Maintaining plant diversity of a species-rich
montane grassland system in the face of global

change

by
Paul Jan Gordijn

(1511172)

Thesis
Submitted in fulfilment of the requirements for the degree

Doctor of Philosophy
in

Ecological Sciences
in the Faculty of Science, University of the Witwatersrand, Johannesburg, South Africa

Supervisor: Prof. Timothy G. O’Connor
Co-supervisor: Francesca Parrini

February, 2022



Abstract

The rich grassland-plant diversity of the Drakensberg mountain region, which has persisted

under heterogeneous fire and herbivory regimes, and significant fluctuations in climate, has

come to face its most formidable threat—land transformation, which involves habitat

destruction and land use intensification. Over the last ca. 200 years, human settlement in the

northern Drakensberg study area, the Mdedelelo-Cathkin landscape in South Africa, has

markedly increased, but the influence of human pressure on plant diversity has been largely

unexplored, leaving an uncertain future for this biodiversity. This thesis aimed to reduce

uncertainty around the influence of socio-ecological systems, represented by communal, private

and protected land tenure systems, on grassland habitat and associated plant diversity. Analysis

of changes in land use and land cover and field-based sampling of plant diversity revealed a

conservation conundrum. Although grassland-plant diversity was maintained at higher levels

on private systems, these systems were especially vulnerable to transformation. And while

communal systems were less vulnerable to transformation, they were associated with

heavy-continuous grazing that transformed grasslands into a novel state depauperate of plant

diversity. Protected systems were, however, largely successful in maintaining primary

grasslands, and their plant diversity. On these systems, where fire is the principal disturbance

agent, the previously unexplored, multidecadal influence of heterogeneous fire regimes, was

explored. A novel characterisation of heterogeneity in fire-return intervals and season of

burning identified a threat of increasing fire regime homogeneity. The influence of

socio-ecological systems, and their particular fire and herbivory disturbance regimes were

framed by a construct termed, “socio-ecological disturbance regimes” (S-EDRs), which

successfully reflected the interconnected nature of human society, disturbance regimes, and

plant diversity. The transformation of a quarter of the landscape’s primary grassland over the

71-year period of assessment should draw urgent attention from conservationists and society.

The S-EDR framework highlighted the critical responsibility society has in maintaining plant

diversity. For the conservation of grassland-plant diversity a cross-societal approach is

necessary, valuable plant diversity was found across the socio-ecological systems evaluated.

Moreover, protecting Drakensberg-grassland ecosystems and their plant diversity will

contribute to the global effort to mitigate the looming influence of climate change.

i



Dedication

To my wife, two daughters, and the mountain breeze ( !ח²וּר ).

ii



Acknowledgements

The privilege of this work would not have been possible without the support of my dear wife,

Nicky, and her enthusiasm for the Drakensberg mountains. Also, the encouragement from my

family, of whom, my parents instilled in me an appreciation for these mountains. The

incredible support from my SAEON colleagues, Kent, Siphiwe, Busii, Michele, and especially

Sue, is much appreciated. Field-sampling adventures with Thami Shezi, Sinethemba

Ntshangase, and Andrew Steele, were just that, an adventure, thank you. I am also grateful for

the support of my EKZNW colleagues, particularly, Debbie Jewitt, Sonja Kruger, Mark

Robertson, Ian Rushworth, Eleste Hadebe, and Clint Carbutt. I am indebted to those who

explored the plant diversity of Drakensberg long before me, in particular, Donald Killick, who

left the legacy of the D.H.K. herbarium. Moreover, Ed Granger, for his establishment of

vegetation plots in the study area. Terry and Colin Everson as well, played an instrumental role

in this area and were open-handed with both their knowledge and coffee. The generosity of

Tim, my chief supervisor, with his time, calculated instruction, and passion for the

Drakensberg, has made this journey of learning especially valuable—thank you. And, thank

you, Francesca, for comments on drafts and ready assistance with administration.

iii



Declaration

I declare that this thesis is my own, unaided work. It is being submitted for the Degree of

Doctor of Philosophy at the University of the Witwatersrand, Johannesburg. It has not been

submitted before for any degree or examination at any other University.

Candidate signature:

Paul J. Gordijn

on the 16th day of February 2022, at Mountain Homes, Hilton.

iv



Contents

1 Introduction 1
1.1 Aims and objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.2 Thesis structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 The ‘eco-story’ of a mountain range: acquiring perspective on Drakensberg plant
diversity and global change 11
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.2 Study area: abiotic template . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.2.1 Geology, soils, and topography . . . . . . . . . . . . . . . . . . . . . . 12

2.2.2 Climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.3 Drakensberg plant diversity: origins, drivers, and disturbance regimes . . . . . 17

2.3.1 Climate and fire: palaeoecological drivers in grassland . . . . . . . . . 18

2.3.2 High-altitude alpine grassland diversity: Cape fynbos elements . . . . . 19

2.3.3 Mid-altitude temperate grasslands with subtropical grassland elements . 20

2.3.4 Low-altitude temperate grasslands with tropical savanna elements . . . 21

2.3.5 Indigenous, low-pressure herbivory regimes . . . . . . . . . . . . . . . 22

2.4 Novel disturbance regimes: intensifying human settlement, from hunter-gatherers

to farmers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.4.1 Hunter-gatherers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.4.2 Farmers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

2.4.3 Establishment of the Cathedral Peak research catchments . . . . . . . . 26

2.4.4 Protected areas in the Drakensberg . . . . . . . . . . . . . . . . . . . . 27

2.4.5 Early colonial impressions of fire . . . . . . . . . . . . . . . . . . . . 27

2.5 Unpacking environmental heterogeneity, fire, and herbivory . . . . . . . . . . . 28

2.5.1 Fire and the grass sward . . . . . . . . . . . . . . . . . . . . . . . . . 29

2.5.2 Fire behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

2.5.3 Herbivory and the grass sward . . . . . . . . . . . . . . . . . . . . . . 30

2.5.4 Forbs, and other plant life forms, in the grass sward . . . . . . . . . . . 31

2.6 Unpacking the influence of socio-ecological systems on grassland-plant diversity 33

2.6.1 Grassland use and land tenure . . . . . . . . . . . . . . . . . . . . . . 34

2.6.2 Grassland transformation and land tenure . . . . . . . . . . . . . . . . 36

2.7 Climate change, atmospheric carbon dioxide enrichment, and grassland-plant

diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

v



Drakensberg plant diversity in the face of global change

2.7.1 Changes in temperature and precipitation . . . . . . . . . . . . . . . . 40

2.7.2 Atmospheric carbon dioxide enrichment . . . . . . . . . . . . . . . . . 41

2.8 Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

3 Socio-ecological legacy effects on grassland transformation in the biodiversity rich
Drakensberg mountains 58
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

3.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

3.2.1 Study area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

3.2.2 Land cover classification . . . . . . . . . . . . . . . . . . . . . . . . . 65

3.2.3 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

3.3.1 Land cover transitions and stationarity . . . . . . . . . . . . . . . . . . 69

3.3.2 Socio-ecological influences on grassland transformation . . . . . . . . 72

3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

3.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

4 Functional perhaps, but depauperate—the fate of Drakensberg grassland plant di-
versity under novel socio-ecological disturbance regimes 86
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

4.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

4.2.1 Study area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

4.2.2 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

4.2.3 Response of plant species and functional diversity to S-EDRs . . . . . 96

4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

4.3.1 The marginal influence of S-EDRs on plant diversity . . . . . . . . . . 97

4.3.2 Response of plant composition to S-EDRs over time . . . . . . . . . . 97

4.3.3 Response of plant species and functional diversity to S-EDRs . . . . . 101

4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

4.4.1 Emulating natural disturbance regimes . . . . . . . . . . . . . . . . . . 105

4.4.2 Novel disturbance regimes . . . . . . . . . . . . . . . . . . . . . . . . 106

4.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

5 Resistance of Drakensberg grasslands to compositional change depends on the in-
fluence of fire-return interval and grassland structure on richness and spatial turnover117
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

5.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

vi



Drakensberg plant diversity in the face of global change

5.2.1 Study site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

5.2.2 Vegetation sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

5.2.3 Quantification of diversity . . . . . . . . . . . . . . . . . . . . . . . . 121

5.2.4 Field sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

5.2.5 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

5.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

5.3.1 Fire-return interval; environmental variable correlation and structure . . 125

5.3.2 Environmental plus fire-return interval influences on grassland structure

and diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

5.3.3 Composition change over time . . . . . . . . . . . . . . . . . . . . . . 130

5.3.4 Compositional—environmental gradient relations over time . . . . . . 130

5.3.5 Species responses to the fire-return interval . . . . . . . . . . . . . . . 131

5.3.6 Directional change in species abundance . . . . . . . . . . . . . . . . 132

5.3.7 Compositional stability in relation to fire regime . . . . . . . . . . . . 132

5.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

5.4.1 Compositional change and the fire-return interval . . . . . . . . . . . . 132

5.4.2 Diversity and the fire-return interval . . . . . . . . . . . . . . . . . . . 134

5.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

6 Multidecadal effects of fire in a grassland biodiversity hotspot: Does pyrodiversity
enhance plant diversity? 140
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

6.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

6.2.1 Study area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

6.2.2 Experimental design . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

6.2.3 Field sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

6.2.4 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

6.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

6.3.1 Composition and structure . . . . . . . . . . . . . . . . . . . . . . . . 150

6.3.2 Plant species diversity . . . . . . . . . . . . . . . . . . . . . . . . . . 150

6.3.3 Partitioning beta diversity . . . . . . . . . . . . . . . . . . . . . . . . 151

6.3.4 Plant groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

6.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

6.4.1 Contribution of pyrodiversity . . . . . . . . . . . . . . . . . . . . . . . 157

6.4.2 Fire return-interval and species diversity . . . . . . . . . . . . . . . . . 158

6.4.3 Burn season and species diversity . . . . . . . . . . . . . . . . . . . . 162

6.4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

vii



Drakensberg plant diversity in the face of global change

7 General discussion 170
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

7.2 Principal limitations of the study design . . . . . . . . . . . . . . . . . . . . . 173

7.3 Socio-ecological disturbance regimes . . . . . . . . . . . . . . . . . . . . . . 175

7.3.1 Land tenure as a component of S-EDRs . . . . . . . . . . . . . . . . . 176

7.3.2 Fire and herbivore disturbance in Drakensberg grassland . . . . . . . . 178

7.4 Climate change and future threats . . . . . . . . . . . . . . . . . . . . . . . . 184

7.4.1 Woody plant encroachment . . . . . . . . . . . . . . . . . . . . . . . . 186

7.4.2 Other future threats . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186

7.5 Guidance for conservation practice . . . . . . . . . . . . . . . . . . . . . . . . 187

7.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190

A Study area: abiotic template 200
A.1 Elevation, geology, slope, and soil moisture . . . . . . . . . . . . . . . . . . . 200

B Land cover classification and analyses 203
B.1 Orthophoto production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

B.2 Land cover classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204

B.3 Wetland delineation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

C Digital Terrain Modelling 216
C.1 Model preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216

D Hierarchical Generalized Additive Models 220
D.1 More on hGAM model structure setup and selection . . . . . . . . . . . . . . . 220

D.2 More on hGAM model results . . . . . . . . . . . . . . . . . . . . . . . . . . 221

E Mdedelelo-Cathkin vegetation survey supplementary information 224
E.1 Correlation structure of soil properties and terrain variables . . . . . . . . . . . 224

E.2 Plant Family contributions to functional groups . . . . . . . . . . . . . . . . . 226

F Brotherton burning trial and plant diversity supplementary material 229
F.1 Brotherton plant diversity: alpha and beta diversity . . . . . . . . . . . . . . . 229

F.2 Brotherton burning trial layout . . . . . . . . . . . . . . . . . . . . . . . . . . 230

viii



List of Figures

2.1 Regional map of the Mdedelelo-Cathkin landscape . . . . . . . . . . . . . . . 13

2.2 3D visualisation of the Mdedelelo-Cathkin landscape . . . . . . . . . . . . . . 14

2.3 Geological formations of the Mdedelelo-Cathkin landscape . . . . . . . . . . . 15

3.1 Land tenure systems in the Mdedelelo-Cathkin landscape . . . . . . . . . . . . 63

3.2 Land cover composition across tenure systems in 1945 and 2016 . . . . . . . . 70

3.3 Cross-tabulation pie matrix: summary of land cover transitions and stationarity

from 1945 to 2016 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

3.4 hGAM smooths showing the influence of key predictors on the vulnerability of

primary grassland to transformation on communal and private tenure systems,

in 1945 and 2016 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

3.5 hGAM predictions of the likelihood of transformation on communal and private

systems, across a 2D topographic profile, in 1945 and 2016 . . . . . . . . . . . 74

3.6 3D visualizations of hGAM predictions: the likelihood of primary grassland

transformation across communal and private systems, in 1945 and 2016 . . . . 75

4.1 Land tenure systems in the Mdedelelo-Cathkin landscape . . . . . . . . . . . . 91

4.2 Venn diagram: partitioning variance in beta diversity across S-EDR components 98

4.3 The magnitude of changes in plant composition across S-EDRs . . . . . . . . . 99

4.4 Ordination depicting changes in plant composition across S-EDRs, environmen-

tal variables, and plant-functional groups, from 1980 to 2016 . . . . . . . . . . 100

4.5 The response of plant species richness and community organisation (Shannon’s

diversity and Simpson’s dominance) to S-EDRs . . . . . . . . . . . . . . . . . 102

4.6 The response of plant-functional group richness and diversity to S-EDRs . . . . 103

4.7 The response of plant-functional group richness to S-EDRs . . . . . . . . . . . 104

5.1 Vegetation monitoring plots in the Cathedral Peak research catchments . . . . . 121

5.2 PCA showing the correlation structure of environmental variables in the Cathe-

dral Peak research catchments . . . . . . . . . . . . . . . . . . . . . . . . . . 126

5.3 Grassland structure, fire-return interval and species richness relations . . . . . . 127

5.4 The response of grass diversity to the fire-return interval and soil texture . . . . 127

5.5 NMDS ordination illustrating the influence of environmental variables and fire-

return interval on grassland-plant composition . . . . . . . . . . . . . . . . . . 130

5.6 The influence of fire-return interval on the abundance, or likelihood of occur-

rence, on grass and forb species . . . . . . . . . . . . . . . . . . . . . . . . . 131

ix



Drakensberg plant diversity in the face of global change

5.7 Predictors of compositional stability between 1986 and 2014 . . . . . . . . . . 133

6.1 NMDS ordination showing variation in composition associated with fire-return

interval treatments and grassland structure (basal cover) . . . . . . . . . . . . . 151

6.2 Forb and graminoid species accumulation curves, for different fire-return inter-

vals and burn season . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

6.3 The response of alpha diversity to fire-return interval and burn season treatments 153

6.4 The response of plant-functional groups to fire-return interval treatments . . . . 154

6.5 The response of plant-functional groups to burn season treatments . . . . . . . 155

6.6 The response of graminoid and forb species richness and turnover to pyrodiversity156

A.1 Study area: abiotic template and land tenure systems . . . . . . . . . . . . . . 200

A.2 Study area: the distribution of geological formations by elevation . . . . . . . . 201

A.3 Study area: the correlation structure of abiotic environment variables and tenure

systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

B.1 Land cover classification no data areas within the Mdedelelo-Cathkin . . . . . 214

C.1 3D visualisations depicting the processing of the DSM to DTM models . . . . 218

D.1 hGAM 2D topographic cross-section (without settlement density) showing asso-

ciated predictions of the likelihood of primary grassland transformation, across

communal and private tenure systems, in 1945 and 2016 . . . . . . . . . . . . 222

E.1 Correlation structure of soil textural and chemical properties, and terrain vari-

ables, as represented by PCA biplots . . . . . . . . . . . . . . . . . . . . . . . 225

F.1 Layout of plot treatments at the Brotherton burning trial . . . . . . . . . . . . . 230

x



List of Tables

2.1 Vegetation type summarised by altitude . . . . . . . . . . . . . . . . . . . . . 12

2.2 Altitudinal variation in temperature regimes over the Mdedelelo-Cathkin landscape 17

3.1 Land cover categories used to evaluate land cover land use change. Non-natural

or human land covers that could support grassland were considered to be trans-

formed in the earliest time period assessed (1944/1945). Indigenous closed

woody, erosion, ‘wetland’, and natural water were natural land covers that were

considered to not support grassland in 1944/1945; thereafter, any departure from

Primary grassland constituted grassland transformation . . . . . . . . . . . . . 66

4.1 Characterisation of S-EDRs by herbivory pressure and fire-return intervals, in

1980 and 2016 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

5.1 Characterisation of fire-return intervals in 1980, 1986, and 2014 . . . . . . . . 123

5.2 The influence of fire-return interval, environmental variables, and spatial hetero-

geneity on grass and forb turnover, from 1980 to 2014 . . . . . . . . . . . . . . 129

5.3 Magnitude and significance of directional changes in grass species abundance . 132

6.1 Significance of predictors for variance in forb and graminoid turnover . . . . . 153

B.1 Air photo set details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

B.2 Details of land cover categories used for classification . . . . . . . . . . . . . . 205

D.1 AIC and adjusted R-squared values used to inform Hierarchical Generalized

Additive Model selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221

E.1 Summary of plant Family composition per functional type . . . . . . . . . . . 227

F.1 Summary of alpha and beta diversity patterns across fire treatments . . . . . . . 229

xi



Drakensberg plant diversity in the face of global change

List of abbreviations

AIC Akaike information criterion

BP (Number of years) Before Present

CI Frequentest Confidence Interval or Bayesian Credible Interval

depending on the mode of statistical analysis

D Simpson’s Dominance

dbRDA distance based redundancy analysis

DSM Digital Surface Model

DT Distance-to-tuft (component of grassland structure)

DTM Digital Terrain Model

GSD Ground Sampling Distance

H Shannon’s diversity

hGAM Hierarchical Generalized Additive Model

IDH Intermediate Disturbance Hypothesis

LULCC Land Use Land Cover Change

ka A thousand years ago

m a.s.l. Meters above sea level (altitude)

Ma A million years ago

MEM Moran’s Eigenvector Maps

PCA Principal Components Analysis

PISR (Topographically modified) Potential Incoming Solar Radiation

RI Fire-return Interval

S Species richness

S-EDR Socio-ecological disturbance regime

SWI Soil Wetness Index

VIF Variance Inflation Factor

xii



Chapter 1

Introduction

The rich-plant diversity of ancient grasslands is highly threatened under global change

(Hoekstra et al. 2005, Bond and Parr 2010, Veldman et al. 2015), that is, the increasing pressure

of humans on the earth’s resources and its natural ecosystems. Covering a little over a third of

the earth’s surface (Suttie et al. 2005), grasslands have been well-recognized for production of

rangeland animals, water yield, nutrient cycling, and climate regulation (Boval and Dixon

2012, Bengtsson et al. 2019). However, their valuable contribution to plant diversity has only

been recognized by mainstream science within the last two to three decades (Bond 2019)—and

perhaps, too late (Hoekstra et al. 2005, Bond and Parr 2010, Carbutt et al. 2017).

Climate change is a looming threat to grassland biodiversity (Polley et al. 2017, Pörtner

et al. 2021), but habitat loss has been the largest driver of terrestrial biodiversity loss, driving

the earth’s sixth mass extinction (Ceballos et al. 2015, Sage 2020, Pörtner et al. 2021).

Grassland, being the primary biome into which humans have expanded their agricultural

activities, is particularly vulnerable to habitat destruction (White et al. 2001, Hoekstra et al.

2005, Pörtner et al. 2021). Once destroyed by human land uses (e.g., tillage agriculture or

plantations) grassland-plant diversity may require over a millennium to recover (Nerlekar and

Veldman 2020).

The protection of habitat has to be prioritized in order to maintain the earth’s biodiversity

under the threat of climate change. In order for the distribution of species to adapt in response

to climate change large, intact and connected tracts of natural habitat needs to be secured

(Jewitt et al. 2017, Dinerstein et al. 2019). Moreover, conserving natural habitat is critical for

maintaining ecosystem services, and mitigating the influence of climate change (Dinerstein

et al. 2019).

The drive to expand protected areas has been critical in an effort to maintain biodiversity in

the face of global change, but the extent of protected areas remains insufficient (Sayre et al.

2020, UNEP-WCMC and IUCN 2021). Temperate grasslands, for example, form the most

extensive grassland biome, but <5% of these are protected (IUCN WCPA 2021) and at least

46% (9 million hectares) have been irreversibly transformed (Hoekstra et al. 2005). This level

of protection has fulfils less half the global target for 2020 (IUCN WCPA 2021). The

insufficient extent of protected areas obligates conservation scientists to explore alternative land

uses for main-streaming conservation efforts. Biodiversity main-streaming is designed to

encourage sustainable use of natural resources and associated biodiversity outside protected

1



Drakensberg plant diversity in the face of global change

areas, through policy and practice (Nagendra et al. 2004, Brooks et al. 2006, O’Connor et al.

2010).

Identifying land uses outside protected areas that are suitable for maintaining

grassland-plant diversity is a challenge. Understanding land transformation, that is, the

alteration of land use and land use intensification, that leads to a change in land cover, does

however, provide insight. Patterns in land transformation, revealed by land use land cover

change analyses (LULCC), are crucial for determining the drivers of land transformation over

time, and consequent vulnerability of habitat (DiGiano et al. 2013, Jewitt et al. 2015, Kleemann

et al. 2017). The vulnerability of different grassland habitats to land transformation may

contrast markedly under different land management strategies. Ostensibly, fertile-lowland

plains are more vulnerable to being transformed to a cultivated state than steeper slopes that are

unsuitable for tillage agriculture (Silveira et al. 2019). However, complex factors tied to the

manner of interaction between people groups, their social, political, and cultural dynamics, and

the environment, that is, socio-ecological dynamics (Francis and Bekera 2014), influence

patterns of transformation across the abiotic environment of a landscape (O’Connor 2005,

O’Connor and Kuyler 2009, DiGiano et al. 2013).

In remaining untransformed grassland outside protected areas, uncertainty around the fate of

grassland biodiversity is exacerbated by the influence of different socio-ecological systems, and

their associated land use and land management strategies (O’Connor 2005). Across

socio-ecological systems, humans have largely modified the disturbance regimes to which

grassland-plant diversity is adapted. This plant diversity evolved under disturbance regimes

characterised by lightning and aboriginal ignited fires, herbivory, and natural climate variability

(Pausas and Keeley 2009, Fuhlendorf et al. 2017, Polley et al. 2017, He and Lamont 2018).

Even within protected areas the effectiveness of management for securing biodiversity can be

highly uncertain (Pörtner et al. 2021). Social and cultural biases, conflicting economic

incentives (Aguilar-Tomasini et al. 2020, Maxwell et al. 2020, Rydén et al. 2020), research

gaps, a lack of monitoring to guide conservation management (Pausas et al. 2009, Gordijn et al.

2018), and poor funding or weak policy can frustrate the success of protected areas (Maxwell

et al. 2020, Pörtner et al. 2021).

Scientists urgently need to reduce uncertainty around the fate of grassland-plant diversity to

guide conservation efforts, within and outside protected areas. Novel approaches are required

to address the lack of research on grassland-plant diversity (Uys et al. 2004, O’Connor et al.

2010, Bond 2019). Grassland science has established a fair understanding of a few important

rangeland grasses valued for animal production, soil conservation and water production using

controlled trials that were undertaken on sites with limited environmental heterogeneity (Fynn

et al. 2005, O’Connor et al. 2010, Ward et al. 2020). However, the full complement of

grassland diversity including rare and uncommon species has largely been ignored (Uys et al.

2004, Bond and Parr 2010, O’Connor et al. 2010, Bond 2019). Specifically, the forb component

of grasslands that contributes 70–80% of grassland-plant diversity (Uys et al. 2004, Gordijn

et al. 2018) requires investigation (O’Connor et al. 2010, Bond 2019). Moreover, application of

the insight derived from controlled microcosm- and mesocosm-type trials is confounded by

‘real world’ heterogeneity in the abiotic environment and disturbances; for example, variation

Chapter 1 2



Drakensberg plant diversity in the face of global change

in fire behaviour or anthropogenic harvesting of plant species across a landscape is not captured

on environmentally uniform controlled trials (O’Connor 2005, Fuhlendorf et al. 2006,

Fuhlendorf et al. 2017, Polley et al. 2017). In order to provide a meaningful contribution to the

conservation of grassland-plant diversity, understanding has to account for ‘real world’

disturbances and environmental heterogeneity which propagate uncertainty across a landscape.

Mountain regions contribute over four-fifths of the earth’s terrestrial biodiversity (Rahbek

et al. 2019), but this biodiversity is particularly vulnerable under global change. Due to their

relatively higher rainfall than surrounding areas, mountain regions have become important

centres of human settlement and associated agricultural activities. Mountains provide half of

the world’s population with water (Viviroli et al. 2003). The grasslands of the Drakensberg

mountain range are the last, most-extensive primary grasslands in southern Africa (Carbutt

et al. 2011).

Conserving a heterogeneous array of remaining Drakensberg grasslands is critical. These

grasslands are known as the ‘hub’ of southern African temperate grassland diversity (Mucina

et al. 2006) containing a little over 2500 plant species with a high level of plant endemism (9%)

(Carbutt 2019). This rich-plant diversity is spread across the complex montane terrain of the

region, and the associated turnover rates, that is, changes in unique species assemblages over

space, is high. Therefore, for effective conservation, a landscape approach is necessary to

ascertain which areas are vulnerable to the pressures associated with increasing human

activities.

This biodiversity-rich montane grassland system requires urgent attention from conservation

scientists. Drakensberg grasslands have remained largely untransformed for millennia, but over

the last ca. 200 years, settlement has increased in the region (Wright and Mazel 2007), and only

a small portion of these grasslands are protected (Carbutt et al. 2011). Farmers have

transformed large tracts of these grasslands to cultivated pasture for livestock production,

tillage agriculture, and plantation forestry (O’Connor and Kuyler 2009). The estimated extent

of irreversibly transformed grasslands of 22.30% (Carbutt et al. 2011) is likely an

underestimate, due to the known constraints with detecting secondary grassland using aerial

imagery (Jewitt et al. 2015). Moreover, the 5.87% level of Drakensberg grassland protection

(Carbutt et al. 2011) is meagre considering the multiple sub-centres of endemism in the region

(Mucina et al. 2006, Carbutt 2019).

Lowland Drakensberg grasslands that have experienced the greatest levels of transformation

are largely unprotected (O’Connor and Kuyler 2009, Carbutt et al. 2011), and differences

between socio-ecological systems could exacerbate uncertainty around their vulnerability to

transformation (Kleemann et al. 2017). Similar to elsewhere in South Africa (Meadows and

Hoffman 2002) and around the world (Reid et al. 2000, DiGiano et al. 2013, Fagin et al. 2016,

Kleemann et al. 2017, Katusiime and Schütt 2020), unprotected Drakensberg grasslands are

composed of two common land tenure systems, specifically, private and communal systems

(O’Connor 2005). Land tenure reflects the complexity of socio-ecological systems, and has

become important for understanding the influence of humans on land transformation and

ecosystems (O’Connor 2005, Fagin et al. 2016, Robinson et al. 2018, Katusiime and Schütt

2020).

Chapter 1 3



Drakensberg plant diversity in the face of global change

Across communal and private land tenure systems, the certainty of access to land can

influence land use and land management practices (O’Connor 2005, Peden 2005, Kleemann

et al. 2017). Uncertain access to communal land has contrasted with secure ownership of

private land. On private systems, more certain access to land and ownership of resources

stimulates investment (Peden 2005). This investment advances land use intensification and land

transformation, but where natural grasslands are retained for animal production sustainable

management practices have been promoted (O’Connor 2005, Kleemann et al. 2017). On

communal areas access to land has been uncertain due to social upheavals that have been

accompanied by the erosion of traditional governance systems (Meadows and Hoffman 2002,

Peden 2005). Grassland management practices associated with this uncertainty, such as

heavy-continuous grazing, can deplete the more nutritious and productive tufted grasses, and

rare plant species, and also expose the soil surface to erosive forces (Meadows and Hoffman

2002, Peden 2005, Fagin et al. 2016).

In the Drakensberg, the long-term influence of disturbance regimes and specific

management practices associated with communal, private and protected systems (e.g., mowing

[Fynn et al. 2005, Ward et al. 2020] or selected harvesting of plants [O’Connor 2004]) on

grassland-plant diversity is largely unexplored (O’Connor 2004, O’Connor et al. 2010).

Although the recovery of grassland-plant diversity is known to be poor (Nerlekar and Veldman

2020) the potential value of secondary grassland in the Drakensberg is also largely unexplored

and requires investigation (O’Connor 2005).

Both fire and grazing are drivers of foremost importance in grassland (O’Connor et al. 2010,

Fuhlendorf et al. 2017, Lunt et al. 2012), but within protected areas of the Drakensberg fire is

the primary management tool (Morris et al. 2020). Herbivory pressure has been naturally

limited in Drakensberg grasslands. Owing to the poor nutrient status of the soils, herbage

quality is low, and the capacity of these grasslands to support herbivores year-round is limited.

In the absence of heavy-continuous grazing, fire is the primary consumer of herbaceous

biomass (Scotcher and Clarke 1981, Morris et al. 2020).

Plant diversity in protected areas of the Drakensberg is threatened by a poor understanding

of the influence of fire (Uys et al. 2004, O’Connor 2005, Gordijn and O’Connor 2021). The

long-term influence of fire regimes, described by their two main components, fire-return

interval and burn season (He et al. 2019), on plant diversity is largely unknown in Drakensberg

grasslands (O’Connor et al. 2010, Gordijn and O’Connor 2021). Similar to other rangelands,

most studies have focused on grasses, largely ignoring the full complement of plant diversity

(Uys et al. 2004, Gordijn et al. 2018).

1.1 Aims and objectives

In order to guide the conservation of the Drakensberg grassland-plant diversity in the face of

global change, the aim of this thesis was to reduce uncertainty around the influence of different

socio-ecological systems on grassland-plant diversity, in the face of global change. The

‘Mdedelelo-Cathkin’ landscape in the northern Drakensberg section in KwaZulu-Natal, South

Africa, with a good representation of communal, private, and protected land tenure systems,

Chapter 1 4



Drakensberg plant diversity in the face of global change

which reflect contrasting socio-ecological systems, provided an opportunity to realise this aim.

Close attention to the influence of fire regimes in protected areas was necessary. Increasing

human pressure has had a dramatic influence on fire regimes (Archibald et al. 2012), and

uncertainty around the influence of fire regime components, such as fire-return interval and

burn season, is possibly the largest threat to grassland-plant diversity. Therefore, in protected

areas where fire is the primary driver used to manage vegetation, the influence of different fire

regimes requires urgent attention (O’Connor et al. 2010).

The species-rich and high-turnover nature of Drakensberg grasslands (Carbutt 2019),

combined with the different influences of socio-ecological systems, necessitated an approach

that considered multiscale processes. Therefore, objectives were structured to provide insight

into landscape-level drivers, down to the influence of drivers at the smaller habitat-level spatial

scales, where plant diversity was quantified and the responses of species were evaluated. The

response to species to the influence of fire, herbivory and other drivers associated with the

contrasting socio-ecological systems, provided the opportunity to improve the understanding of

plant species responses, and functional groups thereof, to disturbance regimes. The first

objective was achieved at the landscape level using remote sensing, and complementary

objectives were realised through field-based vegetation sampling. The specific objectives were:

i) At the landscape-level scale, using remote sensing to undertake an LULCC analysis,

assess the influence of different socio-ecological systems on the spatial extent and drivers

of grassland transformation, and the vulnerability of different grassland habitats to

transformation.

ii) Evaluate the influence of socio-ecological systems and their associated management

practices on grassland-plant diversity, in primary and secondary grassland, across the

landscape.

iii) Assess the influence of fire regimes, defined by short to longer fire-return intervals, on

plant diversity, across the heterogeneous montane environment of protected Drakensberg

grasslands.

iv) Evaluate the treatment effects of fire regimes, defined by different fire-return interval and

burn seasons, on grassland-plant diversity, in protected Drakensberg grasslands.

Uncommon multidecadal data, extending back 30–70 years, were used to gain perspective

on the rate and extent of grassland habitat and plant diversity loss, under the influence of

contrasting socio-ecological systems. The complement of the rare long-term data, plus

thorough assessment of vegetation across the heterogeneous-montane landscape, provided this

opportunity to significantly reduce uncertainty around the fate of grassland-plant diversity in

the face of global change. For the assessment of fire regimes, the combination of the

landscape-scale assessment and controlled mesocosm trial provided a rare opportunity to make

a strong statement regarding the influence of fire in protected Drakensberg grasslands. In a

landscape-scale assessment, the influence of drivers may be obscured by environmental

variation, the uncertainty around which may be reduced through controlled experimentation.

Chapter 1 5



Drakensberg plant diversity in the face of global change

1.2 Thesis structure

In this chapter, the problem that Drakensberg grassland-plant diversity faces under global

change has been highlighted, and this thesis’ contribution to addressing this problem was

introduced. The next chapter, a literature review, unpacks the context necessary for

understanding the nature of this problem in Drakensberg grasslands (Chapter 2). The following

four chapters address the thesis’ objectives i–iv, and are prepared or presented as manuscripts

for publication (Chapters 3–6). Chapters 5 and 6, which addressed the influence of fire regimes,

have been published (Chapter 5 DOI: 10.1016/j.ppees.2018.07.005; Chapter 6 DOI:

10.1002/eap.2391). A general discussion and conclusion is presented in Chapter 7. In order to

limit page numbers in Chapters 3–6 as required for publication, supporting material is

presented in Appendices A–F.

References

Aguilar-Tomasini, M. A., T. Escalante, and M. Farfán (2020). Effectiveness of natural protected

areas for preventing land use and land cover changes of the Transmexican Volcanic Belt,

Mexico. Regional Environmental Change 20 (3):1–9.

Archibald, S., A. C. Staver, and S. A. Levin (2012). Evolution of human-driven fire regimes in

Africa. Proceedings of the National Academy of Sciences 109 (3):847–852.

Bengtsson, J., J. M. Bullock, B. Egoh, C. S. Everson, T. M. Everson, T. G. O’Connor,

P. J. O’Farrell, H. G. Smith, and R. Lindborg (2019). Grasslands—more important for

ecosystem services than you might think. Ecosphere 10 (2):e02582.

Bond, W. J. (2019). Open ecosystems: ecology and evolution beyond the forest edges. Oxford

University Press, Oxford, UK.

Bond, W. J. and C. L. Parr (2010). Beyond the forest edge: Ecology, diversity and conservation

of the grassy biomes. Biological Conservation 143 (10):2395–2404.

Boval, M. and R. Dixon (2012). The importance of grasslands for animal production and other

functions: a review on management and methodological progress in the tropics. Animal 6

(5):748–762.

Brooks, T. M., R. A. Mittermeier, G. A. da Fonseca, J. Gerlach, M. Hoffmann, J. F. Lamoreux,

C. G. Mittermeier, J. D. Pilgrim, and A. S. Rodrigues (2006). Global biodiversity

conservation priorities. Science 313 (5783):58–61.

Carbutt, C., W. Henwood, and L. Gilfedder (2017). Global plight of native temperate

grasslands: going, going, gone? Biodiversity and conservation 26 (12):2911–2932.

Carbutt, C., S. Tau, and B. Escott (2011). The conservation status of temperate grasslands in

southern Africa. Grass Roots 11 (1):17–23.

Carbutt, C. (2019). The Drakensberg Mountain Centre: a necessary revision of southern

Africa’s high-elevation centre of plant endemism. South African Journal of Botany

124:508–529.

Ceballos, G., P. R. Ehrlich, A. D. Barnosky, A. Garcıa, R. M. Pringle, and T. M. Palmer (2015).

Accelerated modern human–induced species losses: Entering the sixth mass extinction.

Science Advances 1 (5):1–5.

Chapter 1 6

https://doi.org/10.1016/j.ppees.2018.07.005
https://doi.org/10.1002/eap.2391


Drakensberg plant diversity in the face of global change

DiGiano, M., E. Ellis, and E. Keys (2013). Changing landscapes for forest commons: linking

land tenure with forest cover change following Mexico’s 1992 Agrarian Counter-Reforms.

Human ecology 41 (5):707–723.

Dinerstein, E., C. Vynne, E. Sala, A. R. Joshi, S. Fernando, T. E. Lovejoy, J. Mayorga,

D. Olson, G. P. Asner, J. E. Baillie, N. Burgess, K. Burkart, R. Noss, Y. Zhang, A. Baccini,

T. Birch, N. Hahn, L. Joppa, and E. Wikramanayake (2019). A global deal for nature:

guiding principles, milestones, and targets. Science Advances 5 (4):eaaw2869.

Fagin, T. D., J. M. Vadjunec, N. M. Colston, K. Wenger, and A. Graham (2016). Land tenure

and landscape change: a comparison of public-private lands in the southern High Plains.

Ecological Processes 5 (1):12.

Francis, R. and B. Bekera (2014). A metric and frameworks for resilience analysis of

engineered and infrastructure systems. Reliability Engineering and System Safety

121:90–103.

Fuhlendorf, S. D., R. W. S. Fynn, D. A. McGranahan, and D. Twidwell (2017). Rangeland

systems. Ed. by D. D. Briske. Springer, Cham, CH. Chap. Heterogeneity as the basis for

rangeland management, p. 169–196.

Fuhlendorf, S. D., W. C. Harrell, D. M. Engle, R. G. Hamilton, C. A. Davis, and D. M. Leslie Jr

(2006). Should heterogeneity be the basis for conservation? Grassland bird response to fire

and grazing. Ecological applications 16 (5):1706–1716.

Fynn, R. W. S., C. D. Morris, and T. J. Edwards (2005). Long-term compositional responses of

a South African mesic grassland to burning and mowing. Applied Vegetation Science 8

(1):5–12.

Gordijn, P. J., T. M. Everson, and T. G. O’Connor (2018). Resistance of Drakensberg

grasslands to compositional change depends on the influence of fire-return interval and

grassland structure on richness and spatial turnover. Perspectives in Plant Ecology,

Evolution and Systematics 34:22–36.

Gordijn, P. J. and T. G. O’Connor (2021). Multidecadal effects of fire in a grassland biodiversity

hotspot: does pyrodiversity enhance plant diversity? Ecological Applications 00 (00):1–19.

He, T. and B. B. Lamont (2018). Baptism by fire: the pivotal role of ancient conflagrations in

evolution of the Earth’s flora. National Science Review 5 (2):237–254.

He, T., B. B. Lamont, and J. G. Pausas (2019). Fire as a key driver of Earth’s biodiversity.

Biological Reviews:1983–2010.

Hoekstra, J. M., T. M. Boucher, T. H. Ricketts, and C. Roberts (2005). Confronting a biome

crisis: global disparities of habitat loss and protection. Ecology Letters 8 (1):23–29.

IUCN WCPA (2021). World Commission on Protected Areas: Temperate Grasslands Specialist

Group. Published on the Internet, URL: https://www.iucn.org/commissions/world-

commission-protected-areas/our-work/temperate-grasslands. Last accessed 11th

December 2021.

Jewitt, D., P. S. Goodman, B. F. N. Erasmus, T. G. O’Connor, and E. T. F. Witkowski (2015).

Systematic land-cover change in KwaZulu-Natal, South Africa: Implications for

biodiversity. South African Journal of Science 111 (9/10):01–09.

Chapter 1 7

https://www.iucn.org/commissions/world-commission-protected-areas/our-work/temperate-grasslands
https://www.iucn.org/commissions/world-commission-protected-areas/our-work/temperate-grasslands


Drakensberg plant diversity in the face of global change

Jewitt, D., P. S. Goodman, B. F. N. Erasmus, T. G. O’Connor, and E. T. F. Witkowski (2017).

Planning for the maintenance of floristic diversity in the face of land cover and climate

change. Environmental Management 59 (5):792–806.

Katusiime, J. and B. Schütt (2020). Linking land tenure and integrated watershed

management—a review. Sustainability 12 (4):1667.

Kleemann, J., G. Baysal, H. N. Bulley, and C. Fürst (2017). Assessing driving forces of land

use and land cover change by a mixed-method approach in north-eastern Ghana, West

Africa. Journal of Environmental Management 196:411–442.

Lunt, I. D., S. M. Prober, and J. W. Morgan (2012). Flammable Australia: fire regimes,

biodiversity and ecosystems in a changing world. Ed. by R. A. Bradstock. CSIRO

Publishing, Melbourne, AU. Chap. How do fire regimes affect ecosystem structure, function

and diversity in grasslands and grassy woodlands of southern Australia, p. 253–270.

Maxwell, S. L., V. Cazalis, N. Dudley, M. Hoffmann, A. S. Rodrigues, S. Stolton, P. Visconti,

S. Woodley, N. Kingston, E. Lewis, et al. (2020). Area-based conservation in the

twenty-first century. Nature 586 (7828):217–227.

Meadows, M. E. and M. T. Hoffman (2002). The nature, extent and causes of land degradation

in South Africa: legacy of the past, lessons for the future? Area 34 (4):428–437.

Morris, C. D., C. S. Everson, T. M. Everson, and P. J. Gordijn (2020). Frequent burning

maintained a stable grassland over four decades in the Drakensberg, South Africa. African

Journal of Range and Forage Science 38 (1):1–14.

Mucina, L. et al. (2006). The Vegetation of South Africa, Lesotho and Swaziland. Ed. by

L. Mucina and M. Rutherford. Strelitzia, Pretoria, ZA. Chap. Grassland Biome,

pp. 350–437.

Nagendra, H., D. K. Munroe, and J. Southworth (2004). From pattern to process: landscape

fragmentation and the analysis of land use/land cover change. Agriculture, Ecosystems and

Environment 101 (2–3):111–115.

Nerlekar, A. N. and J. W. Veldman (2020). High plant diversity and slow assembly of

old-growth grasslands. Proceedings of the National Academy of Sciences 117

(31):18550–18556.

O’Connor, T. G. (2004). Influence of land use on populations of the medicinal plant Alepidea

amatymbica in the southern Drakensberg. South African Journal of Botany 70 (2):319–322.

O’Connor, T. G. (2005). Influence of land use on plant community composition and diversity in

Highland Sourveld grassland in the southern Drakensberg, South Africa. Journal of Applied

Ecology 42 (5):975–988.

O’Connor, T. G. and P. Kuyler (2009). Impact of land use on the biodiversity integrity of the

moist sub-biome of the grassland biome, South Africa. Journal of Environmental

Management 90 (1):384–395.

O’Connor, T. G., P. Kuyler, K. P. Kirkman, and B. Corcoran (2010). Which grazing

management practices are most appropriate for maintaining biodiversity in South African

grassland? African Journal of Range and Forage Science 27 (2):67–76.

Pausas, J. G., J. Llovet, A. Rodrigo, and R. Vallejo (2009). Are wildfires a disaster in the

Mediterranean basin? A review. International Journal of Wildland Fire 17 (6):713–723.

Chapter 1 8



Drakensberg plant diversity in the face of global change

Pausas, J. G. and J. E. Keeley (2009). A burning story: the role of fire in the history of life.

BioScience 59 (7):593–601.

Peden, M. (2005). Tackling ‘the most avoided issue’: communal rangeland management in

KwaZulu-Natal, South Africa. African Journal of Range and Forage Science 22

(3):167–175.

Polley, H. W., D. W. Bailey, R. S. Nowak, and M. Stafford-Smith (2017). Rangeland Systems:

Processes, Management and Challenges. Ed. by D. Briske. Springer International

Publishing, Cham, CH. Chap. Ecological Consequences of Climate Change on Rangelands,

p. 229–260.

Pörtner, H. et al. (2021). IPBES-IPCC co-sponsored workshop report on biodiversity and

climate change. Tech. rep. IPBES and IPCC.

Rahbek, C., M. K. Borregaard, R. K. Colwell, B. Dalsgaard, B. G. Holt, N. Morueta-Holme,

D. Nogues-Bravo, R. J. Whittaker, and J. Fjeldså (2019). Humboldt’s enigma: What causes

global patterns of mountain biodiversity? Science 365 (6458):1108–1113.

Reid, R. S., R. L. Kruska, N. Muthui, A. Taye, S. Wotton, C. J. Wilson, and W. Mulatu (2000).

Land-use and land-cover dynamics in response to changes in climatic, biological and

socio-political forces: the case of southwestern Ethiopia. Landscape Ecology 15

(4):339–355.

Robinson, B. E., Y. J. Masuda, A. Kelly, M. B. Holland, C. Bedford, M. Childress,

D. Fletschner, E. T. Game, C. Ginsburg, T. Hilhorst, et al. (2018). Incorporating land tenure

security into conservation. Conservation Letters 11 (2):e12383.

Rydén, O., A. Zizka, S. C. Jagers, S. I. Lindberg, and A. Antonelli (2020). Linking democracy

and biodiversity conservation: empirical evidence and research gaps. Ambio 49

(2):419–433.

Sage, R. F. (2020). Global change biology: a primer. Global Change Biology 26 (1):3–30.

Sayre, R., D. Karagulle, C. Frye, T. Boucher, N. H. Wolff, S. Breyer, D. Wright, M. Martin,

K. Butler, K. van Graafeiland, J. Touval, L. Sotomayor, M. Jennifer, E. T. Game, and

H. Possingham (2020). An assessment of the representation of ecosystems in global

protected areas using new maps of World Climate Regions and World Ecosystems. Global

Ecology and Conservation 21:e00860.

Scotcher, J. and J. Clarke (1981). Effects of certain burning treatments on veld condition in

Giant’s Castle Game Reserve. Proceedings of the Annual Congresses of the Grassland

Society of Southern Africa 16 (1):121–127.

Silveira, F. A., M. Barbosa, W. Beiroz, M. Callisto, D. R. Macedo, L. P. C. Morellato,

F. S. Neves, Y. R. Nunes, R. R. Solar, and G. W. Fernandes (2019). Tropical mountains as

natural laboratories to study global changes: a long-term ecological research project in a

megadiverse biodiversity hotspot. Perspectives in Plant Ecology, Evolution and Systematics

38:64–73.

Suttie, J. M., S. G. Reynolds, and C. Batello (2005). Grasslands of the World. Food and

Agriculture Organisation, Rome, IT.

UNEP-WCMC and IUCN (2021). Protected Planet Report 2020. Tech. rep. Cambridge UK;

Gland, Switzerland.

Chapter 1 9



Drakensberg plant diversity in the face of global change

Uys, R. G., W. J. Bond, and T. M. Everson (2004). The effect of different fire regimes on plant

diversity in southern African grasslands. Biological Conservation 118 (4):489–499.

Veldman, J. W., E. Buisson, G. Durigan, G. W. Fernandes, S. le Stradic, G. Mahy, D. Negreiros,

G. E. Overbeck, R. G. Veldman, N. P. Zaloumis, F. E. Putz, and W. J. Bond (2015). Toward

an old-growth concept for grasslands, savannas, and woodlands. Frontiers in Ecology and

the Environment 13 (3):154–162.

Viviroli, D., R. Weingartner, and B. Messerli (2003). Assessing the hydrological significance of

the world’s mountains. Mountain Research and Development 23 (1):32–40.

Ward, D., K. P. Kirkman, Z. Tsvuura, C. D. Morris, and R. W. S. Fynn (2020). Are there

common assembly rules for different grasslands? Comparisons of long-term data from a

subtropical grassland with temperate grasslands. Journal of Vegetation Science 00:1–12.

White, R., S. Murray, and M. Rohweder (2001). Grassland ecosystems: pilot analysis of global

ecosystems. World Resources Institute, Washington DC, US.

Wright, J. B. and A. D. Mazel (2007). Tracks in a mountain range: exploring the history of the

uKhahlamba-Drakensberg. Witwatersrand University Press. Johannesburg, ZA.

Chapter 1 10



Chapter 2

The ‘eco-story’ of a mountain range:
acquiring perspective on Drakensberg
plant diversity and global change

2.1 Introduction

Developing local context is key to understanding the influence of global change on

ecosystems and their biodiversity (Fuhlendorf et al. 2017, Polley et al. 2017, Fordham et al.

2020). The response of an ecosystem’s biodiversity to global change depends on complex

interactions between their abiotic template (i.e., physical environment) and the drivers which

shaped their evolution (Bond 2019, He et al. 2019). This review introduces the abiotic

template, upon which Drakensberg plant diversity evolved, and the palaeo drivers that shaped

this evolution. The influence of past climate fluctuations, and insight into the nature of other

historic drivers, provides a framework for predicting the response of plant diversity to global

change (Fordham et al. 2020). In developing this context, the major vegetation communities of

the Drakensberg are described. Then, the early and later influence of humans, and their

interactions with the Drakensberg biodiversity, is discussed.

Once the palaeo and more recent context has been established, novel, human-modified

disturbance regimes are unpacked. These disturbance regimes, of which major components

include fire and herbivory, regulate grassland ecosystems and their biodiversity. The influence

of socio-ecological systems, which are represented by the communal, private and protected

land tenure systems in the study area, on these disturbance regimes and the consequences for

plant diversity are considered. This level of understanding is key to predicting the response of

plant diversity to human-modified disturbance regimes. Importantly, for a landscape-level

perspective, the influence of socio-ecological systems on grassland transformation, which

includes land use intensification, and is the largest driver of biodiversity loss on the earth’s

terrestrial surface (Sage 2020), is discussed.

This local context forms the knowledge-base for realising the aim of this thesis, to reduce

uncertainty around the influence of humans on the Drakensberg’s plant diversity. By reducing

uncertainty around the influence of humans, this thesis will contribute towards the

11



Drakensberg plant diversity in the face of global change

understanding or detection of the looming threat of climate change, on grassland-plant diversity.

2.2 Study area: abiotic template

The study area is situated in the central to northern section of the KwaZulu-Natal

Drakensberg mountains, and is delineated by the catchment areas of four tributaries of a major

river in the area, the Thukela River (28°75’ S, 29°10’ E; 29°25’ S, 29°70’ E; Fig. 2.1).

Specifically, the catchment areas of the uMlambonja, Linquespruit, Sterkspruit, and Little

Thukela Rivers. This study area is termed the ‘Mdedelelo-Cathkin’ area in this thesis.

‘Mdedelelo’ is the isiZulu (southern Bantu language) name for the prominent peak which

protrudes from the main Drakensberg escarpment in the south-west of the area (Fig. 2.1).

‘Mdedelelo’ can be interpreted as, ‘make room for him’, and is a reference to the prominence

of this peak which is visible throughout most of the study area. The English name for the peak,

‘Cathkin peak’, was imported due to the familiarity of a European settler with the Cathkin

Braes hill in Glasgow, Scotland (Pearse 2006).

The Drakensberg has warm, wet summers and cool, dry winters, characteristic of southern

African temperate grasslands (Mucina et al. 2006). However, the climate of the area is variable

but broadly corresponds to the montane altitude gradient and associated topographic features.

The Drakensberg is an erosion escarpment consisting of horizontal igneous, aeolian and

sedimentary strata which have been deeply incised by numerous rivers (King 1944). These

rivers, fed by the high rainfall in the region, contribute significantly to southern Africa’s water

supply (Schulze 1979).

The diverse abiotic template associated with montane terrain, and convergence of different

vegetation lineages in these mountains, have resulted in a speciose flora with a high proportion

of endemic species (Carbutt and Edwards 2006). Nine vegetation types are recognized for the

study area which are broadly delimited by altitude (Table 2.1).

Table 2.1: Vegetation types summarised by altitude, from low to high elevation, in the
Mdedelelo-Cathkin area (Mucina et al. 2006)

Broad Units Vegetation type Median Min Max

Sub-escarpment Grassland KwaZulu-Natal Highland Thornveld 1133 1060 1389
Sub-escarpment Grassland Northern KwaZulu-Natal Moist Grassland 1199 993 1602
Sub-escarpment Grassland Drakensberg Foothill Moist Grassland 1510 1200 1900
Forest Northern Afrotemperate Forest 1626 1141 2189
Drakensberg Grassland Northern Drakensberg Highland Grassland 1817 1346 2180
Drakensberg Grassland Drakensberg-Amathole Afromontane Fynbos 1838 1591 2144
Drakensberg Grassland uKhahlamba Basalt Grassland 2091 1587 3349
Drakensberg Grassland Lesotho Highland Basalt Grassland 2813 1591 2144
Drakensberg Grassland Drakensberg Afroalpine Heathland 3183 2714 3450

2.2.1 Geology, soils, and topography

The geomorphology and related pedogenesis of the Drakensberg is relatively simple. The

older or underlying geology consists of largely undisturbed horizontal sedimentary layers upon

Chapter 2 12



Drakensberg plant diversity in the face of global change

ma
dk

co
ts

do
oW

Spioenkop dam

Lindeq
ues

pr
ui

t

ti
ur

pse
u

qedniL

L i
tt
le
Th

uk

el
a

Litt le
T
h
u
ke
la

Little
Thukela

uMl
am

bo

nja

eM
hlaw

zini

S
te
rk
sp
ru
it

S
te
rk
sp
ru
it

N
je

su

th

i

Thukela

Mdedelelo-Cathkin Peak

Winterton

Bergville

Emmaus

CPRC

Kw
aZu

lu-Nata
l

Lesotho

29°6′E

29°6′E

29°12′E

29°12′E

29°18′E

29°18′E

29°24′E

29°24′E

29°30′E

29°30′E

29°36′E

29°36′E

29°42′E

29°42′E

2
9

°1
2
′S

2
9

°1
2′S

2
9

°6
′S

2
9

°6′S

2
9

°0
′S

2
9

°0′S

2
8

°5
4
′S

2
8

°5
4′S

2
8

°4
8
′S

2
8

°4
8′S

2
8

°4
2
′S

2
8

°4
2′S

0 5 10 15 20 km

Towns
M-C Peak

CPRC

International
boundary
Dam
River
AOI
Protected area

Legend

South Africa

Altitude (m .a.s.l.)
1000
1400
1800
2200
2600
3200

Figure 2.1: Layout of the core study area, that is, the area of interest (AOI) delineated by the
catchment areas of four Thukela River tributaries. Major dams include Woodstock and Spi-
oenkop in the north-west and central north respectively. Protected areas are concentrated in the
south-west, extending from escarpment to Drakensberg foothills. The upper Little Thukela, Lin-
quespruit and uMlambonja areas have been under communal tenure. Privately owned lands are
concentrated around Winterton and extend up the Sterkspruit valley. Winterton and Bergville
are the two major urban centres in the area. CPRC = Cathedral Peak research catchments.

which basaltic lavas erupted. High rainfall in the region plus associated frequent lightning

events have eroded these layers at varying rates, and shaped the formation of the Drakensberg

erosion escarpment (King 1944, Knight and Grab 2014) (Fig. 2.2). The geological formations

and associated soils derived from these are described from the uppermost (youngest) stratum,

to the lowest (oldest) stratum, below.

Chapter 2 13



Drakensberg plant diversity in the face of global change

Figure 2.2: 3D visualisation of the Mdedelelo-Cathkin landscape as depicted from a north-east
(near Winterton) to south-east (the main escarpment) perspective. Rivers and dams are labelled
in blue; place names in black. From left-to-right (south-west to the north-east, respectively): the
Njesuthi and lower Little Thukela catchment areas are largely under communal land tenure; the
next catchment to the right is the Sterkspruit which is largely under private tenure; the upper
reaches of the Linquespruit and uMlambonja river catchments are under communal tenure, and
lower reaches, towards Winterton and Woodstock dam, private tenure. Patches of commercial
cultivated lands are concentrated near Winterton. The dark green areas are afforested grassland,
largely dominated by Eucalyptus species (e.g., left of the Linquespruit label). Protected areas,
at higher altitudes (e.g., the Cathedral Peak Research Catchments) are a lighter green colour
(Landsat-8 image taken in April 2019 courtesy of the U.S. Geological Survey).

The basalt of the uppermost stratum, the Drakensberg Group (182 Ma) (Fig. 2.3), ranges

in thickness up to 2000 m, and constitutes the prominent cliffs (Geological Survey 1981). In

the study area, cliff faces reach up to 250 m. Soils near the escarpment are mostly shallow (<50

cm) (Killick 1961), owing to greater weathering and soil erosion rates on the higher, steeper

slopes of the erosion escarpment (Tyson et al. 1976). The dominant soils derived from this

formation are humic and oxidic (Turner 2007).

Beneath the Drakensberg Basalts are the fine-grained aeolian sandstones of the <200 m thick

Clarens Formation (190–182 Ma), the uppermost layer of the Stormberg Group (Geological

Survey 1981) (Fig. 2.3). Erosion of lower, less-resistant strata and the sandstone has resulted in

the formation of prominent white to cream coloured cliffs and associated caves, which were

used by early hunter-gatherers (Wright and Mazel 2007). The dominant soils derived from this

sandstone are oxidic and have a poorer nutrient status than those derived from basalt. However,

due to their greater sand content they are relatively more structured (Turner 2007).

Underlying the Clarens sandstone is the 210–70 m thick Elliot Formation (210–190 Ma)

which consists of multi-coloured mudstones and sandstones (Geological Survey 1981)

(Fig. 2.3). The older underlying Molteno Formation (237–210 Ma) has alternating beds of

sandstone with thin shale and mudstone layers, and is approximately 200 m thick (Geological

Survey 1981). The dominant soils derived from the Elliot and Molteno formations are oxidic

Chapter 2 14



Drakensberg plant diversity in the face of global change

ma
d	k

cot
sd

oo
W

Spioenkop	dam

Lindequ
esp

ruit

tiurpseuqedniL

L it
tle
	Th

uk

ela

Litt le	Thukela

Little
	Thukela

uMl
am

bon
ja

eMhlawzini

St
er
ks
pr
uit

St
er
ks
pr
uit

Nje
sut

hi

Thukela

Mdedelelo-Cathkin Peak

Winterton

Bergville

Emmaus

CPRC

Kw
aZ

ulu
-Na

tal

Lesotho

29°6′E

29°6′E

29°12′E

29°12′E

29°18′E

29°18′E

29°24′E

29°24′E

29°30′E

29°30′E

29°36′E

29°36′E

29°42′E

29°42′E29
°1
2′S

29°12′S
29

°6
′S 29°6′S

29
°0
′S 29°0′S

28
°5
4′S

28°54′S
28

°4
8′S

28°48′S
28

°4
2′S

28°42′S

0 5 10 15 20 km

AOI
Protected
areas
International
border

Towns

CPRC

Dam
River
M-C Peak

Geological formations
Drakensberg
Clarens
Elliot
Molteno
Tarkastad

Normandien
Masotcheni
Volksrust
Karoo dolerite

Legend

Figure 2.3: Geological formations of the study area. AOI = area of interest; CPRC = Cathedral
Peak research catchments.

(Turner 2007).

The Tarkastad Formation (251–237 Ma) of the Upper Beaufort Group (Catuneanu et al.

2005) underlies the Molteno Formation, and consists of fine- to medium-grained sedimentary

sandstones and mudstones (Geological Survey 1981) (Fig. 2.3). Underlying this is the

Normandien Formation (265–251 Ma) of the Lower Beaufort Group (Catuneanu et al. 2005)

which consists of dark-grey shale, siltstone, and fine- to course-grained sandstone (Geological

Survey 1981). The soils derived from Normandien Formation are largely oxidic (Turner 2007).

The Karoo Dolerite sills and dykes associated with the basalt flows of the Drakensberg

Group are best represented in the Normandien Formation. These have formed the many

conspicuous, weather-resistant rocky outcrops in the lower portion of the study area

(Geological Survey 1981, Turner 2007). Erosion of these dolerite outcrops has produced fertile

Chapter 2 15



Drakensberg plant diversity in the face of global change

oxidic soils in these areas (Turner 2007). The basaltic soils of the upper reaches of the

escarpment are similar in composition to the doleritic soils at lower elevations. Higher rainfall

in the vicinity of the escarpment summit and associated leaching of soil nutrients has made the

basaltic soils of higher elevations comparatively poorer in nutrients than lower doleritic soils

(Turner 2007).

The formation of duplex soils, associated with mudstones from the Tarkastad and

Normandien formations, have resulted in localized erosion-gulley and donga formations

(Turner 2007). The Masotcheni Formation is bounded by the younger Normandien and older

Volksrust Formation of the Ecca Group (.265 Ma) (Catuneanu et al. 2005) (Fig. 2.3). The

Masotcheni contains recent colluvial deposits, and conspicuous gulleys associated with duplex

soils (Catuneanu et al. 2005, Turner 2007).

From the water divide, marked by the prominent basalt cliffs of the main Drakensberg scarp,

numerous streams and rivers flow westwards into Lesotho and eastward into the study area.

The incision of valleys and gorges by the east-flowing river systems has resulted in the

formation of numerous spurs and ridges extending from the main scarp. The basalt followed by

the sandstone cliffs and ridges are the most precipitous. Moving away from the escarpment,

steeper slopes change to gentle-undulating hills and plains associated with the Normandien

Formation (Turner 2007) (Figs 2.1–2.3).

2.2.2 Climate

Changes in climate over palaeo-time scales, in concert with the diverse relief of the

Drakensberg mountain range, have driven the formation of species-rich grasslands in the

Drakensberg. In order to understand the responses of plant diversity to contemporary global

change dynamics, significant shifts in the palaeoclimate are noted in section 2.3. A

contemporary snapshot of meteorological variation over the Drakensberg landscape, which is

key to understanding the influence of the abiotic environment on plant diversity, is outlined

below. For this snapshot, the detailed compilation of climate figures and modelling by Tyson

et al. (1976) which spanned the study area was relied upon. Discontinuity in climate records in

the area have prevented an updated version of this report (MacKellar et al. 2014, Toucher et al.

2016), but trends in climate are noted in section 2.7.

Altitude has a prominent influence on the climate of the Drakensberg. Although

geographically situated within the subtropical latitudes of the Southern Hemisphere, the climate

is temperate, owing to cooling associated with the high altitude of the Drakensberg escarpment.

Over an approximate 2000 m altitude gradient, mean daily minima and maxima range from 7 to

11 °C in mid-winter and mid-summer (Table 2.2) (Tyson et al. 1976). Secondary spatial

patterns in micro- to macro-scale climate variation are largely associated with the

heterogeneous relief of this montane environment (Tyson et al. 1976, Toucher et al. 2016).

Temperature extremes in the Drakensberg range from hot-summer to frosty-winter

conditions. The maximum temperature is largely limited to ±35 °C and for ≤20 days per year.

Frost events are common occurring from April to October near the escarpment and May to

August/September in the lower regions of the study area (Tyson et al. 1976). The actual

number of frost events is difficult to predict due to the diverse relief and associated air

Chapter 2 16



Drakensberg plant diversity in the face of global change

Table 2.2: Temperature regimes over the altitude range of the Mdedelelo-Cathkin landscape.
Low altitude areas are represented by Winterton and Bergville at 1135 and 1026 m a.s.l., respec-
tively. High altitude areas by a site on the escarpment near Organ Pipes Pass, Cathedral Peak
(Tyson et al. 1976)

Mean daily minimum (°C) Mean daily maximum (°C)
July January July January

Low altitude (±1100 m a.s.l.) 4 15 15 29
High altitude (±3000 m a.s.l.) −6 8 8 18

movements. When cool air sinks into valleys and hollows in sub-zero mid-winter, conditions

conducive to frost events may occur on average 25 nights a month in mid-winter. In May and

August, these conditions occur on average 10 and 20 nights respectively (Tyson et al. 1976).

Variation in mean annual precipitation in the study area broadly corresponds to the altitude

gradient, ranging from ±600 mm near Winterton in the east to ±2000 mm at the summit in the

west (Fig. 2.1) (Tyson et al. 1976). Seventy percent of the rainfall is brought by mid- to

late-afternoon thunderstorms in summer (Tyson et al. 1976, Toucher et al. 2016). Cold frontal

systems that move in from the west bring wet and cooler conditions to the Drakensberg from

early Autumn to early Spring (Tyson et al. 1976, Toucher et al. 2016). Snow falls associated

with these frontal systems may occur year-round but are concentrated in mid-winter. An

average of 11.5 snowfalls have been recorded per year, concentrated on the summit of the

Drakensberg escarpment (Grab et al. 2017).

Under clear conditions, winds are dominated by the effects of the escarpment and valley

relief. Katabatic winds drain cool air from higher slopes at night contributing to frosty

conditions. During the day, warmer air moves upslope in anabatic winds (Tyson et al. 1976).

These winds affect the sub-escarpment areas forty plus kilometres away in KwaZulu-Natal

(Tyson et al. 1976). On a meso-scale level, land forms elevated above cool sinking winter air

are important refugia for cold intolerant species (Samways 1990).

Conditions highly conducive to fire ignition and spread are associated with large-scale,

warm, katabatic winds, known in southern Africa, as ‘berg winds’ (Strydom and Savage 2018).

These winds occur from April to September, when preceding the eastward movement of

anticyclones, warm air drains from the central southern African plateau over the Drakensberg

escarpment towards a coastal low. Adiabatic heating, associated with the coastward drainage of

warm air, produces hot, dry, and strong north-westerly winds (Grab and Simpson 2000). ‘Berg

winds’ are most frequent during early spring, when the number of anticyclones moving across

southern Africa peaks (Grab and Simpson 2000). The exposure of bare soil to erosive weather

events and consequent wind and runoff erosion is a management concern (Watson 1984).

2.3 Drakensberg plant diversity: origins, drivers, and disturbance
regimes

The grasslands of the Drakensberg contain phenomenal plant diversity. The study area

(Fig. 2.1) falls within the Drakensberg Mountain Centre of plant endemism (1300 to 3500 m

Chapter 2 17



Drakensberg plant diversity in the face of global change

a.s.l.) which has a total of ±2500 species with a 9% endemism level (Carbutt 2019). The area

ranks as the fourth-richest floral region of southern Africa. Additional centres of endemism in

the foothills of the Drakensberg, that intersect with the study area, are under investigation

(Mucina et al. 2006, Carbutt 2019).

In order to effectively conserve this plant diversity, it is important to understand the drivers

that have shaped patterns of plant diversity across the landscape. The nature of these drivers,

considered as together as disturbance regimes (Pickett et al. 1989), are often debated, influenced

by cultural perceptions, and are dependent on underlying management objectives—leaving an

uncertain future for grassland-plant diversity (Bond and Parr 2010, Bond 2019).

Palaeoecological research, which extends well before the recent influence of human settlement

intensification in the region (ca. 3000–200 BP) (Finch et al. 2021), provides critical insight into

the nature of these drivers, and shaping of the disturbance regimes that regulate Drakensberg

grasslands. Climate, fire, and herbivory have been the primary drivers that shaped the evolution

of grassland-plant diversity in the Drakensberg (Linder 2014). The shaping of the rich-plant

diversity of the Drakensberg over aeons contrasts with the relatively short period of intensifying

human-influence that has engendered novel disturbance regimes (see from section 2.4).

The Drakensberg’s rich-plant diversity is partly attributable to the convergence of alpine,

temperate and tropical biomes in the region (Mucina et al. 2006). The alpine region

(2800–3500 m a.s.l.) is dominated by C3 temperate grasses and contains fynbos or heath

elements (Mucina et al. 2006). The fynbos elements contribute significantly to the richness of

these grasslands having affinities to the richest plant region in the world, the Cape Floral

Kingdom (Carbutt and Edwards 2001). This alpine flora integrates into lower elevation floras,

specifically, the extensive temperate grasslands in the study area, then lower subtropical

savanna elements (Mucina et al. 2006, Linder 2014).

2.3.1 Climate and fire: palaeoecological drivers in grassland

The major uplift of the central interior in the Middle Miocene (ca. 16.0 Ma) that was

associated with aridification and global cooling since the Pliocene (ca. 5.3– 2.58 Ma) (Wichura

et al. 2010, Scott 2002) facilitated the establishment of temperate grasslands in southern Africa

(Bredenkamp et al. 2002). Seasonal drought, and disturbance by lightning-ignited fire and

herbivory favoured the reduced growth forms (Bond 2019) that characterise the contemporary

Poaceae of the Drakensberg. The development of the hemicrytophytic growth form of grasses

and other herbs, with their bud banks held close or at ground level, were an important

development for the spread of grasslands. By protecting their buds at the soil surface from a

harsh climate and fire, they could propagate the herbaceous phytomass to fuel seasonal fires,

thereby promoting their persistence (Bredenkamp et al. 2002, Strömberg 2011, Bond 2019).

The development of the underground storage organs of geophytic life forms, with organs such

as tubers, bulbs, corms, and woody rootstocks, were also important plant traits that engendered

survival in grasslands. Geophytic forbs and shrubs constitute the bulk of contemporary

grassland diversity (Uys et al. 2004, Parr et al. 2014).

Grasslands established from the late Pliocene remained throughout subsequent cool and arid

glacial, then warm and moister interglacial cycles of the Pleistocene (Scott 2002). Over the

Chapter 2 18



Drakensberg plant diversity in the face of global change

climatic instability of the Pleistocene and into the more stable and warmer Holocene

interglacial (<10k BP), changes in temperature and rainfall regimes have corresponded closely

to shifts in the altitudinal distribution of C3 and C4 grassland communities in southern Africa.

Festucoid grasses, which have a C3 photosynthetic pathway, grow optimally when the growing

season is cooler, typified in winter rainfall and some high altitude regions (Scott 2002, Parker

et al. 2011). The C4 photosynthetic mechanism was a tropical adaptation in response to

lowered atmospheric [CO2] and summer rainfall conditions under which these plants have

optimal growth (Scott 2002, Scott and Lee-Thorp 2004).

Palynology and isotope data from the Drakensberg suggest that in the Last Glacial

Maximum (25k–16k BP) of the Pleistocene, regional temperatures were as much as 5–6 °C

cooler. Under these conditions, higher altitude vegetation communities spread to lower

altitudes, as much as 1000 m lower than current extents (Neumann et al. 2014). Vegetation

shifts in the Holocene were comparatively minor compared with those associated with the

glacial-interglacial cycles of the Pleistocene. The early Holocene was marked by rising

temperatures and aridification (likely drier cold seasons, unfavourable for alpine vegetation)

which saw a retreat of C3 grassland elements up the altitudinal gradient of the Drakensberg

(Neumann et al. 2014, Norström et al. 2014). With increased seasonality, favourable for C4

grasses whose herbaceous phytomass cures seasonally, fire frequencies increased as suggested

by increased charcoal deposits towards the middle Holocene. The extent of charcoal deposits

also corresponds to sediment deposition, indicating some fire associated erosion events, along

with transitions from fynbos to grassland vegetation communities (Norström et al. 2014).

From the mid-Holocene (5000 BP), the largest observed shift in C3 and C4 grassland

communities, was a 400 m lowering of alpine C3 into C4 grasslands in Lesotho. This expansion

of C3 grasslands was associated with increasing cool and moist winters from ca. 2960 to 2100

BP (Parker et al. 2011). From ca. 2100 BP, a warmer summer-rainfall regime favoured the

expansion of C4 grasslands. The return of a warmer summer-rainfall climate during this period

of the Holocene (especially from ca. 1800 BP) has also been associated with distinct

anthropogenic influences on vegetation communities in the Drakensberg (February 1994,

Wright and Mazel 2007, Neumann et al. 2014, Lodder et al. 2017). Increased settlement of

farmers in the Drakensberg region from ca. 550 BP corresponds with evidence of increased fire

frequencies in the region (February 1994, Wright and Mazel 2007, Neumann et al. 2014). The

increases in fire frequency have been suggested to have promoted a decline in fire-sensitive

vegetation in the Drakensberg (e.g., Protea spp. [February 1994]). More recently, as indicated

by palynology studies in protected grasslands at Cathedral Peak, increased [CO2] has been

suggested to have resulted in an increase in C3 grassland elements (possibly from C3 grass

genera such as Festuca) (Lodder et al. 2017).

2.3.2 High-altitude alpine grassland diversity: Cape fynbos elements

Fynbos or heath is associated with oligotrophic environments, that is, environments with

nutrient poor soils, limited soil water availability and general environmental stress (Carbutt and

Edwards 2001). This vegetation, is considered the Cape component of Drakensberg flora, and

typical fynbos genera include C3 Poaceae, Iridaceae, Ericaceae and Proteaceae (Carbutt and

Chapter 2 19



Drakensberg plant diversity in the face of global change

Edwards 2001). Phytogeographic studies have shown that Drakensberg fynbos elements have

links to other such communities on high altitude environments throughout Africa (Carbutt and

Edwards 2001, Linder 2014). Origins of the Cape clades have been dated to the Paleogene

(66.0–23.0 Ma), while diversification thereof occurred in the Miocene (23.0–5.3 Ma) (Linder

2014).

Linkages between African fynbos communities are likely to have occurred when

environmental conditions were cooler and drier. Fynbos communities would have likely

descended from their high altitude oligotrophic refugia and formed links centred along Africa’s

escarpments (Carbutt and Edwards 2001, Clark et al. 2011). The high altitude alpine grasslands

of the Drakensberg likely served as one such refugia (Carbutt and Edwards 2001). Divergence

of Drakensberg flora from western clades has been influenced by the east to west rainfall

gradient in southern Africa, which has been present from the Cenozoic (66.0 Ma to current).

Termination of these linkages due to the shifts in climate would have facilitated further

diversification (Linder 2014).

The large scale spread of grasslands and associated increase in fire events is a key event in

the history of plant diversity on the African continent (Bredenkamp et al. 2002). Increases in

fire frequencies likely lead to the erosion of ancestral cool-temperate fynbos diversity over

Africa. Areas protected from fire in cooler climate zones, such as the high altitude Cape fold

mountains and Drakensberg escarpment, would have served as fire-protected areas for

cool-temperate flora (Linder 2014). In field studies over the last half-century, the controlling

effect of fire on fynbos and other woody plants has been observed in the Drakensberg. These

vegetation communities are restricted to fire-protected rocky outcrops, spurs, and valleys

(Granger and Schulze 1977, Carbutt and Edwards 2001). If fire is suppressed for approximately

ten years, fynbos species and temperate-adapted trees (e.g., Erica evansii) may ingress into

Drakensberg grasslands (Granger 1976).

The adaptation of alpine, cool-temperate fynbos clades such as Asteraceae to warmer

grasslands below the escarpment, is an example of the most recent radiation of plant diversity

in Africa (Bentley et al. 2014, Linder 2014). This ecotone is interesting as this flora has adapted

to the more frequent fires associated with a warmer climate (Manry and Knight 1986, Linder

2014). The nature of fire regime required to maintain this flora is, however, largely unexplored

and debated (Gordijn et al. 2018).

2.3.3 Mid-altitude temperate grasslands with subtropical grassland elements

The bulk of the herbaceous phytomass in Drakensberg grasslands is contributed by C4

grasses that originated from the fire assisted movement of temperate grasslands into lower,

tropical latitudes, from the Middle to Late Miocene (16.0–5.3 Ma) (Linder 2014). Temperate

grasslands flourished with the uplifting of the East African Plateau, and consequent cool and

arid conditions (Meadows and Linder 1993, Wichura et al. 2010). Recurrent fire then promoted

the spread of these grasslands into dry forests (subtropical thickets in the southern African

context), shaping novel vegetation communities (Linder 2014).

These novel vegetation communities, characterised by C4 grasses and remaining forest

elements, are associated with higher altitudes throughout Africa. Near the warmer and humid

Chapter 2 20



Drakensberg plant diversity in the face of global change

tropics the forest component of this vegetation dominates, whereas at higher latitudes under

cooler and drier conditions grassland dominates (Meadows and Linder 1993). In the

Drakensberg, patches of these evergreen forests are confined to cooler moist depressions and

areas protected from fire (below ±2000 m a.s.l.) (Adie et al. 2017). Similar to other African

forests surrounded by grasslands, these are species impoverished compared to their adjacent

grasslands (Meadows and Linder 1993, Adie et al. 2017). Examples of subtropical elements

found in the Drakensberg grasslands include Poaceae species such as, Loudetia simplex and

Monocymbium ceresiiforme, and Hypericum and Cussonia genera (Meadows and Linder 1993,

Linder 2014).

Interestingly, the adaptation of alpine Cape elements to these fire-prone grasslands

beginning in the Pliocene (5.3 Ma) is an example of one of the most recent plant radiations on

the continent (Linder 2014). The 900 m uplifting of the Drakensberg during this epoch

provided new montane habitat extremes and following glacial-interglacial swings would have

prompted the radiation of Asteraceae genera (Carbutt and Edwards 2006, Bentley et al. 2014,

Linder 2014).

2.3.4 Low-altitude temperate grasslands with tropical savanna elements

The sub-escarpment temperate grasslands in the lower portion of the study area have

affinities with savanna floral elements (Mucina et al. 2006). African savanna plant diversity is

centred on the watershed between the Congo, Zambezi and Ruaha rivers (Linder 2014).

Acacias, the quintessential African savanna tree, specifically Acacia sieberiana and Acacia

nilotica (alternative names Vachellia sieberiana and Vachellia nilotica) are found in

sub-escarpment Drakensberg grasslands (Mucina et al. 2006). The evolution of savanna flora is

closely linked to the development of C4 grasslands and associated increase in fire frequency in

the Miocene (23.0–5.3 Ma) (Keeley and Rundel 2005). However, typical savanna-clade pollen

and fossils originate from the earlier Palaeogene (66.0–23.0 Ma) (Muller 1981).

Climate and fire are important drivers of savanna extent, and the density of savanna woody

trees and shrubs. In drier cooler conditions, woody plants are largely limited by temperature,

and soil moisture (see section 2.5.4). Under mesic conditions (&650 mm mean annual

precipitation), without fire woody plants dominate and transform grasslands and savannas to

fynbos, thicket and forest (Bond et al. 2003). The influence of herbivory on woody vegetation

is generally positive (Hobbs 1996), however, densities of herbivores were likely too low to

exert an overriding influence on vegetation, in the Drakensberg (Rowe-Rowe and Scotcher

1986) (see section 2.3.5, below).

Temperature and rainfall seasonality have driven the extent of deciduous fire-prone C3 and

C4 grasslands in the Drakensberg. During cooler or winter-wet periods temperate grasslands

and fynbos have tracked the expansion of these conditions into lower altitudes. Similarly, in the

lower reaches of the study area, temperate grasslands have shifted into lower savannas during

cooler and more arid periods, and vice versa (Bredenkamp et al. 2002, Linder 2014). The

deciduous nature of C4 grasslands provided the fuel for lightning-driven fire regimes which

favoured the spread of these (Manry and Knight 1986, Keeley and Rundel 2005, Neumann et al.

2014). The contemporary influence of humans on the fire regime is discussed in section 2.4.

Chapter 2 21



Drakensberg plant diversity in the face of global change

2.3.5 Indigenous, low-pressure herbivory regimes

Drakensberg grasslands only support low, year-round, or at most moderate-herbivore

densities for transient periods, largely owing to a winter nutritional bottleneck (Mentis and

Duke 1976, Rowe-Rowe and Scotcher 1986). Two decades of herbivore estimates, where

herbivore numbers were unconstrained by humans, revealed that population numbers were

naturally restricted (Rowe-Rowe and Scotcher 1986). These grasslands are known as ‘sour

grassveld’ because they are palatable only in spring and early summer and after burns when

fresh regrowth occurs. The low herbage palatability is due to the high rainfall and consequent

leaching of soil nutrients in the area (Mentis and Duke 1976, Tainton 1999). At lower altitudes

(<1000 m a.s.l.) where rainfall decreases, the soils retain more nutrients and can support greater

year-round herbivore densities (Mentis and Duke 1976, Rowe-Rowe and Scotcher 1986).

The historic pattern of herbivore movement in the Drakensberg region has been difficult to

establish due to extirpation under hunting pressure and displacement via human settlement by

the 1860s (Wright and Mazel 2007). Over geological time scales, palaeoecological and

archaeological research has revealed that herbivore densities have fluctuated with climate and

associated C3 and C4 grassland extents. Over these fluctuations, herbivore densities increased

during warmer conditions where more palatable C4 grasslands were dominant (Parker et al.

2011). Historical documentation compiled within the last century suggests that some larger

herbivore species migrated seasonally along the elevational gradient of KwaZulu-Natal, from

the lower coast-ward regions in winter to higher grasslands in spring and summer (Rowe-Rowe

and Scotcher 1986, Rowe-Rowe and Taylor 1996, Wright and Mazel 2007, Boshoff and Kerley

2015). During such migrations larger herds of herbivores would have concentrated in fertile

lowlands, but their increased densities would have been transient. Land transformation, which

has impeded this type of seasonal migration in the Drakensberg, has likely contributed to the

reductions of indigenous herbivore populations (Boshoff and Kerley 2015, Patel et al. 2019).

Many of the smaller ungulates, restricted to specialized habitat patches, such as bushbuck

(Tragelaphus scriptus Pallas, 1766) and grey duiker (Sylvicapra grimmia Linnaeus, 1758) in

forest and thicket, klipspringer (Oreotragus oreotragus E. A. W. Zimmermann, 1783) on rocky

outcrops, mountain reedbuck (Redunca fulvorufula Afzelius, 1815) on steep slopes, common

reedbuck (Redunca arundinum Boddaert, 1785) in vlei and grassland (Rowe-Rowe and

Scotcher 1986), grey rhebuck (Pelea capreolus Forster, 1790) which feed mainly on forbs

(Rowe-Rowe and Scotcher 1986), and oribi (Ourebia ourebi E. A. W. Zimmermann, 1783) to

vlei and plateau have extant populations concentrated in protected areas (Mthimkhulu et al.

2019). Larger herbivores, such as black wildebeest (Connochaetes gnou E. A. W.

Zimmermann, 1780), blesbok (Damaliscus pygargus Pallas, 1767), red hartebeest (Alcelaphus

buselaphus Pallas, 1767) and Burchell’s zebra (Equus quagga Boddaert, 1785) are mostly

restricted to a few protected areas (Rowe-Rowe and Scotcher 1986, Mthimkhulu et al. 2019).

Common eland (Tragelaphus oryx Pallas, 1766) is the only indigenous, large herbivores that

commonly utilize high altitude Drakensberg grasslands (>1400 m a.s.l.), particularly the new

grass growth in spring. Their more selective feeding of woody plants, forbs, and grass make

this seasonal strategy possible for supporting relatively low numbers of eland in the

Drakensberg (0.75 eland·km-2) (Scotcher 1982, Rowe-Rowe and Scotcher 1986). Furthermore,

Chapter 2 22



Drakensberg plant diversity in the face of global change

their capability of traversing steep terrain has enabled their exploitation of high altitude

Drakensberg grasslands (Rowe-Rowe and Scotcher 1986). Other herbivores (and omnivores)

whose foraging habits in the Drakensberg remain largely unexplored include rodents such as

mole rats (e.g., Cryptomys natalensis Roberts, 1913) and Chacma baboons (Papio ursinus Kerr,

1792). Chacma baboon densities are however considered low due to the harsher climate and

relative scarcity of food compared to savanna habitat (Henzi et al. 1992).

The low, year-round densities of indigenous ungulates and other herbivores in Drakensberg

grasslands is expected to have a correspondingly low impact on overall grassland diversity

(Rowe-Rowe and Scotcher 1986, Morris 2017). In contrast, the replacement of these species

with livestock at high stocking rates is known to negatively affect indigenous grassland-plant

diversity in ‘sour grassveld’, especially under continuous grazing. This suggests that these

grasslands evolved under low, year-round grazing pressure (O’Connor et al. 2010, Joubert et al.

2014, Morris 2017).

2.4 Novel disturbance regimes: intensifying human settlement,
from hunter-gatherers to farmers

2.4.1 Hunter-gatherers

80 to 24 ka: Middle to Late Stone Age hunter-gatherers in Lesotho

The first evidence for the presence of humans in the vicinity of the Drakensberg was in

eastern Lesotho (Fig. 2.1) in the Pleistocene and the occupations were strongly pulsed at ca. 80,

60, 50, 46–38 and 24 ka (this pulsed occupation continued into the second millennium BP in

Lesotho) (Stewart et al. 2012). Comparably favourable highland conditions made Lesotho

attractive for settlement, even through the climatically unstable Late Pleistocene to Holocene

transition. The mountainous terrain provided reliable water resources during arid periods, plus

a diversity of natural resources associated with complex topography (Mitchell 1992, Stewart

et al. 2012).

25,000–600 BP: Late Stone Age to Late Iron Age San hunter-gatherers

To the east of the Lesotho highlands, there is some evidence of limited Middle Stone Age

occupation ca. 25,000 ka in the KwaZulu-Natal Drakensberg (Mazel 1982). From this time

until an insubstantial occupation in 8000 BP, no archaeological signs of occupation are evident.

Then from 3000 to 1600 BP there was more intensive San occupation, where after they moved

to the lower central Thukela basin (to the north-east of the study area shown in Fig. 2.1) where

farming communities were already present. After a millennium of absence hunter-gatherers

made a reappearance in the study area, in 600 BP (Wright and Mazel 2007). Based on other

similar hunter-gatherer communities, the minimum population density in the area would have

ranged from 0.01 to 0.06 persons per km2 (Mazel 1989).

The San hunter-gatherers in the Drakensberg likely followed the seasonal movement of

large herbivores along the altitudinal gradient from the KwaZulu-Natal coast to Drakensberg

mountains in summer (Wright and Mazel 2007). Resources for San in the Drakensberg were

Chapter 2 23



Drakensberg plant diversity in the face of global change

plentiful in summer while these were scarcer but sufficient in the cold winter. Over time,

particularly in the lower Thukela basin, a pattern of land use intensification emerged and a less

nomadic culture dependent on food plants, smaller mammals and fish emerged. These patterns

coincided with general social, economic and nutritional stresses. Seeds, bulbs, and corms

became an increasing important part of their diet as evidenced by archaeological excavations

(Mazel 1989, Wright and Mazel 2007). The appearance of isiNtu (Bantu) speaking farmers in

the region (from 600 BP) between the Drakensberg and coast may have inhibited their nomadic

behaviour along this altitudinal gradient (Wright and Mazel 2007).

The relatively low number of San, and their likely seasonal occupation suggests their impact

on the vegetation was low (Hilliard and Burtt 1987), but intensive use of certain plant species

may have had some localized influences especially, within the vicinity of cave shelters. Plants

comprised the most important component of their diet even when animals were relatively

abundant (Mazel 1989, Wright and Mazel 2007). It is possible that San used burns before

spring to promote the growth of perennial grassland plants with underground storage organs

with starches. Local patch burning in the dry season is likely to have been practised to attract

herbivores to the new grass growth (Hilliard and Burtt 1987, Deacon and Deacon 1999).

However, the influence of humans on the fire regime in not traceable until 600 BP. An analysis

of charcoal deposits in a sediment core from the Cathedral Peak research catchments indicates

that throughout the most of the Holocene (10k to 600 BP) charcoal deposition fluctuated over

millennia to a few centuries. These fluctuations correspond to changes in moisture and

temperature variation which affected vegetation growth and fuel for fires (Lodder 2007, Lodder

et al. 2017). This suggests that fire was largely climatically driven during the Holocene, and

that San fire usage corresponded to natural drivers such as lightning (Hilliard and Burtt 1987,

Neumann et al. 2014, Lodder et al. 2017, Finch et al. 2021).

2.4.2 Farmers

1750–200 BP: Zizi farmers

The settlement of agropastoral communities in the Drakensberg has had a pervasive

influence on the plant diversity of the Drakensberg. Early farmers have been suggested to burn

grassland to provide nutritious, year-round herbaceous forage for livestock, and as fire breaks,

for controlling the spread of fire (Finch et al. 2021). Following burns, the flush of nutritious

grass re-growth is valued for maintaining livestock. Evidence from charcoal deposits in

sediment cores and archaeological excavations suggest that with the encroachment of farmers

into the Drakensberg from lower sub-escarpment grasslands (from 600 BP), fire usage in these

areas increased (February 1994, Lodder 2012, Neumann et al. 2014). Fires, especially with

strong easterly winds, would have travelled from Drakensberg lowlands to the escarpment

(Granger 1976).

The Iron Age farmers in the study area concentrated in the savanna-grassland ecotone of

sub-escarpment grasslands (towards Bergville and Winterton, see Fig. 2.1). The lack of trees in

grasslands at higher altitudes would have limited their settlement away from savannas (Maggs

1982, Wright and Mazel 2007). Archaeology suggests that grasslands were more extensive at

lower altitudes at the time of Iron Age settlement; savanna encroachment has occurred since the

Chapter 2 24



Drakensberg plant diversity in the face of global change

arrival of these farmers (Maggs 1982). In the absence of trees in sub-escarpment grassland

(Fig. 2.1), farmers probably used maize cobs and dung to fuel domestic fires (Maggs 1982).

The main crops of these farmers were millets and sorghums. Later on, maize which has a

higher yield and is easier to grow, was introduced probably by Portuguese settlers along the

coa