A giraffe that escapes a wire snare may carry the evidence on its leg long after the trap is gone. That makes surviving animals a possible source of information about a threat that is difficult to observe directly. It also creates a problem: the giraffes available to be photographed are precisely the ones that survived long enough to be seen.
In Uganda’s Murchison Falls National Park, Herbert Kasozi and colleagues photographed 26 giraffes with snare injuries during surveys in 2019. Their population model estimated that visible injuries affected about 1.4% of adults and subadults in the study area. The injuries were not spread evenly across the landscape, and injured animals had different encounter and movement patterns from animals without visible injuries.
The study, published online in 2022 and in the 2023 issue of the Journal of Zoology, showed how an identification survey could reveal where surviving giraffes bore the consequences of snaring. It did not count every snaring event or establish that the overall impact on the population was small.
The animals left outside a poaching tally
A record of carcasses or confiscated animals captures one part of illegal hunting. Injured animals that break free belong to a different category. They may remain alive and mobile, with their later fate unknown.
The researchers investigated this category in the northern part of Murchison Falls, where the park’s giraffe population was concentrated. They interpreted the leg injuries as consequences of non-target snaring: giraffes caught in traps associated with hunting other animals. That interpretation drew on the location of the injuries and the local context described in the paper. The survey itself did not observe the intention behind each individual trapping event.
All 26 injured giraffes had an affected leg, and 24 had injuries on a hind leg. The recorded injuries included wounds, swelling and severe damage. They were visible signs of contact with snares, not a standardized measure of how much each animal’s feeding, reproduction or survival had subsequently changed.
Recognizing this distinction kept the research focused. Its immediate questions were how common such visible injuries were among the living population, which animals carried them, and how their distribution varied across the study area.
Repeated photographs made the estimate possible
Between June and August 2019, teams surveyed five fixed road routes, repeating each 10 times. They photographed encountered giraffes, recorded age class and sex, and noted whether a snaring injury was present. Coat-pattern matching then linked repeated photographs of the same individual.
The resulting catalogue contained 1,402 giraffes: 897 adults, 409 subadults and 96 calves. Of the 26 injured animals, 22 were male and four were female. One of the injured males was a calf.
Those observations were not treated as a complete census. A giraffe that moved widely or spent more time near the survey routes could appear more often than another animal. The researchers therefore fitted a spatial capture–recapture model, which used both identities and encounter locations to estimate abundance and observation probability together.
In ordinary terms, the model inferred an activity centre for each animal and described how the chance of encountering it declined with distance from that centre. It allowed the shape of that relationship to differ with sex, age and injury status. This gave the analysis a way to distinguish the population from the subset that happened to be photographed.
Calves were excluded from the model because their encounters depended on their mothers’ locations. Treating them as independent would conflict with the model’s assumptions. Consequently, the modelled population and injury percentage refer to adults and subadults, even though the field catalogue and the total of 26 injured animals include calves.
That difference in denominators matters. Dividing 26 by 1,402 does not reproduce the model’s 1.4% estimate, and neither quantity is an annual rate of new injuries.
What the 1.4% actually describes
The spatial model estimated 1,939 adults and subadults, with a 95% credible interval of 1,864–2,023. The estimated proportion bearing a visible snaring injury was 1.4%, with a credible interval of 0.9–2.0%.
This was a prevalence estimate: the proportion of living animals carrying a recognizable condition during the study. It was not the probability that a giraffe would encounter a snare during the next year. The researchers did not know when every observed injury had occurred, how long injuries remained recognizable, or how many previously injured giraffes had died before the surveys.
A visible-injury percentage can therefore reflect several processes at once. More encounters with snares could increase it. Faster deaths among injured animals could reduce the number of injured survivors available to be photographed. Injuries that healed or became hard to recognize could disappear from the visible record. The study did not separately estimate these processes.
This is why a low prevalence does not establish a low total death toll. It can still be useful as a repeatable observation, particularly when paired with other evidence, but it must be interpreted as a property of the observed surviving population.
The statistical uncertainty also had a practical qualification. Model diagnostics indicated some lack of fit, potentially causing uncertainty to be understated. The authors suggested that unmodelled differences among individual encounter probabilities could be responsible. Their estimates are informative approximations rather than an exact census with every source of error removed.
Movement and injury were associated
Most injured giraffes were males. The model supplied some evidence that adult males had a greater movement scale than adult females, while subadults generally had smaller movement scales than adults. Injured giraffes also had higher estimated encounter rates and movement scales than animals without visible injuries.
A movement scale here is a model parameter describing the spatial spread of encounters around an inferred activity centre. It is not a directly measured daily walking distance, a GPS track or a complete home range.
The association offers a plausible explanation for exposure: an animal ranging more widely may encounter more opportunities for snaring. But the study could not determine the direction of causation. Injury might change how an animal uses space or how readily observers encounter it. Differences between animals that survived injury and those that did not could also affect the observed pattern.
The authors explicitly acknowledged that fewer injured females might reflect lower exposure, different mortality before detection, or other processes. It would therefore be premature to convert the male majority into a claim that female giraffes were safely insulated from snaring.
Resolving this would require following individuals before and after an injury, rather than comparing injury categories during a short survey. Such work could distinguish traits associated with exposure from changes that followed the injury itself.
The spatial pattern carried another message
The model predicted a concentration of injured giraffes in a river-associated part of the study landscape. That pattern coincided with high giraffe density and with areas the authors discussed as affected by human access and snaring pressure.
Its management value was different from that of the park-wide percentage. An overall average can make a localized threat appear diffuse. A spatial analysis can instead identify where animals bearing evidence of that threat are concentrated, helping conservation teams decide where further investigation and protection may be most useful.
The map still represented the locations of injured survivors and their inferred activity centres. It did not locate each trap that caused an injury. A giraffe could move after becoming entangled, and areas with more giraffes naturally offered more opportunities to encounter injured animals. The researchers’ model accounted for spatial encounter patterns, but it could not retrospectively observe the underlying injury event.
Season also constrained the interpretation. The fieldwork covered one short period that the authors treated as dry-season sampling. The study did not test whether injury distributions, movements or exposure changed across the full annual cycle. A management interpretation extending into other seasons would benefit from repeated observation.
Population consequences require a different measurement
A population model from Cameroon’s Bénoué National Park illustrates why the missing mortality information matters. In that much smaller modelled population, adding occasional adult deaths substantially increased simulated extinction risk. Its outcomes depended on assumed reproduction, survival and isolation.
That study cannot convert Murchison Falls’ 1.4% injury prevalence into an extinction probability. The populations, evidence and quantities differ: one study estimated visible injury among survivors; the other imposed adult removals on a simulated population. Their useful connection is the question between them. To assess the demographic consequences of snaring, researchers need to establish how exposure changes survival and reproduction, not merely how many visible wounds occur.
The Murchison Falls paper raised possible effects on movement, foraging and mating, but did not measure those outcomes. It also did not follow the injured cohort long enough to establish its eventual survival relative to uninjured animals. Our Bénoué modelling commentary explains why those individual outcomes can matter disproportionately when populations are small.
The next useful evidence would connect repeated injury assessments with individual survival, reproductive histories and broader seasonal coverage. Combined with independent records of snaring and removals, that could distinguish the location of surviving injured giraffes from the location and full consequences of exposure.
The 2019 surveys made a largely overlooked group visible: animals that had escaped but still carried an injury. Their distribution offered a warning that a single population-wide percentage could conceal. Understanding the complete cost of snaring requires following what happens to those survivors—and accounting for animals that never reach the camera.
Sources
Kasozi, Herbert; Linden, Daniel W.; Roloff, Gary J.; and Montgomery, Robert A. (2023; published online 2022). “Evaluating the prevalence and spatial distribution of giraffes injured by non-target poaching.” Journal of Zoology, 319, 152–162. Paper and DOI. Accepted full manuscript.
Colston, Kane P. J.; Johnson, Caspian L.; Nyugha, Denis; Mengamenya Goué, Achile; and Penny, Samuel G. (2023). “Viability analysis of Kordofan giraffe (Giraffa camelopardalis antiquorum) in a protected area in Cameroon.” African Journal of Ecology, 61, 929–944. Paper and DOI.
