The influence of sociodemographic and environmental factors on wildlife carcass submissions in urban areas: Opportunities for increasing equitable and representative wildlife health surveillance

Healthy and sustainable wildlife populations are of significant environmental, economic, cultural, spiritual, and recreational importance. Wildlife health surveillance is an essential tool for understanding wildlife health, assessing public health risks (as they relate to wildlife and zoonotic disease), and informing policy decisions related to wildlife management and conservation (Belant and Deese 2010; Ryser-Degiorgis 2013). Surveillance is of particular importance in rapidly expanding urban areas, where wildlife live in close association with humans and face unique challenges and health risks compared to their non-urban counterparts (Murray et al. 2019). Even within urban areas, wildlife health and disease can vary across relatively small spatial scales according to host and environmental factors (e.g., variation in pathogen prevalence in Norway rats, Rattus norvegicus, between city blocks; Himsworth et al. 2013).

In Canada, the Canadian Wildlife Health Cooperative (CWHC) operates a national wildlife health surveillance programme that receives wild animal carcass submissions for diagnostic evaluation. The opportunistic reporting and submission of carcasses through the CWHC allows for data collection at spatial and temporal scales that would not otherwise be financially or logistically feasible. However, this method of data collection may be biased by environmental and sociodemographic factors that can impact whether surveillance outputs are representative of and relevant for the range of wildlife, human, and environmental elements in a given geographic area. Human sociodemographic variables are increasingly recognized for their relevance to wildlife health and citizen engagement. For example, previous research has demonstrated that socioeconomic heterogeneity can influence ecological processes, and both sociodemographic and environmental factors have been found to be associated with citizen engagement in reporting and surveillance related to wildlife (Pickett et al. 2017; Thomas-Bachli et al. 2020; Sánchez et al. 2021).

Owing to the nature of opportunistic surveillance, the majority of CWHC submissions come from ‘urban’ areas, resulting in ‘rural’ areas being commonly recognized as surveillance gaps. Urban areas, however, are not homogeneous (Cadenasso et al. 2007) and surveillance inputs may not be equally distributed across these heterogeneous environments. Urban areas can be evaluated at different scales, and different metrics may be used to define urban boundaries (e.g., city boundary, human population density, building density, land use; Moll et al. 2019); nevertheless, these metrics all tell us something different about the environment and different scales may be relevant depending on the proposed purpose (e.g., management, policy, outreach) and the wildlife species being studied (e.g., home range, distribution). As there is no standardized definition of the word ‘urban’ and it can be interpreted in different ways (French, Giacinti, Robinson et al. 2022), in this article we instead refer to ‘city’ – an administratively relevant boundary that encompasses a complex mosaic of environmental and sociodemographic factors.

It is important to understand the current gaps in the representativeness of wildlife surveillance data within cities in order to inform methodology – the CWHC database provides us with an opportunity to generate this baseline understanding. Our objective was to evaluate the spatial distribution of wildlife carcass submissions to CWHC within Ontario cities and measure the associations between submissions and sociodemographic and environmental factors at different spatial scales.

Records were retrieved for all wild animal carcasses submitted to the Ontario-Nunavut node of the CWHC between January 2015 and December 2019. Submissions with missing dates, location, or species information were excluded. The CWHC does not regularly receive fish specimens in this region; therefore, the small number of fish records were excluded from analysis.

We analysed these data using more than one administrative boundary, to identify factors that are common across scales, generate robust results, and increase the likelihood that the associations may be generalizable to other cities. Furthermore, sociodemographic information is not always available at comparable scales in all cities. Therefore, carcass submissions within city boundaries were mapped to visualize the distribution at two spatial scales that are of relevance for city planning and decision-making: ward and neighbourhood. An analysis at the ward level was conducted because it is a boundary for which census data are available and it is a standard administrative boundary that exists across cities allowing a multiple city analysis. Ward boundaries change over time based on population and other demographic changes. Neighbourhoods are an administratively defined boundary that provide social and demographic data at a smaller and more meaningful scale to assist local government and other community organizations with planning. Neighbourhood boundaries change very infrequently over time allowing for longitudinal comparison but are defined differently across municipalities; therefore, multiple city analysis may not be possible.

Our study presents an evaluation of the factors associated with carcass submissions to the CWHC within two major cities in Ontario, Canada at two different administratively relevant scales, which highlights some of the aspects that may limit public engagement in wildlife health surveillance. The following discussion focuses on those variables that were significantly associated with the rate of submissions at both administrative scales. Our findings suggest that natural and anthropogenic environmental (e.g., parkland and presence of a zoo) and social (e.g., low-income status) variables are associated with submissions at both the ward and neighbourhood levels. City-wide aggregate data are quite broad and may not capture important associations that emerge at finer scales; therefore, it was important to evaluate the data across smaller administrative boundaries within cities. Geographic areas within cities, however, are aggregated at multiple different spatial scales for planning, management, and electoral purposes, and cities may use different methods to divide their administrative boundaries spatially. Surveillance data should ideally be available and assessed at a spatial scale of relevance for wildlife health, city planning, and decision-making.

Geographic features of the environment influence the likelihood of detecting carcasses, human behaviour in these spaces, and suitability for wildlife populations. In our study, there were more samples submitted from areas with a higher proportion of parkland. Parkland may be more likely to have more open space (i.e., fewer visual obstacles), which may aid in the visual detection of carcasses. Additionally, how humans use these spaces, including activities like dog walking, may result in the discovery of animal carcasses in parkland. While parkland habitat may be more suitable and desirable for some species (Ordeñana et al. 2010), we know that some wild species are present in high densities in non-parkland (e.g., raccoons, Procyon lotor; Gross et al. 2012). Due to limitations related to sample size and the availability of wildlife distribution data, we were unable to control for species and density. It would also be useful to further refine the parkland variable to account for the quality and accessibility of parkland, from both human recreational and wildlife habitat perspectives.

When considering parkland, it is also important to be aware of observed inequalities in the distribution of parkland. Previous research has identified that greenspace is unequally distributed across urban landscapes (Statistics Canada 2021; Tooke et al. 2010). In particular, associations have been found between socioeconomic factors, including household income, and proximity and access to green spaces. A study conducted in three major cities in Canada (Montreal, Toronto, and Vancouver) reported that high income had the most consistent and strong positive correlation with vegetation (Tooke et al. 2010). Urban greenspace development is associated with a complex social and historical context; an examination of these factors, while important, was outside the scope of this paper. It is of note, however, that even after controlling for parkland, there was a negative association between the rate of carcass submissions and the proportion of low-income households. There are a number of potential factors that may underlie this association, including the distribution of services and funding related to wildlife in these areas, and/or how people view and value wildlife. Manfredo et al. (2020) found that socioeconomic factors (e.g., higher education and income) were associated with a shift in values from being human-centric towards equality between humans and wildlife. However, this is likely a very complex relationship and it is important to acknowledge this was an ecological study and, therefore, we must be cautious not to infer associations at the individual level based on group-level data (i.e., ecological fallacy).

A range of complex social factors may influence wildlife health surveillance programmes that rely on public participation; however, many are not easily measured by census data including: awareness of the service and wildlife health issues, a knowledge of the importance/relevance of the service, and motivations and barriers to participation. There is currently a limited focus on how these social factors influence wildlife surveillance programmes and how they may lead to bias in surveillance inputs and outputs. Future research that surveys people who have and have not participated in wildlife health surveillance can begin to build this knowledge base and may help to inform future surveillance methodology and messaging.

Wildlife surveillance programmes have made efforts to increase participation of certain public groups (e.g., engagement of hunters and trappers) and help address previously identified gaps in data (e.g., targeting remote vs. populated areas, and particular wildlife species). Similar methods may be useful to address the gaps in intra-city data; however, identifying opportunities for increasing equitable and representative wildlife health surveillance requires a thorough understanding of the influence of environmental and sociodemographic factors on surveillance inputs and outputs. Some examples of methods used by the CWHC include education, messaging, and the development of online tools to facilitate the submission process (i.e., online reporting tool; http://www.cwhc-rcsf.ca/report_and_submit.php). Additionally, targeted relationship building has been a successful approach. For example, collaboration with pest control companies to collect species that are generally underrepresented but of public health importance (e.g., Norway rats; Robinson et al. 2022). Our analysis highlights that this approach is also useful within cities. Based on our knowledge of longstanding relationships between zoo facilities and the CWHC, presence of a zoo in the ward/neighbourhood was included as a fixed effect in our models, and unsurprisingly those wards/neighbourhoods had a higher rate of submissions.

Our study highlights the importance of considering both social and environmental factors when evaluating whether wildlife surveillance data are representative of a given geographic area. The submission of wildlife carcasses in cities likely depends on a number of factors including: the distribution of services and funding related to wildlife, presence of carcasses, detection of carcasses, a willingness or ability to make submissions, but also a knowledge of the service and a knowledge of the importance/relevance of the service. In addition, human population size and geographic proximity to a submission centre may influence carcass submissions (e.g., Hamilton and Toronto are among the most populous cities in Ontario and geographically close to the Ontario node of the CWHC in Guelph, Ontario). An understanding of these relationships can build awareness of potential biases not only in the design and implementation of surveillance programmes but also in the data, analysis, interpretation, and communication of resulting information. Understanding these complex relationships is increasingly pertinent as urbanization continues to increase globally; thus, further research is required to better understand the patterns and processes affecting urban wildlife health throughout the world. As surveillance data may influence public health policy, wildlife management, and other wildlife-related decision-making, it is imperative we increase our ability to identify barriers that may prevent equal opportunity to participate in wildlife health surveillance and to benefit from the associated outputs.

We thank D. Campbell, B. Stevens, and Canadian Wildlife Health Cooperative staff and volunteers who assisted with data collection as well as Nuisance Wildlife Control Inc., other agencies, and private citizens who submitted carcasses to the Canadian Wildlife Health Cooperative. S.K.F., J.A.G., and S.J.R. were supported by the Ontario Veterinary College and the University of Guelph. S.K.F. was also supported by an Ontario Graduate Scholarship. J.A.G. was also supported by the Natural Sciences and Engineering Research Council.