Factors affecting the habitat selection by reintroduced Burchell’s zebra

and blue wildebeest in Maputo Special Reserve

Luís Jr. Comissário Mandlate1,2* and Flávio H. G. Rodrigues1

1
Programa de Pós-graduação em Ecologia, Conservação e Manejo da Vida Silvestre,
Departamento de Biologia Geral, Instituto de Ciências Biológicas, Universidade Federal de
Minas Gerais - UFMG, Belo Horizonte, Minas Gerais, Brazil

2
Curso de Engenharia Florestal, Divisão da Agricultura, Instituto Superior Politécnico de Gaza,
Chókwè, Moçambique;

E-Mail: [email protected]

ABSTRACT

Understanding factors that influence habitat selection by herbivores is one of the most
crucial questions to understand the biological requirements for conservation and
management of animals in conservation areas. This study was aimed at examining the
habitat selection and factors that influence habitat selection by reintroduced plain zebra
and blue wildebeest in Maputo Special Reserve, Southern Mozambique. We calculated
habitat selection indices to assess habitat use by these species during the dry season. We
also use Generalized Linear Models (GLM), with Poisson distribution to construct the
model in order to examine factors that played an important role in habitat selection. The
best candidate model was selected using Akaike information Criterion. We expected
that these two grazers have the same needs in habitat selection, partitioning resource
that allow coexistence among them. The results indicated that the habitat preferred by
zebra was eucalyptus plantation and tree savanna, while blue wildebeest preferred only
tree savanna. Both species preferred more open sites with 0 to 10% shrub and tree
canopy cover, midslope topography, very tall grass, high grass cover and no pattern of
grass greenness was seen. The animals were found at distance of <1km from the water
points. Grass cover, tree cover and shrub cover were predicted to have more influence in
habitat selection by zebra, while for wildebeest the best-supported model included only
habitat type.

Key words: factors of habitat selection, reintroduced Equus burchelli and Connochaetes taurinus.
1. INTRODUCTION
The study of habitat selection is one of the most crucial questions to understand
the biological requirements, conservation and management of animals (Groom
and Harris 2009; De Knegt et al. 2011; Freitas et al. 2008; Conner et al. 2003). It is also
a pre-requisite to understanding the distribution, abundance and determining the
possible species for reintroduction, as well as the density, which they can be stocked
(Dekker et al. 1996). At a landscape scale, the selection of habitat in space and time by
herbivores may be influenced by biotic and abiotic factors (Watson et al. 2011; Bailey
et al. 1996; Redfern et al. 2003, Senft et al. 1987).

Species have varying requirements; with selection of food and shelter being the most
important (Traill 2003). That is why; habitat selection is influenced more by seasonal
changes of forage abundance and quality than other factors (Iversen et al. 2014, Bailey
et al. 1996). Also, differential habitat selection is crucial to coexistence among species
and avoid interspecific competition (Prins et al. 2006, Rosenzweig 1981). Other factors
that influence the habitat selection by herbivores are predation, woody cover, slope and
distance to water (Bailey et al. 1996).

Herbivores select their habitats to maximize forage intake and to minimize their risk of
predation (Burkepile et al. 2013, Riginos and Grace 2008, Dekker et al. 1996). Habitats
with higher vegetation cover and tree density decrease the visibility and ability of
herbivores to detect predators and avoid them (Riginos and Grace 2008). Consequently,
herbivores tend to avoid closed woodland habitats (Riginos and Grace 2008).
Otherwise, large trees improve grass quality by enhance below-canopy grass nutrients,
such as Nitrogen and Phosphorus contents, therefore attracting and benefitting grazers
(Treydte et al. 2009). Seasonal fluctuations in water availability cause changes in the
profitability of habitats and the animals must be able to adapt behavioral and spatial
responses to these fluctuations (Bennitt et al. 2014). The combined influence of biotic
and abiotic factors may be particularly important in determining the distribution patterns
of large herbivores in African savanna ecosystems (Redfern et al. 2003).
Effective monitoring of reintroduced species requires understanding of basic ecological
requirements such as habitat selection (Kelt etal.2014; Muposhi et al. 2014). In
ecosystem context, a clear understanding of the factors influencing habitat selection is
considered a prerequisite for interpretation of interactions between animals and their
environment (Gandiwa 2013, O'Kane 2005).
Maputo Special Reserve (MSR) is one of the conservation areas in Mozambique where
wildlife populations were nearly extirpated as a result of civil war two decades ago
(1977–1992) (Stalmans 2015, Lindsey and Bento 2012). Since 2010, a multi-year
reintroduction programme has been active to restore wildlife populations. Between 2010
and 2016 a population of Burchell’s zebra and Blue wildebeest were reintroduced in the
reserve (Stalmans 2015).

Burchell’s zebra and Blue wildebeest are grazers with similar body weights that prefer
short grass, are water dependent and are often seen in the same foraging site (Redfern et
al. 2003, Apps 2000, Voeten and Prins 1999). Despite this, it is expected that these two
grazers have the same needs in habitat selection, partitioning resource that allow
coexistence among them. The aim of this study was to investigate factors influencing
the habitat selection by reintroduced Burchell’s zebra and Blue wildebeest in MSR. This
research was carried out in the dry season, a critical period for African ungulates.

MATERIALS AND METHODS

Study area

The study carried out in Maputo Special Reserve (MSR), southern Mozambique,
located at the estimated coordinates 26°25’S, 32°45’E (Figure.1). The reserve was
established in 1932 and comprises 800 km2.There are two different seasons – a hot rainy
season (October to March), and a cooler dry season (April to September). The average
annual rainfall is 690–1000 mm. According to De Boer et al. (2000), there are six major
vegetation communities in (MSR): dune, mangrove, riverine, forest, woodland, and
grass plains.
Figure. 1. The location of Maputo Especial Reserve .

Research design and data collection

Data were collected from July 2016 to November 2016. During ten days of each month,
zebra and wildebeest were investigated along road transects. Each road transect was
investigated during the predominant feeding time of ungulates, first in the morning
starting at 6:30 am until 10:00 am and second starting 15:30pm and ending at 18:30 pm.
A vehicle was driven at a speed of 25 km/h with two observers, one observing the
presence of herds out at the right side and another observing out at the left side of the
road transect using binoculars (Zenith TEMPEST 8 X 30 FIELD 7.5o). In total 1089 km
was driven, at an average of 217.8 km per month.

When an animal or herd was observed from 50 to 300 m of the road transect, the
location of the vehicle was recorded using a handheld GPS device (GPS map XL,
Garmin 62) and the distance between the vehicle and the animals was determined using
a range finder (RZ900D Laser Distance Meter 6X). For each road transect the number
of all zebra and wildebeest within the specified area were counted.
After the animal left the sites, the research team approached on foot where the animal
was observed. At each of the sites, freshly grazed grass was searched and confirmed by
a lighter and brighter colour at the surface of the broken grass leaves and stems than old
bites (O’shaughnessy et al. 2014; Arsenault & Owen-Smith 2008). Based on this it was
recorded if the animals had been feeding on the grass or not.

We also recorded the habitat features within a 25m radius plot. The habitat features
recorded included the (1) topography classified as lowland, midslope or upland; (2) tree
(>2.5 m in height) and shrub (<2.5 m) canopy cover, estimated using Walker 8-point
scale (Walker 1976); (3) prevalent grass height, estimated in the following classes short
(≤10 cm), medium (11–20 cm), medium-tall (21–40 cm), tall (41–80 cm) and very tall
(>80 cm); (4) grass greenness and its percentage of cover were estimated using Walker
8-point (Walker 1976); (5) all surface water point were recorded using handheld GPS
and the distance from the center of 25m radius to the nearest point of water was
measured using range finder and computed using Arc GIS 9.3 for point that were
impossible measured using the range finder,(6) the presence of termite mounds were
recorded as presence or absence;(7) habitat type in each sampling site was recorded in
the vegetation map of reserve using the data of forest inventory of Marzoli (2007).

The minimum distance of one sampling point to another was 200 m. A total of 109
points were sampled for zebra and 45 for wildebeest during the period of study in MSR
and the number of sampling points per road transect depended on the location of
animals. The data were collected in four-habitat types where the herds of zebra and
wildebeest were found, namely: Eucalyptus plantation (2781.95 ha) Tree savanna
(127081.19ha), Shrub savanna (131862.1ha), Semi-evergreen forest (16765.49 ha).

Data analysis and statistical methods

Habitat selection of each herd was calculated using the selection index (wi) (Krebs
2014). The proportion locations within a particular habitat was divided by the
proportion availability of that habitat type within the reserve (Jenkins et al. 2015
Djagoun et al. 2014, Canter 2008, Arthur et al 1996; Garshelis 2000; Krebs 1989):

oi
wi=
pi
Where: Wi-selection index, oi- proportion locations of animals within a particular
habitat and pi is the proportion availability of that habitat type. Values of wi above 1.0
indicate preference, values less than 1.0 indicate avoidance and value wi=1 indicate that
the habitats were used in direct proportion to their availability within the reserve (Krebs
2014).

For each habitat, a Chi-square analysis goodness-of-fit was used to test hypothesis that
animals selected all habitats in proportion to their availability (Krebs 2014), followed by
simultaneous Bonferroni confidence intervals at 0.05 level of significance to determine
which the habitats were being preferred or avoided. (Moyo et al 2013, Krebs 2014,
Cherry 1996).

Chi-square test was used to distinguish each habitat features between zebra and
wildebeest during the period of study. The categories of some habitat features were
inadequate for analysis, as a solution we categorized habitat features into fewer
categories to increase the sample size. The following categories were used for grass
greenness: brown (0-25%), mainly brown (26-50%), mainly green (51-75%) and green
(>75); grass cover of the grassland: low (0-25%), medium (26-50%) 51-75% (medium-
high) and >75(high). For shrub and tree cover: open (0-10%), mainly open (11-25%)
and mainly closed (>25%). Topography and grassland height remained with the
categories predetermined during field data collection. Independent T- Test to compare
the distance to the nearest water points between zebra and wildebeest was performed
(Zar 2009).

For each herd, seventeen candidate models were generated in order to screen which
attributes features played an important role in habitat selection were done. The number
of categories within each of habitat feature was used as response variable, while habitat
features were used as categorical predictor variable. Count and categorical data not
generate a normal distribution. Because of that, Generalized Linear Models (GLM),
with Poisson distribution was used to construct the model. (Crowley 2013). All analysis
were performed in R statistical package (R Development Core Team 2015).
 Models
Grass cover
Grass Height
Grass Greenness
Habitat type
Shrub cover
Tree cover
Topography
Habitat type+ Grass greenness + Grass cover+ Grass Height+ Tree cover + Shrub cover +Topography
Habitat type + Topography
Habitat type + Grass cover
Habitat type + Grass Height
Grass Height +Grass greenness + Grass cover
Grass greenness + Grass cover
Grass greenness + Grass Height
Grass greenness + Tree cover+ Shrub cover
Grass greenness + Topography
Grass greenness +Tree cover

Table 1. Priori candidate model for predicting habitat use by introduced Plain zebra and Blue wildebeest
in MSR.

The Akaike Information Criterion (AIC) was used to select the best candidate model
(Table 2) (Burnham et al. 2011; Richards 2005; Burnham and Anderson 2004). The
model with the lowest value of AIC is the best-supported model (Burnham and
Anderson 2004). According to Burnham & Anderson (2004), models with Δ AIC ≤ 2
have substantial support (evidence), those in which 2≤ Δ AIC ≤ 7 have considerably less
support, and models having Δ AIC > 10 have essentially no support.

RESULTS

Habitat selection

The results obtained showed that, the frequency of habitat types used by Burchell’s
zebra differed significantly (χ2 = 44.39, d.f. = 3, P< 0.0001), same as was found for blue
wildebeest (χ2 =28.08, d.f. = 2, P<0.0001). Bonferroni analysis test indicated that
Burchell’s zebra prefer eucalyptus plantation and tree savanna, while blue wildebeest
prefer only tree savanna and used Semi-evergreen forest in proportion of their
availability (Table 2).
Table 2. Habitat selection by Burchell’s zebra and blue wildebeest in MSR, using selection indices and
Bonferroni simultaneous confidence intervals.

Type of habitat Expected Observed Selection Bonferroni Preferenc

proportion proportion indice intervals for (oi) e
of usage of usage (wi)
(Pi) (io)
Eucalyptus 0.009 0.055 5.510 0.012≤P≥ 0.097* +
Plantation
Burchell’s Tree savanna 0.456 0.844 1.849 0.775≤P≥ 0.912* +
zebra Shrub savanna 0.473 0.082 0.174 0.030≤P≥ 0.134* -
Semi-evergreen 0.060 0.018 0.304 -0.006≤P≥ 0.043* -
forest
Tree savanna 0.460 0.777 1.687 0.656 ≤P≥ 0.899* +
Blue Shrub savanna 0.478 0.133 0.278 0.034≤P≥0.236* -
wildebees Semi-evergreen 0.060 0.088 1.461 0.005≤P≥0.172 0
t forest

* Significant selection value: no preference (0), avoidance (-), preference (+)

Factors affecting habitat selection and model selection

The use of areas with different topography did not differ between the two grazers
species (X2= 1.8394, df =2, p=0.39).The two herbivore species was seen more in area
with midslope topography (Fig. 2a). There was no significantly difference between the
two herbivores species in term of use the area of tree canopy cover (X 2 = 2.600, df =2,
p=0.27) and shrub canopy cover (X2=0.065, df = 2, p= 0.96). Both species was seen
frequently in open areas with 0 to 10% of tree and shrub canopy cover (Fig. 2a, c).

In terms of area used by herbivores considering the tree canopy height there were
significant differences significant (X2 = 9.030, df = 2, p= 0.01). It was observed that,
plain zebra occur frequently in areas with short trees, while blue wildebeest was seen
more in areas with tall trees (Fig. 2d).

Significant difference between the two herbivores species in relation to the sites with
grass height (X2 =8.73, df = 6, p= 0.1893), grass greenness (X2=3.28, df=4, p=0.51),
grass cover (X2=1.21, df=3, p= 0.7484) and presence of mounds of termite (X2= 0.062,
df=1, p= 0.802) were not observed. Both species was frequently seen in areas with very
tall grass, high grass cover, no pattern of grass greenness was seen and both species use
more areas with no termite mound (Fig. 2e, f, g, h).
 Significant difference between the two herbivores species in relation to the distance of
water were observed (t = -2.02, def.=155.94, p= 0.044). Plain zebra was found in mean
of 0.98km from the nearest water point, while blue wildebeest was found in mean of
0.68km (Fig. 2i).

The best parsimonious models that explain the factors influencing the habitat selection
by plain zebra was the model assessing grass cover alone, tree cover alone and shrub
alone, all this model are equally supported (Δ AIC ≤ 2). For Blue wildebeest, the best-
supported model included only habitat type. (Table 3).

Table 3. Ten candidate high -ranking priority models for probability of habitat use by reintroduced Plain
Zebra and Blue widebeest in MSR relative to habitat features.

  Models K AICc Delta_AICc AICcWt Cum.Wt LL

Plain zebra              

  Grass Cover 4 206.35 1.52 0.16 0.49 -98.89

  Tree cover 3 206.71 1.88 0.13 0.62 -100.19
  Shrub cover 3 206.75 1.91 0.13 0.75 -100.21
  Topography 3 208.19 3.36 0.06 0.82 -100.93
  Grass Height 5 208.7 3.86 0.05 0.86 -98.91
  Grass Greenness 4 208.81 3.98 0.05 0.91 -100.12
  Habitat type 4 210.33 5.49 0.02 0.93 -100.88
  Grass greenness +Tree cover 6 210.6 5.76 0.02 0.95 -98.68
  Habitat type + Grass cover 7 210.8 5.96 0.02 0.97 -97.56
  Grass greenness + Grass cover 7 211.59 6.76 0.01 0.98 -97.96
Blue              
wildebeest
  Habitat type 2 80.61 1.32 0.25 0.75 -38.11
  Tree cover 3 82.29 4.00 0.07 0.81 -38.25
  Topography 3 83.51 4.22 0.06 0.87 -38.35
  Shrub cover 3 83.8 4.51 0.05 0.93 -38.5
  Habitat type + Topography 4 85.21 5.93 0.03 0.95 -37.92
  Grass Cover 4 85.76 6.48 0.02 0.97 -38.19
  Grass Greenness 4 86.3 7.01 0.01 0.99 -38.46
  Habitat type + Grass cover 5 87.95 8.66 0.01 0.99 -37.9
  Grass Height 5 89.08 9.79 0.00 1.00 -38.47
  Habitat type + Grass Height 6 91.21 11.92 0.00 1.00 -38.05
Figure 2: Comparison of habitat features used by reintroduced Plain Zebra and Blue wildebeest in MSR
DISCUSSION

Reintroduction of species to new habitat is common in Africa savanna ecosystem. In

MSR from 2010 to 2015 a total of 371 Plain zebra and 200 Blue wildebeest were
reintroduced (Hanukom and Cumbane 2014, Bodasing and Cumbane 2012), in order to
restore wildlife that will play an important role in defining the tourism product
(Stalmans 2015). This is the first study analyzing the factors affecting the habitat
selection by these two reintroduced grazers in this area.

Our findings show that zebra and wildebeest prefer more tree savanna habitat, with 0 to
10% of tree and shrub canopy cover, classified as open habitat. The result found in this
study is similar found by other studies in Africa. For example, Brooks (2005), found
zebra selecting more open-acacia vegetation in Makgadikgadi Pans National Park in
Botswana. In Kruger National Park, Macandza et al. (2012), found zebra using bush
savanna more than other habitats. in Serengeti-Mara Ndibalema (2007) found that
wildebeest prefer more open grasslands habitat and Smithers (1983) described that the
blue wildebeest can be associated in habitat like open grassland, floodplain grassland,
open bush savanna and light. The high grass forage available on tree savanna habitat
determines the preference of zebra and wildebeest for this habitat. This could occur
simply because grasses occupy a higher percentage of the habitat where tree cover is
low (Riginos and Grace 2008). In contrast, mainly closed habitat (> 25% of tree and
shrub cover) was not used frequently because dense woodlands tend to decrease grass
cover and productivity due to the decreased light intensity (Treydte et al. 2007).

Habitat features like topography, vegetation, grass, distance to the water influence the
habitat selection by these species. Both species prefer midslope topography. This type
of topography is extremely important for grazers in our area of study because of greater
forage availability. This might appear contrary to much of the literature, which report
that many species of herbivore prefer lowland topography due to its capacity to retain
green forage during the dry season and grazers are expected to focus their foraging
activities in these areas because green grass would be nutritionally advantageous (Roux
2010, Macandza et al. 2004, O’Reagain and Owen-Smith 1996). These results suggest
that, during the dry season, forage quantity is more important than quality for large
grazers, such as zebra and wildebeest, similar to what was reported by other studies
(Groom and Harris 2009, Bell 1971).

Zebra and wildebeest preferred areas where the grass height was very tall (>80cm).
Grass height is an important factor that influence habitat selection by herbivores (Laca
et al. 1992). The food intake rate increases with grass height because with increasing
grass height, bite size increases as animals apprehend a bigger volume of herbage
through increased bite depth and bite area (Illius and Gordon 1990; O’Reagain 2001).
During the dry season when forage quantity declines large grazers, concentrate foraging
on tall grass patches offering high intake rates (Wilmshurst et al. 1999, Sinclair and
Griffiths 1982).

Our results are consistent to the findings of Sinclair (1985) and Sinclair and Griffiths
(1982), who also found zebra selected more areas with tallest grass during the dry
season in the Serengeti. In other studies, Macandza et al. (2012) found zebra usually
foraging in areas where grass height ranged from 41 to 80 cm in foraging in the range
41–80 cm. Helm (2006) found that blue wildebeest prefer grass height ranging from 5 to
50 cm and black wildebeest prefer for sites where the total grass height of 5 to 80cm.
Zebra need to eat considerable quantities of grass to fulfil their nutrient requirements
(Arsenault and Owen-Smith 2008, Bell 1971), thus justifying their preference for areas
with tall grass.

In African ecosystems during the dry season, the locations of water may influence
herbivore distribution. In Kruger National Park, Redfern et al (2003) described the
hypotheses that there are correlation between species distributions and distance to water.
The author emphasizes that during the dry season when the quantity of food is reducing
wildebeest experiment a trade-off between surface-water constraints and nutritional
requirements. While for zebra this strength of trade-off is moderate when either forage
quality or quantity is reduced.

In our study, zebra was found in an average of 0.98 km from the nearest water point,
while blue wildebeest was found an average of 0.68km. Herds of water-dependent
species should occur close to the water sources (Redfern et al. 2003), These two species
are water dependent and drink at least once a day. Therefore, zebra are seldom found
more than 8km from water, while blue wildebeest seldom feed more than 15km from
water (Apps 2000). The limited availability of water during the dry season forces these
species to select areas close to water.

Brooks (2005), studying habitat selection by zebra in Makgadikgadi and Nxai Pan
National Park in Botswana found that the maximum distance from the foraging site to
the surface water was 34.5km and the distance increased from the start of the dry
season. While Helm (2006) in Ezemvelo Nature Reserve found that Blue wildebeest
favoured distances of >50 to 300 m from the nearest water.

This research showed that zebra and wildebeest using more areas with no termite
mounds. However, according Grohmann (2010) and Brody et al. (2010), termite mound
have high level of minerals such as nitrogen, phosphorous, sodium, calcium and
magnesium than in the surrounding soils. Because of this, grass growth at the termite
mound is often enhanced and increase biomass due the high concentration of nutrients,
water availability and good drainage in the termite mound than grass growth to adjacent
soil (Ashad 1982). Therefore, the vegetation communities resulting from termite mound
are considered foraging hotspots because they attract high diversity of grazers and
browsing mammals (Davies et al. 2016, Bonachela et al. 2015, Brody et al. 2010). The
high frequency of areas of use without termites shown by zebra and wildebeest can be
explained by what was reported by Ashad (1982) that the soil nutrient enrichment
effects of termites are long range and have been shown to affect areas above to 25 m
radius used in our study.

Grass cover, tree cover and shrub cover had a strong influence on habitat selection by
zebra (Δ AIC ≤ 2). This observation conforms to the findings of Gandiwa (2013), who
also found grass cover as the only environmental variable strongly correlated with large
wild herbivore density and distribution of zebra and impala in Zimbabwe. As grazers
zebra density is primarily determined by grass cover (Smuth 1974). Other study carried
out in Amboseli-Tsavo ecosystem, in Kenya, showed that herbivore grazers were found
more in areas with higher biomass of grass (Groom & Harris 2009).

For Blue wildebeest, the best-supported model included only habitat type, and we found
that tree savanna as the preferred habitat. This preference reflects the dependence of the
blue wildebeest on open grassland with trees. The importance of trees for herbivores in
African savannas was well documented by Treydte et al. (2007). These authors
concluded that trees improve grass quality, especially in dry savannas grasslands with
nutrient-poor, the greater abundance of high-quality grass species with higher level of N
and P and favorable grass structure beneath trees could attract grazing ungulates. Ben-
Shahar and Coe (1992) found that wildebeest and zebra responded to variations in level
of N and P in grasses by moving to habitat types with grasses communities with high
level of N and P. These nutrients are essential for herbivores (Owen-Smith and Novellie
1982).

This study showed the coexistence of zebra and wildebeest on some features that
influence habitat selection. Two main reasons could be responsible for this coexistence.
First, the difference on food requirement by these two species. Wildebeest is a ruminant,
need grass of high nutrient value, and is more selective than zebra as is hindgut
fermenters and can consume grass with poor nutrient value (Hack et al. 2002) .Second,
facilitation. Some authors, such as Hack et al. (2002), Sinclair and Griffiths (1982) and
Bell (1971), report that zebra play a vital role in herbivore grassland dynamics by
removing tall and fibrous grass facilitating access to short grass for more selective
ruminants, such as wildebeest and small grazers, which concentrate on the tender and
nutritious new growth. However, in our study we did not include quality of food as a
factor influencing habitat selection. Nevertheless, for next studies we suggest that this
factor must be included to better understand his effect on the habitat preferences of
these herbivores.

CONCLUSION

Overall, the present study demonstrates that tree savanna is an important habitat
affecting the distribution of reintroduced zebra and wildebeest in MSR. In addition, the
result showed that grass cover, tree cover and shrub cover has a strong influence in
habitat selection by zebra, while for wildebeest only habitat type has a strong influence.
However, there was no clear separation between the two species, but rather similarities
in the habitat features prevailing on areas selected by these species. This finding perhaps
indicates the potential for facilitation amongst these two grazers when resources are
limited.

In general, the results of this study provided valuable baseline information on habitat
use by zebra and wildebeest in MSR that would help managers of the Reserve to protect
suitable habitat, predict future population distributions and seasonal movements and
detect changes that might be occurring in this aspect in order to make effective
conservation decisions and measures.

ACKNOWLEDGEMENTS

We would like to thanks National Administration of Conservation Areas (ANAC) of

Mozambique for permission to undertake this research. A special thank you to Armando
Guenha, administrator of the Maputo Special Reserve, Natércio Ngovene, Eurico da
Paixão and Rodolfo Cumbane by facilitated our field research. Our acknowledgements
extent to Adao Ndove, Altino Cuna, Bilzia de Souza, Ivan Farinha, Sergio Ngovene and
Vilma Mafanela for their contribution in the fieldwork.

REFERENCES
1. Groom, R., & Harris, S. (2009). Factors affecting the distribution patterns of zebra
and wildebeest in a resource-stressed environment, 15289(509), 159–168.
2. De Knegt, H. J., Van Langevelde, F., Skidmore, A. K., Delsink, A., Slotow, R.,
Henley, S., Prins, H. H. T. (2011). The spatial scaling of habitat selection by African
elephants. The Journal of Animal Ecology, 80(1), 270–8.
3. Freitas, C., Kovacs, K. M., Lydersen, C., & Ims, R. a. (2008). A novel method for
quantifying habitat selection and predicting habitat use. Journal of Applied Ecology,
1365-2664.
4. Kneib, T., & Knauer, F. (2007). A General Approach for the Analysis of Habitat
Selection. Technical Report Number 001,
5. Mwangi, E. M. & Western, D. (1998). Habitat selection by large herbivores in Lake
Nakuru National Park, Kenya. Biodiversity and Conservation 7, 1-8.
6. Dekker, B., Van Rooyen, N. & Bothma, J.Du P. (1995). Habitat partitioning by
ungulates on a game ranch in the Mopani veld. South African Journal of Wildlife
Research,24(4).
7. Watson, L. H., & Chadwick, P. (2007.). Management of Cape mountain zebra in the
Kammanassie Nature Management of Cape mountain zebra in the Kammanassie
Nature Reserve , South Africa, South African Journal of Wildlife Research, 37(1):
31–39.
8. Edfern, J. E. V. R., Rant, R. I. N. A. G., & Iggs, H. A. B. (2003). Surface-water
constraints on herbivore foraging in the Kruger National Park , South Africa, 84(8),
2092–2107.
9. Senft, R.L., Coughnour, M.B., Bailey, D.W., Rittenhouse, L.R., Sala, O.E. & Swift,
D.M. (1987). Large herbivore foraging and ecological hierarchies – landscape
ecology can enhance traditional foraging theory. BioScience 37, 789–799.
10. Traill, L. (2003). Habitat partitioning, and an assessment of habitat suitability using
presence data, of a large herbivore community of a Zimabwean private wildlife
reserve. Master of Thesis, University of Stellenbosch.
11. Burkepile, D. E. E. B., Urns, C. A. E. B., Ambling, C. R. J. T., Mendola, E. L. A.,
Uis, G. R. E. G. M. B., Ovender, N. A. G., Inn, A. N. D. Z. (2013.). Habitat
selection by large herbivores in a southern African savanna: the relative roles of
bottom-up and top-down forces, 4, 1–19.
12. Riginos, C., and J. B. Grace. (2008). Savanna tree density, herbivores, and the
herbaceous community: bottom-up vs. top-down effects. Ecology 89:2228–2238.
13. Bennitt, E., Bonyongo, M. C., & Harris, S. (2014). Habitat selection by African
buffalo (Syncerus caffer) in response to landscape-level fluctuations in water
availability on two temporal scales. PloS One, 9(7).
14. Kelt, D. a., Kelly, P. a., Phillips, S. E., & Williams, D. F. (2014). Home range size
and habitat selection of reintroduced Sylvilagus bachmani riparius. Journal of
Mammalogy, 95(3), 516–524.
15. Gandiwa, E. (2013). Vegetation factors influencing density and distribution of wild
large herbivores in a southern African savannah, 274–283.
16. O’Kane, C.A.J. (2005). Habitat selection, numbers and demographics of large
mammalian herbivores in, ithala game reserve, kwazulu natal. Master thesis,
University of KwaZulu Natal.
17. Rosenzweig, M.L (1981). Theory of habitat selection. Ecology, 62(2), 327-335.
18. Jenkins, E., Silva-opps, M., Opps, S. B., Michael, R., Perrin, M. R. (2015). Home
Range and Habitat Selection of a Reintroduced African Reserve Home range and
habitat selection of a reintroduced African wild dog ( Lycaon pictus ) pack in a
small South African game reserve. African Journal of Wildlife Research 45(2): 233–
246.
19. Djagoun, C. a. M. S., Kassa, B., Djossa, B. a., Coulson, T., Mensah, G. a., & Sinsin,
B. (2014). Hunting affects dry season habitat selection by several bovid species in
northern Benin. Wildlife Biology.
20. Arthur, S. M., Manly, B. F. J., Mcdonald, L. L., & Garner, G. W. (1996). Assessing
Habitat Selection when Availability Changes Published by: Ecological Society of
America Stable. 77(1), 215–227.
21. Moyo, B., Dube, S., Lesoli, M., & Masika, P. (2013). Seasonal habitat use and
movement patterns of cattle grazing different rangeland types in the communal areas
of the Eastern Cape, South Africa, 8(1), 36–45.
22. Cherry, S. 1996. A comparison of confidence interval methods for habitat use–
availability studies. Journal of Wildlife Management 60:653–658.
23. Macandza, V. A., Owen-Smith, N., & Cain III, J. W. (2012). Dynamic spatial
partitioning and coexistence among tall grass grazers in an African savanna. Oikos,
121(6), 891–898.
24. Smithers, R. H. N. 1983. The mammals of the southern African subregion.
University of Pretoria, Pretoria.
25. Eksteen, J.J.& Bornman J.J. (1990). Analysis of natural licks at Loskop Dam and
Suikerbosrand Nature Reserves. South Africa Tyds. Natuurnav. 20 (3).
26. Gandiwa, E. (2013). Vegetation factors influencing density and distribution of wild
large herbivores in a southern African savannah.Afr. J. Ecol., 52, 274–283.
27. Conner, M., L.M, Smith, M.D 2 & BURGER. L.W. (2003). A comparison of
distance-based and classification-based analyses of habitat use, Ecology 84(2), 526–
531.
28. Bailey, D. W., Gross, J. E., Laca, E. A., Rittenhouse, L. R., Coughenour, M. B.,
Swift, D. M., (1996). Mechanisms that result in large herbivore grazing distribution
patterns. J. Range Manage.49:386-400.
29. Treydte, A. C., Heitkönig, I. M. A., & Ludwig, F. (2009). Modelling ungulate
dependence on higher quality forage under large trees in African savannahs. Basic
and Applied Ecology, 10(2), 161–169.
30. Muposhi, V. K., Chanyandura, A., Gandiwa, E., Muvengwi, J., Muboko, N., Taru,
P., & Kupika, O. L. (2014). Post-Release Monitoring of Diet Profile and Diet
Quality of Reintroduced African Buffalo (Syncerus Caffer) in Umfurudzi Park,
Zimbabwe. Tropical Conservation Science, 7(3), 440–456.
31. Iversen, M., Fauchald, P., Langeland, K., Ims, R. a, Yoccoz, N. G., & Bråthen, K.
A. (2014). Phenology and cover of plant growth forms predict herbivore habitat
selection in a high latitude ecosystem. PloS One, 9(6).
32. Prins, H. H. T., De Boer, W. F., Van Oeveren, H., Correia, A., Mafuca, J., & Olff,
H. (2006). Co-existence and niche segregation of three small bovid species in
southern Mozambique. African Journal of Ecology, 44(2), 186–198.
33. Voeten, M.M. & Prins, H. H. T. (1999). Wild and domestic herbivores between
sympatric Resource partitioning in the Tarangire region of Tanzania. Ecology,
120(2), 287–294.
34. O’Shaughnessy, R., Cain, J. W., & Owen-Smith, N. (2014). Comparative diet and
habitat selection of puku and lechwe in northern Botswana. Journal of Mammalogy,
95(5), 933–942.
35. Krebs, C. J. (2014). Ecological Methodology, Benjamin-Cummings Publishing
Company, Inc, San Francisco USA.
36. Garshelis, D. L. (2000). Delusions in habitat evaluation: measuring use, selection,
and importance.Pages 111-164 in L. Boitani e T. K. Fuller, editors. Research
Techniques in Animal Ecology: Controversies and Consequences. Columbia
University Press, New York.
37. ZAR, R.J. (2009). Biostatistical Analysis. Second edition. Prentice Hall, New
Jersey, United States of America.
38. CRAWLEY, M. J. (2013). The R book. Chichester, England, Wiley.
39. R Development Core Team (2015). R: A language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-
900051-07-0, URL http://www.R-project.org.
40. Burnham, K. P., Anderson, D. R., & Huyvaert, K. P. (2011). AIC model selection
and multimodel inference in behavioral ecology: Some background, observations,
and comparisons. Behavioral Ecology and Sociobiology, 65(1), 23–35.
41. Burnham, K. P., & Anderson, R. P. (2004). Multimodel Inference: Understanding
AIC and BIC in Model Selection. Sociological Methods & Research, 33(2), 261–
304.
42. Richards, S.A. (2005) Testing ecological theory using the information-theoretic
approach: examples and cautionary results. Ecology, 86 (10): 2805-2814.
43. Brooks, C. J. (2005). The foraging behaviour of Burchell’s zebra (Equus burchelli
antiquorum) : a dissertation submitted to the University of Bristol in accordance
with the requirements of the degree of Doctor of Philosophy in the Faculty of
Science.
44. Ndibalema, V. G. (2007). Demographic variation, distribution and habitat use
between wildebeest sub-populations in the Serengeti National Park, Tanzania.
Thesis for the degree philosophiae doctor. Department of Biology. Norwegian
University of Science and Technology.
45. Le Roux, E. (2010). Habitat and forage dependency of sable antelope (Hippotragus
niger). In the the Pretorius kop region of the Kruger National Park. A thesis
submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg,
in fulfilment of the requirements for the degree of Masters in Science.
46. Globler, J.H.(1983). Feeding Habits of the Cape Mountain Zebra Equus zebra zebra
LINN. 1758. Koedoe. 26:156-168.
47. Arshad, M.A. 1982. Influence of the termite Macrotermes michaelseni (Sjöst) on
soil fertility and vegetation in a semi-arid savannah ecosystem. Agro-ecosystems 8:
47 – 58.
48. Davies, A. B., Van Rensburg, B. J. , Robertson, M. P., Levick, S. R., & Asner, G. P.
(2016). Seasonal variation in the relative dominance of herbivore guilds in an
African savanna, 97(6), 1618–1624.
49. Marzoli, A (2007) Relatório do inventário florestal nacional. Maputo, Moçambique.
Drecção Nacional de Terras e Florestas. Ministério da Agricultura.
50. Ben-shahar, R., & Coe, M. J. (1992). The relationships between soil factors, grass
nutrients and foraging behaviour of wildebeest and zebra. Oecologia, 90:422-428.
51. Apps, P. (2000). Wild Ways: Field Guide to the Behaviour of Southern African
Mammals.Struik Publishing, South Africa.