# Evaluating persistence and its predictors in a West African carnivore community

> Where have the lions gone? Using cameras, tracks and scat, call-in surveys, interviews and 40 years of ranger patrol records, researchers found only 9 of 16 carnivore species still present in Ghana’s Mole National Park. Surprisingly, size and other traits did not predict which species vanished.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2011-burton-sam-evaluating-persistence-predictors-west/

## Citation and access

- **Authors:** A. Cole Burton, Moses K. Sam, David G. Kpelle, Cletus Balangtaa, Eric B. Buedi and Justin S. Brashares
- **Year published:** 2011
- **Journal or source:** Biological Conservation
- **Full citation:** Burton, A. C., Sam, M. K., Kpelle, D. G., Balangtaa, C., Buedi, E. B. and Brashares, J. S. 2011. Evaluating persistence and its predictors in a West African carnivore community. Biological Conservation 144(9): 2344–2353.
- **DOI:** 10.1016/j.biocon.2011.06.014
- **Access status:** Publisher link only
- **Where to find it:** https://doi.org/10.1016/j.biocon.2011.06.014

## Tags

- **Species:** Mammals > African civet; Mammals > African wild dog; Mammals > Caracal; Mammals > Gambian mongoose; Mammals > Hausa genet; Mammals > Honey badger; Mammals > Large grey mongoose; Mammals > Large-spotted genet; Mammals > Leopard; Mammals > Lion; Mammals > Marsh mongoose; Mammals > Serval; Mammals > Side-striped jackal; Mammals > Slender mongoose; Mammals > Spotted hyena; Mammals > White-tailed mongoose
- **Region:** Africa > Ghana
- **Tracking:** Scat; Sign surveys; Tracking; Tracks; Tracks and signs identification
- **Methods and technology:** Bayesian statistics; Bootstrapping; Camera traps; Detectability; Interviews; Occupancy modelling; Population trends
- **Data type:** Existing data; New field data
- **Observer qualification:** Not reported
- **Discipline:** Behaviour and ecology; Population trends; Wildlife tracking
- **Year published:** 2011

## Detailed keywords

- **Species:** Studied: the historical carnivore community of 16 species. Detected by camera traps (9): spotted hyena (Crocuta crocuta), white-tailed mongoose (Ichneumia albicauda), leopard (Panthera pardus), large-spotted genet (Genetta pardina), African civet (Civettictis civetta), caracal (Caracal caracal), marsh mongoose (Atilax paludinosus), Gambian mongoose (Mungos gambianus) and side-striped jackal (Canis adustus). Not detected and presumed extremely rare or gone from the park (7): lion (Panthera leo), African wild dog (Lycaon pictus), honey badger (Mellivora capensis), serval (Leptailurus serval), large grey mongoose (Herpestes ichneumon), slender mongoose (Herpestes sanguineus, also known as Galerella sanguinea) and Hausa genet (Genetta thierryi). Also mentioned: olive baboon (Papio anubis), as a possible case of “mesopredator release” in other Ghanaian parks; warthog and buffalo distress calls were used as call-in lures.
- **Location and habitat:** Mole National Park (about 4600 km²), Northern Region, Ghana, West Africa: open savanna woodland, mean annual rainfall 1100 mm with over 90% in the April to October wet season, mostly seasonal rivers. About 30,000 people live in 29 villages within 10 km of the park boundary; widespread illegal hunting. The far north of the park was not sampled.
- **Topics:** carnivore persistence and local extinction (extirpation); extinction vulnerability; community disassembly; imperfect detection and false absence; life-history and ecological traits (body mass, home range, gestation, weaning age, density, group size, diet and habitat breadth, activity period, range size); human threats (bushmeat hunting, use of carnivore products, retaliatory killing over livestock, gin or leg-hold traps); law enforcement patrol monitoring; long-term decline of lion and leopard; possible competition between spotted hyenas and lions; trophic cascades and mesopredator release; evidentiary standards for occurrence data.
- **Statistics or models:** Hierarchical multi-species site-occupancy model (Royle and Dorazio framework) fitted with Bayesian methods in R 2.10.1 and WinBUGS 1.4.3, with a “supercommunity” of 20 species; naive occupancy and detections per 100 trap-days; classification and regression trees (R package tree); random forests (R package randomForest, 5000 trees from bootstrap samples); Fisher’s exact test; Spearman’s rank correlation; Wilcoxon rank-sum test; binomial test; patrol catch-per-unit-effort index (individuals per 100 patrols) with lowess trend lines. Possible inconsistencies: the text says historical patrol reports held records of 11 carnivore species, but Table 2 shows historical patrol evidence for 8 (the footnote says ambiguous “genet” and “mongoose” records were left out of the table); and occurrence and detection probabilities are reported as correlated with rs = 0.97 but P = 0.043, while the same rs elsewhere gives P < 0.001.
- **Data type:** New field data: camera trapping at 253 stations (5469 trap-days), sign surveys (330 km on foot, 1400 km by vehicle) and village interviews, October 2006 to January 2009; five call-in surveys, 2 March to 5 April 2007; three night spotlight transects. Existing data: park patrol records from October 2004 to May 2008 (2800 patrols) and from 1968 to 2001 (about 2000 monthly reports, 28,000+ patrols); species traits from the PanTHERIA database and the IUCN Red List; historical occurrence from Grubb et al. (1998).

## Synopsis

### 1. What they did

Carnivores are disappearing worldwide, and losing them can ripple through a whole ecosystem. Most research has focused on large carnivores in East and southern Africa, and very little is known about West Africa, where hunting pressure is high. The authors asked which members of a whole carnivore community had survived in Mole National Park, Ghana’s largest protected area, and whether species traits (such as large body size or a large home range) could predict which ones were lost.

- Historical list. From earlier records they listed 16 carnivore species that lived in the park around the time it was established (about 1960).

- Camera traps. They ran 253 camera stations between October 2006 and January 2009, set about 1 km apart on dirt roads, wildlife trails and at water, for a total of 5469 trap-days.

- Other surveys. They recorded carnivore tracks and scat on foot (330 km) and from vehicles (1400 km), ran five call-in surveys playing prey distress calls and hyena calls at night, and drove three night spotlight transects.

- Patrol records. They searched the park’s law enforcement patrol records from 1968 to 2008 for carnivore sightings and built a yearly index of individuals seen per 100 patrols.

- Interviews. They interviewed 68 people in 27 villages around the park, showing photographs of nine medium and large carnivores.

- Analysis. A multi-species occupancy model used the camera data to estimate how likely it was that a species was present but missed. They then tested 17 species traits as predictors of persistence.

### 2. What they found

Which species remain

- Only 9 of the 16 historical species were detected by the cameras: spotted hyena, white-tailed mongoose, leopard, large-spotted genet, African civet, caracal, marsh mongoose, Gambian mongoose and side-striped jackal.

- Lion, African wild dog, honey badger, serval, large grey mongoose, slender mongoose and Hausa genet were not photographed, and the sign, call-in and spotlight surveys could not confirm them either.

- The occupancy model estimated 9 species in the sampled community (the range of likely values went up to 13), so it was unlikely that cameras had simply missed species. The chance of more than 10 species being present was under 0.2.

- Detection was low (about 6% per site per day), but over all sites and days the chance of missing a species that was there was very low.

Other evidence

- Recent patrol records (2004 to 2008) held only 32 observations of four carnivore species, including three reported lion sightings, the last in April 2007.

- Patrol counts of carnivores fell over 40 years, and lion and leopard counts both declined significantly. Wild dogs were recorded only five times, last in 1995. Spotted hyenas were recorded 17 times, 13 of them from 2004 to 2008.

- In interviews, lion (75%) and leopard (70%) were the best known. A few people reported recent sightings of serval, wild dog, lion and honey badger. Lion (34%), leopard (30%) and wild dog (17%) were most often described as gone or fewer. Use of carnivores for ceremonial, medicinal or food purposes ranged from 5% of respondents (honey badger) to 55% (lion), and lion (45%) and spotted hyena (25%) were most often blamed for killing livestock, with several retaliatory killings reported.

Predictors of persistence

- None of the 17 traits significantly explained which species persisted: not body size, home range, fecundity, density, diet, habitat, range size, threats, conflict with people or local use.

- Among the species still present, larger and wider-ranging species used more sites, but they were not easier to detect where they occurred.

### 3. What they concluded

- Nearly half (7 of 16) of the park’s carnivore species may now be functionally or fully extinct there, including the top predator, the lion. Many of the species that remain have declined.

- Photographs gave the strongest evidence. Patrol sightings and interviews suggest lion, serval, honey badger and even wild dog could persist, but these are harder to confirm, and the weight of evidence suggests they are at best very rare.

- Contrary to their predictions, losses could not be predicted from species traits. Small carnivores, which are less studied, may be more vulnerable than expected.

- Likely threats are illegal hunting, use of carnivore skins and organs, retaliatory killing over livestock, leg-hold traps and the isolation of the park by settlement and farming around it. A rise in spotted hyenas may be linked to the decline of lions. These drivers remain largely speculative and need urgent study.

### 4. How tracking was or was not used

Tracks and scat of carnivores were recorded during foot and vehicle travel around and between the camera sites, both systematically on transects and opportunistically. Sign was used as one line of evidence for which species were present (Table 2), alongside cameras, sightings, patrol records and interviews. Ambiguous tracks or scat of small and medium-sized carnivores that could not be confidently identified to species were left out. The authors judged sign, patrol and interview data less reliable or verifiable than photographs, and the sign data were not used in the occupancy model. No trailing was reported.

### 5. Tracker or observer certification

Not reported. The paper does not say who recorded or identified the tracks and scat, or how their skill was assessed. This matters for the sign evidence, which the authors rated below photographs partly because it is harder to verify.

### 6. Other technologies, skills or innovations

- Passive infrared film camera traps (DeerCam DC-300), set 40 cm high and 3 m from the expected path of animals.

- Hierarchical multi-species occupancy model that accounts for species being missed, to estimate how many species are really present.

- Call-in surveys with horn speakers on a truck and a 2-million candlepower spotlight.

- Long-term law enforcement patrol records turned into an abundance index.

- Decision trees and random forests to test species traits as predictors.

### 7. Implications for conservation (from the paper)

- West Africa’s carnivores need new initiatives to protect and restore them; even protected areas may not be safe refuges.

- Lion declines in other West and Central African parks suggest the lion is in a regional extinction crisis.

- Remaining populations need stronger support for law enforcement, together with efforts to ease conflicts between parks and local livelihoods.

- Because losses were not predictable from traits, more local data on whole carnivore communities are needed for conservation planning.

- Losing carnivores also undermines local customs and beliefs and potential tourism income.

### 8. Implications for tracking (Original Wisdom's observations)

_These are Original Wisdom's observations, not the authors'._

- The authors ranked tracks and scat below photographs as evidence, partly because sign is harder to verify. Reporting who identified the sign and how their skill was tested would make sign evidence easier to weigh against other methods.

## Cite this summary

Lawrence, K. 2026. Summary of: Evaluating persistence and its predictors in a West African carnivore community (Burton and Sam et al., 2011). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2011-burton-sam-evaluating-persistence-predictors-west/