4.1. Interpretation of Trial Results
The trial of the welfare risk assessment process demonstrated its practicality and effectiveness in identifying areas of welfare risk in a wide range of species in a zoo setting. The decrease in average welfare scores at each of the three zoos in 2017 reflects an increase in the number of welfare risks identified. At first glance, this may appear to suggest a deterioration in conditions and welfare level over time. However, upon further consideration, many of the welfare risks newly identified in the 2017 assessments were in fact present in the two previous years but were not identified or scored to the same degree. There appear to be three possible explanations for this outcome: the conditions themselves did in fact deteriorate, the basis for judgement by the staff changed, or both.
Staff experience, education, and awareness of animal welfare are obviously key elements in effective utilization of these processes. It is noteworthy, therefore, that one month prior to data collection for the 2017 round, a three-day staff training workshop was held for keepers and veterinarians across the three zoos. The workshop covered topics on factors that are known to influence welfare in captive animals, scientific assessment of animal welfare and tools to accurately identify and articulate risks to welfare. The majority of staff who completed the assessments were in attendance at this workshop, so it is possible that this training and open communication around animal welfare increased their awareness of, and ability to, identify and report welfare risks leading to lower welfare scores in 2017. Unfortunately, no competency-based assessments were conducted on staff before and after their workshop attendance, so the effect of this form of education on zoo personnel competency and thus the scoring outcomes remains unknown. This is something that is worth investigating in the future as assessing the inter- and intra-observer reliability of the assessment team would both aid in validating the welfare risk assessment process, and evaluating the effectiveness of education or training delivered by the organisation. Nevertheless, any improvement in staff ability to identify and report welfare risks and develop strategies to advance welfare standards proactively is an important factor in driving positive change. As such, this aspect of the results should be considered to be positive from a motivational perspective. Furthermore, as staff become more familiar with the assessment process and gain experience identifying welfare risks, it is expected that their ability to make judgements around welfare risks will be enhanced.
It is also possible that in the first two years of data collection there was a bias towards more positive scoring because of the perception that welfare standards reflect the quality of keepers’ care for their animals. For this reason, it is important that the use of such processes is premised with clear communication around the primary role of the process being to identify opportunities for animal welfare improvements, not to detect deficiencies in staff delivery of care.
Considering the highest and lowest scoring indicators/measures, it is interesting that animal observation time was scored by staff as a common risk factor in their ability to deliver high standards of welfare. This is logical as keepers are expected to be the ‘voice’ for the animals in their care. They have the best understanding of the individual animal’s temperament and preferences and can detect subtle changes in their behaviour [
29]. It is likely that this ability to detect changes in observable welfare indicators in individuals is heavily influenced by the time that keepers spend observing the animals, and ultimately their familiarity with their behaviour. Alternatively, it is possible that keepers gain a somewhat distorted picture of the animal’s behaviour if the animals they work with recognise them as sources of food, enrichment, and/or other positive or negative events. For this reason, it is beneficial to facilitate processes that allow keepers to conduct systematic behavioural observations in ways that do not interfere with the animals’ behaviour. This was the rationale for the installation of various new CCTV networks in enclosures in 2017, based on the welfare priorities identified through the use of this assessment process. Remote monitoring of animals can also allow noninvasive assessment of animals for those at risk of adverse responses to disturbance (e.g., nocturnal species, shy species, and animals in recovery from veterinary procedures).
The sensory environment was also identified as a significant welfare risk factor, suggesting that the keepers believed this to be an environmental feature that posed certain risks to animals. This is in line with the recent increase in published studies investigating the impacts of various sensory conditions on animals in zoo environments such as noise exposure [
39,
51,
52], light conditions [
53], and visual stimuli [
24,
54]. There also appears to be increasing awareness that we do not yet have a good understanding of the breadth of sensory perception across the multitude of species housed in zoos, and as a result the impact of the zoo environment on these animals [
39].
In terms of high-scoring indicators, keeper relationships with animals were scored positively. It is well recognised in other settings that animal caretakers can have profound impacts on welfare [
50], but few studies have examined keeper impacts on animal welfare in the zoo industry. In an investigation into reproductive success in a variety of small felid species in zoos, a positive relationship was observed between reproductive success and a husbandry style where keepers spent a considerable amount of time talking to and interacting with the cats, referring to this keeping style as high-quality [
55]. A study of 72 Clouded Leopards across 12 zoos in the USA revealed that lower faecal glucocorticoid concentrations were associated with keepers spending a greater amount of time with the leopards and fewer keepers caring for the animals [
38]. In both studies the authors linked these husbandry factors to the keepers’ ability to form high quality relationships with their animals. Moreover, many zoo professionals report that they have established bonds with the animals in their care, particularly with primates and carnivores [
56]. It follows that the identification of a high-quality relationship between keepers and animals in this assessment process is a promising trend.
The assessments also highlighted that the main gaps in knowledge existed in the animal behaviour domain, including measures such as behavioural diversity, frequency of abnormal behaviour, and activity budgets. This is likely associated with the low scores for ‘animal observation time’ for keepers. If keepers feel they do not have enough time to observe the animals, they are also likely to feel that they do not have a solid understanding of how the animals spend their time and what behaviours they engage in. Furthermore, some species housed in these zoos are very rare and not well studied in the wild, creating a lack of understanding of behavioural biology for the species, making it very difficult to judge what is considered ‘normal’ behaviour. This is an important gap for zoos to address by investing in both in situ and ex situ research programs.
With regards to taxonomic differences, mammals received the lowest average welfare score compared to birds, fish, invertebrates, and reptiles and amphibians. This may be because there is a greater understanding of factors that can affect the welfare of mammals and therefore a greater capacity to identify risks. Comparatively, there are fewer studies on reptile, bird, and fish welfare in zoos [
57]. Likewise, the lowest scores assigned in the Environment domain across taxa may be due to an enhanced ability to identify risks in this domain. The main environmental features do not change frequently in zoos, so environmental assessments can be made readily and frequently without a need for ongoing observation as is required for animal behaviour assessments.
4.2. The Value of the Process
Safeguarding animal welfare and continuous improvement in housing and husbandry of the animals living in zoos is arguably the biggest opportunity the zoo sector faces with so many exotic species having divergent species-specific and individual requirements. In order to be confident that welfare standards are improving, an evidence-based management approach should be encouraged in zoos [
58], including advanced understanding of diverse species requirements, as well as individual animal preferences and motivations. Research collaborations with Universities or other scientific organizations may prove especially useful in developing this knowledge base. Studies for acquiring the evidence for such approaches can range in scale from more detailed and narrowly focused experiments to broader assessments of welfare risks. It is apparent that a range of tools are needed to assist the zoo sector to understand, assess, and ultimately advance welfare standards. Kagan et al., [
43] proposed an animal welfare framework for zoos that included two tools designed to collect qualitative information about an institution’s processes and programs, with some similar focus areas to the present risk assessment process (e.g., evaluation of climatic conditions, social groups, and enrichment provision). The present process builds upon the Kagan et al. [
43] framework, in the context of the Five Domains Model, and uses numerical scores to facilitate analysis of a qualitative assessment.
The present welfare risk assessment process has demonstrated value by focusing specifically on risk assessments that can be conducted annually to help zoos prioritise areas for action and benchmark progress. Where resources are limited (funding, time, staff, etc.), it is important to be able to appropriately allocate resources in ways that will achieve the most productive results. To facilitate this, systematic processes that collate and analyse relevant information are important. As a result of this type of prioritisation, the three zoos involved in the present trial were able to develop effective welfare work plans and action a substantial number of welfare interventions as well as prioritise development of plans for large-scale enclosure renovations, in a process that places greatest priority on animal needs during decision-making. Furthermore, the ability to track progress over time is also of value. Benchmarking enables institutional welfare goals to be set and then targeted, which is an effective means of integrating animal welfare into the organisation and fostering a culture of continuous improvement. However, it is important that aspirational but realistic targets are set to achieve this outcome.
Specifically, as a result of this trial and the systematic analysis of welfare risks across the three zoos, efforts directed to advance welfare have been focused on four main areas of resource allocation, namely, complete enclosure redesign, in situ welfare interventions, research project investment and boosting staff capacity (development of new positions). Across the three zoos, eight enclosures have either been fully renovated or are in the process of redesign. A further 195 small-scale interventions were actioned including lighting changes, feeding regime adjustment, the addition of more heating, cooling and/or shelter, increase in enclosure furniture for behavioural opportunities, the provision of sound proof retreats, increase in visual barriers, increased enrichment provision, diet reviews, additional visitor barriers, the provision of remote monitoring equipment, and increased keeper time spent in training for proactive health care across species.
The process and its assessment outcomes have also proved useful for determining research priorities. Through the analysis of the ‘unknown’ scores across indicators/measures and across enclosures, gaps in understanding have been identified. This has led to the development of research projects aimed at addressing these gaps. As a result of the assessments and lack of information on the effects of certain environment conditions on animals, 14 research projects have been developed and completed or are presently underway across the three zoos. The majority of this research focuses on animal behaviour analyses to address the ‘unknowns’ in the behaviour domain; others are designed to investigate the impact of various conditions on the animals such as noise levels, lighting regimes, and visitor effects.
Lastly, investment in staff capacity building has resulted in several new positions being created across the three zoos including Animal Training Coordinators at each zoo and two new welfare research assistant positions. The data highlighting the ‘unknowns’ were used to demonstrate a business case to employ the research assistants to help conduct the applied projects described above.
Another indirect benefit of the use of this process has been the creation of an environment of open communication about animal welfare amongst zoo staff. The process enabled facilitated discussions to take place and provided opportunities for staff to raise welfare concerns in a systematic manner. As a result, this process is likely to have benefits for the quality of information gathered, but also benefits for staff education, awareness raising, and the fostering of a positive staff culture around animal welfare. This was not the primary aim of the process so no data were collected on whether or not demonstrable changes occurred, however managers at the three zoos involved anecdotally supported this notion.
4.3. Other Considerations and Limitations of Use
As highlighted above, the welfare assessment process provides zoos with the opportunity to conduct systematic risk assessments of animal welfare and develop evidence-based welfare work plans to prioritise welfare actions. This is an important part of a holistic animal welfare strategy for zoos, but it is important to note that this process should not be used as the sole means for evaluating animal welfare at an institute. Resource assessments can highlight the importance of certain ways of providing for the biological needs of a species, but it is also recognised that the provision of these factors alone does not necessarily translate to good welfare for the animal [
29,
32]. The inclusion of animal-based measures is therefore intended to present a qualitative, yet informative, assessment of how animals utilise the resources provided. However, this process is still predominantly resource-based (75% of the criteria), with a smaller proportion of animal-based measures (25%), and that is why it is referred to as a risk assessment process rather than a welfare assessment process. However as our knowledge improves, further animal-based measures can be incorporated into the process. Moreover, its use may be enhanced by supplementation with additional processes designed to provide more detail on individual animal welfare metrics, supported by applied research focused on behavioural and physiological changes in individuals.
This trial set the welfare threshold for prioritisation relatively high (at 60%), however this should be assessed on a case by case basis for use at other institutions based on their distribution of data. It is also important for benchmarking that the threshold does not shift considerably between years to allow meaningful comparison to occur over time. However, as evidenced in the present dataset, the spread from minimum-to-maximum scores can and would be expected to change, and the goal should ultimately be to increase the benchmark level as progress is achieved in order to foster continuous improvement. In addition to prioritizing the enclosures falling below the set threshold, institutions may also wish to set a standard to improve all enclosures scoring in the bottom 5 or 10%, as this should foster further continuous improvement in welfare regardless of any potential biases in the rating involved in the use of the process. Moreover, it is important that, from the outset, such a process is recognised as adaptable. We will continue to learn more about species needs and preferences and it is critical that these scientific advances inform our consideration of the standards against which we will gauge success for each species. Zoos need to strive for goals well beyond the legislated minimum standards.
Other considerations for the use of this process relate to the personnel and resources required and the potential for inherent bias in the judgements made. Firstly, annual assessment using the process can be resource intensive for zoos because of the number of people and time involved (keepers, veterinarians, managers, and a welfare specialist); the time required will depend on the number of enclosures being assessed. Large zoos with hundreds of enclosures will require a larger portion of time for data collection, analysis, and reporting, meaning annual assessments are likely to be most feasible. However smaller zoos with fewer enclosures may be in a position to conduct more frequent assessments, which has the benefit of providing more data over time which may fasten progress. In addition, the success of the process relies on an effective chair to enable balanced discussion and ensure all participants have the opportunity to meaningfully contribute. Judgements about the level of risk or welfare are necessarily qualitative; as such, the accuracy of such judgements is likely to vary with experience and education/training among the team of experts. There may also be a risk of judgements being too negative or too positive according to the perceptions and motivations of the scorers. For example, if there is a sense that judgement is being made about the level of care offered by the keepers, there may be a risk of a positive bias in scoring. In contrast, if the process is seen to be an effective way to receive resources, there may be risk of a negative bias. Again, this relies on an effective chair and/or welfare specialist to manage.
At a more general level, the lack of sector-wide agreed ‘positive’ welfare standards for many species, as well as a lack of information for some species regarding natural behavioural biology, can make requirements difficult to judge for some species; this predisposes towards assignment of the ‘unknown’ option. Related to this, the major focus of the assessment is on risk identification, i.e., identification and minimization of negative experiences and negative welfare states. This may focus users towards achieving neutral welfare states, rather than striving for positive states for animals. Again, identifying appropriate targets for the assessments relies both on welfare-specific education of scorers and on an effective chair to ensure that sufficient focus is placed on positive welfare.