The periodic censuses of the endangered populations of mountain gorillas help us to understand their population dynamics, to assess the success of conservation programs that are aimed at ensuring their survival, and to ensure that they received continued attention from the global conservation community. The mountain gorillas are highly endangered sub species of primates living in just two populations in the Virunga volcanoes extending to the borders of Rwanda, Uganda and the DRC. Both Bwindi and Virunga are small islands of the forest, surrounded by some of the highest rural population densities in Africa.

The forests have been affected from human impact in the past, in form of timber extraction, gold mining, encroachment, and poaching. Much has been done in past years to improve protection and management, with the creation of national parks and considerable support from international conservation organizations. However, the small size of the remaining forests, coupled with the intense pressure from the surrounding population, still presents considerable challenges to the park managers. Continued immediate threats to the forests and their wildlife include illegal use of forest resources (poaching, firewood collection etc.), encroachment and demand for land, human-induced fires, invasive exotic species and human-wildlife disease transmission.

More so, past hunting and logging has got an impact on the forest with reduced canopy cover and few large herbivores left. These threats raise questions about the ability of these islands of the forests to survive and regenerate in the long term, and emphasise the importance of close monitoring of the forests and the wildlife populations which they support.

The Virunga gorilla population has been the subject of much research and several censuses over the last 30 years.  Most recent census has found out that the population is increasing, despite over a decade of insecurity in the area. The Bwindi population is less well known. The intensive ecological and behavioural research has only been carried out over the last few years. The early surveys estimated the population at less than 200 but these were not based on complete counts and could be underestimates. All the groups were monitored over 7 years to 1993 during which time the population remained stable at around 300 individuals. The only full census of the population using the complete count sweep method developed within the Virungas was carried out in 1997.

This census established a total population of 300 gorillas and concluded that the population size was stable. The gorilla censuses should be carried out every five years under normal circumstances. The main aim of this census carried out in 2002 was to assess changes in the population since 1997 so that to increase our understanding of gorilla population dynamics in Bwindi, levels of human disturbance and their effects on the gorillas and the effectiveness of conservation efforts.

The Methods used

The procedures employed in census of gorillas were based on that previously used in Bwindi and the Virungas. Six teams consisting of trackers and team leaders go through the park systematically from south east to North West. The park wad divided into small sectors, centered on campsites and the access points. One team was assigned to census each sector proceeding such that no more than 3 days were to be left between the completion of work in one sector and the beginning of work in the next sector.

The teams in the near sectors shared campsites, to allow better coordination of movements and comparisons of findings. Each of the sectors was searched by walking an irregular network of investigation routes across the area. When recent gorilla trails was found, it was followed until nest sites were located. The actual direction of reconnaissance routes walked was determined by the terrain and the availability of existing trails, while ensuring that the routes were adequately dense so that no area was missed which could have been large enough for a gorilla group to spend more than one week in it.

So as to achieve this, the distance between adjacent trails was never greater than 500 to 700m. Using the topographic maps, along with GPS, compass and altimeter readings, each census team mapped as accurately as possible all the paths taken and the gorilla trails followed. GPS readings were taken every 250 m along the trail, to ensure that it could be accurately mapped. By covering the area in this way, mapping and dating all gorilla trails and nest sites, and by marking nest sites encountered with cut stakes, it was possible to ensure that all groups were found and that none was counted twice, and to distinguish similar sized but distinct gorilla groups found close to each other.

At each nest site, the nests were counted and dung size measurements, along with the presence of silver hairs, were used to establish the age sex composition of the group. The teams aimed at finding at least three nest sites for each group to confirm the composition of each group, since individual nests or dung could be missed at one nest site. Dung size classes used were as follows:

Adult male (SB): > 7.2 cm (with silver hairs)

Adult female or blackback male (MED): 5.5 – 7.2 cm Juvenile/sub adult (JUV): < 5.5

(Sleeping in own nest) Infant (INF): generally < 4cm (sleeping in mother’s nest)

The Juvenile and sub adult categories were combined since past experience indicated that dung sizes don’t give precise information to differentiate the three categories 9adult, sub adult and the Juvenile). The young gorillas constructing their own nests were always considered here as the combined category juveniles/ sub adults not infants. In the absence of the infant dung, the adult female nests couldn’t be differentiated from those of a comparable sized sub adult male.

The Dung of young infants is rarely found in nests, and so the number of infants in the population is underestimated by these methods. However a correction factor can be calculated for this based on the fact that past census of groups with known composition has shown that one third of infants are missed in this way.

The survey methods of human disturbance

The inspection trails walked while looking for gorilla trail covered a large portion of the park, and provided an excellent opportunity of collecting data on signs of human impacts in the forest and other mammals. As you walk, these trails, signs of human use were recorded, with the GPS location and estimated age of each. These included snares, traps, human tracks, poachers’ camps, pit sawing sites, building poles, firewood and the bamboo cutting.

Age was categorized as recent “recent” < 3 months, “old” > 3 months, and very old> 10 years. This category was very old and was used for pit sawing sites that are visible for many years. The signs showed as very old predated the gazeting of the park. These survey methods were the same as those used in 1997 gorilla census; these survey methods were the same as those used during the 1997 gorilla census, so that direct comparisons could be made in the frequency and distribution of signs of human disturbance over this period.
Signs were analyzed as encounter rates per km walked, which required an accurate estimate of the distance travelled by each census team. Distances were measured for 179 out of 199 reconnaissance trails in this way.

For the remaining 20 trails, the hipchain machines broke, or the thread supplies ran out while the teams were in inaccessible places. The distances for each trail were also estimated from the sum of the straight line distances between each consecutive GPS reading taken on that trail. The distances measured for GPS were closely correlated with distances measured on the ground using the hip chains. (Pearson’s correlation coefficient r = 0.985, p<0.0001), and consistently underestimated the hipchain distance by around 10%. The GPS distance estimations, corrected for this underestimate, were therefore used where hipchain distance measurements were not available.

The total distance walked on each investigation trail was measured using a hip chain with biodegradable thread. Trails walked, animal signs and human use were recorded separately in and outside multiple use zones, where these occur in some of sectors of the park, to allow the comparison between multiple use and non-multiple use areas.

The population results

The five habituated groups in Bwindi contained a total of 72 individuals at the time of the census. More so, 22 unhabituated groups were found along with 10 lone silverback males, giving a total uncorrected population count of 298 individuals. A total of 33 infants were counted in the unhabituated groups, so that we predict that another 17 would have been missed since they were too young for their dung to be visible in the nests. This brings the corrected total to 315 individuals, and as with the previous census in Bwindi, we round this figure to 320 as our best estimate of the population size, since experience shows that a small number of small groups or lone silverbacks can be missed with these methods.

While the total population size has increased slightly, the other population parameters, group size and the percentage of immature in the population, are comparable with those found in 1997. The number of lone silverback found increased from seven to ten, but these individuals are particular difficult to find. More so, the gorillas were concentrated towards the interior and western areas of the southern sector of the park. As with all the past surveys, no gorillas were found in the northern sector. Signs of human disturbance were generally more frequent during this census than in 1997.