© University of Hamburg 2018 All rights reserved

Klaus Hess Publishers Göttingen & Windhoek www.k-hess-verlag.de

ISBN: 978-3-933117-95-3 (Germany), 978-99916-57-43-1 (Namibia)

Language editing: Will Simonson (Cambridge), and Proofreading Pal Translation of abstracts to Portuguese: Ana Filipa Guerra Silva Gomes da Piedade Page desing & layout: Marit Arnold, Klaus A. Hess, Ria Henning-Lohmann Cover photographs: front: Thunderstorm approaching a village on the Angolan Central Plateau (Rasmus Revermann) back: Fire in the miombo woodlands, Zambia (David Parduhn) Cover Design: Ria Henning-Lohmann

ISSN 1613-9801

Printed in Germany

Suggestion for citations: Volume: Revermann, R., Krewenka, K.M., Schmiedel, U., Olwoch, J.M., Helmschrot, J. & Jürgens, N. (eds.) (2018) Climate change and adaptive land management in southern Africa – assessments, changes, challenges, and solutions. Biodiversity & Ecology, 6, Klaus Hess Publishers, Göttingen & Windhoek.

Articles (example): Archer, E., Engelbrecht, F., Hänsler, A., Landman, W., Tadross, M. & Helmschrot, J. (2018) Seasonal prediction and regional climate projections for southern Africa. In: Climate change and adaptive land management in southern Africa – assessments, changes, challenges, and solutions (ed. by Revermann, R., Krewenka, K.M., Schmiedel, U., Olwoch, J.M., Helmschrot, J. & Jürgens, N.), pp. 14–21, Biodiversity & Ecology, 6, Klaus Hess Publishers, Göttingen & Windhoek.

Corrections brought to our attention will be published at the following location: http://www.biodiversity-plants.de/biodivers_ecol/biodivers_ecol .php Biodiversity & Ecology

Journal of the Division Biodiversity, Evolution and Ecology of Plants, Institute for Plant Science and Microbiology, University of Hamburg

Volume 6:

Climate change and adaptive land management in southern Africa

Assessments, changes, challenges, and solutions

Edited by

Rasmus Revermann1, Kristin M. Krewenka1, Ute Schmiedel1, Jane M. Olwoch2, Jörg Helmschrot2,3, Norbert Jürgens1

1 Institute for Plant Science and Microbiology, University of Hamburg 2 Southern African Science Service Centre for Climate Change and Adaptive Land Management 3 Department of Soil Science, Faculty of AgriSciences, Stellenbosch University

Hamburg 2018

RPlease cite the article as follows:

Tshireletso, K., Makhabu, S.W., Nsinamwa, M., Kgosiesele, E., Setlalekgomo, M.R., Seletlo, Z., Majaga, S. &Legwatagwata, B. (2018) Diversity patterns of woody vegetation of Kgatleng District, Botswana. In: Climate change and adaptive land management in southern Africa – assessments, changes, challenges, and solutions (ed. by Revermann, R., Krewenka, K.M., Schmiedel, U., Olwoch, J.M., Helmschrot, J. & Jürgens, N.), pp. 416-423, Biodiversity & Ecology, 6, Klaus Hess Publishers, Göttingen & Windhoek. doi:10.7809/b-e.00353 Diversity patterns of woody vegetation of Kgatleng District, Botswana

Koketso Tshireletso1*, Shimane W. Makhabu2, Mackenzie Nsinamwa1, Ednah Kgosiesele3, Mpho R. Setlalekgomo2, Zibanani Seletlo1, Sipho Majaga1, and Boipuso Legwatagwata2. 1 Department of Animal Science and Production, Botswana University of Agriculture and Natural Resources, Private Bag 0027, Gaborone, Botswana 2 Department of Basic Sciences, Botswana University of Agriculture and Natural Resources, Private Bag 0027, Gaborone, Botswana 3 Department of Agricultural Engineering and Land Use Planning, Botswana University of Agriculture and Natural Resources, Private Bag 0027, Gaborone, Botswana * Corresponding author: [email protected] Biodiversity

Abstract: Vegetation assessments have intensifi ed around the globe to document plant diversity in the wake of climatic shifts and local environmental changes. Botswana also recognises the value of documenting plant diversity, and in this study we present the results of a vegetation survey of Kgatleng District. Between November 2013 and December 2016 we assessed tree species abundance on 77 plots sized 20 m × 50 m. Vegetation classifi cation resulted in the identifi cation of fi ve woody plant communities: Terminalia sericea–Maytenus tenuispina, Dichrostachys cinerea–Combretum apiculatum, Combretum zeyheri–Vachellia tortilis, Senegalia erubescens–Senegalia mellifera, and Grewia fl ava–Rhigozium brevispinosum. One- way analysis of variance indicated that species richness was not diff erent among the fi ve plant communities, nor was tree density. We furthermore provide a vegetation map of the district based on supervised classifi cation of Landsat 8 imagery.

Resumo: Estudos de vegetação intensifi caram em todo o mundo, de modo a documentar a diversidade de plantas na se- quência das alterações climáticas e de mudanças ambientais locais. O Botswana também reconhece o valor de documentar a diversidade de plantas e, neste estudo, apresentamos os resultados do levantamento da vegetação do Districto de Kgatleng. Entre Novembro de 2013 e Dezembro de 2016, avaliámos a abundância de espécies de árvores em 77 parcelas de 20 m × 50 m. A classifi cação da vegetação resultou na identifi cação de cinco comunidades de plantas lenhosas: Terminalia seri- cea–Maytenus tenuispina, Dichrostachys cinerea–Combretum apiculatum, Combretum zeyheri–Vachellia tortilis, Senegalia erubescens–Senegalia mellifera, e Grewia fl ava–Rhigozium brevispinosum. A análise da variância a um factor indicou que nem a riqueza específi ca, nem a densidade de árvores, eram diferentes entre as cinco comunidades de plantas. Fornecemos ainda um mapa da vegetação do districto baseado na classifi cação supervisionada de imagens de Landsat 8.

Introduction species composition (Moore & Attwell, tally friendly and sustainable methods 1999). of farming observing correct rangeland Botswana, like most southern African Bush encroachment threatens the eco- stocking rates have been advocated for in countries, comprises expansive areas of logical integrity of savanna ecosystems Botswana (Kgosikoma et al., 2012). savanna woodlands. Kgatleng District and is of concern in southern Africa (Trol- Species such as Dichrostachys cinerea lies within the woodland and thorn bush lope, 1980; Skarpe 1990; Ward, 2005; can aff ect the development of other savanna ecosystems. Th ese ecosystems Britz & Ward, 2007) and Botswana (Rho- woody vegetation (Mudzengi et al., provide browse for domestic and wild de et al., 2006), as it reduces diversity of 2013), thus reducing species diversity animals, habitat cover for wild animals, vegetation. A number of causal drivers to and richness. A study by Mudzengi et al. wood products, and traditional medi- changes in vegetation composition have (2013) observed signifi cant diff erences cines, among other uses. Th e diversity of been proposed, ranging from low rainfall between invaded and uninvaded sites ecosystems is integral to rural livelihoods, to support grass growth, overgrazing, re- in abundance, density, and richness of as they use veld products to sustain them- duced fi re frequency favouring woody woody species. selves. As such, eff ective management of plant establishment (Kgosikoma et al., Environmental variables such as soil savannas requires proper knowledge and 2012) and climatic shifts (O’Connor and topography infl uence the occur- documentation of vegetation types and et al., 2014). To that end, environmen- rence of plant species and as such govern

416 C A Biodiversity

Figure 1: Location map for Kgatleng District, Botswana. Map on the bottom right shows the location of Kgatleng District in Botswana.

species composition and the diversity of Kgatleng District, with a land area of the results of the phytosociological clas- the vegetation (Miyamoto et al., 2003). 7960 km2, is relatively small and is un- sifi cation, and (4) to determine the spe- A study by Dahlberg (2000) in north- der immense pressure from diff erent cies richness of the tree communities. eastern Botswana observed that woody land uses. Like most communal parts of species richness was higher on red soil Botswana, the district is exposed to both than on white soil, but species richness arable farming and open-access com- Methods did not diff er between ranching and com- munal livestock grazing. For sustainable munal grazing land. The study also found management of vegetation resources, Study area signifi cantly (P < 0.01) more shrubs on documentation of plant communities and The study was conducted between No- communal land than on ranch land, prob- species richness in these areas is of piv- vember 2013 and December 2016. The ably because of less competition with otal importance. Meanwhile, literature is study area covered the Kgatleng District the herbaceous plants, resulting in shrub lacking as far as plant diversity and rich- of Botswana, located between latitude seedling establishment. In Dahlberg’s ness in these areas are concerned. Thus, 23.88°S–24.51°S and longitude 25.89°E– (2000) study conducted in the north- defi ning plant communities in these 26.82°E (Fig. 1). The district covers an east district of Botswana, Senegalia ni- systems in an era of climate-change ef- area of 7,960 km2 and is characterized by grescens, Combretum apiculatum, and fects on vegetation is critical to serve as vegetation dominated by Vachellia spp., Grewia spp. were observed to prefer red reference points for vegetation changes. Senegalia spp., and Grewia spp., among soil whereas Colophospermum mopane Therefore, the objectives of the study others. Annual rainfall for the district preferred white soil. Other environmen- were to (1) determine the abundance of ranges between 450 mm and 550 mm. tal variables such as elevation have also woody (trees and shrubs) species, (2) to The temperatures range between 6°C and been noted to aff ect species and commu- identify tree communities and their eco- 20°C in winter and 22°C and 30°C in sum- nity diversity of vascular plants (Khan et logical drivers, (3) to produce a vegeta- mer (Bhalotra, 1987). Soils are predomi- al., 2011). tion map based on Landsat imagery and nantly arenosols and luvisols (Moganane,

B E 6 2018 417 Table 1. Abundance (individuals/ha) and indicator species for the woody plant communities identifed in Kgatleng District, Botswana.

ďƵŶĚĂŶĐĞц^ /ŶĚŝĐĂƚŽƌ &ĂŵŝůLJ ^ƉĞĐŝĞƐ DĞĂŶц^ ƉΎ ;ŝŶĚŝǀŝĚƵĂůƐͬŚĂͿ sĂůƵĞ;/sͿ hŶŝƚϭdĞƌŵŝŶĂůŝĂƐĞƌŝĐĞĂʹDĂLJƚĞŶƵƐƚĞŶƵŝƐƉŝŶĂĐŽŵŵƵŶŝƚLJ &RPEUHWDFHDH dĞƌŵŝŶĂůŝĂƐĞƌŝĐĞĂ ϵϲ͘ϭцϮϴ͘ϭϰ ϳϬ͘Ϯ ϭϳ͘ϭцϰ͘Ϯϱ Ϭ͘ϬϬϬϮ &HODVWUDFHDH DĂLJƚĞŶƵƐƚĞŶƵŝƐƉŝŶĂ ϯϰ͘ϲцϮϬ͘Ϭϭ ϮϬ͘Ϯ ϭϭ͘ϳцϱ͘Ϯϲ Ϭ͘ϬϳϯϮ &DHVDOSLQLDFHDH ĂƵŚŝŶŝĂƉĞƚĞƌƐŝĂŶĂ ϮϬ͘Ϭцϴ͘Ϭϭ ϭϴ͘ϲ ϵ͘ϯцϰ͘ϳϯ Ϭ͘ϬϱϬϲ &HODVWUDFHDH DĂLJƚĞŶƵƐƐĞŶĞŐĂůĞŶƐŝƐ ϭϮ͘ϴцϳ͘ϰϰ ϳ͘ϰ ϭϬ͘ϲцϱ͘ϭϮ Ϭ͘ϳϬϱϭ  hŶŝƚϮŝĐŚƌŽƐƚĂĐŚLJƐĐŝŶĞƌĞĂʹŽŵďƌĞƚƵŵĂƉŝĐƵůĂƚƵŵĐŽŵŵƵŶŝƚLJ 0LPRVDFHDH ŝĐŚƌŽƐƚĂĐŚLJƐĐŝŶĞƌĞĂ ϭϰϳ͘ϰцϮϮ͘ϵϱ ϱϳ͘ϭ Ϯϯ͘ϭцϯ͘ϵϰ Ϭ͘ϬϬϬϮ &RPEUHWDFHDH ŽŵďƌĞƚƵŵĂƉŝĐƵůĂƚƵŵ ϯϱ͘ϱцϴ͘ϭϭ ϱϰ͘ϯ ϭϰ͘ϳцϰ͘ϲϲ Ϭ͘ϬϬϬϮ 7LOLDFHDH 'ƌĞǁŝĂďŝĐŽůŽƌ ϯϯ͘ϴцϲ͘ϵϲ ϱϭ͘ϵ ϭϴ͘ϱцϱ͘ϱϮ Ϭ͘ϬϬϬϲ Biodiversity 0LPRVDFHDH ^ĞŶĞŐĂůŝĂŶŝŐƌĞƐĐĞŶƐ Ϯ͘Ϭцϭ͘Ϭϯ Ϯϴ ϵ͘ϯцϱ͘Ϭϭ Ϭ͘ϬϬϱϲ $QDFDUGLDFHDH ^ĐůĞƌŽĐĂƌLJĂďŝƌƌĞĂ Ϭ͘ϳцϬ͘ϯϱ ϮϬ ϳ͘ϲцϰ͘ϯϭ Ϭ͘ϬϮϬϰ &DSSDUDFHDH ŽƐĐŝĂĨŽĞƚŝĚĂ Ϭ͘ϯцϬ͘ϮϮ ϭϮ ϲ͘ϳцϰ͘ϬϮ Ϭ͘ϭϲϲ (XSKRUELDFHDH ƌŽƚŽŶŐƌĂƚŝƐƐŝŵƵƐ ϴ͘ϱцϲ͘ϭϬ ϲ͘ϵ ϴ͘ϵцϰ͘ϲϲ Ϭ͘ϲϭϱϵ &RPEUHWDFHDH ŽŵďƌĞƚƵŵŝŵďĞƌďĞ Ϭ͘ϮцϬ͘Ϯϱ ϰ ϲ͘ϱцϭ͘ϴϭ ϭ (XSKRUELDFHDH ^ƉŝƌŽƐƚĂĐŚLJƐĂĨƌŝĐĂŶĂ Ϭ͘ϮцϬ͘Ϯϱ ϰ ϲ͘ϱцϭ͘ϴϬ ϭ 5XELDFHDH sĂŶŐƵĞƌŝĂŝŶĨĂƵƐƚĂ Ϭ͘ϭцϬ͘ϭϮ ϰ ϲ͘ϱцϭ͘ϴϬ ϭ 2ODFDFHDH yŝŵĞŶŝĂĂŵĞƌŝĐĂŶĂ Ϭ͘ϱцϬ͘ϱϭ ϰ ϲ͘ϱцϭ͘ϴϭ ϭ  hŶŝƚϯŽŵďƌĞƚƵŵnjĞLJŚĞƌŝʹsĂĐŚĞůůŝĂƚŽƌƚŝůŝƐĐŽŵŵƵŶŝƚLJ &RPEUHWDFHDH ŽŵďƌĞƚƵŵnjĞLJŚĞƌŝ ϱ͘ϯцϮ͘ϲϰ Ϯϵ͘ϰ ϴ͘ϴцϰ͘ϳϰ Ϭ͘ϬϬϮϴ 0LPRVDFHDH sĂĐŚĞůůŝĂƚŽƌƚŝůŝƐ ϰϬ͘ϲцϭϮ͘Ϯϱ Ϯϵ ϭϴ͘ϭцϱ͘ϯϯ Ϭ͘Ϭϰϭϴ 7LOLDFHDH 'ƌĞǁŝĂƌĞƚŝŶĞƌǀŝƐ Ϯϴ͘ϵцϲ͘ϯϵ Ϯϴ͘ϱ ϭϳ͘ϰцϱ͘ϲϯ Ϭ͘ϬϱϬϲ ŶĂĐĂƌĚŝĂĐĞĂĞ ZŚƵƐƚĞŶƵŝŶĞƌǀŝƐ Ϯ͘ϴцϭ͘Ϭϵ ϭϵ͘ϱ ϭϬ͘ϯцϱ͘Ϭϱ Ϭ͘ϬϱϯϮ WĞůƚŽƉŚŽƌƵŵ &DHVDOSLQLDFHDH ϭ͘ϵцϬ͘ϳϲ ϭϳ͘ϳ ϵ͘ϵцϰ͘ϵϴ Ϭ͘ϬϳϰϮ ĂĨƌŝĐĂŶƵŵ 2FKQDFHDH KĐŚŶĂƉƵůĐŚƌĂ ϮϬ͘ϳцϭϭ͘ϰϬ ϭϮ͘ϰ ϭϮ͘ϵцϲ͘ϭϵ Ϭ͘ϰϯϭϭ (EHQDFHDH ŝŽƐƉLJƌŽƐůLJĐŝŽŝĚĞƐ ϭϲ͘ϮцϭϬ͘ϰϮ ϭϬ͘ϴ ϴ͘Ϭцϰ͘ϯϲ Ϭ͘Ϯϰϳϴ &DHVDOSLQLDFHDH ƵƌŬĞĂĂĨƌŝĐĂŶĂ Ϭ͘ϮцϬ͘Ϯϱ ϴ͘ϯ ϲ͘ϱцϭ͘ϴϭ Ϭ͘ϯϭϰϵ 0LPRVDFHDH sĂĐŚĞůůŝĂĞƌŝŽůŽďĂ ϯ͘ϭцϭ͘ϳϱ ϴ͘ϭ ϴ͘ϳцϰ͘ϱϯ Ϭ͘ϰϯϰϱ 0LPRVDFHDH sĂĐŚĞůůŝĂŬĂƌƌŽŽ ϲ͘ϰцϲ͘Ϯϯ ϱ͘ϴ ϲ͘ϱцϯ͘ϰϰ Ϭ͘ϰϮϭϵ  hŶŝƚϰ^ĞŶĞŐĂůŝĂĞƌƵďĞƐĐĞŶƐʹ^ĞŶĞŐĂůŝĂŵĞůůŝĨĞƌĂĐŽŵŵƵŶŝƚLJ 0LPRVDFHDH ^ĞŶĞŐĂůŝĂĞƌƵďĞƐĐĞŶƐ ϯϲ͘ϭцϵ͘ϰϮ ϱϱ ϭϳ͘ϵцϱ͘ϬϮ Ϭ͘ϬϬϬϮ 0LPRVDFHDH ^ĞŶĞŐĂůŝĂŵĞůůŝĨĞƌĂ ϯϭ͘ϰцϭϮ͘ϭϳ ϯϳ͘ϭ ϭϯ͘ϭцϰ͘ϳϰ Ϭ͘ϬϬϭϰ &DSSDUDFHDH ŽƐĐŝĂĂůďŝƚƌƵŶĐĂ ϭϬ͘ϭцϮ͘Ϭϭ ϭϴ͘ϱ ϭϳ͘ϭцϰ͘ϲϵ Ϭ͘ϯϭϭϳ &RPEUHWDFHDH ŽŵďƌĞƚƵŵŚĞƌĞƌŽĞŶƐĞ Ϯ͘ϮцϬ͘ϴϯ ϭϮ͘ϰ ϭϬ͘Ϯцϰ͘ϴϬ Ϭ͘Ϯϰϳ (EHQDFHDH ƵĐůĞĂƵŶĚƵůĂƚĂ ϭϴ͘ϵцϱ͘ϳϴ ϭϬ͘ϴ ϭϯ͘ϵцϱ͘ϰϭ Ϭ͘ϲϴϰϳ 5KDPQDFHDH ŝnjŝƉŚƵƐŵƵĐƌŽŶĂƚĂ Ϯ͘ϰцϭ͘ϭϱ ϴ͘ϭ ϵ͘Ϭцϰ͘ϰϵ Ϭ͘ϰϲϬϵ 0LPRVDFHDH sĂĐŚĞůůŝĂŶŝůŽƚŝĐĂ Ϭ͘ϯцϬ͘Ϯϴ ϲ͘ϵ ϳ͘ϰцϯ͘Ϯϱ Ϭ͘ϱϵϵϵ  hŶŝƚϱ'ƌĞǁŝĂĨůĂǀĂʹZŚŝŐŽnjƵŵďƌĞǀŝƐƉŝŶŽƐƵŵĐŽŵŵƵŶŝƚLJ 7LOLDFHDH 'ƌĞǁŝĂĨůĂǀĂ ϭϬϯ͘ϳцϭϰ͘ϱϰ ϱϲ͘Ϯ Ϯϯ͘Ϭцϯ͘ϵϮ Ϭ͘ϬϬϬϮ ZŚŝŐŽnjƵŵ %LJQRQLDFHDH Ϯϰ͘ϴцϴ͘ϴϳ ϯϯ͘ϱ ϭϰ͘Ϯцϱ͘ϳϮ Ϭ͘Ϭϭ ďƌĞǀŝƐƉŝŶŽƐƵŵ %XUVHUDFHDH ŽŵŵŝƉŚŽƌĂĂĨƌŝĐĂŶĂ ϰϯ͘Ϯцϴ͘ϴϲ ϯϯ͘ϯ Ϯϭ͘ϵцϱ͘ϵϮ Ϭ͘Ϭϰϴϲ 7LOLDFHDH 'ƌĞǁŝĂĨůĂǀĞƐĐĞŶƐ ϳ͘ϲцϯ͘Ϭϵ Ϯϰ͘ϱ ϭϮ͘ϵцϱ͘ϴϳ Ϭ͘ϬϱϬϰ 0LPRVDFHDH ^ĞŶĞŐĂůŝĂĨůĞĐŬŝŝ ϴ͘ϭцϯ͘ϵϲ ϭϲ͘ϱ ϭϬ͘ϯцϱ͘ϮϬ Ϭ͘ϭϮϱϮ &HODVWUDFHDH DĂLJƚĞŶƵƐďƵdžŝĨŽůŝĂ ϭ͘ϱцϭ͘ϭϱ ϲ͘ϱ ϲ͘ϲцϯ͘ϯϴ Ϭ͘ϰϲϱϳ 2ODFDFHDH yŝŵĞŶŝĂĐĂĨĨƌĂ Ϭ͘ϮцϬ͘ϭϴ ϰ͘ϳ ϲ͘Ϯцϯ͘Ϯϲ Ϭ͘ϲϴϲϯ

418 C A Biodiversity

Figure 2: Cluster analysis showing the fi ve plant communities identifi ed in Kgatleng District, Botswana. Terminalia sericea–Maytenus ten- uispina (1), Dichrostachys cinerea–Combretum apiculatum (2), Combretum zeyheri–Vachellia tortilis (3), Senegalia erubescens–Senegalia mellifera (4) and Grewia fl ava–Rhigozium brevispinosum (5).

1990). The eastern part of the district falls point along the transects, a 50 m × 20 m sation by maximum. The data were then in the hardveld whereas the north-western quadrat, the standard plot size agreed upon subjected to hierarchical cluster analysis part falls partly in the sandveld ecological for SASSCAL vegetation monitoring and (β linkage, β = –0.25, Sorensen distance) zone (Nsinamwa et al., 2005). mapping activities, was established. In (McCune et al., 2002) based on 39 species each vegetation plot, all woody (tree and distributed in the 77 plots. Indicator spe- Vegetation and soil sampling shrub) species rooted in the rectangle were cies analysis (Dufrene & Legendre, 1997) Vegetation sampling involved the use of counted and recorded by species. Fur- was used to defi ne meaningful vegetation transects cutting across diff erent vegeta- thermore, the percent cover of each spe- communities. Indicator values (IVS) were tion types using a geographical position- cies was estimated visually as described assessed for statistical signifi cance using ing system (GPS) receiver. Before fi eld in Bonham (1989). A total of 77 plots the Monte Carlo technique. Sorensen dis- work, satellite imagery from Google Earth were sampled (Fig. 1). Abundance data tance measure was used to examine dif- was used to identify diff erent vegetation for woody species was then expressed as ferences between vegetation communities types by contrasting colours along sys- individuals/ha. Nomenclature for scien- using a multi-response permutation pro- tematically placed transects running in a tifi c names follows van Wyk & van Wyk cedure (McCune et al., 2002). Vegetation south-north direction through the entire (1997) and Kyalangalilwa et al. (2013). communities were plotted in ordination length of the district. At each sampling At each sampling plot, soil samples were space using non-metric multidimensional point a survey pin was thrown over the also collected at a depth of 10–15 cm in scaling (NMDS). And the environmental shoulder, and where it landed it formed the middle of each plot. Soil samples were variables (sand particles [%], clay parti- the centre point of a vegetation plot. Sam- then analysed for particle size, soil pH, cles [%], silt [%], pH, EC, inclination, and pling points were established along tran- and electrical conductivity (EC) following exposition) were fi tted post hoc. All of sects at distances ranging from 5 km to standard laboratory procedures. these statistical analyses were performed 10 km apart depending on the diff erences in PCORD 6 (McCune et al., 2002). in the vegetation and land use (e.g., fenced Data analysis Species richness was calculated as the fi elds). The homogeneity or heterogeneity mean number of species occurring in the of the vegetation determined the spacing Vegetation classifi cation and sta- plots of each community. To test whether of the transects, and distances between tistical analyses species richness diff ers among communi- transects were increased when the vegeta- Cover data (%) of species for all 77 vegeta- ties, the data were analysed with one-way tion was homogenous. At each sampling tion plots were standardised using relativi- analysis of variance using SPSS 16.0.

B E 6 2018 419 106 sifi er. Five vegetation communities were mapped using ENVI 5.1, and the classi- 9DFQLO &RPPXQLWLHV fi cation results were imported into Arc- 7HUVHU0D\WHQ 'LFFLQ&RPDSL Map 10.3 for fi nalization of the map. &RP]H\9HFWRU 6HQHUX6HQPHO *UHZIOD5KLEUH 6SLDIU Results

&RPKHU 9DFHUL Diversity and abundance of woody species

0D\VHQ A total of 15 families, 23 genera, and 39 spe- cies were observed in the district (Tab. 1). =L]PXF 6HQPHO *UHIOV Fabaceae was the most species-rich family (XFXQG in the study area, and there were more spe- &URJUD *UHIOD

$[LV %RVIRH 7HUVHU 0D\WHQ %XUDIU cies of the subfamily Mimosaceae than of 5KLEUH Biodiversity the subfamily Caesalpiniaceae. The Com- 5KXWHV 6$1' &/$< 6,/W(OHFWURQ 2FKSXO S+ bretaceae family had 2 genera with a total

9DFNDU %RVDOE of 5 species. Only 1 species was observed ([SRVLWL 9DFWRU 'LRO\F ,QFOLQDW &RPDIU %DXSHW 'LFFLQ 0D\J\P for each of Bignoniaceae, Burseraceae,

*UHELF 6HQHUX Ochnaceae, Rhamnaceae, and Rubiaceae 0D\ER[ *UHUHW 6HQQLJ families in the district. 3HODIU The most abundant species in their re-

&RPDSL spective communities were Dichrostachys 6FOELU cinerea with 147 individuals/ha followed

;L P DP H by Grewia fl ava (103.7 ± 14.54 indiv./ha) and Terminalia sericea (96 ± 28.1 indiv./ $[LV ha). Overall, abundance of the genera Figure 3: Non-metric multidimensional scaling (NMDS) of the vegetation data, showing the Vachellia and Senegalia was 50.4 and 77.6 species (abbreviations show the fi rst three letters of the genus and the epithet), the plots individuals/ha, respectively (Tab. 1). coloured according to the fi ve identifi ed communities and the environmental variables. Bauhinia petersiana – Bahpet; Boscia albitrunca – Bosalb; Boscia foetida – Bosfoe; Burkea africana – Burafr; Combretum apiculatum – Comapi; Combretum hereroense – Comher; Vegetation classifi cation and Combretum imberbe – Comimb; Combretum zeyheri –Comzey; Commiphora africana – mapping Comafr; Croton gratissimus – Crograt; Dichrostachys cinerea – Diccin; Diospyros lycioides The cluster analysis identifi ed fi ve com- –Diolyc; Euclea undulata – Eucund; Grewia bicolor –Grebic; Grewia fl ava – Grefl a; Grewia munities (Fig. 2). They were named based fl avescens – Grefl a; Grewia retinervis – Greret; Maytenus buxifolia – Maybux; Maytenus senegalensis – Maysen; Maytenus tenuispina – Mayten; Ochna pulchra – Ochpul; Pelto- on the two species with the highest indi- phorum africanum – Pelafr; Rhigozum brevispinosum – Rhibre; Rhus tenuinervis – Rhuten; cator values: Community 1, Terminalia Sclerocarya birrea – Sclbir; Senegalia erubescens – Seneru; Senegalia fl eckii – Senfl e; sericea–Maytenus tenuispina; Communi- Senegalia mellifera – Senmil; Senegalia nigrescens – Sennig; Spirostachys Africana – Spiafr; Terminalia sericea – Terser; Vachellia erioloba – Vaceri; Vachellia karroo – Vackar; ty 2, Dichrostachys cinerea–Combretum Vachellia nilotica – Vacnil; Vachellia tortilis – Vactor; Vangueria infausta – Vaninf; Ximenia apiculatum; Community 3: Combretum Americana – Ximame; Ximenia caff ra – Ximcaf; Ziziphus mucronata – Zizmuc; Terser-May- zeyheri–Vachellia tortilis; Community ten: Terminalia sericea–Maytenus tenuispina, Diccin-Comapi: Dichrostachys cinerea-Com- bretum apiculatum, Comzey-Vactor: Combretum zeyheri-Vachellia tortilis, Seneru-Senmel: 4, Senegalia erubescens–Senegalia mel- Senegalia erubescens-Senegalia mellifera and Grefl a-Rhibre: Grewia fl ava-Rhigozium lifera; and Community 5, Grewia fl ava– brevispinosum. Rhigozium brevispinosum. The indicator species for each community are present- Means were considered signifi cantly dif- (Global Visualization Viewer) at a res- ed in Table 1. NMDS also confi rmed that ferent at P < 0.05. Species richness was re- olution of 30 m. The wet season scenes there were 5 distinct communities (Fig. ported as number of species per 1000 m2. were used because during that season the 3). Though most of the species were clus- Box plots of species richness were gener- vegetation is mature and the images show tered, the dominant indicator species for ated from the data using SPSS 16.0. the vegetation in full development. How- each community were clearly showing. ever, our sampling on the ground covered For example, Ximenia americana, Bos- Vegetation mapping both the dry and wet seasons. The scenes cia foetida, Combretum apiculatum, and A vegetation map was developed to show were atmospherically corrected using Dichrostachys cinerea occurred together, the spatial distribution of plant communi- FLAASH and mosaicked using seamless characterizing the Dichrostachys ci- ties. Four wet season scenes of Landsat 8 mosaic in ENVI 5.1. Supervised classi- nerea–Combretum apiculatum commu- Operational Land Imager of 2016 were fi cation was performed on the Landsat nity (Fig. 3). Terminalia serecea–May- freely downloaded from USGS GloVis images using maximum likelihood clas- tenus tenuispina community occurrence

420 C A Biodiversity

Figure 4: Vegetation map based on supervised classifi cation of Landsat 8 scenes, showing the fi ve communities identifi ed in Kgatleng District, Botswana. was infl uenced by the percentages of clay Senegalia mellifera community was mostly towards the northern parts of the and sand, whereas inclination and soil pH dominant and occurred over most parts district in the sandveld. infl uenced the occurrence of Combretum of the district, followed by the Dichros- zeyheri–Vachellia tortilis communities tachys cinerea–Combretum apiculatum Species richness (Fig. 3). community (Fig. 4). The Terminalia Species richness ranged, on average per Upon mapping the communities, it was sericea–Maytenus tenuispina commu- community, between 6.4 and 7.9 spe- observed that the Senegalia erubescens– nity covered the least area and occurred cies, and no signifi cant diff erences were

Figure 5: Box plots of the species richness of the fi ve communities in Kgatleng District, Botswana. The abbreviations are as explained: Terser-Mayten: Terminalia sericea–Maytenus tenuispina, Diccin-Comapi: Di- chrostachys cinerea-Combretum apiculatum, Comzey-Vactor: Combretum zeyheri-Vachellia tortilis, Seneru-Senmel: Senega- lia erubescens-Senegalia mel- lifera and Grefl a-Rhibre: Grewia fl ava-Rhigozium brevispinosum

B E 6 2018 421 detected among the fi ve communities increase in plant density (Belay et al., zeyheri–Vachellia tortilis, Senegalia

(ANOVA: F4,72 = 1.87, P = 0.125) (Fig. 5). 2013). In our study we observed high erubescens–Senegalia mellifera, and Interestingly, the Senegalia erubescens– woody-plant density in the north of the Grewia fl ava–Rhigozium brevispinosum. Senegalia mellifera community that dom- district, suggesting lessened anthropo- Woody species richness was not signifi - inated the district (Fig. 4) was among the genic impact farther away from major cantly diff erent between the communi- communities that had low species rich- villages in the south. However, species ties. The study provided a map of the ness (Fig. 5), even if it was not signifi cant- richness was similar across all commu- spatial distribution of the communities ly diff erent from other communities. Spe- nities. As species richness focused only in the district. The fi ndings of this study cies richness varied widely between plots on woody vegetation, it would probably will serve as baseline information that for the Combretum zeyheri–Vachellia tor- have been diff erent if all life forms were can be used for determining land suit- tilis community (Fig. 5). considered. ability for diff erent sectors of agriculture. Densities of individuals per hectare Indications of relatively high numbers The information also serves as baseline were 1022, 748, 896, 696, and 787 for of individuals from the genera Vachellia information that can be used in future to the fi ve communities, respectively. Den- and Senegalia may be supported by ear- determine whether encroachment is in- sities of plants did not diff er signifi cantly lier observations by other researchers creasing. Biodiversity among communities (ANOVA: F4,72 = (e.g., Moleele, 1998), who observed en- 1.73, P = 0.15). croacher species such as Senegalia mel- lifera to be dominant in the north-eastern Acknowledgements part of the district. The eastern part of Discussion the district, where bush encroachment The research was carried out in the has been identifi ed, comprises fertile framework of SASSCAL and was spon- Vegetation classifi cation and soils that support the growth of vegeta- sored by the German Federal Ministry mapping tion (Moganane, 1990). In the southern of Education and Research (BMBF) The Terminalia sericea–Maytenus ten- district of Botswana, Mosweu (2008) under promotion number 01LG1201M. uispina community typically occurs on also observed Vachellia spp. encroacher Research permit EWT 8/36/4 XXII (49) the sandveld north-western parts of the species to be more in areas that were was issued by Botswana’s Ministry of district, which is characterized by deep, highly utilized by livestock. Because the Environment, Wildlife and Tourism to sandy soils. Typically, Terminalia sericea current study covered a large area, aver- collect specimens and data for the pro- is deep rooted and occurs on well-drained age plant abundance was fairly low as a ject. Our thanks to Botswana University soils (van Wyk & van Wyk, 1997). Ac- result of many plots sampled with vari- of Agriculture and Natural Resources for cording to Moore & Attwell (1999), able abundances across the district, un- providing vehicles to be used for the pro- coarse-grained soils are associated with like Moleele’s (1998) study, which was ject. Special thanks also go to Andrew the occurrence of tree savannas. The localized in a high-abundance area for Moroka, who helped in the identifi ca- Dichrostachys cinerea–Combretum api- the encroacher species at Oliphants Drift. tion of plants. Consultation on vegetation culatum community dominated on the Earlier studies have indicated that an area classifi cation and mapping expertise of eastern side of the district and occurs is encroached if it has ≥ 2,500 individu- Keoikantse Sianga is also acknowledged. on the hardveld. This area is typical of als/ha (e.g., Richter et al., 2001). Thus, bush savanna, consisting of characteris- in the current study, bush encroachment tic Dichrostachys cinerea thickets mixed was not a concern. In contrast, in these with Grewia species and Combretum ecosystems, Vachellia and Senegalia References apiculatum (Aganga & Omphile, 2000). genera are important sources for browse Aganga, A.A. & Omphile, U.J. (2000) Forage In the centre of the district, the Senega- and fi rewood. Of relevance for the local resources of Botswana. Government Print- lia erubescens–Senegalia mellifera and communities are also the species of the ing and Publishing Services, Gaborone, Bo- Grewia fl ava–Rhigozium communities families Tiliaceae and Combretaceae, as tswana. Belay, T.A., Totland, Ø. & Moe, S.R. (2013) Eco- were dominant. These communities these provide edible berries and fencing system responses to woody plant encroach- comprise species such as S. erubescens, materials, respectively. ment in a semiarid savanna rangeland. Plant Ecology, 214, 1211. S. mellifera, and G. fl ava. Moleele (1998) Bhalotra, Y.P.R. (1987) Climate of Botswana, identifi ed these as encroaching on graz- part II: elements of climate: 1. Rainfall. De- ing areas in the district. Increases in these Conclusion partment of Meteorological Services, Bo- tswana. species may interfere with the predomi- Bonham, C.D. (1989) Measurement for terrestri- nant land use of cattle grazing as grasses Woody vegetation for Kgatleng District al vegetation. John Wiley & Sons, Inc., New are outcompeted. was composed of 15 families, 23 genera, York, USA. Britz, M.L. & Ward, D. (2007) Dynamics of and 39 species. The fi ve communities woody vegetation in a semi-arid savanna, Species richness identifi ed were Terminalia sericea–May- with a focus on bush encroachment. African Studies have observed that plant species tenus tenuispina, Dichrostachys cinerea– Journal of Range and Forage Science, 24, 131–140. diversity and richness increase with an Combretum apiculatum, Combretum

422 C A Dahlberg, A.C. (2000). Vegetation diversity and Skarpe, C. (1990) Structure of the woody vegeta- change in relation to land use, soil and rainfall tion in disturbed and undisturbed arid savanna, — a case study from north-east district, Bot- Botswana. Vegetatio, 87, 11–18. swana. Journal of Arid Environments, 44, Trollope, W.S.W. (1980) Controlling bush en- 19–40. croachment with fi re in the savanna areas of Dufrene, M. & Legendre, P. (1997) Species as- South Africa. Proceedings of the Annual Con- semblages and indicator species: the need for gresses of the Grassland Society of Southern a fl exible asymmetrical approach. Ecological Africa, 15, 173–177. Monographs 67, 345–366. Van Wyk, B. & van Wyk, P. (1997) Field guide to Kgosikoma, O., Mojeremane, W. & Harvie, B.A. trees of southern Africa, 1st ed. Struik Publish- (2012) Pastoralists’ perception and ecological ers, South Africa. knowledge on savanna ecosystem dynamics Ward, D. (2005) Do we understand the causes of in semi-arid Botswana. Ecology and Society, bush encroachment in African savannas? Af- 17, 27. rican Journal of Range and Forage Science, Khan, S.M., Harper, D., Page, S. & Ahmad, H. 22, 101–105. (2011) Species and community diversity of vascular fl ora along environmental gradient in Naran Valley: a multivariate approach through indicator species analysis. Pakistan Journal of Botany, 43, 2337–2346. Kyalangalilwa, B., Boatwright, J.S., Daru, B.H., Biodiversity Maurin, O. & Van Der Bank, M. (2013) Phy- logenetic position and revised classifi cation of Acacia s.l. (Fabaceae: Mimosoideae) in Af- rica, including new combinations in Vachellia and Senegalia. Botanical Journal of the Lin- nean Society, 172, 500–523. McCune, B., Grace, J.B. & Urban, D.L. (2002) Analysis of ecological communities. MjM Software Design, Gleneden Beach, OR, USA. Miyamoto, K., Suzuki, E., Kohyama, T., Seino, T., Miramanto, E. & Simbolon, H. (2003) Habitat diff erentiation among tree species with small-scale variation of humus depth and topography in a tropical heath forest of Central Kaliman, Indonesia. Journal of Tropical Ecol- ogy, 19, 43–54. Moganane, B.G. (1990) Soils and land suitability of Gaborone area. FAO/UNDP/Govternment of Botswana. BOT/85/011. Field Document. Moleele, N.M. (1998) Encroacher woody plant browse as feed for cattle: cattle diet composi- tion for three seasons at Olifants Drift, south- east Botswana. Journal of Arid Environments, 40, 255–268. Moore, A.E. & Attwell, C.A.M. (1999) Geologi- cal controls on the distribution of woody vege- tation in the central Kalahari, Botswana. South African Journal of Geology, 102, 350–362. Mosweu, S. (2008) Soil resources distribution, woody plant properties and land use in a lu- nette dune-pan system in Kalahari, Botswana. Scientifi c Research and Essay, 3, 442–456. Mudzengi, C.P., Kativu, S., Murungweni, C., Dahwa., Poshiwa, X. & Shoko, M.D. (2013) Woody species composition and structure in a semi-arid environment invaded by Dichros- tachys cinerea (l.) Wight and Arn (Fabaceae). International Journal of Scientifi c and Re- search Publications, 3, 1–10. Nsinamwa, M., Moleele, N.M. & Sebego, R.J. (2005) Vegetation patterns and nutrients in re- lation to grazing pressure and soils in the sand- veld and hardveld communal grazing areas of Botswana. African Journal of Range and For- age Science, 22, 17–28. O’Connor, T.G.O., Puttick, J.R. & Hoff man, W.T. (2014) Bush encroachment in southern Africa: changes and causes. African Journal of Range and Forage Science, 31, 67–88. Richter, C.G.F., Snyman, H.A. & Smith, G.N. (2001) The infl uence of tree density on the grass layer of three semi-arid savanna types of southern Africa. African Journal of Range and Forage Science, 18, 103–109.

B E 6 2018 423 References [CrossRef] Moleele, N.M. (1998) Encroacher woody plant browse as feed for cattle: cattle diet Aganga, A.A. & Omphile, U.J. (2000) Forage composition for three seasons at Olifants resources of Botswana. Government Drift, south-east Botswana. Journal of Arid Printing and Publishing Services, Gaborone, Environments, 40, 255–268. CrossRef Botswana. Moore, A.E. & Attwell, C.A.M. (1999) Belay, T.A., Totland, Ø. & Moe, S.R. (2013) Geological controls on the distribution of Ecosystem responses to woody plant woody vegetation in the central Kalahari, encroachment in a semiarid savanna Botswana. South African Journal of rangeland. Plant Ecology, 214, 1211. Geology, 102, 350–362. CrossRef Mosweu, S. (2008) Soil resources distribution, Bhalotra, Y.P.R. (1987) Climate of Botswana, woody plant properties and land use in a part II: elements of climate: 1. Rainfall. lunette dune-pan system in Kalahari, Department of Meteorological Services, Botswana. Scientific Research and Essay, 3, Botswana. 442–456. Bonham, C.D. (1989) Measurement for Mudzengi, C.P., Kativu, S., Murungweni, C., terrestrial vegetation. John Wiley & Sons, Dahwa., Poshiwa, X. & Shoko, M.D. (2013) Inc., New York, USA. Woody species composition and structure in Britz, M.L. & Ward, D. (2007) Dynamics of a semi-arid environment invaded by woody vegetation in a semi-arid savanna, Dichrostachys cinerea (l.) Wight and Arn with a focus on bush encroachment. African (Fabaceae). International Journal of Journal of Range and Forage Science, 24, Scientific and Research Publications, 3, 1– 131–140. CrossRef 10. Dahlberg, A.C. (2000). Vegetation diversity Nsinamwa, M., Moleele, N.M. & Sebego, R.J. and change in relation to land use, soil and (2005) Vegetation patterns and nutrients in rainfall — a case study from north-east relation to grazing pressure and soils in the district, Botswana. Journal of Arid sandveld and hardveld communal grazing Environments, 44, 19–40. CrossRef areas of Botswana. African Journal of Range Dufrene, M. & Legendre, P. (1997) Species and Forage Science, 22, 17–28. CrossRef assemblages and indicator species: the need O’Connor, T.G.O., Puttick, J.R. & Hoffman, for a flexible asymmetrical approach. W.T. (2014) Bush encroachment in southern Ecological Monographs 67, 345–366. Africa: changes and causes. African Journal CrossRef of Range and Forage Science, 31, 67–88. Kgosikoma, O., Mojeremane, W. & Harvie, CrossRef B.A. (2012) Pastoralists’ perception and Richter, C.G.F., Snyman, H.A. & Smith, G.N. ecological knowledge on savanna ecosystem (2001) The influence of tree density on the dynamics in semi-arid Botswana. Ecology grass layer of three semi-arid savanna types and Society, 17, 27. CrossRef of southern Africa. African Journal of Range Khan, S.M., Harper, D., Page, S. & Ahmad, H. and Forage Science, 18, 103–109. CrossRef (2011) Species and community diversity of Skarpe, C. (1990) Structure of the woody vascular flora along environmental gradient vegetation in disturbed and undisturbed arid in Naran Valley: a multivariate approach savanna, Botswana. Vegetatio, 87, 11–18. through indicator species analysis. Pakistan CrossRef Journal of Botany, 43, 2337–2346. Trollope, W.S.W. (1980) Controlling bush Kyalangalilwa, B., Boatwright, J.S., Daru, encroachment with fire in the savanna areas B.H., Maurin, O. & Van Der Bank, M. of South Africa. Proceedings of the Annual (2013) Phylogenetic position and revised Congresses of the Grassland Society of classification of Acacia s.l. (Fabaceae: Southern Africa, 15, 173–177. CrossRef Mimosoideae) in Africa, including new Van Wyk, B. & van Wyk, P. (1997) Field combinations in Vachellia and Senegalia. guide to trees of southern Africa, 1st ed. Botanical Journal of the Linnean Society, Struik Publishers, South Africa. 172, 500–523. CrossRef Ward, D. (2005) Do we understand the causes McCune, B., Grace, J.B. & Urban, D.L. (2002) of bush encroachment in African savannas? Analysis of ecological communities. MjM African Journal of Range and Forage Software Design, Gleneden Beach, OR, Science, 22, 101–105. CrossRef USA. Miyamoto, K., Suzuki, E., Kohyama, T., Seino, T., Miramanto, E. & Simbolon, H. (2003) Habitat differentiation among tree species with small-scale variation of humus depth and topography in a tropical heath forest of Central Kaliman, Indonesia. Journal of Tropical Ecology, 19, 43–54. CrossRef Moganane, B.G. (1990) Soils and land suitability of Gaborone area. FAO/UNDP/Govternment of Botswana. BOT/85/011. Field Document.