely to be affected by the change in visitor presence during zoo opening (P2). Additionally, we predict that engagement in specific behavioral categories may be influenced by the change in visitor presence, with vigilance behavior increasing during the zoo open condition (P3) and den use behavior decreasing throughout the zoo closed condition (P4).

Study Subjects and Housing
Study subjects were 1.1 (1 male, 1 female) Palawan binturong housed at Chester Zoo, U.K. (Figure 1). Both individuals were 11 years old at the time of study and formed part of a mixed species exhibit with 1.2 Malayan sun bears (Helarctos malayanus) in Chester Zoo's 'Islands' area. The enclosure comprised of one indoor on-show exhibit (96 m 2 ), with off-show den areas (7 m 2 ) accessed through custom-built metal tunnels for this species, and two open-air outdoor on-show exhibits (totaling 2290 m 2 ). Despite being a mixed species exhibit, several species-appropriate furnishings for the binturong were included within the environment such as overhead ropes, wooden posts and platforms to provide arboreal transit routes and resting sites. Subjects had 24 h access to the whole enclosure throughout data collection, except during times of keeper maintenance, which lasted for no more than 30 min, usually between the hours of 11:00-13:00. throughout data collection, except during times of keeper maintenance, which lasted for no more than 30 min, usually between the hours of 11:00-13:00.

Data Collection
Behavioral data were collected using continuous focal sampling via networked CCTV cameras (Axis IP Camera, North West Security Group) for 60 min observation periods [26], using a pre-determined ethogram (Table 1). Some observations ended earlier than the 60 min period due to observer availability through reduced staffing. Cameras had infra-red capability, with the footage being automatically saved and uploaded onto a server, allowing footage to be accessed and replayed remotely for up to one month after capture (Milestone XProtect video management software). Due to the nature of the exhibit, approximately 30% of the environment was not covered by CCTV cameras. Data collection aimed to cover the whole 24 h period equally, in order to obtain a comprehensive overview of behavior. Data collection during the 'Zoo open' phase were collected due to routine behavioral monitoring for this species as requested by staff within the animal collections directorate. Chester Zoo closed to the public on 21 March 2020, in response to the COVID-19 outbreak, and behavior observations on the binturong resumed 22 days later.
Data were collected during two conditions:

Data Collection
Behavioral data were collected using continuous focal sampling via networked CCTV cameras (Axis IP Camera, North West Security Group) for 60 min observation periods [26], using a pre-determined ethogram (Table 1). Some observations ended earlier than the 60 min period due to observer availability through reduced staffing. Cameras had infra-red capability, with the footage being automatically saved and uploaded onto a server, allowing footage to be accessed and replayed remotely for up to one month after capture (Milestone XProtect video management software). Due to the nature of the exhibit, approximately 30% of the environment was not covered by CCTV cameras. Data collection aimed to cover the whole 24 h period equally, in order to obtain a comprehensive overview of behavior. Data collection during the 'Zoo open' phase were collected due to routine behavioral monitoring for this species as requested by staff within the animal collections directorate. Chester Zoo closed to the public on 21 March 2020, in response to the COVID-19 outbreak, and behavior observations on the binturong resumed 22 days later.
Data were collected during two conditions:
Total combined observation time for both individuals was 184.6 h (n = 5 days) in the 'Zoo open' condition and 163.2 h (n = 6 days) in the 'Zoo closed condition'. During the 'Zoo open' condition, visitors had access to site from 10:00-16:30. Data were collected digitally, with information inputted directly into a Microsoft Excel spreadsheet by the observers (KF and ML). Study subjects remained in a consistent social grouping, with no management or husbandry routine changes throughout the data collection window. As all observations were conducted remotely using CCTV cameras, there was no observer influence. For visualization of activity levels across a 24 h period, individual behaviors were grouped into 'Active' or 'Inactive' categories (Table 1). For the analysis of specific behaviors in response to zoo closure, individual behaviors were re-grouped into main behavior categories (Table 1). In order to gain a holistic insight into how visitors may influence species engagement in defined behavioral categories, data were split into two distinct time periods '24 hour' which contained all time periods observed 00:00:00-23:59:59 and 'Zoo visitor opening hours' which contained time periods 10:00:00-16:30:00. Due to low engagement in the 'Other' behavioral category (<1% of total activity budget), these behaviors were removed from statistical analysis.

Statistical Analysis
Statistical analyses were conducted in R (Version 3.6.2) [27]. Proportion of time spent engaging in each behavior category were calculated per hour and averaged during each day ('Day' was defined as 00:00:00-23:59:59) of data collection. Day was then treated as the experimental unit. Due to blind spots in the CCTV camera coverage, data were only analyzed from periods of time which were either visible to the observer or their location was undisputed, e.g., time utilizing den areas. Behavior data were found to be non-normally distributed (Shapiro Wilks test for normality, w = 0.516, p < 0.005) and transformations failed to normalize the data, as a result non-parametric tests were used. Paired Wilcoxon signed rank tests were used to compare the mean proportions of time spent engaging in each behavior category between the 'Zoo open' and 'Zoo closed' conditions. No significant difference in time spent in each behavior category between the study subjects was found (Wilcoxon signed rank test: W = 50, p = 0.913), therefore data from each individual were combined and analyzed as one data set throughout. Activity patterns are displayed using descriptive statistics using means ± 1 standard error.

Activity Budget of Zoo Housed Binturong in Relation to COVID-19 Zoo Closure
When considering behavior over a 24 h period during the 'Zoo open' compared to 'Zoo closed' conditions, there were no significant differences in proportion of behavior observed (Wilcoxon signed rank test: Den Use, V = 8.00, p = 1, Travel, V = 6.00, p = 0.81, Vigilance, V = 8.00, p = 1.00, Figure 3). When considering behavior during zoo visitor opening hours only, there were no significant differences in proportion of behavior observed (Wilcoxon signed rank test: Den Use, V = 13.00, p = 0.19, Travel, V = 3.00, p = 0.31, Vigilance, V = 2.00, p = 0.19, Figure 3). These findings were contrary to our prediction that engagement in specific behavioral categories may be influenced by the change in visitor presence (P3). As stated in the ethogram (Table 1), various behaviors were grouped in the 'Other' category.
J. Zool. Bot. Gard. 2022, 3, FOR PEER REVIEW 5 engaging in each behavior category between the 'Zoo open' and 'Zoo closed' conditions. No significant difference in time spent in each behavior category between the study subjects was found (Wilcoxon signed rank test: W = 50, p = 0.913), therefore data from each individual were combined and analyzed as one data set throughout. Activity patterns are displayed using descriptive statistics using means ± 1 standard error.

Activity Budget of Zoo Housed Binturong in Relation to COVID-19 Zoo Closure
When considering behavior over a 24 h period during the 'Zoo open' compared to 'Zoo closed' conditions, there were no significant differences in proportion of behavior observed (Wilcoxon signed rank test: Den Use, V = 8.00, p = 1, Travel, V = 6.00, p = 0.81, Vigilance, V = 8.00, p = 1.00, Figure 3). When considering behavior during zoo visitor opening hours only, there were no significant differences in proportion of behavior observed (Wilcoxon signed rank test: Den Use, V = 13.00, p = 0.19, Travel, V = 3.00, p = 0.31, Vigilance, V = 2.00, p = 0.19, Figure 3). These findings were contrary to our prediction that engagement in specific behavioral categories may be influenced by the change in visitor presence (P3). As stated in the ethogram (Table 1), various behaviors were grouped in the 'Other' category.
the Zoo open condition. During the Zoo closed condition during zoo visitor hours, engagement in two behaviors from the 'Other' category were observed; Feeding (0.01% ± 0.01) and Sniffing (0.04% ± 0.014).

Discussion
In this study we found that Palawan binturong at Chester Zoo exhibit a crepuscular activity pattern (as predicted in P1), which remained consistent when the zoo was both open and closed to visitors (differing from our second prediction, P2). The binturong spent most of their time exhibiting three main behaviors, 'Den Utilization, 'Travel' and 'Vigilance'. Contrary to our predictions P3 and P4, the proportion of time spent exhibiting these behaviors did not significantly change between 'zoo open' and 'zoo closed' conditions.
Whilst studies often report the influence that zoo visitors have on animal behavior [28], some work has highlighted that visitor presence has little to no impact on the activity budget of zoo-housed species [8,19,29]. O'Donovan [30] reported that both visitor number and noise level did not influence the activity budget of cheetahs (Acinonyx jubatus) and Wascher et al. [29] outlined that no negative welfare impacts of visitors were found in relation to visitor presence in captive corvids. When assessed longitudinally, neither season nor visitor density had an influence on the behavior and welfare of five species of zoohoused bear [21]. A neutral response to change in visitor numbers could be attributed to

Discussion
In this study we found that Palawan binturong at Chester Zoo exhibit a crepuscular activity pattern (as predicted in P1), which remained consistent when the zoo was both open and closed to visitors (differing from our second prediction, P2). The binturong spent most of their time exhibiting three main behaviors, 'Den Utilization, 'Travel' and 'Vigilance'. Contrary to our predictions P3 and P4, the proportion of time spent exhibiting these behaviors did not significantly change between 'zoo open' and 'zoo closed' conditions.
Whilst studies often report the influence that zoo visitors have on animal behavior [28], some work has highlighted that visitor presence has little to no impact on the activity budget of zoo-housed species [8,19,29]. O'Donovan [30] reported that both visitor number and noise level did not influence the activity budget of cheetahs (Acinonyx jubatus) and Wascher et al. [29] outlined that no negative welfare impacts of visitors were found in relation to visitor presence in captive corvids. When assessed longitudinally, neither season nor visitor density had an influence on the behavior and welfare of five species of zoo-housed bear [21]. A neutral response to change in visitor numbers could be attributed to a degree of habituation to the presence of zoo visitors in a captive environment [31]. Patel et al. [32] highlight that the likelihood of habituation to visitors is greater if repeated visitor exposure has neither rewarding nor punishing elements and is therefore of no consequence to the species. Sherwen and Hemsworth [2] also outline that the degree to which individuals are likely to habituate to visitor presence can be attributed to multiple factors, such as species-specific biology, enclosure design and individual differences in behavior. Furthermore, Sherwen and Hemsworth [2] suggest that a neutral response to visitors may be far more widespread than the current literature implies, due to a potential current publication bias. This bias may have been created as a consequence of a previous strong focus on highlighting factors which influence the welfare state of zoo-housed animals, rather than those that do not. When considering literature focusing on zoo-closure due to COVID-19, behavioral response was found to vary by species, with some being largely unaffected whilst other species were observed to alter their engagement in specific behavioral categories and enclosure utilization [17,19,20,33].
Similar to the findings in our study, activity levels of several large felid species at Brookfield Zoo were not found to be influenced by visitor presence. Margulis et al. [34] suggested that as the species observed were naturally inactive throughout most of the day, they were less likely to be affected by visitors, which could also be the case for Palawan binturong. Work by Queiroz and Young [9] supports this theory as they concluded diurnal species showed more behavioral changes in relation to visitor presence than nocturnal in a study of 17 mammal species. A suggested reason for this difference was that the diurnal animals were being observed during their normal activity cycle whilst nocturnal animals were largely inactive during the data collection window [9].
However, the human-animal relationship is complex [1]. Many variables could also influence animal behavior in zoos such as weather conditions, keeper visibility, time of day and husbandry routines [35]. Goodenough et al. [36] explored this relationship in a group of captive ring-tailed lemurs (Lemur catta) in a walkthrough exhibit, where potential for visitor effect was suggested to be high. Results revealed that time and weather exerted the strongest effect on behavior, affecting resting, feeding/foraging and locomotion. Similarly, Rose et al. [37] highlighted how any visitor effect is heavily influenced by environmental variables such as temperature and humidity in a pair of captive hornbills. Thus, future directions for work on our study species, the binturong, would be to collect further data allowing for a comprehensive evaluation of how environmental and other variables could influence the neutral visitor effect that we reported.
Little is known in relation to the natural behavioral patterns of Palawan binturong. Research has proved challenging on this elusive species, with individuals known to occupy densely planted, high canopy areas, making them difficult to detect and track in situ [38]. Camera trap images have been used to gather information on individual behaviors [22], revealing their activity pattern to vary from arrhythmic to crepuscular [18]. In light of this information we made the first prediction for this study (P1). P1 was found to be correct as the Palawan binturong at Chester Zoo were found to exhibit a crepuscular activity pattern, with the highest levels of activity observed between 04:01-05:00, 21:01-22:00 and 23:01-00:00. It should be noted however, that patterns of activity were reported to be highly seasonal for wild binturong, with individual travel behavior closely correlated with the timing of fruiting trees [39]. As such, further work may be beneficial during a range of seasons in order to ascertain whether the activity pattern recorded in this study is consistent across the calendar year.
Our finding that visitor presence does not cause behavioral changes in Palawan binturong addresses the gap in literature regarding visitor impacts on non-diurnal zoohoused species. The neutral reported response to visitors by this species contradicts our predictions P2 to P4, but does help to prevent potential publication bias in this field. When accounting for species natural biology, the lack of response to visitor presence or absence, both during visitor opening hours and across a 24 h period, highlights that individuals are continuing to replicate their natural activity patterns. High utilization of the den area also shows that the enclosure design is facilitating individuals to have an element of choice in where they spend their time and provides them with an option to be out of view from visitors. Access to off-show areas was thought to mediate the visitor effect in a group of brown bears, with no visitor influence on behavior found in individuals which had access to an indoor holding space [14].
However, the identification of non-significant trends in activity during time periods where visitors would usually be present should not be overlooked. Other factors such as interaction with other species housed in the same facility (in this case the Malayan sun bear), change in social dynamic of the study subjects or indirect consequence of visitor absence, such as a change in other staff not involved in animal care (guest experience coordinators, etc.) could all warrant further investigation. Furthermore, due to limited researcher capacity through reduced staffing only a short time window during zoo closure conditions was able to be captured. Data collection to investigate both the immediate and longer term responses to zoo closure would have been beneficial in gaining a more comprehensive insight into the visitor effect over a longer timeframe.
It is only by accounting for the full 24 h period that we can conclude that behavior and activity patterns of binturong housed at Chester Zoo were not affected by the COVID-19 closure. Whilst we observed greater variation in activity levels over zoo visitor opening hours, overall activity levels and the three main behaviors tested did not significantly change between the 'Zoo open' and 'Zoo closed' conditions. We therefore encourage the investment in appropriate monitoring equipment for species, particularly those who are biologically less active during zoo opening hours, in order to facilitate a more holistic overview of well-being.

Conclusions
Our findings followed our first prediction (P1) that 24 h activity patterns for Palawan binturong are crepuscular in nature. However, contrary to our second prediction (P2) we found that activity patterns did not change during zoo closure; this was the case over the full 24-h period and when just focusing on visitor opening hours. When considering our final two predictions, P3 and P4, we found them to be incorrect, as, engagement in vigilance and den use behavior of binturong was not affected by visitor presence in this study. We hope that the dissemination of this work will add to the knowledge base on this understudied and elusive species, whilst adding to the body of work surrounding the visitor effect on zoo-housed species in relation to COVID-19 closure, particularly highlighting information on the neutral visitor effect. Institutional Review Board Statement: Ethical review and approval were waived for this study due to purely observational and non-invasive nature of behavioral data collection. Project was internally approved by senior animal and science directorate staff at Chester Zoo.

Informed Consent Statement: Not applicable.
Data Availability Statement: Data available on request.