The conservation value of freshwater habitats for frog communities of lowland fynbos

Introduction

Amphibian conservation has centred around three major themes: habitat change, disease and invasive species (Grant et al., 2019). Freshwater habitats are in particular peril, with declines that are far greater than terrestrial ecosystems (Dudgeon et al., 2006). However, the construction of artificial impoundments is increasing rapidly (Downing et al., 2006), and changing the nature of landscapes especially in arid ecosystems, where these permanent impoundments facilitate invasions of freshwater species (Johnson, Olden & Vander Zanden, 2008). Habitat change and invasive species impact more species globally and are the proximate causes of conservation concern for the majority of amphibian species (Harfoot et al., 2021). Despite a general acknowledgement of these mechanisms, conservation evidence for impacts and their commensurate measures for the conservation of amphibians continues to be low (Meredith, Van Buren & Antwis, 2016).

The impacts of invasive species on amphibians have been assessed qualitatively (Bucciarelli et al., 2014; Falaschi et al., 2020), and quantitatively (Nunes et al., 2019). Invasive freshwater fish are ranked highly in reviews of invasive species as causing severe impacts on many amphibian communities (Hecnar & M’Closkey, 1997; Ficetola & De Bernardi, 2004; Hartel et al., 2007; Holbrook & Dorn, 2016). Excluding invasive fish from sites with threatened frog species has resulted in recovery of anuran populations in Spain and Portugal (Rana iberica Bosch et al., 2019) and California (Rana mucosa Knapp, Boiano & Vredenburg, 2007), leading those workers to identify the proximate role of invasive fish as a threat to amphibian populations. But the impacts of invasive fish are poorly described in the southern hemisphere (except Australia), especially with respect to amphibian communities. However, many amphibian communities are driven by natural environmental factors as well as anthropogenically driven creation and modifications of freshwater habitats (Ficetola & De Bernardi, 2004; Hartel et al., 2007; Kruger, Hamer & Du Preez, 2015).

The low-lying fynbos of South Africa’s Cape region is a biodiversity hotspot (Myers et al., 2000), composed of evergreen, Mediterranean scrub vegetation (Mucina & Rutherford, 2006). The fynbos also holds an important community of amphibians that have high conservation concern (Measey, 2011; Schreiner, Rödder & Measey, 2013; Mokhatla, Rödder & Measey, 2015). Much of the habitat where amphibians and other flora and fauna were once abundant has been transformed for agriculture and more recently for housing (Measey & Tolley, 2011; Rebelo et al., 2011; Measey et al., 2014). Where land has been transformed, temporary wetlands have been infilled and permanent impoundments (dams) or ponds added to the landscape. The addition of permanent water and the introduction of alien fish has been ongoing for ~200 years (Ellender & Weyl, 2014). Angling is a popular pastime in the region, and anglers introduce fish to new impoundments and natural waterbodies (Ellender et al., 2014).

Southern Africa has no salamanders or caecilians, but several major radiations of anurans, many of which specialise in lowland temporary aquatic habitats (Poynton, 1964). The extreme southwestern corner of the continent has a mediterranean climate with winter rains and dry hot summers (Wilson et al., 2020). Several species that rely on temporary water have become threatened, while those that thrive in permanent water have become abundant and ubiquitous even in arid areas. Examples of IUCN threatened species in the area include the Western Leopard Toad Sclerophrys pantherina (EN), the Cape Platanna Xenopus gilli (EN), the Microfrog Microbatracella capensis (CR), and the Flat Caco Cacosternum platys (NT).

In this article, we aim to determine whether habitat characteristics (especially temporary vs permanent water) correlate with Cape lowland amphibian communities. In particular, we were interested to find out whether anthropogenically constructed impoundments are useful sites for threatened amphibian communities, in the presence or absence of invasive fish.

Methods and Materials

Site selection

Using Google Earth imagery from 2017, we classified every waterbody visible within our study area (two catchments) into our nominate freshwater body types being natural: vleis (natural temporary shallow water bodies), natural pools and river edges in the fynbos, and anthropogenically created: small dams (artificial impoundments <2,000 m² including ponds), and large dams (artificial impoundments >2,000 m²). In lowland fynbos, slow moving rivers are effectively temporary water bodies and have the same amphibian species assemblages (Channing, 2019). This gave us a candidate list of 196 sites (see Table S1) all chosen from within the fynbos biome (see Mucina & Rutherford, 2006).

We made our initial stratified sampling selection from within these 196 sites to represent balanced numbers of freshwater body types, equally represented across space, and these were further refined once we requested permission to access sites from landowners. The final 50 sites selected fell within two separate catchments, varied in their spatial proximity and different water body types (see Fig. 1; Supplemental Information; Table S1).

Results

We identified 11 different anuran species (Table 1) across the 50 sites sampled. All sites sampled were found to have at least one species of anuran, with a maximum of nine and a minimum of one. No species were found exclusively at permanent or temporary water sites (Table S2). In addition, we had evidence of three invasive fish species: Large-Mouth Bass Micropterus salmoides, Small-Mouth Bass M. dolomieu and Mozambique Tilapia Oreochromis mossambicus from 14 sites.

The partial RDA model explains 11.06% of the variation in frog community composition across all sites, and is statistically significant (F = 1.583; P = 0.002). From the partial RDA, we found that the best measured environmental determinants of the amphibian community were the presence or absence of invasive fish (F = 1.950; P = 0.035; Fig. 2; Table 2), area (F = 1.828; P = 0.050) and the type of wetland habitat sampled (F = 2.046; P < 0.001). In Fig. 2, only the first axis was found to be significant (F = 6.008; P < 0.001).

Our exploratory analysis on how invasive fish impact amphibian assemblages in permanent water (using 36 of the 50 sites) produced a 2-dimensional NMDS fit with stress of 0.192, which showed significant differences using adonis2 (R² = 0.089; P = 0.025; Fig. 3). The two species most tolerant of the presence of fish were the toads: Sclerophrys capensis and S. pantherina, while the species most intolerant of fish was X. laevis (Fig. 3; Table S3).

Discussion

Our study stresses the importance of freshwater types which determine the type of amphibian community in the southwestern Cape, with an important division between anthropogenically created water bodies (irrespective of size), and those that occur naturally in the fynbos. Permanent water bodies generally hold widespread species, while temporary sites typically hold fynbos endemic species. In addition, we show that the presence of invasive fish in permanent water bodies also impacts amphibian assemblages. Our results indicate that building permanent water bodies, whether they be large impoundments for agricultural water supply or small garden ponds, will favour different amphibian communities from those present in sites with temporary water. Many urban homeowners create permanent small ponds in their gardens with conservation goals. However, our results indicate that trends for increasing biodiversity in urban areas by creating ponds championed in the northern hemisphere (Hassall, 2014; Hill, Lawson & Tuckett, 2017; Hill et al., 2018) are inappropriate in the fynbos where large impoundments already provide for assemblages that require permanent water. Permanent impoundments also promote invasions of both fish and amphibians (Davies et al., 2013; Ellender et al., 2014). Currently, amphibians that rely on temporary water in lowland fynbos are poorly served by anthropogenically created wetlands, but could be better conserved by the promoting construction of temporary water bodies instead of ponds.

While no anuran species was exclusive to permanent water, these types of water bodies were commonly associated with more widespread species: Amietia fuscigula, Hyperolius horstocki, Xenopus laevis and Tomopterna delalandii. These species are not endemic to the area, while those associated with temporary water have much smaller distributions (~20,000 km²). Toads (Sclerophrys capensis and S. pantherina) were most tolerant of the presence of invasive fish, presumably because their eggs and larvae are toxic and adults have prominent parotid glands (Hecnar & M’Closkey, 1997; Crossland & Alford, 1998; Caller & Brown, 2013). The species most intolerant to the presence of fish was X. laevis, which may be because they are principally aquatic and encounter fish more often than other frogs. The area we sampled did not include some threatened species present in the lowland fynbos, for example X. gilli (EN) and Microbatrachella capensis (CR), but these are most commonly associated with temporary water (J. Measey, 2017, Personal Observation).

Historically, these areas of lowland fynbos would have had very few permanent water bodies. The sediment is typically sand or silty soils over young Quaternary sediments, largely derived from weathering Table Mountain sandstones and Cape Supergroup shales (Cawthra et al., 2020). Rivers that flow year round may well have been augmented by the movements of large mammals to increase the permanent water features associated with them (Venter et al., 2020). Away from rivers, most water bodies would have formed through rainfall, or be fed by underground seepages, during the wet winter period, and completely dry out during summer. Many of the lowland fynbos areas have been developed and habitat loss continues to the present day (Skowno, Jewitt & Slingsby, 2021). The Cape Lowland Freshwater Wetlands are considered to be Critically Endangered in the National Ecosystem Status for South Africa (Dayaram et al., 2021).

Because we were not able to sample each site systematically for fishes, there is a chance that we have false negative data among some of the permanent water sites. For example, we found no mosquitofish, Gambusia affinis, in the water bodies we sampled, but they are known to be invasive in many drainages of South Africa (Weyl et al., 2020). Additionally, sites along river edges are likely to have more transient impacts from fish, unlike those of ponds of similar sizes. These false negatives may impact the reported position of anuran species in relation to invasive fish, but they are unlikely to change the overall result. Similarly, we cannot discount the possibility that some anuran species went undetected during our surveys.

We did not include native fish in our scoring. To our knowledge, none of the impoundments that we surveyed contained any native fishes. Sites along the river are reported to have Cape Kurper Sandelia capensis and Galaxias sp. ‘Klein’ (see Chakona, Swartz & Gouws, 2013). Of these, the Cape Kurper may have exerted some predation impact on amphibians. There are other native predatory species that may exert an impact on amphibian communities, such as the Cape Clawless Otter Aonyx capensis, Cape Terrapin Pelomedusa galeata and the Western Cape River Crab Potamonautes perlatus. All of these species are present in the area sampled and further study would be required to interpret their impact on amphibian communities.

Conclusions

Anthropogenically created permanent water bodies (regardless of size) and the presence of invasive fish significantly alter amphibian communities in lowland fynbos by favouring widespread species. Our results question the dogma of creating urban ponds to increase biodiversity (Hassall, 2014; Hill, Lawson & Tuckett, 2017; Hill et al., 2018), at least for amphibian communities but possibly for other species. Recent success in restoring European amphibian populations with pond construction (Moor et al., 2022) needs to be taken in context, and not as a freshwater biodiversity panacea. Rather like the popular fixation on planting trees, the evolutionary and climatic context must take precedence when considering future conservation actions (Bond et al., 2019). While our research is pertinent to low-lying areas of the fynbos, the importance of hydroperiod in structuring aquatic communities, including amphibian communities, has been stressed before (e.g., Pechmann et al., 1989; Welborn, Skelly & Werner, 1996; Werner et al., 2007; Holbrook & Dorn, 2016). This is even more important in mediterranean biomes where permanent water is an unusual natural feature, but anthropogenic need for access to water particularly for agriculture have made it the most abundant of freshwater aquatic features. It may be that other southern African biomes may also have their amphibian communities strongly impacted by hydroperiod, but this remains untested (Kruger, Hamer & Du Preez, 2015). When opportunities arise for mitigation effects that call for creation of wetland habitats in the fynbos, we strongly encourage creation of temporary water features that are allowed to dry out during the summer months. This effectively excludes populations of invasive fish and increases the diversity of amphibian fauna endemic to the southwestern Cape lowlands.

Supplemental Information