Prevalence and drivers of poison use by South African commercial farmers and perceptions of alternative livestock protection measures

Abstract

The use of poison to eliminate predators is causing African vulture populations to collapse. To understand the prevalence and motivations of this practice we conducted an extensive survey with South African commercial farmers. Using a specialised questioning technique and ad hoc quantitative methods we found that an estimated 22% and 31% of farmers used poison over a 1-year and 5-year period, respectively. Poison use hotspots generally coincided with small stock farming areas. The strongest predictor of poison use was whether farmers believed the practice to be common amongst their peers. Our results suggest that farmers’ attitudes to vultures are primarily positive, and farmers are less likely to use poisons if they frequently encounter vultures on their farm. Overall, our findings provide an understanding on poison use that provides leverage points to change farmers’ behaviour and help avert the African vulture crisis and possible cascading ecosystem impacts.

Supplementary information

The online version of this article (10.1007/s13280-020-01461-2) contains supplementary material, which is available to authorized users.

Introduction

Under a human-managed world, changing human behaviour is essential to halt the current biodiversity crisis (Johnson et al. 2017). Broad-scale approaches, such as policy and regulation, may be the most efficient methods to establish pro-environmental practices (Gray and Shimshack 2011), but poor enforcement, especially in developing countries, often results in compliance being voluntary in practice (Rowcliffe et al. 2004). Changing human behaviour at the individual level is thus an important component of conservation efforts seeking to address such issues, but one that is rarely achieved.

Conservation initiatives targeting individual behaviours often rely on the fundamental premise that non-participation in pro-environmental behaviour is because of a lack of awareness (van der Ploeg et al. 2011). However, social and environmental psychology research suggests that information alone is insufficient in producing behavioural change (Abrahamse et al. 2005). Additional approaches that enhance the motivational foundation for pro-environmental behaviour are thus required (Osbaldiston and Schott 2012). To design and effectively implement such alternative approaches, research needs to move beyond the simple quantification of human environmental impacts, to examining the causative behaviours themselves. The realisation that human behaviour is a key component in effective conservation has been slow to penetrate the conservation science field and requires a larger research focus (Cowling 2014).

Environmentally destructive behaviours often take place within a human-wildlife conflict where wildlife cause a range of damages (Nyhus 2016). Responses to such damage can often be overly corrective and in some cases dramatic, such as bombing the offending animals (Cheke and El Hady Sidatt 2019). The use of poison is a widespread and popular method of retaliation or control to target almost any damage causing species (Ogada 2014). Beyond agricultural pests, the most common poisons used to target vertebrate species are agricultural pesticides, such as organophosphates and carbamates, because of their high availability (Ogada 2014). Their legitimate use and their use to target vertebrate species can have a range of both lethal and sub-lethal effects on non-target species, with cascading effects through the entire ecosystem (Ogada 2014).

The use of poison by livestock owners to kill predators that cause livestock losses is currently at the forefront of conservation discussions, specifically concerning its impacts on scavengers (Plaza et al. 2019). Being obligate scavengers, vultures are particularly vulnerable to poisoning and this is one of the main drivers of the population declines that have resulted in an African vulture crisis (Ogada et al. 2016). Currently, seven of the eleven African vulture species are listed as endangered or critically endangered (Ogada et al. 2016). The loss of vultures and their carcass disposal service could have cascading impacts on ecosystems and human wellbeing. Vulture declines have, for example, been linked to increases in mammalian scavengers, such as feral dogs (Canis lupus familiaris), which in turn may result in increases in human rabies infections (Markandya et al. 2008). Furthermore, this associated meso-predator release, may result in increased livestock depredation and further motivate the use of poisons.

Although poisoning predators is currently illegal in most countries, the practice remains widespread in 83% of African countries (Ogada 2014). This is because of the large economic cost associated with livestock depredation, estimated to range between USD 22–171 million in South Africa annually (Van Niekerk 2010; Statistics South Africa 2010). The consequent poison use is estimated to kill more than 500 000 wild animals every year in this country (Endangered Wildlife Trust 2006). Understandably, livestock owners want to avoid such economic losses and, despite some evidence to the contrary (McManus et al. 2015), lethal predator control methods are still viewed as more effective and cheaper than alternative methods (Scasta et al. 2017). Of the lethal options, poisoning is likely the least labour-intensive and is therefore likely to continue if perceptions of the risks associated to this practice are not changed.

Given the large livestock losses associated with predation in South Africa, it is surprising that there are relatively few comprehensive studies investigating livestock protection measures in the region (Kerley et al. 2018). Some information on the prevalence of specific livestock protection measures is available from surveys which used a direct questioning approach, but the authors of these studies highlighted that famers may be apprehensive to answer sensitive questions openly, out of fear of prosecution or confrontation (Badenhorst 2014). Estimating the prevalence of sensitive behaviours is notoriously difficult as social desirability and non-response bias may affect the results (Nuno et al. 2015). Therefore, alternative methods have recently been developed that provide respondents anonymity with regards to their admittance to participating in a sensitive behaviour (Nuno et al. 2015). Such techniques have been effectively used in elucidating the prevalence of illegal bushmeat trading (van Velden et al. 2020) and poison use (Santangeli et al. 2016).

Here, we used specialised social science techniques that ensure respondent anonymity to investigate poison use by livestock farmers in South Africa, a global hotspot for vulture conservation (Santangeli et al. 2019). We first aimed to provide a deep understanding of poisoning behaviour by determining its prevalence among commercial livestock farmers. Secondly, we aimed to identify the context and attitudes that are associated with poison use. Third, we explored the spatial patterns of poison use across the country. Lastly, we investigated farmer perceptions of alternative livestock protection methods that may explain the persistence of poison use. We discuss how our results may be leveraged to curb poison use and how the spatial distribution of poisoning may affect vulture conservation initiatives.

Materials and Methods

Protocol for data collection

To investigate the prevalence of poison use in predator control, we conducted a survey with South African commercial livestock farmers between April and September 2019. Because of South Africa’s political history, which enforced segregation on a racial basis, commercial farming areas tend to be spatially distinct from areas of communal farming. Previous studies have indicated that poisoning is more prevalent among commercial than communal farmers (Santangeli et al. 2016; Craig et al. 2018); because of this and logistical constraints, we focussed this study on commercial farmers. All interviews were conducted in person (by CWB). Commercial farmers were approached at either agricultural retail stores, livestock auctions or agriculture fairs. We restricted our study to South Africa and focussed our sampling efforts to within 100 km of the range of any relevant vulture species in South Africa (African white-backed vultures, Gyps africanus, Cape vulture, Gyps coprotheres, bearded vulture, Gypaetus barbatus, hooded vulture, Necrosyrtes monachus, white-headed vulture, Trigonoceps occipitalis) (Brink et al. 2020a). Interviews took 10 to 30 min to complete and were conducted in either English or Afrikaans, depending on the respondent’s preference. The overwhelming majority of respondents showed good familiarity with either language and we perceive no biases as a consequence of language comprehension in the study. Respondents either filled in the questionnaires by themselves or were read the questions and their answers transcribed.

General questions

We designed the questionnaire to include questions on factors that may be associated with a respondent’s propensity to use poison (Appendix S1). Questions related to basic demography (e.g. age and education level), farming context (e.g. location of farm, size of farm, type and number of livestock on farm, percentage income from livestock farming, depredation numbers), attitudes (e.g. towards predators, game, vultures and farmworkers) and the perception of the effectiveness of alternative predator control methods. Attitude and perception questions were framed on a five-point Likert scale of agreement or effectiveness (ranging from strongly disagree to strongly agree or from very ineffective to very effective). We also asked farmers to indicate the percentage of their peers they believed used poison to control predators. Farmers indicated the locations of their farms on a map and the coordinates were extracted.

The list experiment

Using poison to target predators is illegal in South Africa (Thompson and Blackmore 2020). Therefore, we used an indirect questioning technique that provides anonymity to respondents and does not require them to directly admit to the illegal behaviour, thereby reducing biases inherent in direct questioning (Nuno et al. 2015). The technique used here, referred to as either the unmatched count technique or a list experiment, has been successfully used to quantify illegal hunting prevalence in African communities (Whytock et al. 2018; van Velden et al. 2020). Respondents were randomly assigned to either a treatment or a control group. Each respondent was presented with a list of behaviours containing four non-sensitive behaviours, but in the case of the treatment group a fifth behaviour, the sensitive behaviour under investigation (poisoning), was added to the list (Appendix S1). Respondents were asked to indicate how many of the listed behaviours had been performed on their farm in the last year, and the last 5 years, without indicating the behaviours themselves. This resulted in two datasets, one for poison use over a 1-year period and another for a 5-year period. All behaviours listed were related to farming practices and one very common and one very rare behaviour was included as is the advised experimental design to avoid ceiling and floor effects (Blair et al. 2019). The non-sensitive behaviours, or the control items, were the use of protective collars on livestock (rare behaviour), infrastructure such as fences (common behaviour), hunting of predators and using flock guarding animals such as Anatolian dogs.

Statistical analyses

Data analyses were done in R v.3.6.3 (R Core Team 2019), using the “list” package, which was specifically designed for analysing list experiments (Blair and Imai 2012). Data analyses were performed using the modelling framework provided by Blair and Imai (2012), which performs better than previous multivariate regression analysis approaches used for list experiments (Blair and Imai 2012). Within this framework we used the proposed Non-linear Least Squares (NLS) estimator and tested our models for assumption inherent to list experiments. The 1-year and the 5-year timeframes were analysed separately. Please refer to Appendix S2 for a more detailed description of the statistical approach.

All predictors were tested for collinearity and only largely uncorrelated variables (r < 0.4) were used in our analysis. Below we provide a description of the 15 variables included in the NLS models and the rationale for their inclusion. A subset of variables pertaining to farming context was included on the assumption that this context would influence a farmer’s ability to protect his livestock from predators. These included farm size, and number and type of livestock. Stock type was included as small stock (sheep and goats) have a higher propensity to be predated in South African farmland where the main threat is medium sized predators. Poison use was also strongly related to small stock farming in a previous study (Santangeli et al. 2016). Game in the context of this study refers to medium to large wild ungulates.

High predation rates or the perception that predation is one of the main causes of livestock losses is intuitively expected to trigger predator control. Similarly, farmers with negative attitudes to predators or game were expected to more readily use poison, owing to the potential for these animals to cause financial damages (e.g. predation, crop damage). We expected that poisoning would be more prevalent amongst older farmers who started farming in an era when this behaviour was more widespread and acceptable (Ogada 2014). We assumed that a higher level of education would correlate with increased environmental awareness. We expected farmers to generally be aware that poisoning threatens vultures and thus would be more apprehensive to use poison if they saw vultures regularly on their farms. Similarly, positive attitudes towards vultures were expected to reduce propensity for poison use. Strained relationships between farmers and their workers have been reported to result in unmotivated workers and in some cases vengeful behaviour that can exacerbate livestock losses (Rust et al. 2016). We therefore included a variable gauging farmer relationships with their workers. Lastly, we expected that farmers who perceived a higher prevalence of poison use in their community were more likely to use poison themselves because of the false consensus effect, whereby people are likely to overestimate the prevalence of a behaviour that they participate in (Deutsch 1989). All variables were used as continuous variables except for education level and the main cause of livestock losses which were added as categorical variables with two levels (tertiary education/no tertiary education, predators/other).

Poisoning probability estimates were derived for each respondent and interpolated using Inverse Distance Weighting to show the spatial prevalence of poison use across South Africa, following Santangeli et al. (2016) and Craig et al. (2018).

Results

Respondents characteristics, farming context and farmers attitudes

Of all 1411 people approached 65% indicated that they were livestock farmers and 90% (n = 823) of them participated in the study, nine were excluded because they did not fit the criteria of commercial livestock farmers. Of the 814 respondents that participated, 98% were male, largely Afrikaans (78%), followed by English (12%), the rest were of other less represented groups (Appendix S1). Respondent age was 50 on average (± 14 standard deviation), with 58% having tertiary education.

Respondents largely (76%) were full-time farmers, and 52% earned over 60% of their income from livestock farming. Average farm size was 2773 ha with a mean of 1080 livestock, largely cattle (81%), followed by sheep (56%), goats (16%), game (15%), pigs (3%), chickens (2%) and lastly horses and donkeys (1%). Of the commonly farmed animals, game was the most predated, with an average of 6.7% of stocks being lost to predators, followed by sheep (4.4% of stocks), goats (2.7% of stocks) and cattle (0.5% of stocks). Overall mean predation rate was higher for small stock (sheep and goats; 4.1%) than large stock (cattle, game, horses and donkeys; 3.7%).

Farmers generally reported a positive relationship with their farmworkers (94%), favoured game on their farms (87%) but disliked predators (58%, Appendix S8). Attitudes towards vultures were broadly positive, 84% of respondents wanted vultures on their farm and only 6% of respondents believed that vultures kill livestock (Appendix S8). Although most respondents (57%) agreed that vultures do not spread disease, many (31%) had no opinion on this issue (Appendix S8).

Prevalence and correlates of poison use

Poison use prevalence across the whole sample of 814 respondents was 21.9% (95% CI 11.6–32.2%) with reference to the 1-year period, and 30.8% (95% CI 19.7–41.8%) with reference to the 5 years period. Few respondents (6.6%) admitted directly to using poison. The most common poisons used were pesticides, specifically aldicarb (Appendix S9). Farmers’ estimate of poison use by their neighbours (direct question) was 13.5% on average (95% CI 12.0–15.0%).

Model results indicated that various factors were associated with poison use (Fig. 1, Appendix S10 and S11). Poison use was highest for farmers who perceived this practice to be common among neighbours, for farmers with a positive attitude towards vultures and those who own larger numbers of small stock (Fig. 2). Other factors positively related to poison use were tertiary education, farmer age, proportion of livestock predated, predators being the main cause of livestock loss and negative attitude to predators. Poison use was also lower in areas where vultures were seen more frequently (Fig. 1).

Fig. 1.

Factors associated to the use of poison by South African commercial livestock farmers over a 1-year (blue) and 5-year period (red) in the present study. Variable coefficients (dots) and standard error (lines) are derived from multivariate regression models using a non-linear least squares estimator

Fig. 2.

Partial dependence effects plots showing the association of different factors (x-axis) to the probability of poison use among South African commercial livestock farmers over a 1-year (solid line) and a 5-year period (dashed line) in the present study. All variables have been logged and scaled. Variable effects displayed include farmer perceptions of the prevalence of poisoning under their neighbours (a), attitudes towards vultures (b), number of small stock (c), proportion of livestock predated (d), attitude towards predators (e) and frequency of vulture sightings (f)

Spatial distribution of poison use

Predicted poison use was highest in the Eastern Cape, central Free State and the eastern section of the Northern Cape and was lowest in the northern half of the country (Fig. 3, Appendix S12). While poisoning prevalence patterns in space are broadly consistent for the 1 and 5-year periods, poison use was much more widespread when assessed over a 5-year period. Farmer perceptions of poisoning prevalence were generally lower than predicted by our model but similar in spatial distribution (Appendix S13).

Fig. 3.

Interpolated predicted probabilities of poison use over a a 1-year period, and b a 5-year period, from a multivariate regression analysis of a list experiment involving South African commercial livestock farmers. Poisoning probability is indicated by the colour shading, while protected areas and land under traditional authority (communal farmland) are in grey. Distribution of sampling points (bottom right map) is indicated for both the treatment (red dots) and control group (blue dots). South African provinces are denoted (bottom centre) as WC  Western Cape, NC Northern Cape, NW North West, LP Limpopo, MP Mpumalanga, GT Gauteng, FS Free State, KZN KwaZulu-Natal, and EC Eastern Cape (LS indicates the country Lesotho). To assist visualisation and interpretation the colour ramp was scaled to include 2–98% of the data values

Perceptions of alternative predator control methods

One of four respondents (73.6%) believed infrastructures, such as fences and enclosures, are most effective in reducing livestock depredation (Fig. 4). Lethal control, guarding animals and removing problem individuals were also regarded as effective by half of respondents, while other methods where broadly considered as less effective (Fig. 4).

Fig. 4.

Farmer perceptions of the effectiveness of various livestock depredation avoidance methods. The percentage of farmers listing each method on a Likert scale varying from very ineffective to very effective is indicated

Discussion

Here, we use a specialised survey technique to characterise, quantify and map poison use by commercial farmers across South Africa. We found that more than one in five farmers used poison over a 1-year period and almost a third of farmers in a 5-year period. This practice was positively associated with small stock farming and with the social environment, i.e. higher poison use prevalence among neighbours favours this practice at the individual level. Poison use was negatively associated with the presence of vultures. The continued widespread belief in the effectiveness of lethal control methods, over alternative livestock predation avoidance methods, likely contributes to the continued prevalence of poison use in predator control.

Poison use prevalence

Although poison use has likely declined in the past 30 years, it remains a relatively common practice (Nattrass and Conradie 2018). Interestingly, our 5-year estimate was very similar to the 34% poisoning prevalence recorded around 1988 among farmers in the southern and central Drakensberg region (Brown and Piper 1988). This estimate, however, originates from a mail survey which used direct questioning, and therefore, likely underestimated poison use. Poison use seems to have shifted away from strychnine, used by 54% of poison users in the Brown and Piper (1988) study, and was only mentioned by a single self-proclaimed poison user in our study. Although few respondents indicated which poisons they used, more readily available pesticides such as carbamates were mentioned most in the present study (Appendix S9). Over a 1-year period our poisoning prevalence estimate (22%) was very similar to the 20% poisoning prevalence recorded for Namibian commercial farmers using a similar questioning technique (Santangeli et al. 2016).

The additional 10% increase in poisoning prevalence between our 1-year and 5-year estimates highlights the importance of incorporating longer time scales in such surveys. Poison use is reactionary and if the specific conditions, such as a particularly problematic predator, were not present in the immediate past, then poisoning is unlikely to be recorded. Thus, this longer time scale may provide a more accurate representation of the spatial distribution of the threat of poison in the landscape.

Correlates of poison use

Of the factors associated with poison use, farmers perceptions of the prevalence of poison use in their community was the strongest predictor. Farmers may thus be more likely to use poison if they believe the practice to be locally common. Alternatively, farmers that use poison may overestimate the number of their peers that use poison as a self-justification for their own behaviour (Deutsch 1989).

In our study, as in similar studies (Brown and Piper 1988; Santangeli et al. 2016), small stock farming was more closely associated with poison use than other types of livestock farming. Small stock, because of their size, are more prone to predation by medium sized and widespread predators (e.g. black-backed jackal, Canis mesomelas, and caracal, Caracal caracal), which are responsible for the majority of livestock losses in South Africa (Van Niekerk 2010). Our results also provide evidence that poison use increases with the proportion of livestock that is predated and a negative attitude towards predators. This suggests that the intensity of human-carnivore conflict triggers the use of poison.

The relationship between a positive attitude towards vultures and poisoning is counter intuitive. One explanation for this may be that farmers who use poison may try to compensate for their detrimental behaviour by showing a positive attitude towards vultures. Interestingly, as for Namibian commercial farmers (Santangeli et al. 2016), farmers listed a range of measures they use to avoid mortalities in non-target species when using poison. These include using poison at night only, with small parcels of meat, and placed in strategic locations which are perceived to be less accessible to non-target species. These measures are still associated with high risks to non-target species (Santangeli and Arkumarev, pers. comm.).

Older farmers were more likely to use poison. This finding may be explained by younger generations being more environmentally conscious (Pérez Urdiales et al. 2016) or because older farmers, as some have indicated, are no longer able, or keen, to implement more labour intensive means of predator control, such as shooting, thus reverting to poison use.

Farmers who frequently observe vultures on their farms are less likely to use poison, likely because they are aware of its potential impacts on vultures. However, this result may also underscore a potential risk to vultures in areas where they are rarer and where farmers may feel a false sense of complacency for using poison.

Spatial distribution of poisoning

Predicted poisoning prevalence was highest in the southern parts of South Africa where small stock farming is most common (Van Niekerk 2010). Because of this spatial pattern, the species at highest risk to poisoning are the Critically Endangered white-backed vulture, the Endangered lappet-faced and Cape vulture, and the only extant bearded vulture population of southern Africa (SABAP2 2007). Additionally, the high poison use in the Eastern Cape represents an important threat for a planned bearded vulture reintroduction, two sites in this region have recently been identified as candidates for this reintroduction (Brink et al. 2020b).

Poisoning and attitudes to alternative predator control

There was widespread support among farmers for the effectiveness of fences, enclosures and other infrastructures in keeping animals safe from depredation. Denying predators access to livestock would present a simple solution for this human-wildlife conflict. Unfortunately, the large financial investment required to implement this measure effectively makes it unattainable for many farmers (Landman 2016).

Among other means of predator control, lethal methods are still perceived to be more effective than non-lethal ones. This perception likely supported the persistence of poison use as reported in this study. Encouragingly, half of the farmers supported a problem individual approach rather than an indiscriminate one, likely stemming from the shift from the “vermin” or “pest species” mentality towards the idea that specific individuals disproportionately cause depredation (Linnell et al. 2008; Swan et al. 2017). Evidence suggests that selective predator management strategies can be effective (Swan et al. 2017). This strategy likely promotes co-existence between livestock farmers and predators which may reduce the use of poison.

Non-lethal predator control methods, such as translocation, did not carry much favour among farmers. Translocation of problem animals has not been very successful in the past (Linnell et al. 1997), which may explain the farmers’ scepticism for this method. Guarding animals (Marker et al. 2005; Rust et al. 2013) and deterrents (Miller et al. 2016b) have been shown to effectively reduce livestock predation. However, these means require expertise, financial resources, and are logistically challenging, factors adding to the risk of failure (Rust et al. 2013).

Financial compensation when administered appropriately can be an effective measure for reducing retaliatory killings of predators (Miller et al. 2016a) but fails when administered ineffectively (Karanth et al. 2012) because of resulting growing intolerance in uncompensated livestock owners (Karanth et al. 2013). A lack of faith in the ability of national government to employ such schemes was a commonly communicated sentiment, with regards to this question, and explains why farmers believed that financial compensation would not be effective.

Conservation solutions and behaviour change

The vast majority of interviewed farmers of this study had positive attitudes towards vultures. South African farmers, as also those from Namibia, in general seem to have an appreciation for the cleaning services provided by vultures and many make use of them for carcass disposal (Santangeli et al. 2016; Brink et al. 2020c). Conservation work with these farmers may thus yield positive outcomes, i.e. aid the reduction of poison use, by increasing their awareness of vultures’ declines stemming from poisoning, and by suggesting alternative predator control means.

Among those farmers who use poison, many perceived they could do it in a responsible and selective way. It is thus important to make those farmers aware of the high potential risks that poison use entails for any species in the surrounding landscape (humans, pets and livestock included), irrespective of the way it is used. This could be done by running awareness campaigns particularly in those areas of the southern half of South Africa which we identified as having a high prevalence of farmers using poison.

There is currently a large degree of distrust between farmers and conservation agencies in South Africa. Thus, a possible channel to work with farmers on issues related to poison use could be through trusted bodies that are close to farmers interests, such as farmers associations and support bodies. This approach has previously proved to be beneficial in conflicting situations between landowners and nature conservation (Santangeli et al. 2012).

Conclusions

An understanding of the motivations and perceptions leading to environmentally destructive behaviours is still largely lacking. Economic gain is often understood as the only relevant component and leverage points for behavioural change therefore remain undiscovered. This is particularly salient in the context of this study, which suggests that farmers who use poison may in general have a positive disposition to vultures and likely lack awareness of, or underestimate, the risks associated with their behaviour. For commercial farmland to become viable vulture habitat, guidance from conservationists may be needed, but a bottom up approach with farmers as key stakeholders is a necessity (Redpath et al. 2013). Any campaign aiming to address poison use prevalence should incorporate the environmental sensibilities uncovered in this study. Central to such a campaign would be convincing farmers of the risks that this behaviour, in all its forms, poses to highly endangered avian scavengers, the potential effectiveness of alternative methods if correctly applied and the potential ecosystem disruptions caused by the loss of vultures.

Our study suggests that the presence of vultures represent a strong deterrent for farmers to minimise the use of poison. If vultures are lost, there may be no constraint for farmers in using poisons, with catastrophic consequences on a wealth of other scavenger and non-scavenger species. This may ultimately compromise the ecosystem balance to the detriment of wildlife and humans alike (Markandya et al. 2008).

Supplementary information

Below is the link to the electronic supplementary material.

Biographies

Christiaan Willem Brink

is a Ph.D. candidate at the FitzPatrick Institute of African Ornithology. His main research interest is in identifying and assessing the effectiveness of conservation actions.

Robert Leslie Thomson

is a Senior Lecturer at the FitzPatrick Institute of African Ornithology. His main research falls within behavioural and community ecology but he is expanding into more applied themes in conservation biology.

Arjun Amar

is an Associate Professor at the FitzPatrick Institute of African Ornithology. His research interests lie in understanding the processes that regulate animal distributions, demography and population dynamics, and applying this understanding to the conservation of declining populations.

Marco Girardello

is based at the Azorean Biodiversity Group, Portugal. His research interests include macroecology, biogeography and conservation biology.

Andrea Santangeli

is a postdoctoral fellow at the Finnish Museum of Natural History. His research interests include applied ecology and conservation science with particular focus on bottom-up approaches to solve conservation issues.

Compliance with ethical standards

Ethical approval

Informed consent was provided before each interview. To ensure anonymity of respondents this consent was obtained verbally. This study was approved by the Faculty of Science Research Ethics Committee at the University of Cape Town (Approval code: FSREC 19—2019).