Climatic niche shift and possible future spread of the invasive South African Orchid Disa bracteata in Australia and adjacent areas

Introduction

The study of biological invasions has been called “one of the hottest current topics in ecology” (Sol, 2001), mostly because together with habitat destruction and climate change the spread of non-native organisms is considered to be a major threat to biodiversity. In Australasia, invasive (= not native) species are a major problem; for example, the number of species of plants reported as introduced, that have been released and may or may not have become naturalized in Australia exceeds 28,000 (Randall, 2007). A subset of them have naturalized and are a threat to the rich endemic flora of that continent (Coutts-Smith & Downey, 2006; Randall, 2007; Duursma et al., 2013) as the populations are self-sustaining and spreading without human assistance. In addition, the eradication of invasive weeds is costly (Sinden et al., 2004). Governmental agencies and private landowners invest large amounts of money in controlling the spread of weeds using various methods to assess risk of their further spread. One of the emerging techniques is to evaluate the potential future ranges of invasive species using climatic niche modeling (Peterson, 2003).

In this research we implemented modeling approach to evaluate the possible further spread of invasive orchid species, Disa bracteata Sw., which was first reported in Australia relatively recently, in 1944. This plant is listed in the Global Compendium of Weeds (GCW; http://www.hear.org/gcw) and is the only weedy representative of the mainly sub-Saharan genus Disa P.J. Disa bracteata is classified in the GCW as an environmental weed (species that invade native ecosystems; (Blood, 2001) or a naturalized species (self-sustaining and spreading populations but not necessarily affecting the environment; Barker et al., 2005; Hussey et al., 1997). D. bracteata is a South African endemic plant found in both the Eastern and Western Cape (Foden & Potter, 2005), where it is widespread and common, especially in areas subject to mild disturbance. In undisturbed vegetation it is somewhat less frequent. D. bracteata is included on the Red List of South African plants as a taxon of Least Concern (Raimondo et al., 2009; Foden & Potter, 2005). In the mid-20th-century it was brought to Australia where it became naturalized (Groves et al., 2003). This orchid was first formally recorded near Bacchus Marsh, west of Melbourne, in 1944 (Wileman, 2015; Land Management Team, 2015) and since 1945 additional reports came from areas of the Great Southern Region in Western Australia (vicinity of Albany). Later it was also recorded in South Australia (in 1988) and Victoria (in 1994). Recently it was found in Tasmania (Viridans Biological Databases). In Australia D. bracteata was probably accidentally introduced and it is now growing along roadsides. Invasive populations are large with up to almost 80 mature individuals in one square meter (Tucker, 2006; Trees For Life).

The aim of this research was to evaluate the similarities in the bioclimatic niches occupied by invasive and native populations of D. bracteata and to estimate the potential further spread of this species in Australia and adjacent areas. The comparison of the bioclimatic conditions experienced by African and non-native plants was conducted to explain the nature of this invasion by exploring the possible changes in the bioclimatic preferences of D. bracteata in early and present stage of its spread in Australia. Currently it is unclear whether the studied species inhabits similar climatic niches in Australia and Africa or whether it was able to colonise new niches in Australia. The negative effect of the introduction of exotic orchid on the native flora was reported only once so far (Recart, Ackerman & Cuevas, 2013), but the actual impact of invasive Orchidaceae on local plant communities remains poorly recognized. We do believe that results of our analyses of possible future spread of D. bracteata will be valuable information that can be used in planning conservation actions.

Material and Methods

Results

Potential distribution of suitable niches under current climatic conditions

The model of the current distribution of suitable bioclimatic niches for D. bracteata was calculated based on all uncorrelated variables and a model in which only climatic factors were considered exclusively were visually congruent (Figs. 1A, 1B, 2A and 2B). The main difference concerns the transitional zone between the western and eastern Cape in Africa and the Nullarbor Plain in Australia. The first region was not indicated as suitable in the model based on climatic factors, whereas the second one was shown as suitable for this orchid in this analysis. Otherwise the difference may be seen in the scale of suitable habitat—while the general trend is very similar, both models differ slightly in the extent of the predicted suitable niche. Currently, the species is found in most of the suitable habitats in both Africa and Australia. Yet there are some areas in Australia that are not colonized—the Eyre Peninsula and some smaller areas detached from the main distribution. A vast area of Tasmania was predicted to be suitable but so far the occurrence of D. bracteata has been reported only from the northern part of the island. Also, there were no records of this orchid from New Zealand (especially the North and South Islands) although the models indicate the existence of suitable habitat in this area.

Niche overlap and identity

The overlap between the studied environmental requirements of invasive and the natural populations is moderate or even low. Statistics calculated according to Broennimann et al. (2012) were D = 0.35 and I = 0.58. These values were slightly higher using the method developed by Warren, Glor & Turelli (2008): D = 0.44, and I = 0.73. Also, niche identity tests reveal that the bioclimatic niches occupied by invasive and natural populations were different (p < 0.01). Yet visualization using PCA (Fig. 3) indicates a large overlap between the niches with only a small proportion of which is different.

Analyzed environmental conditions recorded at the sites occupied by this species are illustrated in Fig. 4. To measure the dissimilarity between invasive and native populations a bias metric was calculated. It indicates that all invasive populations occupy relatively different habitats compared to populations in Africa (Chi-squared = 46.031, p = 4.9 × 10⁻⁴). The same result is obtained when the invasive records are considered as one dataset or when the records are divided into those for Eastern and Western Australia (Chi-squared = 73.807, p = 4.9 × 10⁻⁴). Nevertheless, the suitable niches of both groups of Australian populations are also different (Table 1). There are, however, two similarities in the site characteristics of these two Australian regions: there is a greater sand content and lower precipitation in the warmest quarter in the areas occupied relative to that in areas occupied by natural populations. Others have either a medium difference (e.g., soil pH) or a significant dissimilarity (e.g., temperature annual range, mean temperature in wettest quarter, precipitation seasonality, precipitation in coldest quarter and soil organic carbon content; Table 1).

Discussion

Orchids are usually not regarded as weeds although some, e.g., Epipactis atrorubens, E. helleborine and Dactylorhiza majalis, colonize secondary habitats in temperate Europe (Adamowski, 2006; Rewicz, Kołodziejek & Jakubska-Busse, 2015). Even fewer species are reported as invasive (e.g., Ackerman, 2007) and so far only one, Spathoglottis plicata, has been shown to negatively affect native plants (Recart, Ackerman & Cuevas, 2013).

The invasive success of D. bracteata has not been thoroughly investigated and the mechanisms of this phenomenon remain unclear. The only predictive models are in Weed Futures (http://www.weedfutures.net; Duursma et al., 2013). Most ground orchids usually do not have great potential for spreading because of their very specific ecological requirements. Orchidaceae often have specific insect pollinators and specific mycorrhizal associations that need to be present in the soil to enable seed germination (e.g., Batty et al., 2002; Cozzolino & Widmer, 2005; McCormick & Jacquemyn, 2014). Many species of orchids will only germinate with the aid of one or a few species of fungus, so their distribution, and hence ecological success, is heavily dependent on suitable conditions for the fungus. Disa bracteata appears to be able to form an association with a large number of fungal partners, especially those that can survive in disturbed soils (Bonnardeaux et al., 2007), thus it is much less limited in terms of the places and conditions in which it may become established. For this reason, we did not incorporate the distribution of mycorrhizal fungi in our analyses. It is noteworthy that the incorporation of a fungal factor in any analysis of orchid distribution is extremely difficult. Most orchid mycorrhizal fungi belong to the genus Rhizoctonia, a diverse polyphyletic group that is difficult to classify and molecular methods have become the standard means of assigning these orchid fungi to groups within the Rhizoctonia alliance (Bonnardeaux et al., 2007). Because of the lack of data on the distribution of specific fungi it is not possible to use such data in ecological niche modeling.

As seed production in Disa bracteata is pollinator-independent there was no need to incorporate the potential distribution of any pollinator in order to get a more realistic potential distribution of this orchid. Disa bracteata appears to be self-pollinating as a result of the breakup of pollinia in the anther (Kurzweil & Johnson, 1993). Generally, this is not beneficial for genetic variability, however it does enable it to produce a large number of seeds. High propagule pressure greatly enhances the chances of the establishment of invasive species (Colautti, Grigorovich & MacIsaac, 2006). In fact, genetic variability seems to be less important as many successful invaders reproduce vegetatively. This is frequently the only mode of reproduction in some invasive species. This is often influenced by environmental conditions that are not suitable for the full development of a plant e.g., maturation of seeds. Ideally, even in such cases an invasive species may couple vegetative propagation with occasional sexual reproduction in order to respond to a suite of selective pressures and propagate efficiently (Atwater et al., 2017).

The environmental similarity between Australia and South Africa enabled numerous African plants to naturalize in Australia. 15% of naturalized flora of South Australia consists of species native to South Africa (Kloot, 1986; Scott & Panetta, 1993). Slightly higher contribution of African plants was observed in Western Australia (17%; Scott & Delfosse, 1992; Scott & Panetta, 1993). It is worth to notice that the two species considered as major environmental weeds in Australia and New Zealand, Chrysanthemoides monilifera (Asteraceae) and Asparagus asparagoides (Asparagaceae), are of African origin (Thorp & Lynch, 2000).

Invasive species often experience release from biotic interactions and dispersal barriers in their non-native ranges (Torchin et al., 2003; Colautti et al., 2004; Jiménez-Valverde & Peterson, 2011). While the visualization using PCA indicate a large overlap between the niches of African and invasive populations of D. bracteata they are not the same and that small difference is important as it influences the results of the niche identity test.

Another interesting result is that the niche of invasive populations has changed over time as the colonization process has progressed. At the time of the first introduction, which was around 1944 (date of the first recorded specimen), only a few localities were known and the conditions there were similar to those in its natural range. However, over time more populations were established, which eventually gave rise to the colonization of the eastern part of Australia. The first georeferenced specimen collected in 1989 was found on the southern Adelaide Plains. Dispersal to the western part of the continent involved colonizing novel habitats or those that are not available to Disa in its native range (File S3). This shift is congruent with the Köppen–Geiger climate classification system (Peel, Finlayson & McMahon, 2007), which indicates that the difference in niches may be influenced by the climate availability or different preferences within both ranges. In Africa, six climatic types were occupied while in Australia two of them were not populated and the majority of established populations occurred within the Csb climate (Coastal Mediterranean). The pattern of shifting niche of invasive species was detected in some previous studies in cases when available evidence suggests use of novel environments by alien species in the invaded range (Medley, 2010; Petersen, 2012; Di Febbraro et al., 2013) when those conditions could be unavailable or inaccessible in the native range (Broennimann & Guisan, 2008; Godsoe, 2010; Guisan et al., 2014; Qiao, Escobar & Peterson, 2017). Additionally, there was a difference between Eastern and Western Australia: in the former second most frequent climate type is the Cfb (Marine With Mild Winter), while in the latter this position belongs to the Csa climate (Interior Mediterranean; Fig. 5). Thus it is not an entirely new climate but rather a shift of the climatic preferences that may be attributed to the establishment of invasive populations in areas possessing different composition of available climates. However, it may be related to novel preferences as well since all climates that are present in Africa are present in Australia. Interestingly, areas of some types of climate are less frequent in Africa and apparently are sparsely populated by D. bracteata but mainly by other Disa s.l. species. This may indicate that in its native distribution there are factors that prevent D. bracteata occupying this niche, such as biotic interactions (possibly with other closely related species) while in Australia this constraint is absent and D. bracteata is able to colonize these new sites. Studies on niche shifting species indicate this occurs in various areas (Broennimann et al., 2007; Elith, Kearney & Phillips, 2010) but it is relevant to fewer than 15% of plant invaders (Petitpierre et al., 2012). There are significant differences between the climatic niches occupied by invasive and native populations of another invasive species of orchid, E. helleborine (Kolanowska, 2013), but in this case, no shift within the Köppen–Geiger climate was found. A study of the niche shifts of plants introduced into Australia found that none of 26 species included in this study changed their Köppen–Geiger climatic niche (Gallagher et al., 2010).

In the case of D. bracteata a significant difference in the proportion among climates in areas occupied in Africa and Australia was recorded (Fisher’s Exact Test p = 1.5 × 10⁻⁵). The same result is obtained when Africa is compared with two regions of Australia considered separately (Africa vs. Eastern Australia: Fisher’s Exact Test p = 1.4 × 10⁻⁴, Africa vs. Western Australia: Fisher’s Exact Test p = 1.3 × 10⁻³). This suggests that irrespective of invaded region, D. bracteata occurs in Australia in habitats characterized by different climatic conditions (in terms of climatic zones) than these recorded in Africa. Even though both contain the same climatic types, they differ in their abundance and relative areas. This may be influenced by the total available land area which is much smaller in Southern Africa that is stretched along meridians while Australia is rather stretched along parallels.

Disa bracteata may be the first record of such a shift within Australia. When estimating niche conservatism and changes in future range, one needs to bear in mind that correct estimation of the latter is tricky because acquiring a new niche is an active process possibly linked to novel adaptations that may not be known at the time of a study (Thuiller et al., 2008; Elith, Kearney & Phillips, 2010). Alexander & Edwards (2010) suggested that the probability of a niche shift in invasive species depends primarily upon the ecological and genetic processes limiting the species in its native range. Unfortunately, during our studies we did not have access to a sufficient amount of molecular data to explore genetic differences between African and invasive populations of D. bracteata.

Conclusions

A South African D. bracteata has become invasive in Australia and it is already present on a large part of the continent. Created models suggest that area with the suitable niche for this species is larger than that currently occupied by the studied orchid thus; the spread of this species will continue. How this expansion will proceed depends on future changes in the factors influencing its distribution and primarily on the magnitude of the climate modification. As demonstrated here, it is very likely that a niche shift has occurred in this case so the further spread of studied orchid should be monitored. Altogether, the results of this study indicate the need of further research on the spread of D. bracteata, especially analyses of genetic differences between native and invasive populations.

Supplemental Information