Behavioral Coding of Captive African Elephants (Loxodonta africana): Utilizing DeepLabCut and Create ML for Nocturnal Activity Tracking

This study investigates the possibility of using machine learning models created in DeepLabCut and Create ML to automate aspects of behavioral coding and aid in behavioral analysis. Two models with different capabilities and complexities were constructed and compared to a manually observed control period. The accuracy of the models was assessed by comparison with manually scoring, before being applied to seven nights of footage of the nocturnal behavior of two African elephants (Loxodonta africana). The resulting data were used to draw conclusions regarding behavioral differences between the two elephants and between individually observed nights, thus proving that such models can aid researchers in behavioral analysis. The models were capable of tracking simple behaviors with high accuracy, but had certain limitations regarding detection of complex behaviors, such as the stereotyped behavior sway, and displayed confusion when deciding between visually similar behaviors. Further expansion of such models may be desired to create a more capable aid with the possibility of automating behavioral coding.

From left to right all through the night: Characteristics of lying rest in zoo elephants

Despite increased research during the past years, many characteristics of resting behavior in elephants are still unknown. For example, there is only limited data suggesting elephants express longer lying bouts and increased total nightly lying durations on soft substrates as compared to hard surfaces. Additionally, it has not been investigated how frequently elephants change body sides between lying bouts. Here we present these characteristics based on observations of nighttime lying behavior in 10 zoo elephants (5 African Loxodonta africana and 5 Asian Elephas maximus elephants) living in five different European facilities. We found that elephants housed on soft substrates have significantly increased total lying durations per night and longer average lying bouts. Furthermore, at 70%-85% of all bouts, a consistently higher frequency of side change between lying bouts occurred on soft substrates, leading to an overall equal laterality in resting behavior. Deviations from this pattern became evident in elephants living on nonsand flooring or/and in nondominant individuals of nonfamily groups, respectively. Based on our findings, we consider elephants to normally have several lying bouts per night with frequent side changes, given an appropriate substrate and healthy social environment. We encourage elephant-keeping facilities to monitor these characteristics in their elephants' nighttime behavior to determine opportunities for further improvements and detect alterations putatively indicating social or health problems in individual elephants at an early stage.

What’s Black and White and Pink All Over? Lesser Flamingo Nocturnal Behaviour Captured by Remote Cameras

The study of animal behaviour is important for the development of husbandry and management practices for zoo-housed species. Yet, data are typically only collected during daylight hours, aligning with human work schedules rather than animal activity patterns. To remedy this, 24 h data collection is needed. This study investigated the behaviour of a captive flock of lesser flamingos to understand temporal changes in their time-activity patterns. Two remote camera traps were placed around the birds’ outdoor enclosure and one within the indoor house. Counts of birds visible within specific enclosure zones were recorded from photographic data. Behaviour was defined as active or inactive, and modified Spread of Participation Index (SPI) was used to calculate enclosure zone occupancy. Results indicated that lesser flamingos are active overnight, and to a similar amount as in the daytime. Proportions of birds observed as active were significantly higher at later times of the day (i.e., dusk) when compared to the number of active birds in the morning. Enclosure usage was diverse and indoor and outdoor zones could be used by different numbers of birds at different times of the day. Variation in enclosure usage may indicate the changing needs of the flamingos when housed indoors overnight and when they have night-time access to an outdoor enclosure. This research has identified the need for further research into the nocturnal behaviour and space use of lesser flamingos and suggests the need for 24 h research in captive birds, and other zoo-held species, especially when species are locked indoors or face behavioural restriction overnight due to biosecurity measures surrounding zoonoses outbreaks, e.g., Avian Influenza.

The historical development of zoo elephant survivorship

In the discussion about zoo elephant husbandry, the report of Clubb et al. (2008, Science 322: 1649) that zoo elephants had a "compromised survivorship" compared to certain non-zoo populations is a grave argument, and was possibly one of the triggers of a large variety of investigations into zoo elephant welfare, and changes in zoo elephant management. A side observation of that report was that whereas survivorship in African elephants (Loxodonta africana) improved since 1960, this was not the case in Asian elephants (Elephas maximus). We used historical data (based on the Species360 database) to revisit this aspect, including recent developments since 2008. Assessing the North American and European populations from 1910 until today, there were significant improvements of adult (≥10 years) survivorship in both species. For the period from 1960 until today, survivorship improvement was significant for African elephants and close to a significant improvement in Asian elephants; Asian elephants generally had a higher survivorship than Africans. Juvenile (<10 years) survivorship did not change significantly since 1960 and was higher in African elephants, most likely due to the effect of elephant herpes virus on Asian elephants. Current zoo elephant survivorship is higher than some, and lower than some other non-zoo populations. We discuss that in our view, the shape of the survivorship curve, and its change over time, are more relevant than comparisons with specific populations. Zoo elephant survivorship should be monitored continuously, and the expectation of a continuous trend towards improvement should be met.

Behaviour and Welfare Impacts of Releasing Elephants from Overnight Tethers: A Zimbabwean Case Study

Within the southern African elephant tourism industry, chaining or tethering elephants is still a relatively routine practice, despite the known negative impacts. Cited reasons for chaining include fear of aggressive interactions between elephants when handlers are absent, or a general increase in expression of aggressive behaviours (both to other elephants and to their human handlers). In Zimbabwe, concerns expressed include the danger of elephants escaping and entering human-inhabited areas. Four male semi-captive elephants at a Zimbabwe tourist facility were taken off overnight (~12 h) tethers and were placed in small pens (‘bomas’), approximate sizes from 110 m2 to 310 m2), as part of a strategy to improve elephant welfare. Behavioural data were collected from overnight videos from December 2019 to March 2020, between 18:00 to 06:00, using focal, instantaneous sampling (5-min interval). Data were collected for three nights at three time periods: (i) Tethered; (ii) approximately four weeks post-release; (iii) approximately eight weeks post-release. Behavioural change over these time points was analysed using general linear models with quasibinomial error structures. Behavioural changes indicative of improved welfare were observed following these management changes, and no significant increases in aggression were observed either between elephants, or towards their human handlers. Proportion of time engaging in lying rest was higher in the first month after release from tethering (mean ± SD, 50 ± 14%) than when elephants were tethered (20 ± 18%) (p < 0.05). Additionally, although not statistically significant, stereotypies were reduced when elephants were no longer tethered (4 ± 6% observations tethered compared to 2 ± 2% off tethers), and positive social behaviour also increased (1 ± 1% on tethers, 2 ± 2% off tethers), with the greatest improvements seen in the pair-housed elephants. To improve elephant welfare in southern African tourism facilities we strongly advocate that less restrictive management practices which enable greater choice and freedom of movement overnight are implemented.

Nighttime behavior and the length of social relationships in male Asian elephants

To evaluate elephant welfare, it is important to understand their use of time both during day and night. The length of social relationships can influence how much time they spend in different activities. We assessed daytime and nighttime activity budgets of male Asian elephants at Denver Zoo and examined how length of relationships influenced nighttime behavior. Using scan sampling we investigated activity budget and proximity to a conspecific, and used General Estimating Equations to compare them across day and night and across new and established dyads at night. During daytime, elephants spent significantly more time exhibiting affiliative and agonistic behaviors, and in proximity to a conspecific, and less time resting, compared to night. Overnight, the odds of resting were significantly lower in new social dyads compared to established dyads, and new dyads spent more time exhibiting agonistic behavior and in proximity to a conspecific compared to established dyads. Our study suggests that male elephants at Denver Zoo have developed strong relationships and highlights the importance of systematically observing elephants overnight so that managers make decisions that improve animal welfare.

Working to Supply the Demand: Recent Advances in the Science of Zoo Animal Welfare

If animal welfare scientists were economists, we would be saying that the demand for knowledge is increasing faster than the supply [...]