The identification and verification of optimal 
reintroduction sites for the Southern Ground 
Hornbill Bucorvus leadbeateri in the Musina area of 
the Limpopo Province, South Africa  
 
 
 
Francette Jerling 
 
Research Unit for Environmental Sciences and Management, North-West 
University, Potchefstroom 
 
  
2011 
 
 
The identification and verification of optimal reintroduction sites 
for the Southern Ground Hornbill Bucorvus leadbeateri in the 
Musina area of the Limpopo province, South Africa 
 
 
Francette Jerling 
 
Dissertation submitted in partial fulfilment of the requirements for the 
degree  
 
Magister Scientiae  
 
in the Research Unit for Environmental Sciences and Management  
North-West University 
Potchefstroom  
 
 
Supervisor 
Mr. H.C. Coetzee 
Research Unit for Environmental Sciences and Management, North-West 
University, Potchefstroom 
 
Co-supervisor 
 Prof. L. Van Rensburg 
Research Unit for Environmental Sciences and Management, North-West 
University, Potchefstroom 
 
November 2011
i 
 
TABLE OF CONTENTS 
Acknowledgements..................................................................................................................iv 
Preface......................................................................................................................................vi 
Summary.................................................................................................................................vii 
List of figures.........................................................................................................................viii 
List of tables..............................................................................................................................x 
 
CHAPTER 1: INTRODUCTION AND LITERATURE REVIEW...................................1 
1.1 CONTEXT AND MOTIVATION FOR STUDY ....................................................2 
 1.1.1 Description of the Southern Ground Hornbill...........................................2 
 1.1.2 Current distribution of the Southern Ground Hornbill..............................4 
 1.1.3 Biology of the Southern Ground Hornbill.................................................4 
  1.1.3.1 Social and feeding behaviour.....................................................4 
  1.1.3.2 Nesting and breeding behaviour................................................5 
  1.1.3.3 Habitat requirements..................................................................6 
1.1.4 Conservation status and reasons for decline..............................................7 
 1.1.4.1 Conservation status....................................................................7 
 1.1.4.2 Reasons for decline in population numbers...............................7 
 1.1.5 Previous Southern Ground Hornbill reintroduction attempts....................8 
 1.1.6 Aim and objectives of this study...............................................................9 
 1.1.7 Importance of this study............................................................................9 
 
CHAPTER 2: STUDY SITE, MATERIALS AND METHODS........................................11 
 2.1 DESCRIPTION OF STUDY SITE........................................................................12 
  2.1.1 Location...................................................................................................12 
  2.1.2 Climate....................................................................................................13 
   2.1.2.1 Rainfall.....................................................................................13 
   2.1.2.2 Temperature.............................................................................13 
  2.1.3 Geology...................................................................................................13 
  2.1.4 Vegetation...............................................................................................14 
   2.1.4.1 Musina Mopane Bushveld........................................................15 
   2.1.4.2 Limpopo Ridge Bushveld..........................................................15 
 2.1.5 Plant communities...................................................................................16 
ii 
 
  2.1.5.1 Site A: White syringa-mopane-Lowveld cluster-leaf  
bushveld........................................................................17 
  2.1.5.2 Site B: Mopane-slender three-hook thorn low dense 
Bushveld.......................................................................18 
  2.1.5.3 Site C: Umbrella thorn-bushveld signal grass old fields.........19 
 2.2 MATERIALS AND METHODS...........................................................................20 
2.2.1 Selection of optimal reintroduction sites ................................................20 
2.2.2 Methods to determine the habitat profile (floristic survey).....................22 
2.2.3 Methods to determine availability of prey items (faunal survey)...........23 
2.2.3.1 Methods to identify and quantify (frequency and volume) prey  
            items.........................................................................................23 
2.2.4 Methods to determine the availability of trees for nesting and roosting 
        (presence and spacing/ density and distribution of suitable trees)...........24 
2.2.5 Pre-release site evaluation/stakeholder engagement...............................25 
 
CHAPTER 3: RESULTS.......................................................................................................26 
3.1 OPTIMAL REINTRODUCTION SITES..............................................................27 
3.2 HABITAT PROFILE ............................................................................................28 
 3.2.1 Species composition................................................................................28 
  3.2.1.1 Site A........................................................................................28 
  3.2.1.2 Site B........................................................................................30 
  3.2.1.3 Site C........................................................................................32 
 3.2.2 Change in ground cover..........................................................................33 
  3.2.2.1 Site A........................................................................................33 
  3.2.2.2 Site B........................................................................................34 
  3.2.2.3 Site C........................................................................................34 
3.3 AVAILABILITY OF PREY ITEMS.....................................................................35 
 3.3.1 Species composition................................................................................35 
  3.3.1.1 Site A........................................................................................35 
  3.3.1.2 Site B........................................................................................37 
  3.3.1.3 Site C........................................................................................39 
  3.3.1.4 Combination of all three sites, A, B and C...............................41 
 3.3.2 Quantification (frequency and volume) of prey items............................44 
3.4 AVAILABILITY OF TREES FOR NESTING AND ROOSTING .....................45 
iii 
 
3.5 PRE-RELEASE SITE EVALUATION/STAKEHOLDER ENGAGEMENT......45 
 3.5.1 Demographic information of stakeholders..............................................46 
 3.5.2 Attitudes regarding the SGH...................................................................46 
 3.5.3 Sightings of the SGH on farm or in area.................................................46 
 3.5.4 Willingness to reintroduce SGH onto farm.............................................46 
 3.5.5 Presence of nest on farm.........................................................................47 
 3.5.6 Dynamics regarding other stakeholders in the area.................................47 
 3.5.7 Reasons for decline according to stakeholders........................................47 
 3.5.8 Sufficient farm size for reintroductions..................................................47 
 3.5.9 Active conservation officer.....................................................................48 
 3.5.10 Local vet in the area..............................................................................48 
 3.5.11 Predators in the area..............................................................................48 
 3.5.12 Problems regarding window-breaking..................................................48 
 
CHAPTER 4: DISCUSSION.................................................................................................49 
4.1 DISCUSSION: OPTIMAL REINTRODUCTION SITES....................................50 
4.2 DISCUSSION: HABITAT PROFILE...................................................................50 
 4.2.1 Forbs and grasses....................................................................................50 
  4.2.2 Woody structure......................................................................................51 
4.2.3 Change in ground cover..........................................................................53 
4.3 DISCUSSION: AVAILABILITY OF PREY ITEMS...........................................54 
4.4 DISCUSSION: AVAILABILITY OF TREES FOR NESTING AND 
      ROOSTING............................................................................................................56 
4.5 DISCUSSION: PRE-RELEASE SITE EVALUATION/STAKEHOLDER 
      ENGAGEMENT....................................................................................................56 
 
CHAPTER 5: CONCLUSION AND RECOMMENDATIONS........................................58 
 
REFERENCES.......................................................................................................................61 
 
APPENDIX 1..........................................................................................................................68 
APPENDIX 2..........................................................................................................................70 
iv 
 
ACKNOWLEDGEMENTS 
Firstly, I would like to give a big hug and thanks to my parents who gave me the opportunity 
to study and encouraged me to get where I am today. Thank you for all your support over the 
years and for believing in me.  
Prof Leon van Rensburg, thank you for seeing the potential in me, and for your trust in me 
when I didn’t have any in myself. Thank you for all your encouragement and for your 
generosity; for the financial and moral support. I would never have been able to have done it 
without your help. 
Hendri Coetzee, thank you for everything that you have done for me and have taught me. 
Thank you for your guidance and your support. You always had faith in me to do the things I 
never thought I could possibly have achieved, and for pushing me to do things I had never 
thought I would be capable of doing. Without your support and your knowledge I would not 
have come to where I am now. 
Duan Steyn and Henry Alers and the owners of Greyghost safaris. Thank you for the freedom 
of using your farm for my fieldwork. I have lived like a queen over the past year in the luxury 
accommodation; I felt so spoilt! Thank you for your hospitality and for always making me 
feel welcome. I will miss the fireplace under the big baobab and the animals on the lawn, and 
the game drives and late night chats under the stars. Thank you to the staff of Greyghost 
Safaris; to Mogadi, Jan, Felicitas, and Shepherd. 
Retha and Oom Tuba van der Walt, thank you both for your hospitality and your kindness. 
Retha, I cannot thank you enough for your amazing book on plants. Without it I don’t know 
what I would have done. Your knowledge and passion for the bush has inspired me indeed! 
Oom Tuba, thank you for always being friendly and willing to help. Thank you for getting us 
out of the thickest mud Musina has ever seen, even though it took two tractors, a Land 
Cruiser and 10 men. 
Dirk Cilliers, thank you for your input with your knowledge of GIS and your ever-welcoming 
and friendly attitude and patience.  
Thank you to the Mabula Ground Hornbill Research and Conservation Project for introducing 
me to the world of hornbills and for being so hospitable and helpful.  
v 
 
Last, but most definitely not least, thank you to all the farmers of Musina for your hospitality 
and kindness. You were all willing to open your homes to me and for participating in my 
study. Thank you for your willingness and eagerness to help save the Southern Ground 
Hornbill. Their survival greatly depends on you. 
vi 
 
PREFACE 
The research reported in this dissertation was conducted from February 2010 to November 
2011 in the School of Environmental Sciences and Development, North-West University, 
Potchefstroom Campus, Potchefstroom, South Africa. The first part of the research was 
purely literary research which was conducted during 2010, and the remainder of the research 
period was focused on field research in the selected study area.  
 
The research conducted in this dissertation represents original work undertaken by the author 
and has not previously been submitted for degree purposes to any university. Where use has 
been made of the work of other researchers, it is duly acknowledged in the text. 
 
The reference style used in this dissertation is according to the specifications given by the 
Harvard method subscribed to by the International Standards Organisation and the HSRC. 
 
Any opinions, findings, and conclusions or recommendations expressed in this dissertation 
are those of the author and therefore the North-West University, Potchefstroom Campus, does 
not accept any liability thereto.  
vii 
 
SUMMARY 
The Southern Ground Hornbill (SGH) (Bucorvus leadbeateri) has recently been listed 
globally by the IUCN Red list as Endangered or Critically Endangered. In South Africa, the 
declining population of the SGH has led it to be listed nationally as Vulnerable and falls 
under the “Protected Species” legislation. Research into the habitat preferences and 
identification of suitable reintroduction areas have been few since the initiation of 
reintroduction attempts, therefore resulting in mixed outcomes of previously-reintroduced 
SGH. The aim of this study was to identify and verify optimal reintroduction sites for the 
SGH, in the Musina area in the Limpopo Province. Five main objectives were determined, 
namely: to identify optimal reintroduction areas; to develop a habitat profile of the designated 
area;  to determine the availability of prey items; to determine the availability of large trees 
needed for nesting and roosting; to evaluate pre-release sites by involving a number of 
stakeholders. Three study sites were selected on the farm Greyghost Safaris (Ludwigslust), 
each made up of a different plant community. Sampling was carried out a total of four times 
during the year: one sampling bout in each season. A niche-based modelling technique was 
used to describe the suitability of a particular habitat (“ecological space”) then used to project 
it onto another geographical space. Floristic and faunal surveys were conducted to determine 
the species composition of prey items and composition and frequency of the herbaceous and 
woody layer. High-resolution, natural-colour aerial imagery was used in order to determine 
the availability of trees for nesting and roosting. Pre-release site evaluations and stakeholder 
engagements were conducted by means of interviews with landowners on and around the 
study sites. The Limpopo River Valley and across to the KNP on the eastern border of the 
Limpopo Province, is suitable re-introduction sites for the SGH. This study site provided a 
good opportunity to explore what an optimal site for reintroduction should resemble.  
Keywords: Southern Ground Hornbill; habitat profile; prey abundance; vegetation structure; 
plant community; conservation; reintroduction; habitat viability assessment 
viii 
 
LIST OF FIGURES 
Figure 1.1: Left shows a female with red facial skin with a blue patches under throat; top 
right shows juvenile bird being fed (right) by adult male; bottom right shows male with all-
red facial skin.............................................................................................................................3 
 
Figure 1.2: Current distribution of the Southern Ground Hornbill in South Africa (Cilliers,     
2011)...........................................................................................................................................4 
 
Figure 2.1: Topocadastral map of Greyghost Safaris (Van Rooyen, 2008)............................12 
 
Figure 2.2: Musina Mopane Bushveld and Limpopo Ridge Bushveld forming the main 
vegetation types for Greyghost Safaris (Mucina & Rutherford 2006).....................................14 
 
Figure 2.3: Different plant communities found on Greyghost Safaris....................................17 
 
Figure 2.4: Illustration of the four-step process for identifying optimal reintroduction sites in 
the Limpopo Province..............................................................................................................21 
 
Figure 2.5: Diagram illustrating 4 x 50 m cross transect with measurements and samples 
taken at 1 m intervals starting 10 m from the centre point.......................................................23 
 
Figure 2.6: Diagram illustrating 4 x 50 m cross transect with measurements and samples 
taken at 1 m intervals starting 10 m from the centre point.......................................................24 
 
Figure 3.1: Map illustrating the areas that are from least to most probable areas of optimal 
habitat for the SGH in the Limpopo Province..........................................................................27 
 
Figure 3.2: Map of farmland areas in the Musina, Pondrif and Alldays vicinity. Those farms 
in red are those that were identified as the most optimal 
sites...........................................................................................................................................28 
 
Figure 3.3: Illustration of the change in the quantity of prey items depicted against the 
different months for each site A, B and C................................................................................42 
ix 
 
 
Figure 3.4: Map illustrating the availability of large trees on the farm Greyghost Safaris.....45 
 
x 
 
LIST OF TABLES 
Table 2.1: Rainfall readings for Greyghost Safaris from the year 2008-2011........................13 
 
Table 3.1: Species list of forbs and grasses for Site A............................................................29 
 
Table 3.2: Basic descriptive statistics for forbs and grasses for Site A (STATISTICA 10)...29 
 
Table 3.3: Species list, frequency and size of the woody structure for Site A: small (up to 5 
m), medium-sized (5-15 m), large (15-30 m) and very large  
(more than 30 m)......................................................................................................................30 
 
Table 3.4: Species list of forbs and grasses for Site B............................................................30 
 
Table 3.5: Basic descriptive statistics for forbs and grasses for Site B (STATISTICA 10)...31 
 
Table 3.6: Species list, frequency and size of the woody structure for Site B: small (up to 5 
m), medium-sized (5-15 m), large (15-30 m) and very large  
(more than 30 m)......................................................................................................................31 
 
Table 3.7: Species list of forbs and grasses for Site C............................................................32 
 
Table 3.8: Basic descriptive statistics for forbs and grasses for Site C (STATISTICA 10)...32 
 
 
Table 3.9: Species list, frequency and size of the woody structure for Site C: small (up to 5 
m), medium-sized (5-15 m), large (15-30 m) and very large  
(more than 30 m)......................................................................................................................33 
 
Table 3.10: Friedman ANOVA analysis for determining the average change in ground cover 
for Site A (p = 0.49039)...........................................................................................................34 
 
Table 3.11: Friedman ANOVA analysis for determining the average change in ground cover 
for Site B (p = 0.20797)...........................................................................................................34 
 
Table 3.12: Friedman ANOVA analysis for determining the average change in ground cover 
for Site C (p = 0.06456)...........................................................................................................35 
 
xi 
 
Table 3.13: Species composition and frequency of prey obtained from both pitfall and 
sweepnet traps for each month at Site A..................................................................................35 
 
Table 3.14: Descriptive statistics for the species composition sampled in both the pitfalls and 
sweepnet traps for Site A.........................................................................................................36 
. 
Table 3.15: Friedman ANOVA analysis for determining the change in species composition 
for Site A (p = 0.06656)...........................................................................................................37 
 
Table 3.16: Species composition and frequency of prey obtained from both pitfall and 
sweepnet traps for each month at Site B..................................................................................37 
 
Table 3.17: Descriptive statistics for the species composition sampled in both the pitfalls and 
sweepnet traps for Site B..........................................................................................................38 
 
Table 3.18: Friedman ANOVA analysis for determining the change in species composition 
for Site B (p = 0.03725)...........................................................................................................38 
 
Table 3.19: Species composition and frequency of prey obtained from both pitfall and 
sweepnet traps for each month at Site C..................................................................................39 
 
Table 3.20: Descriptive statistics for the species composition sampled in both the pitfalls and 
sweepnet traps for Site C..........................................................................................................40 
 
Table 3.21: Friedman ANOVA analysis for determining the change in species composition 
for Site C (p = 0.27361)...........................................................................................................40 
 
Table 3.22: Grouped descriptive statistics for the species composition sampled in both the 
pitfalls and sweepnet traps for Site A, B and C........................................................................41 
 
Table 3.23: Grouped analysis (Site A, B and C) (Friedman ANOVA) for the determination of 
the average change in species composition (p = 0.01950).......................................................42 
 
xii 
 
Table 3.24: Independent grouping (Kruskal-Wallis test) results using multiple comparisons 
for January (p = 0.4982)...........................................................................................................43 
 
Table 3.25: Independent grouping (Kruskal-Wallis test) results using multiple comparisons 
for April (p = 0.5021)...............................................................................................................43 
 
Table 3.26: Independent grouping (Kruskal-Wallis test) results using multiple comparisons 
for August (p = 0.8335)............................................................................................................43 
 
Table 3.27: Independent grouping (Kruskal-Wallis test) results using multiple comparisons 
for October (p = 0.4476)..........................................................................................................44 
 
Table 3.28: Combined data reflecting the number of invertebrates from all three site and the 
volume which was obtained from the invertebrates...............................................................44
1 
 
CHAPTER 1 
 
INTRODUCTION AND LITERATURE 
REVIEW 
2 
 
CHAPTER 1: INTRODUCTION 
 
1.1 CONTEXT AND MOTIVATION FOR STUDY 
This chapter gives a brief overview of the following topics: description of the Southern 
Ground Hornbill (Bucorvus leadbeateri, Vigors, 1825) (SGH), its current distribution, basic 
biological information, current conservation status, reasons for decline, previous 
reintroduction attempts, and the objectives as well as importance of this study. 
 
1.1.1 Description of the Southern Ground Hornbill 
The SGH is the largest hornbill (Bucerotiformes) in the world. There are approximately 54 
hornbill species that are grouped together in the order Bucerotiformes. Two of these, namely 
the Southern and Northern (Abyssinian) Ground Hornbill (NGH), differ from most other 
hornbills and are grouped in their own family, Bucorvidae (Kemp, 1995:94).   
 
The adult male SGH is approximately 100 cm in body length and has an all black plumage 
with white primary feathers. It has a large, black bill with a low casque at the ridge on the 
base of the bill. Around its eyes and throat, it has bare red facial skin with an inflatable red 
throat pouch below its bill. Its eyes are yellow, and it has black legs and feet. The adult 
female is similar in appearance but is slightly smaller and has violet-blue patches on the 
throat, sometimes extending down the side of the neck (Fig. 1.1). The NGH differs from the 
SGH by having predominantly blue facial skin, a cowl-like casque and a yellow patch at the 
base of upper mandible. It also has a slightly different call to the SGH (Kemp, 1995). 
 
 
3 
 
 
Figure 1.1: Left shows a female with red facial skin with a blue patches under throat; top 
right shows juvenile bird being fed (right) by adult male; bottom right shows male with all-
red facial skin. 
 
The immature SGH differs from adult birds by having a more sooty-brown colouration with 
black flecks in the white primaries. The bill of the immature bird is also greyer in colour and 
is not as large as the adult bird’s (Fig. 1.1). Its eyes are grey with pale grey-brown facial skin. 
Once the bird has reached a sub-adult stage, which can be from the age of three years, it will 
start to gain adult colouration around its face and neck (Kemp,1995).  
 
The SGH makes a deep four-note booming call, or hoo hoo hoo-hoo, followed by three 
bodily contractions, the last one producing the double note. Males and females often call in a 
duet and mostly at dawn and dusk. The sound of the call carries over large distances of up to 
5 km and is presumably used to advertise their territories (Kemp, 1995). 
 
4 
 
1.1.2 Current distribution of the SGH 
The SGH is distributed throughout Rwanda, Burundi, southern Kenya, Zaire, Tanzania, 
Angola, Zambia, Malawi, Namibia, Botswana, Zimbabwe, Mozambique and South Africa 
(Kemp, 1995). In South Africa (Fig. 1.2), the SGH can be found in the north-western and 
northern parts of the Limpopo Province (e.g. Limpopo River Valley), in the Kruger National 
Park and its surrounding game reserves in the Limpopo and Mpumalanga Provinces, where 
the majority of the population occur, down into the northern parts of KwaZulu-Natal, into its 
midlands areas and all the way down into the Eastern Cape.  
 
 
Figure 1.2: Current distribution of the Southern Ground Hornbill in South Africa (Cilliers,     
2011). 
 
 
1.1.3 Biology of the SGH 
 
 1.1.3.1 Social and feeding behaviour 
A SGH group territory can stretch up to a distance of 100 km2 per group from estimations 
made in the Kruger National Park (KNP) (Kemp & Kemp, 1980), Eastern Cape (Vernon, 
5 
 
1984) and KwaZulu-Natal (Knight, 1990) but can reach up to 200 km2 in semi-arid areas 
outside of protected areas (Theron, 2011). A group primarily consists of numbers ranging 
between three and five birds with one alpha breeding pair. The remainder of the group will 
then consist of a combination of male and female juveniles and adult male helpers (Kemp, 
1980). Social interactions between the birds include cooperative hunting, play, cooperative 
breeding, allo-feeding and allo-preening (Kemp & Kemp, 1980).  
 
According to Kemp (1995), foraging takes place during the day, with birds walking an 
average distance of 11 km. Walking is preferred (up to 70% of the day) unless unsuitable 
habitat has to be crossed (Kemp, 1995). The diet of the hornbill, being primarily carnivorous 
(Kemp, 1987; Kemp, 1988; Msimanga, 2004), consists of anything it can overpower, 
including snakes, tortoises, rabbits, rodents, lizards, smaller birds and invertebrates (Kemp & 
Kemp, 1978). The groups forage in areas with low plant growth densities, such as open 
grasslands in KZN (Knight, 1990), and do so by capturing prey in the low vegetation or by 
digging in the ground with their large bills (Kemp, 1995).  
 
 1.1.3.2 Nesting and breeding behaviour 
According to Jones (1969), nest-spacing between groups usually ranges from 4.0-8.9 km, the 
latter being in lesser dense areas. In more dense vegetation, as is found in the KwaZulu-Natal 
reserves and farmland, nest-spacing was found to have averaged a distance of up to 10 km. 
Hornbill groups usually remain sedentary within their territories throughout the year as long 
as there is adequate food and trees available for nesting and roosting (Kemp, 1980; Kemp, 
1988).  
 
Breeding behaviour of the SGH is primarily cooperative breeding (Du Plessis, Siegfried & 
Armstrong, 1995), as mentioned above, with one monogamous pair per group and several 
helpers (Ranger, 1931). Breeding takes place at the start of the rainy season from September 
to November, usually following within 1-2 weeks after the first heavy rainfall (Kemp & 
Kemp, 1991) Late rainfall for example in May, will result in a missed breeding endeavour. 
Nest sites consist of large tree cavities of at least 40 cm diameter in the KNP (Kemp & Begg, 
1996).  
 
The SGH’s preferred nesting sites in the KNP, due to its large size, include cavities with a 
median internal diameter of 40 cm in a tree (96%) or on a rock face (4%) (Kemp & Begg, 
6 
 
1996). Popular nesting trees for the SGH include: Comretum imberbe, Ficus sycomorus, 
Diospyros mespiliformis, Sclerocarya birrea and Adansonia digitata (Kemp, 1995). The 
availability of suitable nesting sites are a primary factor limiting the SGH population in the 
KNP (Kemp, 1995). Once a nest has been selected, the female will spend approximately five 
hours a day in the cavity, preparing the inside of the nest. Helpers will help to bring in dried 
leaves as well as food items for the female (Kemp & Kemp, 1980). The alpha male will 
encourage the female towards the nest by giving a high-pitched call and by depositing food 
items for the female outside the nest cavity (Ranger, 1931).  
 
Once the female is ready to lay her eggs, she will lay a clutch of one to three eggs, each a few 
days apart, and incubate them for 42 days (Kemp, 1976) followed by an 86 day nestling 
period (Kemp & Kemp, 1991). If a second chick is hatched, it will usually die of starvation 
within a week due to parental neglect as well as sibling competition (Kemp & Kemp, 2007). 
Once the first egg has hatched, it takes another 86 days before the fledgling will be old 
enough to leave the nest. Feeding during this time is assisted by the other members of the 
group. The chick will remain a juvenile for five to six years before it reaches sexual maturity, 
and even then it might not be ready to breed (Kemp, 1976).  
 
It has been estimated that in South Africa, the SGH will start to breed for the first time only 
after nine years (Kemp, 1995). Females, after having obtained their full facial colouring, tend 
to leave the group in search of a new group or to start a group themselves. Because of their 
slow reproductive rate, Kemp (2000b) suggested that by harvesting the second-hatched 
chicks and rearing them in captivity, the chicks can be used as a resource that could be used 
to manage the conservation of the species. Attempts of this sort have been taking place in the 
Kruger National Park (KNP) since 1969 (Kemp, 2000b). 
 
1.1.3.3 Habitat requirements of the Southern Ground Hornbill 
A suitable habitat site is an important aspect to be considered before reintroducing ground 
hornbills back into the wild. The need to clearly understand habitat utilisation incurred by the 
SGH lies in the aspiration of contributing to the overall fitness and the survival of the 
individual (Block & Brennan, 1993). Once this is understood, conservationists can gain 
knowledge of the factors that are limiting and affect the size and selection of territories. 
 
7 
 
SGH are prone to choose the savannah biome as their nesting and foraging territories, which, 
according to Tainton (1999), is characterised by seasonal rainfall and dry winters. SGH’s can 
also be found in some woodland and grassland (Kemp & Webster, 2007; Wilfred, 2007; 
Engelbrecht et al, 2007). The savannah biome has the richest abundance of bird species, but 
only 8.5% is protected (Rutherford & Westfall, 1994), hence the emphasis on the cooperation 
from landowners to maintain ecological integrity (Theron, 2011). Alongside the SGH, 
competing for prey, the savannah and grassland biomes includes many other terrestrial 
carnivorous bird species, such as cranes (now Anthropoides paradiseus), herons (Ardea 
melanocephala), storks (Leptoptilos crumeniferus), Otis bustards and the Secretarybird 
(Sagittarius serpentarius) (Kemp & Kemp, 1978).  
 
1.1.4 Conservation status and reasons for decline  
 
1.1.4.1 Conservation status 
According to the South African Red Data Book (Kemp, 2000), the species has been listed as 
Vulnerable. However, more recent studies have determined that with a 66% decline in range 
over the last 115 years, the species may require re-classification as Endangered or Critically 
Endangered (Kemp & Webster, 2008). An IUCN global status review of the birds has 
precipitated a change from Least Concern to Vulnerable by BirdLife International due to the 
increase in knowledge of the existing threats posed to the SGH (BirdLife International, 
2010). This has caused some concern regarding the decline in population numbers and 
distribution of the SGH (Kemp, 1980; Vernon & Herremans, 1997; Cyrus & Robson, 1980).  
 
During the 1960’s and 1970’s, the SGH population in South Africa was much larger; 
however, over the last 20-30 years, the population numbers declined rapidly (Theron, 2011). 
Conservation strategies in South Africa have been ongoing over a period of 15-20 years, 
including education and awareness campaigns among children, farmers and the general 
public. A number of different activities have been taking place, such as harvesting and hand 
rearing of second-hatched chicks, captive breeding, reintroduction programmes, as well as the 
erection of artificial nests 
 
 1.1.4.2 Reasons for decline in population numbers 
The main reasons for their decline have been attributed to a number of human-induced threats 
such as the removal of large trees needed for nesting and roosting, habitat alteration and 
8 
 
destruction, direct and indirect poisoning, persecution, increase in human population (Kemp, 
2000; Engelbrecht et al, 2007; Morrison et al, 2007) and the illegal trade in live specimens 
and traditional practices (Derwent & Mander, 1997; Kemp, 2000).  
 
Along with human-induced causes, there are numerous biological factors that play a role in 
the decline of population numbers as can be seen from exploring their slow reproductive rate 
as a result of slow sexual maturity and their large territories. Furthermore, the ideal habitat 
site for a SGH requires numerous factors in order for it to breed regularly (Vernon, 1984). 
The birds’ natural breeding cavities for nesting are usually limited, so breeding site selection 
usually takes a lot of time, hence contributing to the slow reproductive rate (Kemp & Kemp, 
1991).  
 
1.1.5 Previous Southern Ground Hornbill reintroduction attempts 
 
A SGH Working Group was established in 2003 and has compiled a Species National 
Recovery Plan (SNRP) in 2011 which includes strategies, relevant fieldwork techniques, 
application of research activities and investigations into the threats to the species and its 
habitat. Reintroduction is considered to be a viable option that needs further and more intense 
research attention. 
 
Previous attempts to reintroduce hand-reared SGHs back into the wild have had mixed 
results. The first number of attempts to reintroduce hand-reared SGH chicks took place 
between 1999 and 2002 at Mabula Game Reserve (see Mabula website for more information 
[http://www.ground-hornbill.org.za]). More recent reintroductions attempts have taken place 
since then, still exhibiting varied results with different natural- and human-induced causes for 
failure and success. These reintroduction projects proved to be highly labour intensive as 
reintroduced groups need individual experience and training as juveniles for their survival in 
the wild. Constant guarding against predators and supplementary feeding is usually carried 
out by a shepherd after the release. The use of shepherds has proved successful in the sense 
that the young birds are led along a specific route for a two-day period, with them finally able 
to walk the same route staying within the boundaries for up to two years. The shepherding 
unfortunately also meant that the birds became habituated to humans which could cause 
potential problems for the birds in future. In order for reintroductions to be successful, the 
9 
 
elimination of possible poisoning, excessive human contact, feasible causes of injury and 
inadequate housing are to be attended to (Kemp et al, 2007).  
 
A series of releases subsequently followed in the Limpopo, Eastern Cape and North West 
Provinces between 2003 and 2009. One group of hand-reared birds from Mabula was 
released in the Marakele National Park and flew back to Mabula within three months. Many 
of the released juvenile birds found it difficult to adapt to their new environments, especially 
near human settlements, and preferred the company of humans to their wild counterparts. In 
other cases, the members of the group were lost due to poisoning or human error or died 
because of Newcastle’s disease. An ongoing programme to harvest redundant, second-
hatched SGH chicks from the KNP and surrounding Associated Private Nature Reserves 
(APNR) in the Limpopo and Mpumalanga Province is currently undertaken by the Mabula 
Ground Hornbill Action Group, and attempts to further develop and perfect a reintroduction 
protocol for SGHs are underway.  
 
1.1.6 Aim and objectives of the study 
The aim of this study was to identify and verify optimal reintroduction sites for the SGH, 
specifically in the Musina area in the Limpopo Province. Five main objectives were set out to 
be determined, namely: 1) to identify optimal reintroduction areas, 2) to develop a habitat 
profile of the designated area, 3) to determine the availability of prey items, 4) to determine 
the availability of large trees needed for nesting and roosting and 5) to evaluate pre-release 
sites by involving a number of stakeholders.  
 
1.1.7 Importance of the study 
The identification of optimal reintroduction sites is of utmost importance, due to the fact that 
SGH reintroduction projects are an expensive and time-consuming process that requires a lot 
of manpower and resources. It will contribute to the conservation of the SGH in South Africa 
by aiming to eliminate problems regarding the four basic prerequisites for feasible SGH 
reintroductions, namely basic environmental variables, the availability of food and the 
availability of trees for nesting and roosting, as well as the attitudes of relevant stakeholders 
regarding SGH reintroduction.  
 
The survival of the SGH is of vital importance as it plays an important role in the functioning 
of the ecosystem. Being one of the so-called ‘Big 6’ birds in South Africa, the SGH is of 
10 
 
great economic importance in the eco-tourism industry and is also a flagship species for the 
savannah biome and, therefore, acts as an important indicator species (Engelbrecht et al, 
2007; Wilfred, 2007).  With the ever increasing human population and subsequent increasing 
number of disturbed natural habitat, it is important to be able to identify areas which are 
suited for the SGH and to be able to distinguish between different habitat types and different 
niches during different seasons. The preferred habitat type during different seasons and the 
types of habitat that allow for the greatest prey abundance are also of great significance. This 
study can be a first step towards providing a platform for future reference when in the process 
of reintroducing the SGH back into the wild, whether it be in protected areas or not.  
 
11 
 
 
CHAPTER 2 
 
STUDY SITE, MATERIALS AND METHODS 
 
12 
 
 
CHAPTER 2: STUDY SITE, MATERIALS AND METHODS 
 
This chapter gives a brief description of the study site and the methods and materials used. 
 
2.1 DESCRIPTION OF STUDY SITE 
2.1.1 Location 
The study was conducted between January and November 2011 on the game farm Greyghost 
Safaris, approximately 30 km west of Mucina at 22º 17.77’ S latitude and 29º 47.05’ E, south 
of the road to Mapungubwe National Park (Fig. 2.1). The farm is approximately 2000 ha in 
size. 
 
 
Figure 2.1: Topocadastral map of Greyghost Safaris on the Farm Ludwigslust 22º 17.77’ S 
latitude and 29º 47.05’ E longitude (Van Rooyen, 2008) 
 
 
13 
 
2.1.2 Climate 
2.1.2.1 Rainfall  
According to Van Rooyen (2008), the mean annual rainfall in the area is 348 mm (min = 155 
mm and max = 539 mm) according to readings from 12 stations in and around Musina. 
However, looking at the rainfall specifically for Greyghost Safaris from 2008 to 2011 (table 
2.1), it is clear that in 2010-2011, the highest rainfall compared to the other years occurred. 
The months October to February are predominantly the rainy season (wet), with the driest 
months occurring from May to September. The rainy season can deliver up to 83% of the 
annual rainfall.   
 
Table 2.1: Rainfall readings for Greyghost Safaris from the year 2008-2011. 
Season 
2008/2009 
 
2009/2010 
 
2010/2011 
 
2011/2012 
October 0 40.75 15 30.5 
November 60.25 130.75 60  
December 168.5 25.5 85.5  
January 199 30 355.5  
February 31.5 25.5 0  
March 9.5 0 10  
Total 468.75 252.50 526 30.5 
 
 
 2.1.2.2 Temperature 
At the station Macauville (22º 16’ S; 29º 54’ E), an annual mean temperature for a 29-year 
period was recorded as 22.5ºC, with an extreme maximum of 43.5ºC and an extreme 
minimum of -3.8ºC. A mean daily maximum temperature for the area was recorded as 33.5ºC 
during January and 24.9ºC during July. The mean minimum daily temperature during January 
is 21.3ºC and 6.9ºC during July. Frost may occur from June to August. The area’s highest 
relative percentage air humidity occurs from February to July, resulting in dew precipitation 
in the mornings. During December, January and February the area has the highest cloud 
cover, while the least cover occurs during July, August and September.  
 
2.1.3 Geology 
According to Van Rooyen (2008), the geology of the farm consists of predominantly 
sedimentary coarse-grained sand of the Quaternary System (Qs) and different types of gneiss 
(granite). However, on the surface of the ground in the central-western part of the farm, 
quartzite and calc-silicate can be found. That specific area is comprised mainly of meta-
14 
 
anorthosite metagabbro of the Messina suite. In the south-western corner of the farm, one can 
find a small section comprising of dark to light-grey or pink porphyroblastic biotite gneiss 
(Alldays gneiss). Within the north-eastern corner, leucocratic quartzo-feldspatic gneiss and 
metapelite of the Malala Drift Group of the Beit Bridge Complex can be found. 
Porphyroblastic gneiss (Bulai gneiss) in the form of rocky outcrops can be seen on the south-
eastern corner.  
 
2.1.4 Vegetation  
According to Acocks (1953) and Low and Rebelo (1996), Greyghost Safaris is located in an 
area of Mopane Veld and Mopani Bushveld. More recent studies, however, conducted by 
Mucina and Rutherford (2006) described the area as being dominantly Musina Mopani 
Bushveld, with a small section of Limpopo Ridge Bushveld in the south-eastern corner of the 
farm (Fig. 2.2).  
 
 
Figure 2.2: Musina Mopani Bushveld and Limpopo Ridge Bushveld forming the main 
vegetation types for Greyghost Safaris (Mucina & Rutherford, 2006). 
 
 
 
15 
 
2.1.4.1 Musina Mopane Bushveld  
According to Mucina & Rutherford (2006), characteristics of the Musina Mopane Bushveld 
include undulating and uneven plains with a number of scattered hills with altitudes reaching 
up to approximately 600 m. Situated near the Beit Bridge Complex, Greyghost Safari’s 
geology consists mainly of gneisses and metasediments, rocks having undergone 
metamorphism and intense deformation. This then leads to the various types of soil which 
can be found, ranging from deep red/brown clays to freely-drained sandy soils as well as 
shallow Glenrosa and Mispah soils often formed as a result of rock weathering and clay 
alleviation (Rossouw & Van der Waals, 2010).  
 
Trees commonly found in the area include the Mopane (Colosphospermum mopane), baobab 
(Adansonia digitata), knobthorn (Acacia nigrensis), red bushwillow (Combretum 
apiculatum), slender three-hook thorn (Acacia senegal var. leiorhachis) and velvet corkwood 
(Commiphora mollis).  
 
Some of the more conspicuous small trees and shrubs include the white raisin (Grewia 
bicolor), Lowveld cluster leaf (Terminalia prunioides), stink shepherds tree (Boscia foetida 
subsp. rehmanniana) and velvet raisin (Grewia flava). Grass species commonly found 
include bushveld signal grass (Urochloa mosambicensis), grey sour grass (Enneapogon 
cenchroides), false signal grass (Brachiaria deflexa), three-awn (Aristida sp.), sand quick 
(Schmidtia pappophroides) and silky bushman grass (Stipagrostis uniplumis).  
 
2.1.4.2 Limpopo Ridge Bushveld (Mucina & Rutherford, 2006) 
Making up the south-eastern corner of the farm, the Limpopo Ridge Bushveld is 
characterised as having more irregular hills and ridges which reach up to altitudes of between 
300 to 1000 m. The area consists of poorly developed ground resulting in a moderate 
savannah vegetation structure. The soils found here are more calcareous gravel and calc-
silicate soils, which provide perfect growing environments for the white syringa (Kirkia 
acuminata) and the baobab (Adansonia digitata). Thriving in the calcareous gravel, the shrub 
Catophractes alexandri is dominant. Near the Mapungubwe National Park, more sandstone 
of the Clarens Formation can be seen.  
 
The vegetation structure found here is primarily in a frost-free area. Primary species of trees 
found include the before-mentioned baobab, marula (Sclerocaya birrea), mopane 
16 
 
(Colophospermum mopane), tall common corkwood (Commiphora glandulosa), Lowveld 
cluster-leaf (Terminalia prunoides), shepherd’s tree (Boscia albitrunca) and various fig 
species (Ficus spp.).  
 
2.1.5 Plant Communities (Van Rooyen, 2008) 
Within these vegetation structures on Greyghost Safari are 11 identified plant communities, 
namely (Fig. 2.3): 
 
1. Umbrella thorn-bushveld signal grass old fields 
2. White syringa-red bushwillow bushveld 
3. Red bushwillow bushveld 
4. White syringa-mopane-Lowveld cluster-leaf bushveld 
5. White syringa-velvet corkwood bushveld 
6. Mopane-Lowveld cluster-leaf dense bushveld 
7. Mopane-slender three-hook thorn low dense bushveld 
8. Apple-leaf-umbrella thorn drainage lines 
9. Rocky outcrops 
10. Disturbed areas 
11. Mopane erosion patches 
 
17 
 
 
Figure 2.3: Different plant communities found on Greyghost Safaris (Van Rooyen, 2008). 
 
 Three plant communities were selected for the study, namely white syringa-mopane-lowveld 
cluster-leaf bushveld (4), mopane-slender three-hook thorn low dense bushveld (7) and 
umbrella thorn-bushveld signal grass old fields (1) and will, therefore, be discussed in more 
detail. These sites were selected based on where the SGHs have frequently been observed by 
the landowners and the farm labourers during the year 2010. The afore-going sites will be 
referred to as site A, B and C respectively.  
 
2.1.5.1 Site A: White syringa-mopane-Lowveld cluster-leaf bushveld (4) (Van 
Rooyen, 2008) (Kirkia acuminata-Colophospermum mopane-Terminalia prunoides 
bushveld) 
Most of the western part of the farm is found to be comprised of the white syringa-mopane-
lowveld cluster leaf bushveld, covering an area of approximately 706 ha. This part of the 
farm is relatively flat with occasional slopes of 3º with surface rocks covering 20% of the 
area. The soils are more deep red loamy to deep yellow-red to sandy clay-loam soils which 
are formed as a result of the gneiss.  
18 
 
Diagnostic tall trees that are associated with this plant community include baobab, white 
syringa, velvet corkwood, false marula and the slender three-hook thorn. Other trees 
commonly found are the Lowveld cluster-leaf, blue thorn (Acacia erubescens), mopane 
(Colophospermum mopane), umbrella thorn (Acacia tortilis), shepherd’s tree (Boscia 
albitrunca), common star chestnut (Sterculia rogersii) and the common corkwood 
(Commiphora glandulosa).  
 
Characteristic shrubs in this community include raisin bushes, glossy-leaved corkwood 
(Commiphora schimperi) and sickle bush (Dichrostachys cinerea). The average cover of 
shrubs is approximately 12%. The grass layer, on the other hand, covers on average 52% of 
the area and includes species such as the grey sour grass (Enneapogon cenchroides), false 
signal grass (Brachiaria deflexa), silky bushman grass (Stipagrostis uniplumis), Lehman’s 
love grass (Eragrostis lehmanniana), annual three-awn (Arristida adscensionis), carrotseed 
grass (Trachus berteronius) and flaccid finger grass (Digitaria velutina). Less commonly 
found are the Natal red top (Melenis repens), bushveld signal grass (Urochloa 
mosambicensis), blue buffalo grass (Cenchrus ciliaris) and large-seeded three-awn (Aristida 
rhiniochloa).  
 
The most commonly found forbs in this plant community, covering 12% of the area, include 
Euphorbia neopolycnemoides, Amaranthus praetermissus, Dicoma tomentosa, Geigeria 
acaulis, Ptycholobium contortum, Pavonia burchellii, Crotalaria sphaerocarpa, Vigna 
frutenscens, Hermannia modesta, Achyranthus aspera, Hibiscus sidiformis, Hermannia, 
boraginiflora, Tephrosia purpurea, Monsonia senegalensisAcalyphya indica, Kyphocarpa 
angustifolia, Evolvulus aslinoides and Solanum kwebense.  
 
2.1.5.2 Site B: Mopane-slender three-hook thorn low dense bushveld (7) (Van 
Rooyen, 2008) (Colophospermum mopane-Acacia Senegal var. Leiorhachis dense 
bushveld) 
Towards the north-west corner of the farm is a slightly more hilly area with a rocky low 
dense surface. This area makes up approximately 412 ha and is characterised by calcrete 
gravel and rocks. These rocks make up approximately 40% of the 412 ha. The slightly hilly 
surface slopes up to about 3º.  Due to the high lime content in the soil, the majority of the 
soils found here are more whitish shallow loam soils. 
 
19 
 
This type of plant community can be classified by taking note of the most diagnostic species, 
namely the eight-day grass (Enneapogon desvauxii), blue bushman grass (Stipagrostis 
hirtiglumma), Lantana rugosa, Tribulus zeyheri, Cleome augustiflora and Commicarpus 
pentadrus.  
 
The main tall tree species (>6 m) are the knobthorn, the shepherd’s tree and the slender three-
hook thorn. The mean canopy cover is 2% with a tree occurrence density of approximately 4 
individuals/ha. The average tree density (<6 m) is 112 individuals per ha and covers 13% of 
the area. Trees such as the tall trees mentioned above together with the mopane, umbrella 
thorn, Lowveld cluster-leaf and the red bushwillow make up this community. There are 
numerous shrubs found here, making up 18% of the area, namely stink shepherd’s tree, white 
raisin, sickle bush, blue sourplum (Ximenia Americana), white-berry bush (Flueggia virosa) 
and the knobbly creeper (Combretum mossambicense). The density of the shrubs makes up 
more of the area than do the trees, with a mean density of between 704 individuals/ha to 1110 
individuals/ha.  
 
Because of the high density rock cover in the area, the grass layer is relatively low, covering 
on average 41%. Characteristic species of this plant community include grey sour grass, 
Lehman’s love grass, blue buffalo grass and carrotseed grass.  
 
Common forbs found in this plant community include Hermbstaedtia odorata, Corchorus 
asplenifolius, Phyllanthus angolensis, Kohautia caespitose, Acalypha indica, Monechma 
divaricatum and Chascanum pinnatifidum.  
 
2.1.5.3 Site C: Umbrella thorn-bushveld signal grass old fields (1) (Van Rooyen, 
2008) 
 (Acacia tortilis-Urochloa mosambicensis old fields) 
This type of plant community consists of old ploughed fields set in the deep red sandy-clay-
loam soils of the gneissic plains. A number of vegetation species found here characterise this 
type of plant community, namely Solanum incanum, Alternanthera pungens and the common 
crowfoot (Dactyloctenium aegyptium). This area has been found to have a lack of tall trees 
(between 3 and 600> m) but, instead, has a 3% cover of shrubs, including the mopane and the 
umbrella thorn. As a result of previous cultivation practices and heavy grazing, grass cover is 
only approximately 15% in this area. Common grass species found, however, include the 
20 
 
common crowfoot, spreading three-awn (Aristida congesta subsp. barbicollis), false signal 
grass, annual three-awn, grey sour grass, carrotseed grass, Lehmans love grass and the 
bushveld signal grass. Forbs such as Althernanthera pungens, Solanum incanum, Geigeria 
acaulis, Waltheria indica, Indigofera vicioides, Limeum sulcatum and Heliotropium steudneri 
cover approximately 8% of the basal cover.  
 
2.2 METHODS AND MATERIALS 
2.2.1 Selection of optimal reintroduction sites 
One of the primary determinants for reintroduction success is habitat quality/suitability 
(Moorhouse et al, 2009; Tavecchia et al, 2009; Michel et al, 2010; Hebblewhite et al, 2011). 
However, these habitats (specific sites) first need to be identified and later verified. One 
option is to use a niche-based modelling technique that describes the suitability of a particular 
habitat (“ecological space”) in terms of its environmental attributes and then to project this 
onto another geographical space (Phillips et al, 2006). Maxent 3.3.3e was used which is 
obtainable from http://www.cs.princeton.edu/~schapire/maxent. Maxent is a niche-based 
model as it represents an approximation of a species’ environmental niche in the examined 
environment. It uses a predictive modelling approach that allows for the estimation of the 
target distribution of a species based on environmental conditions at sites where the species is 
known to occur (Phillips et al, 2004, 2006).  
 
A species’ fundamental niche is all those factors and conditions that enable it to survive, 
while its realized niche is that subset of the fundamental niche which in reality it occupies 
(Hutchinson 1957). This realized niche will often be smaller that the fundamental niche due 
to external influences such as human activities and geographic barriers.  
 
If assumed that occurrence localities, as discussed above, were extracted from source habitat 
(habitat which meets the requirements of the species to survive), the environmental 
conditions at these localities should be made up of samples of the realised niche (Phillips et 
al, 2006). A niche-based model will then represent an approximation of the species’ realised 
niche within the environmental variables being studied (Phillips et al, 2006). The origins of 
Maxent lie in statistical mechanics and are based on the maximum entropy principle 
discussed by E.T. Jaynes (Jaynes 1957; Phillips et al, 2006). It uses machine learning 
21 
 
techniques to find maximum entropy distribution and, thus, habitat suitability (Phillips et al, 
2004, 2006).  
 
For this reason, it was decided to follow a four-step process to identify optimal reintroduction 
sites in the Limpopo area. Figure 2.4 depicts the basic process that was followed. 
 
 
Figure 2.4: Illustration of the four-step process for identifying optimal reintroduction sites in 
the Limpopo Province. 
 
Step 1: Geographic Information Science (GISc) in combination with the Maximum Entropy 
(MAXENT) algorithm in the “openModeller” software application was first used to evaluate 
and compare various environmental datasets at all available SGH sighting points within the 
Kruger National Park with a view to compiling a basic profile. The following environmental 
variables from three different classes, namely climate, topography and habitat, were used: 
? Average maximum temperature (tmax) 
? Average minimum temperature (tmin) 
? Average rainfall (rain) 
? Elevation (dem) 
? Slope (slope) 
? Vegetation (veg) 
? Annual Natural Diversified Vegetation Index (ndvi) 
? MODIS tree cover (tree) 
? Soil types (soil) 
 
All datasets were re-sampled to a cell size of 1000 meters and projected on the Albert Equal 
Area Projected Coordinate system after which they were converted to ASCII Grids. Land 
cover was not used as some of the occurrence records were older than the available national 
land-cover dataset and would not be suitable for modelling purposes (Phillips et al, 2006). 
 
22 
 
Step 2: This profile was then extrapolated to the rest of the Limpopo Province to identify 
specific sites with a similar profile as the sites in the Kruger National Park.  
 
Step 3: These results were then compared to a control set of sighting data for the Limpopo 
Province to determine the accuracy of the projection.  
 
Step 4: Finally, the results were then transferred from “openModeller” to ArcInfo for 
mapping purposes. 
 
2.2.2 Methods to determine habitat profile (floristic survey) 
In order to create a habitat profile, the ground cover and species composition was determined 
with regard to forbs and grasses and the woody composition. These parameters were sampled 
four times per year in order to determine any changes. During each survey, a 4 x 50 m cross 
transect was used to record the frequency (point to tuft distance), ecosystem integrity (at each 
1 m interval the presence of erosion, rock, litter or bare patches are recorded) and woody 
vegetation (Sutherland, 2000:42). Measurements and samples were taken at 1 m intervals 
starting 10 m from the centre point so as to eliminate the possibility of oversampling. The 
frequency and basal cover of the herbaceous layer was recorded by identifying the 
herbaceous species that is closest to the point (descending point method) (Sutherland, 2000) 
and calculating the number of hits of herbaceous species (Fig. 2.5).  
 
For the woody component, the species name of the tree is given, accompanied by the 
frequency of hits during the four months that sampling was done and the size of the tree. The 
sizes of the trees were categorized into four classes as follows: small (up to 5 m), medium-
sized (5-15 m), large (15-30 m), and very large, (more than 30 m). 
 
 
23 
 
 
Figure 2.5: Diagram illustrating 4 x 50 m cross transect with measurements and samples 
taken at 1 m intervals starting 10 m from the centre point. 
 
2.2.3 Methods to determine availability of prey items (faunal survey) 
2.2.3.1 Methods to identify and quantify (frequency and volume) of prey items  
A nested cross-array method (Perner & Schueler, 2004: 467-477) was used to survey 
invertebrate prey. A total of 13 pitfall traps – one in the centre, at 5 m, 15 m and 35 m – were 
used at each site (Fig. 2.6). The traps were constructed from 15 cm-long drainpipes, with a 
diameter of 20 cm, with a plastic cup inside it that is lowered into the ground and levelled 
with the surface. The cups were filled up to approximately 2 cm with non-hazardous 
propylene glycol. The invertebrate sampling was done for four consecutive days four times 
during the year 2011. Two 200 m sweep-net surveys were conducted per day at each site to 
capture low-flying invertebrates found in the grasses and forbs. All captured invertebrates 
were preserved in containers containing 70% alcohol. Invertebrates were sorted and identified 
up to order level as far as possible and classified into size classes so as to measure the volume 
using the volumetric water-displacement method (Magnusson, 2003).  
 
50m 
 
10m 
24 
 
 
Figure 2.6: Diagram illustrating the 4 x 50 m cross-transect and the spacing between each 
pitfall trap. 
 
The statistical methods used to analyse all the data for the vegetation and the prey items, a 
non-parametrical test was carried out using the statistical package Friedman ANOVA (Field, 
2009). 
 
2.2.4 Methods to determine the availability of trees for nesting and roosting (presence 
and spacing/ density and distribution of suitable trees) 
High-resolution, natural-colour aerial imagery was used in order to determine the availability 
of trees for nesting and roosting. Firstly, 50 large baobab trees in and around the area were 
plotted using a GPS and then read into ArcMap 10. Twelve of the points were held back so as 
to be used for testing purposes. The remaining points were used to identify large trees on 
aerial photographs of the farm used for the study by means of an Image Analysis tool in 
ArcMap 10. Once the trees were identified on the aerial photographs, all other aspects were 
identified with the same Image Analysis tool in ArcMap 10, such as dams, small bushes and 
roads. Once all this information was extracted, all the large trees were extracted specifically, 
with the use of Interactive Supervised Classification. Included here were all the surface areas 
larger than 100 m² which were derived from the sizes of the trees plotted with the GPS. In 
order to test the effectiveness of the results, the remaining 12 plots held back at the 
35m 15m 5m 
25 
 
commencement of the process were used. Ten out of the 12 points were revealed, thus 
indicating that there is an 83.3% certainty that the results are satisfactory.  
 
2.2.5 Pre-release site evaluations/ stakeholder engagement 
Ten interviews were conducted with local landowners in the area by means of a semi-
structured questionnaire in order to determine the attitudes of local landowners regarding the 
reintroduction of SGHs on their farms or in the surrounding area. Interviews are useful to 
gather data in a person’s own, verbalised words. The benefits of using interviews as a data 
gathering method are that it permits immediate follow-up questions that can be used for 
clarifications and can, thus, be regarded as a flexible mode of data collection that not only 
contributes to the trustworthiness of the data but the general credibility of the study.  An 
example of the questionnaire can be found in the Appendix 2.  
 
26 
 
 
CHAPTER 3 
 
RESULTS 
27 
 
 
CHAPTER 3: RESULTS 
This chapter gives a brief overview of the results for the identification of the optimal 
reintroduction sites, habitat profile, availability of prey items, availability of trees for nesting 
and roosting, and the evaluations of the pre-release sites and stakeholder engagement. 
 
3.1 OPTIMAL REINTRODUCTION SITES 
As illustrated by the map in Figure 3.1, the areas in bright red to orange are areas more suited 
for SGH according to the biophysical aspects of its required habitat (results of Step 2). As can 
be seen from the map, high priority areas can be found all the way from south of Thabazimbi, 
stretching across to Ellisras (Lephalale), and all the way past Alldays along the Limpopo 
River Valley, north of the Soutpansberg. The area on the north-eastern and eastern border of 
the Limpopo and into the KNP also appear to be suitable. Other areas which are also suitable 
are south of Louis Trichardt, to the east of Tzaneen, stretching all the way to the western 
border of the KNP. Except for smaller pockets around Roedtan, Marble Hall and Groblersdal, 
the central parts of the province appears to be less suitable for re-introduction.  
 
Figure 3.1: Map illustrating the areas that are from least to most probable areas of optimal 
habitat for the SGH in the Limpopo Province 
28 
 
 
Figure 3.2: Map of farmland areas in the Musina, Pondrif and Alldays vicinity. The farms in 
red are those that were identified as optimal sites. 
According to the map, 14 farms were identified as meeting the optimal habitat requirements 
for the SGH (Figure 3.2, results of step 4), namely Denstaat, Verbaard 53 MT, Samaria, 
Greefswald, Kongo, Ceon, Coventry, Dresden, Leeds, Frey, Tovey, Oostenryk, Verbaard, 
Faure and, finally, Ludwigslust. As mentioned before, Greyghost Safaris forms part of the 
larger Ludwigslust. All these farms occur in the Pontdrif, Alldays and Musina region. This 
was subsequently verified by recent reported SGH sightings on some of the farms.  
 
3.2 HABITAT PROFILE  
 
3.2.1 Species composition 
(Appendix 1) 
 
3.2.2.1 Site A 
A total number of 13 species, including grass and forbs, were recorded at site A (see Table 
3.1).  
 
29 
 
 
Table 3.1: Species list of forbs and grasses for Site A 
Aristida adscensionis 
Cenchrus ciliaris 
Endostemon tereticaulis 
Eragrostis lehmanniana 
Evolvulus alsinoides 
Geigeria acaulis 
Geigeria burkei 
Heliotropium nelsonii 
Hermannia modesta 
Ocimum americanum 
Panicum deustum 
Waltheria indica 
Xyrophyta humulis 
 
Table 3.2: Basic descriptive statistics for forbs and grasses for Site A (STATISTICA 10) 
 
Variable Mean Median Minimum Maximum Standard deviation 
January 6.1 6.000000 0 33 9.0 
April 8.3 4.000000 0 40 14.4 
August 9.2 1.000000 0 66 18.9 
October 5.5 0.000000 0 32 9.4 
 
in January, the average number of hits was 6.1 with a minimum number of 0 and a maximum 
number of 33 out of the maximum possible number of hits of 200 calculated from the 4 x 50 
m cross-transect. The standard deviation for January is 9.0. In April, the average number of 
hits was 8.3 with a minimum number of 0 and maximum number of 40 (standard deviation = 
14.4). In August, the highest average hits were obtained with 9.2 hits and a minimum of 0 
and maximum of 66 (standard deviation = 18.9). October had the lowest number of hits with 
5.5 and a minimum of 0 and a maximum of 32 (standard deviation = 9.4).  
 
The woody component of Site A consists of six different tree species (Table 3.3).  
 
 
 
 
 
30 
 
Table 3.3: Species list, frequency, and size of the woody structure for Site A: small (up to 5 
m), medium-sized (5-15 m), large (15-30 m) and very large (more than 30 m). 
 
Species Frequency Height (Class) 
Acacia senegal 5 Medium 
Colophospermum mopane 27 Medium-large 
Combretum apiculatum 9 Small-medium 
Grewia bicolour 19 Small 
Grewia flava 4 Small 
Kirkia acuminate 6 Medium 
 
As depicted by Table 3.3, there was a total count of five Acacia senegal (medium) species at 
Site A. Colophospermum mopane (medium-large) had a total count of 27 trees, Combretum 
apiculatum (small-medium), 9 trees, Grewia bicolor (small), 19 trees, grewia flava (small), 
four trees and Kirkia acuminate (medium), six trees. The dominant specie found at Site A is 
Colophospermum mopane. 
 
 3.2.2.2 Site B 
A total number of 25 species including grass and other forbs were recorded at site B (see 
Table 3.4).  
 
Table 3.4: Species list of forbs and grasses for Site B 
Aristida adscensionis 
Cenchrus ciliaris 
Corchorus asplenifolius 
Endostemon tereticaulis 
Eragrostis lehmanniana 
Heliotropium nelsonii 
Hermbstaedtia fleckii 
Hermannia modesta 
Hermbstaedtia odorata 
Indigofera ingrate 
Indigofera trita 
Indigofera vicioides 
Microcharis galpinii 
Monechma divaricatum 
Ocimum americanum 
Panicum deustum 
Phylanthus angolensis 
Seddera capensis 
31 
 
Sida chrysantha 
Stipagrostis uniplumis 
Syncolostemon canescens 
Tephrosia semiglabra 
Tragus berteronianus 
Tragus racemosus 
Waltheria indica 
 
Table 3.5: Basic descriptive statistics for forbs and grasses for Site B (STATISTICA 10) 
 
Variable Mean Median Minimum Maximum Standard deviation 
January 4.9 1 0 32 7.9 
April 3.8 0 0 19 6.2 
August 3.7 0 0 21 6.7 
October 2.5 0 0 22 6.1 
 
As can be seen from the basic descriptive statistics calculated for Site B (Table 3.5), it is clear 
that in January, the highest average number of hits was obtained with a number of 4.9 and a 
minimum number of 0 and a maximum number of 32 (standard deviation = 7.9). In April, the 
average number of hits was an average of 3.8, with a minimum number of 0 and maximum 
number of 19 (standard deviation = 6.2). In August, an average of 3.7 hits was obtained with 
a minimum of 0 and maximum of 21 (standard deviation = 6.7). October, once again, had the 
lowest number of hits with an average of 2.5, minimum= 0 and a maximum=22 (standard 
deviation = 6.1).  
 
The woody component of Site B consists of nine different tree species (Table 3.6). The 
species name of the tree is given, accompanied by the frequency of hits during the four 
months that sampling was done, the height of the tree and the width of the canopy structure.  
 
Table 3.6: Species list, frequency and size of the woody structure for Site B: small (up to 5 
m), medium-sized (5-15 m), large (15-30 m) and very large (more than 30 m). 
 
Species Frequency Height (Class) 
Acacia nigrensis 9 Large 
Acacia senegal 14 Small-medium 
Acacia tortilus 10 Medium-large 
Colophospermum mopane 38 Medium-large 
32 
 
Dichrostachys cinerea 6 Small-medium 
Dombeya rotundifolia 1 Small-medium 
Grewia bicolour 46 Small 
Grewia flava 26 Small 
Terminalia prunoides 17 Small-medium 
 
As depicted by Table 3.6, there was a total count of nine Acacia nigrensis (large), 14 counts 
of A. senegal (small-medium), 10 counts of A. tortilus (medium-large), 38 counts of 
Colophospermum mopane (medium-large), six counts Dichrostachys cinerea (small-
medium), one count of Dombeya rotundifolia (small-medium), 46 counts of Grewia bicolor 
(small) , 26 counts of Grewia flava (small), and 17 counts of Terminalia prunoides (small-
medium). The dominant specie found at Site B is Grewia bicolor.  
 
 3.2.2.3. Site C  
A total number of 13 species including grass and other forbs species were recorded at site C 
(see Table 3.7).  
 
Table 3.7: Species list of forbs and grasses for Site C 
Acanthosperma hispidum 
Alternantha pungens 
Aristida adscensionis 
Boerhavia erecta 
Brachiaria deflexa 
Endostemon tereticaulis 
Eragrostis lehmanniana 
Geigeria acaulis 
Panicum deustum 
Tragus berteronianus 
Tribulus terrestris 
Uruchloa pancoides 
Waltheria indica 
 
Table 3.8: Basic descriptive statistics for forbs and grasses for Site C (STATISTICA 10) 
 
Variable Mean Median Minimum Maximum Standard deviation 
January 7.1 5 0 22 8.0 
April 8.7 0 0 52 15.8 
August 5.7 0 0 46 13.0 
33 
 
October 2 0 0 13 4.2 
 
As can be seen from the basic descriptive statistics calculated for Site C (Table 3.8), it is clear 
that in January, the average number of hits was calculated as 7.1 with a minimum number of 
0 and a maximum number of 22 (standard deviation = 8.0). In April, the highest average 
number of hits was calculated as 8.7, with a minimum number of 0 and maximum number of 
46 (standard deviation = 15.8). In August, an average of 5.7 hits was obtained with a 
minimum of 0 and maximum of 46 (standard deviation = 13.0). October, once again, had the 
lowest number of hits with an average of 2 hits, minimum= 0 and a maximum=13 (standard 
deviation = 4.2).  
The woody component of Site C consists of five different tree species. This can be seen in 
Table 3.9. The species name of the tree is given, accompanied by the frequency of hits during 
the four months that sampling was done, the height of the tree and the width of the canopy 
structure.  
 
Table 3.9: Species list, frequency and size of the woody structure for Site C: small (up to 5 
m), medium-sized (5-15 m), large (15-30 m) and very large (more than 30 m). 
 
Species Frequency Height (Class) 
Acacia tortilus 2 Medium-large 
Colophospermum mopane 7 Medium-large 
Combretum apiculatum 2 Small-medium 
Grewia bicolour 4 Small 
Terminalia prunoides 2 Small-medium 
 
As depicted by Table 3.9, there was a total count of two A.tortilus (medium-large), seven 
counts of Colophospermum mopane (medium-large), two counts of Combretum apiculatum 
(small-medium), 4 counts of Grewia bicolor (small), and two counts of Terminalia prunoides 
(small-medium). The dominant tree species for Site C is Colophospermum mopane.  
 
3.2.2 Change in ground cover 
3.2.2.1 Site A 
In Table 3.10, below, the mean change in ground cover is given for Site A; the highest 
occurring in August. Following August, April had the second highest mean change in ground 
cover with a mean value of 8.3. January and October followed after with the values of 6.5 and 
5.5, respectively. According to the p-value, which is bigger than 0.1 (0.49039) there is no 
34 
 
statistical significance for the differences between the four months regarding change in 
ground cover for Site A.  
 
Table 3.10: Friedman ANOVA analysis for determining the average change in ground cover 
for Site A (p = 0.49039). 
 
Variable Mean  Standard deviation 
January 6.5 9.0 
April 8.3 14.4 
August 9.2 18.8 
October 5.5 9.4 
 
 3.2.2.2 Site B 
In Table 3.11, below, the average change in ground cover is given for Site B for each month. 
The highest average change in ground cover occurred in January, with a mean value of 4.9. 
Following January, April had the second highest change in ground cover with a value of 3.8. 
August has the third highest change in ground cover with a value of 3.7. The least average 
change in ground cover occurred in October with a value of 2.5. According to the p-value, 
which is bigger than 0.1 (0.20797) there is no statistical significance for the differences 
between the four months regarding change in ground cover for Site B.  
 
Table 3.11: Friedman ANOVA analysis for determining the average change in ground cover 
for Site B (p = 0.20797). 
 
Variable Mean  Standard deviation 
January 4.9 7.9 
April 3.8 6.2 
August 3.7 6.7 
October 2.5 6.1 
 
 3.2.2.3 Site C 
In Table 3.12, below, the average change in ground cover is given for Site C for each month. 
The highest average change in ground cover occurred during April, with a value of  8.7. 
Following April, January had the second highest average change in ground cover with a value 
of 8.7. August has the highest average change in ground cover with a value of 5.7. The least 
35 
 
change in ground cover occurred in October with a value of 2.0. According to the p-value, 
which is smaller than 0.1 (0.06456) there is a statistical significance for the differences 
between the four months regarding change in ground cover for Site C.  
Table 3.12: Friedman ANOVA analysis for determining the average change in ground cover 
for Site C (p = 0.06456). 
 
Variable Mean  Standard deviation 
January 7.2 8.0 
April 8.7 15.8 
August 5.7 13.0 
October 2.0 4.2 
 
3.3 AVAILABILITY OF PREY ITEMS 
3.3.1 Species composition 
3.3.1.1. Site A 
Below is a Table (3.13) which portrays the species composition and number of species caught 
in both the pitfall and sweepnet traps during all four months at Site A. Species were identified 
up to order level as far as possible. One can see that in January, the dominant prey species 
available was Coleoptera with a number of 38 insects. In April and August, the same can be 
found with a number of 38 and 29 Coleoptera insects respectively. In October, the order 
Hemiptera was found to be the dominant order of insects. 
 
Table 3.13: Species composition and frequency of prey obtained from both pitfall and 
sweepnet traps for each month at Site A. 
 
Order  January April August October 
Blattodea 0 0 0 2 
Centipede 0 0 1 0 
Coleoptera 38 38 29 17 
Dermaptera 0 0 0 0 
Diptera 1 2 1 1 
Gastropoda 0 0 0 0 
Hemiptera 2 2 0 59 
Hymenoptera 2 4 5 16 
Isoptera 0 0 0 0 
Larvae 0 1 0 0 
Lepidoptera 2 6 0 2 
Mantodea 0 2 0 0 
36 
 
Millipede 0 0 0 12 
Neuroptera 0 0 0 0 
Orthoptera 17 6 0 1 
Phasmatodea 4 0 0 0 
Scorpiones 0 1 0 1 
Solifugae 0 1 1 4 
Araneae 7 7 3 5 
Thysanoptera 0 0 0 0 
Thysanura 0 0 0 0 
 
In Table 3.14, the descriptive statistics are given for the species compositions which were 
sampled in both the pitfall and the sweepnet traps during the four months for Site A. It is 
clear from referring to the Table that October had the highest average species composition 
with a number of 5.7 insects (min = 0; max = 59; standard deviation = 13.3). Second to 
October is January with an average of 3.5 species (min = 0; max = 38; standard deviation = 
8.8). April has the second lowest average species composition with a value of 3.3 species 
(min = 0; max = 38; standard deviation = 8.3). The month with the least number of species 
was August with a number of 1.9 insects (min = 0; max = 29; standard deviation = 6.3).  
 
Table 3.14: Descriptive statistics for the species composition sampled in both the pitfalls and 
sweepnet traps for Site A 
 
Variable Mean  Median  Minimum Maximum Standard deviation  
January 3.5 0 0 38 8.8 
April 3.3 1 0 38 8.3 
August 1.9 0 0 29 6.3 
October 5.7 1 0 59 13.3 
 
In Table 3.15, the average change in species composition for Site A for each month is 
depicted. The highest average change in species composition occurred in October, with a 
value of 5.7. Following October, January has the second highest average change in species 
composition with a value of 3.5. The second lowest change in species composition falls in the 
month of April with a value of 3.3. The least average change in species composition falls in 
August with a value of 1.9. According to the p-value, which is smaller than 0.1 (0.06656) 
there is a statistical significance for the differences between the four months regarding 
species composition for Site A.  
 
37 
 
Table 3.15: Friedman ANOVA analysis for determining the change in species composition 
for Site A (p = 0.06656). 
 
Variable Mean  Standard deviation 
January 3.5 8.8 
April 3.3 8.3 
August 1.9 6.3 
October 5.7 13.3 
 
 3.3.1.2. Site B 
Below is a Table (3.16) which portrays the species composition and number of species caught 
in both the pitfall and sweepnet traps during all four months at Site B. Species were identified 
up to order level as far as possible. One can see that in January, the dominant prey species 
available was Orthoptera with a number of 43 insects. In April, the dominant order is 
Coleoptera with a number of 112 insects. The same applies for August with Coleoptera being 
the dominant order with a number of 54. In October, the order Hemiptera was found to be the 
dominant order of insects. 
 
Table 3.16: Species composition and frequency of prey obtained from both pitfall and 
sweepnet traps for each month at Site B 
 
Order January April August October 
Centipede 0 0 1 3 
Coleoptera 6 112 54 5 
Dermaptera 0 0 0 0 
Diptera 2 1 0 0 
Gastropoda 0 0 0 0 
Hemiptera 9 2 0 14 
Hymenoptera 15 11 1 4 
Isoptera 0 0 0 0 
Larvae 15 0 0 0 
Lepidoptera 3 5 0 0 
Mantodea 5 0 0 1 
Millipede 1 0 0 3 
Neuroptera 0 0 0 0 
Orthoptera 43 4 3 2 
Phasmatodea 11 0 0 0 
Scorpiones 0 0 0 2 
Solifugae 0 0 0 0 
38 
 
Araneae 9 3 4 2 
Thysanoptera 0 0 0 0 
Thysanura 0 0 0 0 
 
In Table 3.17, the descriptive statistics are given for the species compositions which were 
sampled in both the pitfall and the sweepnet traps during the four months for Site B. It is 
clear from referring to the Table that April had the highest average species composition with 
a number of 6.6 insects (min = 0; max = 112; standard deviation = 24.3). Second to April is 
January with an average of 5.7 species (min = 0; max = 43; standard deviation = 10). August 
has the second lowest average species composition with a value of 3 species (min = 0; max = 
54; standard deviation = 11.7). The month with the least number of species was October with 
a number of 1.7 insects (min = 0; max = 14; standard deviation = 3.2).  
 
Table 3.17: Descriptive statistics for the species composition sampled in both the pitfalls and 
sweepnet traps for Site B 
 
Variable Mean  Median  Minimum Maximum Standard deviation  
January 5.7 1 0 43 10.0 
April 6.6 0 0 112 24.3 
August 3.0 0 0 54 11.7 
October 1.7 0 0 14 3.2 
 
In Table 3.18, the average change in species composition for Site B for each month is 
depicted. The highest average change occurred in April, with a value of 6.6. Following April, 
January has the second highest change in species composition with a value of 5.7. The second 
lowest change in species composition occurred in the month of August with a value of 3.0. 
The least average change in species composition occurred in October with a value of 1.7. 
According to the p-value, which is smaller than 0.1 (0.06656) there is a statistical 
significance for the differences between the four months regarding species composition for 
Site B.  
 
Table 3.18: Friedman ANOVA analysis for determining the change in species composition 
for Site B (p = 0.03725). 
 
Variable Mean  Standard deviation 
January 5.7 10.0 
39 
 
April 6.6 24.3 
August 3.0 11.7 
October 1.7 3.2 
 
 3.3.1.3. Site C 
 
Below is a Table (3.19) which portrays the species composition and number of species caught 
in both the pitfall and sweepnet traps during all four months at Site C. One can see in 
January, the dominant prey species available was Orthoptera with a number of 55 insects. In 
April, the dominant order is Coleoptera with a number of 52 insects. The same applies to 
August with Coleoptera being the dominant order with a number of 30. In October, the order 
Hemiptera was found to be the dominant order of insects. 
 
Table 3.19: Species composition and frequency of prey obtained from both pitfall and 
sweepnet traps for each month at Site C 
 
Order January April August October 
Blattodea 0 0 1 0 
Centipede 0 0 0 0 
Coleoptera 20 52 30 11 
Dermaptera 0 0 0 0 
Diptera 4 1 5 0 
Gastropoda 1 0 0 0 
Hemiptera 3 1 0 46 
Hymenoptera 1 6 1 10 
Isoptera 0 0 0 0 
Larvae 1 0 0 1 
Lepidoptera 0 8 0 0 
Mantodea 0 0 0 0 
Millipede 4 0 0 2 
Neuroptera 1 0 0 0 
Orthoptera 55 25 1 0 
Phasmatodea 0 0 0 0 
Scorpiones 0 0 0 1 
Solifugae 0 0 1 0 
Araneae 14 1 4 3 
Thysanoptera 1 0 0 0 
Thysanura 0 0 0 0 
 
40 
 
In Table 3.20, the descriptive statistics are given for the species compositions which were 
sampled in both the pitfall and the sweepnet traps during the four months for Site B. It is 
clear from referring to the Table that April had the highest average species composition with 
a number of 6.6 insects (min = 0; max = 112; standard deviation = 24.3). Second to April is 
January with an average of 5.7 species (min = 0; max = 43; standard deviation = 10). August 
has the second lowest average species composition with a value of 3 species (min = 0; max = 
54; standard deviation = 11.7). The month with the least number of species was October with 
a number of 1.7 insects (min = 0; max = 14; standard deviation = 3.2).  
 
Table 3.20: Descriptive statistics for the species composition sampled in both the pitfalls and 
sweepnet traps for Site C 
 
Variable Mean  Median  Minimum Maximum Standard deviation  
January 5 1 0 55 12.5 
April 4 0 0 52 11.6 
August 2 0 0 30 6.5 
October 4 0 0 46 10.2 
 
In Table 3.21, the average change in species composition for Site C for each month is 
depicted. The highest average change in species composition occurred in January, with a 
value of 4. Following January, April and October has the second highest change in species 
composition with values of 4. The least average change in species composition occurred in 
August with a value of 2. According to the p-value, which is bigger than 0.1 (0.27361) there 
is no statistical significance for the differences between the four months regarding species 
composition for Site C.  
 
Table 3.21: Friedman ANOVA analysis for determining the change in species composition 
for Site C (p = 0.27361). 
 
Variable Mean  Standard deviation 
January 5 12.5 
April 4 11.6 
August 2 6.5 
October 4 10.2 
 
 
41 
 
 3.3.1.4 Combination of all three sites A, B and C 
In Table 3.22, grouped descriptive statistics are given for the species compositions which 
were sampled in both the pitfall and the sweepnet traps during the four months for Site A, B 
and C. It is clear from referring to the Table that January had the highest average species 
composition with a number of 4.7 (min = 0; max = 55; standard deviation = 10.4). Second to 
January is April with an average of 4.6 species (min = 0; max = 112; standard deviation = 
16.1). October has the second lowest average species composition with a value of 3.7 species 
(min = 0; max = 59; standard deviation = 9.8). The month with the least number of species 
was August with a number of 2.3 insects (min = 0; max = 54; standard deviation = 8.5).  
 
Table 3.22: Grouped descriptive statistics for the species composition sampled in both the 
pitfalls and sweepnet traps for Site A, B and C 
 
Variable Mean  Median  Minimum Maximum Standard deviation  
January 4.7 0 0 55 10.4 
April 4.6 0 0 112 16.1 
August 2.3 0 0 54 8.5 
October 3.7 0 0 59 9.8 
 
Below is an illustration (Fig. 3.3) of what is portrayed in Table 3.22. The change in the 
quantity of prey items is depicted against the different months for each site A, B and C. 
42 
 
 Site
 B
 Site
 A
 Site
 C
Jan Apr Aug Okt
SEASON
-6
-4
-2
0
2
4
6
8
10
12
14
16
Q
u
a
n
t
i
t
y
 
o
f
 
p
r
e
y
 
i
t
e
m
s
 
Figure 3.3 Illustration of the change in the quantity of prey items depicted against the 
different months for each site A, B and C. 
 
In Table 3.23, the average change in species composition for Sites A, B and C for each month 
is depicted. The highest average change in species composition occurred in January, with a 
value of 4.7. Following January, April had the second highest change in species composition 
with a value of 4.6. The second least average change in species composition occurred in 
October with a value of 3.7. The least average change in species composition occurred in 
October. According to the p-value, which is smaller than 0.1 (0.06456) there is a statistical 
significance for the differences between the four months regarding species composition for 
all the sites grouped together.  
 
Table 3.23: Grouped analysis (Site A, B and C) (Friedman ANOVA) for the determination of 
the average change in species composition (p = 0.01950) 
 
Variable Mean  Standard deviation 
January 4.7 10.4 
April 4.6 16.1 
Oct 
43 
 
August 2.3 8.5 
October 3.7 9.8 
 
In Table 3.24, the results of the change in species composition with regard to each site for 
January is given. With a p-value of 0.4982 (bigger than 0.1), there is no statistical significant 
difference between the sites in January. 
 
Table 3.24: Independent grouping (Kruskal-Wallis test) results using multiple comparisons 
for January (p = 0.4982) 
 
January A B C 
A  0.8 1.0 
B 0.8  1.0 
C 1.0 1.0  
 
In Table 3.25, the results of the change in species composition with regard to each site for 
April is given. With a p-value of 0.5021 (bigger than 0.1), there is no statistical significant 
difference between the sites in April. 
 
Table 3.25: Independent grouping (Kruskal-Wallis test) results using multiple comparisons 
for April (p = 0.5021) 
 
April A B C 
A  1.0 1.0 
B 1.0  1.0 
C 1.0 1.0  
 
In Table 3.26, the results of the change in species composition with regard to each site for 
August is given. With a p-value of 0.8335 (bigger than 0.1), there is no statistical significant 
difference between the sites in August. 
 
Table 3.26: Independent grouping (Kruskal-Wallis test) results using multiple comparisons 
for August (p = 0.8335) 
 
August A B C 
A  1.0 1.0 
44 
 
B 1.0  1.0 
C 1.0 1.0  
 
In Table 3.27, the results of the change in species composition with regard to each site for 
October is given. With a p-value of 0.4476 (bigger than 0.1), there is no statistical significant 
difference between the sites in October. 
 
Table 3.27: Independent grouping (Kruskal-Wallis test) results using multiple comparisons 
for August (p = 0.4476) 
 
October A B C 
A  1.0 0.8 
B 1.0  1.0 
C 0.8 1.0  
 
 
3.3.2 Quantification (frequency and volume) of prey items 
Table 3.28: Combined data reflecting the number of invertebrates from all three sites and the 
volume which was obtained from the invertebrates 
  Pitfalls  Sweepnets  
Month 
Invertebrate 
numbers 
Invertebrate 
volume (ml) 
 
Invertebrate 
numbers Invertebrate volume (ml) 
Mean 
individual 
invertebrate 
volume 
January 378 14.8 
 
454 6.8 
 
0.03 
April 241 10.9  40 5.9 0.06 
August 140 13 
 
6 0.6 
 
0.09 
October 232 97.3 
 
7 0.5 
 
0.41 
TOTAL 991 136 
 
507 13.8 
 
0.1 
 
 
 
 
 
 
45 
 
3.4 AVAILABILITY OF TREES FOR NESTING AND ROOSTING 
 
Figure 3.4: Map illustrating the availability of large trees on the farm Greyghost Safaris. 
 
The results show that there are relatively a large number of large enough trees on the farm 
which could be suitable for nesting and roosting regarding size preference for nesting. 
However, a calculation of the number of trees shows less than one (0.605) tree per hectare. 
Suitable trees on the farm appear to occur in clusters on the eastern and western side of the 
farm.   
 
3.5 PRE-RELEASE SITE EVALUATION/STAKEHOLDER ENGAGEMENT 
 
A series of interviews were conducted with the local landowners regarding SGH’s and the 
landowners’ attitudes towards and opinions regarding potential reintroductions either on their 
farm or in the surrounding area. 
46 
 
3.5.1 Demographic information of stakeholders 
Ten interviews were conducted using a purposeful sampling strategy with landowners in and 
around the study site. One of the interviewees was not a landowner as such but a reserve 
manager. Out of the ten interviewees, nine were male and one female. Their ages ranged from 
30 years to 55 years with all the interviewees being Caucasian. Seventy percent of the 
farmers interviewed have been living on their farms for more than 10 years, some attaining a 
stay of between 20 and 30 years. Only one interviewee had been living in the area for less 
than a year. The type of farming activities taking place on these farms is primarily wildlife 
and hunting; however, four out of the 10 landowners have stock and practise arable farming 
activities, too. Sizes of the farms ranged between 330 and 9300 ha. These farms had various 
numbers of people working or living there, ranging between five and 500 people.  
 
3.5.2 Attitudes regarding the SGH 
All the interviewees claimed that they felt positive towards having SGHs on their farms and 
said that they would be glad if the birds were seen more frequently. Two of the interviewees, 
however, stated that even though they are for having the birds on their farms, they are 
concerned about the tortoises that act as prey for the SGH, since introducing the birds may 
result in a decline in their numbers.  
 
3.5.3 Sightings of the SGH on farm or in area 
Nine out of ten interviewees had seen the birds in or around their farms before. Three of the 
interviewees stated that they see the birds as often as once every week, either on their farm or 
on the roadside just outside of Musina. The same group seen by these observers consisted of 
two females and one male. One interviewee claimed that the birds are seen by him at least 
once a month, and another farmer claimed he sees the birds at least three times per month. 
Two interviewees claimed that the birds were sighted seasonally, especially during March or 
April. Other farmers claimed that they have not seen the birds for at least two years, whereas 
one farmer even claimed that he has not seen the birds for over eight years.  
 
3.5.4 Willingness to reintroduce SGH onto farm 
The question regarding whether or not the farmers are positive to have SGHs reintroduced 
onto their farm had an all-round positive response. Everyone was keen, if approached in the 
future, to allow reintroduced birds onto their farms. There was one concern from one farmer, 
47 
 
however, and that was that there is a lot of activity taking place on some parts of his farm 
which could cause some disturbance and may cause the birds to fly off.  
 
3.5.5 Presence of nest on farm 
There was one farmer in particular who claimed that the birds were seen at least once a week 
on his land for a number of years. This could give rise to the possibility of there being a 
breeding pair, and even a nest, on this particular farm although the farmer was not aware of a 
nest. Fifty percent of the interviewees claimed that they are not aware of any nests on their 
farms, and 40% declared they have artificial nests erected although all were unoccupied at the 
time.  
 
3.5.6 Dynamics regarding other stakeholders in the area  
All the farmers stated that the relationships between the farmers are all positive and that there 
are no underlying issues as such which may create problems if the SGHs were to be 
reintroduced into the area. 
 
3.5.7 Reasons for decline according to stakeholders 
The main reasons for decline in SGH population numbers, according to the interviewees, are 
developments, agricultural activities, habitat alteration and degradation, poisoning (indirect), 
poaching and the birds’ slow reproductive rates. Two of the interviewees stated that they had 
no idea why the decline is occurring, and one interviewee stated that the population numbers 
are not declining but are still the same as many years ago. 
 
3.5.8 Sufficient farm size for reintroductions 
In order to reintroduce SGHs into an area, one must be sure that there is a large enough area 
for them to forage and to claim as their territory. SGHs, as mentioned in chapter 1, require an 
area ranging between 100 and 200 km². When asked whether or not their farm will be 
suitable with regard to size for reintroducing the birds, six farmers claimed that their farms 
were sufficient in size. The remaining four interviewees who had not such a large area 
available claimed that their neighbours will gladly cooperate in allowing the birds to move 
about freely from one area to the next. 
 
 
 
48 
 
3.5.9 Active conservation officer 
Five out of the ten interviewees were aware of the local fauna and flora conservation officer 
in Musina. The landowners of the area have incorporated their own group, called the Wildlife 
Study Group, which meets once a month outside hunting season. Here they discuss problems 
or aspects regarding wildlife and the conservation thereof.  
 
3.5.10 Local vet in the area 
According to the interviewees, there is a local veterinarian in Musina who deals mostly with 
domestic animals. When in need of a wildlife specialist, vets from Louis Trichardt and Bela-
Bela and Lephalale (Ellisras) are called upon and transported to the farms. This, they say, is 
very costly to the farmers, thus the vets are only called upon when absolutely necessary. 
Every farmer was keen on taking an injured SGH to the veterinarian to be revived, if they 
ever come across one. During one interview, it came to the fore that one farmer had actually 
found an injured bird approximately two years ago and had tried to bring it back to good 
health; however, due to the severity of its injuries, the bird unfortunately did not survive.  
 
3.5.11 Predators in the area 
According to all the farmers, they all seem to have the same predator species on the farms, 
namely leopards, hyenas (brown and spotted), caracal, serval, African wildcat and occasional 
cheetah. 
 
3.5.12 Problems regarding window-breaking 
None of the interviewed farmers have had problems with SGHs breaking windows, although 
two farmers claimed that they have indeed heard of such incidents on neighbouring farms and 
near Lephalale.  
 
 
49 
 
 
CHAPTER 4 
 
Discussion 
50 
 
 
CHAPTER 4: DISCUSSION  
 
The aim of this study was to identify and verify optimal reintroduction sites for the Southern 
Ground Hornbill (SGH) in the Musina area of the Limpopo Province. To achieve this, five 
main objectives were identified, namely to: 1) identify optimal reintroduction areas, 2) 
develop a habitat profile, 3) determine the availability of prey items, 4) determine the 
availability of large trees, and 5) evaluate pre-release sites by involving a number of 
stakeholders. This study can be viewed as the foundation for similar work in other parts of 
the SGH’s range.  
 
4.1 OPTIMAL REINTRODUCTION SITES 
It is clear that the Limpopo River Valley, north of the Soutpansberg and stretching across to 
the KNP on the eastern border of the Limpopo Province, is a potentially suitable re-
introduction area for the SGH. Other areas in the central parts of the province seem to be less 
suitable, probably because of low minimum temperature which has an effect on the 
availability of prey items. Greyghost Safaris, the farm that was used as the primary study site 
which forms part of the larger farm called Ludwigslust, is one of the farms in the Limpopo 
River Valley and, therefore, a good representative of the larger area west of Musina. The 
SGH is also regularly observed on the farm in the habitats that were chosen for further 
analysis. Furthermore, the area has a high density of large baobabs and other large tree 
species and further meets the SGH’s habitat requirements regarding vegetation, climate and 
prey abundance which makes it a very suitable area for re-introduction and a good study site.  
 
4.2 HABITAT PROFILE 
4.2.1 Forbs and grasses 
In terms of the composition of forbs and grass species, Site A and Site C, had the same 
number of forbs and grass species, while Site B had the highest species composition. 
According to Van Rooyen’s (2008) description of the different plant communities for the 
specific sites, Site A should have the highest species composition (52% grass cover and 12% 
forbs), while Site B should have a lower species composition (41% grass cover) than Site A 
as the area has a high rock density which should affect the percentage ground cover. Site C, 
on the other hand, should have the least percentage ground cover as it is an area of old 
51 
 
ploughed fields lacking in tall trees. The area has previously been exposed to heavy grazing, 
and Site C has certain forbs species which are in high abundance, such as Alternanthera 
pungens and Tribulus terrestris. A possible reason for these phenomena could be the high 
rainfall on the farm which occurred during 2010 and 2011 relative to the years 2008 and 2009 
(Table 2.1) which could have affected the average plant growth for the area at the different 
sites. The reason for Site A having a much lower species composition compared to Site B 
could be ascribed to poor land-use management, such as overgrazing. The presence of larger 
tree species at Site A could also have an adverse affect on the grass cover, compromising the 
composition, spatial distribution and productivity of grasses in the savannah (Scholes & 
Archer, 1997). 
 
 By referring to the descriptive analysis of the forbs and grass species composition for all the 
sites during the four different months (Tables 3.2, 3.5 and 3.8), it is clear that for Site A, the 
highest composition of species was collected during August. This can be due to the presence 
of a higher abundance of annual grasses and forbs which are adapted to survive throughout 
the winter months. At Site B, the highest composition of species was recorded during 
January. This could be a result of the high rainfall that occurred during January 2011 (Table 
2.1). At Site C, the highest composition of species recorded was during April. April is usually 
a colder month, although Site C, having a disturbed soil and vegetation structure due to the 
previous use of the land, has a high abundance of vegetation which is commonly found on 
disturbed area, such as the before-mentioned Alternanthera pungens. Due to the lack of trees, 
this could also mean that compared to the other three sites, Site C would have the highest 
percentage ground cover during this month as the forbs and grasses composition cannot be 
adversely affected by trees and their roots, as could have been the case at the other sites. 
Although April is low in rainfall and temperatures relative to the other months of the year, 
Site C is still capable of providing enough food sources for the SGH. October provided the 
least number of species composition for all three sites. October is known to be one of the 
hottest and driest months as it is post-winter and precedes the rainfall season which usually 
runs from November to March (Van Rooyen, 2008). As a result, plant species composition is 
low.  
 
4.2.2 Woody structure 
Together with having the highest species composition of forbs and grasses, Site B had the 
highest tree species composition, compared to the other two sites, with nine different species 
52 
 
(Table 3.6). Site A followed with six main tree species (Table 3.3) and Site C with five 
species (Table 3.9). Most of the species found at all three sites are classified as small (up to 5 
m) or medium (5-15 m) (Van Wyk & Van Wyk, 2007). This is not ideal for SGH nesting or 
roosting as they require larger trees, falling in the classes of large (15-30 m) to very large 
(more than 30 m). Trees which fall more in this size range include Combretum imberbe, 
Ficus sycomorus, Diospyros mespiliformis, Sclerocarya birrea and Adansonia digitata, 
which all are preferred by the SGH for nesting in the KNP (Kemp, 1995). 
 
For all three Sites A (Table 3.3), B (Table 3.6) and C (Table 3.9), the dominant tree species 
recorded was Colophospermum mopane. The main reason for this is the fact that the sites all 
fall under the Musina Mopane Bushveld vegetation type (Fig. 2.2) (Mucina & Rutherford, 
2006), so the chances of mopane being the dominantly found tree species along the transects 
are high, as the farm has large areas of dense mopane bushveld. Mopane can usually be found 
growing in deep red/brown clays to freely-drained sandy soils, as well as shallow Glenrosa 
and Mispah soils often formed as a result of rock weathering and clay alleviation (Rossouw 
& Van der Waals, 2010; Walker et al, 1978).   
 
When comparing the species composition of all three sites with each other for the woody 
structure, it is clear that Site A and Site B have more in common with regard to tree species 
(two commonly-shared species, namely Acacia senegal and Grewia flava) than Site A with 
Site C. Site C has more in common with Site B, however, with two species occurring at both 
Sites (Acacia tortilis and Terminalia prunioides). Site C only has one species in common 
with Site A; that being Combretum apiculatum. Sites A and B both have species not 
occurring at any of the other sites (Site B Acacia nigrescens, Dichrostachys cinerea and 
Dombeya rotundifolia; Site A Kirkia acuminata).  
 
According to the description of the different plant communities in which the three sites are 
classified (Fig. 2.3), Kirkia acuminata is common for Site A’s specific plant community, 
together with the raisin bushes (Grewia flava and G. bicolor). The remaining characteristic 
trees and bushes of the plant community in which Site A is classified were not recorded. This 
could be due to the method of sampling which was used. The cross-transect of 200 m 
recordings does not provide 100% effectiveness for identifying all the tree species that are 
found in the area but, instead, only along the narrow transect lines.  
 
53 
 
For Site B, a greater similarity was found between the vegetation species composition 
identified during sampling and the description of the plant community (Fig. 2.3) according to 
Van Rooyen (2008). Seven of the species listed in the plant community description for Site B 
were amongst the nine species sampled through the cross-transect (Table 3.6). The 
classification of Site B’s plant community as “mopane-slender three-hook thorn low dense 
bushveld” is, therefore, considered to be best-suited.  
 
Acacia tortilus and Colophospermum mopane were recorded at Site C (Table 3.9), coinciding 
with the description of the plant community in which this site is classified (Fig. 2.3). 
According to the plant community description, “umbrella thorn-bushveld signal grass old 
fields” (Van Rooyen, 2008), the area does not consist of tree species but, instead, has a 3% 
cover of shrubs. This is due to the fact that the area was previously used for arable farming; 
consequently, any large tree species were possibly removed and the area heavily grazed.  
 
4.2.3 Change in ground cover 
In terms of the results for the change in ground cover per month for each site, Site A had the 
highest average change in ground cover occurring in August (Table 3.10). For Site B (Table 
3.11), the highest average change in ground cover occurred during January, and for Site C 
(Table 3.12), the highest average change in ground cover occurred during April. The reason 
for the high change in ground cover at Site A during August could be due to the fact that 
during this month, Site A had the highest forbs and grass species composition compared to 
the other three months consequently more species commonly subject to the changing of the 
seasons occur at this site.  Site B obtained the highest change in ground cover in January, and 
this could be due to the high species composition for Site B compared the other two sites. 
Since January falls in the rainy season, more species available on Site B are subject to change 
as a result of the changing of the seasons. October, being one of the driest and hottest months, 
is the month with the lowest change in ground cover for all three sites. The difference 
between Sites A and B with C, regarding the change in ground cover, is that Site C’s results 
do have statistical significance for the difference between the months (p-value smaller than 
0.1), while the other two did not. This could be due to the fact that Site C did not have a big 
species composition to begin with.  
 
 
 
54 
 
4.3 AVAILABILITY OF PREY ITEMS 
The site with the highest prey species composition is Site A (3.13) with a total of 16 orders. 
Site C had a total of 15 different prey orders (Table 3.13), while Site B (Table 3.16) had the 
lowest composition of prey. The dominant prey order for all three sites is Coleoptera. 
However, when comparing the dominant prey order for each site between the different 
months, it is clear that Site B and Site C are more similar as they both have the same 
dominant orders for each month (January = Orthoptera, April = Coleoptera, August = 
Coleoptera and October = Hemiptera).  
 
The only difference between Sites B and C and Site A is the dominant species in January 
(Site A = Coleoptera). At Site A, a member of the order Squamata was captured and released 
during April. This was unfortunately the only species of this order which was captured in the 
traps. It is clear that at Site C, the order Anura was recorded to appear during the month of 
October. The first rains of the season occurred during sampling in October (30.5 mm) and 
thus gave rise to the reappearance of amphibians. During October, at all three sites, the 
highest appearance of the order Scorpiones occurred. The size and abundance of millipedes 
were also much higher than during the previous months. Large millipedes tend to appear after 
the rainfall. This, too, is the reason why SGH would be able to survive in this area, even 
during the hotter, drier periods of the year, as each month has its own provision of prey items 
in its own form. The open grass field at Site C provided a large number of low-flying insects 
such as grasshoppers, butterflies and moths, especially during the rainy season. The SGH has 
been known to prefer open areas such as this for foraging as they have clear visibility for 
potential threats. Furthermore, because of their primarily walking nature, they prefer the 
openness as it is easier to forage with less density of bushes and trees.  
 
During the rainy season, Site C was more prone to have stagnant water puddles after high 
rainfall. This gave rise to the presence of amphibians. Frogs provide high protein content for 
the SGH to prey upon. These water puddles together with the frogs can be an attraction for 
other potential prey species of the SGH, such as snakes, rodents, other insects and even 
species of smaller birds. The soft ground also allows for digging in search of food, especially 
during dry months. 
 
With regard to the highest average prey species composition, Site A (Table 3.14) differs from 
Site B (Table 3.17) and C (Table 3.20) by having the highest species composition during 
55 
 
October. As the temperature becomes hotter, prey such as the mentioned scorpions, 
amphibians, reptiles and possible rodents appear due to the increase of insect abundance after 
August. October is near the start of the rainy season, which could give the SGH the 
opportunity to move back to this area in instances where they may have migrated to other 
areas during August. For Sites B and C, the month with the highest average species 
composition occurs in April. For Sites B and C, the least average species composition occurs 
in October, compared to August for Site A. 
 
The highest average change in prey species composition, when it comes to change in species 
composition between the different months for Site A (Table 3.15), October has the highest 
change, compared to Site B (Table 3.18) which had the highest average change in April. Site 
C (Table 3.21) had the highest change in composition during January. For Sites A and C, the 
least average change in species composition occurred during August, compared to October 
for Site B. August was the month which could have the highest adverse effect on the feeding 
habits of the SGH. This would mean that the SGH groups are more likely to migrate to other 
areas; hence their large territory sizes outside protected areas. Foraging in cultivated fields or 
flying across to be closer to the Limpopo River is a possibility.  
 
When comparing the different months regarding species composition for the three sites 
combined (Table 3.22), January has the highest average change in species composition for all 
three. This could be due to the high rainfall which had occurred during January (Table 2.1). 
Second is April, and third is October. The month with the least species composition is 
August. There is no significant statistical difference between the sites for each month 
separately (p-value bigger than 0.1). October has a higher species composition compared to 
August due to the fact that October is a warmer month and 30.5 mm rainfall was recorded 
during that month which could have initiated the return of prey species. Regarding the change 
in species composition for all three sites combined between the different months (Table 
3.23), January showed the highest change, second is April, third is October and the lowest 
average change occurred during August.  
 
The volumes of the prey which were measured for each month (Table 3.28) for the pitfall and 
sweepnet traps had appealing results. When combining the results for the sweepnet and the 
pitfall traps, it is clear that January had the highest number of prey which was collected; 
however, it did not offer the most volume. This indicates that the birds, during the summer 
56 
 
months, will forage more and consume a higher number of smaller prey items, except when 
referring to larger prey such as rodents and tortoises and snakes which should be in greater 
abundance during this time. The SGHs are more likely then to travel longer distances in 
search of prey.  January also had the highest composition of forbs and grasses at Site B. This 
is an indication that just because there is a higher number of insects offered during that 
month, it does not necessarily mean that the volume of food available whilst foraging will 
provide the sufficient amount of food to meet the SGH’s feeding requirements. Surprisingly, 
October gave the highest volume of food items, although it was third when comparing the 
number of prey species caught. The large volume came from the large size of the insects as 
well as the large scorpions which were collected from the traps. As mentioned before, 
October also had the presence of amphibians which were collected in the traps. October, as it 
seems, would be the month which would provide the most sufficient volume of food within a 
smaller area. This is advantageous to the SGH as they would not need to walk long distance 
to forage during the drier months. This could also prepare them for the next breeding season 
which should occur as soon as the first big rains have occurred as of from the start of 
November. They could then spend more of their energy in preparing the nest for the female. 
August provided the least number of insects, and the least volume. This is partly due to the 
fact that during winter, less prey is available. April had the second highest number of species 
although it came third when comparing the volume it provided.  
 
 
4.4 AVAILABILITY OF TREES FOR NESTING AND ROOSTING 
As seen on the map in Figure 3.4, it is clear that a large number of large trees are available on 
Greyghost Safaris, although calculations have determined that there is less than one tree per 
hectare, which could mean that the quantity of the large trees is not as important as the 
availability of trees with suitable nest cavities. 
 
4.5 PRE-RELEASE SITE EVALUATION/STAKEHOLDER ENGAGEMENT 
Finally, the reaction from the local landowners in the Musina area was mainly positive with 
regard to the SGH being a local resident on their farms. The majority of the farms (60%) 
belonging to the farmers interviewed are wildlife farms ranging between 330 and 9300 ha. 
The general perception of the farmers of reintroducing SGHs into the area, if ever decided, is 
positive. Veterinary assistance is available in the area, as well as a local Fauna and Flora 
officer. The farmers have also established their own Wild Study Group which deals with 
57 
 
everyday aspects regarding wildlife management. Conservation of wildlife species is a 
concern amongst the majority of the farmers. As was identified by means of the interviews, it 
is clear that there is an existing population of SGH (group of three) which is spotted 
regularly. The most sightings and voice calling of the group occur on the farm neighbouring 
Greyghost Safaris. This could be a sign that they have a nest in the vicinity and also that the 
habitat is preferred by the birds. Forty percent of the farmers claimed that they had artificial 
nests erected on their farms, although they have never been utilised by SGH.  
 
All the landowners claimed to have the same species predators which may influence 
reintroduction site selection, namely leopards, brown and spotted hyena, jackal, caracal and 
the occasional cheetah.  
 
A concern for reintroducing SGH into an area is the act of breaking windows. SGH are 
highly territorial (Kemp, 1995), and will thus attack anything that resembles an imposter, 
even though it is simply the bird’s own reflection. Of the landowners interviewed, none had 
ever had a problem with window-breaking although it has been heard of elsewhere. This 
causes concern for the health of the SGH as they may be injured in the process of breaking 
the windows, and it affects the farmers financially.  
 
58 
 
CHAPTER 5 
 
CONCLUSION AND RECOMMENDATIONS 
59 
 
 
CHAPTER 5: CONCLUSION AND RECOMMENDATIONS  
 
This study site provided a good opportunity to explore what an optimal site for reintroduction 
should resemble. Site A, with the highest composition of prey items for all four months 
combined, suggests that this type of plant community represents the most ideal habitat, 
including food variety and availability. As this area is considered to be woody in the sense of 
small bushes and trees (second to Site B), small mammals such as rodents are prone to 
habituate the area. Squirrels and other tree-living species are more abundant, as well as other 
smaller birds which make their nests in the branches of the trees. The presence of these 
species may likely attract other prey for the SGH such as snakes which are after the birds and 
their eggs, as well as small rodents.  
 
Even though the presence of large trees  in the sampling area were few, this does not indicate 
that there is no possible solution for the birds to make their nests, as artificial nests could be 
implemented and further research into the architecture and positioning of the nest-box can be 
undertaken. It is recommended that this type of habitat is further mapped and explored to 
identify similar areas.  
 
The results obtained from the study are not 100% effective when compared to the actual 
sightings and observations of prey abundance and tree abundance. An aspect to consider 
when measuring the effectiveness of the study, with regard to the model used to determine 
the niche preference in the current study area, is the fact that most of the previous studies 
done on the habitat of the species were conducted in the KNP; that which the selected study 
site does not represent entirely. Thus, extrapolating environmental data from the KNP to the 
Limpopo River Valley ecosystem is not entirely a suitable method to determine niche or 
habitat preference in the current study area.  
 
The cross-transect can provide a vague indication of the actual vegetation species abundance. 
Decisions should not be based wholly on this study or on studies resembling this one but, 
instead, should include actual observations and should be conducted on a larger area, possibly 
over a longer time period. The effectiveness of the large tree abundance can be questioned as 
only baobabs were identified; along the water courses other large Fig tree species could also 
60 
 
be suitable for nesting. Also, ground-truthing needs to be done so as to identify trees with 
large enough cavities in them which will make it suitable for nesting.  
 
The prey abundance results are not 100% effective either, as the pitfalls could have been 
deterred by external factors such as rainfall, the area selected, time of sampling and length of 
sampling, amongst others. The pitfall traps were there simply to give an indication as to what 
species are in the area which could be a possible food source for the SGH, although the 
majority of the prey captured in the traps was small-sized insects which only contribute 
slightly to the diet of the SGH. There are other sampling methods which could be 
implemented for a more detailed analysis. The scale of the sampling, relative to the size of 
the SGH territory also acted as a potential problem in the effectiveness of the results. As a 
solution to this problem, it is recommended to carry out similar studies but on larger scale. It 
is recommended by Sutherland (2000) to use long thin quadrats instead of the square, cross 
quadrants that were used in the study. Thin long quadrats can contain 18% more species than 
if using a square quadrat.  
 
As was noted by the variation in prey abundance and vegetation during the different seasons, 
seasonal and breeding behaviour as well as seasonal requirements should be considered 
before reintroducing the birds. It is important not to ignore the vegetation composition in 
future modelling of SGH-habitat relationships as it could be a good indicator of food 
availability in reintroduced SGH populations. Habitat quality may also be assessed better 
within an interacting ecosystem and as a functional-resource-based concept of habitat 
(Michel et al, 2010). The distribution of resources and suitable habitats across the available 
landscape could influence the survival and reproduction of reintroduced bird populations. 
 
With regard to the stakeholder interviews, it was not feasible to interview members of the 
local subsistence farmers as the area selected did not include any of the above mentioned 
farmer type; hence the limited data on the perception of farmers regarding the species. Future 
studies should include a larger area thus including other members of the surrounding 
communities as they also play an important role in the survival of the species.  
61 
 
REFERENCES 
 
Acocks, J.P.H. 1953. Veld types of South Africa. Mem. Bot. Surv. S. Afr 57: 1-146 
 
BirdLife International 2010. BirdLife International. [Web]: www.birdlife.org.za  Retrieved 
 November, 2010. 
 
Block, W.M. and Brennan, L.A. 1993. The habitat concept in ornithology: Theory and 
applications. Current Ornithology 11: 35-91. 
 
Campbell, H. W.,  & Christman, S. P. 1982. Field techniques for  herpetofaunal community 
analysis.  Pages  193-200.  In:  N.  J.  Scott,  Jr.,  (ed.) Herpetological communities.  
U.S. Fish and Wildl. Serv.,  Wildlife Reserve Report 13. 
 
Cyrus, D. & Robson, N. 1980. Bird Atlas of Natal. University of Natal, Pietermaritzburg. 
 
Derwent, S., & Mander, M. 1997. Twitchers bewitched. The use of birds in traditional 
 healing. Africa birds and Birding 2(1): 22-25 
 
Du Plessis, M.A., Siegfried, W.R. and Armstrong, A.J. 1995. Ecological and life-history 
correlates of cooperative breeding in South African birds. Oecologia 102(2): 180- 
188. 
 
Engelbrecht, D., Theron, N., Turner, A., Van Wyk, J., Pienaar, K. 2007. The status and 
conservation of Southern Ground Hornbills, bucorvus leadbeateri, in the Limpopo 
Province, South Africa. In: Kemp, A.C. & Kemp, M.I. (eds.), The Active 
Management of Hornbills and their habitats for conservation, pp. 252-266. CD-
ROMM Proceedings of the 4th International Hornbill Conference, Mabula Game 
Lodge, Bela-Bela, South Africa. Naturalists & Nomads, Pretoria, South Africa.  
 
Field, A. 2009. Discovering statistics using SPSS. 3rd ed. London: Sage Publications.  
 
62 
 
Hebblewhite, M., Miquelle, D. G., Murzin, A. A., Aramilev, V. V., & Pikunov, D. G. 
2011.  Predicting potential habitat and population size for reintroduction of the Far Eastern 
leopards in the Russian Far East. Bioligical Conservation. doi:10.1016/j.biocon.2011.03.020 
 
Hutchinson, G.  E. 1957. Concluding remarks. Cold Spring Harbor Symp. Quant. Biol  22:  
 415-427. 
 
IUCN 1987. The IUCN position statement on translocation of living organisms: 
Introductions, Reintroductions and Restocking. 22nd Meeting of the IUCN Council, 
Gland, Switzerland 
 
Jaynes, E.T. 1957. Information Theory  and  Statistics  Methods,  I. Physical  Rev.  
106 (1957a): 620-30. 
 
Jones, B.C. 1969. Cliff nesting. Lammergeyer 10:103-4 
 
Kemp, A.C. 1976. Factors affecting the onset of breeding in African hornbills. Proceedings 
of the 16th International Ornithological Congress. Canberra City, Australia. 248- 
257. 
 
Kemp, A.C. 1980. The importance of the Kruger National Park for bird conservation in the  
 Republic of South Africa. Koedoe 23: 99-122. 
 
Kemp, A.C. & Kemp, M.I. 1980. The Biology of the Southern African Ground Hornbill 
 Bucorvus leadbeateri (Vigors) (Aves: Bucerotidae). Ann. Transvaal Mus. 32: 65-100. 
 
Kemp, A.C. & Kemp, M. I. 2007. What proportion of Southern Ground Hornbill nesting 
attempts fledge more than one chick? Data from the Kruger National Park. In: Kemp, 
A C. & Kemp, M. I. (eds). The active management of Hornbills and their Habitats for 
Conservation, pp. 288-297. CD-ROM Proceedings of the 4th International Hornbill 
Conference, Mabula Game Lodge, Bela-Bela, South Africa. Naturalists & Nomads, 
Pretoria. 
 
Kemp, A. C. 1987. Ground Hornbill under pressure in South Africa. Afr. Wildl. 41: 293. 
63 
 
Kemp, A.C. 1988. The behavioural ecology of the Southern Ground Hornbill: Are  
competitive offspring at a premium? Proceedings of the International 100 DO-G 
Meeting, Current Topics in Avian Biology. Bonn: 267-271. 
 
Kemp, A.C. 1995. The Hornbills. Oxford University Press: Oxford. 
 
Kemp, A.C. 2000a. ‘Southern Ground Hornbill’. In: Barnes, K.N. (ed.), The Eskom red data 
book of birds of South Africa, Lesotho and Swaziland, Birdlife South Africa, 
Johannesburg, South Africa.  Kemp, 2000. 
 
Kemp, A. C. 2000b. The sustainable utilisation of birds. Emu 100: 355-365 
 
Kemp, A.C., Joubert, S.C.J. & Kemp, M.I. 1989. Distribution of Southern Ground Hornbills 
in the Kruger National Park in relation to some environmental factors. South African 
Journal for Wildlife Research 19(3): 93-98 
 
Kemp, A.C. and Begg, S.G. 1996. Nest sites of the Southern Ground Hornbill Bucorvus 
leadbeateri in the Kruger National Park, South Africa, and conservation implications. 
Ostrich. 67: 9-14. 
 
Kemp, M.I. and Kemp, A.C. 1978. Bucorvus and Sagittarius: Two modes of terrestrial 
predation. In: Kemp, A.C. (Ed). Proceedings of the Symposium on African Predatory 
Birds. Northern Transvaal Ornithological Society, Pretoria. Pp. 13-16. 
 
Kemp, A.C. and Kemp, M.I. 1991. Timing of egg laying by Southern Ground Hornbills 
Bucorvus leadbeateri in the central Kruger National Park, South Africa. Ostrich. 62: 
80-82. 
 
Kemp, A. C., Kemp, M. I., & Turner, A. 2007. What has become of eggs and chicks of  
Southern Ground Hornbills harvested from the Kruger National Park? In: Kemp, A.C. 
& Kemp, M.I. (eds.), The Active Management of Hornbills and their Habitats for 
Conservation, pp. 252-266. CD-ROMM Proceedings of the 4th International Hornbill 
Conference, Mabula Game Lodge, Bela-Bela, South Africa. Naturalists & Nomads, 
Pretoria, South Africa. 
64 
 
Kemp, A.C. and Webster, R. 2007. Latest analysis of Southern Ground Hornbill (SGH) 
distribution and population in South Africa. Unpublished Report. Mabula Ground 
Hornbill Project. 
 
Knight, G.M. 1990. Status, distribution and foraging ecology of the Southern Ground 
Hornbill (Bucorvus cafer) in Natal. Unpublished MSc dissertation, University Natal, 
Durban. 
 
Low, A.B. & Rebelo, A.G. (Eds.), 1996. Vegetation of South Africa, Lesotho and  
 Swaziland. Department of Environmental Affairs and Tourism, Pretoria. 
 
Magnusson, W. E., Lima, A. P., Alves da Silva, W., Carmozina de Araújo, M., & 
Montgomery, W. L. 2003. Use of Geometric Forms to Estimate Volume of 
Invertebrates in Ecological Studies of Dietary Overlap. Copeia: (1):13019.  
 
Mabula ground hornbill conservation project. 1999. [Web]:  
http://www.mabulagroundhornbill conservationproject.org.za. Retrieved 7 March 
2010.  
 
Michel, P., Dickinson, K. J. M., Barrett, B. I. P. and Jamieson, I. G. 2010. Habitat 
selection in reintroduced bird populations: a case study of Stewart Island robins and 
South Island saddlebacks on Ulva Island. New Zealand J. Ecol. 34: 237–346. 
 
Moorhouse, T.P., Gelling, M. & Macdonald, D.W. 2009. Effects of habitat quality upon 
reintroduction success in water voles: evidence from a replicated 
experiment. Biological Conservation, 142:53–60. 
 
Morrison, K., Daly, B., Burden, D., Engelbrecht, D., Jordan, M., Kemp, A., Kemp, M.,  
Potgieter, C., Turner, A., & Friedmann, Y. 2007. A conservation plan for the 
Southern Ground Hornbill Bucorvus leadbeateri. In: South Africa. In: Kemp, A.C. & 
Kemp, M.I. (eds.), The Active Management of Hornbills and their Habitats for 
Conservation, pp. 252-266. CD-ROMM Proceedings of the 4th International Hornbill 
Conference, Mabula Game Lodge, Bela-Bela, South Africa. Naturalists & Nomads, 
Pretoria, South Africa. 
65 
 
Msimanga, A. 2004. Breeding Biology of Southern Ground Hornbill Bucorvus leadbeateri in 
 Zimbabwe: impacts of human activities. Bird Conservation International, 2004. 
Cambridge University Press. 
 
Mucina, L. & Rutherford, M.C. 2006. The Vegetation of South Africa, Lesotho and 
Swaziland. Strelitzia 19. South African National Biodiversity Institute, Pretoria, 
South Africa. 
 
Perner, J. & Schueler, S. 2004. Estimating the density of ground-dwelling arthropods with 
 pitfall traps using a nestedcross array. J. Anim. Ecol. 73:469–477. 
 
Phillips, S. J., Dudik, M. and Schapire, R. E. 2004. A maximum entropy approach to  
species distribution modeling. – In: Proc. of the 21st International Conference on 
Machine Learning, Banff, Canada, 2004. 
 
Phillips, S. J., R. P. Anderson, and R. E. Schapire. 2006. Maximum entropy modeling of 
species geographic distributions. Ecol. Modell. 190:231–259. 
 
Ranger, G. 1931. The Ground Hornbill at home. Blythswood Review, 8:9-34 
 
Rossouw, P.S. & Van der Waals, J.H. 2010. Soil and Agricultural potential of portions of  
the farms Matjiesrivier 41/2, Annashoek 41/3, Karos 959/0 and Zandemm 944/0 in 
the Northern Cape Province. Scoping Report: Terrasoil Science 2010; pp 3. The South 
African Council for Natural Scientific Professions; Reg no. 400106/08. 
 
Rutherford, M.C. and Westfall, R.H. 2004. Biomes of southern Africa: An objective 
categorization. Memoirs of the Botanical Survey of South Africa no. 63. National 
Botanical Institute, Pretoria. 
 
Scholes, R. J. & Archer, S. R. 1997. Tree-grass interactions in savannas. Annu. Rev. Ecol. 
Syst. 28:517-44 
 
Sutherland, W. J. 2000. The conservation handbook: Research, management and policy. 
 Blackwell Science Ltd. United Kingdom. 
66 
 
Tainton, N.M. 1999. The savanna biome. In: Tainton, N.M. (Ed). Veld Management in 
South Africa. University of Natal Press, Pietermaritzburg. Pp. 33-36. 
 
Tavecchia, G., Besbeas, P., Coulson, T., Morgan, B. J. & Clutton-Brock, T. H. 2009.  
Estimating population size and hidden demographic parameters with state-space 
modeling. American Naturalist 173:722–733. 
 
Theron, N. T. 2011. Genetic connectivity, population dynamics and habitat selection of the 
Southern Ground Hornbill (Bucorvus leadbeateri) in the Limpopo Province. 
Unpublished MSc dissertation, Bloemfontein. 
 
Van Rooyen, N. 2008. Ecological evaluation and management plan for Greyghost Safaris. 
 Unpublished report. Unversity of Pretoria.  
 
Van Wyk, B & Van Wyk, P. 2007. How to identify trees in southern Africa. Struik  
 Publishers. South Africa. 
 
Vernon, C.J. 1984. The Ground Hornbill at the Southern extremity of its range. Ostrich. 
57: 16-24. 
Vernon, C. J. & Herremans, M. 1997. Ground Hornbill. In: Harrison, J. A., Allan, D. G.,  
Underhill, L. G., Herremans, M., Tree, A. J., Parker, V. & Brown, C. J. (eds). The 
Atlas of Southern African Birds. Vol. 1: Non-passerines, pp. 708-709. Birdlife South 
African. Johannesburg. 
 
Vigors, N. A. 1825. XXII. Observations on the Natural Affinities that connect the Orders  
And Families of Birds. Transactions of the Linnean Society of London 14: 395–517. 
doi: 10.1111/j.1095-8339.1823.tb00098.x. 
 
Walker, B. H., Norton, G. A., Conway, G. R., Comin, H. N., & Birley, M. 1978. A  
procedure for multidisciplinary ecosystem research: with reference to the South 
African savanna ecosystem project. Journal of Applied Ecology 15: 481-502. 
 
 
67 
 
Wilfred, P. 2007. Habitat viability for the Southern Ground Hornbill in Tanzania: the case 
for the Malagarasi-Moyovozi Ramsar Site. In: Kemp, A.C. & Kemp, M.I. (eds.), The 
Active Management of Hornbills and their Habitats for Conservation, pp. 252-266. 
CD-ROMM Proceedings of the 4th International Hornbill Conference, Mabula Game 
Lodge, Bela-Bela, South Africa. Naturalists & Nomads, Pretoria, South Africa.  
 
68 
 
APPENDIX 1 
 
Woody component 
Scientific name English name 
Acacia nigrensis Knob thorn 
Acacia senegal Slender three-hook thorn 
Acacia tortilus Umbrella thorn 
Colophospermum 
mopane 
Mopane 
Combretum apiculatum Red bushwillow 
Dichrostachys cinerea Sickle bush 
Dombeya rotundifolia Common wild pear 
Grewia bicolor White raisin 
Grewia flava Velvet raisin 
Kirkia acuminata White syringe 
Terminalia prunoides Lowveld cluster-leaf 
 
Grasses and forbs 
 
Scientific name English name 
Aristida adscensionis Annual three-awn 
Acanthospermum 
hispidum 
Upright starbur 
Alternantha pungens Khakiweed/ Paperthorn 
Boerhavia erecta Erect spiderling 
Brachiaria deflexa False signal grass 
Cenchrus ciliaris Buffalo grass 
Corchorus 
asplenifolius 
Wild jute 
Endostemon 
tereticaulis 
Small purple keepsakes 
Eragrostis 
lehmanniana 
Lehmann’s love grass 
Evolvulus alsinoides Blue haze 
Geigeria acaulis Rosulate geigeria 
Geigeria burkei (Afr. = knoppiesvermeerbos) 
Heliotropium nelsonii Common string of stars 
Hermbstaedtia fleckii (Afr. = katstert) 
Hermbstaedtia odorata Cat’s tail 
Hermannia modesta Fairy lights 
Indigofera ingrata Unknown 
69 
 
Indigofera vicioides Unknown 
Microcharis galpinii Grassy false indigo 
Monechma divaricatum Wild lucerne 
Ocimum americanum Wild basil 
Panicum deustum Broad-leaved panicum 
Phylanthus angolensis Unknown 
Seddera capensis Small white seddera 
Sida chrysantha Golden sida 
Stipagrostis uniplumis Silky bushman grass 
Syncolostemon 
canescens 
White-tipped hemizygia 
Tephrosia semiglabra Unknown 
Tragus berteronianus Carrot-seed grass 
Tragus racemosus Large carrot-seed grass 
Tribulus terrestris Devil’s thorn 
Uruchloa pancoides Garden signal grass 
Waltheria indica Unknown 
Xyrophyta humulis Unknown 
 
70 
 
APPENDIX 2 
Example of questionnaire: 
 
1. Name 
 
2. Age 
 
3. Gender 
 
4. Race 
 
5. Farm name 
 
6. Farm size 
 
7. Type of farming activities 
 
8. Number of people living on farm 
 
9. How long have you lived on this farm? 
 
10. How do you feel about ground hornbills? 
 
11. How often do you see hornbills on your farm/other areas? 
 
12. What are your views on having hornbills reintroduced onto your farm? 
 
13. Are you aware of any nests on your farm? 
 
14. What are the dynamics around the area regarding farmers’ relationships or any other 
issues? 
 
15. What do you think are the main reasons for their decline in the area? 
 
16. Hornbills need a territory size of about 100-200km2. . Would your farm be big enough, or 
which of your neighbours would you recommend who will be willing to cooperate? 
 
17. Do you know the local conservation officer in your area? 
 
18. Do you have a local vet in your area and would you ever be willing to take a wounded 
bird to the vet? 
 
19. What kind of predators do you have on your farm? 
 
20. What will be your reaction towards window breaking?