Bonobo (Pan paniscus) Conservation Strategy 2012–2022

6

3. Status of Bonobo Populations

3.1 Current Knowledge 

Information on the distribution and abundance of bonobos is fragmented, as much of their geo-

graphical range has not been surveyed. Speculative population estimates have varied from 29,500 

(Myers Thompson 1997) to 50,000 (Dupain & van Elsacker 2001). Fruth et al. (2008) advised caution 

on the use of these figures, because of the wide confidence intervals – the estimate for Salonga 

alone was 7,100–20,400 (Grossmann et al. 2008). Through analysis of all available data from recent 

surveys (2003–2010), the modelling group ascertained that less than one third of the bonobo’s 

range has been surveyed. Figure 1 shows the bonobo’s range as modelled for suitable conditions 

(Hickey et al. 2012). Figure 2 shows the areas surveyed for bonobos between 2003 and 2010. The 

total area surveyed (139,537 km²) represents almost 25% of the historic range (564,542 km²), thus 

it was not possible to produce a rangewide estimate of bonobo density or abundance.

Although quantitative data are patchy, the sites that have been surveyed give a minimum population 

estimate of 15,000–20,000 individuals (see Table 1). All available bonobo nest survey data collected 

between 2003 and 2010 were used to develop a model to predict the spatial distribution of potentially 

suitable conditions for bonobos throughout the area between the Kasaï and Congo rivers (Hickey et 

al. 2012). The modelling software used for this exercise was MaxEnt (Phillips et al. 2006).

The predictive environmental variables used in the final analysis were:

•  percent forest land cover

•  forest edge density (a measure of forest fragmentation)

•  distance from rivers

•  distance from agriculture

The main caveats of the resulting predictive model, which were systematically addressed

2

, are:

•  bias may exist due to some sites and habitat types being sampled more intensively 

than others

•  nest location errors could be present due to a possible inconsistency in GPS settings 

used across sites

•  environmental predictor variables were limited to those available across the full range 

in raster format, because MaxEnt requires spatially complete data. Highly detailed 

biotic and abiotic data relevant to bonobos are lacking at this scale.

Table 1. Minimum bonobo population estimates

Landscape

Bonobo population size

Salonga-Lukenie-Sankuru

>5,000

Tshuapa-Lomami-Lualaba

>5,000

Lac Télé-Lac Tumba Swamp Forest Landscape

<5,000

Maringa-Lopori-Wamba Forest Landscape

<5,000

Outside protected areas

>1,000

Minimum total estimate

15,000-20,000

2 Iterative modelling with data from a single site at a time, followed by modelling with all data minus one site at a 

time, provided an assessment of the potential bias. Based on this sensitivity analysis, any bias in the final model is 

expected to be low. The resolution of predictor variables in the final model was 100-m pixels, and several predictor 

variables were developed using a neighbourhood analysis such that the conditions (e.g., edge density or percent 

forest) in neighbouring pixels were incorporated into a given pixel’s value. These steps reduced the effect of any 

potential GPS location errors. Data points that did not meet basic quality assessment/quality control rules were 

excluded from the final analysis. The unavoidable limitation of spatially complete environmental predictor variables 

likely has the most influence on the model outcomes.

---

Figure 1. CARPE landscapes and officially designated protected areas overlapping the bonobo range, as modelled for suitable conditions.

---

Figure 2. Map of generalized survey areas within which 2003–2010 bonobo surveys were conducted. [Unlike the nest-only data used to model suitable conditions for bonobos, not all grid squares were 

surveyed or contained nests]. Landcover layer: WRI & MECNT 2010

---

9

Figure 3 indicates the relative probability of occurrence of suitable conditions for bonobos. The 

model identified areas likely to provide suitable conditions for bonobos that have not yet been 

surveyed. The model also indicates that, on a rangewide scale, the principal factors determining 

the distribution of bonobo nests are a) distance from agriculture and b) forest edge density, both 

of which suggest that bonobos avoid areas of higher human activities (Hickey et al. 2012). The 

poaching associated with these measures of human activities is considered to be the common 

determinant of current bonobo distribution, as for most other large species in the Congo Basin.

3.2 Priority Populations

The modelling exercise identified four strongholds (Fig. 4) that harbour the majority of bonobos 

known to remain. We refer to these as the ‘northern block’ (Maringa-Lopori-Wamba), ‘eastern 

block’ (Tshuapa-Lomami-Lualaba, TL2), ‘southern block’ (Salonga) and ‘western block’ (Lac 

Tumba-Lac Mai Ndombe). Each of these forest blocks contains at least one proposed or exist-

ing protected area and large expanses of forests where most conservation and research projects 

currently operate. The concentration of research and conservation activities in these blocks may 

have introduced a bias and overestimation of the importance of these areas compared with other 

less-intensively surveyed sites. The model also predicts that some unsurveyed areas outside of 

these four strongholds likely contain suitable conditions for bonobos, and these will be priorities 

for future survey efforts.

Northern block (Maringa-Lopori-Wamba)

This block corresponds approximately to the Maringa-Lopori-Wamba landscape adopted by 

the Congo Basin Forest Partnership (CBFP) and the Commission des Forêts d’Afrique Centrale 

(COMIFAC). Situated in the Maringa and Lopori river basins in Equateur Province, this landscape 

covers approximately 74,000 km². The area is very far from urban centres, experiences high levels 

of poverty and people are extremely dependent on natural resources obtained through slash-and-

burn agriculture, fishing and hunting. The human population is estimated at 586,700 inhabitants 

with densities of 2–4 people/km² where the existing or proposed protected areas are located, and 

up to 32 people/km² in agro-pastoral zones and urban centres. Forest covers approximately 67% 

of the block with swamps covering an additional 26%. The rest is young secondary forest and rural 

complex (Dupain et al. 2009). The block contains four areas offering various levels of protection:

•  Lomako Yokokala Forest Reserve (3,625 km²)

•  Luo Scientific Reserve (225 km²)

•  Kokolopori Bonobo Reserve (4,000 km²), a community-based natural resource man-

agement (CBNRM) area

•  Iyondji Community Bonobo Reserve (1,100 km²), an area of intact forest adjacent to 

the Luo Scientific Reserve.

Since 2006, a major participative land-use planning exercise has been carried out, covering 

approximately 70% of the landscape. This work has designated a mosaic of protected areas, 

CBNRM areas, sylvo-agro-pastoral zones and logging concessions

3

. The objective is to maintain 

forest cover and connectivity between ecologically important habitats in order to reconcile conser-

vation needs and human activities.

Bonobos are found in varying densities throughout this landscape. Although current data do not 

allow estimation of the total number of bonobos in this stronghold, surveys suggest that this block 

contains some of the most important populations in the bonobo’s range. 

3 As of August 2012, one concession was operational (TRANS-M) and two more are planned (K7 and K2 have 

been attributed to SIFORCO).

---