Eastern Chimpanzee's habitat fragmentation in Nyungwe
National Park (NNP), Rwanda
Dominique Mvunabandi1 , Iris van Duren2, Tiejun Wang2
1
Department of Land Survey, Faculty of Applied Fundamental Sciences, Institut d'Enséignement
Supérieur de Ruhengeri, P.O. Box: 155 Ruhengeri, Rwanda
2
Natural Resource Department, Faculty of Geo-information Sciences and Earth Observation
(ITC), University of Twente, Enschede 7500 AE, The Netherlands
[email protected], [email protected], [email protected]
Abstract
Habitat fragmentation, arising from anthropogenic activities, is one of the main threats to
biological diversity. Chimpanzees are rare and endangered, and with a low fecundity they
are much affected by it. This study assessed chimpanzee habitat fragmentation around
Cyamudongo forest, Rwanda, in the light of potential habitat restoration to reconnect the
isolated forest patch with the contiguous forest of Nyungwe. Landsat satellite imagery and
aerial photographs were used to quantify land cover changes and analyse fragmentation
patterns between 1989, 2005 and 2013. Based on spatial chimpanzee habitat suitability
criteria followed by least cost path analysis, different scenarios for reconnecting the forest
patches were generated and compared.
Keywords
Ecological corridor, GIS, habitat fragmentation, Pan troglodytes, Remote Sensing, suitability
model.
1. Introduction
Chimpanzee populations have declined dramatically in the last decades and disappeared in the
wild in some African Countries due to improper forest management, habitat loss and hunting
(Kormos et al. 2003). Franklin et al. (2002) define habitat fragmentation as “ reduction and
isolation of patches of natural habitat” which affect many wildlife species including chimpanzees
(Basabose 2005). As a result, chimps are classified as endangered species by IUCN and
considered as species of global concern (Plumptre et al. 2010).
In Rwanda, chimpanzees are found in three nature reserves Nyungwe National Park,
Cyamudongo forest and Gishwati natural forest and these reserves are mainly composed of
mountain rainforest (Barakabuye et al. 2007). Also here forests are facing rapid degradation, loss
and change following deforestation for settlement, farming, agricultural activities and soil
erosion. The Wildlife Conservation Society (WCS) in Nyungwe National Park (NNP) showed
that the chimpanzee population decreased from 382 individuals in 2004-2005 to 306 individuals
in 2009. Forest fires contributed to the decline (Barakabuye et al. 2007; Plumptre et al. 2010). In
Gishwati forest reserve only 9-19 chimpanzees are left (Chancellor et al. 2012) while this reserve
was once the largest forest in Rwanda. However, it has significantly decreased in size over the
last 30 years as the forest reduced from 280km2 to 6 km2 (Guinness et al. 2014). Another small
population of 37-41 chimpanzees (pers. comm.) currently reside in the Cyamudongo forest
which is a disconnected forest patch at a distance of around 8 km from other populations in the
Nyungwe National Park.
To judge the risks for the isolated chimpanzee populations and to plan for reconnecting forest
areas, reliable and up to date baseline information about the state of the forests is required.
Remote sensing techniques have proven to be successful tools in mapping forest cover changes.
Pintea (2007) used Landsat imageries to detect and monitor changes of forest structures at the
landscape. He also investigated the use of Landsat data to map chimpanzee habitat in Gombe
National Park between 1972-1999. Several methods and techniques have been developed and
applied to quantify fragmentation. Hickey et al. 2012 developed and applied four landscape
metrics edge density, cohesion, contagion and class area, to characterize potential suitable habitat
of Bonobos, close relatives of the chimpanzees.
Spatially modelling potential chimpanzee habitat involves mapping environmental factors that
define chimpanzee habitat suitability. Such habitat suitability models and insight in where
destroyed chimpanzee habitat potentially can be restored into suitable chimp habitat are inputs
for ecological corridor design. Habitat factors seen as important for chimpanzee habitat
restoration are: land cover, protected areas, distance to roads and distance to villages (Saad et al.
2013; Nangendo et al. 2010). Riverbanks protected by country’s wetlands Organic Law N°
04/2005 (REMA 2011) potentially can be used to connect different patches. This law
recommend 20 meters from rivers as free zone to human activities. However this will require
political will for law enforcement and participation of local farmers. Agro-forestry and tree
planting activities may be motivating to local community to participate in restoring degraded
chimpanzee habitat. In Tanzania, Lake Tanganyika Catchment Reforestation and Education
(TACARE) Progam initiated in 1994 is a good example. People around lake Tanganyika and
Gombe national reserve are encouraged to actively protect chimpanzee habitat through
reforestation (JGI 2014).
In Rwanda (Saad et al. 2013) elaborated on a reforestation suitability map and an ecological
corridor that would be beneficial for national parks managers for future plans to connect
Nyungwe forest, Mukura forest and Gishwati to enable free mobility of chimpanzees. A similar
study in Uganda, considered chimpanzees habitat specific required to build a corridor model that
will enable mammals including chimpanzees to move across many patches of the MurchisonSemliki landscape (Nangendo et al. 2010).
This research seeks to assess the forest cover changes over the last 30 years, to examine the
fragmentation status of the forest patches between Cyamudongo and Nyungwe forest, to assess
habitat suitability in this area and determine the most promising location of an ecological
corridor when habitat restoration is considered to take place.
2. Materials and methods
Description of the study area
Nyungwe National Park is a high-altitude, mountainous tropical rainforest situated in the Southwestern of Rwanda between the latitude of 02°15'-02°55'S and longitude 29°00'-29°30'E. It
covers a total area of 970 km and together with neighbouring Kibira National Park in North of
Burundi, it forms one of the largest tropical mountain forests in Africa (Plumptre et al. 2002).
Temperatures are general cool (average minimum of 10.9 °C and maximum of 19.6°C) with
annual precipitation average of 1744 mm (Kaplin et al. 1998). Nyungwe forest is characterized
by various unique vegetation types and rich in wildlife among which many rare and endemic
species.
Cyamudongo forest (02°33.12’S 28°59.49’E) is a small dense forest patch around eight km away
from Nyungwe National Park (NNP). Historically, Cyamudongo forest was connected to
Nyungwe in its northeastern side. However, the increasing human population, converted forest
into farmlands and tea plantation and Cyamudongo forest was disconnected and progressively
reduced in size up to 410 ha remaining today. Its altitude ranges between 1,700 to 2,000 m above
sea level and it harbours a wide range of flora and fauna species. A group of some 40
chimpanzees (Pan troglodytes schweinfurthii), baboons, mona monkeys, important bird species
and many plants species find their home here.
Land cover change analysis
Landsat satellite images (TM 1989, ETM+ 2005 and ETM+ 2013) were classified into four land
cover types: forest, tea shrubs, crops, bareland and settlements using the Maximum Likelihood
Classification (MLC) algorithm in ERDAS imagine10.1 software. Digital aerial photographs
(2008 and 2010) were used for field observation. In a sample of 150 field locations, vegetation
recordings were collected during field work. These were used to classify (45 observations) and
validate (105 observations) the images. A 3x3 majority filter was applied to the classified images
to eliminate and replace isolated pixels; providing a more realistic representation of the forested
areas. The classification accuracy of the classified images was assessed in an error matrix and the
Kappa coefficient was used to compare classification accuracies of classified images. To identify
the land cover changes between 1989-2005 and 2005-2013, post classification comparison was
applied in ArcGIS 10.2 software.
Fragmentation analysis
Due to the shortage of publication on fragmentation analysis for primates, other publications on
Panda (Tong 2011, Sun 2011) and bonobos in DRC (Hickey et al. 2012) were used to identify
relevant landscape metrics that could best describe the habitat fragmentation for chimpanzees.
First, the land cover maps were reclassified in “forest” and “non-forest”. Next, FRAGSTATS
(raster version) was applied to quantify the forest fragmentation patterns. In the moving window
mode, fragmentation metrics were assessed for the forest cover maps of 1989, 2005 and
2013.The metrics are (1) Patch density: PD (2) Mean patch area index: Area MN (3) largest
patch index: LPI and (4) Mean proximity index: PROX_MN. A short description of these
metrics is provided in table 1:
Table 1: Landscape pattern metrics description
Index (unit)
LPI (%)
AREA_MN (ha)
PROX_MN (m)
PD
Description
The percentage of the landscape comprised by the largest patch
Average size of patches
Average proximity index for all patches in a class
Number of corresponding patches divided by total landscape area
Habitat suitability modelling
Chimpanzee habitat requirements are described by Basabobe (2002), Pintea (2007 ) and Koops et
al. (2012). They mention as most important spatial landscape variables: land cover type,
protected areas, slope, distance to rivers, distance to roads, distance to villages and forest
patches. These variables were mapped and standardized with scores from 0 = unsuitable to 100 =
highly suitable (see table 2). In a weighted summation a final suitability score was calculated
based on their relative importance to the chimpanzees (Nangendo et al. 2010). However, since
differences in opinion about relative importance of different variables exist between experts
(Nangendo et al. 2010; Saad et al. 2013; local park managers (pers. comm.), we created 4
different suitability scenarios applying different relative importance weights and standardization.
The scenarios are described as follows:
Scenario 1 is based on Nangendo et al. (2010) combining the landscape variables: land cover,
protected areas, distance to roads, distance to rivers, distance to settlements.
Scenario 2 and 3 were basically the same in terms of relative importance of the different factors
and was based on the publication by Saad et al. (2013). They gave steeper slopes higher
suitability scores because chimpanzees reside in areas with steeper slopes. However, we felt that
chimpanzee may be found on steeper slopes not because they prefer to reside there but because
human activities are more intense in relative flatter areas pushing the chimps into steeper areas.
Therefore, scenario 3 was created to generate a habitat suitability map based on lower suitability
scores for steeper slope.
Scenario 4 assumed that existence of remnant forest patches nearby in the landscape are utmost
important in the light of restoring the possibility for chimps to travel from the Cyamudongo
forest patch to the main Nyungwe forest area and vice versa. Therefore a new spatial layer was
created showing the presence of nearby forest patches within a distance of 300 m.
Table 2: Suitability criteria and standardisation
Land
cover
Protected areas
Forest
100
Tea
0
Crops
0
Bare
0
Protected
100
Non protected
0
Slope (degrees)
set 1
set 2
Distance to
roads (m)
0 – 10
0 – 50
10
51 –100
20
101–500
30
501 – 1000
40
> 1000
100
10
50
11 – 20
20
75
21 – 30
30
100
31 – 40
40
100
41 – 50
50
100
51 – 60
60
100
61 – 70
70
100
71 – 80
80
75
81 – 90
90
50
91–100
100
0
Forest in
surroundings
(300 x 300 m)
0 – 10
10
11 – 20
20
21 – 30
30
31 – 40
40
41 – 50
50
51 – 60
60
61 – 70
70
71 – 80
80
81 – 90
90
91 – 100
100
Distance
to rivers
(m)
0–50
100
51–100
90
> 100
80
Distance to
Settlements
(m)
0 –100
0
101-500
20
501-1000
40
1001 – 5000
60
> 5000
100
In table 3 the relative weights applied in the different scenarios are shown.
Table 3: Relative weights for habitat suitability factors
Factors
Land cover
Protected Areas
Slope
Dist. to roads
Dist. to rivers for
drinking water
Dist. to settlements
Forest patches
Scenario 1
(Nangendo et
al., 2010)
0.51%
0.31%
----0.04%
0.11%
Scenario 2
(Saad et al.,
2013) set 1
0.40%
0.13%
0.20%
0.04%
0.20%
Scenario 3
(Saad et al.,
2013) set 2
0.40%
0.13%
0.20%
0.04%
0.20%
Scenario 4
Nearby forest
patches
0.30%
0.10%
0.10%
0.04%
0.20%
0.03%
-----
0.03%
-------
0.03%
-------
0.03%
0.23%
Ecological corridor modelling
The suitability maps produced in the previous section served as inputs to model potential ecocorridor paths and analyse the sensitivity of the location of the corridor to the different habitat
suitability maps. To undertake this process, a software called ''corridor designer'' (Beier et al.
2007) was downloaded and added to the ArcGIS toolbox. The forest patches Cyamudongo and
Nyungwe were defined as start and end of the corridor and we assumed a corridor width of 200
m to be sufficiently wide.
3. Results
Land cover and land cover changes
The three thematic maps presented in figure 1 show land cover types of the study area in years
1989, 2005 and 2013. Four land cover classes are displayed namely, forest, crops, tea shrubs and
bare land and settlements. The relative sizes of the areas within the study area are presented in
Table 4 while table 5 tabulates the changes in land cover types between 1989-2005 and 20052013. The built up area expanded over time. A remarkable observation is that the forest coverage
reduced between 1989 and 2005 while it increased again in the period from 2005 to 2013. Near
the main NNP forest reserve several patches became larger and denser while in the more
intensively used agricultural area the forest remnants further disappeared. Between 1989 and
2005 the tea plantations were progressively converted in crops and between 2005 and 2013 many
new tea plantations were established. The overall accuracy of classification was 87.3% for the
1989 TM image, 88.2% for 2005 TM image, and 91.2 % for the 2013 ETM+ image and the
Kappa coefficients were 0.78, 0.79 and 0.85 respectively.
Table 4: Areas of land cover types in 1989, 2005 and 2013
Land cover classes
Forest
Tea shrubs
Crops
Bare land &
settlements
Total Area in Ha
1989
54100
5235
124317
471
184122
%
29.4
2.8
2005
%
47786 26.0
2191 1.2
2013
49972
6026
%
27.2
3.3
97.5 133882 75.4 126429
0.3
262 0.1
1347
98.8
0.7
184122
184122
A: Land cover map 1989
B: Land cover map 2005
C: Land cover map 2013
Land cover types
Figure 1: Land cover types in 1989, 2005 and 2013
Table 5: Magnitude of change in land cover types between 1989-2005 and 2005-2013
To
Forest
Tea shrubs
1989-2005
2005-2013
Tea shrubs 1989-2005
2005-2013
1989-2005
Crops
2005-2013
Bare land 1989-2005
&
settlements
2005-2013
34337
34196
345.6
214.2
13099
15562
5.4
151.2
1006.2
1096.2
1137.6
943.2
3882.6
0.9
19609
12369
3793.5
837
110075
113113
404.1
1.8
6.3
0
1.8
199.8
1188
60.3
0
0
110.7
151.2
Forest
From
Crops Bare land
Fragmentation analysis
The fragmentation varies among the forest coverage of the area, as highlighted in table 6. The
statistics indices showed that the forest patch density remained almost the same with from 1989
to 2005 but decreased from 2005 to 2013. At the same time, the size of the individual patches
decreased from 1989 – 2005 and increased again between 2005 and 2013 with an increase of the
largest patches. However, there was a serious decline in the proximity mean which did not
reverse in the period afterwards. This means that the expansion of agricultural land, human
settlements and other physical infrastructures worsened seriously the degree of isolation and
fragmentation of forest patches from 1989 – 2005 and this situation did not improve in the period
afterwards.
Table 6: Four landscape metrics values calculations
Years
PD
LPI
AREA_MN
PROX_MN
1989
16.98
10.85
1.73
349.91
2005
16.32
8.61
1.59
123.23
2013
8.369
12.50
3.24
107.68
Where PD = Patch density, LPI = Largest patch index, AREA_MN = Mean patch area and
PROX_MN = Proximity mean
Habitat suitability modelling
Four suitability maps, based on the different descriptions of habitat suitability by different
experts are shown in figure 2. The nature reserves are highly suitable whereas tea shrubs and
settlements appeared to be unsuitable in all the four maps. Scenario 1 shows a more extreme
contrast in suitability between the forest areas and the agricultural land in between compared to
scenario 2 and 3, where agricultural land became moderately suitable and rivers traces are clearly
shown. Scenario 4 shows contrast of rivers and forest patches in areas adjacent to forest of
Nyungwe.
Figure 2: Habitat suitability maps based on habitat suitability models by different experts.
Scenario 1 was based on Nangendo et al., 2010, Scenario 2 and 3 was based on Saad et al., 2013
with different standardisation for “slope” and scenario 4 was based on the assumption that forest
and presence of nearby forest patches are of utmost importance to chimp habitat.
Ecological corridor modelling
Three proposed corridor paths were created from suitability maps 2, 3 and 4 and are presented in
figure 3. The habitat suitability model based on Nangendo et al., 2010 did not allow creating an
ecological corridor between the Cyamudongo forest and the Nyungwe National Park because of
the too long distances of very unsuitable habitat. The corridors path 2 and 3 pass in the central
parts of the landscape through the agricultural land and mainly follow the river traces. It should
be noted that in these areas the slopes are somewhat steeper but virtually no forest patches are
available. The corridors of scenarios 2 and 3 are 8.44 km and 8.49 km long respectively.
Corridor path 4 was the longest with 9.98km, followed by corridor path 3 traverses through
agricultural land towards north direction of the landscape, areas with forest patches close to
Nyungwe National Park. None of the corridor paths 2, 3, and 4 avoided roads and only corridor 4
crossed human settlements.
A: No corridor path possible
A: Corridor path 2 based on suitability map 2
B: Corridor path 3 based on suitability map 3
C: Corridor path 4 based on suitability map 4
Figure 3: Corridor paths based on different suitability maps
4. Discussion
In this research, chimpanzee habitat suitability maps were the base of creating a potential
ecological corridor between Cyamudongo forest and the forest of Nyungwe National Park.
Habitat suitability depends on various factors and these factors all need standardisation before
they can be combined in a quantitative assessment. Some arguments for standardisation and
combination of the factors could be derived from literature (Saad et al. 2013; Nangendo et al.
2010) and other arguments came from discussion with experts like the park managers, research
officers in Rwanda and the Executive Director of Jane Goodall Institute in the Netherlands.
Suitability map 1 used the same biophysical variables, scores and weights as were used for the
Murchison-Semliki landscape in Uganda (Nangendo et al. 2010). However, based on this map,
no connection could be made between the Cyamudongo and Nyungwe forest because of too long
unsuitable stretches. Although being the longest route, corridor path 4 looks most promising for
chimpanzee habitat restoration because it appears to recognise best the importance of forest
patches as essential stepping stones included in the corridor route.
Chimpanzees cannot cross wide rivers unless there is a bridge as they cannot swim (JGI 2014a).
However, rivers are an opportunity to design ecological corridors as river banks are protected by
laws (Mclennan 2008). Mclennan (2008) observed that chimpanzees of Hoima in Western
Uganda were using forest patches along water courses. A similar approach can be applied for
forest patches and rivers found in the Cyamudongo-Nyungwe landscape. However, in the current
agricultural landscape in between the Cyamudongo and Nyungwe forest, there are virtually no
forest patches left. So, in its current state, this would hamper chimps to actually make use of such
a corridor. More ambitious implementation of protection and replanting river banks is required
when corridor establishment is really planned for.
Corridor paths 2 and 3 (based on Saad et al., 2013) pass through the central part of the landscape
characterized by relatively steeper slopes. Since this area is more susceptible to soil erosion
when used agriculturally, reforestation of river banks and other slopes is an environmental
measure that can contribute to a more sustainable landscape. This would go hand in hand with
shortening the ecological corridor for chimps. In case reforestation of riverbanks and other
currently used agricultural lands cannot be foreseen, we feel that the territory with steeper slopes
in between the two main forests is overvalued in the suitability maps 2 and 3. Forest and forest
patches are a boundary condition for chimps and if this boundary condition is not met it is
unrealistic to expect chimps travelling 8 km. Also, Chen et al. (2008), looking at possibilities to
reconnect isolated habitat patches of another threatened and rare species, the giant panda, stated
that to transform unsuitable habitat into suitable and to improve the landscape connectivity,
vegetation restoration needs to be applied. On top of that, chimps are probably found more
frequently on steeper slopes not only because they prefer it but also because human activities are
less intense compared to flat areas. Researchers revealed that steeper slope is favourable habitat
for chimpanzee especially for nesting (Koops et al. 2012; Saad et al. 2013, Ukpong et al. 2013).
However, the findings Ukpong et al. (2013) indicated that chimpanzees do construct nests on flat
lands.
None of the corridor paths 2, 3, and 4 avoided roads and some crossed human settlements. So
probably in the current set of weights used, the factor “distance to settlements” may be
undervalued. In the study by Saad et al. (2013), corridors paths did not avoid roads crossing
between the Nyungwe forest and Mukura forest in Rwanda. Therefore roads became a constraint
to the dispersal and free movement of chimpanzees between these two forests. As the roads are
paved, Saad et al. (2013) proposed an infrastructure solution to the Government of Rwanda to
ensure the safe crossing of chimpanzees. In the Netherlands, there is much experience with
designing eco-corridors enable animals to cross the roads. Fauna bridges or fauna tunnels let
animals cross safely (Jongman 2008) ). However, these expensive solutions are not necessary in
the Cyamudongo-Nyungwe landscape as the traffic density is rather low and the roads are not
paved.
It is difficult to find evidence to judge what width of a chimpanzee habitat corridor would be
scientifically sound. In the Murchison-Semliki landscape, Nangendo et al. (2010) used a 200m
wide corridor to enable free movement of chimpanzees. After consultation of park managers in
Rwanda, the same width of 200 meters was judged to be sufficient for a corridor path of around
8 kilometres from Cyamudongo to Nyungwe National Park.
If a corridor path is to be established in the future, there is the necessity of laws, regulations and
policies to be implemented. In this sense, well-established forestry and wetlands policies in
Rwanda could be followed. The country’s wetlands are protected by Organic Law N°04/2005
highlighting modalities of conservation and protection of the environment. The law recommends
a creation of 20 meters free buffer zone from river banks to ensure the protection of wetlands and
its biodiversity (REMA, 2011). Although one could argue if 20 m width is sufficient for the
corridor, this law will be helpful in delineating areas suitable for such a corridor. In Indonesia,
the law (Keputusan Presiden No 32/1990) requires a 50 meters buffer zones on both side along
small rivers. This law was applied to design an eco-corridor for the Borneo Orangutan (Persey,
2011) and the orangutans do profit of this forest corridor. It would be, therefore, a good
suggestion to expand the buffer required by the law from 20 meters to 50 meters. It will provide
safer mobility of chimpanzees. In 2004, the Government of Rwanda has put in place a forestry
policy with the aim of increasing forest cover up to 30% of national total area (MINIRENA,
2010). This policy supports forest restoration alongside rivers and corridor path 4. According to
Plumptre et al. (2010), restoration of degraded habitat and encourage agroforestry are among the
top priorities to protect chimpanzees in Rwanda and were included in the National Action Plan
for great apes. In this sense, the forest restoration should include fast growing trees which are
identified as chimpanzees food and nesting tress. Therefore, three levels of intervention can be
applied in Rwanda: (1) Involving local community to undertake agro-forestry practices by
planting trees such as Grevillea trees which provide fuel wood and timber to local community (2)
Planting fast growing trees of their preferences such fruit trees: avocado, papaya, orange and
citrus providing fruits (3) to plant nesting and food tress for chimpanzees such as
Tabernaemontana stapfiana, Prunus africana, Schefflera goetzenii, Salacia erecta, Coccinea
mildbraedii, Cleistanthus polystachyus, Croton megalocarpus, Casearia runssorica, Ekebergia
capensis, Macaranga kilimandschalica Cymphonia groblefela, Syzygium guineensis, Dombea
gotseenii, Ficus sp., Carapa grandiflora, etc. (Gross-camp et al. 2009).
Similar initiatives have been applied successfully in different countries to protect chimpanzee
and their natural habitat. In Tanzania the Lake Tanganyika Catchment Reforestation and
Education (TACARE) project was initiated to address poverty and support livelihoods in villages
around Lake Tanganyika and Gombe National Park. TACARE focuses on community socioeconomic development and provides training and education program to local communities to
increase their knowledge in forest conservation and natural resources in general (JGI, 2014b).
The same experience and success story from TACARE in forest restoration have an added value
in Rwanda during implementation of reforestation program in Cyamudongo-Nyungwe
landscape.
5. Conclusion:
In this research, land cover change analysis of the landscape between Cyamudongo and
Nyungwe forest was performed in relation to protection and potential restoration of chimpanzee
habitat. Fragmentation patterns of a period of 25 years were calculated and analysed. Different
chimpanzee habitat suitability models were compared we discussed to what extent these models
seem to provide a realistic base for design of an ecological corridor. The paths were modelled
based on the following set of biophyisical variables: land cover, protected areas, slope, distance
to rivers, distance to settlements, forest patches. After evaluation of how these factors were
standardised and what relative weights were assigned to the individual factors we concluded that
the most realistic suitability map resulting from this study is suitability map 4. This also provides
the most realistic routing for a potential chimpanzee corridor to connect the isolated
Cyamudongo forest with the forest of the Nyungwe National Park
6. Acknowledgments
We are sincerely grateful the staff of Wildlife Conservation Society and Rwanda Development
Board-Tourism as well as to the Conservation Department for their permission, logistics
assistance and support to conduct research in Nyungwe National Park. We would like to thank
Mr. Diederik Visser from Jane Goodall Institute -Netherlands for his invaluable expert
knowledge on chimpanzee that helped us to successfully conduct this research. We are also
grateful to Netherlands Fellowship Programme for providing the funding for the research.
7. Bibliography
Basabose, A. K. (2005). Ranging Patterns of Chimpanzees in a Montane Forest of Kahuzi ,
Democratic Republic of Congo. International Journal of Primatology, 26(1).
doi:10.1007/s10764-005-0722-1
Beier, P, Majka, D, Jenness, J. (2007). Conceptual steps for designing wildlife corridors.
Chancellor, R. L., & Rundus, A. S. (2012). The Influence of Seasonal Variation on Chimpanzee
( Pan troglodytes schweinfurthii ) Fallback Food Consumption , Nest Group Size , and
Habitat Use in Gishwati , a Montane Rain Forest Fragment in Rwanda. International Journal
of Primatology, 115–133. doi:10.1007/s10764-011-9561-4
Chen, L. D., Liu, X. H., Fu, B. J., & Qiu, J. (2008). Identification of the potential habitat for
Giant Pand in the Wolong Nature Reserve by Using Landscape Ecology methodology. In R.
C. for E.-E. Sciences (Ed.), National Key Lab of Systems Ecology (pp. 95–112). Beijing:
Chinese Academy of Sciences. Retrieved from
http://download.springer.com/static/pdf/766/chp%3A10.1007%2F1-4020-54882_7.pdf?auth66=1423968945_14b2d973b01f1875e5376d64c11b7d80&ext=.pdf
Franklin, A. B., Noon, B. R., & George, T. L. (2002). What is habitat fragmentation/. Stud Avian
Biol, (25), 20–29. Retrieved from
http://www.globalrestorationnetwork.org/uploads/files/LiteratureAttachments/368_what-ishabitat-fragmentation.pdf
Gross-camp, N, Masozera, M, and Kaplin, B. . (2009). Chimpanzee Seed Dispersal Quantity in a
Tropical Montane Forest of Rwanda. American Journal of Primatology, 911(November
2008), 901–911. doi:10.1002/ajp.20727
Guinness, S. M., & Taylor, D. (2014). Human Dimensions of Wildlife : An Farmers ’
Perceptions and Actions to Decrease Crop Raiding by Forest- Dwelling Primates Around a
Rwandan Forest Fragment. Human Dimensions of Wildlife, 19(2), 179–190.
doi:10.1080/10871209.2014.853330
Hickey, J. R., Carroll, J. P., & Nibbelink, N. P. (2012). Applying Landscape Metrics to
Characterize Potential Habitat of Bonobos ( Pan paniscus ) in the Maringa-Lopori-Wamba
Landscape , Democratic Republic of Congo. International Journal of Primatology, 381–400.
doi:10.1007/s10764-012-9581-8
JGI. (2014a). Chimpanzees FAQs About Us the Jane Goodall Institute of Canada. The Jane
Goodall Institute. Retrieved January 01, 2015, from http://www.janegoodall.ca/about-faqschimps.php#q6
JGI. (2014b). Jane Goodall Institute - Lake Tanganyika Catchment Reforestation and Education
(TACARE) Project, Tanzania. Retrieved January 01, 2015, from
http://www.ehproject.org/phe/jgi-tanzania_final.html
Jongman, R. and M. B. (2008). KEN - Knowledge for Ecological Networks : Catalysing
Stakeholder Involvement in the Practical Implementation of Ecological Networks Current
status of the practical implementation of ecological networks in the Netherlands (pp. 1–32).
Alterra. Retrieved from
http://www.ecologicalnetworks.eu/documents/publications/ken/NetherlandsKENWP2.pdf
Kaplin, B. A., Munyaligoga, V., Frugivore, P., & Moermond, T. C. (1998). The Influence of
Temporal Changes in Fruit Availability on Diet Composition and Seed Handling in Blue
Monkeys ( Cercopithecus. BIOTROPICA, 30(August 1995), 56–71.
Koops K, McGrew WC, de Vries H, M. T. (2012). Nest-building by chimpanzees (Pan
troglodytes verus) at Seringbara, Nimba Mountains: anti-predation, thermoregulation and
anti-vector hypotheses. International Journal of Primatology, 33, 356–380.
doi:10.1007/s10764-012-9585-4
Kormos, R., Boesch, C., Bakarr, M.I. and Butynski, T. (eds. ). (2003). West African
Chimpanzees. Status Survey and Conservation Action Plan. (p. 219). Gland,Switzerland
and Cambridge, UK: IUCN/SSC Primate Specialist Group. Retrieved from
http://www.primate-sg.org/storage/PDF/WACAP.English.pdf
Mclennan, M. R. (2008). Beleaguered Chimpanzees in the Agricultural District of Hoima ,
Western Uganda. Primate Conservation, 23(1), 45–54.
doi:http://dx.doi.org/10.1896/052.023.0105
MINIRENA. (2010). National Forestry Policy. KIGALI-Rwanda. Retrieved from
http://rnra.rw/uploads/media/final_national_forestry_policy_2011f.pdf
Nangendo, G., Plumptre, A.J., and Akwetaireho, S. (2010). Identifying Potential Corridors for
Conservation in the Murchison-Semliki Landscape. Unpublished Report to the UNDP/GEF
Conservation of Biodiversity in the Albertine Rift Forests of Uganda Project. (Vol. 256, pp.
0–20).
No Title. (2007), (December).
Persey, S, Imanuddin and Sadikin, L. (2011). A Practical Handbook for Conserving High
Conservation Value Species and Habitats within oil palm landscapes. (Z. S. of L.
Conservation, Ed.) (pp. 1–82). Indonesia: ZSL Indonesia’s Biodiversity & Oil Palm Project.
Retrieved from https://www.hcvnetwork.org/resources/folder.2006-09-29.6584228415/ZSL
Practical Handbook for Conserving HCV species - habitats within oil palm landscapes_Dec
2011.pdf
Pintea, L. (2007). Applying Remote Sensing and Geographic Information System for
Chimpanzee habitat, Change detection, Behavious and Conservation: PhD thesis. University
of Minnesota.
Plumptre, A., Rose Nangendo, G., Williamson, E., Didier, K., Hart, J., Mulindahabi, F., …
Bennett, E. (2010). Eastern Chimpanzee ( Pan troglodytes schweinfurthii ): Status Survey
and Conservation Action Plan 2010-2020 (p. 52). Gland,Switzerland. Retrieved from
https://portals.iucn.org/library/efiles/documents/2010-023.pdf
Plumptre, A.J, Masozera, M, Fashing, P.J, McNeilage, A, Ewango, C, Kaplin, B.A and Liengola,
I. (2002). Biodiversity Surveys of the Nyungwe Forest Reserve In S.W. Rwanda. WCS
Working Papers. Retrieved from Available for download from http://www.wcs.org/science/
REMA. (2011). ATLAS OF RWANDA’S CHANGING ENVIRONMENT: Implications for
Climate Change Resilience. Rwanda Environment Management Authority (p. 110).
KIGALI-Rwanda. Retrieved from https://na.unep.net/siouxfalls/publications/REMA.pdf
Saad, A. M., Regan, J., Muzungu, E., Mwiza, F., Paulson, C., Weiser, J., & Willis, A. (2013).
Highway for Apes : Using NASA EOS for wildlife corridor planning in Rwanda. Rwanda
Eco Team (p. 2013). Virginia. Retrieved from http://earthzine.org/2013/04/03/highway-forapes-using-nasa-eos-for-wildlife-corridor-planning-in-rwanda/
Sun, Y. (2011). Reassessing Giant Panda Habitat with Satellite-derived Bamboo Information : A
Case Study in the Qinling Mountains , China. MSc Thesis. University of Twente.
Tong, B. E. I. (2011). Three Decades of Change in Giant Panda Habitat around and within
Foping Nature Reserve , China. MSc Thesis. University of Twente.
Ukpong, E. E., Jacob, D. E., Ibok, P. B., & Nelson, I. . (2013). Nest Building Behaviour of
Chimpanzee ( Pan Troglodytes Blumenbach 1799 ) At Filinga Range of Gashaka Gumti
National Park , Nigeria. ARP Journal of Science and Technology, 3(7), 2011–2014.