Project no. FP6-018505
Project Acronym FIRE PARADOX
Project Title FIRE PARADOX: An Innovative Approach of Integrated Wildland Fire
Management Regulating the Wildfire Problem by the Wise Use of Fire:
Solving the Fire Paradox
Instrument Integrated Project (IP)
Thematic Priority Sustainable development, global change and ecosystems
Deliverable 3.5-1-40
Relationships between fire and grazing:
Determining the relationship between fire and savanna in Central and Southern
Africa
Due date of deliverable: Month 40
Actual submission date: Month 38
Start date of project: 1st March 2006
Duration: 48 months
Organisation name of lead contractor for this deliverable: Silva Forest Services
Revision (1000)
Project co-funded by the European Commission within the Sixth Framework Programme
(2002-2006)
Dissemination Level
PU
Public
PP
Restricted to other programme participants (including the Commission Services)
RE
Restricted to a group specified by the consortium (including the Commission Services)
CO
Confidential, only for members of the consortium (including the Commission Services)
PU
TABLE OF CONTENTS
Short Note
…………………………………………………………… 5
Abstract
…………………………………………………………… 5
INTRODUCTION
…………………………………………………………… 7
OBJECTIVE 1
…………………………………………………………… 7
1.1 Introduction
…………………………………………………………… 7
1.2 Fire a Natural Factor of the Environment in African Grassland & Savannas 8
1.3 Fire Effects in African Grassland & Savannas …………………………………………. 9
1.4 Ignition Sources of Fires in African Grasslands & Savannas …………………….. 10
1.5 Fire Ecology of African Grasslands & Savannas ……………………………………….10
1.5.1 Type of Fire
…………………………………………………………… 11
1.5.2 Fire Intensity
…………………………………………………………… 13
1.5.3 Season of Burning
…………………………………………………………… 15
1.5.4 Frequency of Burning
…………………………………………………………… 18
OBJECTIVE 2
…………………………………………………………… 20
OBJECTIVE 3
…………………………………………………………… 20
3.1 Introduction
…………………………………………………………… 20
3.2 Fire Herbivory Interaction Related to Domestic Livestock Grazers …………… 22
3.2.1 Humid Grasslands
…………………………………………………………… 22
3.2.1.1 Effects of Cattle & Sheep Grazing After Burning ………………………… 22
3.2.1.1.1 Treatments
…………………………………………………………… 22
3.2.1.1.2 Results
……………………………………………………………… 24
3.2.1.1.3 Conclusions
…………………………………………………………… 28
3.2.1.2 Post Fire Grazing Effects After Burning ……………………………………… 28
3.2.1.2.1 Treatments
…………………………………………………………… 29
3.2.1.2.2 Results
……………………………………………………………………... 29
3.2.1.2.3 Conclusions
…………………………………………………………… 31
3.2.2 Arid Grasslands
…………………………………………………………… 31
3.2.2.1 Treatments & Measurements ……………………………………………………. 32
3.2.2.1.1 Burning Treatment
…………………………………………………………… 33
i) Type of Fire
…………………………………………………………… 34
ii) Fire Intensity
…………………………………………………………… 34
iii) Season of Burn
…………………………………………………………… 34
3.2.2.1.2 Post-Burn Grazing Management ……………………………………………… 34
3.2.2.2 Results & Conclusions – Growing Season 2007/2008 ………………….. 35
3.3 Fire Herbivory Interaction Related to Wild Ungulate Grazing …………………… 36
3.3.1 Moist & Humid Savannas
…………………………………………………………..… 36
3.3.1.1 Effects of Wild Ungulate Grazing After Burning ………………………….. 36
3.3.1.1.2 Experimental Burn Plot Trial (EBP’s) ………………………………………. 36
3.3.1.1.2.1 Treatments
…………………………………………………………… 39
3.3.1.1.2.2 Measurements …………………………………………………………… 41
3.3.1.1.2.2.1 Pre-Treatment Measurements …………………………………… 41
3.3.1.1.2.2.2 Fire Behaviour Measurements …………………………………….. 41
3.3.1.1.2.2.3 Post-Treatment Measurements …………………………………… 41
3.3.1.1.2.3 Results
…………………………………………………………… 42
3.3.1.1.2.3.1 Grass Forage & Fuel Potential …………………………………….. 42
D3.5-1-40-1000
Page 2 of 78
3.3.1.1.2.3.2 Grass Forage Potential ……………………………………………… 44
3.3.1.1.2.3.3 Grass Fuel Potential …………………………………………………… 48
3.3.1.1.2.3.4 Basal Cover …………………………………………………………… 52
3.3.1.2 Effects of Fire & Grazing on the Botanical Composition of
the Grass Sward in Moist & Arid Savannas …………………………………. 58
3.3.1.3 Overall Conclusions on the Effects of Grazing with Wild Ungulates
After Burning on Range Condition in Moist & Arid Savannas ………… 64
3.3.1.3.1 Grass Forage Potential ………………………………………………………. 64
3.3.1.3.2 Grass Fuel Potential …………………………………………………………. 64
3.3.1.3.3 Basal Cover
…………………………………………………………… 65
3.3.1.3.4 Botanical Composition of the Grass Sward ………………………… 65
3.4 Effects of Fire & Grazing On Animal Ratios in the Ngorongoro Crater,……. 67
4.
General Discussion & Conclusions ……………………………………………………… 70
5.
References
…………………………………………………………… 74
OBJECTIVE 4
…………………………………………………………… 79
OBJECTIVE 5
…………………………………………………………… 79
D3.5-1-40-1000
Page 3 of 78
Author:
Winston Smuts Watts Trollope (P34: SFS)
Reference
Trollope, W.S.W., 2009. Relationship between fire and grazing : Report on determining the
relationship between fire and savanna in Central and Southern Africa. Deliverable D3.5-1 of the
Integrated project “Fire Paradox”, Project FP6-018505. European Commission, p 62.
Short Note:
The relationships between fire and grazing in African grasslands and savannas differ greatly in
moist and arid rangelands requiring different fire regimes for maintaining the condition of the grass
sward. This has necessitated the development of ecological criteria that provide practical guidelines
for prescribed burning in African rangelands.
Abstract:
As a means of providing an objective perspective of the relationships between fire and grazing in
African grasslands and savannas a detailed review is presented on the effects of fire per se in
terms of type and intensity of fire and season and frequency of burning on the grass sward and
tree and shrub vegetation. This provided the means of being able to differentiate between the
effects of fire alone and the effects of the burning grazing interaction on the grass sward and tree
and shrub vegetation.
Research on the effects of burning and grazing with domestic livestock (cattle/ sheep) is
generally limited to moist grasslands where it has been found that to promote animal performance
grazing must commence as soon as possible after burning in order to derive maximum benefit from
the highly nutritious and palatable regrowth of the grass sward. However, this intense grazing
gradually reduces the vigour of the grass sward and must be accompanied by extended rest
periods of a year to restore plant reserves and maintain seed reserves.
While there is limited research information on the effects of African wild ungulate grazers a
long term burning and grazing experiment in the Kruger National Park in South Africa provides a
valuable overview of the effects of season and frequency of burning and grazing on the condition
of the grass sward in terms of its forage and fuel potential and resistance to soil erosion in moist
and arid savannas.
The research results reflecting the results of the treatments after 47 years clearly indicate
that moist savannas are well adapted to fire and grazing in contrast to arid savannas that are very
sensitive to fire and grazing irrespective of season of burn and very negatively affected by frequent
fires. Nevertheless prescribed burning is an essential management practice in African grasslands
and savannas that are used for domestic livestock production and wildlife management.
Consequently a set of criteria have been developed to provide guidelines for the use of fire
as a range management practice that effectively control the season and frequency of burning that
should be applied to maintain the productivity and sustainability of the grass sward in African
grasslands and savannas.
D3.5-1-40-1000
Page 4 of 78
Finally the withdrawal of fire as a management tool in the Ngorongoro Crater in Tanzania in
the 1970’s emphasise the importance of considering the ratio of bulk to concentrate grazers when
formulating range management programs in African grasslands and savannas.
Keywords:
Fire-grazing interaction, fire ecology, fire behaviour, Africa, grassland, savanna, fire regime,
burning season, burning frequency, fire intensity, type of fire.
D3.5-1-40-1000
Page 5 of 78
INTRODUCTION
The specific objectives of the Fire Paradox project WP3.5: Relationship between fire and
grazing are as follows:
1.
To determine the relationship between fire and savanna in Central and Southern Africa
with regard to biomass, age, climate and season of burn by region;
2.
To quantify the fire/savanna relationships between the various Central and Southern
(climatic) regions;
3.
To determine the relationship between fire and different vegetation types (savanna in
Central and Southern Africa, degraded and man-treated in Chaco) with regard to biomass,
age, wildlife and domestic herbivory, climate and season of burn;
4.
To quantify the fire-tree, shrub and grass species relationships in these vegetation
types;
5.
To check about similarities between the southern and northern hemispheres in the
above matters.
The objectives will be dealt with individually.
OBJECTIVE 1: Determine the relationship between fire and savanna in central and
southern Africa with regard to biomass, age, climate and season of burn.
1.1
Introduction
This objective has been achieved by reviewing the effects of fire in African grasslands and
savannas. The effects of biomass on the fire ecology of savannas are referred to in the section on
fire intensity and its effects on the grass sward and tree/shrub vegetation. This includes
information on the effects of season and frequency of burning thereby addressing the effect of fire
relative to the age of the vegetation.
Considering the relationship between fire and savanna with regard to climate this has been
dealt with by dividing the vegetation in Central and Southern Africa into arid and moist savannas
and grasslands. Fire research in these regions of Africa has not been systematically conducted
according to climate and region hence the review of the literature and interpretation is based on
personal research experience. I considered it to be adequate and more meaningful to deal with fire
effects with this broad climatic classification.
Experience in formulating fire management plans based on the fire ecology that is described
in the review of literature have proven successful in practice using this broad classification. It must
be noted though that in the review of literature on the effects of fire in African grasslands and
savannas, this has not been limited to the effects of season of burn but has also included the
effects of type and intensity of fire and frequency of burning.
D3.5-1-40-1000
Page 6 of 78
As a result of the 11th Tall Timbers Fire Ecology Conference held in Tallahassee in 1971, it
became apparent that it is fundamentally essential to recognise that the effects of fire are
dependent on considering all the different components of the fire regime viz. Type and Intensity of
fire and Season and Frequency of burning. Personal experience in formulating and implementing
fire management plans in southern and central African grasslands and savannas has clearly shown
that it is essential to consider all four components of the fire regime when using fire as a range
management practice and interpreting its effects in grassland and savanna ecosystems. In
addressing the effects of fire in southern and central African grasslands and savannas the following
topics have been dealt with:

fire a natural factor of the environment;

fire effects;

ignition sources of fires;

fire ecology of African grasslands and savannas with respect to:
o
type of fire;
o
fire intensity;
o
season of burn;
o
frequency of burn.
This is necessary to illustrate that fire is recognised as a natural factor of the environment in
African grasslands and savannas and provides an overview of the current state of scientific
knowledge on the effects of the fire regime in these major vegetation types.
1.2
Fire a Natural Factor of the Environment in African Grasslands and Savannas
Africa is referred to as the Fire Continent (Komarek, 1965) due to the widespread occurrence of
biomass burning, particularly in the grassland and savanna biomes particularly in central and
southern Africa where it is recognised as a natural factor of the environment in these vegetation
types. The early Portuguese explorers, who rounded the Cape of Good Hope in the fifteenth
century, referred to the interior of South Africa in their ships logs as "Terra dos fumos" - the land
of smoke and fire (Scott, 1971).
This capacity of Africa to support fire stems from the fact that climatic factors are the
driving force of fire ecology and the main requirement for fire to occur anywhere on earth is to
have lightning as the primary ignition source and climatic conditions that will permit the burning of
vegetation and the spread of fires caused by lightning strikes. Africa is one of the continents that is
highly prone to lightning storms and has a fire climate comprising distinct dry and wet periods
during which times fires can burn the plant fuels produced and accumulated during the wet rainy
period, during the dry flammable period (Komarek, 1971).
The role of lightning is to balance the electrical equilibrium of the earth. As a result of the
atmosphere being able to conduct electricity to a certain extent there is a constant leakage of
electricity from the earth to the atmosphere, creating an electrical potential. When the potential is
D3.5-1-40-1000
Page 7 of 78
great enough, electricity discharges back to earth in the form of lightning. It has been estimated
that the earth would loose its electrical charge in less than an hour (48 minutes) unless it is
replenished through lightning.
Thunderstorms are therefore both a thermodynamic and electromagnetic necessity. It is
estimated that thunderstorms produce more than 8 million lightning strokes per day globally which
is equivalent to more than two thousand million kilowatt hours of electricity i.e. approximately 4,9
times the amount of electricity produced in South Africa per year (Anonymous, 1992). While
recognising the primary ignition role of lightning in causing vegetation fires in Africa the stage has
now been reached that in most regions of the world humans have become more important than
lightning as sources of ignition (Goldammer & Crutzen, 1993). This is well illustrated in the savanna
areas comprising the Kruger National Park in South Africa, where anthropogenic fires have become
the dominant ignition source of fires in that savanna community (Trollope, 1993).
The other requirement for natural fires is a fire climate which comprises dry and wet
periods so that fires can burn the plant fuels during the dry period that have been produced and
accumulated during the wet rainy period (Komarek, 1971). These climatic conditions are
characteristic of the grasslands and savannas of Africa which receive rainfall during the growing
season followed by an extended dry period during the dormant season. Lightning is further
enhanced as an ignition source by the climatic characteristic in Africa of having dry lightning storms
at the end of the dry winter period during which time the plant fuels have a very low moisture
content and are highly inflammable. Finally Africa has the most extensive area of tropical savanna
in the world which is characterised by a grassy understory that becomes extremely inflammable
during the dry season.
The afore mentioned reasons explain why Africa is regarded as the “Fire Continent” and
why fire is recognised as a natural factor of the environment and an important ecological factor in
the grassland and savanna ecosystems of the continent. This has led to research being conducted
since the early period of the twentieth century on the effects of the fire regime on the biotic and
abiotic components of grassland and savanna ecosystems. This in turn has led to a general
understanding of the effects of type and intensity of fire and season and frequency of burning on
the grass and tree components of the vegetation. This information has clarified the use of fire as a
range management practice in Africa and viable burning programs have been developed for
livestock production, game farming and nature conservation in African grasslands and savannas
(Tainton, 1999). Its use is best summed up by Phillips (1965) who described it as “a bad master
but a good servant”.
1.3
Fire Effects in African Grasslands and Savannas
Besides human activities related to urban living and agricultural production, fire is the most
widespread ecological disturbance in the world. From the artic boreal forests to the tropical
grasslands and savannas of the world, fire consumes enormous quantities of plant biomass. It has
been estimated that 2 700-6 800 million tons of plant carbon are consumed annually through the
burning of savanna vegetation and through its use in shifting agriculture. It is concluded that
human beings have used fire for over a million years and in Africa fire has extended the grasslands
and savannas at the expense of evergreen forests.
This reinforces the fundamental conclusion that fire is a general and influential ecological
phenomenon throughout the world (Bond & van Wilgen, 1996) and cannot be ignored when
D3.5-1-40-1000
Page 8 of 78
considering the management of rangeland ecosystems. It is therefore of primary importance to
obtain a clear understanding of the effects of fire in African grasslands and savannas in order to
provide a sound ecological basis to the management of these types of vegetation for both
domestic livestock and wildlife purposes.
1.4
Ignition Sources of Fires in African Grasslands and Savannas
Africa is where fire and humanity first interacted and the factor that makes fire on this continent
distinctive from other regions is the antiquity of anthropogenic fire (Pyne, 1995) If one accepts that
Homo sapiens originated through evolution in Africa then humans have evolved in a fire
environment in Africa and have had to become fire ecologists (Komarek, 1971). Be that as it may,
the earliest evidence of the use of fire by man is 1.5 million years B.P. and since then natural fire
regimes have been successively altered by humans in response to increases in the human
population. For example the majority of the tropical savannas of the world have been shaped and
maintained by anthropogenic fires. Furthermore the stage has now been reached that in most
regions of the world humans have become more important than lightning as sources of ignition
(Goldammer & Crutzen, 1993) and modern fire regimes that are not affected by anthropogenic
fires are extremely rare (Bond & van Wilgen, 1996). The dominant role of anthropogenic fires in
contemporary in Africa is illustrated by the area of savanna that was burnt by different ignition
sources in the Kruger National Park in South Africa during the period 1985 to 1992 (see Figure 1).
LIGHTNING
10%
CONTROLLED
BURNS
47%
POACHERS
20%
REFUGEES
23%
Figure 1: The percentage area burnt in the Kruger National Park in South Africa by fires ignited as
controlled burns and by refugees, poachers and lightning during the period 1985 to 1992.
The dominant and minor roles played respectively by anthropogenic (90%) and lightning (10%)
fires in the savannas in the Kruger National Park is clearly illustrated in Figure 1 and is generally
representative of the ignition sources of fires in the savanna and grassland areas of Africa.
1.5
Fire Ecology of African Grasslands and Savannas
Fire ecology refers to the response of the biotic and abiotic components of the ecosystem to the
fire regime i.e. type and intensity of fire and the season and frequency of burning (Trollope, et al,
1990). Research on the effects of fire has been conducted the grassland and savanna areas of
D3.5-1-40-1000
Page 9 of 78
Africa since the early period of this century. West (1965) reviewed the topic and found that the
first burning plots were established at Groenkloof, Pretoria in South Africa in 1916, at Olokomeji in
Nigeria in West Africa in 1929, at Ngong near Nairobi in Kenya in East Africa in 1931, at Ndola in
Zambia in Central Africa in 1933 and at the Matopos in Zimbabwe in Southern Africa in 1947.
An interesting feature about these early investigations and subsequent research up until
1971, was that it focused on addressing the two key questions of what are the effects of season
and frequency of burning on the forage production potential of the grass sward and the ratio of
bush to grass in savanna areas (West, 1965; Rose-Innes, 1971; Scott, 1971; Gill, 1981). This was
undoubtedly in response to requests from mainly agricultural scientists and livestock farmers
involved with range management who wanted to know when is the correct time to burn rangeland
and how often should the rangeland be burnt in order to maintain its forage potential and to
control bush encroachment?
Thus until recently, fire research in Africa, and in particular South Africa, was conducted
with an agricultural objective in mind rather than with the ecological objective of determining the
effect of fire on all the biotic and abiotic components of the ecosystem. This was in contrast to fire
research in other fire prone habitats like the United States and Australia where the emphasis was
on studying fire behaviour as a means of controlling wild fires. However, in 1971 a conference was
convened in the United States of America by the Tall Timbers Research Station at Tallahassee in
Florida, on the theme of "Fire in Africa". This congress was attended by fire ecologists from
throughout Africa.
The major benefit that accrued from this conference was the realization that in Africa the
study of fire behaviour and its effects on the ecosystem, as described by type and intensity of fire,
had been largely ignored in all the fire research that had been conducted up until that time. In
contrast detailed knowledge on and models for predicting fire behaviour had been developed by
the United States Forest Service (Byram, 1959; Rothermel, 1972; Brown & Davis, 1973) as a
means for controlling wildfires in the extensive forested areas of the country. A similar situation
existed in Australia where McArthur (1966), a forest fire researcher in New South Wales, had
developed procedures based on fire behaviour for decreasing the fire hazard in highly flammable
Eucalyptus forests by reducing fuel loads through controlled burning. The outcome of this congress
proved to be a turning point in fire research in the savanna and grassland areas in South Africa and
a research program was initiated to determine the effect of all the components of the fire regime
on the vegetation i.e. effects of type and intensity of fire and season and frequency of burn.
Unfortunately a similar research program was not initiated elsewhere in Africa as far as is
known, but nevertheless the aforementioned program has gone a long way in describing the
effects of the entire fire regime on the vegetation in the grassland and savanna areas of the
continent. To follow will be an overview of the known effects of the fire regime on grass and tree
and shrub vegetation in African grasslands and savannas based on research results.
1.5.1 Type of fire
The most common types of fire in grassland and savanna areas are surface fires (Trollope, 1983)
burning either as head or back fires. Crown fires do occur in savanna but only under extreme fire
conditions. Generally under these conditions they occur as passive crown fires characterised by the
“torching” of individual trees rather than as active crown fires that are sustained by more abundant
and continuous aerial fuels. Ground fires burning accumulated organic material below ground level
do occur in African grasslands and savannas but are generally rare and are more an exception
D3.5-1-40-1000
Page 10 of 78
rather than a rule. The significance of the effect of type of fire on plants is that it determines the
vertical level at which heat energy is released in relation to the location of bud tissues from which
meristematic sites the plants recover after burning.
Trollope (1978) investigated the effects of surface fires, occurring as either head or back
fires, on the grass sward in the arid savannas of the Eastern Cape Province in South Africa. The
results showed that back fires significantly depressed the regrowth of grass in comparison to head
fires because a critical threshold temperature of approximately 95O C was maintained for 20
seconds longer during back fires than during head fires. It was also found that more heat was
released at ground level during the back fires compared to the head fires, therefore the shoot
apices of the grass plants were more adversely affected during the back fires than during the head
fires.
Bush is very sensitive to various types of fires because of differences in the vertical
distribution of the release of heat energy. Field observations in the Kruger National Park and in the
Eastern Cape indicate that crown and surface head fires cause the highest topkill of stems and
branches as compared with back fires. Unfortunately there are only limited quantitative data to
support these observations. Research results were obtained from a burning trial at the University of
Fort Hare in the False Thornveld of the Eastern Cape (arid savanna) in South Africa, where a field
scale burn was applied to an area of 62 hectares to control bush encroachment. The effect of
surface head and back fires on the topkill of stems and branches of bush is presented in Table 1.
The data were collected in two metre wide belt transects laid out in the areas burnt as head and
back fires.
Table 1: The effect of surface head and back fires on the topkill of bush in the False Thornveld of
the Eastern Cape in South Africa expressed as the reduction in the number of tree equivalents - TE
(TE = tree or shrub one and a half metres high).
Transect
Type of Fire
Bush Phytomass
Bush Phytomass
TE/ha
TE/ha
Reduction
%
Length (m)
Width (m)
Before
After
Head Fire
940
2
3525
888
75
Back Fire
560
2
3407
1991
42
The majority of the trial area was burnt as a head fire and the results in Table 1 indicate that the
phytomass of bush was reduced by 75% in the area burnt as a head fire in comparison to 42% in
the area burnt as a back fire. The explanation for this is that the flame height of head fires can be
up to three times greater than for back fires, resulting in higher temperatures being generated
above ground level (Trollope, 1978). Therefore the above ground growing points of these plants,
which are located in the canopies of the trees and shrubs, are subjected to greater heat loads and
resultant damage during head fires than during back fires. This clearly illustrates the effects
different types of fire have on tree and shrub vegetation.
D3.5-1-40-1000
Page 11 of 78
Similar results were obtained in the Scattered Tree Grassland: Acacia-Themeda range type
(Edwards & Bogdan, 1951) in Kenya by Trollope & Trollope (1999) on the effects of head and back
fires on the topkill of bush (see Figure 2).
99 99
100
91
89
90
80
75
69
TOPKILL - %
70
59
60
49
50
43
HEAD FIRE
40
BACK FIRE
29
30
27
20
9
10
0
0.5
1
2
3
4
5
HEIGHT - m
Figure 2: Effect of head and back fires on the topkill of trees and shrubs of all species on the
Lewa Wildlife Conservancy and Hopcraft Ranch in the central highlands of Kenya.
The results in Figure 2 indicate that head and back fires have different effects on the topkill of bush
with head fires generally causing a greater topkill than back fires. Initially both types of fires cause
a high topkill of stems and branches when the bush is short but as the trees and shrubs increase in
height back fires cause a lower topkill compared to head fires. This trend becomes more
pronounced with trees greater than two metres in height. The reason for this is that head fires
generate greater flame heights than back fires thus resulting in the fire susceptible growing points
of taller trees and shrubs being above the flaming zone of combustion during back fires as
compared to head fires.
1.5.2 Fire intensity
Fire intensity refers to the release of heat energy per unit time per unit length of fire front (kJ/s/m)
(Byram, 1959). There have been very limited attempts in African savannas and grasslands at
quantitatively measuring the intensity of fires and relating fire intensity to the response of
herbaceous and woody plants in terms of mortality and changes in physical structure. Such
research appears to be limited to studies conducted in the savanna areas of South Africa.
The effect of fire intensity on the recovery of the grass sward after burning was
investigated in the arid savannas of the Eastern Cape Province. After a series of fires ranging in
intensity from 925 to 3 326 kJ/s/m (cool to extremely intense) there were no significant differences
in the recovery of the grass sward at the end of the first or second growing seasons after the burns
(Trollope & Tainton, 1986) leading to the conclusion that fire intensity has no significant effect on
the recovery of the grass sward after a burn. This is a logical result as otherwise intense fires
would not favour the development and maintenance of grassland.
D3.5-1-40-1000
Page 12 of 78
The effect of fire intensity on bush has been studied in the arid savannas of the Eastern
Cape Province (Trollope & Tainton, 1986) and the Kruger National Park (Trollope, Potgieter &
Zambatis, 1990) in South Africa. This comprised determining the mortality of plants and secondly
the total topkill of stems and branches of bush of different heights. The results indicated that bush
is very resistant to fire alone and in the Eastern Cape the mortality of bush after a high intensity
fire of
3 875 kJ/s/m was only 9,3 per cent. In the Kruger National Park the average mortality
of 14 of the most common bush species subjected to 43 fires ranging in fire intensity from 110 to 6
704 kJ/s/m was only 1,3 per cent. In both areas the majority of the trees that suffered a topkill of
stems and branches coppiced from the collar region of the stem. Therefore it can be concluded
that, generally, the main effect of fire on bush in the savanna areas is to cause a topkill of stems
and branches forcing the plants to coppice from the collar region of the stem.
The detailed results of this study are illustrated in Figure 3.
79
80
72
67
70
64
62
61
TOPKILL - %
60
51
52
53
2000
3000
4000
53
47
50
40
40
30
20
10
0
500
1000
FIRE INTENSITY - kJ/s/m
5000
EASTERN CAPE
KRUGER NATIONAL PARK
Figure 3: Effect of fire intensity on the topkill of bush two metres high in the Eastern Cape
Province and Kruger National Park in South Africa.
The results in Figure 3 show that there was a significantly greater topkill of bush with increasing
fire intensities. However, the research also showed that the bush became more resistant to fire as
the height of the trees and shrubs increased and this is illustrated in Figure 4.
D3.5-1-40-1000
Page 13 of 78
100
90
92
87
TOPKILL - %
80
70
70
60
48
50
40
29
30
16
20
13
10
8
6
4
0
.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00
1 - .01 - .51 - .01 - .51 - .01 - .51 - .01 - .51 0.5
1
1
2
2
3
3
4
4
0
0-
HEIGHT CLASSES - m
Figure 4: Effect of height on the topkill of bush subjected to a fire intensity of 3 000 kJ/s/m in the
Kruger National Park in South Africa.
Similar responses were obtained in the arid savannas of the Eastern Cape Province in South Africa
(Trollope & Tainton, 1986) and in the Scattered Tree Grassland: Acacia-Themeda savanna in the
central highlands of Kenya (Trollope, & Trollope, 1999).
1.5.3 Season of burning
The temporal and spatial distribution of savanna fires in Africa has been studied by Cahoon, et al,
(1992) using night time satellite imagery. The results indicate that most fires are left to burn
uncontrolled so that there is no strong diurnal cycle in the fire frequency. The analysis of monthly
satellite images for the period 1986 to 1987 indicates that January is the peak season for African
savanna fires burning north of the equator when all the savannas in these areas receive less than
25 mm of rain.
These fires occur in a wide band stretching across the savannas south of the Sahara desert
and the majority of them are initiated by human activities. Rainfall then increases in both
hemispheres and savanna burning is reduced to a minimum during April. After April the rainfall
decreases in the southern hemisphere and the frequency of savanna fires increases initially to a
maximum during June in the western regions of southern Africa. From July to October burning in
the savannas increases in the eastern portions and during the dry conditions prevailing during
August and September fire activity reaches a maximum in East Africa and northern Mozambique.
An interesting difference in the season of burning exists between fires ignited by lightning
and anthropogenic sources. Studies in the Kruger National Park in South Africa for the period 19801992 showed that lightning fires occurred most frequently during spring and summer (October to
January) when thunderstorms are most frequent. Conversely anthropogenic fires occurred mainly
D3.5-1-40-1000
Page 14 of 78
during the mid-winter to early spring period (June to September) (Trollope, 1993). These results
are clearly illustrated in Figure 5.
25
24
23
21
PERCENTAGE
20
16
15
15
14
14
13
10
10
7
6
6
5
5
5
3
3
33
1
1
1
1
1
J
J
A
2
0
J
F
M
A
M
MONTH
S
O
N
D
LIGHTNING
ANTHROPOGENIC
Figure 5: The season of burning of lightning and anthropogenic ignited fires recorded in the
Kruger National Park during the period 1980 to 1992.
Very little published quantitative information is available on the effect of season of burning on the
grass sward. West (1965) stressed the importance of burning when the grass is dormant. Scott
(1971) quoted data from the Southern Tall Grassveld of the KwaZulu/ Natal Province in South
Africa where the mean grass basal cover of plots burnt in autumn, late winter and after the first
spring rains for a period exceeding 20 years, was 12.8, 13.0 and 14.4 per cent respectively. The
absence of large differences in the mean basal cover obtained with these different seasons of
burning indicated that for all practical purposes burning when the grass sward is dormant in late
winter or immediately after the first spring rains had very little difference in effect on the grass
sward.
This conclusion is supported by Tainton, Groves and Nash (1977), Dillon (1980) and
Everson, Everson, Dicks & Poulter (1988) who also found that burning before or immediately after
the first spring rains in KwaZulu/Natal had essentially the same effect on the recovery of burnt
veld. Conversely, if the veld is burnt later in the season when it is actively growing it causes a high
mortality of tillers of Themeda triandra, resulting in a significant reduction in the abundance of this
species (Dillon, 1980; Everson, Everson, Dicks & Poulter, 1988).
D3.5-1-40-1000
Page 15 of 78
The effect of season of burning on the recovery of grass was also investigated in the arid
savannas of the Eastern Cape (Trollope, 1987). This comprised determining the effect of burning
the grass sward in late winter, spring, late spring and early summer. The results are presented in
Figure 6.
3777
STANDING CROP GRASS - kg/ha
4000
3194
3500
3000
3117
2806
2554
2418
2500
2126
1719
2000
1500
1000
500
0
LATE WINTER
SPRING
LATE SPRING
BURNING SEASON
EARLY
SUMMER
FIRST SEASON
SECOND SEASON
Figure 6: Effect of season of burning on the grass sward in the arid savannas of the Eastern Cape.
Standing crop of grass sward expressed in kilograms per hectare.
The results in Figure 6 show that burning in late winter consistently resulted in a significantly better
recovery in the grass sward during the first growing season after the burn than the other
treatments. This effect was still present during the second growing season but was not as evident
as during the first growing season. Conversely the early summer burns that were applied when the
grass was actively growing had a significantly depressive effect throughout the recovery period on
the regrowth of the grass sward in relation to the other treatments.
Thus the overall effect of the treatments was that burning when the grass was actively
growing adversely affected the recovery of the grass sward when compared with burning when the
grass was dormant. A possible explanation for these results is that observations and measurements
made at the time of the early summer burn showed that the grass tillers were actively growing and
the shoot apices were therefore probably elevated and in a vulnerable position to be damaged by
the fire. Furthermore the mean rate of spread for the early summer burns was 0,11 m/s compared
to 0,31 m/s for the later winter burns. This would suggest that the slow moving early summer burn
resulted in a longer duration of critical threshold temperatures compared to the fast moving winter
burn and therefore had a greater damaging effect on the exposed shoot apices (Trollope, 1987).
Subsequent investigations have confirmed this in that it was found that rate of spread was
significantly positively correlated with the recovery of the grass sward during the first growing
season after the burn. These different sources of evidence lend support to the view that the effect
D3.5-1-40-1000
Page 16 of 78
of season of burning on the grass sward depends upon the physiological state of the grass at the
time of the fire.
Season of burning also has an effect on the botanical composition of the grass sward. It
was found in Kwazulu/Natal Province in South Africa that Themeda triandra declined after burning
in autumn in comparison to burning in winter and spring whereas Tristachya leucothrix responded
in the exact opposite manner (Bond & van Wilgen, 1996).
It is difficult to ascertain the effect of season of burning on bush because generally it is
confounded with fire intensity. This is because when the trees are dormant in winter the grass is
dry and supports intense fires whereas when the trees are actively growing during summer the
grass is green and the fires are much cooler. Suffice it so say that West (1965) postulated that
trees and shrubs are probably more susceptible to fire at the end of the dry season when the plant
reserves are depleted due to the new spring growth. However, the results of Trollope, Potgieter &
Zambatis (1990) showed that the mortality of bush in the Kruger National Park was only 1.3
percent after fires that had been applied to bush ranging from dormant to actively growing plants.
Therefore it would appear that bush is not sensitive to season of burn.
1.5.4 Frequency of burning
The effect of frequency of burning on vegetation is influenced by event-dependent effects and
interval-dependent effects (Bond & van Wilgen, 1996). The event dependent effects occur at the
time of the fire and are influenced by the type and intensity of the burn and the physiological state
of the vegetation at the time of the fire. The interval dependent effects are influenced by the
treatment and growing conditions that occur during the interval between the burns. These two
overall effects tend to confound the interpretation of the effect of frequency of burning and must
be borne in mind when reporting on the effect of frequency of burning.
Frequency of burning has a marked effect on the botanical composition of the grass sward
with species like Themeda triandra being favoured by frequent burning and Tristachya leucothrix
being favoured by infrequent burning in the moist grasslands of Kwazulu/Natal Province in South
Africa (Scott, 1971; Dillon, 1980). Similar results have been obtained in the arid savannas of the
Eastern Cape in South Africa where it was found that frequent burning favours an increase in
Themeda triandra and a decrease in Cymbopogon plurinodis (Robinson, Gibbs-Russel, Trollope &
Downing, 1979; Forbes & Trollope, 1991). In East Africa Pratt & Gwynne (1977) reported that
Themeda triandra is a common constituent of grasslands in the Central Highlands of Kenya on
undulating plateaux and mountain flanks where fires are regular occurrences and the grazing
pressure is not too high. Where fires are infrequent or lacking the upland grassland tends to
become dominated by Pennisetum schimperi and Eleusine jaegeri which are coarse tufted species
of limited grazing value. These are interval dependent effects of frequency of burning because T.
triandra is sensitive to low light conditions that develop when the grass sward is not defoliated and
this species rapidly becomes moribund during extended intervals between fires. Conversely species
like T. leucothrix and C. plurinodis are not as sensitive to low light conditions and survive extended
periods of non-defoliation.
Conflicting results have been obtained on the effect of frequency of burning on bush.
Kennan (1971) in Zimbabwe and van Wyk (1971) in the Kruger National Park in South Africa, both
found that there were no biologically meaningful changes in bush density in response to different
burning frequencies. In the False Thornveld of the Eastern Cape Province in South Africa Trollope
D3.5-1-40-1000
Page 17 of 78
(1983) found that after ten years of annual burning the density of bush increased by 41 per cent,
the majority of which were in the form of short coppicing plants. Conversely Sweet (1982) in
Botswana and Boultwood & Rodel (1981) in Zimbabwe found that annual burning resulted in a
significantly greater reduction in the density of bush than less frequent burning. It is difficult to
draw any general conclusions from these contradictory results except to note that in all cases
significant numbers of trees and shrubs bushes were present even in the areas burnt annually,
irrespective of whether they had decreased or increased after burning. These very variable results
would also suggest that the effect of frequency of burning on woody vegetation is more an eventdependent effect where factors like the type and intensity of fire have had highly significant
individual effects overshadowing the effect of frequency of burning per se.
On the contrary the withdrawal of fire for extended periods of time appears to have a more
predictable effect. For example on the Accra Plains in south eastern Ghana protection of moist
savanna from fire for 29 years resulted in the development of a forest type vegetation with a fairly
closed canopy. The fire sensitive tree species Ceiba pentandra became dominant (Carson & Abbiw,
1990). Similar results have been obtained in the Lamto Reserve in the Ivory Coast which receives a
high mean annual rainfall of 1 300 mm and forms part of the Guinea savanna immediately adjacent
to the deciduous rain forest. This savanna vegetation is subjected to annual burning during the
middle of the dry season. In a study investigating the exclusion of fire for 13 years it was found
that after eight years the open savanna rapidly changed into a dense closed formation and after 13
years the first signs of forest developing occurred in the form of seedlings and saplings. This led to
the conclusion that in all the burnt savannas of Lamto the pressure of forest elements on savanna
vegetation is very high and the exclusion of fire initiates the development of forest (Menaut, 1977).
Similar trends have been found in the more arid savannas (500-700 mm p.a.) in southern Africa
where in the Kruger National Park the exclusion of fire caused both an increase in the density and
size of tree and shrub species (van Wyk, 1971).
The effect of frequency of burning on forage production has not been intensively studied in
South Africa and only limited quantitative data are available. The general conclusion is that the
immediate effect of burning on the grass sward is to significantly reduce the yield of grass during
the first growing season after burning but the depressive effect disappears during the second
season (Tainton & Mentis, 1984; Trollope, 1984).
The effect of frequency of burning on the quality of forage is that generally frequent fires
improve and maintain the nutritional quality of grassland particularly in high rainfall areas making it
highly attractive to grazing animals. This phenomenon has been recorded throughout the savanna
and grassland areas of Africa (West, 1965; Tainton et al, 1977; Moe, et al., 1990; Munthali &
Banda, 1992; Schackleton, 1990). West (1965) stated that the fresh green shoots of new growth
on burnt grassland are very high in protein and quotes Plowes (1957) who found that the average
crude protein content of 20 grasses after burning at the Matopos Research Station in Zimbabwe
was 19%. This is approximately twice the protein content of mature grasses that have not been
burnt at the end of the dry season.
There is apparently no information available on the effect of frequency of burning on the
production and quality of browse by bush in the savanna areas.
D3.5-1-40-1000
Page 18 of 78
OBJECTIVE 2: Quantify the fire/savanna relationships between the various Central and
Southern (climatic) regions.
After serious consideration it was decided that this objective would be more objectively and
adequately achieved through the review of literature and presentation of results that will be dealt
with in the following section in OBJECTIVE 3 which deals with the relationship between fire and
grazing in different vegetation types in central and southern Africa.
OBJECTIVE 3: Determine the relationship between fire and grazing in different
vegetation types (savanna in Central and Southern Africa, degraded and
man-treated in Chaco) with regard to biomass, age, wildlife and
domestic herbivory, climate and season of burn.
This objective is regarded as the primary objective and central task in the Fire Paradox Project
WP3.5: Relationship between fire and grazing. Consequently the report has been structured and
appropriate information gathered that is specifically related to different aspects of the interaction
of fire and grazing in the savanna and grassland areas of central and southern Africa. The
relationship between fire and grazing in degraded and man-treated Chaco has not been dealt with
because this is being attended to by the fire ecologists delegated to this task in the Fire Paradox
Project in South America. The following aspects have been dealt with in achieving OBJECTIVE 3:

Fire/ Herbivory Interaction related to Domestic Livestock Grazers;

Fire/ Herbivory Interaction related to Wild Ungulate Grazers.
3.1
Introduction
The unfortunate reality about investigating the interaction of fire and ungulate herbivory in African
grasslands and savannas is that very limited controlled research has been conducted and what has
been undertaken has largely been undertaken in South Africa. The basic reason for this is that it is
logistically complex, time consuming and expensive both in terms of research personnel and
financial resources. Nevertheless it is essential information that is necessary for the formulation of
viable management practices pertinent to the ecological and economic sustainability of savanna
and grassland ecosystems that are used for both domestic livestock production and wildlife
management. The significance of this interaction arises from the necessity to burn vegetation as a
range management practice in both systems of land use. The current view amongst range
scientists and progressive land users on the permissible reasons for burning rangeland is that fire
can be used to:

remove moribund and/or unacceptable grass material (Hardy, Barnes, Moore & Kirkman,
1999; Trollope, 1989)

control and/or prevent the encroachment of undesirable plants (Trollope, 1989).
These are the primary reasons for burning grassland and savanna vegetation in Africa and
are both applicable to areas used for domestic livestock husbandry and wildlife management. In
D3.5-1-40-1000
Page 19 of 78
this context the fire herbivory interaction relates in particular to the effects of the timing, intensity
and frequency of defoliation of the vegetation by grazing and browsing animals after the fire.
The general effects of post-fire grazing or browsing is that the sooner, more intensely and
more frequently the regrowth of the recovering burnt plants is grazed or browsed after a fire, the
greater is the decrease in their rate of growth and vigour. This can lead to their eventual mortality
if this form of defoliation is maintained for extended periods of time resulting in a significant
decrease in plant density and cover which in turn can cause in grasslands, increased runoff and
accelerated soil erosion. In the case of grass dominated communities a very useful conceptual
model and procedure has been developed to describe and categorise different grass species
according to their reaction to the grazing gradient ranging from low to high intensities and
frequencies of grazing (Foran et al, 1978; Trollope, 1990). For this purpose grass species have
been classified into the following categories:
DECREASER SPECIES – Palatable and productive grass & herbaceous forage species which
dominate when grassland is subjected to moderate intensities and
frequencies of grazing viz. when rangeland is neither under nor over
grazed i.e. correctly grazed;
INCREASER I SPECIES – Less palatable and productive grass & herbaceous forage species which
dominate when grassland is subjected to low intensities and
frequencies of grazing i.e. when rangeland is under or selectively
grazed;
INCREASER II SPECIES – Less palatable and productive grass & herbaceous forage species which
dominate when grassland is subjected to high intensities and
frequencies of grazing i.e. when rangeland is over grazed.
In general, Decreaser grass species are both palatable and productive forage species while
Increaser I & II grass species are less palatable and productive forage species. The value of this
conceptual model is, that in the case of the fire herbivory interaction, the dominance of different
grass categories provides an indication of the levels of intensity and frequency of grazing the grass
sward has been subjected to after repeated burning treatments. It must also be borne in mind
though, that when assessing the effects of the interaction of fire and herbivory in African
grasslands, particularly humid grasslands, these plant communities are themselves products of a
major interaction of fire and herbivory that occurred in the past involving the conversion of wooded
vegetation into grassland through the effects of fire alone or in the presence of animals. This is
well illustrated by examples of the effects of excluding fire and herbivory in the extensive humid
fire climax grasslands located in the higher lying moist eastern regions of South Africa (Tainton,
1999). Classical examples are of plots protected from fire and grazing established by Professor J.D.
Scott in the 1930’s in the Highland Sourveld (Acocks, 1988) at Estcourt in KwaZulu-Natal and in
1950 on the Ukulinga Research Station of the University of KwaZulu-Natal, South Africa (Fynn,
Morris & Edwards, 2004) – see Figure 7.
D3.5-1-40-1000
Page 20 of 78
Figure 7: Two examples of the effects of excluding fire and grazing from grassland for extended
periods of time leading to the development and dominance of tree and shrub vegetation. On the
left is an exclosure plot established in the Highland Sourveld (Acocks, 1988) near Estcourt,
KwaZulu-Natal, South Africa in the 1930’s and photographed in 1971. On the right an exclosure
plot established in 1950 on the Ukulinga Research Station of the University of KwaZulu-Natal,
South Africa and photographed in 2004.
3.2
Fire Herbivory Interaction Related to Domestic Livestock Grazers
3.2.1 Humid Grasslands
3.2.1.1
Effects of Cattle and Sheep Grazing After Burning
An intensive study of the individual effects of cattle and sheep on the proportional species
composition of humid grasslands (sourveld) after burning was investigated between 1992 and 1997
at the Nooitgedacht Agricultural Development Centre near Ermelo (26032’E: 29058’S) in the North
Eastern Sandy Highveld (Acocks, 1988) of South Africa. This is located in the eastern sector of the
central plateau at an altitude of 1694 metres above sea level in an area with sandy soils, a mean
annual rainfall of 719mm falling mainly in summer (November – April) and dry winters
accompanied by severe frosts.
3.2.1.1.1 Treatments
The grass sward was burnt biennially at the end of winter (August/September) with a cool fire to
remove residual, dead material commencing in 1992. The following grazing treatments with cattle
and sheep were replicated three times and applied during the growing season from October to
April over a five year period (1992/93 to 1995/96) (Kirkman, 2002) – see Table 2.
Table 2: The range of burning, grazing and resting treatments with sheep and cattle applied to
investigate the fire grazing interaction involving domestic livestock in the North Eastern Sandy
Highveld humid grasslands2 of South Africa (Kirkman, 2002a).
TREATMEN
T
D3.5-1-40-1000
SEASON
SEASON
SEASON
SEASON
SEASON
1992/93
1993/94
1994/95
1995/96
1996/97
Page 21 of 78
Sheep 1
Burnt
August –
Rest
season
Graze Sheep
Sheep 2
Sheep 3
Burnt
August –
Cattle 3
D3.5-1-40-1000
full Burnt
September –
Conducted
grass surveys to
assess
treatment
effects
Graze sheep Burnt
early season September –
until Dec.–
Graze Sheep rest
late Graze Sheep
season until
April.
Graze sheep
early season
until
Dec.–
rest
late
season until
April.
Burnt
August –
Graze sheep Burnt
full season
September –
Burnt
August –
Graze Cattle
Cattle 2
Rest
season
Graze Sheep
Graze sheep Burnt
full season
September –
Graze Sheep
Cattle 1
full Burnt
September –
Burnt
August –
Graze Sheep
Rest
season
full Burnt
September –
Burnt
September –
Conducted
grass surveys to
assess
treatment
effects
Conducted
grass surveys to
assess
treatment
effects
Rest
season
full Burnt
September –
Graze Cattle
Conducted
grass surveys to
assess
treatment
effects
Graze Cattle
Graze cattle Burnt
early season September –
until Dec.–
rest
late Graze Cattle
season until
April.
Graze cattle
early season
until
Dec.–
rest
late
season until
April.
Burnt
Graze cattle Burnt
Graze
Burnt
September –
Conducted
grass surveys to
assess
treatment
effects
cattle Burnt
Page 22 of 78
August –
full season
Graze Cattle
September –
full season
Graze Cattle
September –
Conducted
grass surveys to
assess
treatment
effects
3.2.1.1.2 Results
The overall effects of the interaction of burning and grazing with sheep and cattle combined with
different seasons and periods of rotational resting on the relative proportions of Decreaser and
Increaser grass species are presented in Figure 8.
60
53
48
PERCENTAGE
50
43
37
40
42
38
34
28
30
19
20
19
18
DECREASER
INCREASER I
INCREASER II
20
10
0
SHEEP Season 1
SHEEP Season 5
CATTLE Season 1
CATTLE Season 5
SHEEP & CATTLE GRAZING AND RESTING
Figure 8: The overall effects of the interaction of burning and grazing with sheep and cattle
combined with different seasons and periods of rotational resting on the relative proportions of
decreaser and increaser grass species (Kirkman, 2002).
The results in Figure 8 indicate that the overall effects of the interaction of burning and grazing
with sheep combined with different seasons and periods of rotational resting over the five year
period of the trial resulted in:

a significant reduction in the proportion of Decreaser grass species (28 to 20%) with Themeda
triandra, a highly palatable and productive species in this category, showing the greatest
proportional decrease;

a marked increase in the proportion of Increaser I grass species (19 to 37%) with Aristida
recta, a highly unpalatable forage species in this category, increasing from 5.4% to 23.4%;
D3.5-1-40-1000
Page 23 of 78

a maintenance of the dominance of Increaser II grass species (53 vs 43%) with Eragrostis
curvula, a significantly less palatable and productive species in this category, increasing from
8.4% to 17.3%;

the overall decrease in the proportion of Decreaser species and the co-dominance of Increaser
I and II grass species is an apparent anomaly suggesting that the grass sward is being both
under and over grazed simultaneously. It is however, a classical example of the grass sward
being selectively grazed by sheep which are highly selective grazers and which overgrazed the
highly palatable Decreaser species (Themeda triandra) and grazed the less palatable Increaser
I and II grass species less intensively less resulting in their co-dominance (Kirkman, 2002);
In contrast the results in Figure 8 indicate that the overall effects of the interaction of burning and
grazing with cattle combined with different seasons and periods of rotational resting during the trial
resulted in:

a significant increase in the proportion of Decreaser grass species (34 to 42%) with Themeda
triandra, the highly palatable and productive species in this category, showing the greatest
increase (27 to 34%);

no significant change in the proportion of Increaser I grass species (18 vs19%);

a significant decrease in the proportion of Increaser II grass species, particularly in forbs and
Heteropogon contortus (48 to 38%);

this change in the proportional species composition of the grass sward to a dominance of
Decreaser grass species and a non-co-dominance of Increaser I and II grass species
indicates that the grass sward was not being selectively grazed by the cattle, a well known
grazing characteristic of this type of grazing ungulate (Kirkman, 2002).
The effects of the different seasons and periods of rotational resting on the proportions of
Decreaser and Increaser grass species was another important result of this trial investigating the
individual effects of sheep and cattle after burning. The effects of the rotational resting treatments
involving sheep are presented in Figure 9.
57
60
51
51
PERCENTAGE
50
40
30
44
37
36
37
31
27
26
21
20
43
43
16
19
23
19
19
DECREASER
INCREASER I
INCREASER II
10
0
BURN/GRAZE
REST FULL
SEASON
BURN/GRAZE
REST LATE
SEASON
BURN/GRAZE
GRAZE FULL
SEASON
BURNING, GRAZING AND RESTING TREATMENTS
D3.5-1-40-1000
Page 24 of 78
Figure 9: The effects of the different seasons and periods of rotational resting on the proportions
of Decreaser and Increaser grass species when grazed with sheep after burning (Kirkman, 2002).
The results in Figure 9 indicate that the overall effects of the different seasons and periods of
rotational resting on the proportions of Decreaser and Increaser grass species when grazed with
sheep were:

that in all cases, the rotational resting treatments all had similar effects on the proportions of
Decreaser and Increaser grass species viz. there was a significant decrease in the proportion
of Decreaser species, a significant increase in the proportion of Increaser I species and a
significant decrease in the proportion of Increaser II species.
The effects of the rotational resting treatments involving cattle are presented in Figure 10.
60
54
48
PERCENTAGE
50
40
40
40
49
41
38
39
38
35
35
30
30
25
20
20
16
14
23
15
DECREASER
INCREASER I
INCREASER II
10
0
BURN/GRAZE
REST FULL
SEASON
BURN/GRAZE
REST LATE
SEASON
BURN/GRAZE
GRAZE FULL
SEASON
BURNING, GRAZING AND RESTING TREATMENTS
Figure 10: The effects of the different seasons and periods of rotational resting on the proportions
of Decreaser and Increaser grass species when grazed with cattle after burning (Kirkman, 2002).
The results in Figure 10 indicate that the overall effects of the different seasons and periods of
rotational resting on the proportions of Decreaser and Increaser grass species when grazed with
cattle were:

that in all cases the rotational resting treatments resulted in an increase in the proportions of
Decreaser grass species but to a greater extent when there was either a full season or late
season rest applied during the second season after the burn;

that the in all cases the rotational resting treatments had no significant effects on the
proportions of Increaser I species;

That in all cases the rotational resting treatments resulted in a significant decrease in the
proportion of Increaser II grass species.
When comparing the effects of the different rotational resting treatments on the proportions of
Decreaser and Increaser grass species when grazed with cattle or sheep, the results show that in
the case of Decreaser and Increaser II grass species the type of animal rather than the resting
D3.5-1-40-1000
Page 25 of 78
treatment had the greatest effect on the botanical composition of the grass sward. These results
clearly demonstrate that generally sheep have a greater detrimental effect on the species
composition of the grass sward than cattle resulting in a less productive grass sward in terms of
providing forage for grazing animals (Kirkman, 2002).
Finally another component of this investigation into the individual effects of cattle and sheep
grazing after burning were on the production and condition of the grass sward in terms of plant
vigour. This was determined by measuring the regrowth of grass species during the season
following the application of the grazing and resting treatments and compared with ungrazed
controls. The results showed that:

the vigour of the grass sward grazed by sheep declined rapidly compared to that grazed by
cattle;

the vigour of palatable Decreaser grass species like Themeda triandra was significantly
reduced whereas the vigour of unpalatable Increaser I species like Aristida recta increased
very markedly with sheep grazing;

Resting improved the vigour of the grass sward in both the sheep and cattle treatments but
the results showed that the productivity of the sward remained at a lower level when grazed
by sheep compared to when grazed by cattle (Kirkman, 2002) indicating that an extended
rest period may be required for maintaining the vigour of the grass sward.
The question of the necessity for resting to maintain the vigour of the grass sward had been
addressed earlier by Haschke & Kirkman (1994) while assessing the importance and significance of
long term rests associated with humid grasslands that had been burnt in autumn, a practice that
had been found to result in a decrease in the abundance of Decreaser species like Themeda
triandra (Bond & van Wilgen, 1996).
It was found that long term rests were necessary to compensate for the reduction in vigour
of preferred grass species and in this light the practice of autumn or late summer burning of
grassland which had always been considered detrimental to the condition of the grass sward in
terms of species composition, was re-evaluated. The species composition of grassland burnt in
autumn followed by an extended rest versus late winter/spring burning followed by short rests in
the form of periods of absence associated with rotational grazing were assessed and compared.
The results showed that the grassland burnt in autumn followed by an extended rest had a
significantly superior species composition than the grassland burnt in late winter/spring followed by
short periods of absence associated with rotation grazing. It was concluded that the long term
rests associated with the autumn practice enabled the preferred grass forage species to maintain
their vigour and compete effectively with the less preferred grass forage species.
D3.5-1-40-1000
Page 26 of 78
3.2.1.1.3
Conclusions
The effects of the interaction of burning and grazing with sheep and cattle combined with different
seasons and periods of rotational resting lead to the following overall conclusions of grazing with
sheep compared to cattle after burning in late winter/spring:

Sheep resulted in a significant decrease in the proportion of productive and palatable
Decreaser grass species like Themeda triandra and a significant increase in both Increaser I
and II grass species as a result of selective grazing, which are less productive and palatable
to domestic grazing livestock;

The type of animal rather than the resting treatment had the greatest effect on the botanical
composition of the grass sward clearly demonstrating that generally sheep have a greater
detrimental effect on the species composition of the grass sward than cattle resulting in a
less productive grass sward in terms of providing forage for grazing animals;

the vigour of the grass sward grazed by sheep declined rapidly compared to that grazed by
cattle;

the vigour of palatable Decreaser grass species like Themeda triandra was significantly
reduced whereas the vigour of unpalatable Increaser I species like Aristida recta increased
very markedly with sheep grazing;

Resting improved the vigour of the grass sward in both the sheep and cattle treatments but
the results showed that the productivity of the sward remained at a lower level when grazed
by sheep compared to when grazed by cattle;

Grazing after burning, followed by regular long term rests, result in a significantly superior
grass species composition than burning and grazing followed by short periods of absence
associated with rotational grazing of livestock.
3.2.1.2
Post-fire Grazing Effects After Burning
A great deal of scientific debate exists on the recommended grazing management to apply after a
prescribed burn because stocking too early after a burn has been regarded as a major reason for
degraded rangeland. The general recommendation for moist grasslands in South Africa has been
that grazing after burning at the end of the dormant winter season should only commence when
the grass sward has regrown to a height of 10 – 15 cm in order to maintain the vigour of the grass
plants (Hardy et al, 1999; Trollope, 1989). These guidelines were hotly debated by livestock
farmers who believed that delaying grazing until the grass sward had regrown to that extent
resulted in a significant decrease in the nutritional value of the grazing and defeated the objective
of burning to remove moribund and poor quality forage for their livestock. Arising from this
situation Zacharias (1994) initiated an intensive investigation into the fire/grazing interaction in the
Dohne Sourveld (Acocks, 1988) in the Eastern Cape Province of South Africa. The research was
conducted at the Dohne Research Station (27030’E 32035’S) at an altitude of approximately 900 m
and located in moist grassland receiving an annual rainfall of 756 mm with light frost occurring in
winter. The objective of the study was to determine the effect of various grazing management
practices after burning on animal performance and vegetation response.
D3.5-1-40-1000
Page 27 of 78
3.2.1.2.1 Treatments
The treatments in the trial comprised:

Annual burning immediately after the first rain in spring (September) in excess of 12.5mm;

Biennial burning immediately after the first rain in spring (September) in excess of 12.5mm;

Early Graze applied as soon after the burn when there was sufficient forage to graze;

Late Graze applied as soon as the grass sward had recovered after the burn to a height of
100 – 150mm;

Continuous Grazing with mature Dohne Merino wethers for the duration of the grazing period
in each season;

Rotational Grazing with mature Dohne Merino wethers initially for a grazing period of one
week rotated around four camps and extended to two weeks as the season progressed.
A uniform stocking rate of 1.5 ha per animal unit applied in all the grazing treatments.
Animal performance was recorded using animal mass to determine seasonal weight gains, grazing
days per treatment and energy consumption.
The response of the vegetation involved recording annually the species composition of the
grass sward using 250 nearest plant points for each treatment. The basal cover of the sward was
monitored annually by recording living rooted plant cover along permanent line transects
(Zacharias, 1994).
3.2.1.2.2 Results
The effects of the treatments on animal performance in terms of livemass gain are presented in
Figure 11:
D3.5-1-40-1000
Page 28 of 78
17
15
LIVEMASS GAIN - kg
15
13
13
12
13
11
11
10
10
9
8
9
7
7
6
EARLY GRAZING
LATE GRAZING
6
5
4
5
3
1
-1
1987
1988
1989
1990
1991
1992
MEAN
YEAR
Figure 11: Effects of early and late grazing after annual burning on the livemass gains of sheep
recorded at the end of five growing seasons expressed in kilograms per small stock unit (Zacharias,
1994).
The results in Figure 11 clearly illustrate the significantly higher livemass gains obtained
from the sheep with early grazing after the annual spring burns resulting in a mean overall higher
livemass gain of 57.1% over the six growing seasons. A similar trend but lower livemass gains
were obtained from the early grazing after the biennial spring burns. These results clearly illustrate
why livestock farmers are attracted to the practice of wanting to graze their livestock as soon as
possible after burning at the end of the dormant winter period (Zacharias, 1994).
The effects of early grazing compared to late grazing after annual and biennial burning with
continuous versus rotational grazing all resulted in an increase in the proportion of Themeda
triandra (Decreaser) and a decrease in Tristachya leucothrix (Increaser I) constituting an
improvement in the condition of the grass sward in terms of its botanical composition (Zacharias,
1994).
Regarding the basal cover of the grass sward there were no significant differences recorded
between the start (17.3%) and the end of the experiment (14.38%) in all the treatments, with
both values indicating a dense grass cover (Zacharias, 1994). Nevertheless the results did show
that annual burning and early grazing did increase the level of runoff and soil loss. However, the
levels of soil loss were less than 2 tons per hectare per annum which would suggest that they may
match soil genesis. Nevertheless the fact that frequent burning and heavy grazing did result in
increased runoff and soil loss suggests that this result must not be disregarded (Zacharias, 1994)
and managers must be sensitive to accelerated levels of soil erosion associated with this type of
burning and grazing management on their properties.
Regarding the vigour of the grass sward the results showed that there was an overall
decline in the vigour of the grass sward during the period of the trial irrespective of early versus
late grazing, annual versus biennial burning and continuous versus rotation grazing (Zacharias,
1994). This decline in vigour possibly provides the reason for the decrease in animal performance
recorded during the latter portion of the trial (see Figure 11).
D3.5-1-40-1000
Page 29 of 78
3.2.1.2.3 Conclusions
The following overall conclusions can be drawn from the results of the trial:

Early grazing with sheep immediately after burning resulted in significantly higher livemass
gains than late grazing with sheep applied as soon as the grass sward had recovered after
the burn to a height of 100 – 150mm;

Early grazing compared to late grazing after annual and biennial burning with continuous
versus rotational grazing all resulted in an increase in Decreaser grass species and a decrease
in Increaser I grass species, constituting an improvement in the condition of the grass sward
in terms of its botanical composition;

Both early grazing and late grazing after burning resulted in no significant differences in the
basal cover of the grass sward between the start and the end of the experiment all with
values indicating a dense grass cover. Nevertheless the results did show that annual burning
and early grazing did increase the level of runoff and soil loss. However, the levels of soil loss
were less than 2 tons per hectare per annum suggesting that this level of soil loss may be
equal to soil genesis. Nevertheless managers must be sensitive to accelerated levels of soil
erosion associated with this type of burning and grazing management in practice;

Both early grazing and late grazing after burning resulted in an overall decline in the vigour
of the grass sward and possibly a decline in overall animal performance during the period of
the trial indicating the necessity of including a regular long rest where burning and early
grazing after the burn are used to manage moist grasslands (sourveld) in Africa.
3.2.2 Arid Grasslands
In terms of the aforementioned reasons for prescribed burning (Hardy, Barnes, Moore & Kirkman,
1999; Trollope, 1989) arid grasslands (<500 mm p.a.) in Africa seldom require burning to remove
moribund unpalatable herbage material because of low rainfall and their susceptibility to periodic
droughts and, until recently in South Africa, this practice was either discouraged or prohibited by
law. Consequently very limited research attention has been given to the effects of the interaction of
fire and grazing in arid savannas.
A burning trial has recently been initiated in the southern Free State Province in South
Africa to develop a management strategy to improve the condition of the grass sward in the
Cymbopogon Themeda Veld (Acocks, 1988) range type. This rangeland has become dominated by
Cymbopogon plurinodis, Elionurus muticus, Aristida spp and Eragrostis spp as a result of selective
grazing by sheep since settled livestock farming commenced in the region in the late 1800’s. This
has resulted in a significant reduction in the grazing capacity of the grass sward causing a marked
reduction in the economic viability of livestock farming in the affected areas. Research conducted in
the False Thornveld (Acocks, 1988) of the Eastern Cape Province in the Alice and Bedford areas
indicates that veld in the aforementioned condition can be significantly improved by applying
controlled burning, appropriate forms of rotational grazing and strategic rotational resting
(Trollope,1999).
It was therefore postulated that a combination of controlled burning, high utilization
grazing, high production grazing and rotational resting will result in a reduction in the dominance of
Cymbopogon plurinodis, Elionurus muticus, Aristida spp and Eragrostis spp and an increase in
D3.5-1-40-1000
Page 30 of 78
productive and palatable species like Themeda triandra. These range management practices are
being tested in the camp “Boonste Braak” on the farm “Weenkop” owned by Mrs S. Templeton
located in the Rouxville district of the south eastern region of the Free State Province ( 30025’S
26050’E) in an area with a mean annual rainfall of 534 mm falling mainly in summer. While the area
is not extremely arid and truly representative of arid grasslands, it is prone to periodic droughts
and provides the only example available of the effects of grazing after burning in non-humid
grasslands. In this case an intensive study is being conducted where prescribed burning is being
applied under specific conditions together with a particular regime of grazing and resting
treatments that will continue over a period of several years. The performance of the livestock
grazing the burnt camp is also being recorded in terms of the number of grazing days obtained
during each period of occupation in each camp and the mass of the livestock is being measured
after each grazing period. The trial was initiated during September, 2006.
3.2.2.1 Treatments & Measurements
A field survey of the “Boonste Braak” camp showed that the grassland comprised two distinct
grassland communities in approximately equal proportions. The first area was devoid of stones and
dominated by the grass species Cymbopogon plurinodis and the second area was spread over a
rocky ridge dominated by Elionurus muticus. However, in both grass communities there is still a
significant amount of Themeda triandra present in the camp to provide an adequate potential for
testing the hypothesis that burning and appropriate grazing treatments can be applied to improve
the condition of the grass sward. As a consequence of the existence of these two grassland
communities two separate monitoring sites in the camp were developed to determine the effect of
burning, grazing and resting on the two grass species C. plurinodis and E. muticus (see Figure 13).
Figure 12: Location of the Cymbopogon and Elionurus monitoring sites in the camp “Boonste
Braak” on the farm “Weenkop” in the Rouxville district of the Free State Province in South Africa..
The following treatments are being applied in the two monitoring sites in the camp “Boonste
Braak” on the farm “Weenkop”.
D3.5-1-40-1000
Page 31 of 78
3.2.2.1.1 Burning Treatment
The camp “Boonste Braak” was burnt after abundant spring rains on the 27th and 28th September,
2006. The controlled burning was conducted by the “Working on Fire” program professionally
trained fire crew (see Figure 13).
Figure 14: The “Working on Fire” program applying the controlled burning in “Boonste Braak”
camp ably assisted by a professionally trained fire crew.
The standing crop of grass representing the grass fuel load was estimated with a Disc Pasture
Meter (DPM) in each of the fenced and unfenced plots prior to the application of the initial
controlled burns (see Figure 14).
Figure 14: The Disc Pasture Meter used for estimating the standing crop of grass in the
Cymbopogon plurinodis and Elionurus muticus communities, in the “Boonste Braak” camp on the
D3.5-1-40-1000
Page 32 of 78
farm “Weenkop” in the Rouxville district, recorded prior to the application of the initial burning and
grazing treatments.
The controlled burn in “Boonste Braak” camp was applied according to the following fire regime:
i) Type Of Fire
Except for the construction of the fire break around the perimeter of the camp the major portion of
the camp was burnt with a head fire burning with the wind (see Figure 15).
Figure 15: Applying a low intensity head fire with the wind which resulted in the major portion of
the camp being successfully burnt.
ii) Fire Intensity
The primary objective of the controlled burn was to apply a low intensity fire to remove the
moribund grass material in the camp thereby resulting in the production of highly palatable and
nutritious regrowth of the grass sward. The controlled burn was applied under relatively cool and
moist conditions both in terms of low air temperatures, high relative humidities, a moderately high
fuel moisture content and moderate wind speeds resulting environmental conditions best suited for
burning to remove unpalatable, moribund grass material.
iii) Season Of Burn
In accordance with scientifically accepted recommendations for controlled burn (Trollope, 1989)
the “Boonste Braak” camp was burnt during spring (27th/28th September, 2007) immediately after
substantial rainfall events of >100 mm.
3.2.2.1.2 Post-Burn Grazing Management
The following post-burn grazing management treatments were applied in the “Boonste Braak”
camp during the 2007/ 2008 growing season:

High utilization grazing with cattle was initiated as soon as there was grazeable material
available after the burn and re-grazing of the camp was applied whenever the grass sward had
recovered to a grazeable height of approximately 10 cm i.e. the burnt camp was treated as a
D3.5-1-40-1000
Page 33 of 78
priority grazing camp where the primary objective was to keep the grass sward in as short and
uniformly grazed condition as possible. Animal performance was monitored by recording the
number, class and mass of all the livestock grazing “Boonste Braak” camp for each grazing period
(date in & date out). This was done to enable animal performance to be quantified in terms of
grazing days and changes in animal mass per unit area.

High utilization grazing was applied for 18 months (two growing seasons) after the
application of the controlled burn;

At the commencement of the third growing season high production grazing will be applied
where the grazing animals are removed from the burnt camp as soon as the palatable grass
species like Themeda triandra (red grass) and Digitaria eriantha (finger grass) have been grazed to
a height of approximately 5 cm.
3.2.2.2
Results & Conclusions - Growing Season 2007/2008
The effects of the burning and grazing treatments on the botanical composition, basal cover and
forage and grass fuel production potential of the grass sward were estimated in the Cymbopogon
plurinodis and Elionurus muticus replicates of the experiment in botanical surveys completed on the
19th December, 2007. The overall effects of the burning and grazing treatments in the presence
and absence of grazing during the 2006/2007 growing season in the Cymbopogon and Elionurus
dominated grass communities were as follows:

In both cases the frequency of Cymbopogon plurinodis and Elionurus muticus did not
change in response to the different treatments indicating that these two grass species are highly
resilient and resistant to change even when subjected to extreme defoliation treatments like
burning followed by intensive grazing;

Also in both cases the basal cover of the grass sward was not significantly affected by the
different treatments with the point to tuft distances all remaining at low levels, indicating a high
potential for resisting accelerated soil erosion. Interestingly this result is similar to that recorded by
Zacharias (1994) in the humid grassland in the Eastern Cape Province in South Africa, where high
utilization grazing after the burn did not result in a significant decrease in the basal cover of the
grass sward. Nevertheless, as was cautioned previously, managers must be sensitive to accelerated
levels of soil erosion associated with burning followed by high utilization grazing on their
properties.

Similarly the forage and fuel production potential of the grass sward were not affected and
remained at the same levels as prior to the application of the different burning and grazing
treatments;

The only noteworthy effect was the significant increase in the degree of utilization of the
grass sward by cattle that were stocked immediately after the grass sward had recovered
sufficiently after the burn to enable grazing to occur. This is a very important practical result as it
provides quantitative evidence that increased livestock production can be obtained by grazing
immediately after the application of a burning program, without any detriment to the condition of
the grass sward. Again this result is similar to that recorded by Zacharias (1994) in the humid
grassland in the Eastern Cape Province of South Africa .
D3.5-1-40-1000
Page 34 of 78

Although there has been no decline in the abundance of the two problem grass species
Cymbopogon plurinodis and Elionurus muticus, it must be appreciated that their dominance in the
grass sward is the result of decades of management practices involving continuous and selective
grazing, particular with sheep and the virtual withdrawal of fire from the ecosystem. It is therefore
unrealistic to expect positive results after only one growing season and the future application of a
years rest to promote the seeding and vigour of Themeda triandra (red grass) is expected to have
significant effects in the improvement of the condition of the grass sward. The results also
emphasise the fact that the rehabilitation of rangelands, particularly arid grasslands, is a long term
process that requires patience and commitment and what has taken decades to bring about will
not be rectified in a short period of time..
3.3
Fire Herbivory Interaction Related to Wild Ungulate Grazers
3.3.1 Moist and Arid Savannas
3.3.1.1
Effects of Wild Ungulate Grazing After Burning
3.3.1.1.2 Experimental Burn Plot Trial (EBP’s) – Kruger National Park, South Africa
No formal and controlled scientific investigation could be found on the effects of grazing after
burning by wild ungulates in humid grasslands in Africa. The closest allied scientific trial to this
objective is the Experimental Burn Plot (EBP) trial in the Kruger National Park in South Africa being
conducted in humid and arid savannas in the north eastern region of South Africa contiguous to
the international border with Mozambique in the east and Zimbabwe in the north. However, before
describing and assessing the effects of the treatments on the grass sward associated with this trial,
it is necessary to give an overview of the experiment and place it in the perspective of the period
during which it was initiated. Prior to the commencement of the experiment in 1954 prescribed
burning had become a highly contentious management practice in South Africa. The injudicious use
of fire had been identified as one of the important contributing factors responsible for widespread
soil erosion in the country. This resulted in strict regulations being included in the Soil Conservation
Act No 45 of 1946 governing the use of fire in range management which in turn led to the
development of a negative attitude towards controlled burning as a range management practice in
South Africa. It undoubtedly influenced the attitude of Colonel J.A.B. Sandenbergh, the newly
appointed warden of the Kruger National Park in 1946, judging by the annual report he submitted
to the National Parks Board in 1950. Brynard (1971) states that almost all controlled burning was
abandoned during Colonel Sandenbergh’s first years in office and in 1949 the National Parks Board
passed a resolution:

“that no veld (rangeland) shall be burnt more often than once every five years; that all such
burning shall only be done after the first good spring rains; and that by every means at our
disposal accidental fires must be avoided”.
 Arising from the effects of the de facto fire suppression policy applied in the Park during the
period 1946 to 1954 the following concerns were expressed (van der Schijff, 1957; Brynard,
1971):
 there was a serious deterioration in the nutritional quality of the grazing, which led to serious
undergrazing of less preferred moist savanna (sourveld), particularly in the vicinity of
Pretoriuskop, and the overgrazing of preferred arid savannas (sweetveld);
D3.5-1-40-1000
Page 35 of 78
 a significant increase in the density and volume of woody vegetation resulting in bush
encroachment;
 the development of a serious fire hazard as a result of the excessive accumulation of grass
fuel leading to the occurrence of extensive large scale high intensity fires. Figure 16
illustrates the extensive areas burnt by wild fires in the Kruger National Park during 1953, a
year which was particularly problematic with regard to wild fires.
N
Figure 16: The extent of uncontrolled wild fires in the Kruger National Park during 1953,
estimated to total approximately one quarter of the Park (van der Schijff, 1957; Brynard, 1971):
As a possible means of addressing these aforementioned problems, it was decided that a
comprehensive fire research program be conducted in the major vegetation types of the Kruger
National Park. The stated objective of the fire research program was to determine the effect of
season of burning on range condition in the major vegetation types and while the effect of
frequency of burning on range condition was not specifically stated by van der Schijff (1957), it
was implied.
The research program was initiated in 1954 and comprised laying out replicated treatments in the
four major vegetation types of the Kruger National Park (Gertenbach, 1983; Trollope et al, 1998)
namely, the Sour Bushveld of Pretoriuskop (sandy granitic soils); the Combretum Woodland (sandy
granitic soils) near Skukuza; the Knobthorn/Marula Savanna (clay basaltic soils) in the vicinity of
Satara; and the Mopane Shrub (clay basaltic soils) north of Letaba. The location of the different
replicates of the trial is illustrated in Figure 17.
D3.5-1-40-1000
Page 36 of 78
Figure 17: The location of the four different replicates of the Experimental Burn Plot trial in the
four major vegetation types in the Kruger National Park viz. the Sour Bushveld at Pretoriuskop, the
Combretum Woodland near Skukuza, the Knobthorn/Marula Savanna near Satara and the Mopane
Shrub north of Letaba near Mopane.
The rainfall profile of the areas where the Experimental Burn Plots are located is presented in
Figure 18.
D3.5-1-40-1000
Page 37 of 78
RAINFALL - mm
800
700
600
500
400
300
200
100
0
727
539
521
249
224
218
30%
Pretoriuskop
489
MEAN
SDEV
CV %
173
32%
43%
Skukuza
Satara
51%
Mopane
VEGETATION LANDSCAPES
Figure 18: The mean annual rainfall, standard deviations and coefficients of variation in the
annual precipitation recorded at the weather stations located at Pretoriuskop, Skukuza, Satara and
Mopane. Data recorded for >15 years.
The Sour Bushveld at Pretoriuskop receives the highest rainfall (727mm p.a.) and can be regarded
as moist savanna. The Mopane Shrub at Mopane receives the lowest rainfall (489mm p.a.) and has
an extremely high coefficient of variation of 51% that is characteristic of arid savannas. Satara with
a mean annual rainfall of 539mm p.a. can also be regarded as arid savanna with its high coefficient
of variation. Unfortunately the rainfall data for Skukuza is not representative of the Combretum
Woodland vegetation landscape which is contiguous to the Sour Bushveld vegetation type. It
receives less rainfall than the Sour Bushveld but its rainfall profile is not correctly represented by
the Skukuza data and can be regarded as being intermediate between the rainfall data for
Pretoriuskop and Skukuza.
3.3.1.1.2.1
Treatments
The Experimental Burn Plot trial was laid out as a randomised block design with four replications
comprising different seasons and frequencies of burning. All treatments are replicated four times at
each site giving a total of 208 plots approximately 370m x 180m (+ 7 ha) in size and a total area of
approximately 1456 hectares.
Details of the treatments are presented in Table 3.
D3.5-1-40-1000
Page 38 of 78
Table 3: Season and frequency of burning treatments being applied in the Experimental Burn Plot
trial in the Sour Bushveld at Pretoriuskop and the Combretum Woodland near Skukuza since 1954
and the Knobthorn/Marula Savanna at Satara and the Mopane Shrub near Mopane since 1958.
PRETORIUSK
OP
SKUKUZA
SATARA
MOPANE
Oct B2
Oct B2
Oct B2
Oct B2
Oct B3
Oct B3
Oct B3
Oct B3
Dec B2
Dec B2
Oct B4
Oct B4
Dec B3
Dec B3
Oct B6
Oct B6
Feb B2
Feb B2
Dec B2
Dec B2
Feb B3
Feb B3
Dec B3
Dec B3
Apr B2
Apr B2
Feb B2
Feb B2
Apr B3
Apr B3
Feb B3
Feb B3
Aug B1
Aug B1
Apr B2
Apr B2
Aug B2
Aug B2
Apr B3
Apr B3
Aug B3
Aug B3
Aug B1
Aug B1
K
K
Aug B2
Aug B2
Aug B3
Aug B3
K
K
Where:
B1 = annual burn; B2 = biennial burn; B3 = triennial burn; B4 = quadrennial burn;
B6 - sexennial burn; K = no burn;
Oct. = October (spring); Dec. = December (early summer); Feb. = February (mid-summer);
D3.5-1-40-1000
Page 39 of 78
Apr. = April (autumn); Aug. = August (mid-winter).
3.3.1.1.2.2
Measurements
3.3.1.1.2.2.1 Pre-Treatment Measurements
Prior to the application of the initial burning treatments botanical surveys comprising 500 point
quadrats were conducted along the two diagonals of each plot with a Levy Bridge apparatus giving
a total of 1000 points per plot. The number of strikes of living, rooted plant material was recorded
for the different grass and herbaceous species and the data expressed as the percentage relative
frequency. The data were expressed on an individual species basis for the grasses but generally as
an undifferentiated single group for the non-grasses designated as forbs. The number of strikes
was expressed as the percentage basal cover of the grass sward.
3.3.1.1.2.2.2 Fire Behaviour Measurements
Initially during the application of the burning treatments qualitative descriptions of the condition of
the vegetation and atmospheric conditions were recorded. However, with the development of
research on fire behaviour in South Africa, from 1982 onwards quantitative measurements were
made of the grass fuel load, fuel moisture, air temperature, relative humidity and wind speed. The
rate of spread and flame height of the fires were also recorded during the burning of the plots and
the fire intensity calculated for each plot (Trollope et al, 1998).
3.3.1.1.2.2.3 Post-Treatment Measurements
Arising from a Fire Workshop held in the Kruger National Park in 1997 it was decided to conduct a
complete re-survey of the herbaceous grass and tree and shrub vegetation in each plot in the
different replicates of the EBP trial (Braack, 1997). These were completed in 1998 in Satara, 2000
in Mopane and 2001 in Pretoriuskop and Skukuza for the grass sward and Higgins et al (2007)
completed a peer reviewed paper describing the treatment effects on the woody tree and shrub
vegetation. The re-surveys of the grass sward in each plot comprised a 200 point quadrat survey
arranged along the two diagonals of each plot where the nearest rooted grass or other herbaceous
plant was recorded together with the point to tuft distance expressed in centimetres. This latter
measurement is a surrogate measure for the basal cover of the grass sward and was developed by
Hardy and Tainton (1993) and found by Vetter (2003) to be significantly negatively correlated with
the degree of accelerated soil erosion i.e. the smaller the point to tuft distance the greater is the
resistance of the grass sward to accelerated soil erosion and vice versa. It was decided to deviate
from the original procedure used for surveying the herbaceous grass layer because it is too
laborious and tedious. Furthermore research experience had shown that recording the nearest
rooted plant to a recording point together with the point to tuft distance provided adequate and
ecologically meaningful descriptions of the botanical composition and basal cover of the grass
sward that could be used to interpret treatment effects.
With the completion of the follow-up surveys in the Experimental Burn Plot trial it was
realised that these follow up surveys provide a unique and ecologically valuable set of data
reflecting the long term effects of season and frequency of burning on the vegetation in southern
African savannas. At a workshop held at the Scientific Network Meeting in the Kruger National Park
in April, 2008 on the future of the EBP trial, it was decided that a special effort be made to
specifically address the earlier mentioned objectives of the trial i.e. determine the effects of season
and frequency of burning on the grass sward. On further reflection during November, 2008, it was
D3.5-1-40-1000
Page 40 of 78
also decided to interpret the effects of season and frequency of burning in relation to some
functional characteristics pertinent to the utilization and management of the grass sward for
wildlife management. The functional characteristics of the grass sward that were selected were:

the forage and grass fuel production potential of the grass sward as reflected by the forage
and fuel scores developed by Trollope (1990a) for the Kruger National Park;

the resistance of the grass sward to accelerated soil erosion as represented by the basal
cover and or the point to tuft distance of the grass sward;
It was believed that by broadening the analysis and interpretation of the data reflecting the
treatment effects on the herbaceous vegetation, this would succeed in addressing the original
objectives for initiating the Experimental Burn Plot trial. In addition it would provide essential
information pertinent to the utilization and management of African savannas for grazing wild
ungulate populations.
3.3.1.1.2.3
Results
In presenting the results of the different fire and grazing treatments in the Experimental Burn Plot
trial spanning the period 1954 to 1998/2001 it was decided to highlight the results obtained in the
replicates of the experiment located at the rainfall extremes represented by the moist and arid
savannas i.e. the effects of the fire and grazing treatments in the Sour Bushveld (moist savanna) in
the south of the Kruger National Park at Pretoriuskop and in the Mopane Shrub (arid savanna)
located in the arid north of the Park at Mopane (see Figure 18). This procedure will apply to the
detailed effects of the season and frequency of burning and grazing on the functional
characteristics of the grass sward. This is because a study of the results clearly indicates that
rainfall has a major effect on the response of the vegetation to fire and grazing thereby illustrating
the interacting effects of fire and grazing in moist and arid savannas in Africa. Nevertheless, the
overall effects of fire and grazing on the condition of the vegetation in the different vegetation
landscape will also be included to provide an inclusive overview of the results of the Experiment
Burn Plot trial.
3.3.1.1.2.3.1 Grass Forage and Fuel Potential
Range condition is defined as the condition of the vegetation in relation to some functional
characteristics (Trollope et al, 1990) and its potential to produce grass forage as nutrition for
grazing ungulates and grass fuel to support fires to influence the balance between grass and
woody vegetation are fundamental functional characteristics of rangeland in both grassland and
savanna areas. These functional characteristics of the grass sward can be estimated by allocating
forage and fuel factors on a scale of 0-10 to the different species making up the grass sward, and
represent their genetic potential to produce grass forage and fuel. The resultant grass forage and
fuel potentials, as represented by the Forage and Fuel Scores, are calculated by multiplying the
percentage frequency for each recorded grass species by its respective forage and fuel factor and
then totalling the scores for each survey site. It must be clearly understood though, that these
forage and fuel scores do not reflect the actual and current amounts of available grass forage and
fuel but their potential provided that conditions for plant growth are favourable, particularly
moisture in the form of rainfall. In the case of forage, this potential production also translates into
actual available forage depending on the intensity and frequency of defoliation by grazing animals
i.e. grazing management.
D3.5-1-40-1000
Page 41 of 78
In addition, the forage factors represent the potential of a grass species to produce forage
for consumption by bulk feeders like buffalo, zebra and elephant because another important factor
influencing range condition are animal ratios of bulk to concentrate grazers which influences the
degree of selective grazing. Mentis (1981) recommended that the metabolic mass of concentrate
grazers should not be permitted to exceed that of bulk grazers in any grazing unit. He proposed a
maximum ratio of 1 animal unit bulk grazers : 1 animal unit concentrate grazers. This is based on
research experience with domestic livestock that indicates that the ratio should not be greater than 1
Bovine (bulk grazer) : 6 sheep (concentrate grazer). However, experience in the arid savannas of the
Eastern Cape Province in South Africa suggests that a narrower ratio is preferable in low rainfall areas
where the condition of the grass sward is more sensitive to selective grazing. Consequently it is
recommended that in wildlife situations a maximum ratio of 1 animal unit bulk grazers : 1 animal unit
concentrate grazers be applied in humid savannas, and a maximum ratio of 1 animal unit bulk grazers
: ½ animal unit concentrate grazers in arid savannas (Trollope, 1990b).
In conclusion these functional characteristics describing the potential of the grass sward to
perform functions pertinent to the system of land use are very useful in quantifying the current
condition of the vegetation, and in this case the condition of the grass sward. Forage and fuel
factors have been allocated to all the different grass species in the Kruger National Park and these
were estimated in consultation with experienced plant scientists and wild life managers in the Park
together with pasture scientists and livestock farmers familiar with the ecology of the savanna
communities in this region. This procedure of assessing the grass forage and fuel potential of the
grass sward has been successfully used across a great diversity of types of savannas in southern
and east Africa e.g. in the Kruger National Park of South Africa by Trollope & Potgieter (1986); in
the central highlands of Kenya (Trollope & Trollope, 1999); in the Caprivi region of Namibia
(Trollope et al, 2000); in the Ngorongoro Crater in Tanzania (Trollope & Trollope, 2001a); the Gile
National Reserve in Zambesia Province in Mozambique (Trollope & Trollope, 2001b) and the
Okavango Delta in Botswana (Trollope et al, 2006).
In using forage and fuel scores to assess the condition of the grass sward in terms of the
effects of grazing by wildlife during different seasons and frequencies of burning, the following
guidelines have been developed for use in the Kruger National Park (Trollope et al, 1989) and will
be used in the interpretation of the treatment effects on the grass sward in the Experimental Burn
Plot trial – see Table 4.
Table 4: Guidelines for interpreting the grass forage and fuel potentials of the grass sward as
represented by forage and fuel scores recorded in the Kruger National Park in South Africa
(Trollope et al, 1989).
FORAGE/ FUEL SCORES
GRASS
FORAGE
POTENTIALS
<200
Very low
200 - 300
Low
301 – 400
Medium
D3.5-1-40-1000
&
FUEL
Page 42 of 78
401 – 500
High
>500
Very high
3.3.1.1.2.3.2 Grass Forage Potential
The overall effects of fire and grazing on the grass forage potential of the grass sward in the Sour
Bushveld, Combretum Woodland, Knobthorn/Marula Savanna and Mopane Shrub replicates of the
Experimental Burn Plot trial in the Kruger National Park are presented in Figure 19.
753
800
652
FORAGE SCORE
700
600
500
400
532
508
480
421
352
379
416
410
348
279
FIRE + GRAZING
GRAZING
300
200
100
0
SOUR
BUSHVELD
1954
SOUR
BUSHVELD
2001
Combretum
WOODLAND
1954
Combretum KNOBTHORN/ KNOBTHORN/
MOPANE
WOODLAND
MARULA
MARULA
SHRUB 1954
2001
SAVANNA 1954 SAVANNA 1998
MOPANE
SHRUB 2000
VEGETATION LANDSCAPES
Figure 19: The overall effects of fire and grazing on the grass forage potential in the Sour
Bushveld at Pretoriuskop, the Combretum Woodland at Skukuza, the Knobthorn/Marula Savanna
near Satara and the Mopane Shrub at Mopane in the Experimental Burn Plot trial in the Kruger
National Park during the period 1954 to 2001.
Comparing the overall grass forage potential of the grass sward in 1954 versus that recorded
in the follow-up surveys conducted between 1998 and 2001 in the different vegetation landscapes,
fire and grazing has resulted in an increase in the forage potential in the more humid Sour Bushveld
and Combretum Woodland vegetation types at Pretoriuskop and Skukuza. In contrast, in the more
arid vegetation landscapes fire and grazing has resulted in an overall decrease in the forage potential
in the Knobthorn/Marula Savanna and the Mopane Shrub at Satara and Mopane, with the latter
vegetation type showing a marked decrease. This is a clear indication that moist savannas are better
adapted to burning than arid savannas in terms of the effect of fire and grazing on the grass forage
potential. In all cases grazing in the absence of fire resulted in a marked increase in the forage
potential of the grass sward except for the Mopane Shrub vegetation type in the extremely arid
savanna at Mopane in the north of the Park. It must be noted though, that the grass sward in the
treatments with high forage potentials (>400) associated with only grazing and no fire, was generally
D3.5-1-40-1000
Page 43 of 78
in a moribund and less palatable condition, characterised by high tick populations clearly indicating
that some form of defoliation is necessary for utilizing areas with a high forage potential.
The interacting effects of season of burning and grazing on the grass forage potential in moist
savanna in the Kruger National Park is presented in Figure 20.
GRASS FORAGE
POTENTIAL
600
500
400
512
477
409
372
338
516
510
358
344
340
300
200
100
0
SPRING
1954
SPRING
2001
EARLY
SUMMER
1954
EARLY
SUMMER
2001
LATE
SUMMER
1954
LATE
SUMMER
2001
AUTUMN
1954
AUTUMN
2001
WINTER
1954
WINTER
2001
SEASON OF BURN
Figure 20: The interacting effects of season of burning and grazing on the grass forage potential
in the moist Sour Bushveld at Pretoriuskop in the Experimental Burn Plot trial in the Kruger National
Park during the period 1954 to 2001.
The results in Figure 20 show that in all cases the season of burn and subsequent grazing
resulted in an increase in the forage potential of the grass sward. This effect was particularly
marked in all the seasons of burn except for the winter burning treatment where the increase was
substantially less. The effect of the winter treatment resulting in a less marked increase in the
forage potential should not be ascribed to the fire having a negative effect on the condition of the
grass sward. It was shown in the review of literature on the effects of fire alone on the recovery of
the grass sward after a fire, winter burning when the grass sward is dormant resulted in a
significantly better recovery in the grass sward. This would suggest that the main reason for the
smaller improvement in the forage potential of the grass sward was the result of consistent,
intense grazing of the grass plants after the fire following the first spring rains at the
commencement of the growing season.
At the end of winter high quality grazing is not abundant and the succulent and highly
palatable regrowth of the burnt grass in a small plot of seven hectares surrounded by extensive
areas of generally unburnt rangeland would be highly prone to heavy grazing. This would have a
negative effect on the vigour of the grass sward and its botanical composition and therefore a
negative effect on the forage potential of the grass sward.
The interacting effects of season of burning and grazing on the grass forage potential in arid
savanna in the Kruger National Park is presented in Figure 21.
D3.5-1-40-1000
Page 44 of 78
GRASS FORAGE POTENTIAL
450
420
395
400
415
406
394
354
350
301
300
239
250
254
242
200
150
100
50
0
SPRING
1954
SPRING
2001
EARLY
EARLY
LATE
LATE
AUTUMN AUTUMN
SUMMER SUMMER SUMMER SUMMER
1954
2001
1954
2001
1954
2001
WINTER
1954
WINTER
2001
SEASON OF BURN
Figure 21: The interacting effects of season of burning and grazing on the grass forage potential
in the arid Mopane Shrub at Mopane in the Experimental Burn Plot trial in the Kruger National Park
during the period 1954 to 2001.
The results in Figure 21 are in complete contrast to the effects of season of burning and
grazing in the moist savanna in the Sour Bushveld at Pretoriuskop. In the arid savanna in the Mopane
Shrub all the treatments resulted in a decrease in the forage potential of the grass sward. This effect
was particularly marked in the fire grazing interaction associated with the late summer, autumn and
winter burns in contrast to the spring and early summer burns. Again it is believed that these effects
are mainly due to the effects of grazing rather than the season of burning. This is because the spring
and early summer burns are normally associated with more reliable and higher rainfall than the latter
part of the growing season and winter when the grass sward is drying off and forage is becoming less
abundant resulting in increased grazing pressure in the burn plots.
This arid savanna in the Mopane Shrub is also characterised by highly variable annual rainfall
with a coefficient of variation of 52%. This undoubtedly further exacerbates the ecological sensitivity
of the grass sward to intensive grazing at the end of the growing season and winter thereby
explaining the marked decrease in the forage potential associated with these treatments.
The interacting effects of frequency of burning and grazing on the grass forage potential in
moist savanna in the Kruger National Park is presented in Figure 22.
D3.5-1-40-1000
Page 45 of 78
600
498
488
FORAGE SCORE
500
400
346
346
346
359
300
200
100
0
ANNUAL 1954
ANNUAL 2001 BIENNIAL 1954 BIENNIAL 2001 TRIENNIAL 1954 TRIENNIAL 2001
FREQUENCY OF BURN
Figure 22 The interacting effects of frequency of burning and grazing on the grass forage
potential in the moist Sour Bushveld at Pretoriuskop in the Experimental Burn Plot trial in the
Kruger National Park during the period 1954 to 2001.
The results in Figure 22 clearly indicate that annual burning and grazing for 47 years
resulted in no change in the forage potential of the grass sward in the moist savanna whereas the
less frequent biennial and triennial burning combined with grazing caused a marked improvement
in the forage potential of the grass vegetation. This is further evidence that moist savanna benefits
from burning as a grazing resource for wild grazing ungulates, but not as frequently as annual
burning. This result supports the general experience of land users that in moist savannas fire is an
important and essential management practice in maintaining the forage potential of the grass
sward for grazing animals.
The interacting effects of frequency of burning and grazing on the grass forage potential in
arid savanna in the Kruger National Park is presented in Figure 23.
D3.5-1-40-1000
Page 46 of 78
FORAGE SCORE
453
500
450
400
350
300
410
279
401
296
262
250
200
150
100
50
0
ANNUAL
1954
ANNUAL
2001
BIEN4NIAL
1954
BIENNIAL
2001
TRIENNIAL
1954
TRIENNIAL
2001
FREQUENCY OF BURN
Figure 23: The interacting effects of frequency of burning and grazing on the grass forage
potential in the arid Mopane Shrub at Mopane in the Experimental Burn Plot trial in the Kruger
National Park during the period 1954 to 2001.
Again the complete contrast in the interacting effects of fire and grazing in arid savanna
compared to moist savanna with annual, biennial and triennial burns all causing a marked decrease
in the forage potential of the grass sward. In this case all three frequencies of burning had similar
effects on decreasing the grass forage potential clearly indicating that even triennial burning
combined with grazing is too frequent as a management practice in arid savannas.
3.3.1.1.2.3.3
Grass Fuel Potential
The overall effects of fire and grazing on the grass fuel potential of the grass sward in the Sour
Bushveld, Combretum Woodland, Knobthorn/Marula Savanna and Mopane Shrub replicates of the
Experimental Burn Plot trial in the Kruger National Park are presented in Figure 24.
GRASS FUEL POTENTIAL
800
700
600
632
668
638
612
571
558
466
500
487
522
556
463
385
400
300
FIRE + GRAZING
GRAZING
200
100
0
SOUR
BUSHVELD
1954
SOUR
BUSHVELD
2001
Combretum
WOODLAND
1954
Combretum KNOBTHORN/ KNOBTHORN/
WOODLAND MARULA
MARULA
2001
1954
1998
MOPANE
SHRUB
1954
MOPANE
SHRUB
2001
VEGETATION LANDSCAPE
Figure 24: The overall effects of fire and grazing on the grass fuel potential in the Sour Bushveld
at Pretoriuskop, the Combretum Woodland at Skukuza, the Knobthorn/Marula Savanna near Satara
D3.5-1-40-1000
Page 47 of 78
and the Mopane Shrub at Mopane in the Experimental Burn Plot trial in the Kruger National Park
during the period 1954 to 2001.
The results in Figure 24 show that fire and grazing resulted in an increase in the fuel
potential of the grass sward in the more moist Sour Bushveld and Combretum Woodland and a
decrease in the arid Knobthorn/Marula Savanna and Mopane vegetation landscapes when
compared to its condition in 1954 at the initiation of the Experimental Burn Plot trial. Grazing in the
absence of fire also had a similar effect on the grass fuel potential except that in the arid Mopane
Shrub the fuel potential remained similar to its condition in 1954. These results are very similar to
the effects of fire and grazing had on the grass forage potential which is not surprising in that the
same grass species are being used to estimate the forage and fuel potentials using their respective
forage and fuel factors. These forage and fuel factors are in numerous cases very similar for the
different grass species and only differ for unpalatable species like Bothriochloa radicans (Forage =
2; Fuel = 7), Cymbopogon plurinodis (Forage = 3; Fuel = 7) and Hyperthelia dissoluta (Forage =
4; Fuel = 10). The coefficient of determination (r2) describing the statistical relationship between
the forage and fuel factors for the different grass species in the Kruger National Park is 0.6393
which illustrates the significant but not exact relationship between the forage and fuel factors.
Nevertheless these results also show that in respect of the production of grass fuel for generating
either prescribed fires or wildfires, moist savannas are better adapted to burning than arid
savannas in terms of the effect of fire and grazing on the fuel potential of the grass sward.
GRASS FUEL POTENTIAL
The interacting effects of season of burning and grazing on the grass fuel potential in moist
savanna in the Kruger National Park is presented in Figure 25.
800
655
700
600
561
593
611
568
615
599
539
562
574
500
400
300
200
100
0
SPRING
1954
SPRING EARLY
EARLY
LATE
LATE AUTUMN AUTUMN WINTER WINTER
2001
SUMMER SUMMER SUMMER SUMMER
1954
2001
1954
2001
1954
2001
1954
2001
SEASON OF BURN
Figure 25: The interacting effects of season of burning and grazing on the grass fuel potential in
the moist Sour Bushveld at Pretoriuskop in the Experimental Burn Plot Trial in the Kruger National
Park during the period 1954 to 2001.
The results in Figure 25 show that in all cases the season of burn and subsequent grazing
resulted in an increase in the fuel potential of the grass sward in the moist savanna at
Pretoriuskop, as was the case for the grass forage potential. The only difference was that in terms
D3.5-1-40-1000
Page 48 of 78
of the fuel potential the effects of all the seasons of burning were very similar all resulting in high
fuel scores in excess of 500 illustrating the high potential for intense fires in moist savanna.
The interacting effects of season of burning and grazing on the grass fuel potential in arid
savanna in the Kruger National Park is presented in Figure 26.
GRASS FUEL POTENTIAL
700
600
576
577
548
559
528
500
391
404
400
302
325
302
300
200
100
0
SPRING
1954
SPRING EARLY
EARLY
LATE
LATE
AUTUMN AUTUMN WINTER
2001
SUMMER SUMMER SUMMER SUMMER
1954
2001
1954
1954
2001
1954
2001
WINTER
2001
SEASON OF BURN
Figure 26: The interacting effects of season of burning and grazing on the grass fuel potential in
the arid Mopane Shrub at Mopane in the Experimental Burn Plot trial in the Kruger National Park
during the period 1954 to 2001.
As was the case with the grass forage potential all the different seasons of burning and
grazing resulted in a marked decrease in the fuel potential of the grass sward in the arid savanna
in the Mopane Shrub vegetation landscape. The fuel potential was reduced from being very high in
1954 down to a medium potential in the late summer, autumn and winter burns. Again this effect
can be attributed to the heavier grazing associated with these times of burning rather than the
direct effects of burning on the grass sward for the same reasons given for the treatment effects
on the grass forage potential.
The interacting effects of frequency of burning and grazing on the grass fuel potential in moist
savanna in the Kruger National Park is presented in Figure 27.
D3.5-1-40-1000
Page 49 of 78
GRASS FUEL POTENTIAL
700
600
628
565
554
566
609
497
500
400
300
200
100
0
ANNUAL
1954
ANNUAL
2001
BIENNIAL
1954
BIENNIAL
2001
TRIENNIAL
1954
TRIENNIAL
2001
FREQUENCY OF BURN
Figure 27: The interacting effects of frequency of burning and grazing on the grass fuel potential
in the moist Sour Bushveld at Pretoriuskop in the Experimental Burn Plot trial in the Kruger National
Park during the period 1954 to 2001.
The results in Figure 27 indicate that annual burning and grazing since 1954 caused a
marginal decline in the fuel potential of the grass sward in the moist savanna in the Sour Bushveld
but biennial and triennial burning and grazing resulted in a marginal increase in the grass fuel
potential. Again these results are similar to the effects of frequency of burning on the grass forage
potential except that in this case the effects were marginal whereas in the case of the forage
potential the effects were more pronounced. A striking feature about the fuel potential of the grass
sward in the moist savanna at Pretoriuskop was that irrespective of treatment the grass sward had
a very high potential to produce grass fuel and resultant high intensity fires, a result strongly
supported by research (Govender et al, 2006) and field experience in this region of the Park.
The interacting effects of frequency of burning and grazing on the grass fuel potential in arid
savanna in the Kruger National Park is presented in Figure 28.
D3.5-1-40-1000
Page 50 of 78
700
591
574
535
600
FUEL SCORE
500
333
400
353
331
300
200
100
0
ANNUAL
1954
ANNUAL
2001
BIEN4NIAL 1954
BIENNIAL
2001
TRIENNIAL 1954
TRIENNIAL 2001
FREQUENCY OF BURNING
Figure 28: The interacting effects of frequency of burning and grazing on the grass fuel potential
in the arid Mopane Shrub at Mopane in the Experimental Burn Plot trial in the Kruger National Park
during the period 1954 to 2001.
Similar to the effects of frequency of burning on the grass forage potential, the frequency
of burning and grazing in the arid savanna caused a marked decrease in the fuel potential of the
grass sward from very high levels (>500) to medium levels (331 to 353). Again this can be
ascribed to the effects of frequent heavy grazing after the fires rather than the effects of frequency
of burning on the grass sward and the results clearly indicate that even triennial burning in an open
grazing system is too frequent to maintain the grass fuel potential of arid savanna.
3.3.1.1.2.3.4 Basal Cover
The basal cover of the grass sward is one of the fundamentally important functional characteristics
of the grass sward that influences the rate of accelerated soil erosion. Areas with a high basal
cover have a low soil erosion potential and vice versa (Snyman, 1999). As mentioned earlier, the
point to tuft distance measured during a point quadrat survey of the grass sward in the
Experimental Burn Plot trial was used as a surrogate measure of basal cover and to serve as an
indicator of the potential for accelerated soil erosion as influenced by the different burning and
grazing treatments.
A covariance analysis of the effects of the different burning and grazing treatments on the
basal cover of the grass sward showed that the initial basal cover in the different plots had no
statistically significant effect on the subsequent point to tuft distances recorded during the followup surveys conducted 47 years after the initiation of the trial. Consequently, the effects of the
different burning and grazing treatments on the potential for soil erosion in the different plots will
be assessed in terms of the point to tuft distance recorded during the follow-up surveys. Personal
experience with such data has shown that the following general guidelines can be used in the
Kruger National Park to assess the potential for accelerated soil erosion as influenced by point to
tuft distance – see Table 5.
D3.5-1-40-1000
Page 51 of 78
Table 5: Guidelines for assessing the soil erosion potential in the Kruger National Park in South
Africa.
POINT TO TUFT DISTANCE - ACCELERATED
cm
POTENTIAL
<5 cm
SOIL
EROSION
Low potential
5 – 10 cm
Moderate potential
>10 cm
High potential
The overall effects of fire and grazing on the Point To Tuft Distance (PTTD) of the grass
sward in the Sour Bushveld, Combretum Woodland, Knobthorn/Marula Savanna and Mopane Shrub
replicates of the Experimental Burn Plot trial in the Kruger National Park are presented in Figure 29.
POINT TO TUFT DISTANCE - cm
9.0
8.0
7.0
8.0
6.5
7.0
5.2
6.0
5.0
4.0
4.0
3.0
FIRE + GRAZING
GRAZING
3.9
2.8
3.0
2.0
1.0
0.0
SOUR BUSHVELD
2001
Combretum
WOODLAND
2001
KNOBTHORN/
MOPANE SHRUB
MARULA SAVANNA
2000
2001
VEGETATION LANDSCAPES
Figure 29: The overall effects of fire and grazing on the point to tuft distance of the grass sward
in the Sour Bushveld at Pretoriuskop, the Combretum Woodland at Skukuza, the Knobthorn/Marula
Savanna near Satara and the Mopane Shrub at Mopane in the Experimental Burn Plot trial in the
Kruger National Park during the period 1954 to 2001.
The results in Figure 29 show that there is a definite trend in the PTTD’s as influenced by
increasing rainfall in the different vegetation landscapes with the moist Sour Bushveld in the south
of the Kruger National Park at Pretoriuskop having the least PTTD’s and the extremely arid Mopane
Shrub in the north of the Park at Mopane having the greatest PTTD’s. Fire and grazing and grazing
alone had virtually similar effects on the PTTD in the more humid Sour Bushveld and Combretum
Woodland vegetation landscapes, both having a low potential for accelerated soil erosion as
represented by the PTTD. Conversely, the fire and grazing resulted in significantly lower PTTD’s
D3.5-1-40-1000
Page 52 of 78
than grazing alone in the Knobthorn/Marula Savanna and Mopane Shrub vegetation landscapes.
This can be explained in terms of the fire and grazing treatments maintaining the grass
sward in these vegetation landscapes in a less moribund and actively growing condition, resulting
in a greater density of grass plants and therefore a lower PTTD. In contrast, in the plots that were
not burnt and were only grazed, the grass sward would have been less attractive to grazing
animals and therefore less intensively grazed tending to develop into a moribund and overgrown
condition characterised by widely spaced large grass tufts with a low basal cover and significant
PTTD. Nevertheless it is surprising that this occurred in the more arid vegetation landscapes rather
than in the higher rainfall vegetation types with a greater potential for plant growth in the unburnt
plots leading to the development of a moribund and overgrown grass sward with a low basal cover.
Finally the results in Figure 29 indicate that fire and grazing have resulted in a low overall effect on
accelerated soil erosion in the moist savannas decreasing to a moderate level in the arid savannas.
Unfortunately PTTD’s are not available for 1954 when the Experimental Burn Plot trial was initiated
so it is not possible to determine whether the overall potential for accelerated soil erosion has
increased or decreased in response to fire and grazing during this period. Suffice it to say that the
percentage basal cover recorded in 1954 for the Sour Bushveld was 10%, the Combretum
Woodland was 8%; in 1960 the Knobthorn/Marula Savanna was 8% and the Mopane Shrub was
9%. Based on personal field experience these basal covers for the grass sward in these vegetation
landscapes are high and represent rangeland that would not be prone to accelerated soil erosion as
influenced by basal cover. This would therefore suggest that the overall effect of fire and grazing in
the moist and arid savannas has not had a marked effect on the potential of the grass sward with
respect to accelerated soil erosion. The interacting effect of season of burning and grazing on the
point to tuft distance of the grass sward in moist savanna in the Kruger National Park is presented in
Figure 30.
4.1
POINT TO TUFT DISTANCE - cm
4.5
4.0
3.1
3.5
3.0
2.8
2.8
2.5
2.5
2.0
1.5
1.0
0.5
0.0
SPRING 2001
EARLY
SUMMER 2001
LATE SUMMER
2001
AUTUMN 2001
WINTER 2001
SEASON OF BURN
Figure 30: The interacting effect of season of burning and grazing on the point to tuft distance of
the grass sward in the moist Sour Bushveld at Pretoriuskop in the Experimental Burn Plot trial in
the Kruger National Park during the period 1954 to 2001.
D3.5-1-40-1000
Page 53 of 78
The results in Figure 30 indicate that the season of burning and grazing have had no
marked effect on the PTTD of the grass sward with the PTTD in all treatments being less than 5cm
suggesting that the season of burning has had minimal effect on the potential for soil erosion in the
moist savanna. This is not a surprising result because the Sour Bushveld in the south of the Kruger
National Park receives adequate and regular rainfall thereby ensuring significant growth of the
grass sward after a burn and the maintenance of a high basal cover.
The interacting effect of season of burning and grazing on the point to tuft distance of the
grass sward in arid savanna in the Kruger National Park is presented in Figure 31.
8.4
POINT TO TUFT DISTANCE cm
9.0
8.0
7.6
7.2
6.6
7.0
6.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
SPRING
2001
EARLY
SUMMER
2001
LATE
SUMMER
2001
AUTUMN
2001
WINTER
2001
SEASON OF BURN
Figure 31: The interacting effect of season of burning and grazing on the point to tuft distance of
the grass sward in the arid Mopane Shrub at Mopane in the Experimental Burn Plot trial in the
Kruger National Park during the period 1958 to 2001.
As was the case in the moist savanna, the results in Figure 31 indicate that the season of burning
and grazing had a minimal effect on the PTTD of the grass sward and therefore the potential for
accelerated soil erosion. Nevertheless because of the lower annual rainfall in this vegetation
landscape the grass sward does have an inherently lower PTTD and therefore a moderate potential
for accelerated soil erosion.
The interacting effect of frequency of burning and grazing on the point to tuft distance of the
grass sward in moist savanna in the Kruger National Park is presented in Figure 32.
D3.5-1-40-1000
Page 54 of 78
3.3
3.1
POINT TO TUFT DISTANCE - cm
3.5
2.8
3.0
2.5
2.0
1.5
1.0
0.5
0.0
ANNUAL 2001
BIENNIAL 2001
TRIENNIAL 2001
FREQUENCY OF BURN
Figure 32: The interacting effect of frequency of burning and grazing on the point to tuft distance
of the grass sward in the moist Sour Bushveld at Pretoriuskop in the Experimental Burn Plot trial in
the Kruger National Park during the period 1954 to 2001.
The results in Figure 32 show that the frequency of fire and grazing had a minimal effect on
the PTTD of the grass sward and therefore the potential for accelerated soil erosion which was low
when estimated in 2001 i.e. PTTD <5cm.
POINT TO TUFT DISTANCE - cm
The interacting effect of frequency of burning and grazing on the point to tuft distance of the
grass sward in arid savanna in the Kruger National Park is presented in Figure 33.
7.4
9.0
7.0
8.0
7.0
5.6
6.0
5.0
4.0
3.0
2.0
1.0
0.0
ANNUAL 2001
BIENNIAL 2001
TRIENNIAL 2001
FREQUENCY OF BURN
Figure 33: The interacting effect of frequency of burning and grazing on the point to tuft distance
of the grass sward in the arid Mopane Shrub at Mopane in the Experimental Burn Plot trial in the
Kruger National Park during the period 1958 to 2001.
D3.5-1-40-1000
Page 55 of 78
The results in Figure 33 also show that as in the moist savanna, the frequency of burning
and grazing in the arid savanna has not had a marked effect on the PTTD of the grass sward. The
PTTD measurements recorded during the follow-up surveys in 2000 indicated that the potential for
accelerated soil erosion was estimated to be moderate. However, this result for the effect of the
frequency of fire and grazing in the arid savanna must be treated with caution. During an
inspection of the annual burn plot in the arid savanna in the Knobthorn/Marula Savanna vegetation
landscape at Satara during November 2008, the visual assessment of the PTTD of the grass sward
indicated a very sparse cover of grass and other herbaceous plants. This is clearly illustrated in an
aerial and ground level photograph of the annually burnt plot in the Nwanedzi replicate of the
Experimental Burn Plot trial in the Knobthorn/Marula Savanna at Satara – see Figure 34.
Aerial view- annual burn
Nwanedzi replicate
Ground view- annual burn
Nwanedzi replicate
Figure 34: Aerial and ground views of the plot burnt annually since 1958 in the Nwanedzi replicate of
the arid Knobthorn/Marula Savanna at Satara in the Experimental Burn Plot trial in the Kruger
National Park. Note the extremely low basal cover associated with this fire and grazing treatment
recorded on the 14th October, 2008.
This visual assessment of the PTTD of the effects of annual burning and grazing in the arid
savanna in the Kruger National Park does not provide support for the effects of this fire and
grazing treatment recorded in the follow-up survey conducted in 1998 in this replicate. There are
two possible explanations for these conflicting results i.e. the PTTD distance in the plot illustrated
in Figure 35 has decreased very significantly since the follow-up survey was conducted in 1998, or
the PTTD measurement representing the basal cover of the grass sward is not a reliable surrogate
measure of basal cover in arid savannas. This latter explanation is possibly due to the high
proportion of annual herbaceous species that are greatly influenced by a highly variable rainfall in
arid savannas. Consequently more research attention needs to be focused on these conflicting
results on the effects of frequency of fire and grazing on the basal cover of the grass sward in arid
savanna.
D3.5-1-40-1000
Page 56 of 78
3.3.1.2 Effects of Fire and Grazing on the Botanical Composition of the Grass Sward
in Moist and Arid Savannas
The effects of wild ungulate grazing species on the botanical composition of the grass sward in
terms of Decreaser and Increaser grass species in moist and arid savannas, using the results of the
Experimental Burnt Plot trial in the Kruger National Park, was determined.
The overall effects of fire and grazing on the proportions of Decreaser and Increaser grass
species in moist savanna in the Experimental Burn Plot trial in the Kruger National Park are
presented in Figure 35.
90
77
PERCENTAGE - %
80
70
60
54
48
FIRE + GRAZING
GRAZING
50
40
30
30
20
27
25
24
13
20
10
0
DECREASER
1954
DECREASER
2001
INCREASER I
1954
INCREASER I
2001
INCREASER II
1954
INCREASER 11
2000
GRASS CATEGORIES
Figure 35: The overall effects of fire and grazing on the proportions of Decreaser and Increaser
grass species in the moist Sour Bushveld at Pretoriuskop in the Experimental Burn Plot Trial in the
Kruger National Park during the period 1954 to 2001.
The results in Figure 35 show that both fire and grazing, and grazing alone, have both
caused a marked increase in the proportion of Decreaser grass species, a significant reduction in
the proportion of Increaser I grass species and a marginal decrease in the proportion of Increaser
II grass species. These results suggest that there has been an increase in the intensity of grazing
since 1954 as indicated by the decrease in the proportion of Increaser I grass species that are not
well adapted to intensive grazing. This can be ascribed to the application of annual, biennial and
triennial burns that would have significantly increased the intensity and frequency of grazing of the
grass sward in the burnt plots. This also may have indirectly increased the grazing pressure in the
unburnt plots as higher numbers of grazing animals were attracted to the adjacent burnt plots.
This in turn has resulted in a positive response from the Decreaser grass species that are better
adapted to grazing hence their significant increase. The Decreaser species that have significantly
increased are Setaria sphacelata and Panicum maximum and the Increaser I species that has
decreased is Hyperthelia dissoluta.
The overall effects of fire and grazing on the proportions of Decreaser and Increaser grass
species in the arid savanna in the Experimental Burn Plot trial in the Kruger National Park are
presented in Figure 36.
D3.5-1-40-1000
Page 57 of 78
82
PERCENTAGE - %
90
80
70
60
58
55
44
42
50
40
30
20
FIRE + GRAZING
GRAZING
18
10
1
0 0
0
DECREASER
1954
DECREASER
2001
INCREASER I
1954
INCREASER I INCREASER II INCREASER II
2001
1954
2000
GRASS CATEGORIES
Figure 36: The overall effects of fire and grazing on the proportions of Decreaser and Increaser
grass species in the arid Mopane Shrub at Mopane in the Experimental Burn Plot Trial in the Kruger
National Park during the period 1954 to 2000.
As has already been determined on the effects of fire and grazing on the functional
characteristics in the moist and arid savannas the results in Figure 36 show the opposite effects of
fire and grazing on the botanical composition of the grass sward in the arid savanna with the
proportions of Decreaser grass species declining significantly and Increaser II grass species
increasing markedly. The results show that the Decreaser grass species that have decreased
markedly were Panicum maximum and Panicum coloratum and the Increaser II grass species that
had increased significantly were Aristida congesta and Enneapogon cenchroides. In the absence of
fire, grazing alone did not cause as marked a reduction in the Decreaser grass species and increase
in the Increaser II grass species.
This clearly illustrates that in the arid savanna the grass sward is very sensitive to the
combined effects of fire and grazing resulting in a marked decrease in the productive and palatable
Decreaser grass species and a dominance of less productive and palatable Increaser II grass
species. These results provide the reason why fire and grazing causes an increase in the forage
potential of the grass sward in the moist savanna and a decrease in the arid savanna presented in
Figure 19. Finally an interesting feature of the botanical composition of arid savannas is the almost
complete absence of Increaser I grass species in both the 1954 and follow-up survey data recorded
between 1998 and 2001. This phenomenon was observed in all the grass surveys conducted in the
different replicates of the arid savannas recorded at Satara and Mopane and also to a degree in the
Combretum Woodland at Skukuza.
For the sake of simplicity and understanding, the effect of season of burning and grazing on
the botanical composition of the grass sward will be limited to the proportion of Decreaser grass
species in moist and arid savannas in the Kruger National Park. The effects of season of burning and
grazing on the proportion of Decreaser grass species in moist savanna are presented in Figure 37.
D3.5-1-40-1000
Page 58 of 78
60
PERCENTAGE - %
54
53
50
50
41
37
40
30
20
24
22
18
16
20
10
0
SPRING
1954
SPRING
2001
EARLY
EARLY
LATE
LATE AUTUMN AUTUMN WINTER
SUMMER SUMMER SUMMER SUMMER
1954
2001
1954
1954
2001
1954
2001
WINTER
2001
SEASON OF BURN
Figure 37: The interacting effects of season of burning and grazing on the proportion of Decreaser
grass species in the moist Sour Bushveld at Pretoriuskop in the Experimental Burn Plot trial in the
Kruger National Park during the period 1954 to 2001.
The results in Figure 37 indicate that there was a marked increase in the proportion of
Decreaser grass species in the grass sward in all the treatments. However the increase was noticeably
less in the winter burning and grazing treatment and as was discussed in the effects of fire and
grazing on the grass forage potential, where the same trend emerged, this is best ascribed to the
effect of grazing rather than the season of burn. It was shown in the review of literature that burning
in winter when the grass is dormant resulted in a significantly better recovery in the grass sward.
This would therefore suggest that the main reason for the smaller increase in the
proportion of Decreaser grass species was the consistent heavy grazing of the grass plants after
the winter burn following the first spring rains at the commencement of the growing season. At the
end of winter high quality grazing is not abundant and the succulent and highly palatable regrowth
of the burnt grass in a small plot of seven hectares, surrounded by extensive areas of generally
unburnt rangeland would be prone to intense grazing. This would have a negative effect on the
vigour of the grass sward resulting in a lower proportion of Decreaser grass species. The effects of
season of burning and grazing on the proportion of Decreaser grass species in arid savanna are
presented in Figure 38.
D3.5-1-40-1000
Page 59 of 78
PERCENTAGE - %
70
60
60
55
55
54
53
50
40
30
30
21
20
13
17
12
10
0
SPRING
1954
SPRING
2001
EARLY
EARLY
LATE
LATE
AUTUMN
SUMMER SUMMER SUMMER SUMMER
1954
1954
2001
1954
2001
AUTUMN WINTER
2001
1954
WINTER
2001
SEASON OF BURN
Figure 38: The interacting effects of season of burning and grazing on the proportion of Decreaser
grass species in the arid Mopane Shrub at Mopane in the Experimental Burn Plot trial in the Kruger
National Park during the period 1954 to 2000.
In contrast to the moist savanna, all seasons of burning and grazing caused a significant
reduction in the proportion of Decreaser grass species in the arid savanna in the Mopane Shrub. The
decrease in the proportion of the Decreaser grass species provides an explanation for the decrease in
the forage potential of the grass sward in the arid savanna presented in Figure 19 i.e. a decrease in
the more palatable and productive Decreaser grass species is the reason why the forage potential of
the grass sward decreased with fire and grazing in the arid savanna.
The effects of frequency of burning and grazing on the proportion of Decreaser grass species
in moist savanna are presented in Figure 39.
D3.5-1-40-1000
Page 60 of 78
PERCENTAGE - %
60
50
45
50
40
32
30
20
23
18
17
10
0
ANNUAL
1954
ANNUAL
2001
BIENNIAL
1954
BIENNIAL
2001
TRIENNIAL
1954
TRIENNIAL
2001
FREQUENCY OF BURN
Figure 39: The interacting effects of frequency of burning and grazing on the proportion of
Decreaser grass species in the grass sward in the moist Sour Bushveld at Pretoriuskop in the
Experimental Burn Plot trial in the Kruger National Park during the period 1954 to 2001.
The results in Figure 39 clearly show that frequency of burning and grazing had a marked
positive effect on the proportion of Decreaser grass species in the moist Sour Bushveld at
Pretoriuskop. Annual burning and grazing had the least effect and this increased with a decrease in
the frequency of burning. Again these effects can be ascribed mainly to the effects of intensity and
frequency of grazing after the burns because the annual burns were applied in winter when the grass
sward was dormant and when research has shown that fire has the least depressing effect on the
recovery of the grass sward (Tainton et al, 1977; Dillon, 1980; Trollope, 1987 and Everson et al,
1988). These results would suggest that the intense and frequent grazing after the annual burns
caused the lowest proportion of Decreaser grass species to develop with these treatments, rather
than the effect of annual burning per se. The results also show diminishing returns in the rate of
increase in the proportion of Decreaser grass species after biennial burning, indicating that the grass
sward was beginning to show the ill effects of self shading and was in the initial stages of becoming
moribund. This provides the explanation for the effects of burning and grazing on the grass forage
potential which increased at a slower rate after triennial burning compared to biennial burning – see
Figure 22.
The effects of frequency of burning and grazing on the proportion of Decreaser grass
species in arid savanna are presented in Figure 40.
D3.5-1-40-1000
Page 61 of 78
PERCENTAGE - %
70
60
55
60
54
50
40
30
14
20
21
16
10
0
ANNUAL
1954
ANNUAL
2001
BIENNIAL
1954
BIENNIAL
2001
TRIENNIAL
1954
TRIENNIAL
2001
FREQUENCY OF BURN
Figure 40: The interacting effects of frequency of burning and grazing on the proportion of
Decreaser grass species in the grass sward in the arid Mopane Shrub at Mopane in the
Experimental Burn Plot trial in the Kruger National Park during the period 1954 to 2001.
In complete contrast the results in Figure 40 indicate that in the arid Mopane Shrub at
Mopane the frequency of burning and grazing treatments resulted in a marked decrease in the
proportion of Decreaser species in the grass sward, no matter what the frequency. There was a
slight trend for the annual, biennial and triennial burning and grazing treatments to have a positive
linear effect on the proportion of Decreaser grass species but this was very marginal. These results
support the earlier conclusion on the effects of frequency of burning on the grass forage potential
in arid savanna that even triennial burning combined with grazing is too frequent as a management
practice in arid savannas.
D3.5-1-40-1000
Page 62 of 78
3.3.1.3
Overall Conclusions on the Effects of Grazing with Wild Ungulates after
Burning on Range Condition in Moist and Arid Savannas
The following overall conclusions can be drawn on the long term effects of fire and grazing by wild
ungulates on the functional characteristics and botanical composition of the grass sward in the
moist and arid savannas in the Kruger National Park in South Africa viz. the grass forage and fuel
potential, the basal cover and resistance to accelerated soil erosion and the proportion of
Decreaser and Increaser grass species in response to different grazing intensities and frequencies
along a grazing gradient:
3.3.1.3.1 Grass Forage Potential:

Fire and grazing resulted in a marked overall increase in the forage potential of moist
savanna and a significant decrease in arid savanna. This is a clear indication that moist
savannas are better adapted to burning and grazing than arid savannas;

Grazing in the absence of fire resulted in a greater increase in the forage potential than fire
and grazing in the moist savanna but the resultant herbaceous material was generally in a
moribund and unpalatable condition, characterised by high tick loads indicating that some
form of defoliation was necessary for improving the quality of the grass forage. In the arid
savanna grazing alone resulted in the maintenance of the original high forage potential of the
grass sward;

In all cases the season of burn resulted in an overall increase in the forage potential of the
grass sward in moist savanna. However, burning and grazing in winter resulted in a less
marked increase in the forage potential, but this cannot be ascribed to the season of burning
alone but rather to the effects of the consistently intense and frequent grazing after the fire
following the first spring rains. In the arid savanna all the seasons of burn and grazing
resulted in a decrease in the forage potential of the grass sward with the spring and early
summer burns causing a lower decrease, possibly due to higher and more reliable rainfall at
this time of the year;

Decreasing frequency of burning and grazing resulted in an increase in the grass forage
potential in the moist savanna with triennial burning resulting in the highest forage potential.
In the arid savanna annual, biennial and triennial frequencies of burning all reduced the
forage potential of the grass sward equally, clearly indicating that even triennial burning with
grazing is too frequent in arid savannas. A subsequent inclusion of quadrennial and sexennial
burning frequencies were subsequently tested in the arid savanna but these had no
significant effect on improving the forage potential of the grass sward again showing that
even more infrequent fires than sexennial burning are necessary for maintaining the forage
potential of arid savanna.
3.3.1.3.2 Grass Fuel Potential

The overall effect of fire and grazing on the fuel potential of the grass sward was similar to
that which occurred with the forage potential in both the moist and arid savannas. This result
is ascribed to the fact that in many cases the forage and fuel factors associated with the
genetic potential of the different grass species is similar except for notable exceptions. Again
this result clearly shows that moist savannas are better adapted to burning than arid
savannas with respect to producing adequate grass fuel to sustain a fire;
D3.5-1-40-1000
Page 63 of 78

All seasons of burn had a similar effect on the fuel potential of the grass sward in moist
savanna, all maintaining high fuel potentials illustrating the inherent ability of moist savannas
to generate intense fires. In contrast, all seasons of burn caused a marked decrease in the
fuel potential of the grass sward with, as in the case of forage potential, the spring and early
summer burns showing less of a decrease in fuel potential;

All frequencies of burn maintained a high grass fuel potential in moist savanna as was the
case with the forage potential. In contrast, frequency of burn in arid savanna caused a
marked decrease in the grass fuel potential with annual, biennial and triennial having similar
effects, again as was the effect on the grass forage potential. The effect of frequency of burn
and grazing in moist and arid savannas clearly illustrates the inherent contrasting potentials
of these types of savanna to support high intensity fires.
3.3.1.3.3 Basal Cover

The overall effects of fire and grazing on the basal cover of the grass sward and its potential
for accelerated soil erosion, as represented by the point to tuft distance (PTTD), followed a
distinct linear trend according to annual rainfall with moist savanna having the lowest PTTD
and arid savanna the greatest PTTD. This would imply contrasting low and higher potentials
respectively for accelerated soil erosion with moist savanna having a low potential and arid
savanna a moderate potential in response to fire and grazing;

Season of burn had no marked differential effect on the PTTD in both moist and arid
savannas and therefore had a minimal effect on the potential for accelerated soil erosion.

Frequency of burn also had minimal effect on PTTD in moist savanna with annual, biennial
and triennial burning and grazing treatments all resulting in a low potential for accelerated
soil erosion. Frequency of burn also had minimal effect on PTTD in arid savanna but this
result must be treated with caution. This is because a visual assessment during November,
2008 of the basal cover of the grass sward in one of the annual burn and grazing replicates
of the Experimental Burn Plot trial in the arid Knobthorn savanna at Satara indicated that it
had a very high PTTD and sparse basal cover. This suggests that possibly where there is a
high component of annual herbaceous plants that can, and do occur with heavy grazing in
arid savanna, the PTTD is not a reliable surrogate measure for basal cover under these
circumstances and the results must be treated with caution.
3.3.1.3.4 Botanical Composition of the Grass Sward

Both the overall effects of fire and grazing, and grazing alone, caused a significant increase in
the proportion of Decreaser grass species and a marked decrease in Increaser I grass species
in moist savanna. This is ascribed to a significant increase in the intensity and frequency of
grazing in response to the introduction of the annual, biennial and triennial burning
treatments since 1954. The increase in the grazed only plots is ascribed to a higher grazing
pressure resulting from the attraction of grazing animals to the adjacent burnt plots. In the
arid savanna the opposite effect occurred with a dramatic decrease in the Decreaser grass
species but in this case, a marked increase in Increaser II grass species in response to the
excessively heavy grazing associated with the annual, biennial and triennial burning
treatments. This explains why the overall effects of fire and grazing alone resulted in marked
increases in the forage potential of the grass sward in moist savanna and vice versa in arid
savanna;
D3.5-1-40-1000
Page 64 of 78

In moist savanna all seasons of burning and grazing caused a significant increase in
Decreaser grass species and in arid savanna a marked decrease again providing an
explanation for the effect of fire and grazing on the forage potential of the grass sward;

Frequency of burning and grazing caused a linear increase in the proportion of Decreaser
grass species with annual, biennial and triennial burning in moist savanna. In contrast, these
frequencies of burning caused a uniform decrease in the proportion of Decreaser species in
the grass sward again explaining the effects of frequency of burning on grass forage
potential.
These general conclusions provide a valuable insight and understanding to the overall
effects of fire and grazing on the functional characteristics of the grass sward in terms of forage
production to support wild grazing ungulates, fuel production for sustaining fires and the basal
cover of the grass sward to prevent accelerated soil erosion. These data provide the general
guiding principles that must be considered for formulating prescribed burning programs for wildlife
management in African grasslands and savannas.
D3.5-1-40-1000
Page 65 of 78
3.4
Effects of Fire & Grazing On Animal Ratios in the Ngorongoro Crater, Tanzania
A significant example of the effects of fire and grazing on the ratio of bulk grazers to concentrate
grazers is found in the Ngorongoro Crater located in the Ngorongoro Conservation Area in north
western Tanzania. The Crater is approximately 30 000ha in size of which 25 000ha comprise the
floor and 5000ha the steeply rising slopes of the rim. The walls of the Crater rise 500m (1 600 ft)
above the floor and it is largely a self-contained ecological unit but is linked to the adjacent
Serengeti Plains and the Ngorongoro Highlands (Trollope & Trollope, 1995). The vegetation of the
Crater comprises primarily grassland with woody vegetation being limited to the Lerai Forest,
dominated by Acacia xanthophloea and occurring in the southern sector, and dry savanna
woodland scattered along the steeply rising walls of the Crater.
The grassland comprises basically two types , the short and long grass plains. The indicator
species in the Short Grass Plains are Sporobolus ioclados, Digitaria macroblephera, Eragrostis
species, Andropogon greenwayi, Cynodon dactylon and Cynodon nlemfuensis . In the Tall Grass
Plains the indicator species are Chloris gayana, Hyparrhenia rufa, Themeda triandra, Aristida
kenyensis and Pennisetum mezianum (Hanby & Bygott, 1992). The mean annual rainfall in the
Crater is approximately 600 mm in the Tall Grass Plains and 500 mm in the Short Grass Plains
(Runyuro, 1995). The Crater has one of the largest and most diverse concentrations of herbivores
occurring anywhere in Africa having 31 different ungulate species ranging from the dainty dik dik
to the imposing elephant. The most abundant species are the wildebeest but other common
ungulate species include buffalo, zebra, Grant’s gazelle, Thompson’s gazelle and Kongoni (Trollope
& Trollope, 1995).
During the early 1970’s, burning of the grasslands in the Crater was prohibited (Trollope et
al,2003) and replaced the previous fire management practices of the resident Masaai pastoralists
who burnt the Crater regularly to provide nutritious grazing for their livestock and to control ticks.
This change in fire management involving regular burning and grazing by both domestic livestock
and wild ungulate grazing species to no burning and grazing only by wildlife has since the 1970’s
resulted in major changes in the composition of the wild ungulate species utilizing the Ngorongoro
Crater. These changes have been investigated by Estes et al (2006) and clearly show the
significant decline in the wildebeest population between 1964 and 2005 and the marked increase in
the buffalo population between 1970 and 2005 – see Figure 41.
D3.5-1-40-1000
Page 66 of 78
NUMBER
Wildebeest Population - 1964 to 2005
20000
18000
16000
14000
12000
10000
8000
6000
4000
2000
0
1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
YEARS
Buffalo Population - 1970 to 2005
7000
NUMBER
6000
5000
4000
3000
2000
1000
0
1970
1975
1980
1985
1990
1995
2000
2005
YEARS
Figure 41: Changes in the wildebeest and buffalo populations in the Ngorongoro Crater in
Tanzania in response to the withdrawal of burning as a management practice in the 1970’s
D3.5-1-40-1000
Page 67 of 78
Relating the results in Figure 41 which illustrate the increase in the buffalo population and
the decrease in the wildebeest population since the 1970’s to changes in the condition of the grass
sward, an assessment of the the condition of the grasslands in the Crater in 1995 showed that the
standing crop of grass in the Tall Grass Plains in the northern sector of the Crater was
approximately 3500 – 4000 kg/ha. In these northern areas the grass sward was dominated by
Chloris gayana, Pennisetum mezianum and Andropogon greenwayi and except for A greenwayi all
the aforementioned species were showing aerial tillering and becoming moribund. During a followup visit to the Crater in 1998 a repeated assessment of the northern Tall Grass Plains showed that
the grass sward was generally in a moribund condition and therefore provided a plausible
explanation for the decrease in the wildebeest and an increase buffalo populations illustrated in
Figure 41. These results suggest that with the general prohibition on controlled burning that had
been in force since the early 1970’s, the grass sward in the Crater had been able to grow out fully
on a regular basis, particularly in the northern Tall Grass Plains which receive a higher rainfall. This
in turn had resulted in the condition of the grass sward becoming more favourable for buffalo,
which are bulk grazers and prefer long grass, and less favourable for wildebeest which are
concentrate grazers and prefer shorter grass. This preference of different ungulate grazing species
for grassland with different amounts of standing crop was investigated and the results are
presented in Figure 42.
Figure 42: Effect of standing crop of grass on habitat selection by grazing ungulates in the
Ngorongoro Crater in Tanzania.
D3.5-1-40-1000
Page 68 of 78
The results in Figure 42 clearly show that there is a clear tendency for wildebeest and other
smaller ungulate species to select shorter grassland than larger ungulates like buffaloes that prefer
tall growing grassland. Therefore these data provide evidence for the possible effects the
withdrawal of fire and grazing can have on animal ratios involving bulk to concentrate grazers. This
effect is clearly illustrated using the data collected by Estes et al (2006) on the changes that have
occurred in the animal ratios of concentrate to bulk grazers in the Ngorongoro Crater from the
1970’s to 2005 – see Figure 43.
2.4
ANIMAL RATIO:
CONCENTRATE: BULK GRAZERS
2.5
2
1.5
0.8
0.8
1987/ 1995
1996/ 2005
1
0.5
0
1971/ 1986
PERIODS
FIGURE 43: Animal ratios of bulk to concentrate grazers for the periods 1971/ 1986, 1987/ 1995
and 1996/ 2005 in the Ngorongoro Crater in Tanzania (Estes et al, 2006)
The results in Figure 43 indicate that the animal ratio of bulk to concentrate grazers was
significantly greater (2.4) during the period 1971 to 1986 but that it decreased and stabilized at 0.8
from 1987 to 2005. The aforementioned circumstantial evidence would suggest that this decrease
in the animal ratio was primarily the result of the decline in the wildebeest population (concentrate
grazers) and the increase in the numbers of buffaloes in response to the increase in the general
height of the grass sward particularly in the northern Tall Grass Plains, as a result of the
withdrawal of fire as a management tool in the Crater in the 1970’s.
4.
General Discussion and Conclusions
The review of the relationships between fire and grazing in African grasslands and savannas
provides valuable information that can be used in formulating and adapting current range
management practices for domestic livestock husbandry and wildlife management. While
information on the effects of fire and grazing with domestic livestock is limited to moist grasslands
and savannas, clear guidelines have emerged on how to apply appropriate grazing practices that
D3.5-1-40-1000
Page 69 of 78
will maximise livestock production on a sustained basis without resulting in accelerated soil
erosion. For maximising animal performance it is both ecologically permissible and necessary to
graze burnt areas as soon as possible after burning to ensure that the livestock derive maximum
benefit from the highly palatable and nutritious regrowth of the grass sward (Zacharias, 1994).
However the research also showed that it is necessary to apply regular prolonged rest
periods to maintain the vigour and therefore the grazing capacity of the rangeland. Practical field
experience obtained on commercial ranching operations in South Africa have shown that
withdrawing the burnt and heavily grazed areas for a year every four years successfully maintains
the vigour, basal cover and grazing capacity of the grass sward. The other important information
emerging from this review is that the vigour and condition of the grass sward is better maintained
when utilized by cattle rather than sheep even if prolonged rest periods are included in the grazing
program (Kirkman, 2002). The solution to this problem is to ensure that the livestock ratio of cattle
(bulk grazers) to sheep (concentrate grazers) does not exceed one animal unit to six small stock
units. This supports earlier research that had shown that this livestock ratio resulted in superior
animal performance and utilization of the grass sward and minimised selective grazing (Hardy et
al., 1999).
As indicated in the review there is no research information available on the effects of fire
and grazing with domestic livestock in true arid grasslands primarily as a result that controlled
burning was frowned upon in these ecologically sensitive rangelands where the perception was
that burning was highly detrimental to the condition of the grass sward and promoted accelerated
soil erosion. Personal experience with fire and grazing in arid grasslands and savannas in the
Eastern Cape Province in South Africa, the Ngorongoro Crater in Tanzania and the central
highlands of Kenya has shown that prescribed burning can also be used to remove moribund grass
material after above average rainfall periods without any detrimental effect on the condition of the
grass sward provided the burnt area is not continuously grazed for extended periods (years) after
burning and the frequency of subsequent burning is low (>10 years).
Currently Information on the effects of controlled burning and grazing with wild ungulate
grazers in moist and arid grasslands and savannas is generally limited to the Experimental Burn
Plot trial in the Kruger National Park in South Africa. Nevertheless the results from this trial
initiated in 1954 and comprehensively assessed in 1998 to 2001 provide comprehensive and
unique information on the effects of burning and grazing in moist and arid savannas. The results
clearly show that moist savannas are well adapted to fire compared to arid savannas. Assessing
the effects of season and frequency of burning and grazing on the forage, fuel and soil erosion
potential of the grass sward the results showed that season of burning and grazing generally
promoted the forage and fuel potential of the grass sward in moist savanna but to a lesser extent
in the winter burning and grazing treatments. This latter result was ascribed to the subsequent
heavy grazing associated with the winter burning rather than the effects of winter burning per se.
Conversely all seasons of burn resulted in a marked reduction in the forage and fuel potential of
the grass sward in both moist and arid savannas.
Season of burning and grazing had no marked differential effect on the soil erosion
potential of the grass sward in the moist and arid savannas with moist savannas having a low
potential and the arid savannas a moderate potential arising from differences in mean annual
rainfall. The effects of frequency of burning and grazing showed that less frequent burning
promoted the forage potential in moist savanna but uniformly depressed the forage potential in
arid savanna indicating that this type of savanna is not well adapted to frequent burning and
grazing. Frequency of burning had no marked effect on the grass fuel potential in moist savanna
D3.5-1-40-1000
Page 70 of 78
but significantly depressed it uniformly in arid savanna again indicating that this type of savanna is
not well adapted to frequent fires and grazing. In terms of the effect on soil erosion potential
frequency of burning and grazing had no effect on this parameter in moist savanna but had
contradictory results in arid savanna where annual burning resulted in the lowest potential for
accelerated soil erosion. This result is possibly the result of the point to tuft distance being an
unsatisfactory surrogate measure of basal cover in situations where there is an abundance of
annual grasses and forbs.
Finally the results showed that all seasons of burn caused a significant increase in
productive and palatable grass species (Decreasers) in moist savanna and a marked decrease in
these species in arid savanna. Interestingly frequency of burning and grazing caused a linear
increase in productive and palatable grass species (Decreasers) with increasing frequencies in
moist savanna but a uniform overall decrease in arid savannas. Again these results clearly illustrate
that moist savannas are well adapted to fire and arid savannas sensitive to fire particularly with
frequent burning and grazing.

The challenge arising from all these different effects of burning and grazing with domestic
livestock and wild ungulate species in moist and arid rangelands is to develop a set of
guidelines that can be used to formulate and implement a prescribed burning program.
Practical field experience with the development and implementation of prescribed burning
programs in southern and east African grasslands and savannas has resulted in the
development of the following criteria that can be used to objectively decide whether
rangeland needs to be burnt or not when grazed by Prescribed burning should not be applied
if the grass sward is in a pioneer condition dominated by Increaser II grass species caused
by overgrazing. Burning is generally not recommended when rangeland is in this condition in
order to enable it to develop to a more productive stage dominated by Decreaser grass
species;

Conversely when the grass sward is in an under grazed condition dominated by Increaser I
species, it needs to be burnt to increase the better fire adapted and more productive
Decreaser grass species;

Prescribed burning is necessary when the grass sward has become overgrown and moribund
as a result of excessive self-shading arising from undergrazing and above average rainfall
conditions and this condition develops when the standing crop of grass is generally >4 000
kg/ha;

Prescribed burning must be applied when the grass sward is dormant in order to avoid any
detrimental effects on the regrowth and basal cover of the sward and the burning window
can extend over the entire dry season;

Finally limits must be set to the area burnt as a precaution against over burning to prevent
inadequate supplies of forage being available for herbivorous wildlife. Field experience
indicates that not more than 50% of the rangeland be burnt in moist grassland and savanna
ecosystems (>700 mm p.a.) and not more than 33% in arid (<500 mm p.a.) grassland and
savanna ecosystems.
either domestic livestock or wild ungulate species:
D3.5-1-40-1000
Page 71 of 78
The great advantages of using the aforementioned criteria is that they are equally
applicable and operational in moist and arid African grasslands and savannas and are able to
regulate the frequency of burning in accordance with the grazing pressure of both domestic
livestock and wild ungulate grazers. The recommended season of burn will also ensure that the
grass sward is maintained in a vigorous and productive condition.
In conclusion the effects of the withdrawal of controlled burning in the Ngorongoro Crater
in Tanzania clearly indicate the importance of applying animal ratios in management programs that
are compatible with the condition of the grass sward and the grazing preference of the available
wild ungulate grazing species in an area.
D3.5-1-40-1000
Page 72 of 78
5.
References
Acocks, J.P.H, 1988. Veld types of South Africa. Mem. Bot. Surv. S. Afr. No. 57. Botanical Research
Institute, Dept. Agric & Water Supply, South Africa.:1-146.
Anonymous, 1992. ESCOM Annual Report. Electricity Supply Commission. South Africa: 1-25.
Bond, W.J. & van Wilgen, B.W., 1996. Fire and plants. Chapman & Hall, London. P:1-263. ISBN 0
8050 14721.
Boultwood, J.N. & Rodel, M.G.W., 1981. Effects of stocking rate and burning frequency on
Brachystegia/Julbernadia veld in Zimbabwe. Proc. Grassl. Soc Sth. Afr. 16: 111-115.
Braack, L.E.O., 1997. An objectives hierarchy for the management of the Kruger National Park. Vol.
VII Management Plan, Kruger National Park. Skukuza.
Bransby, D.I. & Tainton, N.M. 1977. The disc pasture meter : possible applications in grazing
management. Proc. Grassl. Soc. Sth. Afr. 12 : 115-118.
Brown, A.A. & Davis, K.P., 1973. Forest fire: control and use. McGraw Hill Book Co., New York.
Brynard, A.M., 1971. Controlled burning in the Kruger National Park - history and development.
Proc. Annual Tall Timbers Fire Ecology Conference. 11: 219-231.
Byram, G.M., 1959. Combustion of forest fuels. In: Forest fire: control and use. Ed. K.P. Davis.
McGraw Hill Book Co., New York.
Cahoon, D.R., Stocks, B.J., Levine, J.S., Cofer, W.R. & O’Neill, Katherine, P., 1992. Seasonal
distribution of African Savanna fires. Nature, 359: 812-815.
Carson W.P. & Abbiw, D.K., 1990. The vegetation of a protection site on the northern Accra Plains,
Ghana. Afri. J. Ecol. 28: 143-146.
Dillon, R.F., 1980. Some effects of fire in the Tall Grassveld of Natal. M.Sc. (Agric.) Thesis, Univ.
Natal. Pietermaritzburg.
Edwards, D.C. & Bogdan, A.V., 1951. Important grassland plants of Kenya. Sir Isaac Pitman &
Sons, Ltd., London: 1-124.
Estes, R.D., Atwood & Estes, A.B., 2006. Downward trends in Ngorongoro Crater ungulate
populations 1986–2005: Conservation concerns and the need for ecological research. Biological
Conservation, 131: 106-120.
Everson, T.M., Everson, C.S., Dicks, H.M. & Poulter, A.G., 1988. Curing rates in the grass sward of the
Highland Sourveld at Cathederal Peak in the Natal Drakensberg. S Afr. Forestry J. 145: 1-8.
Forbes, R.G. & Trollope, W.S.W., 1991. Veld management in the rural areas of Ciskei. J.Grassl. Soc.
South Afr. 8, 4: 147-152.
Foran, B.D., Tainton, N.M. & Booysen, P.de V., 1978. The development of a method for assessing
veld condition in three grassveld types in Natal. Proc. Grassl. Soc. Sth. Afr. 13: 27-33.
D3.5-1-40-1000
Page 73 of 78
Fynn, R.W.S., Morris, C.D. & Edwards, T.J., 2004. Effect of burning and mowing on grass and forb
diversity in a long-term grassland experiment. Applied Vegetation Science 7:1-10.
Gertenbach, W.P.D., 1983. Landscapes of the Kruger National Park. Koedoe 26:9-21.
Gill, A.M., 1981. Fire Ecology in South Africa. S. Afr. J. Sci. 77: 62-64.
Goldammer, J.G. & Crutzen, P.J., 1993. Fire in the environment: Scientific rationale and summary
of results of the Dahlem Workshop. In: Crutzen, P.J. & Goldammer, J.G. (eds). Fire in the
environment: The ecological, atmospheric and climatic importance of vegetation fires. John Wiley &
Sons, Toronto: 1-14.
Govender, N., Trollope, W.S.W. & van Wilgen, B.W., 2006. The effect of fire season, fire frequency,
rainfall and management on fire intensity in savanna vegetation in South Africa. Journal Applied
Ecology, 43: 748-758.
Hanby, J. & Bygott, D., 1992. Nogorongoro Conservation Area - Guide Book. Tanzania Printers
Limited, Tanzania.
Hardy, M.B. & Tainton, N.M., 1993. Towards a technique for determining basal cover in tufted
grasslands. African Journal of Range & Forage Science, 10(2): pp. 77-81
Hardy, M.B., Barnes, D.L., Moore, A. & Kirkman, K.P., 1999. The management of different types of
veld - sour grassveld. In: Tainton, N.M. (ed). Veld management in South Africa. University Natal
Press, Pietermaritzburg, South Africa: 280-288.
Haschke, S.W. & Kirkman, K.P., 1994. The importance of long term rests in sourveld management
with particular reference to autumn burning practices. Proc. South African Society, Agricultural
Extension, South Africa: 103-108.
Higgins, S.I., Bond, W.J., February, E.C., Bronn, A., Euston-Brown, D.I.W., Enslin, B., Govender, N.,
Rademan, L., O’Regan, S., Potgieter, A.L.F., Scheiter, S., Sowry, R., Trollope, L.A. and Trollope,
W.S.W., 2007. Effects of four decades of fire manipulation on woody vegetation structure in savanna.
Ecology, 88, 5: 1119-1125.
Kennan, T.C.D., 1971. The effects of fire on two vegetation types of Matopos. Proc. Tall Timbers
Fire Ecology Conf. 11: 53-98.
Kirkman, K.P., 2002. The influence of various types and frequencies of rest on the production and
condition of sourveld grazed by sheep or cattle. 1. Proportional species composition. African
Journal Range & Forage Science 19: 55-62.
Komarek, E.V.,1965. Fire ecology - grasslands and man. Proc. Tall Timbers Fire Ecology Conf. 4:
169-220.
Komarek, E.V., 1971. Lightning and fire ecology in Africa. Proc. Tall Timbers Fire Ecology Conf. 11:
473-511.
McArthur, A.G., 1966. Weather and grassland fire behaviour. Commonw. Aust. For. Timb. Bur.
Leafl. 100.
D3.5-1-40-1000
Page 74 of 78
Menaut, J.C., 1977. Evolution of plots protected from fire since 13 years in a Guinea savanna of Ivory
Coast. Actas Del IV Simposium Internacional De Ecologia Tropical Tomo II, Universidad De Panama,
Panama: 543-558.
Mentis, M.T. 1981. The animal as a factor in pasture and veld management pp. 289 - 311. In :
Tainton, N.M. (ed). Veld And Pasture Management In South Africa. Shuter & Shooter,
Pietermaritzburg.
Moe, S. R.,Wegge, P. & Kapela, E.B., 1990. The influence of man-made fires on large wild
herbivores in Lake Burungi area in northern Tanzania. Afr. J. Ecol. 28: 35-43.
Munthali, S.M. & Banda, H.M., 1992. Distribution and abundance of the common ungulates of
Nyika National Park, Malawi. Afr. J. Ecol. 30: 203-212.
Phillips, J.F.V., 1965. Fire - as a master and servant: its influence in the bioclimatic regions of Trans
- Saharan Africa. Proc. Tall Timbers Fire Ecology Conf. 4: 7-10.
Plowes, D.C.H., 1957. The seasonal variation of crude protein in twenty grasses at Matopos,
Southern Rhodesia, and related observations. Rhod. Agric. J. 54: 33-55.
Pratt, D.J. & Gwynne, M.D., 1977. Rangeland Management & Ecology in East Africa. Hodder &
Stoughton, London.
Pyne, S.J., 1995. World Fire: the Culture of Fire on Earth. Henry Holt & Company, New York.
Robinson, E.T., Gibbs-Russel, G.E., Trollope, W.S.W. & Downing, B.H., 1979. Assessment of short
term burning treatments on the herb layer of False Thornveld of the Eastern Cape. Proc. Grassld.
Soc. Sth. Afr. 14: 79-83.
Rose-Innes, R., 1971. Fire in west African vegetation. Proc. Tall Timbers Fire Ecology Conf. 11:
147-173.
Rothermel, R.C., 1972. A mathematical model for predicting fire speed in wildland fuels. U.S.D.A.
For. Serv. Res. Paper I.N.T. 115.
Runyuro, V., 1995. Annual Report, Ngorongoro Conservation Area Authority. Ngorongoro, Tanzania
Scott, J.D., 1971. Veld burning in Natal. Proc. Tall Timbers Fire Ecology Conf. 11: 33-51.
Snyman, H.A., 1999. Soil erosion and conservation . In : Tainton, N.M. (ed). Veld And Pasture
Management In South Africa. Shuter & Shooter, Pietermaritzburg: 355-380.
Sweet, R.J., 1982. Bush control with fire in Acacia nigrescens/ Combretum piculatum savanna in
Botswana. Proc. Grassl. Soc. Sth. Afr. 17: 25-28.
Tainton, N.M., Groves, R.H. & Nash, R., 1977. Time of mowing and burning veld: short term effects
on production and tiller development. Proc. Grassl. Soc. Sth. Afr. 12: 59-64.
Tainton, N.M., 1999. The ecology of the main grazing lands of South Africa. In: Tainton, N.M. (ed).
Veld management in South Africa. University Natal Press, Pietermaritzburg, South Africa: 23-53.
D3.5-1-40-1000
Page 75 of 78
Tainton, N.M. & Mentis M.T., 1984. Fire in grassland. In: Booysen, P.de V. & Tainton, N.M. (eds).
Ecological effects of fire in South African ecosystems. Ecological Studies No 48: 115-197.
Trollope, W.S.W., 1978. Fire behaviour - a preliminary study. Proc. Grassl. Soc. Sth. Afr. 13: 123-128.
Trollope, W.S.W., 1983. Control of bush encroachment with fire in the arid savannas of south eastern
Africa. PhD thesis, University Natal, Pietermaritzburg.
Trollope, W.S.W., 1984. Fire in savanna. In: Booysen, P.de V. & Tainton, N.M.
effects of fire in South African ecosystems. Ecological Studies No 48: 149-175.
(eds).
Ecological
Trollope, W.S.W., 1987. Effect of season of burning on grass recovery in the False Thornveld of the
Eastern Cape. J. Grassl. Soc. South. Afr. 4, 2:74-77.
Trollope, W. S. W., 1989. Veld burning as a management practice in livestock production. In: Veld
Management in the Eastern Cape. (eds) J E Danckwerts & W R Teague. Govt. Printer, Pretoria: 6773.
Trollope, W S W., 1990a. Development of a technique for assessing veld condition in the Kruger
National Park. J. Grassld. Soc. South. Afr.7, 1:46-51.
Trollope, W S W., 1990b. Veld management with specific reference to game ranching in the grassland
and savanna areas of South Africa. Koedoe 33, 2:77-86.
Trollope, W.S.W., 1993. Fire regime of the Kruger National Park for the period 1980 – 1992. Koedoe,
36, 2: 45-52.
Trollope, W S W and Potgieter, A L F., 1986. Estimating grass fuel loads with a disc pasture meter in
the Kruger National Park. J. Grassld. Soc. Sth Afr. 3, 4: 148-152.
Trollope, W S W, Trollope, L.A. & Bosch, O J H., 1990. Veld and pasture management terminology in
southern Africa. J. Grassld. Soc. South Afr. 7,1:52-61.
Trollope, W.S.W. & Trollope, L.A., 1995.
Report On The Fire Ecology Of The Ngorongoro
Conservation Area In Tanzania With Particular Reference To The Ngorongoro Crater. Department
Livestock & Pasture Science, University Fort Hare, Alice:1 - 21.
Trollope, W.S.W., Potgieter, A.L.F., Biggs, H.C. & Trollope, L.A., 1998. Report on the long term
burning trial determining the effects of season and frequency of burning in the major vegetation types
of the Kruger National Park. Report submitted to Scientific Services, Kruger National Park, South
African National Parks, Skukuza: 1- 40.
Trollope, W.S.W., 1999. Veld burning: Savanna. In: Tainton, N.M. (ed). Veld management in South
Africa. University Natal Press, Pietermaritzburg, South Africa: 236-242.
Trollope, W.S.W. & Tainton, N.M., 1986. Effect of fire intensity on the grass and bush components of
the Eastern Cape Thornveld. J. Grassl. Soc. Sth. Afr. 2: 27-42.
Trollope, W.S.W., Potgieter A.L.F. & Zambatis, N., 1989. Assessing veld condition in the Kruger
National Park using key grass species. Koedoe 32, 1: 67-94.
D3.5-1-40-1000
Page 76 of 78
Trollope, W.S.W. & Trollope, L.A., 1999. Report on the assessment of range condition and the fire
ecology of the savanna vegetation on the Lewa Wildlife Conservancy in Kenya – 1998. Final Report.
Department Livestock & Pasture Science, Faculty of Agriculture, University Fort Hare, Alice: 1-54.
Trollope, W.S.W, Hines , C.J.H., & Trollope, L.A., 2000. Simplified Techniques for Assessing Range
Condition in the East Caprivi Region of Namibia. Final Report of the Namibia-Finland Forestry
Programme in the east Caprivi Region of Namibia. Department Livestock & Pasture Science, Faculty
of Agriculture, University Fort Hare, Alice: 1-67.
Trollope, W.S.W. & Trollope, L.A., 2001a. Relationship between range condition and the incidence
of ticks in the Ngorongoro Crater in Tanzania. Final Report. Department Livestock & Pasture
Science, University Fort Hare, Alice, South Africa: 1-16.
Trollope, W.S.W. & Trollope, L.A., 2001b. Fire management program for the National Reserve of
Gile in Zambesia Province. Final Report. Department Livestock & Pasture Science, University Fort
Hare, Alice, South Africa: 1-24.
Trollope, W.S.W., Fyumagwa, R. & Trollope, L.A., 2003.Relationship between range condition and
the incidence of ticks in the Ngorongoro Crate
r, Tanzania.Proceedings VII International Rangeland Conference, Durban, South Africa.
Trollope, W.S.W., Trollope, L.A., Austin, C. de Bruno, Held, A.C., Emery, A. & Hines, C.J.H., 2006. A
fire management plan for the Okavango Delta Ramsar Site in Botswana. Department Forestry &
Range Resources, Botswana Government, Gabarone, Botswana: 1-207.
van der Schijff, H.P., 1957. Ekologiese studie van die flora van die Nasionale Krugerwildtuin. D.Sc
Tesis, Universiteit Potchefstroom.
van Wyk, P., 1971. Veld burning in the Kruger National Park - An interim report of some aspects of
research. Proc. Tall Timbers Fire Ecology Conf., 11: 9-31.
Vetter, S., 2007. What are the costs of environmental degradation to South African communal
livestock farmers: the case of the Herschel district in the Eastern Cape Province. PhD Thesis,
University Cape Town, South Africa.
West, O., 1965. Fire in vegetation and its use in pasture management with special reference to
tropical and subtropical Africa. Mem. Pub. Commomw. Agric. Bur., Farnham Royal, Bucks., England.
No. 1.
Zacharias, P.J.K., 1994. The fire/ grazing interaction on Dohne Sourveld. PhD thesis. University Natal,
Pietermaritzburg.
D3.5-1-40-1000
Page 77 of 78
OBJECTIVE 4: Quantify the fire-tree, shrub and grass species relationships in these
vegetation types.
This objective has been achieved by reviewing the effects of fire in African grasslands and
savannas where the effects of type and intensity of fire and season and frequency of burning have
been specifically related to their effects on the grass sward and trees and shrubs in these plant
communities.
OBJECTIVE 5: Check similarities between the southern and northern hemispheres in
the above matters.
This objective has not been addressed because firstly it has not been specified for what type of
land use the comparison should be based on. Secondly there is very limited information available
on the specific effects of the different components of the fire regime on the grass sward and tree
and shrub vegetation in the northern hemisphere that could be meaningfully compared with the
interacting effects of the fire regime and herbivory involving the unique and diverse spectrum of
wildlife occurring in African grasslands and savannas. Finally it is believed that this objective is too
far removed from the central objective focussing specifically on the relationships between fire and
grazing in African grasslands and savannas.
D3.5-1-40-1000
Page 78 of 78