Endemic to the Upper Guinean Forests of West Africa, the pygmy hippopotamus Choeropsis liberiensis has lost an estimated 75% of its former range and now occurs in seven remnant fragments across Sierra Leone, Guinea, Côte d’Ivoire and Liberia (Lewison & Oliver, Reference Lewison and Oliver2008; Fig. 1). The species is categorized as Endangered on the IUCN Red List and the most recent population estimate, of c. 3,000, was made in the early 1990s (Lewison & Oliver, Reference Lewison and Oliver2008). The main threats to the species include habitat fragmentation, land conversion and hunting (Roth et al., Reference Roth, Hoppe-Dominik, Muhlenberg, Steinhauer-Burkart and Fischer2004). Its threatened status, long independent evolutionary heritage, and ongoing threats make the pygmy hippopotamus a candidate for priority conservation attention (Isaac et al., Reference Isaac, Turvey, Collen, Waterman and Baillie2007).

Fig. 1 The 2008 (dark grey) and historical (light grey) range of the pygmy hippopotamus Choeropsis liberiensis. This map excludes the Nigerian subpopulation, which is thought to be extinct (Lewison & Oliver, Reference Lewison and Oliver2008). Inset indicates the location of the main figure in West Africa.

Conservation of the pygmy hippopotamus is hampered by a lack of basic biological knowledge, including details of distribution, population trends and ecology. Observational records of the species are scant because of its cryptic nature but the most acute period of range decline (inferred from habitat loss) has been over the past 3 decades (Lewison & Oliver, Reference Lewison and Oliver2008). The largest fragment of the species’ former range is the contiguous Upper Guinea forest in Liberia and Côte d’Ivoire and this area probably harbours the majority of the remaining population (IUCN Hippo Specialist Group, 2008). Priorities for the conservation of the pygmy hippopotamus include establishing a reliable method for assessing the sizes of the various populations, and monitoring the species in protected areas using census techniques (Lewison & Oliver, Reference Lewison and Oliver2008).

Cost effective and accurate monitoring of rare and cryptic species is problematic. Transect methods using spoor and droppings must deal with differences in skill levels of those conducting the survey, and decay and defecation rates, all of which contribute to inaccuracy and increased variance in abundance estimates (Plumptre, Reference Plumptre2000). Although the use of cameras triggered by the animals themselves is not new (Champion, Reference Champion1927) it is only since the development of camera traps for wildlife monitoring in the early 1990s that their use has become more widespread. Camera-trap monitoring is rapidly gaining acceptance (Rowcliffe & Carbone, Reference Rowcliffe and Carbone2008) and new standardized camera-trap methods have been advocated for landscape-scale monitoring (O’Brien et al., Reference O’Brien, Baillie, Krueger and Cuke2010).

To begin to address conservation priorities for the pygmy hippopotamus we carried out camera-trap monitoring in Sapo National Park (Fig. 1), a stronghold of the species (IUCN Hippo Specialist Group, 2008). During the Liberian civil wars (1989–1996 and 1999–2003) many hundreds of hunters and gold miners inhabited the Park. Although the amount of wildlife killed was not recorded, pygmy hippopotamuses were reported anecdotally to be targets. A recent government led initiative has resulted in an estimated several thousand miners leaving the protected area. Their presence had resulted in habitat destruction, pollution of water courses, and commercial and subsistence hunting by the miners and their families.

Although logging has ceased within Sapo National Park it continues elsewhere in the pygmy hippopotamus’s range. The forests of Liberia declined by 2.9% between 1986 and 2000 (Christie et al., Reference Christie, Steininger, Juhn and Peal2007) but rates were higher in Côte d’Ivoire. Land conversion continues at a rapid pace, principally because of agricultural expansion, wood extraction and infrastructure development (Norris et al., Reference Norris, Asase, Collen, Gockowski, Mason, Phalan and Wade2010).

We conducted a survey using infrared heat and motion triggered digital camera traps from 24 January to 23 March 2008. The survey was designed to detect wide ranging and cryptic species. Cameras were spaced at 2-km intervals (O’Brien et al., Reference O’Brien, Baillie, Krueger and Cuke2010) in Zone 1 of the Park (the Park has three administrative zones). We used 32 cameras set 40 cm from the ground (O’Brien et al., Reference O’Brien, Baillie, Krueger and Cuke2010). The centre of each grid square was located using a global positioning system, and cameras secured in an optimal location (e.g. recently used animal trail), in a 100-m radius from the centre of the grid square. A total of 1,247 camera trap days were achieved.

Seven camera-trap events provided photographic records of pygmy hippopotamuses (Plate 1). All records were from the same camera trap station, in seasonally inundated primary forest on the inside of a river meander, just inside the 2005 boundary extension to the Park. Footprint and scat signs of the species were recorded in two further locations, both of which were in similar habitat. Lack of photographs from the other cameras could be because other cameras were located in different habitat (higher altitude, secondary forest). Although methods to estimate absolute density without the need for individual recognition continue to be developed (Rowcliffe et al., Reference Rowcliffe, Field, Turvey and Carbone2008) the number of trap events in this study were not sufficient to employ such methods. A more targeted survey, rather than the general scheme used here, is required.

Plate 1 Camera-trap photograph of Choeropsis liberiensis in Sapo National Park (Fig. 1).

To optimize survey design for estimation of change in occupancy, or to calculate density of the pygmy hippopotamus, the grid size used should maximize the likelihood of encounter. An ideal grid to track changes in occupancy would have cells with an area approximately equal to home range size, such that changes in the population would generate changes in the proportion of area occupied (MacKenzie et al., Reference MacKenzie, Nichols, Lachman, Droege, Royle and Langtimm2002). Average home range for C. liberiensis is estimated to be 0.4–0.5 km² for females and 1.5 km² for males (Roth et al., Reference Roth, Hoppe-Dominik, Muhlenberg, Steinhauer-Burkart and Fischer2004). Confirmation of pygmy hippopotamus home range sizes using radio telemetry would help guide the optimal grid size for measuring change in occupancy, which would provide an index of any change in abundance (O’Brien et al., Reference O’Brien, Baillie, Krueger and Cuke2010). If cells are smaller than average home range then the measure is a measure of intensity of habitat use (Royle & Nichols, Reference Royle and Nichols2003).

Refining and developing this camera trapping technique, in combination with sign surveys, could yield information on the status of the pygmy hippopotamus and on any population trends in Liberia and across its other range states, thus addressing the recommendations of the IUCN Hippo Specialist Group. A range-wide survey of the pygmy hippopotamus is required to establish a robust baseline from which future conservation efforts can be developed. It is important to gain an understanding of whether this species is able to cope with the long-term effects of habitat fragmentation across its range. Repeat surveys are now underway in Sapo National Park, and an extension of the survey area to other key sites across the species’ range will begin in 2011.