Invasive alien species are considered a major pressure on the current state of biodiversity globally (Butchart et al. 2010). Invasive alien plants (IAPs), in particular, have spread rapidly and extensively in many regions of the world, impacting resident species diversity, ecosystem processes and people’s livelihoods (Levine et al. 2003; Pyšek et al. 2012; Schirmel et al. 2016; Vilà et al. 2011). South Africa is no exception and nearly two million hectares of land have been invaded by alien plants (Van Wilgen et al. 2012), with well-known impacts on hydrology, nutrient cycling and fire regimes (Kraaij, Cowling & Van Wilgen 2011; Le Maitre, Gush & Dzikiti 2015; Richardson & Van Wilgen 2004). This estimate of alien plant coverage includes 27 species, without incorporating arid and transformed land except for Prosopis trees in the arid northwest of the country (Kotzé et al. 2010; Van den Berg, Kotze & Beukes 2013). Alien Acacia species cover most of the dense areas of invasion, followed by Eucalyptus and Pinus trees, Opuntia cacti and Chromolaena odorata shrubs. These invasions extend across the country, with higher concentrations in the southwestern, southern and particularly eastern coastal belts and the adjacent interior (Henderson 2007; Kotzé et al. 2010; Van Wilgen et al. 2012). Overall, there is reasonable knowledge of alien plant occurrence in South Africa, especially at a coarse spatial resolution. Whereas their effects on native plant diversity have been fairly well assessed (e.g. Gaertner et al. 2009; Richardson & Van Wilgen 2004), fewer studies have focused on the impacts of alien plants on native animal communities (Richardson et al. 2011, but see Breytenbach 1986).

Species population and community metrics such as abundance, richness and composition can provide useful baseline data as indicators of animal diversity change between invaded and uninvaded areas. The direction and magnitude of effects of alien plant invasions on animal communities can, however, depend on a variety of factors, including the scale of the plant invasion (extent and density), the stage of invasion, and the region and taxonomic group affected (Kumschick et al. 2015; Ricciardi et al. 2013). These context dependencies make extrapolations of the effects of IAPs on resident biota from region to region a substantial challenge, especially for the development of generalised management frameworks across diverse habitats. The invasion of alien plants into natural or previously uninvaded habitats involves a number of significant changes to the habitat, often negatively affecting resident fauna and sometimes in counterintuitive or non-obvious ways (Figure 1). Alien plants may directly modify the structure and complexity of the physical environment and, thus, restrict the opportunities for the animal to thermoregulate or hydroregulate within its microenvironment or impose barriers to essential functions such as moving, creating nests or finding refuges. Alien plants can also directly or indirectly affect food resources for animal communities (e.g. Groot, Kleijn & Jogan 2007). For example, a change in plant composition will affect herbivores directly by reducing the amount or quality of plant hosts whereas a change in habitat structure, microenvironment and litter or soil properties can indirectly affect prey availability or predator abundance and alter trophic interactions (Figure 1; e.g. Pearson 2009).

Whereas reports of negative impacts of invasive plant species are pervasive in the literature, positive effects have also been reported, for example, via increases in suitable habitat or net resources to recipient fauna (e.g. Schlaepfer, Sax & Olden 2010). Whether the latter represent rare case studies or whether any general patterns can be drawn from these is presently unclear. Thus, knowledge of the underlying mechanisms should improve the understanding of the consequences of alien plant invasions on native diversity. More importantly, knowledge of proximate and ultimate causes of species declines may enhance our ability to predict responses of animal communities in newly invaded areas of South Africa with similar characteristics to those studied previously, or facing concomitant pressures such as global climate change or habitat transformation (Ricciardi et al. 2013). Together, these are essential elements of the scientific framework that will allow invasion science to robustly predict the impacts of new and developing invasions, and not simply be viewed as a series of unique invasion case studies.

In this study, we aim to synthesise studies that have examined the effects of IAPs on animal diversity in South Africa. We concentrate this review on ectotherms (reptiles, amphibians and invertebrates) for several reasons. Firstly, their energy budgets are more directly influenced by the environment compared with endotherms (Gates 1980). Consequently, environmental factors likely play a major role in determining a suite of physiological and behavioural attributes and, at least partly, influence life history and timing of key phenological events (e.g. mating and reproduction) in the group. Secondly, they typically have smaller dispersal abilities than mammals and birds (Endler 1977), likely reducing their capacity to move away from disturbance or suboptimal conditions. Finally, they make a large contribution to overall animal diversity that is mostly explained by high insect diversity and abundance (Wilson & Peter 1988), playing a central role in food webs and ecosystem function. To present a comprehensive view of the status of knowledge of impacts of IAPs on resident animal species in South Africa and identify information gaps, we ask three key questions: (1) How many studies have addressed the effects of IAPs on terrestrial ectotherm diversity and what general patterns can we draw from these?, (2) What is the taxonomic and spatial coverage of IAPs and affected native species studied so far? and (3) How much is known about the mechanisms underlying these impacts? These questions are central to assess the status of knowledge of alien plant impacts on animal species diversity and inform invasive plant monitoring and control programmes in South Africa (Wilson et al. 2017).

We searched the ISI Web of Science for relevant studies comparing abundance, richness or composition of terrestrial ectotherms between invaded and uninvaded sites in South Africa. Our search combined terms for (1) invasive plants; (2) native reptiles, amphibians and terrestrial invertebrates; and (3) effects on species abundance, richness or composition (shown in detail in Appendix 1). We included studies comparing sites with native vegetation or cleared of alien vegetation to sites invaded by alien plants or with plantations of alien species. A second search targeted studies addressing only the mechanisms underlying the potential effects on ectotherm species, without necessarily quantifying changes in species diversity. This second search thus replaced the search terms for effects with terms for mechanisms such as altered thermoregulatory behaviour, prey availability or reproductive output (Appendix 1). We retrieved articles, reviews or book chapters for all years available on the Web of Science Core Collection on 10 June 2016. The articles considered relevant for our review were then screened for additional references.

We gathered information from the studies on the location of the field sites and respective biomes (Mucina et al. 2014), and on the native animal and IAP species included. We classified the studies according to the phyla of native animals potentially affected by plant invasions (Arthropoda, Annelida, Chordata and Onychophora). For the IAP species included, we used three classifications. Firstly, we assigned IAP species to six major growth forms: tree, shrub or bush, vine, forb, grass and aquatic plant. Secondly, we used the categories of the Alien and Invasive Species (A&IS) Lists that were published under the National Environmental Management: Biodiversity Act (NEM:BA) in 2014. The NEM:BA A&IS categories include eradication targets (category 1a), widespread invasive species where a national species management programme is required (category 1b) and invasive species that can be kept under managed circumstances (categories 2 and 3). Thirdly, to assess the extent to which the existing studies focus on the invasive plants with the largest potential impacts on ectotherms, we also considered a published classification of the most prominent invasive plant species in South Africa, according to their estimated impacts on native biodiversity (Van Wilgen et al. 2008). For all classifications above, we assigned studies to multiple categories when they presented individual comparisons for more than one biome, native phylum or IAP category.

To assess the impacts of alien plants on native animal diversity, we focused on comparisons of species abundance or richness between sites with native and alien vegetation and classified the effects as positive (i.e. an increase in abundance or richness in sites with alien vegetation), negative (decrease) or neutral effect. Native vegetation sites included sites cleared of alien plants in two studies where authors indicated that sufficient time had elapsed for recovery of native vegetation. We incorporated comparisons based on original data, accumulation and rarefaction curves, and richness estimators (e.g. Chao1 and Chao2) but did not include diversity indices (e.g. Shannon’s or Simpson’s index) as their use was very variable across studies. For composition, comparisons were classified as alien vegetation resulting in a change or not in species composition. These comparisons were typically the result of ordination techniques based on similarity measures or cluster analyses. When studies presented the effects of alien plants on fauna for several functional or taxonomic groups (e.g. herbivores vs. predators and beetles vs. spiders) or for different habitat types (e.g. invaded by Eucalyptus vs. Pinus), we considered each comparison separately in the synthesis unless there were no statistical tests associated with these. Therefore, the number of comparisons was typically higher than the number of studies assessed. If available, we extracted information on the growth form, stand age and spatial coverage of the IAP and on mechanisms driving the impacts such as changes in habitat structure, thermal opportunities, food resources, predators and refuge or nest site availability.

Our first search for studies addressing the effects of invasive plants on ectotherm diversity in South Africa yielded 358 studies. Of these, only 42 were relevant for this review (Appendix 2). Our second search for articles studying the mechanisms underlying the effects of alien plants on ectotherms in South Africa yielded 702 papers. Among these, we only retained six relevant studies (Appendix 2), partly because a large portion of the articles found investigated the viability of biological agents for invasive plant control which was outside the focus of this study.

The 42 papers reviewed were published between 1985 and 2016, with a slight increase in the annual number of publications since 2000 (Figure 2a). The vast majority of studies focused on arthropods (Figure 2b), particularly in the Insecta (mostly Coleoptera, Hymenoptera, Diptera, Lepidoptera and Odonata) and Arachnida (mostly Araneae) classes. A single study included Onychophora and two studies focused on earthworms. Only three studies addressed vertebrate ectotherms, covering amphibians (Order Anura), lizards and snakes (Order Squamata).

Most studies compared sites along the coastal belt and adjacent interior in the Western Cape and KwaZulu-Natal provinces (Figure 2c). The majority of studies took place in the Fynbos and Grassland biomes, whereas a few studies covered the Savanna, Indian Ocean Coastal Belt and Albany Thicket biomes (Figure 2c). Native arthropods remained the focal native class across biomes (Figure 2c). Trees, particularly those belonging to the genera Acacia, Eucalyptus and Pinus, were the most common growth form of invasive plants (Figure 3a), followed by shrubs such as Chromolaena odorata, Hakea species, Lantana camara and Solanum mauritianum. These alien trees and shrubs are listed as major invaders in the country (Henderson 2007; Le Maitre, Versfeld, & Chapman 2000; Van Wilgen et al. 2012). Many of them are species that need to be controlled according to the NEM:BA A&IS regulations (category 1b; Figure 3a) and are estimated to have high impacts on native biodiversity (Van Wilgen et al. 2008; Figure 3b).

Alien plants had a larger decreasing effect on native species abundance compared to species richness (Figure 4a and Table 1). Most studies found that the effects of alien plants were either neutral or decreasing for native animal species richness, but increasing effects of alien plants were rare for both species richness and abundance. Alien plants had a substantial impact on species composition (Figure 4b and Table 1). Among the 15 IAPs listed on NEM:BA, only 3 species (Arundo donax, Chromolaena odorata and Passiflora edulis) showed no negative effects on the arthropod species studied. However, these results refer to a single study conducted for each alien plant, highlighting the data deficiency for these species (Table 1). Acacia mearnsii was the IAP, most commonly studied, with negative or neutral effects on arthropod species found in four studies (Table 1). Although most studies incorporated standard metrics of species abundance and richness, accumulation and rarefaction curves were presented (or mentioned) in only 13 and 7 studies, respectively, and 10 studies provided comparisons of metrics between invaded and uninvaded sites without presenting accompanying statistical analyses.

Seventy percent of the studies investigating differences in species diversity between invaded and uninvaded habitats referred to potential mechanisms underlying the patterns found. These mechanisms included changes in habitat structure (n = 15 of 29 studies), microclimates (n = 13 of 29 studies) and food resources (including host specificity for herbivores, n = 14 of 29 studies), and the degree of species’ ecological breadth (e.g. generalists vs. specialists, n = 3 of 29 studies). In the 6 publications that solely referred to mechanisms (Appendix 2), the authors highlighted habitat structure, microclimate and food resources as mechanisms affecting functional diversity (according to McGill et al. 2006) and animal behaviours such as the ability to create nests, flight dynamics and flower visitation rates. Of a total of 35 studies that addressed mechanisms, 21 studies used an observational (correlative) approach to assess the link between impact and mechanism, 4 studies followed a manipulative experimental approach (including a study that incorporated the clearing history of the invasive plant) and 10 studies speculated on potential mechanisms.

IAPs have been recognised as a threat to South Africa’s native biodiversity for more than two decades, with efforts to manage invasions underway through the Working for Water Programme since 1995 (Van Wilgen et al. 2016). Our review underscores the importance of controlling IAPs to reduce their impacts on terrestrial ectotherm diversity in the country. Although not always statistically significant, reductions in abundance and richness of native ectothermic species were found in 56% and 41% of the comparisons presented in the studies reviewed, respectively, and changes in species composition were reported in almost 75% of the comparisons. These findings echo those of global reviews and meta-analyses, which have reported significant decreases in native animal species abundance, diversity and fitness as impacts of plant invasions (Pyšek et al. 2012; Schirmel et al. 2016; Van Hengstum et al. 2014; Vilà et al. 2011). In the following three sections, we discuss key directions for future research, the context-dependent nature of the problem and issues relating to methodologies used to assess impacts.

The current state of knowledge of the impacts of IAPs on resident ectothermic animal species in South Africa relies heavily on a few key studies, with distinct biases in geographic locations and taxonomic groups. In cases where detailed information is available, it is nevertheless clear that there are pronounced negative impacts of IAPs on terrestrial animal (ectotherm) species diversity. The mechanisms underlying these impacts are unclear, but here we highlight a few key abiotic and biotic processes that could be examined in future, especially if microenvironments determine key behaviours and life-cycle timing that lead to changes in population abundance. Such an integrated approach to the question of IAPs and their impact on native animal species diversity would be of direct value to monitoring and conservation efforts, such as alien plant removal and control programmes. At present, it is wholly unclear whether the removal of IAPs will be sufficient to allow recovery of native ectotherm biodiversity.