|dc.description.abstract||In Tsavo National Park, Kenya, there has existed for some years
an acute problem that pivots on the interactions between elephants and trees.
Elephant problems are a common feature of African National Parks, but the
effects of the one in Tsavo are heightened by the marginal nature of the
environment. This study describes an attempt to throw light on certain aspects
of the ecology of Tsavo that are directly related to the elephant problem.
All the major large herbivore species are considered, with particular reference
to their numbers and distribution within the range of the Tsavo elephants and
also with reference to their exploitation of areas both affected and more or \
less unaffected by elephant destruction of woodlands.
Chapter 1 is a general introduction, both to the Park itself and
to the study. The historical background to the formation of the Park is
described, followed by a presentation of the first symptoms, in the 1950's, of
the onset of an elephant problem. The background of other research that has
been conducted in Tsavo over the last twelve years is discussed. Within the
framework of that previous research, which had mostly been on single herbivore
species, the objectives of the present work are described, being to define
the geographical functioning of the large herbivore populations, to ascertain
how all the species make use of the land both within the National Park and
around it and to use this and other information to describe the practical
implications for the conservation of large herbivores in Tsavo. The study
area, being the total range occupied by the Tsavo elephant population, is
defined, as is the large herbivore community. Data collection, using mainly
2 aerial and ground survey techniques appropriate to the large area (43,000 km ),
Chapter 2 describes the environment in which the large herbivores
live.A description of the landscape includes a presentation of the major
geological features of the area, their expression in terms of topographical
variation, the nature and distribution of the major soil types and a description
of permanent water availability. Lengthy analyses of climate are performed to
discern both spatial and temporal variation in rainfall, and also spatial
variation in the drying power of the atmosphere. The spatial analysis, using
both contour plotting and trend surface analysis of variance, shows that there
are considerable differences across the study area, the north-east being
hotter, drier and less predictable than elsewhere. Temporal analysis 6f
rainfall by Fourier analysis shows that there has been cyclical variation in
Tsavo during the course of this century, some of the energy in the power spectra
being attributable to phenomena of 36 years or more in wavelength.
Studies were performed on the spatial change across the study area,
of the physical structure of the vegetation. Attention was paid to mapping
the areas of woodland, shrub land and grassland and also the areas of tree
damage. All the information on permanent attributes of the environment,
including those variables used to describe the landscape, climate and vegetation,
is synthesized in a regional classification using principal components and
cluster analyses; the outcome of this is that there are two distinct regions
in Tsavo, one East and one West. It is stressed that these do not coincide
with the two administrative halves of the National Park, Tsavo East and Tsavo
Chapter 3 describes the animal populations. Methods of deriving
population and density estimates, on the ground and in the air, are presented.
Population estimates with confidence limits are derived for each species, being
the mean of six separate aerial censuses. Ground census work is used for the
calibration of aerial density estimates. The aerial census data on the
density of each species is separated into eight components, which are extracted
by division into two seasons, two regions and the land inside and outside the
National Park. These eight subsets of data are tested, for each species, for
significant differences. The results are presented, alongside wet and dry
season distribution maps, for each species. The structure of the community
as a whole is synthesized in an examination of biomass density in the eight
subsets of data. Significant differences are shown between them, as is the
remarkable dominance by elephants of the communities inside the Park.
Chapter 4 brings together the '. data on the environment and those on tne QeDsities of each animal species, and, by a series of multiple regression
analyses on the data, divided into eight subsets as before, attempts to
examine the relationship between them. The results of this are combined for
six of the species for which there was sufficient data, with information on
reproductive strategies. Then, for each species" a synthesis is presented of
the key features of its survival strategies in Tsavo, drawing on the data from
Chapter "3 and from this chapter, and combining it with published information
on that species, from Tsavo arid elsewhere.
Chapter 5 is a general discussion. First, the suitability of
the methods is discussed, both those of large-scale data collection and of
multivariate analysis of that data. It is concluded that the broad
geographical aims of the project justify the methods at both the collection
and analysis stages.
The influence on the animals of gradients of rainfall and of the
other climatic factors are discussed and it is shown that the distributions of
three species are directly affected by climate (two of them being absent from
the arid end of the study area and one of them increasing in density there).
The densities of most of the other species are influenced indirectly by climate
in that rainfall has a direct effect on the primary production and thence on
the carrying capacity of African savannahs.
The carrying capacity is considered further in the context of the
very reduced dry season ranges that are shown to apply to most of the large
herbivore community. It is concluded that the effective biomass densities
are very much higher than if the entire annual range is considered. In the
light of recent findings on the carrying capacity of African savannahs,
suggestions are made about the carrying capacity of Tsavo.
The cyclical interaction between elephants and trees is still an
unverified hypothesis, but data on climatic fluctuations and the patterns of
tree damage are presented which sit reasonably with certain properties of two
published mathematical models of the interaction.
The effect that the elephant damage to woodlands may have had on
herbivore density and distribution is discussed, using the multiple regression
data from the previous chapter and evidence from early this century. It is
concluded that elephants have invaded the Tsavo West grasslands during the last
70 years, displacing the other species in so doing, but altering the structure
of the habitat rather little. In the woodlands, drastic modification of tree
structural has clearly taken place, though not apparently to the detriment of
The densities and occupance of the areas inside and outside the
National Park are considered and it is concluded that the areas outside all
support substantial wildlife populations, some of them arriving there in the
wet season from inside the Park. For most species, though, the park serves a
vital role and densities inside are higher than those outside.
of this, particularly to elephant and rhino, is discussed.
The role of the human populace around the Park boundaries as
competitors for land, trees and as a hunter of elephants is both presented and
discussed. Human settlement patterns around the Park are shown to be uneven,
as are the strategies of exploitation of the land and the animals. The foci
of elephant hunting and the rate of it are shown and, from published data on the
reproductive rate of Tsavo elephants, it is demonstrated that the population
must be declining at 4% per annum, probably faster.
The foregoing information is borne in mind in a consideration of
the role of the National Park in the conservation of the large herbivore
community. Suggestions are made as to how the conservation goal might better
survive the decades ahead, both by integration of the interests of the Park and
neighbouring ranchers, and by exploitation of further areas for tourism. In
conclusion, it is suggested that the large-scale geographical approach to the
problems of herbivore populations in Tsavo has provided results that are both
valid and of considerable practical importance.||en