Ecological profiles of potential bush encroacher in the Manyeleti Game Reserve

G.J. Bredenkamp Department of Botany, Potchefstroom University for C.H.E., Potchefstroom

The habitat preferences of twelve potential bush encroacher Introduction species in the Manyeleti Game Reserve, situated in the Arid Low· veld veld type in the eastern Transvaal, were investigated. These An undesirable increase of woody in savanna, wood­ species have wide distribution patterns and owing to their land and bush veld has for a long time been an important presence in many sample plots, high frequencies were recorded in problem facing pasture scientists, cattle farmers and conser­ almost all classes of various habitat variables. Application of the Ecological Profiles Technique therefore resulted in disappointing vationists (Van der Schijff 1964). The economical implications results, as habitat preferences could not be established. The of bush encroachment are enormous, not only as far as actual technique was modified to process quantitative density data (individuals per hectare) and the resulting profiles were expressed combat against encroachment is concerned (Scott 1967; in terms of relative mean density per habitat class. The results Mostert et al. 1971; Werger 1974), but especially concerning were significant and habitat preferences were positively identified. meat production (Roux 1976; Van Vuuren 1979). Grouping of species with similar profiles lead to the identification of the following five indicator groups: Natural increase in density of woody plants in bush veld sericea, zeyheri, C. apiculatum and vegetation may be a result of higher rainfall (Walter 1971, exuvialis with preference to sandy, acid, leached, non-sodic soils with low conductivity. 1973; Bredenkamp 1976; Werger 1977; Van Vuuren 1979) Pterocarpus rotundifolius with preference to clayey, acid, medium during wet weather cycles (Dyer & Tyson 1977; Gertenbach leached, non-sodic soils with low conductivity. 1980). However, an abnormal undesirable increase in the Acacia gerrardii, A. nigrescens and Albizia harveyi with preference to clayey, alkaline, non-leached, sodic soils with high conductivity. density of woody species is probably caused mainly by a Oichrostachys cinerea, Oalbergia melanoxylon and Ormocarpum biotic factor such as overgrazing and trampling, and often trichocarpum with preference to very clayey, alkaline, moderately also by the incorrect application of fire (compare inter alia leached, slightly sodic soils with medium conductivity. May tenus senegalensis with preference to sandy, neutral (pH), Pienaar 1956; Van der Schijff 1964; Werger 1973, 1974; moderately leached soils with medium conductivity. Van Wyk 1974; Bredenkamp 1976; Gertenbach & Potgieter S. Afr. J. Bot. 1986, 52: 53 - 59 1979; Van Vuuren 1979; Van Vuuren & Bredenkamp 1980; Die habitatvoorkeure van twaalf potensiele bosindringerspesies in Dyer 1983). die Manyeleti-wildtuin, gelee in die Dorre Laeveld veldtipe in die Little is known about the ecology of bush encroacher Oos-Transvaal, is ondersoek. Hierdie spesies het 'n wye versprei­ ding en weens hulle teenwoordigheid in 'n groot aantal monster­ species in southern Africa. A knowledge of the habitat con­ persele, is hoe frekwensies in byna al die klasse van die habitat­ ditions favourable for the encroachment of undesirable species veranderlikes aangeteken. Die resultate van die Ekologiese can enable the land manager to devise ecologically sound Profiele-tegniek was dus teleurstellend aangesien habitat-voorkeure nie vasgestel kon word nie. Die tegniek is uitgebrei om management strategies for the prevention and control of bush kwantitatiewe digtheidsdata (individue per hektaar) te verwerk en encroachment. This study was therefore undertaken to deter­ die verkree profiele word in terme van relatiewe gemiddelde mine some of the relationships between twelve potential bush digtheid per habitatklas uitgedruk. Die resultate is betekenisvol en habitatvoorkeure is positief ge"ldentifiseer. Die groepering van encroacher species found in the Manyeleti Game Reserve and spesies met soortgelyke profiele het tot die identifisering van die certain soil variables. This Reserve is situated in the Arid volgende vyf indikatorgroepe gelei: Lowveld veld type (Acocks 1975) between 24°29' and 24°42'S, Terminalia sericea, Combretum zeyheri, C. apiculatum en Acacia exuvialis met 'n voorkeur vir sanderige, suur, geloogde, nie-natriese and 31 °03' and 31 ° 16'E. gronde met lae geleiding. From these relationships habitat preferences for the Pterocarpus rotundifolius met 'n voorkeur vir kleierige, suur, medium-geloogde, nie-natriese gronde met lae geleiding. particular species can be determined and distribution patterns Acacia gerrardii, A. nigrescens en Albizia harveyi met 'n voorkeur may be explained. As only scanty data on the grazing and vir kleierige, alkaliese, nie-geloogde, natriese gronde met hoe geleiding. fire history of the study area were available, it was decided Oichrostachys Cinerea, Oalbergia melanoxylon en Ormocarpum to exclude these factors from the habitat survey. trichocarpum met 'n voorkeur vir baie kleierige, alkaliese, effens geloogde, effens natriese gronde met 'n medium geleiding. Species included in this survey are Acacia exuvialis May tenus senegalensis met 'n voorkeur vir sanderige, neutrale, Verdoorn, A. gerrardii Benth. var. gerrardii, A. nigrescens medium-geloogde gronde. Oliv., Albizia harveyi Fourn., Combretum apiculatum Sond. S.-Afr. Tydskr. Plantk. 1986,52: 53-59 subsp. apiculatum, C. zeyheri Sond., Dalbergia melanoxylon Keywords: Bush encroachment, density data, ecological profiles Guill. & Perr., Dichrostachys cinerea (L.) Wight & Arn. subsp. nyassana (Taub.) Brenan, Maytenus senegalensis G.J. Bredenkamp (Lam.) Exell, Ormocarpum trichocarpum (Taub.) Engl., Department of Botany, Potchefstroom University for C.H.E., Potchefstroom, 2520 Republic of South Africa Pterocarpus rotundifolius (Sond.) Druce subsp. rotundifolius, and Terminalia sericea Burch. ex DC. Infraspecific taxa and Accepted 23 August 1985 author citations are not repeated in the text. 54 S.-Afr. Tydskr. Plantk., 1986, 52(1)

Methods class. In constructing an ecological profile of a species for a particular habitat variable, the normal procedures are: 1. Ecological Profiles Technique using frequency data 1. Establish the classes into which the variable may be The Ecological Profiles Technique (Godron 1965; Guillerm divided. 1971) is very useful to synthesize floristic and habitat data 2. Determine which sample plots are represented in each and to provide information on the ecological reactions of variable class. species to particular environmental factors (Morris & Guillerm 3. Determine the frequency of occurrence of the species in 1974; Bosch 1974; Bredenkamp 1982). A detailed description the sample plots represented in each variable class. of this technique is given by Morris & Guillerm (1974). Although application of the Ecological Profiles Technique The technique requires quantitative habitat data to establish resulted in valuable quantitative information on the ecological the different habitat classes for each variable, and qualitative behaviour of many of the herbaceous species that occur in (presence/absence) floristic data to calculate the frequency the study area (Bredenkamp 1982), the results concerning the (relative or corrected frequency) of the species for each habitat potential bush encroachers were, except for Tenninalia sericea,

Percentage clay Conductivity pH Potassium >- 600 t- 300 300 300 300 (f) - Z - t- W 200 200 200 200 0

100 100 100 100 Z

> - 200 200 200 t- -

Figure 1 Ecological profiles of Combretum zeyheri.

Percentage clay Conductivity pH Potassium >­ t-

(f) 300 Z

w 200 o

100 Z

300 300 > t­ 200 200

disappointing as habitat preferences could not be positively the ordinate scale of all the diagrams, all the absolute mean identified. This is ascribed to the fact that the technique density values were transformed to relative mean density processes only presence/absence floristic data to obtain (RMD), expressed as a percentage of the normal density. frequencies, and as the ecologically aggressive encroacher Consequently, a RMD of 100 may be considered as normal species were present in too many sample plots, no significant whereas a RMD of larger than 100 usually indicates habitat habitat preferences could be established. preference or may in some cases suggest encroachment. The ecological profiles (Figures 1 - 5) are all expressed in 2. Adaptation of the Ecological Profiles Technique to terms of RMD. Species with more or less similar profiles are utilize density data grouped into indicator groups (Morris & Guillerm 1974). An During the field survey it was observed that despite the wide indicator group is a collection of species with the same or distribution of the potential bush encroacher species, their similar ecological requirements. In this study indicator groups density and vigour varied considerably under different habitat were identified by means of tabular comparison. conditions. Vigorous plants were abundantly present in certain sample plots, whilst poorly developed individuals occurred 3. Data collection singly or scattered in other sample plots. Consequently the Floristic and habitat data were obtained from 245 stratified Ecological Profiles Technique was modified to utilize the random sample plots. These sample plots were also used for density data which were available for all woody species. The a Braun-Blanquet inventory-type vegetation survey (Breden­ normal procedures of the 'technique were followed to deter­ kamp 1982) and a survey and analysis of woody vegetation mine habitat classes, comprehensive profiles, equitability of structure (Bredenkamp & Theron 1985). sampling and mutual information shared by species and By using the variable plot size method of Coetzee & Gerten­ habitat variables. However, in the construction of the bach (1977) density data (rooted individuals per hectare) were ecological profiles, frequency of occurrence of species in the obtained in each sample plot for each woody species. As some individual habitat classes of a variable was not used, but the of these species may be multi-stemmed, care was taken to mean density of each species was calculated for each habitat identify rooted individuals as far as practically possible. Soil class. data used in this study included clay content, pH, conductivity In a large area a theoretical normal density of a species and exchangeable potassium, sodium, magnesium and calcium can be seen as the mean density of the particular species under of soil samples taken at a depth of 200 - 300 mm in each the habitat conditions where the species is normally found. sample plot. High density values, exceeding the theoretical normal density of a species would normally be indicative of the particular Results and Discussion habitat preference of the species within the relevant area. In 1. Habitat class'ils some cases, however, high densities could also indicate possible After the compilation of comprehensive profiles and the test encroachment. As the potential encroacher species were for equitability, habitat classes were distinguished for each present in many sample plots from a relatively large area, a variable. To enable convenient tabular comparison of normal density for each species was defined as the mean ecological profiles, these classes were grouped into low, density calculated from all sample plots. The frequency of medium and high categories. The results are given in Table 1. releves in the different classes of a variable, and especially the absolute mean density values per class may vary con­ 2. Ecological profiles of the species and indicator groups siderably, resulting in ecological profile diagrams with irregular The use of density data resulted in significant ecological ordinate (y-axis) scales. To smooth out variations caused by profiles. The tabular comparison of these ecological profiles great differences in absolute mean density, and to standardize resulted in the recognition of five indicator groups. These

>­ Percentage clay Conductivity pH Potassium r 300

(f) - z 200 w o 100

z onl <{ 01234567 01234567 012345 012345 w Sodium Calcium Magnesium ~ 300

w > 200 r 100 <{ ....J W o~ 0123456789 o I 2 345 6 7 8 9 o I 234 5 6 7 8 9 ~ a:: HABITAT C LAS S E S (see text for explanation)

Figure 3 Ecological profiles of Albizia harveyi. 56 S.-Afr. Tydskr. Plantk., 1986, 52(1)

Percentage clay Conductivity pH Potassium >­ 300 f-

(f) 200 Z w 100 o

z 01234567 01234567 0123 4 5 012345 « w Sodium Ca lcium Magnesium ~ 6001 w 300 > 200 f­ « -.J 100 w Il:: 0123 456789 123456789 012345678910 HABITAT C LAS S E S (see text for explanation)

Figure 4 Ecological profiles of Dalbergia melanoxylon.

>- Pe rcentoge cloy Conductivity f- 300 300

(f) Z 200 W

o 100

z o t-'-,-'--r'-,-'-"""'-'--r'-r'- « 0 12 34567 W

Sodium Calcium Magnesium 300 300 w > 200 200

r- f- r- r-r-r- « 100 100 -.J W o ,- l Il:: o I 2 345 6 7 8 9 Ol23 43678St 012345678910 HABITAT C LAS S E S (see text for explanation)

Figure 5 Ecological profiles of Maytenus senegalensis. groups as well as a summary of the habitat preferences of calcium and magnesium content, but the former two species the 12 potential bush encroacher species are given in Table 2. especially C. zeyheri give preference to potassium-rich soils. As the species within an indicator group exhibit similar Acacia exuvialis shows no significant preference with regard profiles, the profiles of only one representative species per to potassium and calcium content, but in contrast seems to indicator group are given (Figures 1 - 5). obtain a high RMD on soils with a high magnesium content. Results of Werger & Coetzee (1978), Bredenkamp (1982) and Indicator group 1 Coetzee (1983) indicate that Combretum apiculatum, C. This group consists of Terminalia sericea, Acacia exuvialis, zeyheri and Terminalia sericea have quite similar ecological Combretum apiculatum and C. zeyheri (Table 1). The proftles requirements, causing their associated occurrences. However, of C. zeyheri are given in Figure 1. All these species give a variation in their distribution pattern is related to a gradient definite preference to sandy, acid, leached, non-sodic soils, from deep sandy leached soils to shallow coarse leached soils. and severe encroachment occurs locally under these habitat Terminalia sericea is dominant on the deep sandy soils on the conditions especially with incorrect grazing and fire manage­ one end of the gradient and Combretum apiculatum is ment. Terminalia sericea, Combretum zeyheri and C. dominant on the shallow sandy soils on the other end of the apiculatum occur abundantly on granitic soils with low gradient. Intermediate situations on moderately deep sands S. Afr. J. BOL, 1986,52(1) 57

Table 1 The habitat classes distinguished for each habitat variable and the grouping of habitat classes into the categories low (L), medium (M) and high (H)

Habitat classes Variable 2 3 4 5 6 7 8 9 10 Clay (070) 0- 10 11 - 15 16 - 20 21 - 30 31 - 40 41 - 50 > 50 L L L M M M H

pH (H2O) 5,6 5,6-6,4 6,5 -7,4 7,5-8,4 >8,4 L L M H H Conductivity (mSm - I) 0 - 5,0 5,1 - 10,0 10,1 - 20,020,1 - 37,5 37,6-97,597,6- 200,0 > 200,0 L L L M M H H Potassium (mg 100 g - I) 0- 100 101 - 200 201 - 300 301 - 500 >500 L M M H H Sodium (mg 100 g-I) 0- 50 51 -75 76-100 101 - 125 126 - 175 176 - 250 251 - 400 401 - 600 >600 L L L M M M H H H Calcium (mg 100 g- I) 0 1- 125 126 - 250 251 - 375 376-500 501 - 625 626-1000 1001 - 2000 >2000 L L L M M M H H H Magnesium (mg 100 g - I) 0- 125 126-250 251-375 376 - 500 501 -750 751 - 875 876 - 1375 1376 - 2000 2001 - 3000 >3000 L L L L M M M H H H

Table 2 A summary of the habitat preferences of the twelve potential bush encroacher species, and the five indicator groups

Conduc- Factor Clay tivity pH K Na Ca Mg

Relative value L low L M H L M H L M H L M H L M H L M H L M H (see Table I) M = medium H = high Indicator groups I Terminalia sericea X X X X X X X Combretum zeyheri X X X X X X X Combretum apiculatum X X X X X X X Acacia exuvialis X X X X X

2 Pterocarpus rotundifolius X X X X X X X

3 Acacia gerrardii X X X X X X X X X Albizia harveyi X X X X X X X X X Acacia nigrescens X X X X X X

4 Dichrostachys cinerea X X X X X X X X X X X Dalbergia melanoxylon X X X X X X X X Ormocarpum trichocarpum X X X X X X X X

5 May tenus senegalensis X X X X X X X X are typified by the dominance of C. zeyheri (Werger and on the very clayey but well structured soils of the Shortlands Coetzee 1978). Acacia exuvialis attains prominence at the form, which is associated with well drained positions on shallow and moderately deep part of the gradient and is often gabbro. Sample plots from this area were included in the associated with Combretum· apiculatum and C. zeyheri. survey and this explains the high RMD values of this species in clay class 7 (Figure 2). Indicator group 2 A single species namely Pterocarpus rotundifolius represents Indicator group 3 this group (Table 2 and Figure 2). This species tends to have The microphyllous Acacia gerrardii, A. nigrescens and Albizia high RMD values on acid, very clayey soils with a medium harveyi represent this group (Table 2). The profiles of A. potassium, sodium and magnesium content. The very high narveyi are given in Figure 3. These species attain the highest RMD value on the very clayey soils is surprising as Ptero­ RMD values on alkaline, sodic, medium textured clay soil with carpus rotundifolius is often amongst the dominant or medium to high conductivity. These soils normally have subdominant woody plants on habitats with deep sandy soils medium calcium and magnesium contents. Acacia gerrardii (Werger & Coetzee 1978; Bredenkamp 1982; Coetzee 1983). gives preference to soils with low potassium content, while When clay class 7 (Figure 2) is ignored, the ecological profile Albizia harveyi prefers soils with a medium potassium content. of this species against clay content suggests a decrease in RMD These conditions are often associated with the poorly drained with an increase in clay content, with maximum RMD on soils of the Sterkspruit, Valsrivier, Kroonstad and Mayo soil soils with a clay content of 11-15070 (clay class 2, Figure 2). forms in the lowlands of the rolling granitic landscape (Werger In the study area, however, extremely high densities and & Coetzee 1978; Bredenkamp 1982; Coetzee 1983), or with severe encroachment of Pterocarpus rotundifolius are found the Arcadia soil form found on gabbro. Encroachment of 58 S.-Afr. Tydskr. Plantk., 1986, 52(1) the species of this group could be expected under these habitat BOSCH, O.J.H. 1974. Die wisselwerking tussen die habitat en 'n conditions and has been observed locally but here the aantal grasgemeenskappe in die suidoostelike Oranje-Vrystaat. D.Sc. thesis, Potchefstroom University for Christian Higher individual plants are mostly dwarfed. Education. BREDENKAMP, G.J. 1976. Bosindringing in die Manyeleti­ Indicator group 4 wildtuin - 'n voorlopige verslag. Committee for Develop­ This group includes Dichrostachys cinerea, Dalbergia mental Research, Department of Co-operation and Develop­ melanoxylon and Ormocarpum trichocarpum. The profiles ment, Pretoria. of Dalbergia melanoxylon are given in Figure 4. These species BREDENKAMP, G.J. 1982. 'n Plantekologiese studie van die give preference to alkaline soils with a medium to high clay Manycleti-wildtuin. D.Sc. thesis, University of Pretoria. BREDENKAMP, G.J. & THERON, G.K. 1985. A quantitative content and medium conductivity. The sodium content of approach to the structural analysis and classification of the these soils is medium but the potassium, magnesium and vegetation of the Manyeleti Game Reserve. S. Afr. J. Bot. 51: calcium contents are usually high (Table 1). These habitat 45 -54. conditions are often associated with the Arcadia soil form COETZEE, B.J. 1983. Phytosociology, vegetation structure and found on gabbro. Serious encroachment of the species of this landscapes of the central district, , South Africa. Dissertationes Botanica 69: I - 456. group seems to occur locally. As in the case of indicator group COETZEE, B.J. & GERTENBACH, W.P .D. 1977. Technique for 3, the individual plants are often dwarfed on the very clayey describing woody vegetation composition and structure in soils. Dense impenetrable bush is often found on clayey but inventory type classification, ordination and animal habitat better drained red structural soils. surveys. Koedoe 20: 67 - 75 . Werger & Coetzee (1978) mention various examples to DYER, C. 1983 . Factors associated with the encroachment of bush indicate that Dichrostachys cinerea subsp. africana finds its into semi-arid savanna grazing lands in southern Africa. M.Sc. thesis, University of the Witwatersrand. optimum habitat on deep, sandy but trophic soils, usually DYER, T.G.J. & TYSON, P.D. 1977. Estimating above and situated at the bottom of sandy slopes, towards nutrient"rich below normal rainfall periods over South Africa, 1972 - 2000. lowlands or plains. In these situations the B"horizons of the Jnl appl. Met. 16: 145 - 147. soils are usually more clayey. D. cinerea subsp. africana is GERTENBACH, W.P.D. 1980. Rainfall patterns in the Kruger also abundant on strongly red coloured soils, indicative of National Park. Koedoe 23: 35 - 43. good aeration and drainage and these two soil conditions GERTENBACH, W.P .D. & POTGIETER, A.L.F. 1979. Veld­ brandnavorsing in die struikmopanieveld van die Nasionale together with at least a fair amount of soil nutrients appear Krugerwildtuin. Koedoe 22: 1 - 28. to be the common denominators for typical optimal D. cinerea GODRON, M. 1965. Les principaux types de profil ecologique. subsp. africana habitats (Werger & Coetzee 1978). CNRS, CEPE. GUILLERM, J.-L. 1971. Profile ecologiques et information Indicator group 5 mutuelle entre especes et facteurs ecologiques. Decol. . 6: Maytenus senegalensis is the only species in this group (Table 1). 209 - 225. The ecological profiles are given in Figure 5. This species MORRIS, J.W. & GUILLERM, J .-L. 1974. The ecological prefers sandy but neutral to alkaline soils with low to medium profiles technique applied to data from Lichtenburg, South Africa. Bothalia 11: 355 - 364. conductivity. The potassium, calcium and magnesium contents MOSTERT, J.W.c., ROBERTS, B.R., HESLINGA, C.F. & of these soils are in the medium range, but no preference can COETZEE, P.G.F. 1971. Veld management in the OFS be indicated with regard to the sodium content. Region. Dept. Agric. Techn. Servo Bull. 39: I - 97 . PIENAAR, A.J. 1956. Veld-burning in the grassveld areas of the Conclusion Union of South Africa. S.A.R.C.C.U.S. Feb. 1956. Despite the imperfect nature of relative mean density values ROUX, P.W. 1976. Opening address. Proc. Grassl. Soc. S. Afr. per se, significant ecological profiles resulted from the 1'1: 11-13. procedures using density data. Habitat preferences were SCOTT, J.D. 1967. Bush encroachment in South Africa. S. Afr. J. Sci. 63: 311-314. established for all the potential bush encroacher species VAN DER SCHUFF, H.P. 1964. 'n Herevaluasie van die investigated. The results do not indicate directly whether or probleem van bosindringing in Suid-Afrika. Tydskr. Natuurw. not bush encroachment has taken place as 'normal' densities 4: 67 -80. of the different species under the normal range of habitat VAN VUUREN, D.R.J. 1979. Kom ons gesels oor die conditions are not known. Furthermore, no definite criterion bosveldbome. Borne S. Afr. 31: 9 - 17. exists for the determination or measurement of encroachment. VAN VUUREN, D.R.J. & BREDENKAMP, G.J . 1980. The role of Botany in developing communities in South Africa. In: However, a species may probably obtain an undesirable high Where theory meets practice: The Faculty of Science ih service density under habitat conditions most favourable for its of the community, ed. Els, W.J. Publications of the University specific ecological requirements, especially when competition of the North Series B19: 118 - 132. is reduced by for example severe grazing or fire. The results VAN WYK, P. 1974. Veldbrand as weiveldbeheermetode. Custos of this study may contribute to a management strategy as 3: 6-9, 3: 12-21,3: 17-21. areas susceptible to encroachment by the particular species WALTER, H. 1971. Ecology of tropical and subtropical vegetation. Oliver & Boyd, Edinburgh. may be identified and managed to prevent increase of WALTER, H. 1973. Die Vegetation der Erde in undesirable species. okophysiologischer Betrachtung. Bd. 1. Die tropischen und subtropischen Zonen. 3 Aufl. Fischer Verlag, Stuttgart. Acknowledgements WERGER, M.J.A. 1973. Phytosociology of the Upper Orange This study formed part of a wider study on plant ecology River Valley, South Africa. V & R, Pretoria. of the Manyeleti Game Reserve and was partly funded by WERGER, M.J.A. 1974. Effects of game and domestic livestock the Committee for Developmental Research of the Depart" on vegetation in East and Southern Africa. In: Handbook of Vegetation Science 13, ed. Krause, W. Dr. W. Junk, The ment of C~operation and Development. Hague. WERGER, M.J.A. 1977. Environmental destruction in southern References Africa: the role of overgrazing and trampling. In: Vegetation ACOCKS, J.P.H. 1975. Veld Types of South Africa. Mem. bot. Science and environmental protection, eds Miyawaki, A. & Surv. S. Afr. 40: 1-128. Tiixen, R.M. Proc. Int. Symp. Prot. Envir. Tokyo 1974: S. Afr. J. Bot., 1986, 52(1) 59

301 - 305 . Maruzen Co., Tokyo. Zambezian Region: In: Biogeography and Ecology of Southern WERGER, M.l .A. & COETZEE, B.l. 1978. The Sudano- Africa, ed . Werger, M.l.A. Dr. W. lunk, The Hague.