University of Groningen
The Kenyan hippo
Kanga, Erustus Mutembei
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to
cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2011
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Kanga, E. M. (2011). The Kenyan hippo: Population dynamics, impact on riparian vegetation and conflicts
with humans Groningen: s.n.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.
Download date: 18-06-2017
Chapter four
4
Hippopotamus and livestock grazing near
water points: consequences for vegetation
cover, plant species richness and
composition in an East African Savannah
Erustus M. Kanga, Joseph O. Ogutu, Hans-Peter Piepho and Han Olff
Abstract
Grazing large mammals shape landscape heterogeneity and the composition
and structure of vegetation in savannas. The riverine systems of the Masai
Mara National Reserve of Kenya are inhabited and grazed by the common
hippopotamus all year, which seems to be especially important in vegetation
modification. However, Maasai pastoralists inhabiting pastoral rangelands
adjoining the Mara reserve have been progressively settling down from a
formerly semi-nomadic lifestyle, thus increasing concentration of livestock
grazing along the local rivers, potentially altering these effects of hippo on
the riparian vegetation. We examined if landscapes grazed predominantly by
hippopotamus and livestock differ in vegetation structure, plant species
richness and composition along distance from water sources in the wet and
dry seasons. We used 25 transects, each 5 km long and having 13 sampling
plots each measuring 10 × 10 m2 and radiating from rivers in the reserve
(n = 16) and pastoral ranches (n = 9). In each plot, we measured the height
and visually estimated the percent cover of grasses, forbs, shrubs and bare
ground, plant species composition and richness and grazing intensity. The
riverine areas in pastoral ranches were more heavily and homogonously
grazed than the reserve. The mean plant species richness was similar in
both landscapes but varied in response to spatial gradients in grazing intensity, mostly due to variation in forb and shrub species. By contrast, the plant
species composition differed strikingly between the reserve and the ranches,
while species similarity indices declined with increasing distance from
water, reflecting differential modification of vegetation structure and spatial
heterogeneity by gradients in grazing intensity. Intense and sustained
grazing shifted vegetation composition towards annuals and invasive weedy
species and grasses with short stature that characteristically exclude fires,
thus stimulating densification of woody species. Although the precise differences between hippopotamus and livestock grazing were complex and
varied with landscape and seasons, impacts of livestock grazing altered
plant species composition, the structure and spatial heterogeneity of vegetation and declined less rapidly with increasing distance from water relative
to hippopotamus grazing.
48
Chapter 4
Introduction
Grazing large herbivores influence and are reciprocally influenced by vegetation, an
interaction that shapes landscape heterogeneity, nutrient cycling, vegetation
composition, productivity and structure (McNaughton 1985; Eckert and Spencer
1987; Whicker and Detling 1988; Belsky 1992; Ritchie and Tilman 1995; Frank et
al. 1998; Olff and Ritchie 1998; Dahlberg 2000; Eccard et al. 2000; Shackleton 2000;
Thrash 2000). The impact of grazing on vegetation varies with the species, density,
body size and spatial and temporal distribution of mammalian herbivores (Belsky
1983; Bakker 1985; McNaughton 1986; Olff and Ritchie 1998; Ritchie and Olff
1999). Continuous and intense gazing can exert greater negative impacts on vegetation than can intense but periodic grazing that allows sufficient time for vegetation
recovery. Traditional pastoralists in East Africa have long been aware of this and
thus distributed the impacts of their livestock grazing by moving them periodically
between pastures and allowing sufficient time for vegetation recovery from heavy
utilization. Such mobility also enabled livestock herds to more efficiently exploit
grazing resources distributed unevenly in space and time (Homewood and Rodgers
1991; Oba et al. 2000).
Intense herbivore grazing and trampling creates and widens gaps in vegetation,
changes species colonization matrix of rangelands and differentially influences vegetation structure and species composition (Whicker and Detling 1988; Fusco et al.
1995; Zerihun and Saleem 2000; Oba et al. 2001; Thrash 2000; Van Wieren and
Bakker 2008). Moreover, the effect of grazing on species richness varies regionally
with soil productivity (Olff and Ritchie 1998), such that ecosystems with moderate to
high soil productivity have higher plant species diversity than those with lower
productivity (Olff and Ritchie 1998; Bakker 1998; Milchunas et al. 1988, 1998; Proulx
and Mazumder 1998; Bakker et al. 2006). Plant species diversity can be enhanced by
low to moderate grazing intensity (Bakker 1989; Van Wieren 1995; Hodgson and
Illius 1996) but reduced by high grazing intensity (Milchunas et al. 1988; Hobbs and
Huenneke 1992). As a result, moderate disturbances induced by intermediate grazing
intensities can support high levels of diversity in grassland communities (Grime
1973; Zeevalking and Fresco 1977; Milchunas et al. 1988; Puerto et al. 1990; Hobbs
and Huenneke 1992). Other important consequences of grazing include trampling
that enhances creation of niches for colonizing species and propagule dispersal (Olff
and Ritchie 1998; Bullock and Marriott 2000; Bakker and Olff 2003). As well, grazing
by mixed herds of different herbivore species can enhance vegetation productivity
and plant species diversity, with the effect varying depending on the complement of
herbivore species involved and the status of resources limiting plant growth, such as
water, nutrients and light (Bell 1971; Jefferies et al. 1994; Ritchie at al. 1998; Pausas
and Austin 2001). Thus, the effect of grazing on vegetation can be complex and vary
with the environment, production system, spatial and temporal scales.
Contrasting effects of hippo and livestock grazing
49
Riparian savanna habitats grazed heavily by hippopotamus (Hippopotamus
amphibious Linnaeus 1758) or livestock, experience ecological stresses through
depletion of herbaceous vegetation and increased denudation which can be detrimental to habitat suitability for other herbivores (Lock 1972; Eltringham 1999;
Fleischner 1994; Thrash 2000). Hippopotamuses and livestock sharing the same
landscape may compete for grazing resources, especially close to water, differentially modify vegetation structure and produce contrasting impacts on riparian
habitats as a result. However, despite grazing being an important evolutionary
force shaping vegetation biodiversity and structure in African savanna ecosystems
(McNaughton 1979; 1983; 1984; 1988), a comparative evaluation of the effects of
hippopotamus and livestock grazing on herbaceous plant species richness and
composition in protected and pastoral landscapes has received relatively scanty
attention. Consequently, we investigated how grazing by hippopotamus and livestock modify vegetation structure, species richness and composition along distance
from water sources in a protected and adjoining pastoral landscapes in the Mara
Region of Kenya.
We expected hippopotamus to cause fewer patches of bare ground than livestock
because of differences in their grazing strategies. Hippopotamus pluck grasses and
create patches of short grass lawns whereas livestock pull up shallow-rooted
grasses as a result of their bulk-feeding style. We further expected the varying
intensity of hippopotamus and livestock grazing to exert differential impacts on
vegetation, resulting in differences in vegetation structure, plant species abundance, richness and composition and dominance of particular plant species. More
particularly, we expected vegetation height and cover to increase with increasing
distance from water sources, due to declining grazing intensity away from water, in
both hippopotamus and livestock dominated areas. In contrast, we expected plant
species richness to increase with increasing distance from water, owing to declining
grazing intensity farther from water, and to be higher in areas grazed predominantly by hippopotamus than by livestock. Finally, we expected both hippopotamus
and livestock grazing to encourage the establishment and spread of invasive and
annual plant species through competitive release, especially in heavily grazed
areas. Testing predictions of these hypotheses is essential to enhancing our understanding of the long-term consequences of the ongoing sedenterization by pastoralists and their livestock on the biodiversity of riparian habitats, and the associated
forage resources in savanna rangelands of the Mara and possibly elsewhere.
Materials and Methods
Study area
The Mara Region (Mara) is located in south western Kenya, between latitudes
50
Chapter 4
a
o
orr
Kenya
Ol
rou
Oi
Ch
N
Lemek
20
Nairobi
19
Tanzania
18
Olkinyei
15
24
23
14
21
17 9
25
13
7
11
Koyiaki
22
12
4
10
Se
5 km transects
ren
1
get
iN
atio
5
6
3
2
nal
Pa
rk,
MMNR
Tan
z
ani
a
16
Siana
Kenya
20 km
Figure 4.1 Map of Masai Mara National Reserve and the adjoining pastoral ranches showing the
transects (numbered) radiating from rivers, sampled during 2007–2008. The study area encompassed the protected Masai Mara National Reserve and the Koyaki, Lemek and Ol Chorro Oirua
pastoral ranches.
34º45´ E and 36º00´ E and is bounded by the Serengeti National Park (SNP) in
Tanzania to the south and Siria escarpment to the west (Fig. 4.1). This region forms
the northernmost limit of the Serengeti-Mara ecosystem, covering some 25,000 km2
and straddling the Kenya-Tanzania boundary. The ecosystem comprises several
wildlife conservation administrations and conservation-pastoralist multiple land
use zones in each of the two countries (Sinclair and Arcese 1995). The Mara covers
about 5,500 km2, with the Masai Mara National Reserve (MMNR) covering some
1,530 km2 while the adjacent pastoral ranches, including Koyiaki (931 km2),
Olkinyei (787 km2), Siana (1316 km2), Lemek (717 km2) and Ol Chorro Oiroua (59
km2) cover a combined total of about 4,000 km2. The Mara receives an annual rainfall of about 600 mm in the south east, rising to about 1300 mm at the north
western edge (Norton-Griffiths et al. 1975; Ogutu et al. in press). Rainfall is
bimodal, with the short rains falling from November to December and the long
Contrasting effects of hippo and livestock grazing
51
140
long term (1990–2005)
sample period (2006–2008)
mean rainfall (mm)
120
100
80
Figure 4.2 Periodic and seasonal
rainfall distribution for the Mara
region of Kenya, during the
sampling period and long term
averages.
60
40
20
0
early dry
late dry
early wet
late wet
seasonal periods
rains from January to June, though January and February are often dry. During
the sampling period, rainfall approximated the multiyear norm (Fig. 4.2). The Mara
plains are blanketed by extensive flows of Tertiary phonolitic lava, but dark,
nutrient-rich deep clay vertisols, also called black cotton soils, are more extensive
(Williams 1964; Glover 1966). The vegetation is predominantly grassland, with
isolated scrublands and woodlands, especially along the drainage lines and on hill
tops (Epp and Agatsiva 1980).
The Mara has had a long history of harmonious coexistence between pastoralists and wildlife dating back to several millennia (Marshall 1990; McCabe 2003).
Traditional pastoralism remains the major form of land use to date, and the Mara
still supports a robust and diverse assemblage of large herbivores, with livestock
numbers having been largely stable between 1977 and 2002 despite marked
seasonal fluctuations (Broten and Said 1995; Serneels and Lambin 2001; Lamprey
and Reid 2004). However, marked declines by large wild herbivores in the Mara
during 1977-2003 have been attributed to progressive exclusion from the pastoral
ranches by land use changes, including expansion of large-scale mechanized
commercial and small-scale subsistence agriculture and settlements, human population growth, illegal wildlife harvests and marked climatic variability (Homewood
et al. 2001; Serneels and Lambin 2001; Ottichilo et al. 2001; Ogutu et al. 2009).
These changes progressively degrade, fragment and contract wildlife habitats,
thereby intensifying competition between livestock and wild herbivores in the
pastoral ranches. Together with land subdivision and privatization of land tenure
(Kimani and Prickard 1998) and sedenterization of the formerly semi-nomadic
Maasai pastoralists and the associated intensification of land use in Masailands
(Western et al. 2009), these processes progressively compress wildlife and livestock
distributions and heighten conflicts among the competing and incongruent land
uses, including along riparian habitats.
52
Chapter 4
Sampling design
We selected two landscapes; a protected conservation reserve, the Masai Mara
National Reserve (MMNR), and the adjoining community pastoral ranches of
Koyiaki, Lemek and Ol Chorro Oiroua (Fig. 4.1). Livestock grazing is prohibited in
the reserve except for illegal incursions but livestock and wildlife graze together in
the pastoral ranches. We established 25 random transects, each 5 km long and radiating from the Mara, Talek and Olare Orok Rivers. Sixteen transects were located in
areas grazed by hippopotamus and other wild herbivores, while another nine transects were placed in areas grazed by livestock, hippos and other wild herbivores
(Fig. 4.1). However, along the 5 km riparian strip, hippos and livestock are the main
resident grazers in the MMNR and the pastoral ranches, respectively. Topography
increased rather gently away from rivers within the 5 km distance sampled by
transects in both the reserve (range 1668 m to 1718 m) and the pastoral ranches
(1773 m to 1836 m).
Along each transect, we established 13 sampling plots each measuring 10 × 10 m2
at distances of 0, 100, 250, 500, 750, 1000, 1250, 1500, 2000, 2500, 3000, 4000 and
5000 meters from rivers. Because we expected vegetation to change more rapidly
close to, than far from rivers, we reduced the intensity of sampling away from
rivers to capture this gradient. In each plot, we visually estimated the percent cover
of three growth forms of vegetation (grasses, forbs and shrubs) and bare ground.
Grasses were further subdivided into three height classes: less than 10 cm tall,
10–30 cm, and greater than 30 cm. The cover measurements provided a simple,
quick and efficient method for assessing rangeland conditions. We estimated plant
species richness and abundance by identifying and counting individuals of each
plant species in each plot. Plant specimens that could not be definitively identified
in the field were taken to the East African Herbarium for further identification.
Plant nomenclature follows Agnew and Agnew (1994) and Beentje (1994). Additionally, we recorded and scored the intensity of grazing within a radius of 30 m from
the centre of each plot. The grazing score was set on a scale of one (heavily grazed)
to four (no clear evidence of grazing), based on visual assessment. Finally, because
hippopotamus can deplete herbaceous vegetation and increase denudation in areas
that they heavily utilize, we counted all hippopotamus trails within a 50 m radius
from the center of each plot. Field samplings were carried out during the early dry
(July-August) and late dry (September-October) seasons of 2007 and 2008 and in the
late wet season (March-April) of 2008. We were unable to access the study area to
obtain samples for the early wet (January-February) season of 2008 as scheduled
due to the outbreak of widespread post-election violence in Kenya at the time.
Transects were treated as the unit of replication. The total of 650 samples (n = 325
plots × 2 seasons) dropped to 634 as 16 samples (n = 8 for the reserve and n = 8 for
the ranches) were discarded because the associated plots were either burned in the
dry season, or were inaccessible in the wet season due to heavy rainfall. The 16
Contrasting effects of hippo and livestock grazing
53
transects used in the reserve therefore produced 406 samples during the wet and
dry seasons combined whereas the 9 transects used in the pastoral ranches
produced 228 samples over the same period.
Statistical analysis
Percentage cover of the five vegetation components and bare ground were arcsine
square-root transformed prior to analyses. We then used Mann-Whitney U-test to
relate grazing intensity scores to distance from water and landscapes. Further, we
used a multivariate general linear model to relate the proportions of bare ground,
vegetation cover and species richness to distance from water, landscapes, seasons
and their interactions. We calculated the mean frequency of each species at each
sampling distance from water per landscape and estimated the contribution of each
plant life form to the total species richness between landscapes, seasons and along
distance from water. To estimate the proportion of plant species shared between
landscapes, we used a simple binary measure of species presence and absence, the
Jaccard’s coefficient. This index measures the similarity in species composition
between two communities (A and B), as C = j/(a+b-j), where j is the number of
species common to both communities and a and b are the numbers of species occurring only in communities A and B, respectively. The index ranges from zero, for two
communities with no species in common, to one, for two communities with identical sets of species (Magurran 1988). Lastly, we used gradient analysis to examine
plant species compositions between the two landscapes based on species frequency
of occurrence along distance from water. First, a Detrended Correspondence
Analysis (DCA) was used to determine the length of the gradient, which was 1.3,
and therefore a Principal Components Analysis (PCA) was used for further analyses
(ter Braak 1985). Preliminary analyses showed that samples from plots on river
banks had strong effect on ordination and we therefore removed these samples
from our analysis, to show more continuous effects of gradients. We used grazing
intensity, number of hippopotamus trails and distance from water as passive environmental variables to correlated species composition along axes. All models were
fitted in Statistica (version-8, StatSoft 2007) and in PC-ORD (McCune and Mefford
2006).
Results
Grazing intensity, distribution of hippopotamus trails and vegetation cover
Grazing intensity was higher in the pastoral ranches (mean rank = 397.33, n = 228)
than in the MMNR (mean rank = 272.67, n = 406; Mann-Whitney U = 28083,
P < 0.001). The intensity of grazing did not change significantly along distance from
water in the pastoral ranches (Kendall’s tau_b = -0.069, P > 0.05, Fig. 4.3), but
54
Chapter 4
mean grazing intensity
(increasing factor 1–4)
4
3
pastoral ranches
2
1
Masai Mara National Reserve
0
0–0.75
1–2
2–3
Figure 4.3 Mean grazing intensity
as a function of distance from
water in the Mara region of Kenya
during 2007–2008. Solid and
dashed lines denote the Masai
Mara National Reserve and
pastoral ranches, respectively.
4–5
distance from water (km)
declined significantly with distance from water in the MMNR (Kendall’s tau_b =
-0.401, P < 0.001, Fig. 4.3). The mean number of hippopotamus trails was significantly higher in the MMNR (F1, 618 = 17.6, P < 0.001; 0.31±0.02, n = 406) than in
the pastoral ranches (0.14±0.01, n = 228) and declined significantly with distance
from rivers (F3, 618 = 36.2, P < 0.001), with the decline being steeper for the MMNR
(P = 0.013). Hippopotamus actively utilized a strip within 2.5 km on either side of
the rivers in both landscapes, and this strip was patchily grazed in the MMNR.
The mean percentage cover of the five components of vegetation differed significantly across seasons (F6, 616 = 11.0, P < 0.001), landscapes (F6, 616 = 19.7,
P < 0.001) and along distance from water (F18, 1742 = 8.8, P < 0.001, Fig. 4.4). The
mean percentage cover of bare ground was similar between the dry and wet
seasons but varied between landscapes and with increasing distance from water
(P < 0.001, Fig. 4.4A), such that it was lower in the pastoral ranches than in the
MMNR within 750 m from rivers but became higher in the pastoral ranches than
the MMNR at greater distances from rivers. The mean percentage cover of grasses
shorter than 10 cm was higher in the dry (0.81±0.02, n = 312) than the wet
(0.59±0.02, n = 322, P < 0.001) season and in the pastoral ranches (0.92±0.03, n =
228) than the MMNR (0.57±0.02, n = 406, P < 0.001). It declined significantly with
increasing distance from rivers (P < 0.001), but the decline was only significant for
the MMNR (Fig. 4.4B). The mean percentage cover of grasses 10–30 cm tall was
similar in both seasons and landscapes and increased significantly with distance
from water similarly in both landscapes (P < 0.001, Fig. 4.4C). The mean percentage
cover of grasses taller than 30 cm was higher in the wet (0.59±0.03, n = 322) than
the dry (0.32±0.02, n = 312, P < 0.001) season, as well as in the MMNR (0.58±0.03,
n = 406) than in the pastoral ranches (0.24±0.03, n = 228, P < 0.001). It also
increased with increasing distance from rivers (P < 0.001), but the increase was
only significant for the MMNR (Fig. 4.4D). The mean percentage cover of forbs was
higher in the wet (0.14±0.01, n = 322) than the dry (0.11±0.01, n = 312, P = 0.05)
season but similar in both landscapes and declined with increasing distance from
Contrasting effects of hippo and livestock grazing
55
water similarly in both landscapes (P < 0.001, Fig. 4.4E). The mean percentage
cover of shrubs was comparable between the dry and wet seasons but was higher
in the pastoral ranches (0.07±0.01, n = 228) than in the MMNR (0.04±0, n = 406,
P = 0.012). Furthermore, it declined significantly with distance from water, similarly in both landscapes (P < 0.001, Fig. 4.4F). A multivariate test of significance
showed that grazing intensity was significantly related to percent cover of bare
ground and vegetation categories (Wilks’ Lambda = 0.19, F18, 1768 = 76.8, P <
0.35
bare ground
0.30
0.25
acd
0.6
acd
a
0.4
d
a
d
a
a
a
0.2
0.05
0.00
mean cover (%)
b
cd
0.10
0.0
grass <10 cm
B
c
0.20
forbs
E
a
abc
abc
0.16
ab
ab
bd
bc
0.8
ab
bc
0.12
0.6
cd
d
0.08
d
0.2
0.04
0.0
0.00
grass 10–30 cm
C
0.5
pastoral ranches
Masai Mara National Reserve
0.4
cd
0.2
bcd
bcd
d
0.08
abc
bd
0.12
shrubs
F
a
abc
a
0.06
bc
ab
c
0.04
c
c
0.02
0–0.75
1–2
2–3
distance from water (km)
c
0.10
bcd
0.1
0.0
bcd
d
0.4
0.3
b
ac
0.15
1.0
b
0.8
a
D
1.0
0.20
1.2
grass >30 cm
A
b
4–5
0.00
0–0.75
1–2
2–3
c
4–5
distance from water (km)
Figure 4.4 Mean percent cover of vegetation and bare soil and interactions between landscape and
distance from water in the Mara region of Kenya during 2007–2008. Grey and white bars denote
the pastoral ranches and the Masai Mara National Reserve, respectively. Bars with similar letters
in each category denote means that did not differ significantly.
56
Chapter 4
0.001), such that the cover of bare ground (P < 0.001), grasses shorter than 10 cm (P
< 0.001), forbs (P < 0.001) and shrubs (P < 0.001) increased significantly with
increasing grazing intensity, whereas the cover of grasses between 10-30 cm (P <
0.001) and taller than 30 cm (P < 0.001) declined with increasing grazing intensity.
Species richness and floristic composition
We recorded 145 plant species belonging to 98 genera and 40 families and
comprising 31% forbs, 28% grasses, 23% trees and 18% shrubs. Plant species richness varied between landscapes, seasons and along distance from water (F13, 620 =
9.35, P < 0.001, Table 4.1). The mean number of plant species per 100 m2 plot was
higher in the wet (15.44±0.25, n = 322) than the dry (11.89±0.26, n = 312) season.
There was no significant difference in plant species richness between the MMNR
(13.86±0.21, n = 406) and the pastoral ranches (13.74±0.32, n = 228; Table 4.1;
Fig. 4.5). However, the number of plant species declined with distance from water
in both landscapes (P = 0.012, Fig. 4.5) and was significantly influenced by grazing
intensity (P < 0.001). In particular, species diversity was lower in patches of dense
vegetation cover or bare ground than in patch mosaics supporting mixed short,
medium and tall vegetation.
Growth form characterization
The mean number of forb, grass, shrub and tree species differed between landscapes (F4, 617 = 9.04, P < 0.001), seasons (F4, 617 = 29.66, P < 0.001) and along
distance from water (F12, 1632 = 4.06, P < 0.001). Grazing intensity significantly
modified the number of forb species (P < 0.001) but only marginally influenced the
grass species (P = 0.085, Table 4.2). More forb species were recorded in the wet
Table 4.1 Results of statistical tests of the effects of season, landscape, distance from water and
their interactions on plant species richness in the Mara Region of Kenya. NDF is the numerator and
DDF the denominator degrees of freedom.
Effects
NDF
DDF
Intercept
1
620
Landscape
1
620
Season
1
620
F
P
10697.12
631.17
<0.001
62.46
3.69
0.055
1294.13
76.36
<0.001
MS
Distance
3
620
62.69
3.70
0.012
Landscape × Season
1
620
15.17
0.89
0.345
Landscape × Distance
3
620
18.94
1.12
0.341
Season × Distance
3
620
12.80
0.76
0.520
Grazing intensity
1
620
183.38
10.82
0.001
Contrasting effects of hippo and livestock grazing
57
mean species richness (n/10 m2)
16
15
Figure 4.5 Mean number of
plant species per 10 m2 as a
function of distance from water
in the Mara region of Kenya
during 2007–2008. Solid and
dashed lines denote the Masai
Mara National Reserve and
pastoral ranches, respectively.
14
13
12
Masai Mara National Reserve
pastoral ranches
11
0
0.1 0.25 0.5 0.75 1 1.25 1.5
2
2.5
3
4
5
distance from water (km)
(4.54±0.13, n = 322) than the dry (3.61±0.09, n=312, P < 0.001) season, in the
MMNR (4.26±0.10, n = 406) than the pastoral ranches (3.78±0.13, n = 228,
P < 0.001) and the number of forb species declined with distance from water, similarly in both landscapes (P = 0.010, Table 4.2). The number of grass species was
higher in the wet (9.37±0.15, n = 322) than the dry (7.08±0.10, n = 312, P < 0.001)
season, marginally higher in the MMNR (8.40±0.14, n = 406) than the pastoral
ranches (7.98±0.16, n = 228, P = 0.046, Table 4.2) and did not change with
distance from water in either landscape. The number of shrub species was similar
in both seasons, higher in the pastoral ranches (1.68±0.12, n =228) than the MMNR
(1.00±0.05, n = 406, P = 0.001) and declined significantly with distance from
water in both landscapes (P < 0.001, Table 4.2). The decline was more marked in the
pastoral ranches (P = 0.008). The mean number of tree species was similar between
landscapes but declined significantly with increasing distance from water (P =
0.018) and the rate of decline was steeper for the pastoral ranches (P = 0.025, Table
4.2).
Species similarity index and grazing impact signatures
The reserve and the pastoral ranches had an overall plant species similarity index
of 0.57. This index declined consistently with distance from water such that it was
0.50, 0.49, 0.41 and 0.29 at 0–0.75, 1–2, 2–3 and 4–5 km from water, respectively.
The first two ordination (PCA) axes explained 31% of the variation in plant species
composition between the two landscapes. Grazing intensity was the most important factor influencing species composition along the first axis independent of the
number of hippo trails, while distance from water and number of hippo trails influenced species composition along the second axis (Fig. 4.6). The PCA ordination space
clearly separated the reserve and pastoral ranches based on plant species composition, with four species clusters apparent: (a) species associated with high number of
hippo trails and close to water, were exemplified by Aristida adoensis, Loudetia
58
Chapter 4
simplex, Sida ovate and Tragus berteronianus ; (b) species associated with distance
from water in areas of low number of hippo trails clustered together and included
Eragrostis racemosa, Acacia drepanolobium, and Athroisma psylloides; (c) species
responding to heavy grazing irrespective of number of hippo trails and distance
from water were represented by Cynodon dactylon, Microchloa kunthii, Sida
massaica and Solanum incanum; and (d) species common in low or moderately
grazed areas irrespective of distance from water were exemplified by Chloris
gayana, Cymbopogon spp and Setaria incrassate (Fig. 4.6).
Table 4.2 Results of statistical tests of the effects of landscape and distance from water and their
interactions on the mean number of forb, grass, shrub and tree species in the Mara region of
Kenya. NDF is the numerator and DDF the denominator degrees of freedom.
Growth form
Effects
Forbs
Shrubs
Grasses
Trees
NDF
DDF
F
P>F
Intercept
1
633
165.99
<0.001
Landscape
1
633
18.11
<0.001
Season
1
633
37.85
<0.001
Distance
3
633
3.81
0.010
Landscape × Distance
3
633
0.69
0.582
Grazing level
1
633
15.40
<0.001
Intercept
1
633
46.54
<0.001
Landscape
1
633
12.44
Season
1
633
0.005
Distance
3
633
9.53
0.001
0.936
<0.001
Landscape × Distance
3
633
3.93
0.008
Grazing level
1
633
0.48
0.492
Intercept
1
633
738.66
<0.001
Landscape
1
633
4.08
0.046
Season
1
633
101.54
<0.001
Distance
3
633
1.01
0.386
Landscape × Distance
3
633
0.50
0.684
Grazing level
1
633
2.97
0.085
Intercept
1
633
5.26
0.016
Landscape
1
633
0.32
0.495
Season
1
633
1.64
0.108
Distance
3
633
3.28
0.018
Landscape × Distance
3
633
3.18
0.025
Grazing level
1
633
2.79
0.121
Contrasting effects of hippo and livestock grazing
59
R_0.5
R_0.25
Hippo trails
Axis 2
R_0.1
MMNR
Pastoral ranches
TraguBER
SidaOVA
LoudeSIM
R_1.25
Landscape legend
AristADO
R_0.75
SidaMAS
R_1
R_1.5
MicroKUN
CymbopSP
SolanINC
P_0.5
P_0.1
Axis 1
R_2
P_0.75
ChlorGAY
SetarINC
P_1 Grazing intensity
P_1.25 CynodDAC
P_0.25
P_3
P_1.5
P_2.5
R_2.5
R_3
R_4
Distance
P_5
P_4
P_2
AcaciDRE
R_5
AthroPSY
Eragr.RAC
Figure 4.6 Ordination diagram (PCA) based on plant species frequency, showing the influence of
distance from water, grazing intensity and number of hippo trails on plant species composition in
the Mara region of Kenya. Abbreviation refer to sampling locations and species abbreviation [R and
P refer to reserve and pastoral ranches respectively and the preceding number is the location (km)
from water, while species are abbreviated with the first 5-letter and first 3-letters in caps for Latin
names].
Discussion
The high grazing intensity recorded close to water in both landscapes is characteristic of gradients created by herbivores in areas of concentrated resource utilization
(Thornton 1971; Lock 1972; Eltringham 1999; Fleischner 1994; Thrash 1998;
Thrash 2000; Oba et al. 2001), such as areas close to water, especially in the dry
season when water is limiting. However, grazing intensity declined steeply away
60
Chapter 4
from water in the reserve but not in the pastoral ranches, indicating more homogeneous distribution of grazing impact in the pastoral than the protected landscape.
The heavily grazed strip right next to water had the highest number of
hippopotamus trails in the reserve due to concentrated utilization by hippopotamus
leaving and returning to water daily. In contrast, the more widespread impact of
heavy grazing in the pastoral ranches was due to year-round grazing by resident
livestock, occurring at high densities in the pastoral ranches. The difference in
grazing intensity between the reserve and the pastoral ranches suggest that the
impact of grazing by wild herbivores is more diffusely distributed over the landscape and fades faster away from water sources, resulting in the lower grazing
intensity farther from water recorded for the reserve. However, the intense grazing
impacts within the 750 m radius from rivers in the reserve implicate repeated
grazing and trampling by hippopotamus (Thornton 1971; Lock 1972; Eltringham
1999). Higher concentrations of livestock reduced vegetation cover and plant height
over longer distances from water in the pastoral ranches than the reserve.
Repeated grazing and trampling by hippopotami leaving and returning to water
daily in the reserve created sacrificial zones devoid of most vegetation cover and
extending for nearly 750 m from river banks, similar to findings of earlier studies
(Lock 1972; Eltringam 1999; Thrash and Derry 1999). These sacrificial zones were
characteristically denuded and had the highest grazing intensity scores and percent
cover of short grasses and forbs. In contrast, the higher percent cover of bare
ground beyond 750 m from water in the pastoral ranches than the reserve can be
largely attributed to heavy and sustained grazing and trampling by large herds of
livestock. Grass height was markedly shorter closer to water in both landscapes
and declined significantly with increasing distance from water in the reserve but
not in the pastoral ranches, where short grasses dominated up to 5 km from water.
Forbs were well represented 2-3 km from water in both landscapes but their contribution to ground cover decreased sharply thereafter, implying that forbs favored
short grass lawns created and maintained by heavy grazing close to water. Shrub
cover was higher closer to water in both landscapes and farther from water in the
ranches than in the reserve, implying that heavy grazing enhanced thickening and
spread of shrubs, through suppression of fires and reduced competition with
grasses (Moleele and Perkins 1998; Thrash 1998; Oba et al. 2000; Sharp and Whittaker 2003; Lunt et al. 2007).
Grazing patterns and impacts varied spatially, creating discernible spatial
heterogeneity in vegetation structure and species composition (McNaughton 1983;
Olff and Richie 1998), especially in the reserve. Notably, areas closer to water were
typically more heavily grazed than distant areas in the reserve, similar to other
observations of utilization gradients emanating from points of countertraded use
known as piospheres (Andrew 1988; James et al. 1999). The differences in the
patterns of spatial heterogeneity in vegetation structure away from rivers between
Contrasting effects of hippo and livestock grazing
61
the two landscapes could reflect differences in hippopotamus and livestock grazing
styles and their resultant impacts on vegetation. Although high grazing intensity
maintained grass cover at relatively low levels, grasses nevertheless constituted the
major fraction of the herbaceous layer. Also, the observed seasonal variation in
vegetation cover, suggest that forage availability is limiting in the dry season as
indicated by intensification of impacts of grazing close to water.
Despite their differences in grazing intensity and vegetation structure, both
landscapes had similar plant species richness. Species richness was highest at intermediate grazing intensity scores in the reserve, consistent with predictions of the
intermediate disturbance hypothesis for mesic savannah rangelands (Grime 1973),
suggesting that moderate grazing pressure enabled coexistence of many plant
species thereby enhancing herbaceous species richness (Whicker and Detling 1988;
Olff and Richie 1998; Oba et al. 2001). However, at high and low grazing intensity
scores, species richness declined in the reserve due to domination by a few species,
as has also been observed elsewhere (Del-Val and Crawley 2005; Lunt et al. 2007).
This was not the case for the pastoral ranches where the intermediate level of
grazing intensity was not evident but species richness still followed a humped
distribution from water. This contrast further reflects the contrasting grazing
strategies of hippopotamus and livestock, the two dominant constituents of the
grazing large mammal community in the riparian-edge habitats of Masai Mara. In
the reserve, the impact of hippopotamus grazing created discrete shifting mosaics
of tall, medium and short vegetation cover that promoted the coexistence of herbaceous species, especially at moderately grazed locations from water. In the pastoral
ranches, by contrast, livestock grazing was heavy, creating wide patches of bare
ground and homogenized the structural vegetation pattern, thus enhancing the
establishment of annual and invasive species, which also elevated species richness
at intermediate distances from water similar to patterns found elsewhere (Zerihun
and Saleem 2000). Therefore, grazing of the riparian zone by hippopotamus and
livestock in the Mara had contrasting impacts but allowed diverse herbaceous
species to coexist.
The numerical predominance of forbs within heavily grazed patches is consistent with findings of several other studies (Whicker and Detling 1988; Del-Val and
Crawley 2005), while the high number of plant species recorded in the wet season
probably reflects germination of annual species (Zerihun and Saleem 2000). The
number of forb species reduced with increasing distance from water implying that
grasses outcompeted forbs in areas of low grazing intensity (Oba et al. 2001; Del-Val
and Crawley 2005). Surprisingly, the number of grass species did not change along
the distance-to-water gradient in both landscapes, suggesting that it was insensitive to average vegetation height, and was only marginally responsive to grazing
intensity. Woody species (shrubs and trees) were consistently more abundant closer
to water and in the pastoral ranches, implying that intense grazing close to water
62
Chapter 4
by hippopotamus in the reserve and livestock in the ranches kept grasses short all
year, thereby suppressing fires and stimulating encroachment of woody species
(Moleele and Perkins 1998; Oba et al. 2000; Sharp and Whittaker 2003; Lunt et al.
2007). Our results thus demonstrate that both high and low intensities of grazing
can reduce plant species richness, most notably by altering the species composition
of forbs and shrubs.
Although plant species richness was similar between the reserve and the
pastoral ranches, differences in species composition sufficiently separated the two
landscapes in the PCA ordination space. Additionally, the Jaccard’s similarity index
showed that both landscapes shared 57% of the species recorded; the similarity
index also declined considerably away from water in response to differences in
grazing gradients between the landscapes. Variations in species composition in the
reserve are explained by distance from water, number of hippo trails and grazing
intensity. In contrast, variations in species composition in the pastoral ranches are
less explained by distance from water but mostly by grazing intensity. Notably,
areas far away from water (4–5 km) in the pastoral ranches have similar species
composition to areas close to water (0.1–0.5 km) in the reserve, while species
compositions within 1–5 km in reserve and 0.1–3 km in the pastoral ranches are
distinctive to each landscape (Fig. 4.6). The differences in species composition
between the reserve and pastoral ranches are clearly discernable; and are probably
because of the added effects of heavy year-round livestock grazing in the pastoral
ranches. Large herds of migratory herbivores only use the reserve half-year,
allowing sufficient time for vegetation recovery. Therefore, we attribute differences
in species composition to the different modes and intensities of hippopotamus and
livestock grazing on vegetation structure and spatial heterogeneity between the
reserve and ranches (Thrash 2000).
Certainly, different factors are driving plant species composition between the
reserve and ranches resulting to species clusters in the landscapes. Heavily grazed
areas especially in the pastoral ranches retained species like Cynodon dactylon
which is a grazing lawn species, Microchloa kunthii which establishes well in
degraded areas, an increaser species like Sida massaica and invasive species like
Solanum incanum (Fig. 4.6). Areas highly utilized by the hippos and mostly in the
reserve had typical species such as Aristida adoensis which is common in highly
grazed areas; an increaser species like Sida ovate, weeds such as Tragus berteronianus, and a low palatable grass Loudetia simplex. Species associated with distance
from water and in areas less utilized by hippos but commonly in the pastoral
ranches included Eragrostis racemosa, Athroisma psylloides and Acacia drepanolobium. However, areas in the reserves that were moderately grazed irrespective of
distance from water had a typical species cluster that included Setaria incrassate
which only survives in moderately grazed areas; Cymbopogon spp and Chloris
gayana, a high quality pasture grass. Therefore, the data demonstrate that differen-
Contrasting effects of hippo and livestock grazing
63
tial modes and intensities of grazing by hippopotamus and livestock along riparianedge habitats of Masai Mara have significantly and differentially modified plant
species composition. Furthermore, changes in vegetation structure, plant species
richness and composition induced by hippopotamus and livestock created habitat
and forage diversity that may influence the suitability of riparian habitats for other
mammalian herbivore species in the Mara.
Conclusion
We assumed that the effects of soils and other herbivores on vegetation structure
and grazing intensity would be comparable between the reserve and pastoral
ranches because both landscapes are broadly similar, are adjacent to each other and
are not separated by physical barriers to animal movements. Therefore, patterns in
grazing intensity and vegetation structure detected along the riparian zones in the
reserve and the pastoral ranches could be, either directly or indirectly related to
impacts of hippopotamus and livestock grazing. The gradients in impacts of grazing
we observed altered vegetation structure, species richness and composition.
Despite more intense grazing and associated reduction in vegetation cover in the
pastoral ranches, species richness was similar between the reserve and the ranches,
suggesting that vegetation structure and species composition responded differentially to grazing impacts. Species richness indices concealed important shifts in
plant species composition that occurred along the distance-to-water gradient in
response to varying impacts of grazing. The increasing sedenterization of the Masai
pastoralists and their livestock in the pastoral ranches of the Mara, will progressively increase both the intensity and extent of the impact of livestock grazing,
thereby shifting the vegetation composition towards annuals and invasive weedy
species and further suppressing fires and promoting the establishment and spread of
woody species encroachment, and lowering the quality of the pastoral rangelands
for wildlife. The long-term consequences of these impacts are still unclear and hard
to predict but continued densification of woody species may reduce the suitability
of the pastoral landscapes for livestock and other grazing wild herbivores.
Acknowledgements
We thank Charles Matankory and Sospeter Kiambi for assistance with field work and for
arranging field logistics. We also thank the Kenya Wildlife Service rangers for providing
security during field work, wardens of the Masai Mara National Reserve and the management of the Koyiaki, Lemek and Ol Chorro Oiroua pastoral ranches for allowing us unlimited
access to the study area. EK was supported by the Netherlands Fellowship Program (NFP)
and the University of Groningen through the Government of Kenya and by the Frankfurt
Zoological Society (FZS).
64
Chapter 4
Contrasting effects of hippo and livestock grazing
65