Woody plants surrounding treeless lawns on sodic substrates in southern Kruger National Park, part 1: floristic composition

In the southern parts of Kruger National Park in South Africa, the vegetation features nearly treeless patches of vegetation on apparently sodic substrates (https://koedoe.co.za/index.php/koedoe/article/view/1112/1467 and https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1365-2028.2004.00532.x and https://www.researchgate.net/publication/227939381_Proposed_mechanism_for_the_origin_of_sodic_patches_in_Kruger_National_Park_South_Africa and http://www.scielo.org.za/scielo.php?script=sci_arttext&pid=S0075-64582020000200008 and https://www.semanticscholar.org/paper/A-comparison-of-grass-production-and-utilisation-in-Alard/e6b3c536c6fe3a756997ddc1c7228d34f3d35049 and https://wiredspace.wits.ac.za/handle/10539/7797 and https://www.krugerpark.co.za/krugerpark-times-4-1-krugers-secrets-24069.html and file:///C:/Users/Antoni%20Milewski/Downloads/land-11-01148-v2%20(1).pdf).

These are associated with nocturnal gregariousness of the impala (Aepyceros melampus, https://www.inaturalist.org/taxa/42278-Aepyceros-melampus).

The nearly treeless patches are effectively grazing lawns, free of wildfire. The impala seems to congregate on them at night, throughout the year.

In Sept.- Nov. 2016, I participated in an intensive study of the woody plants in and adjacent to these lawn-like patches (https://onlinelibrary.wiley.com/doi/10.1111/aje.12898?af=R and https://onlinelibrary.wiley.com/doi/abs/10.1111/aje.12898 and https://ouci.dntb.gov.ua/en/works/4KnpeZq7/).

The sampling area lay along the N’waswitshaka River, southwest of Skukuza (https://en.wikipedia.org/wiki/Skukuza).

Our sampling of the sodic patch complex consisted of sets of plots, one set being located in the patches most devoid of woody plants, and the other set being located in the densest stands of woody plants immediately adjacent. This approach ensured that the full contrast was captured in our data.
 
A long list of woody species was found in the sample plots. However, the main woody genera are Spirostachys, Euclea, Senegalia, and Vachellia.
 
Combretum is present but not typical. In this way the sodic patches and their woody surrounds deviate from the typical granitic vegetation of the region, which is typically combretaceous.
 
In terms of plant height, the greatest contrast with the treeless lawns is provided by Senegalia nigrescens (and Combretum imberbe where it occurs in this system, which is not typical).

In terms of the density of woody cover, the greatest contrast is provided by Spirostachys africana, which forms patches of virtual forest adjacent to the treeless lawns.

In terms of suppression of woody plants by herbivores, the greatest contrast is provided by Vachellia grandicornuta. The growth-form of this species differs remarkably between the lawns and their surrounds (see part 2 of this series.)
 
The formation of sodic patches in the sampling area seems to resolve, at least in part, to an intense contest between S. africana and lawn graminoids, e.g. Sporobolus nitens (https://www.inaturalist.org/taxa/594677-Sporobolus-nitens).

Spirostachys africana (http://pza.sanbi.org/spirostachys-africana) is extreme as a woody plant, because it is both toxic and clonal. It parallels lawn-forming grasses in being able to spread laterally by vegetative means. It also has the advantage of being so toxic that this protects even its dead wood.

The impala, acting directly, can do no more than hedge S. africana slightly. This is because it

• has no means of breaking the stems,
• cannot reach the branches higher than 1.5 m,
• cannot eat the toxic leaves in more than small quantities, and
• may eat this deciduous species mainly in the form of shed leaves on the ground.

Sprirostachys africana should be able to exclude grasses by means of clonal propagation. However, it seems to lose the contest with graminoids, via rhizal competition.

Spirostachys looks like an ectomycorrhizal plant, but is not, instead having similar arbuscular mycorrhizae (https://en.wikipedia.org/wiki/Arbuscular_mycorrhiza) to those of grasses and sedges.

The rhizal system of the lawn seems powerful enough to exclude that of S. africana, despite the fact that the graminoids are kept so short year-round, by the impala, that the photosynthetic biomass of the lawn is minimal.

WOODY FLORA
 
Spirostachys africana is abundant, and in some plots dominant, adjacent to the patches of lawn in the sodic complex. However, it did not penetrate any of the treeless sampling plots.

Spirostachys africana seems largely immune to consumption by the African bush elephant (Loxodonta africana), owing to its toxicity. It compensates for any breakage by cloning, and it can partly defend itself from the impala by chemical means. However, it is incapable of penetrating the lawns more than marginally with its mode of vegetative spreading.

At the edges of the patches of lawn, S. africana is sparsely/patchily present as hummock-like ‘shrubs’ suppressed to < 50 cm high by the impala (see part 2).
 
Euclea divinorum is patchily common in the sodic patch complex. Euclea natalensis is also common, although never found away from its more general congener. However, neither species occurred in or next to our treeless plots in the grazing lawns. Euclea divinorum is typical of smaller sodic patches elsewhere in the southern half of Kruger National Park, where it seems to abut the lawns.
 
Senegalia nigrescens is fairly common as an emergent tree in the sodic patch complex. It occurred also as miniaturised, suppressed saplings (see part 2) in some of our treeless plots.

Senegalia nigrescens can survive disfigurement and felling by the African bush elephant. Indeed, it seems to be particularly persecuted by the proboscidean, which is common in the area, judging from the amount of dung we observed, and the obvious breakage of the boles/branches of various spp. of trees and shrubs.
 
Vachellia grandicornuta, which has the potential to grow into a tree about 6 m high, is present on these lawns. However, this is only as suppressed saplings about 10 cm high (see part 2).

Vachellia grandicornuta is dominant in patches, and common in miniaturised form in the treeless plots. Its presence in the latter is typical, although it nowhere managed to make up a woody cover of >1 % in our plots on the grazing lawns.

Vachellia grandicornuta gives the impression of being a short-lived pioneer that grows rapidly, and soon dies and collapses. The main factor limiting V. grandicornuta seems to be browsing by the impala at the sapling stage. This acacia defends itself with straight, stipular spines. However, it seems to surpass herbivory only periodically, in cohorts decades apart.
 
Vachellia tortilis is scarce throughout the sodic patch complex, but conspicuous in places. It occurs mainly in the vegetation intermediate between treeless lawns and the nearby woody vegetation.

Dichrostachys cinerea is common, but mainly as suppressed saplings (see part 2).

Manilkara mochisia, although occurring only sporadically, dominated one plot. It has a noteworthy growth-form, indicating great persistence. I.e. it occurs in a few remarkably developed stands, but it is far from being the kind of plant that could usurp the areas of lawn.
 
Grewia bicolor is common as small, suppressed plants, extending into the patches of lawn, and represented within the treeless plots as a few individuals. The shrubs of G. bicolor (which do reach reproductive size) are small and sparse. Grewia hexamita and G. flavescens are common as small plants, but seem incapable of persisting in the treeless lawns in the way performed by G. bicolor.
 
Pappea capensis is a feature of the sodic patch complex. However, it does not dominate even small patches of vegetation, being better-described as conspicuous, stout trees standing more or less alone, in the area intermediate between the treeless lawn and the surrounding wooded vegetation. Pappea capensis was not important in any plot.

Diospyros mespiliformis is fairly common, but not well-developed. It is far from the case that mature trees occurred in our plots.
 
Combretum imberbe occurred sparsely. However, such specimens as I did encounter in the plots are well-grown.
 
The following species, while taking much time to record in detail in our plots, amount to little other than to show the floristic diversity of the woody vegetation associated with the sodic patches:

Berchemia discolor, Gymnosporia maranguense, G. senegalensis, Elaeodendron transvaalense, Sclerocarya birrea (extremely scarce and always only saplings), Cordia sp., Ormocarpum trichocarpum, Gardenia volkensii, Strychnos madagascariensis, Allophyllus sp., Zizyphus mucronata, Searsia gueinzii, Philenoptera violacea, Pyrostria hystrix, Schotia brachypetala, Pavetta sp., Balanites maughamii, Combretum hereroense, Flueggea virosa, Dalbergia sp., Commiphora africana, Capparis tomentosa, and possibly Maerua parvifolia.

Noteworthy for their complete absence are Peltophorum africanum, Terminalia spp., and Euphorbia spp.

A species of faboid legume is common as low individuals (<20 cm) in the treeless plots (see part 2). It is semi-herbaceous rather than strictly woody, but was included in all assessments of woody cover in the treeless plots, because it did possess a significantly lignified stem.
 
The minimal presence of Sclerocarya birrea in the sodic patch complex is noteworthy. This species is excluded almost as categorically as is Combretum apiculatum. 

The treeless lawns give way abruptly to any of three main spp. of trees. These are as follows.

• Vachellia grandicornuta forms open savanna about 6 m high;
• Senegalia nigrescens  can reach >15 m, and is emergent above tall shrubs such as Euclea spp.; and
• Spirostachys africana is awkward to describe in terms of stratification, because it is clonal, and suckers to form dense stands that constitute patches of low open-forest.

It is remarkable that none of these formidable woody species seems capable of usurping space in the patches of lawn maintained by the impala.  
 
In summary so far:
The woody communities adjacent to the treeless lawns ranged from Vachellia grandicornuta open savanna and Euclea divinorum ‘thicket’, to Spirostachys africana open-forest, with Manilkara mochisia copses as an interesting variation.

I have named each of these types here as if it is a different vegetation type. However, the scale of the patches is so small that the real vegetation type is this whole complex, constituting a single patchy/heterogeneous entity.
 
HEIGHT OF WOODY PLANTS:

Senegalia nigrescens occurs in the woody plots as an emergent, reaching far greater heights than does Vachellia grandicornuta.

It is noteworthy that Senegalia, Vachellia, Dichrostachys, and Spirostachys, all typical of the granitic catena in southern Kruger National Park, all penetrate the treeless patches in extremely suppressed form (continually clipped by the impala).

However, Euclea does not manage to penetrate the treeless patches. This genus, although characteristic of the immediate vicinity of the lawns, is noteworthy in failing to survive (even as suppressed plants) in the treeless vegetation.

Also failing to survive in the treeless lawns, of course, is Combretum imberbe.

The tallest plants in each of the 20 woody plots are as follows:
 
Plot 11: Senegalia nigrescens 18 m 
12: Spirostachys africana 9 m
13: Senegalia nigrescens 15 m 
14: Senegalia nigrescens 16 m, over Spirostachys africana about 6 m 
15: Spirostachys africana 8.5 m 
16: Senegalia nigrescens 8 m 
17: Spirostachys africana 7 m 
18: Euclea divinorum 3.5 m (this species is a multi-stemmed tall shrub) 
19: Euclea divinorum 2 m (reduced from the original woody stems owing to dieback in drought) 
20: Combretum imberbe 10 m, over Euclea divinorum about 3.5 m 
10 b (which replaces a plot done mistakenly by Zurelda & Jessica): Spirostachys africana 6.5 m 
22: Senegalia nigrescens 13 m 
23: Senegalia nigrescens 14 m 
24: Spirostachys africana 10.5 m 
25: Euclea natalensis 6 m (this species is a multi-stemmed tall shrub) 
26: Combretum imberbe 16 m, over Euclea divinorum 4-5 m 
27: Combretum imberbe 10 m, over Pappea capensis 7.5 m 
9 b: Spirostachys africana 8.5 m 
28: Vachellia grandicornis 7.5 m 
29: Manilkara mochisia 5 m (would be taller if the massive trees were not prostrate)
 
Please note that Sclerocarya birrea is virtually absent from the sodic patch complex, although re-entering immediately adjacent, where the usual vegetation on granitic substrates resumes.
 
The tallest plants in our plots vary from 2 m to 18 m, which is a large range. However, the vegetation was consistently dense. Thus, if one considers both the height and cover of woody plants, the contrast in vegetation structure in these pairwise comparisons, of lawn and nearby woody stands, is as great as occurs anywhere in Kruger National Park.

I had the impression that, if the impala were to be suddenly removed, the suppressed acacias in the treeless lawn would soon grow out and take over the vegetation.
 
Summarising the findings, from a slightly different perspective:
 
On granitic catenas in southern Kruger National Park, it is common for the tallest plants to be Senegalia nigrescens and Sclerocarya birrea, leading to the ‘knobthorn-marula’ designation. These trees grow as emergents over Combretum, typically apiculatum. On basalt in the central part of the Park, Senegalia nigrescens is common, but the tallest plant is often Combretum imberbe rather than S. birrea.

I suspect that the sodic patches correspond to dolerite dykes, which might explain why, in some ways, the emergents in the immediate vicinity have affinities with both granite and basalt. The woody vegetation around the treeless lawns certainly retains S. nigrescens. However, C. imberbe replaces S. birrea.

Combretum apiculatum was completely absent from sodic patch complex. Furthermore, it may be significant that C. hereroense (which I found as an occasional individual in the woody plots around the treeless patches) is more typical of gabbro than of granite in southern Kruger National Park.

Spirostachys africana, which occurs on the granitic catena, is typical of the clay-rich soils of the lower part of this catena. Therefore, its abundance in the woody vegetation around the treeless patches would be consistent with a possibly doleritic bedrock here.
 
To summarise further:
 
On and near the treeless patches in southern Kruger National Park, the normal savanna trees seem, at first glance, to be replaced by multi-stemmed tall shrubs of Euclea (which resembles mallee in Australia, https://en.wikipedia.org/wiki/Mallee_(habit)).

However, what is more important is that the vegetation dominated by tall (>4 m) woody plants, adjacent to the lawns, differs from the vegetation of the granitic catena in lacking Combretum apiculatum and Sclerocarya birrea, and in containing e.g. Pappea capensis and Manilkara mochisia.
 
I suspect that the origin of the sodic patches is related to the incidence of dolerite dykes in the granitic landscape. Certain aspects of the floristic composition of the woody vegetation around the treeless lawns, found in thisnstudy, do seem consistent with a doleritic origin.

Senegalia nigrescens, the tallest species in many of our woody plots, would of course occur on dolerite, just as it does on both granite and basalt. Spirostachys africana would of course be expected to be abundant on dolerite. And the absence of Combretum apiculatum would be consistent with a doleritic origin, because that species is restricted to the relatively sandy profiles of granite.

CANOPY COVER OF WOODY PLANTS 

Please bear in mind that we arbitrarily chose plots to represent dense woody vegetation, e.g. centering a plot on a copse of Manilkara mochisia when an adjacent plot, located just to the side of the same copse, would have had far lesser canopy cover of woody plants.

The following summarises the canopy cover values in the woody plots of the pairwise comparisons of sodic patches along the N’waswitshaka River in the southern Kruger National Park. (These values need checking from my field notes.)
 
Plot 11: 80%, 12: 85%, 13: 65%, 14: 87%, 15: 65%, 16: 77%, 17: 80%, 18: 62%, 19: 35%, 20: 37%, 10b: 62%, 22: 70%, 23: 80%, 24: 65%, 25: 75%, 26: 65%, 27: 40%, 9b: 40%, 28: 43%, 29: 62%.
  
The range for canopy cover in these plots is 35-87%. The lowest value referring to a plot dominated by Euclea divinorum, and the highest value referring to a plot dominated by Spirostachys africana.

Nevertheless, it is clear that the stands of Euclea near the treeless lawns of sodic patches do not achieve anything close to the density of patches of Spirostachys.

The ‘mallee’ of Euclea would not even qualify as thicket, whereas the densest clonal stands of Spirostachys would certainly qualify as forest if they occurred on a broader scale, as opposed to just a patchy aspect of the total complex of woody plants near the treeless lawns.

There is a difference between this ‘mallee’ of Euclea and the Spirostachys in all three: growth-form, plant height and canopy cover. In the ‘mallee’ stands, the plants are tall shrubs, not trees, because they are virtually always multi-stemmed, and died back to the lignotuberous base in the recent drought.

In the clonal patches of ‘forest’ of Spirostachys, each sucker tends to be single-boled at ground level.

Both species may prove clonal. However, the point is that the modes of regeneration and emergence from the ground are different.

Euclea provides a handy visual marker for sodic patches in Kruger National Park. However, it is Spirostachys that really epitomises the contrast between the potential woody cover of this environment and the suppression/exclusion of woody plants from the treeless lawns.

What we have, in the final analysis, is a contrast between treeless lawns, some plots in which contained zero woody plants and thus had canopy cover 0% for woody plants, and monopolisation by clonal Spirostachys in tree form, with something approaching 90% canopy cover in a 10 m X 10 m plot.

This is surely about as extreme a contrast in the incidence of woody plants as one can expect to find, within a single landform and substrate type, anywhere in Kruger National Park.
 
It is worth checking whether even Euclea is technically clonal. However, even if Euclea and Spirostachys are both mainly clonal regenerators (using suckers as opposed to relying on germinative regeneration followed by vegetative regeneration from well-grown woody structures), their growth forms differ.

Thus, in the environment of the sodic patches in southern Kruger National Park, the distinction between patches of ‘mallee’ of Euclea and patches of Spirostachys correlates with different degrees of coverage of the ground by the crowns of the woody plants.

Both Euclea and Spirostachys are extremely defended from herbivory. Therefore, their competitive relationship with graminoids is arguably more important than their ability to withstand, and recover from, gross breakage by the African bush elephant and other large herbivores (which is more important in the case of the acacias).
 
The bottom line:
Within the sodic patch system, we can go from complete exclusion of all woody plants (including even the semi-woody Indigofera and the pliable-stemmed low quasi-karoid which has yet to be identified) in some plots, to something approaching 90% coverage of the area by the crowns of not just woody plants, but what are to all intents and purposes trees (of Spirostachys) – and with dense wood at that.

I do not know which grasses and other herbaceous manage to occur in these densest stands of Spirostachys. However, these are few indeed. This really is the crux of this study. What we have in the sodic patch complex is a contrast between 100% herbaceous in some 10 m X 10 m plots, and virtually zero herbaceous (even including lianes) in some clonal stands of Spirostachys nearby, on what is ostensibly the same landform and substrate.

PREDOMINANCE OF SPIROSTACHYS (https://en.wikipedia.org/wiki/Spirostachys):
 
Spirostachys africana proved to be important in our contrast between the treeless lawns and the tall, dense woody vegetation associated with sodic patches. This species spreads clonally, and tends to crowd out other plant taxa, both woody and herbaceous.

Of a particular set of 20 plots in woody vegetation near the treeless lawns, 8 were dominated by S. africana.
 
What this means is that about 40% of the woody plots in the sodic patch complex were located in what are effectively patches of forest of S. africana.

I have analysed the plot data with the following questions in mind:

• in these 8 plots, which other species of woody plants, other than S. africana, occurred, and
• how do these fit into the picture, in terms of their niches?

The following lists the spp. present in these plots, besides the dominant S. africana:

• mimosoid legumes: Senegalia nigrescens, Vachellia grandicornuta, Dichrostachys cinerea 
• faboid legumes: Ormocarpum trichocarpum, Philenoptera violacea 
• Caesalpinioid legumes: Schotia brachypetala
• Ebenaceae: Euclea divinorum, Euclea natalensis, Diospyros mespiliformis
• Celastraceae: Elaeodendron transvaalense, Gymnosporia maranguense
• Anacardiaceae: Searsia gueinzii
• Burseraceae: Commiphora africana
• Sapindaceae: Pappea capensis
• Boraginaceae: Cordia ovalis or monoica, Ehretia obtusifolia
• Combretaceae: Combretum hereroense
• Rubiaceae: Pavetta sp., Pyrostria hystrix, Gardenia volkensii
• Tiliaceae: Grewia hexamita, Grewia bicolor, Grewia flavescens
• Brassicaceae: possibly Maerua parvifolia (easily confused with juvenile Pyrostria hystrix).

Given the length of this list, readers may be wondering:
How is such floristic diversity compatible with the dominance by one species, namely S. africana?
 
The short answer is that, apart from emergent large trees of Senegalia nigrescens, most of the other species occur only as juvenile or suppressed individuals, contributing little to canopy cover.
 
It may be helpful to categorise the above species by nutritional modes and dispersal strategies.
 
The first category is the N-fixers.

This is relevant because

• the dominant species, S. africana, has prevailed without N-fixation, implying that the local soils have ample N relative to other nutrients, and
• similar landforms nearby, e.g. along the Sabie River, have Thorn Thicket dominated by N-fixers.

The three spp. of mimosoid legumes, which can be loosely referred to as acacias, are important N-fixers in this vegetation. They achieve, in these plots, biomass approaching that of S. africana. (They may even exceed the biomass of S. africana in particular 10 m X 10 m areas, in the case of the large, emergent trees of S. nigrescens.)

So, although the dominant S. africana is not an N-fixer, the leguminous N-fixers coexisting with this euphorbiaceous plant are an important part of the community.

The two spp. of faboid legumes, although also N-fixers, contribute negligibly. This is because one (P. violacea) is scarce and suppressed, and the other (O. trichocarpum) occurs only as small individuals.
 
The second category is fleshy fruit-producing shrubs and trees.

These tend to be dispersed by birds or, in the case of D. mespiliformis, carnivores such as the civet, Civettictis civetta (the faeces of which, by the way, I never observed in this whole complex of vegetation associated with sodic patches).
 
In this endozoochorous (but never N-fixing) category we have E. transvaalensis, Pappea capensis, Pyrostria hystrix (and Maerua parvifolia), Euclea spp., Diospyros mespiliformis, Pavetta sp., Grewia spp., Cordia ovalis/monoica, Commiphora africana, Searsia guenzii, and Gymnosporia maranguense.

Most of these spp. are dispersed by fruit-eating birds.

Gardenia volkensii and Schotia brachypetala are also probably endozoochorous, but more specialised for larger dispersers.

However, none of the species in this category are much dispersed by ungulates in the way N-fixing Dichrostachys is.

The above may seem like an impressive list. However, my interpretation is that most of these ca 15 spp. are represented by what are no more than ‘hopeful hangers-in'. I refer to individuals that have germinated from seeds dropped by birds, and are able to persist as saplings/juveniles/suppressees for long periods, in the hope that the monopoly of S. africana may, one day, somehow be relaxed.

In fact, few of these individuals are destined ever to grow to maturity. For example, D. mespiliformis and S. brachypetala belong not here, but >0.5 km away, on the banks of the N'waswitshaka River - where they dominate the riparian strip of forest as stately trees.
 
The third category is the typical, wind-dispersed, non N-fixing trees of granitic bushveld in Kruger National Park.

In this category we have only Combretum hereroense, and even this species was so scarce in these plots that its presence is negligible.

Bearing in mind also the fact that Combretum imberbe does not occur in any of these 8 plots as an emergent tree in the way S. nigrescens does, what we find is that S. africana has almost completely replaced Combretaceae in the densest vegetation associated with the sodic patches.
 
To summarise:

Spirostachys africana has dominated these patches of vegetation, without excluding some important N-fixing acacias in the way it has excluded Combretaceae.

There are more than 15 spp. of associated, endozoochorous trees and shrubs, that can be found by poking around in the dense stands of clonal S. africana. However, these amount to little in functional terms. They merely adding floristic spice, rather than mattering much in the ecosystem.

For example, although E. transvaalense is capable of growing into a substantial tree, I doubt that, in the entire complex of vegetation associated with the sodic patches along the N’waswitshaka River, there is a single mature individual of this species to be found.

to be continued in https://www.inaturalist.org/journal/milewski/68846-woody-plants-surrounding-treeless-lawns-on-sodic-substrates-in-southern-kruger-national-park-part-2-adaptations-to-intense-herbivory-at-the-edges-of-the-grazing-lawns#...

Posted on July 7, 2022 06:35 AM by milewski