Cynhyene

Population estimates drew upon the best available evidence for each area of resident range. Where on-the-ground monitoring projects were in place, these projects were asked to provide up-to-date subpopulation size estimates (numbers of packs, and numbers of animals aged ≥12 months), for their areas of operation, these estimates drew on a range of data sources including tracking known individuals and packs with GPS-collars, visual monitoring, camera trap surveys, tourist photos, spoor surveys, and extrapolation from intensive study areas to nearby areas. Where there was no alternative data source, subpopulation size estimates were derived by multiplying the area of the resident range polygon by a conservative estimate of adult population density, with the number of packs then estimated using a mean pack size of 10 adults (McNutt and Silk 2008, Woodroffe et al. 2019).

Mature individuals
For African wild dogs, the estimation of numbers of mature individuals is complicated by the species’ cooperative breeding behaviour, which means that many sexually mature individuals do not reproduce and hence do not contribute genetically to the next generation (Malcolm and Marten 1982, Creel et al. 1997, Girman et al. 1997a, McNutt and Silk 2008, Woodroffe et al. 2019). For this assessment, the number of mature individuals was estimated from the number of packs, by assuming that every pack contained a single alpha female and a single alpha male, while a proportion of packs also contained subdominant breeders in addition to the alphas. We used published data to calculate, separately, the average number of subdominant females, and the average number of subdominant males, reproducing each year, in addition to the alpha pair. Across seven study areas in Botswana, Kenya, Zimbabwe, Tanzania, and South Africa, behavioural observations indicated that 32 of 343 litters were born to subdominant mothers (Frame et al. 1979, Girman et al. 1997b, Creel and Creel 2002, Spiering et al. 2010, McNutt 2019). As there were 32 litters born to subdominants and 311 born to alpha females, we inferred that, on average, there were 32/311 (0.103) subdominant breeding females for every alpha female. This value compared well with available genetic evidence (from two of the study subpopulations, Girman et al. 1997b, Spiering et al. 2010) which showed that, in South Africa, 10 of 137 pups had subdominant mothers, indicating that subdominant females produced 0.079 pups for every pup born to an alpha female.

Male reproductive success can be measured only using genetics. In the same two genetic studies (Girman et al. 1997b, Spiering et al. 2010) 27 of 115 pups had subdominant fathers, indicating that subdominant males fathered 0.307 pups for every pup fathered by an alpha male.

On the basis of this evidence, we estimated the number of mature individuals per subpopulation from the number of packs as:

Mature individuals = Number of alpha females (= number of packs) + Number of subdominant breeding females (= number of packs x 0.103) + Number of alpha males (= number of packs) + Number of subdominant breeding males (= number of packs x 0.307).

The estimate for each subpopulation was then rounded to the nearest whole number, and summed across all subpopulations to give a global population estimate.

Trend in population size
The global estimate for the number of African wild dogs in the current assessment (2025: 1,676 mature individuals) exceeds the estimate for the previous assessment (2012: 1,409 mature individuals). However, this difference reflects two important changes, unrelated to real population trajectories. First, several additional wild dog subpopulations were discovered since the previous estimate, including those in south-eastern Angola (Funston et al. 2017), western Angola (Overton et al. 2020), and eastern parts of the Central African Republic (Aebischer et al. 2020). The addition of these subpopulations contributed to an increase in the global population estimate, but this change reflected a more accurate estimate rather than actual population growth.
Second, for this assessment we slightly altered our method for estimating number of mature individuals. In contrast with the previous (2012) assessment, we based our estimate on the numbers of packs per site, rather than the numbers of adults, both because pack number can be easier to estimate in the field, and because it is more meaningful as a measure of population size (Woodroffe, O’Neill and Rabaiotti 2019). Additionally, the number of genetic studies available to inform our estimates of the number of subdominant breeders per pack increased from one (Girman et al. 1997a) to two (Girmanet al. 1997a, Spiering et al. 2010). For females, the rate of subdominant reproduction reported in the newer study (7.0% of pups with subdominant mothers, Spiering et al. 2010) was similar to that from the older study (7.8% of pups with subdominant mothers, Girman et al. 1997a), and was also consistent with behavioural data from seven sites. In contrast, for males the rate of subdominant reproduction could only be measured from genetics (rather than from behaviour as for females), and was found to be much higher in the newer study (27.9% of pups with subdominant fathers, Spiering et al.2010) than in the older study (10.3% of pups with subdominant fathers, Girman et al. 1997a). Unlike the older study (Girman et al. 1997a), the newer study came from a small, isolated subpopulation (Spiering et al. 2010) where limited opportunities to disperse and form new packs may have encouraged subdominant reproduction, and it is not clear to what extent it was representative of wild dog subpopulations more generally. However, as both studies had small sample sizes, we chose to include them both to draw upon the widest possible array of data to inform our estimates. However, we are aware that, in so doing, we may have inflated the estimated number of mature individuals, the global total (1,676 mature individuals) would have been 8% lower (1,543 mature individuals) had we drawn only on the older study of subdominant paternity. Indeed, the global total would have been 17% lower (1,392 mature individuals) had we ignored subdominant reproduction, as in the 2001 assessment.

To avoid bias introduced by these methodological differences, we estimated recent population change by (i) restricting comparisons over time to sites with subpopulation size estimates in both 2012 and 2025, and (ii) comparing the estimated numbers of adults (animals aged ≥12 months) between the two time points, rather than the numbers of mature individuals, as this measure was estimated at both time points.

On this basis, we estimated a 17% decline over the past three generations (15 years). The comparison period includes the population crash in Kenya’s Ewaso ecosystem (Mutinda et al. 2017), loss of wild dogs from Zambia’s Liuwa Plain (Zambia Carnivore Programme 2016) and the WAP Complex in West Africa (IUCN/SSC in review), and multiple whole-pack deaths in South Africa (Du

Plessis 2016, Loots et al. 2017) and Tanzania (Grumeti Fund 2018). Across Africa, threats to wild dog populations are growing rather than declining, as human impacts on formerly wildlife-friendly habitats expand. As time passes, we expect there to be less habitat, fewer prey, more snares, more roads, and more domestic dogs and so more disease transmission. Facing this mass of threats, we expect wild dogs to continue the pattern of decline observed in recent years.

Extreme fluctuations
Wild dogs experience dramatic “boom and bust” population dynamics. Litter sizes are large, allowing populations to grow rapidly when conditions are favourable (e.g., Pole 2000, Woodroffe 2011a, Davies-Mostert et al. 2015). However, episodes of very high mortality are also relatively frequent, and are typically linked to disease outbreaks (e.g., Alexander et al. 2010, Goller et al. 2010). Since the last assessment there have been multiple episodes of this type (e.g., Du Plessis 2016, Zambia Carnivore Programme 2016, Grumeti Fund 2018, Loots et al. 2018, van Schalkwyk et al. 2019), including an epidemic of canine distemper virus which killed an estimated 20 packs in Kenya’s Laikipia County (part of the Ewaso ecosystem) within a three-month period in 2017 (Mutinda, Cook and Kinya 2017). For perspective, 23 of the 31 wild dog subpopulations remaining in Africa are estimated to number et al. 2000, Hofmeyr et al. 2004, Grumeti Fund 2018), as well as prompting two reintroductions to replace populations lost to disease (Masenga et al. 2017, African Parks 2021).

Historical accounts show that disease-related wild dog die-offs are not a recent phenomenon. The earliest accounts document the loss of wild dogs from what is now the northern part of South Africa’s Kruger National Park during an epidemic attributed to “pulmonary distemper” in 1920, thought to have been transmitted by domestic dogs and to have entered from neighbouring Zimbabwe (Stevenson-Hamilton 1939). This die-off in the northern part of Kruger was followed in 1926 by the emergence of another disease (thought to be rickettsiosis) which led to wild dogs’ disappearance from the southern part of the park by 1931 (Stevenson-Hamilton 1939). Wild dogs were reported to have recolonised from neighbouring Mozambique in 1934 (Stevenson-Hamilton 1947), but rapidly disappeared again, with population recovery only beginning in earnest in 1948 (Reich 1981). A similar pattern has been observed elsewhere. Hoier (1955), noted wild dogs’ sudden appearance in, and equally sudden disappearance from, Albert National Park (now Virunga National Park in DRC), and proposed infectious disease as the cause. Wild dogs disappeared from Kenya’s Laikipia County during the 1980s in association with a disease outbreak, before recolonising two decades later (Woodroffe 2011a). Likewise, wild dogs disappeared from the Serengeti-Mara ecosystem on the Kenya-Tanzania border in a disease outbreak in 1990-1 (Gascoyne et al. 1993, Alexander and Appel 1994), returning after a decade (Marsden et al. 2011). The disappearance of wild dogs from the northern sections of Kruger National Park in the late 1990s may also have been caused by disease (Endangered Wildlife Trust 2017), this subpopulation recovered recently, after remaining low for two decades (Nicholson et al. 2020).

These periods of very rapid disease-related decline, followed by prolonged absence and then recovery, are consistent with two other elements of African wild dogs’ life history. First, the very long-range movements of dispersers (Fuller et al. 1992, Davies-Mostert et al. 2012, Masenga et al. 2016, Cozzi et al. 2020, Woodroffe et al. 2020, Sandoval-Seres et al. 2022, Beytell et al. 2024) make it possible for them to detect and colonise (or recolonise) suitable habitat, even when it is located at some distance from resident populations. Second, wild dogs are capable of rapid population growth. Because litters are large, packs can easily double or even triple in size within a single breeding attempt (e.g., Mills 1993, Woodroffe et al. 2019) and, where environmental conditions are favourable, populations may grow rapidly. For example, wild dogs recolonising Kenya’s Laikipia County grew from one pack to 19 packs in eight years (Woodroffe 2011a), those establishing South Africa’s managed metapopulation grew from 17 individuals to 202 in seven years (Davies-Mostert et al. 2015), and two packs reintroduced to Mozambique’s Gorongosa National Park in 2018-9 (Bouley et al. 2021) had grown to 11 packs by late 2024 (Mercia Angela, pers. comm.). Other species with excellent dispersal abilities and high reproductive rates are adapted to exploit environmental conditions which are unpredictable and ephemeral (e.g., pioneer tree species which colonise and rapidly exploit gaps in rainforest canopy), and it may be the case that African wild dogs’ facility for (re)colonisation may indicate a similar adaptation. While the ephemeral resource to be exploited is likely to involve habitat with abundant prey and/or few competitors, one source of such habitat is land vacated due to disease mortality.

Wild dogs’ “boom and bust” life history is relevant to their conservation (and hence to their red list assessment) for three reasons. First, rapid die-offs due to disease have occurred in large populations (e.g., Kruger in 1920 and Laikipia in 2017) as well as small populations, and inside protected areas (e.g., Serengeti in 1991 and Kruger in 2016) as well as outside, showing that neither population size nor protected status is sufficient to avoid sudden die-offs. Accounting for this additional risk is an important element of evaluating red list status. Second, although long-distance dispersal would, in the past, have allowed wild dogs to recolonise habitat vacated following disease die-offs, the habitat fragmentation typical of today’s human-dominated landscape acts as a barrier to dispersal (Cozzi et al. 2013, Jackson et al. 2016, Cozzi et al. 2020, O'Neill et al. 2020, O’Neill et al. 2022), impeding natural recolonisation and thus intensifying the impact of infectious disease on metapopulation persistence. Third, even if natural or assisted dispersal can return wild dogs to sites devastated by disease, other anthropogenic threats (including human-wildlife conflict, prey loss, climate change, and other infectious diseases) are all likely to hinder subpopulation recovery.

In our view, this evidence of repeated disease-related die-offs, followed by natural recolonisation and recovery, makes the “extreme fluctuations” criterion appropriate to this species.