Content uploaded by Hervé Fritz
Author content
All content in this area was uploaded by Hervé Fritz on Feb 28, 2018
Content may be subject to copyright.
Changes in home range size of African lions in relation to pride size
and prey biomass in a semi-arid savanna
Andrew J. Loveridge*, Marion Valeix*, Zeke Davidson, Felix Murindagomo, Herve´ Fritz and
David W. Macdonald
A. J. Loveridge (andrew.loveridge@zoo.ox.ac.uk), M. Valeix, Z. Davidson and D. W. Macdonald, Wildlife Conservation Research Unit,
Zoology Dept, Oxford Univ., Tubney House, Abingdon OX13 5QL, UK. F. Murindagomo, Zimbabwe Parks and Wildlife Management
Authority, PO Box CY140, Causeway, Harare, Zimbabwe. H. Fritz, Univ. de Lyon, CNRS Univ. Claude Bernard Lyon 1 UMR 5558,
Laboratoire Biome
´trie et Biologie Evolutive, Ba
ˆt Gregor Mendel, 43 Bd du 11 novembre 1918, FR69622, Villeurbanne cedex, France.
Within-population studies are needed to investigate the extent of, and the factors underlying, intraspecific variation in
home range size. We used data from 12 female and 8 male adult lions instrumented with GPS radio-collars to describe
the ranging behaviour of lions in a population from a dystrophic semi-arid savanna, Hwange National Park, Zimbabwe.
We measured prey availability at the home range scale in 2003, 2004, and 2005. For females, home range size increased
as pride biomass increased, which is strongly suggestive of expansionism. Once controlled for pride biomass, home range
size decreased as prey biomass increased. Pride ranges responded to changes in food abundance on an annual timescale
rather than on a seasonal timescale. Female home range size was influenced by the abundance of kudu in the early dry
season, whereas it was influenced by buffalo and young elephant abundance in the late dry season. This study shows that
female home range size is mainly driven by the size of the pride, but also by prey abundance. Furthermore, female
seasonal home range size may be determined, not only by prey abundance, but also by prey dispersion in the landscape.
Home range size of males was driven by both prey biomass and the density of female prides.
Home range (sensu Burt 1943) behaviour is a common
pattern of space use and understanding variation in animal
home range size, and identifying the factors that underlie
this variation, are fundamental to understanding the
distribution and abundance of animals, and ultimately
their population regulation (Wang and Grimm 2007),
habitat selection (Rhodes et al. 2005), community structure
(Fagan et al. 2007), and for management and conservation
of ecosystems (Woodroffe and Ginsberg 2000). Interspe-
cific variation in home range size has been thoroughly
explored, particularly in terms of allometric relationships
(McNab 1963, Kelt and Van Vuren 2001). At the
intraspecific level, there has been considerable work on
variation in home range size (Kruuk and Parish 1987,
Geffen et al. 1992, Patterson and Messier 2001, Said et al.
2005) but the extent of, and mechanisms underlying,
intraspecific variation are less well understood (but see
Bo
¨rger et al. 2006a, 2008). The ranging behaviour of
animals is likely to be affected by a number of ecological
(for food see McLoughlin and Ferguson 2000, for landscape
structure see Said and Servanty 2005), demographic (for
population density see Benson et al. 2006), and behavioural
factors (for competition for territories see Rohner and Krebs
1998). Furthermore, mammals frequently exhibit distinct
inter-sexual differences in ranging behaviour. These differ-
ences are manifestations of separate selection pressures:
female reproductive success is closely tied to an ability to
exploit resources, whereas male reproductive success is
coupled with an ability to find and mate with females
(Clutton-Brock 1989). Thus, female ranging behaviour is
expected to be configured around the distribution of
resources while that of males is expected to be largely
influenced by the distribution of females.
An animal’s home range is often considered to be
mediated by the abundance of resources, as well as by their
dispersion, and the predictability with which they are
available (Macdonald and Carr 1989). Indeed, increases
in overall resource abundance may lead to smaller ranging
areas (Mills and Knowlton 1991) but, when resource
availability is heterogeneous, larger areas may still be needed
to encompass the spatial variability (e.g. patches of high
habitat quality) and temporal variability (e.g. grass renewal,
birth peaks, arrival of migratory prey) of these resources
(Macdonald and Carr 1989). The quantity, quality, and
distribution of resources in African savanna ecosystems
show wide variation in space and time, and are influenced
by the interaction between rainfall and soil characteristics
(Bell 1982, East 1984, Fritz and Duncan 1994). This
Ecography 32: 953962, 2009
doi: 10.1111/j.1600-0587.2009.05745.x
#2009 The Authors. Journal compilation #2009 Ecography
Subject Editor: Douglas A. Kelt. Accepted 7 March 2009
*Both authors contributed equally to this work
953
influences both herbivore biomass (Coe et al. 1976) and
herbivore community structure (East 1984, Fritz and
Duncan 1994), and therefore indirectly carnivore numbers
(East 1984, Grange and Duncan 2006) and potentially
carnivore home ranges.
While access to resources (food, females) is thought to be
a key determinant of home range size, in social species
ranging behaviour may also be influenced by social factors.
For instance, various relationships between group size and
home range size exist (Macdonald 1983). Kruuk and
Macdonald (1985) predict that home range behaviour in
group living species should conform to one of two
strategies. They should either expand their range to
encompass sufficient additional resources to support more
group members (expansionism) or increase group size only
up to the size that can be sustained on the resources within
the smallest economically defendable range needed
to support the minimum social unit (contractionism).
Expansionism is likely to occur where the sociological
benefits of a larger group outweigh the costs of securing and
maintaining the larger range needed to sustain it. The
dispersion of resources will surely influence the costs of
expansionism, and thus the net marginal advantage of
recruiting additional group members (Macdonald and Carr
1989). The goal of this study is to assess the role of
ecological and social factors in shaping the size of home
ranges using the example of a social large carnivore, the
African lion Panthera leo.
Intraspecific variation in home range size has been
reported for lions and home range size varies markedly
between populations across a spectrum of ecosystems (from
20 to 404 km
2
van Orsdol et al. 1985, Stander 1991,
Hanby et al. 1995, Funston and Mills 2006). Lion social
group size also varies markedly (from 2 to 35 Schaller
1972, van Orsdol et al. 1985) and there has been debate as
to the contributions of various selective pressures favouring
their sociality (for communal hunting see van Orsdol 1981
and Cooper 1991; for advantages in territoriality and
cooperative cub defence see Packer et al. 1990). Macdonald
(1983) drew attention to evidence of expansionism in the
relationship between the size of lion groups and the home
ranges they occupied. However, an inter-population com-
parison suggested that lion pride home range size was not
correlated with group size, but was negatively correlated
with lean season prey abundance (van Orsdol et al. 1985).
Furthermore, lions are opportunistic predators and seasonal
shifts in prey selection have recently been demonstrated
(Owen-Smith 2008). It is therefore possible that home
range size may respond to different factors at different
timescales. Ranging behaviour in lions is expected to differ
between the sexes as it has been suggested that male ranges
are more likely to depend on both food resources and the
need to defend and access female prides, while female
ranges are configured around access to resources (Schaller
1972). The high variability in prey abundance across
African savannas may add to the difficulty in comparing
systems, and most behavioural studies of lions have been
undertaken in eutrophic savannas (high nutrient soil
quality) such as the Serengeti-Mara ecosystem, Tanzania-
Kenya (Schaller 1972, Packer et al. 1990, Hanby et al.
1995), and to a lesser extent in mesotrophic savannas
(medium nutrient soil quality) such as the Kruger National
Park, South Africa (Funston and Mills 2006). Less is known
about lion populations in dystrophic savannas (low nutrient
soil quality) (but see Stander 1991) where the low nutrient
soil content leads to a low density herbivore community
dominated by large (African buffalo Syncerus caffer) to very
large (giraffe Giraffa camelopardalis and African elephant
Loxodonta africana) herbivores (Fritz et al. 2002).
We investigated the impact of group social structure and
spatial and temporal variation in food resources (prey
biomass) on lion home range size within a single population.
We developed a large and detailed data set of both lion
movements and the seasonal variation in abundance of their
ungulate prey in a semi-arid dystrophic savanna, Hwange
National Park, Zimbabwe, to: 1) explore the effects of group
size and prey biomass on social carnivore home range size; 2)
provide insights into the role of annual and seasonal variation
in prey biomass; 3) assess whether some prey species have a
greater seasonal influence on home range size than do others;
4) assess whether male home range size is related to prey
biomass or the need to defend access to females. Through this
study, we aim at identifying whether either or both social
factors (group size) or ecological factors (prey biomass)
determine home range size in social carnivores.
Materials and methods
Study site
The study was carried out in the northern part of Hwange
National Park (HNP). HNP covers ca 15 000 km
2
of
dystrophic savanna in north-western Zimbabwe (19800?S,
26830?E). Altitude varies from 800 to 1100 m. The vegeta-
tion is primarily woodland and bushland savanna (64%) and
vegetation communities are dominated by Colophospermum
mopane,Combretum spp., Acacia spp., Baikiaea plurijuga and
Terminalia sericea (Rogers 1993). HNP is a semi-arid
ecosystem. No surface water remains in the southern area
of the park during the dry season, except in years of
exceptionally high rainfall. In the northern area of the park,
water is artificially supplied to some waterholes (ca 50) during
the dry season. HNP research staff recorded rainfall data
daily. The long-term (19282005) mean annual rainfall is
606 mm but is highly variable (CV :30%). Annual rainfall
was calculated as the rainfall that fell between October and
September the following year. Annual rainfall was 362.6 mm
in 2003, 695.8 mm in 2004 and 287 mm in 2005, which are
the years corresponding to our study. Even though annual
rainfall was higher in 2003 than in 2005, 2003 endured a
more severe drought (it was the second of two consecutive dry
years; 2002 received 476 mm). Three seasons are distin-
guished in this study: wet season (NovemberFebruary),
early dry season (MarchJune), and late dry season (July
October). The wet season of year Y corresponds to the period
between November of year Y1 and February of year Y. Lion
density was ca 2.7 lions 100 km
2
in the northern region of
HNP (Loveridge et al. 2007a).
Lion data
From 2002 to 2005, female prides and male coalitions
were closely monitored in the northern part of HNP
954
(ca 7000 km
2
). We captured and instrumented 17 female
and 10 male adults with GPS Simplex radio-collars (female:
900 g, male: 950 g; Televilt Positioning, Lindesberg,
Sweden; see Loveridge et al. 2007a for details). Only data
from12 female and 8 male adults fitted with GPS Simplex
radio-collars were used either because data were not
available for a whole 4-month study season (the difficulty
of recapturing individuals to replace telemetry equipment
combined with a high mortality rate (cf. Loveridge et al.
2007a) led to incomplete data for some individuals) or to
avoid pseudo-replication when two females had been
collared within the same pride. Captured lions were, where
possible, weighed using a canvas stretcher attached to a
scale, suspended from poles attached to the front of a 4 4
vehicle. Mean weight (9SD) was 199910 kg for adult
males (n4) and 143911 kg for adult females (n 6).
These weights are consistent with those published for
southern African populations (Smuts et al. 1980). Posi-
tional data from the GPS Simplex radio-collars were
downloaded regularly (for each individual, one location
was available hourly from 18:00 to 7:00), and animals’
locations were available from November 2002 to October
2005. Preliminary analyses revealed that lionesses from the
same pride stay together most of the time in HNP, with
females from a pride sighted together in 89.297.4% of
sightings. Consequently, we assume that individual ranges
do generally represent pride ranges. We calculated pride
biomass using the composition of each pride based on
monthly observation. We used field data for adult weight
(see details above). For cubs and sub-adults, we used growth
equations from Smuts et al. (1980): body mass (kg)
4.21age (months)5.29 for males (r
2
0.98) and
body mass (kg)3.31 age (months) 6.64 for females
(r
2
0.99).
Analysis of home ranges
We defined home ranges and cores (sensu Powell 2000) as
the 90% (recommended by Bo
¨rger et al. 2006b) and 50%
probability contour of location distribution using the fixed
kernel density estimator, a method with recognized
strengths (Worton 1989, Powell 2000), and the reference
smoothing factor h
ref
(recommended by Hemson et al.
2005 and Bo
¨rger et al. 2006a). Home range estimates are
subject to many uncertainties. We consequently calculated
home ranges using an alternative method: the local convex
hull (LoCoH) nonparametric kernel method (Getz and
Wilmers 2004, Getz et al. 2007) with the heuristic value
kân (n is the number of points in the set). All
subsequent analyses were carried out with the two home
range estimators but when LoCoH is not specified, home
range is taken to mean the 90% kernel home range. We
investigated home range size for wet, early dry, and late dry
season. We preliminary found a very strong correlation
between the home range estimates based on all available
fixes per day and those using only one fix per day (F
1, 41
2579.1; pB0.0001; r
2
0.985). Consequently, in all
subsequent home range size analyses, we arbitrarily used
the locations taken at 00:00 h or the GPS location taken
closest to 00:00 h on that night to compare home range
estimates based rigorously on the same number of locations
(we used the same number of points, ca 120 fixes 1fixd
1
for 4 months to construct each seasonal home range).
Only animals whose GPS receiver was operational during a
whole season were used in the analyses. Home-range
analyses were undertaken using Ranges 7 (ver. 0.811, South
and Kenward 2006) for the kernel density estimator and
using the extension LoCoH v.2.1 for ArcView (ver. 3.2,
Environmental Systems Research Inst., Redlands, USA) for
the LoCoH estimator.
Prey abundance data
Many studies resorted to using habitat types as surrogate
proxies for resource availability; in contrast, we measured,
where possible, prey availability directly at the home range
scale. Given the opportunistic foraging behaviour of lions in
HNP (Loveridge et al. 2006, van Kesteren 2006), all large
herbivore species present in the study area were included in
our analysis, including elephants 54 yr-old (frequently
recorded as prey during drought years in HNP, Loveridge
et al. 2006). Species included are: African buffalo, African
elephant, blue wildebeest Connochaetes taurinus, Burchell’s
zebra Equus quagga, eland Taurotragus oryx, giraffe, greater
kudu Tragelaphus strepsiceros, impala Aepyceros melampus,
roan antelope Hippotragus equines, sable antelope Hippo-
tragus niger, warthog Phacochoerus aethiopicus, and water-
buck Kobus ellipsiprymnus. The abundance of herbivores in
the northern part of HNP has been monitored since
December 2002 with road counts (see transect coverage
in Fig. 1). Monitoring sessions were carried out in
December (wet season), MayJune (early dry season) and
SeptemberOctober (late dry season). The Main Camp area
was monitored in the three seasons, the Sinamatella area in
the early and late dry season, and the Ngamo area in the late
dry season (see areas in Fig. 1).
Calculation of prey kilometric biomass
We calculated the kilometric abundance index for each prey
species (Vincent et al. 1991, Maillard et al. 2001), which
represents an encounter rate per kilometre of road driven
and was taken as a proxy for the rate at which lions
encounter individuals of each species of prey. We then
converted the kilometric abundance index into kilometric
biomass by multiplying the unit mass (i.e. the average mass
of individuals in a population (Cumming and Cumming
2003)) by kilometric abundance for each of the prey
species. We also calculated the total prey kilometric biomass
by adding the kilometric biomasses of all the species
encountered (this measure of prey abundance is referred
to as the prey kilometric biomass hereafter). For most home
ranges (65%), we used the data from line transects
encompassed by their home ranges. However, some home
ranges included insufficient transect coverage (e.g. lionesses
F6 and F8 in Fig. 1), or corresponded to periods when no
road count data were collected (e.g. in May for animals in
the Ngamo area). For these seasonal home ranges (n 14
for females and 9 for males), we calculated the mean prey
biomass of each habitat type, and then extrapolated the prey
kilometric biomass for each home range based on the
habitat composition of the range. Habitat composition was
955
extracted from the vegetation map layer (Rogers 1993)
using ArcView 3.2. This approach was validated by
comparing the calculated prey kilometric biomasses mea-
sured by direct transects and those estimated by using
habitat type as a surrogate. In home ranges for which both
measures were made, they were positively correlated
(F
1,26
50.78; pB0.0001; r
2
0.67). No seasonal home
range overlap occurred between the marked males but some
home range overlap occurred between marked females
(21 of 42 seasonal home ranges concerned; mean9SE
1793%; range342%). We considered that resource
availability in areas of overlap was inversely proportional to
the number of prides using the area. We therefore took into
account depletion of resources by competing female groups
in the analyses by dividing the prey kilometric biomass in
the area of overlap by the number of prides potentially in
competition.
Analysis of home range size variation
Data gathered on males and females were analysed separately.
For females, we fitted a mixed linear model with restricted
maximum likelihood estimation, using seasonal home range
size as the dependent variable and individual identity as a
random factor. Explanatory variables were pride biomass
(log-transformed for linearity), prey kilometric biomass, and
the interaction between these two variables. We selected the
most likely model using AICc (Akaike information criteria
corrected for small sample size) (Burnham and Anderson
2002). Relative strength of evidence of each model was
assessed using Akaike weights (referred to as w and calculated
as exp(0.5DAICc); models giving relative strength of
evidence of 0.5 or greater can be considered as strong
contenders for the model providing the best fit). In addition,
we report R
2
values, which provide the proportion of
variance explained by the model (fixedrandom effects).
The effect of pride biomass had to be controlled before
testing for the effect of other variables so we used a type 1
approach and pride biomass was the first covariate to be
entered. We further regressed female seasonal home range
size against the density of waterholes to gain insights into the
influence of such key habitat feature. To investigate the
influence of each prey species on home range size per season,
we could not justifiably perform a mixed-model procedure
because of the low number of degrees of freedom, so we
regressed female mean seasonal home range size against the
mean kilometric biomass of each prey species.
For males, due to the smaller sample size, we also
regressed the mean seasonal home range size separately
against 1) the mean number of males in the coalition, 2) the
mean prey kilometric biomass, 3) the mean density of
waterholes in the home range, and variables linked to
females such as 4) the mean number of prides of females
encompassed, 5) the mean number of females encompassed,
6) the mean density of prides of females within the male
home range (number of prides encompassed/male home
range size), and 7) the mean density of females within the
male home range (number of females encompassed/male
home range size). When we examined the influence of
females, 8 male seasonal home ranges were excluded from
the analysis because no corresponding information was
available for females. Statistical analyses were performed
with SAS software (ver. 8.2) (SAS Inst. 1999), using REG
Figure 1. Map of the location of the mean seasonal home range for females (each shaded area represents one mean seasonal home range
for one female and labels indicate the identity of the female). Black lines represent the transects along which prey abundance was
monitored. Black dots represent pumped waterholes.
956
and MIXED procedure for normally distributed data
(Kolmogorov-Smirnov: D 0.11; p 0.15).
Results
Prey biomass
The mean total prey biomass density for the whole northern
part of HNP in October was 2312 kg km
2
(95% CI:
17593119 kg km
2
). It is noteworthy that prey abun-
dance was constant in the wet season, but prey abundance
in the study area increased in the late dry season as annual
rainfall decreased (Fig. 2a). Prey abundance was consistently
higher in the late dry season for a given year, particularly for
a dry year (Fig. 2a).
Home ranges
Mean seasonal home range size was 388 km
2
for females
(SE35 km
2
; range35981 km
2
; Table 1 for details)
and 478 km
2
for males (SE50 km
2
; range71
1002 km
2
; Table 1 for details). Home range size did not
change significantly between seasons of the same year for
either sex (paired-sampled Friedman test: females: Q
0.25; p0.88; males: Q 0.67; p 0.72; Fig. 2b, c).
Our data were sufficient to assess the effect of year on home
range size for only 6 females for which seasonal home ranges
were averaged and compared between 2003 and 2004.
Home range size for these females was significantly larger in
2004 than in 2003 (Paired sample Wilcoxon test S6; p
0.014; Fig. 2b). LoCoH seasonal home range size was
396 km
2
for females (SE33 km
2
; range101925 km
2
)
and 534 km
2
for males (SE51 km
2
; range109
1156 km
2
). Mean seasonal core size was 153 km
2
for
females (SE15 km
2
; range13384 km
2
) and 182 km
2
for males (SE24 km
2
; range20426 km
2
). The size of
LoCoH home ranges and 50% kernel cores showed the
same pattern of annual and seasonal variations as the size of
90% kernel home ranges.
Female home range size variation
The confrontation of alternative models (Table 2) did not
clearly allow us to identify whether the ‘‘pride’’ model or the
additive model ‘‘prideprey’’ was the most likely to explain
seasonal home range size (DAICc 52 and w 0.5).
Parsimony would dictate consideration of the ‘‘pride’’ model
as the best model to explain seasonal home range size. The
relationship between pride biomass and home range size was
indeed very strong, in particular for prides for which total
pride biomass is 5800 kg (the weakened relationship for
prides 800 kg is attributable largely to the behaviour of
one female F7 and to the possible existence of a threshold at
higher values of pride biomass) (Fig. 3a). However, once
controlled for pride biomass, prey kilometric biomass had a
significant effect on seasonal home range size (estimate9
SE0.09990.036; F
1,27
11.15; p0.0025). Hence,
even though pride biomass had the strongest effect on
seasonal home range size, we considered an additive model
with a positive relationship with pride biomass (Fig. 3a), and
a negative relationship with prey kilometric biomass (Fig.
3b). Because different females were sampled in different
years, our conclusions regarding ‘‘year’’ effects can only be
tentative, but it is noteworthy that 1) slopes of the relation-
ship between seasonal prey biomass and seasonal home range
size differed between wet years (2004) and dry years (2003
and 2005), and 2) for a given prey biomass, home range size
differed between years (Fig. 4b). Seasonal home range size
was negatively related to the density of waterholes in the
seasonal home range (F
1,27
14.43; p0.0007; r
2
0.67).
We found similar results for seasonal LoCoH home ranges
and 50% cores (Table 2), which also increased as the pride
biomass increased (F
1,27
24.70; pB0.0001; and F
1,27
22.60; pB0.0001 respectively) and decreased as the prey
kilometric biomass increased (F
1,27
8.83; p0.0062; and
F
1,27
6.99; p0.0135 respectively). F3 and F17 have only
a single record, so we checked that this did not bias the
(a)
(b)
(c)
Mean prey kilometric
biomass (kg km–1)
200
400
600
800
1000
1200
1400
1600
1800
2000 Wet season
Early dry season
Late dry season
2005 2003
2004
Mean +/– SE
female home range size (km2)
0
100
200
300
400
500
600
2003
2004
2005
Annual rainfall (mm)
200 300 400 500 600 700 800
Mean +/– SE
male home range size (km2)
0
200
400
600
800 2005 2003
2004
Figure 2. Relationship between annual rainfall and (a) mean prey
kilometric biomass; (b) mean female seasonal home range size (no
information is provided for the wet season 2003 because positional
data from the GPS Simplex radio-collars for female began in
February 2003); (c) mean male seasonal home range size.
957
results of the mixed model; the model excluding data from
these two females provided the same results.
A focus on prey species
Mean female home range size was clearly associated with the
mean kilometric biomass of kudu (F
1,10
10.02; p
0.012; r
2
0.53) in the early dry season, whereas it was
influenced by the mean kilometric biomass of buffalo
(F
1,10
7.89; p0.020; r
2
0.47) and juvenile elephant
(B4 yr-old) (F
1,10
5.69; p0.041; r
2
0.16) in the late
dry season; for all other species: p 0.05. There was no
significant relationship between female mean home range
size and the mean kilometric biomass of any particular prey
species in the wet season (all p]0.05).
Male home range size variation
Mean seasonal home range size for males was not influenced
by the number of males in the coalition but the relationship
approached statistical significance (F
1,6
4.05; p0.09).
There was a negative relationship between mean seasonal
home range size and mean prey kilometric biomass (F
1,6
10.52; p0.0176; r
2
0.64; Fig. 4a). There was no
significant relationship between mean seasonal home range
size and mean kilometric biomass of any particular prey
species (all p]0.05). Mean seasonal home range size was
not influenced by the density of waterholes (F
1,6
0.66;
p0.45). When we examined the influence of females on
male mean seasonal home range size, we found that the
home range size decreased as the density of prides
encompassed within the male home range increased
(F
1,4
12.87; p0.0230; r
2
0.76; Fig. 4b).
Discussion
Our results strongly suggest that pride biomass is the main
determinant of pride seasonal home range size. This is very
strongly suggestive of expansionism, with larger prides
inhabiting larger home ranges (Macdonald 1983, Kruuk
and Macdonald 1985). Because spatial and temporal
distribution of resources in semi-arid savannas is often
unpredictable, it is possible that species which form social
groups may, in these environments, adopt an expansionist
strategy. It is noteworthy that the relationship between lion
pride biomass and home range size is especially strong for
prides whose biomass is 5800 kg. This relationship
Table 1. Seasonal home range size for African lions in Hwange National Park, Zimbabwe.
Individual 90% kernel home range size (km
2
)
2003 2004 2005
wet early dry late dry wet early dry late dry wet early dry late dry
Females
F1 115 292 287 450 384 522
F3 376
F4 96 79 145 35
F5 91 171 392 509 673
F6 283 789 778
F7 159 426 366 276 189 385 450
F8 369 467 742 689
F9 981 836 656
F11 240 429 406
F15 620 347 267
F16 268 442
F17 96
Males
M1 541 785
M2 519 673 666 152
M4 833 280
M5 566 470
M6 71 1002 478 584 449
M7 346 395 554 483 187
M8 306
M9 109 336 687
Table 2. Summary statistics for models of seasonal home range size
for female African lions in Hwange National Park, Zimbabwe.
AICc DAICc w R
2
90% kernel home range
Pride 532.0 0.9 0.64 0.68
Prey 560.4 29.3 0.00 0.59
Prideprey 531.1 0 1.00 0.66
Prideprey(preyyear) 533.6 2.5 0.29 0.64
LoCoH home range
Pride 534.2 0.6 0.74 0.62
Prey 558.2 24.6 0.00 0.52
Prideprey 533.6 0 1.00 0.62
Prideprey(preyyear) 535.5 1.9 0.39 0.62
50% kernel core
Pride 468.9 0 1.00 0.66
Prey 495.5 26.6 0.00 0.54
Prideprey 470.9 2 0.37 0.62
Prideprey(preyyear) 475.9 7 0.03 0.61
Pridelog(pride biomass).
Preyprey kilometric biomass.
958
appears to reach a threshold at pride biomass equivalent to
800 kg ( 4 adult females). There are two possible reasons
for this: one is that the benefits of larger prides (e.g. for
communal hunting see van Orsdol 1981 and Cooper 1991;
for advantages in territoriality and cooperative cub defence
see Packer et al. 1990) outweigh the costs of home range
expansion up to prides of 4 adult females, but less so
thereafter, and the other is that patrolling a range of greater
than 800 km
2
becomes disadvantageous in terms of
defence, energy and maintenance of local knowledge.
Having controlled for pride biomass, pride seasonal
home range size was inversely related to prey biomass within
each home range, mirroring the general relationship
previously reported between ecosystems (van Orsdol et al.
1985). Our results further demonstrate that female seasonal
home range size was mainly determined by the abundance
of kudu in the early dry season, and of buffalo and young
elephant (B4 yr-old) in the late dry season. These findings
corroborate recent evidence showing seasonality in prey
consumed in Kruger National Park, South Africa (Owen-
Smith 2008). Buffalo and kudu are the two main prey
species in HNP (Loveridge et al. 2007b) and young
elephants make up an unusually large proportion of lion
prey in HNP in the dry season of years of low rainfall
(Loveridge et al. 2006). Both buffalo and young elephant
are dangerous species for lions to hunt. However, both are
more vulnerable to predation in the late dry season when
nutritional and water deprivation further weaken suscep-
tible individuals (young and old). Furthermore, both
elephant and buffalo aggregate in the vicinity of waterholes
during the late dry season, but disperse widely when water is
available throughout the park. Thus in the late dry season,
buffalo and young elephant can be predictably found near
waterholes and thus lions configure their ranges to ensure
access to these areas. In the wet season, when buffalo and
elephant herds have dispersed, lions must rely on more
sedentary species such as kudu, which although resident
year round are less predictably distributed, forcing lions to
alter their foraging patterns and home range use during
these periods of less predictable resource abundance.
When resources vary in availability over both the short
and the long term, it poses difficult decisions for long-lived,
territorial species whose ranges persist for longer than the
periodicity of changes in resource availability. In arid and
semi-arid savannas, such as HNP, the distribution of
herbivores in the landscape is constrained by the distribu-
tion of surface water (Redfern et al. 2003, Valeix et al.
2009), which is largely influenced, seasonally and annually,
by annual rainfall (Chamaille
´-Jammes et al. 2007, 2008). In
HNP, changes in local herbivore abundance are likely to
arise from movements of some herbivore species between
the southern sector of HNP where there is nearly no surface
water under dry conditions, but where water is widely
available in the wet season, and the northern sector of HNP
where artificially supplied waterholes ensure provision of
water throughout the year. This may explain why herbivore
(a)
(b)
Lion pride biomass (kg)
0 200 400 600 800 1000 1200 1400 1600 1800
Seasonal home range size (km
2
)
0
200
400
600
800
1000
F7 F7
F7
F7 F7
F7
Pre
y
kilometric biomass (k
g
km–1)
0 500 1000 1500 2000 2500 3000
Seasonal home range size (km
2
)
0
200
400
600
800
1000 2003
2004
2005
Figure 3. (a) Relationship between lion pride biomass and female
home range size. F7 is indicated as we believe this female can be
regarded as an outlier. (b) Relationship between prey kilometric
biomass and female home range size (see Table 3 for model
details).
(b)
Mean prey kilometric biomass (kg km
–1
)
Mean seasonal
home range size (km²)
0
100
200
300
400
500
600
700
Mean pride densit
y
(number of prides km
–2
)
0 200 400 600 800 1000 1200
0.000 0.002 0.004 0.006
Mean seasonal
home range size (km²)
0
200
400
600
800
1000
(a)
Figure 4. Regression of the mean male seasonal home range size
against (a) the mean prey kilometric biomass and (b) the mean
pride density encompassed within the male home range.
959
abundance in the ca 7000 km
2
study area was higher during
late dry seasons of years of low rainfall, and, within each
year, herbivore abundance was higher in the late dry season
than in the wet and early dry seasons. There was no
consistent seasonal difference in female home range size but
home ranges were smaller when annual rainfall was lower. It
is likely that in wet years, larger home ranges will lead to
larger overlaps between pride seasonal home ranges. These
results suggest that pride ranges respond to changes in food
abundance between years, but in the shorter term, they
remain constant between seasons, adapted perhaps to the
worst of recent conditions. Over short time-scales, it may be
too energetically costly to adapt home ranges to short-term
variations in resource availability (Macdonald and Carr
1989). Additionally, there are fewer waterholes retaining
water during the dry season in dry years than in wet years in
HNP (Chamaille
´-Jammes et al. 2007, 2008). Under dry
conditions, herbivores tend to aggregate around the few
scarce water sources (Thrash et al. 1995), where predators
have a greater chance of encountering their prey. Thus,
through its influence on surface water availability, annual
rainfall influences not only the abundance of herbivores,
but also their dispersion and the predictability with which
lions encounter them around waterholes, and ultimately
ranging behaviour of carnivores. In the case of lions in
HNP, the fact that slopes of the relationship between prey
biomass and seasonal home range size tended to differ
between years and that for a given prey biomass, home
range size tended to differ between years, suggested that the
variation cannot be attributed solely to variation in
herbivore abundance. Our findings corroborate those
from previous studies that showed that climatic variability
explains variance around the seasonal pattern in home range
size (Fisher and Owens 2000, Bo
¨rger et al. 2006a, b). For
carnivores, the effect of rainfall on home range size appears
to be indirect through prey abundance and dispersion
(Marker and Dickman 2005).
As suggested in this study, pride biomass is likely to be
the main determinant of seasonal home range size, and the
abundance, distribution and dispersion of herbivores (and
particularly of kudu in the early dry season, and of buffalo
and young elephant in the late dry season) within the
landscape, primarily determined by the distribution of
waterholes, may also significantly influence the configura-
tion of home ranges. However, it is likely that home range
size is not only determined by intra-group and ecological
factors, but also by interactions with con-specifics and
intraspecific competition for space. For instance many
studies have demonstrated that home range size in territorial
species appears to scale inversely with population density
(Krebs 1971, Erlinge et al. 1990, Makarieva et al. 2005,
Marker and Dickman 2005). Since lions are exclusively
territorial and, males particularly, will not tolerate the
presence of non-members in their territories (McComb
et al. 1994, Heinsohn 1997), it is highly likely that the
location and size of home ranges is influenced by the
interactions of resident groups with their neighbours, and
home range size is expected to be inversely related to
population density. Unfortunately, we could not investigate
these patterns in this study since most home ranges
bordered unstudied areas. However, if interactions with
conspecifics had an effect on home ranges, we suspect that
this effect was not strong because the study was carried out
in a low density lion population.
It is noteworthy that male lion density was artificially
low because of trophy hunting off-take over the study
period (Loveridge et al. 2007a), which might have allowed
males to range further in the absence of intra-sexual
competition. Still, our results showed that males differed
from females in their ranging behaviour as previously
demonstrated in several carnivore studies (Dillon and Kelly
2008). Male home range size was larger than that of females
and appeared to be affected by both prey abundance and
pride density. It is likely that home range sizes of males were
determined indirectly by prey abundance, through associa-
tion with female prides, and larger ranges are a male
reproductive strategy because they configure their home
ranges not only to secure food but also to acquire and
defend access to females (Schaller 1972, Bygott et al. 1979).
This is corroborated by the fact that males have larger home
ranges when the density of female prides is lower,
suggesting that male lions extend their home range to cover
a minimum number of prides.
Our study clearly shows that home range size in social
carnivores is primarily influenced by social factors, such as
group size which is intrinsically linked to the group
metabolic needs. It also emphasizes the role of ecological
factors, and more particularly the interaction between prey
abundance and distribution, which may further explain
home range size, shedding light on the ecological algebra of
intraspecific variation in home range size.
Acknowledgements The Director General of the Zimbabwe Parks
and Wildlife Management Authority is acknowledged for provid-
ing the opportunity to carry out this research and for permission to
publish this manuscript. We are indebted to CIRAD as herbivore
data were collected in the frame of a CIRAD-CNRS project, the
HERD project (Hwange Environmental Research Development),
funded by the French ‘‘Ministe
`re des Affaires Etrange
`res’’, the
‘‘Ambassade de France au Zimbabwe’’, the CIRAD, the CNRS,
the IFB ‘‘Global Change and Biodiversity’’, and the ANR
Biodiversite
´‘‘BioFun project’’. This work was made possible
with grants from The People’s Trust for Endangered Species, The
Darwin Initiative for Biodiversity Grant 162/09/015, The Eppley
Foundation, Disney Foundation, Marwell Preservation Trust,
Regina B. Frankenburg Foundation, Panthera Foundation, and
the generosity of Joan and Riv Winant. We thank all the people
that participated in the fieldwork, particularly Jane Hunt, Simon
Chamaille
´-Jammes and the rangers and ecological staff of HNP.
We thank Simon Chamaille
´-Jammes, our colleagues from the
WildCRU and reviewers for their fruitful comments on previous
drafts of this manuscript. Relevant animal care protocols were
followed, and approval received from the appropriate agencies.
References
Bell, R. H. V. 1982. The effect of soil nutrient availability on
community structure in African ecosystems. In: Huntley,
B. J. and Walker, B. H. (eds), Ecology of tropical savannas.
Springer, pp. 193216.
Benson, J. F. et al. 2006. Regulation of space use in a solitary felid:
population density or prey availability? Anim. Behav. 71:
685693.
960
Bo
¨rger, L. et al. 2006a. An integrated approach to identify
spatiotemporal and individual-level determinants of animal
home range size. Am. Nat. 168: 471485.
Bo
¨rger, L. et al. 2006b. Effects of sampling regime on the mean
and variance of home range size estimates. J. Anim. Ecol. 75:
13931405.
Bo
¨rger, L. et al. 2008. Are there general mechanisms of animal
home range behaviour? A review and prospects for future
research. Ecol Lett. 11: 637650.
Burnham, K. P. and Anderson, D. R. 2002. Model selection and
multimodel inference: a practical information-theoretical
approach. Springer.
Burt, W. H. 1943. Territoriality and home range concepts as
applied to mammals. J. Mammal. 24: 346352.
Bygott, J. D. et al. 1979. Male coalitions in large coalitions gain
reproductive advantages. Nature 282: 839841.
Chamaille
´-Jammes, S. et al. 2007. Climate-driven fluctuations in
surface-water availability and the buffering role of artificial
pumping in an African savanna: potential implications for
herbivore dynamics. Austral Ecol. 32: 740748.
Chamaille
´-Jammes, S. et al. 2008. Resource variability, aggrega-
tion and direct density-dependence in an open context: the
local regulation of an African elephant population. J. Anim.
Ecol. 77: 135144.
Clutton-Brock, T. H. 1989. Mammalian mating systems. Proc.
R. Soc. B 236: 339372.
Coe, M. J. et al. 1976. Biomass and production of large African
herbivores in relation to rainfall and primary production.
Oecologia 22: 341354.
Cooper, S. M. 1991. Optimal hunting group size: the need for
lions to defend their kills against loss to spotted hyaenas. Afr.
J. Ecol. 29: 130136.
Cumming, D. H. M. and Cumming, G. S. 2003. Ungulate
community structure and ecological processes: body size, hoof
area and trampling in African savannas. Oecologia 134: 560
568.
Dillon, A. and Kelly, M. J. 2008. Ocelot home range, overlap and
density: comparing radio telemetry with camera trapping. J.
Zool. 275: 391398.
East, R. 1984. Rainfall, soil nutrient and biomass of large African
savanna mammals. Afr. J. Ecol. 22: 245270.
Erlinge, S. et al. 1990. Density related home-range size and
overlap in adult field voles (Microtus agrestis) in southern
Sweden. J. Mammal. 71: 597603.
Fagan, W. F. et al. 2007. Population and community conse-
quences of spatial subsidies derived from central-place fora-
ging. Am. Nat. 170: 902915.
Fisher, D. O. and Owens, I. P. F. 2000. Female home range size
and the evolution of social organization in macropod
marsupials. J. Anim. Ecol. 69: 10831098.
Fritz, H. and Duncan, P. 1994. On the carrying capacity for large
ungulates of African savanna ecosystems. Proc. R. Soc. B
256: 7782.
Fritz, H. et al. 2002. Megaherbivores influence trophic guilds
structure in African ungulate communities. Oecologia 131:
620625.
Funston, P. and Mills, M. G. L. 2006. Behavioural ecology of
lions in Kruger National Park. In: Loveridge, A. J. et al.
(eds), Lion research conservation Workshop 3 and 4 from
conflict to socioecology. Wildlife Conservation Research Unit,
Oxford, pp. 5456.
Geffen, E. et al. 1992. Habitat selection and home range in the
Blandorf’s fox, Vulpes cana: compatibility with the resource
dispersion hypothesis. Oecologia 91: 7581.
Getz, W. M. and Wilmers, C. C. 2004. A local nearest-neighbor
convex-hull construction of home ranges and utilization
distributions. Ecography 27: 489505.
Getz, W. M. et al. 2007. LoCoH: nonparamteric kernel methods
for constructing home ranges and utilization distributions.
PlosOne 2: e207.
Grange, S. and Duncan, P. 2006. Bottom-up and top-down
processes in African ungulate communities: resources and
predation acting on the relative abundance of zebra and
grazing bovids. Ecography 29: 899907.
Hanby, J. P. et al. 1995. Ecology, demography, and behavior of
lions in two contrasting habitats: Ngorongoro Crater and the
Serengeti plains. In: Sinclair, A. R. E. and Arcese, P. (eds),
Serengeti II: dynamics, management, and conservation of an
ecosystem. Univ. of Chicago Press, pp. 315331.
Heinsohn, R. 1997. Group territoriality in two populations of
African lions. Anim. Behav. 53: 11431147.
Hemson, G. et al. 2005. Are kernel the mustard? Data from global
positioning system (GPS) collars suggests problems for kernel
home-range analyses with least-squares cross-validation. J.
Anim. Ecol. 74: 455463.
Kelt, D. A. and Van Vuren, D. H. 2001. The ecology and
macroecology of mammalian home range area. Am. Nat.
157: 637645.
Krebs, J. R. 1971. Territory and breeding density in the great tit,
Parus major L. Ecology 52: 222.
Kruuk, H. and Macdonald, D. W. 1985. Group territories of
carnivores: empires and enclaves. In: Sibly, R. M. and Smith,
R. H. (eds), Behavioural ecology. Blackwell, pp. 521536.
Kruuk, H. H. and Parish, T. 1987. Changes in the size of groups
and ranges of the European badger (Meles meles L.) in an area
in Scotland. J. Anim. Ecol. 56: 351364.
Loveridge, A. J. et al. 2006. Influence of drought on predation of
elephant (Loxodonta africana) calves by lions (Panthera leo)in
an African wooded savannah. J. Zool. 270: 523530.
Loveridge, A. J. et al. 2007a. The impact of sport-hunting on the
lion population in a protected area. Biol. Conserv. 134: 548
558.
Loveridge, A. J. et al. 2007b. Hwange Lion Project Annual Report
2007. Zimbabwe Parks and Wildlife Management Author-
ity, Zimbabwe.
Macdonald, D. W. 1983. The ecology of carnivore social
behaviour. Nature 301: 379384.
Macdonald, D. W. and Carr, G. M. 1989. Food security and the
rewards of tolerance. In: Standen, V. and Folley, R. A. (eds),
Comparative socioecology: the behavioural ecology of humans
and other mammals. Blackwell, pp. 7599.
Maillard, D. et al. 2001. The kilometric index as a monitoring tool
for population of large terrestrial animals: a feasibility test in
Zakouma National Park, Chad. Afr. J. Ecol. 39: 306309.
Makarieva, A. N. et al. 2005. Why do population density and
inverse home range scale differently with body size? Implica-
tions for ecosystem stability. Ecol. Complex. 2: 259271.
Marker, L. L. and Dickman, A. J. 2005. Factors affecting leopard
(Panthera pardus) spatial ecology, with particular reference to
Namibian farmlands. S. Afr. J. Wildl. Res. 35: 105115.
McComb, K. et al. 1994. Roaring and numerical assessment in
contests between groups of female lions, Panthera leo.Anim.
Behav. 47: 379387.
McLoughlin, P. D. and Ferguson, S. H. 2000. A hierarchical
sequence of limiting factors may help explain variation in
home range size. Ecoscience 7: 123130.
McNab, B. K. 1963. Bioenergetics and the determination of home
range size. Am. Nat. 97: 133140.
Mills, L. S. and Knowlton, F. F. 1991. Coyote space use in
relation to prey abundance. Can. J. Zool. 69: 15161521.
Owen-Smith, N. 2008. Changing vulnerability to predation
related to season and sex in an African ungulate assemblage.
Oikos 117: 602610.
Packer, C. et al. 1990. Why lions form groups: food is not enough.
Am. Nat. 136: 119.
961
Patterson, B. R. and Messier, F. 2001. Social organization and
space use of coyotes in eastern Canada relative to prey
distribution and abundance. J. Mammal. 82: 463477.
Powell, R. A. 2000. Animal home ranges and territories and home
range estimators. In: Boitani, L. and Fuller, T. K. (eds),
Research techniques in animal ecology controversies and
consequences. Columbia Univ. Press, pp. 65110.
Redfern, J. V. et al. 2003. Surface-water constraints on herbivore
foraging in the Kruger National Park, South Africa. Ecology
84: 20922107.
Rhodes, J. R. et al. 2005. A spatially explicit habitat selection
model incorporating home range behavior. Ecology 86:
11991205.
Rogers, C. M. L. 1993. A woody vegetation survey of Hwange
National Park. Dept of National Parks and Wildlife
Management, Zimbabwe.
Rohner, C. and Krebs, C. J. 1998. Response of great horned owls
to experimental ‘‘hot spots’’ of snowshoe hare density. Auk
115: 694705.
Said, S. and Servanty, S. 2005. The influence of landscape
structure on female roe deer home-range size. Landscape
Ecol. 20: 10031012.
Said, S. et al. 2005. Ecological correlates of home-range size in
spring-summer for female roe deer (Capreolus capreolus)ina
deciduous woodland. J. Zool. 267: 301308.
SAS Inst. 1999. SAS/STAT software, ver. 8. Carey, NC.
Schaller, G. B. 1972. The Serengeti lion: a study of predator prey
relations. Chicago Univ. Press.
Smuts, G. L. et al. 1980. Comparative growth of wild male and
female lions (Panthera leo). J. Zool. 190: 365373.
South, A. B. and Kenward, R. E. 2006. Ranges7 v1.0: for the
analysis of tracking and location data. Anatrack, Wareham,
UK.
Stander, P. E. 1991. Demography of lions in the Etosha National
Park, Namibia. Madoqua 18: 19.
Thrash, I. et al. 1995. Dry season herbivore densities around
drinking troughs in the Kruger National Park. J. Arid
Environ. 29: 213219.
Valeix, M. et al. 2009. Behavioral adjustments of African
herbivores to predation risk by lions: spatiotemporal variations
influence habitat use. Ecology 90: 2330.
van Kesteren, F. 2006. Analysis of the diet of African lions
(Panthera leo) in Hwange National Park, Zimbabwe. M.Sc.
thesis, Univ. of Oxford.
van Orsdol, K. G. 1981. Lion predation in Rwenzori National
Park, Uganda. Ph.D. thesis, Univ. of Cambridge.
van Orsdol, K. G. et al. 1985. Ecological correlates of lion social
organization (Panthera leo). J. Zool. 206: 97112.
Vincent, J. P. et al. 1991. Kilometric index as biological indicator
for monitoring forest roe deer populations. Acta Theriol. 36:
315328.
Wang, M. and Grimm, V. 2007. Home range dynamics and
population regulation: an individual-based model of the
common shrew Sorex ayaneus.Ecol. Model. 205: 397409.
Woodroffe, R. and Ginsberg, J. R. 2000. Ranging behaviour and
vulnerability to extinction in carnivores. In: Gosling, L. M.
and Sutherland, W. J. (eds), Behaviour and conservation.
Cambridge Univ. Press, pp. 125140.
Worton, B. J. 1989. Kernel methods for estimating the utilization
distribution in home-range studies. Ecology 70: 164168.
962
