A targeted survey for the Durban Dwarf Burrowing Skink Scelotes inornatus (Smith 1849) at Bluff Nature Reserve and Treasure Beach in Durban, KwaZulu-Natal, South Africa, with notes on sympatric herpetofauna

ABSTRACT

A survey for the Critically Endangered Durban Dwarf Burrowing Skink Scelotes inornatus (Smith 1849) was conducted in two protected areas in Durban, South Africa, in August and September 2021. Twelve sites, each encompassing a combination of vegetation type, elevation, slope and aspect, were sampled for S. inornatus using dug quadrats, pitfall-and funnel-trap arrays, and coverboard arrays. Seven S. inornatus were recorded, at least one by each sampling method. Individuals were captured at the forest edge and in the grassland part of the forest-grassland ecotone and none were captured in the interior of the forest. None of the sampling methods were suitable for the long-term monitoring of S. inornatus, although a modification of the coverboard array using terracotta tiles instead of corru-board tiles should be tested. Five S. inornatus captured were introduced to the Johannesburg Zoo’s ex-situ insurance and captive-breeding population. Nine other sympatric species of herpetofauna were recorded during the survey, with the Bush Squeaker Arthroleptis wahlbergii Smith, 1849 the most caught and S. inornatus ranking fifth. This study provides information that is useful for addressing some of the conservation actions and research needed for S. inornatus, but more research on the biology of this skink (particularly on its life history and population size) and habitat management interventions (restoration and rehabilitation of its habitat) are required to assist with improving its conservation status.

KEYWORDS:

• critically endangered reptile
• fossorial lizard
• capture methods
• habitat use
• sympatric herpetofauna

Introduction

The Durban Dwarf Burrowing Skink Scelotes inornatus (Smith, 1849) is a primarily fossorial limbless species and one of the larger members of the Scelotes Fitzinger, 1826 genus. It is endemic to a very narrow coastal strip between Durban and Scottburgh in southern KwaZulu-Natal (Broadley 1994; Marais and Bauer 2014). It is listed as Critically Endangered on the IUCN Red List owing to the destruction and fragmentation of its habitat through urbanisation, industrialisation, and agriculture (Marais and Bauer 2014; Alexander et al. 2022).

As a result of the precarious status of S. inornatus, a Memorandum of Understanding was signed between Johannesburg City Parks and Zoo and Ezemvelo KZN Wildlife to create a captive insurance and breeding population housed at the Johannesburg Zoo. One of the aims of the captive breeding is to provide offspring for release into the species’ natural range or adjacent to that in suitable habitat to increase the size of the wild population and re-establish S. inornatus populations at rehabilitated areas where the species has gone locally extinct.

Owing to the difficulty of obtaining wild parental stock for the captive breeding programme from unprotected areas, because of the widespread habitat destruction, a decision was made to survey two protected areas, namely Bluff Nature Reserve and Treasure Beach, for S. inornatus. The aims of the study were to: (1) ascertain what habitats had the highest abundances of S. inornatus; (2) develop a suitable monitoring protocol for the species; (3) collect additional breeder stock for the insurance population; and (4) record what other herpetofauna occur sympatrically with S. inornatus.

Materials and Methods

Bluff Nature Reserve and Treasure Beach (Figure 1) were stratified using the Idrisi TerrSet 19.0.4 GIS (Clark Labs, Clark University, Worcester, MA, U.S.A.) according to vegetation type (Scott-Shaw and Escott 2011), elevation, slope, and aspect, to delimit habitats. The vegetation types were KwaZulu-Natal Coastal Belt Grassland, Southern Moist Coastal Lowlands Forest and East Coast Dune Forest. Four elevational bands (m a.s.l.: 1 = 0–20, 2 = 21–40, 3 = 41–60, 4 = 61–81), three slope bands (°: 1 = 0–5, 2 = 6–11, 3 = 12–17) and four aspects (°: 1 = 316–45, 2 = 46–135, 3 = 136–225, 4 = 226–315) were used in the stratification. The projection of all the variable coverages was the Transverse Mercator lo 31 on the WGS84 datum and the pixel size was 20 m × 20 m.

Figure 1. Sampling sites in Bluff Nature Reserve and Treasure Beach in KwaZulu-Natal province, South Africa, where reptiles and amphibians were sampled from 16 August to 14 September 2021. Inset: Durban Dwarf Burrowing Skink Scelotes inornatus, Photo: PR Jordaan. Top Left: Location of the sampling sites within South Africa. Top right: Location of the sampling sites in relation to Durban.

Twelve sampling sites (Figure 1; Table 1), representing eleven unique variable combinations, were selected a priori after considering time and resource constraints. The actual sites sampled (Table 1) were those from the list of potential sites that were accessible, amenable to sampling by not being excessively wooded, did not have homeless people living nearby, and were within a fenced portion of each protected area to avoid human and domestic animal disturbance of the sampling sites. Sampling was conducted from 16 August to 14 September 2021, falling within the period of heightened S. inornatus surface activity (Alexander 1987). Three sampling techniques were employed to compare the effectiveness and feasibility of each method for the sampling and potential annual monitoring of S. inornatus at Bluff Nature Reserve and Treasure Beach. These were excavated quadrats, pitfall-and funnel-trap arrays, and coverboard arrays.

Table 1. Combinations of environmental variables used to stratify the Bluff Nature Reserve and Treasure Beach in Durban, KwaZulu-Natal province, South Africa, to guide the survey for Durban Dwarf Burrowing Skinks Scelotes inornatus, the code for each sampling site, and the soil type and texture at each site. Elevation (m a.s.l.): 1 = 0–20, 2 = 21–40, 3 = 41–60, 4 = 61–81; slope (°): 1 = 0–5, 2 = 6–11, 3 = 12–17; aspect (°): 1 = 316–45, 2 = 46–135, 3 = 136–225, 4 = 226–315; *grassy area adjacent to Southern Moist Coastal Lowlands Forest vegetation.

Protected area	Vegetation type	Elevation	Slope	Aspect	Sampling site code	Soil type	Soil texture
Bluff Nature Reserve	Southern Moist Coastal Lowlands Forest	1	1	4	I	Loamy sand	Medium to moderately fine
		2	1	1	H*	Loam	Medium to moderately fine
		2	1	2	E	Loamy sand	Medium to moderately fine
		2	2	2	G	Loam	Medium
		3	2	2	C and F	Sandy loam	Moderately fine
		3	3	2	D	Sandy loam	Medium
		4	2	2	B	Loam	Medium
		4	2	3	A	Loamy sand	Medium to moderately fine
Treasure Beach	KwaZulu-Natal Coastal Belt Grassland	5	2	3	J	Loam	Medium
	East Coast Dune Forest	4	4	2	L	Sandy loam	Medium
	Southern Moist Coastal Lowlands Forest	5	2	3	K	Sandy loam	Medium

Fossorial herpetofaunal quadrat sampling was conducted following the excavation and processing methodology of PR Jordaan et al. (unpublished data). At each sampling site, two 4 m² quadrats (2 m × 2 m) were marked using an 8 m long rope held in place by droppers at each corner. Each quadrat was then excavated using spades to a depth of 0.3 m by two persons on opposite sides, starting along the perimeter of the demarcated quadrat. Once the perimeter was fully excavated and cleared of loose sand, the rest of the quadrat area was excavated. The excavated sand was stored in 20-litre buckets until completion of the excavation, after which the sand from each bucket was individually emptied onto a sifting sheet and searched by hand for S. inornatus and other herpetofauna whilst removing plant material for replanting after sampling was complete. The sand was then passed through an enamelled metal mesh sieve with a mesh dimension of 2 mm × 2 mm to expose any remaining herpetofauna in the sample. All herpetofauna were recorded, measured, weighed, photographed, and released except those S. inornatus that were retained for the captive insurance population and breeding project at Johannesburg Zoo. During the fossorial quadrat sampling, the phase responsible for exposing amphibians and reptiles (excavation, sorting, or sifting), was recorded for each specimen. Afterwards, the sand was returned to the excavated quadrat pit and uprooted plants replanted. The soil surface in the quadrats was covered mainly by leaf litter in the forest sites and by forbs and grass in the grassland sites, with the percentage cover varying from 90–100%. The quadrats were sited within 20 m of the pitfall-and funnel-trap arrays.

In addition to fossorial quadrat sampling, pitfall- and funnel-trap arrays and coverboard arrays were employed. The pitfall- and funnel-trap arrays consisted of a three-armed drift fence structure converging at a central pitfall trap with additional pitfall traps installed at the end of each of the three 5-m long arms. Pitfall traps consisted of 20-litre plastic buckets dug into the soil. Drift fences were constructed by linking four 1.25 m long corru-board sheets to each other using Velcro® links. The bottom edge of each drift fence and the pitfall bucket openings were sunken 0.2 m below the soil surface, guiding fossorial reptiles moving through the upper soil column, such as the target species, toward the pitfall traps (Henderson et al. 2016). Bucket lids were cut in half and two halves were placed on wooden stakes on either side of the drift fence that ended above each bucket to shelter trapped animals from exposure. Some plant litter and sand were added to each pitfall bucket to provide additional shelter for fossorial herpetofauna that fell into the bucket. Funnel traps, consisting of inward-facing plastic funnels enclosed with metal enamelled steel mesh netting with an aperture size of 2 mm × 2 mm, were deployed on either side of and midway along each drift fence. Funnel traps were partially covered with vegetation to shelter any trapped animals. All pitfall- and funnel-trap arrays were inspected at least once and usually twice a day to collect captured herpetofauna for processing and to release small mammal and invertebrate bycatch. Following the survey period, all survey materials and equipment were removed, and the pitfall holes were filled with sand.

Each coverboard array was placed outside the imaginary circle encompassing the arms of the pitfall- and funnel-trap array in the closest large-enough area devoid of obstructing vegetation. The coverboard array consisted of eight 0.4 m × 0.4 m corru-boards placed on the soil surface, arranged in a grid of four boards in two rows with all boards separated by 0.1 m from each other. Each of the eight corru-board sheets were fixed in place with wooden stakes to prevent them from blowing or washing away. The coverboard arrays were left in the field and were checked for S. inornatus on 29 November 2021 and on 28 February, 26 May and 14 November 2022, and then removed.

Captured specimens were identified using Branch (1998) for reptiles and du Preez and Carruthers (2017) for amphibians, recorded, measured, and photographed where possible to facilitate identification and released near the capture site, except those S. inornatus that were to be retained for the captive breeding project. Measurements taken were snout-vent length (SVL, from tip of snout to the posterior edge of the cloaca; mm), tail length (mm, from the posterior edge of the cloaca to the tip of the tail; for reptiles), total length (mm) and mass (g). Vernier callipers were used to measure the length of the animals to 0.1 mm precision and a digital scale (AE Adam SKS, Milton Keynes, United Kingdom) was used to weigh the captured animals up to 0.1 g accuracy.

The soil texture at each sampling site was assessed in the field by the mud-ball test (FAO 2020), and the soil type (textural class) by the manipulative test and the textural triangle diagram (FAO 2006). Hourly meteorological data for Virginia Airport in Durban, situated approximately 20 km north of the sampling sites and the closest weather station, were obtained from the South African Weather Service. The following calculated metrics were used in the analyses: mean daily wind speed (m/s; Wind), mean daily temperature (°C; Tmean), daily rainfall (mm; Rain) and mean daily relative humidity (%; RH). The data were explored and analysed in R (R Core Team 2021) using the R Commander interface (Fox 2005).

Results

Ninety-seven reptiles and amphibians of ten species were recorded during the survey (Table 2). The highest number of species were captured in forest at its edge or immediately adjacent to forest (sites A and H with five and seven species, respectively). The fewest species were captured at some forest interior sites (sites B, D and E) with one species each. The greatest number of herpetofauna were captured at grassland sites (sites H and J) with 18 and 15 individuals, respectively. In terms of the pitfall- and funnel-trap arrays, funnel traps captured more species than pitfall traps (eight vs six species), but pitfall traps captured the most individuals (n = 36). The fossorial quadrats captured fewer species (six) than the pitfall- and funnel-trap arrays and produced fewer specimens (n = 28). The focal species, S. inornatus, was captured at four sites (Table 2), in pitfall- and funnel-trap arrays at three sites and during fossorial quadrat sampling at two sites.

Table 2. Numbers of each herpetofaunal species captured during the survey using pitfall- and funnel-trap arrays and excavated quadrats at the Bluff Nature Reserve and Treasure Beach in Durban, KwaZulu-Natal province, South Africa, from 16 August to 14 September 2021. A = Amphibian, R = Reptile, F = Funnel trap and P = Pitfall trap in the pitfall- and funnel-trap array, Q = excavated Quadrat.

Species	Site												Total
	A	B	C	D	E	F	G	H	I	J	K	L
Arthroleptis wahlbergii (A)	1F	3P1Q	2F2P1Q		2F5P	1F6P1Q	2P4Q		1Q		3P1Q	2P	38
Breviceps sp. (A)										2F2P			4
Hyperolius pusillus (A)	1F			1Q		2Q		1Q					5
Leptopelis natalensis (A)							1Q	1Q	2P3Q			1Q	8
Aparallactus capensis (R)	3Q		2P								1P		6
Gerrhosaurus flavigularis (R)								2F		1F			3
Lycodonomorphus rufulus (R)								1F					1
Panaspis wahlbergii (R)							1Q	3F		6F3P1Q	1P		15
Philothamnus hoplogaster (R)	1F							7F	2F				10
Scelotes inornatus (R)	1P							3Q	1F1Q		1P		7
Total species	5	1	2	1	1	2	3	7	4	3	4	2	10
Total abundance	7	4	7	1	7	10	8	18	10	15	7	3	97
Total species (F)	3		1		1	1		4	2	3			8
Total abundance (F)	3		2		2	1		13	3	9			33
Total species (P)	1	1	2		1	1	1		1	2	4	1	6
Total abundance (P)	1	3	4		5	6	2		2	5	6	2	36
Total species (Q)	1	1	1	1		2	3	3	3	1	1	1	6
Total abundance (Q)	3	1	1	1		3	6	5	5	1	1	1	28

Scelotes inornatus (n = 7 individuals) was the fifth most frequently captured species, comprising 7.2% of the total captures. Most S. inornatus were recorded during fossorial quadrat sampling (57.1% of the captured specimens). This species was found in forest at its edge and in grassland immediately adjacent to forest, but none were recorded in the forest interior (Table 2). Scelotes inornatus was captured in loamy sand of medium to moderately fine texture, loam of medium to moderately fine texture, and sandy loam of medium texture. The excavated quadrats yielded a mean ± 1 SD density of 0.167 ± 0.4815 S. inornatus per quadrat, translating to an estimated density of 417 individuals per hectare.

The Bush Squeaker Arthroleptis wahlbergii (Smith, 1849) was the most commonly caught species (Table 2) representing 39.2% of the total captures, followed by Wahlberg’s Snake-eyed Skink Panaspis wahlbergii (Smith, 1849) (15.5% of total captures) and the Southeastern Green Snake Philothamnus hoplogaster (Günther, 1863) (10.3% of total captures). Arthroleptis wahlbergii was the most widespread species, occurring at nine (75%) of the sites, and was captured at all but one forest site. The excavated quadrats yielded a mean ± 1 SD of 0.375 ± 0.8754 Arthroleptis wahlbergii per quadrat, translating to an estimated density of 938 individuals per hectare.

There was no significant correlation between any weather variable and the total number of frogs, reptiles, or herpetofauna captured per day by the pitfall- and funnel-trap arrays (Figure 2), although a slight increase in the number of herpetofauna captured was apparent the day after the main rainfall event during the sampling period (28 August 2021). Mensural data for many of the herpetofauna captured are presented in Table 3. The S. inornatus specimens captured during the study varied markedly in size but the individuals could not be aged or sexed. The largest S. inornatus had a SVL of 87.3 mm, close to the maximum recorded (90 mm; Broadley 1994).

Figure 2. Number of herpetofauna (Herps), reptiles and amphibians captured in pitfall- and funnel-trap arrays at Bluff Nature Reserve and Treasure Beach in Durban, KwaZulu-Natal province, South Africa, in relation to weather variables from 22 August 2021 to 4 September 2021.

Table 3. Morphological measurements of specimens that were captured during the survey conducted at the Bluff Nature Reserve and Treasure Beach in Durban, KwaZulu-Natal province, South Africa, from 16 August to 14 September 2021. SVL = snout-vent length, TL = total length, SD = standard deviation, A = Amphibian, R = Reptile.

Species	SVL (mm)			Tail (mm)			TL (mm)			Weight (g)			N
	Mean	SD	Range	Mean	SD	Range	Mean	SD	Range	Mean	SD	Range
Arthroleptis wahlbergii (A)	23.7	3.49	16.9–28.9							1.3	0.61	0.3–2.4	35
Breviceps sp. (A)	34.3	5.55	25.2–40.4							6.0	3.09	2.4–10.5	5
Hyperolius pusillus (A)	15.7	2.53	13.6–19.2							0.3	0.17	0.1–0.5	4
Leptopelis natalensis (A)	33.7	18.95	18.9–68							6.9	9.92	0.5–27.2	9
Aparallactus capensis (R)	162.3	39.38	111–204	39.2	9.45	24.5–47.4	201.5	47.95	141.2–247.7	2.8	1.56	1–4.4	6
Gerrhosaurus flavigularis (R)	90.0	35.00	49.6–111.8	131.0	25.25	103–152	221.0	59.47	152.6–260.5	21.1	16.06	2.6–31.8	3
Lycodonomorphus rufulus (R)	317.0			115.0			432.0			13.6			1
Panaspis wahlbergii (R)	39.0	2.06	34.3–42	40.2	13.72	37–61	79.2	13.83	59.4–101.6	1.0	0.20	0.7–1.3	14
Philothamnus hoplogaster (R)	384.7	29.77	356–444	182.4	19.79	155–204	567.1	46.50	511–648	16.7	2.56	13.4–20.9	9
Scelotes inornatus (R)	68.1	17.58	48.6–87.3	50.5	12.98	28.2–59	118.6	28.88	76.8–144.1	1.6	1.14	0.5–2.9	5

Most individuals of the herpetofauna found during the fossorial quadrat sampling were exposed during the digging phase (n = 16) in contrast to the sorting phase (n = 9) and the sieving phase (n = 3). No species were recorded in fossorial quadrats that were not recorded in the pitfall- and funnel-trap arrays. However, like S. inornatus, more Water Lily Frogs Hyperolius pusillus (Cope, 1862) and Natal Tree Frogs Leptopelis natalensis (Smith, 1849) were recorded by the fossorial quadrat method. The predominantly soil-living rain frog Breviceps sp. was not recorded during the fossorial quadrat survey but was captured in a pitfall- and funnel-trap array.

Only one S. inornatus was recorded under a coverboard, at site I on 29 November 2021. This site was underwater on 26 May 2022 and so could no longer be monitored. No other herpetofauna were recorded under coverboards.

Discussion

The Critically Endangered S. inornatus appears to favour forest-grassland ecotone habitat and did not seem to avoid any of the soil types and soil textures present at the sampling sites (this study; Alexander and Marais 2007). The skink’s remaining intact habitat is small and declining in size in its extremely limited geographic distribution, due to the almost total transformation of the Critically Endangered KwaZulu-Natal Coastal Belt Grassland that borders the skink’s coastal forest habitat (Jewitt 2018). Only 11% of the total original area of this grassland remained in 2016, and with land transformation proceeding apace in the province of KwaZulu-Natal, far less will remain untransformed this decade (Jewitt 2018; Jewitt et al. 2015). Another threat is the invasion of alien plant species at the forest-grassland ecotone and the wooding up of coastal grasslands due to suppression of fire, the lack of large browsing herbivores in urban protected areas, and the increased carbon dioxide levels in the atmosphere (O’Connor et al. 2014; Stevens et al. 2016). Pollution also may be a threat to the species. At least eight of the 24 excavated quadrats contained anthropogenic litter such as plastic, glass, and iron. The litter is either dumped in the protected areas or enters through surface water run-on or through wind-blow. This litter may fragment the soil habitat for the fossorial species and their prey and may adsorb or release toxic chemicals during decomposition that might have impacts on the health of the soil ecosystem (Teuten et al. 2009; Sajjad et al. 2022).

The relatively low density and litter size of S. inornatus further highlights the vulnerability of the species (Alexander and Marais 2007). Little is known of the biology and ecology of S. inornatus in comparison to some of the other species captured, such as A. wahlbergii (e.g., Schweiger et al. 2017; Tolley et al. 2018), and this hampers the development of scientifically based conservation actions for S. inornatus. Although a viable population of S. inornatus could potentially be maintained in a suitable habitat in the order of a few tens of hectares, this potential might not be realised at present owing to the degraded state and fragmentation of much of the skink’s remaining habitat. The protection and appropriate conservation management of the remaining habitat within the geographic distribution of S. inornatus and the restoration or creation of additional suitable habitat for the species are priority conservation actions. However, further research on the biology and ecology of S. inornatus is required to provide a solid scientific foundation for the development of such conservation actions.

The method used for the monitoring of S. inornatus populations must be efficient in terms of resources (cost, time, personnel) and environmentally friendly while still providing adequate data to measure trends in the sub-population size. Four individuals were recorded at two sites during quadrat sampling in comparison to three at three sites by the pitfall- and funnel-trap arrays (Table 2) and one under the coverboard arrays. The capture rate of all three methods was low, and therefore none of the methods assessed appear suitable for the long-term monitoring of this Critically Endangered species. Apart from the time and effort required to implement the excavated quadrat method, the potential for the accidental killing of S. inornatus and the uprooting of low vegetation mitigate against the widespread use of this method.

One of the disadvantages of using corru-boards was their thinness and lightness, which resulted in some boards being disturbed and displaced over time while their light weight meant that the entire ventral surface did not necessarily make contact with the soil surface. However, the coverboard array was the least intensive and intrusive method and should be tested using large terracotta tiles instead of corru-boards to assess whether the trapping success rate might thereby be improved. Alternatively, the use of detection dogs may be a viable option to survey for the presence of S. inornatus or even monitor populations as has been conducted for other fossorial skink species (e.g., Nielsen et al. 2016). Recently, the presence of well-developed epidermal glandular tissues was reported for male Lowveld Dwarf-Burrowing Skinks Scelotes bidigittatus FitzSimons, 1930 with postulated scent production (Jordaan and Steyl, in press). If the presence of similar glands is confirmed for S. inornatus, the use of detection dogs to locate the species should be investigated as a possible monitoring tool.

Five S. inornatus individuals were added to the Johannesburg Zoo captive-breeding and insurance population. The maintenance of an ex-situ breeding population appears necessary for the survival of the species and to eventually increase the size and area of occupancy of its overall population (Alexander et al. 2022). The development of a husbandry manual for the species should assist in the establishment and breeding of other captive populations, provided that the source populations are not detrimentally impacted through removal of founder stock (IUCN SSC 2013).

Regarding the sympatric herpetofauna captured during the survey, some interesting data were collected. The largest A. wahlbergii measured had a SVL of 28.9 mm, exceeding the maximum size (25 mm) noted by du Preez and Curruthers (2017), as did the largest Leptopelis natalensis (68 mm versus 65 mm). The discovery of five adult L. natalensis, one mature (SVL 68 mm) and four immature (SVL 18.9–25.3 mm) individuals, in the sandy soil indicates that this species may spend time underground not only for the purpose of oviposition. No consistent effect of the weather on the numbers of each higher taxon captured could be ascertained, apart from a slight peak in the number of herpetofauna captured after the main rainfall event during the sampling period. This may be a result of the weather in early summer being generally favourable for the activity of all the species encountered. Although most herpetofauna recorded during the fossorial quadrat sampling were observed during the digging phase, some were still recorded while sorting through the excavated material by hand and when sifting the soil through a sieve. The time and effort required to implement these three phases may therefore be warranted when this method is used to sample herpetofauna.

This study provided specimens and information useful for addressing some of the conservation actions and research needed for S. inornatus as listed under ‘Species management’ and ‘Research’ in the Red List assessment (Alexander et al. 2022). This study provided: (1) five S. inornatus for the ex-situ population at Johannesburg Zoo (Conservation Action 3.4 Ex-situ conservation); (2) some information on the ecology of S. inornatus (Research Action 1.3 Life history and Ecology); (3) awareness of the potential threat of pollution in the context of Research Need 1.5; and (4) recommendations in the context of Research Need 3: Monitoring. However, much more research on the biology of this skink (particularly on its life history and population size) and habitat management interventions (restoration and rehabilitation of its habitat) are required to assist with improving its conservation status. We hope that this study has provided impetus for addressing these research and conservation needs.

Acknowledgements

We thank Mr Kevin Lakani of the Wildlife and Environmental Society of South Africa and Mr Errol Douwes and Dr Colin Pillay of the eThekwini Metro for permission to do the research at Treasure Beach, and Ms Dhirusha Raghunandan, Client Liaison Officer: Durban, of the South African Weather Service for the weather data collected at Virginia Airport for the monitoring period. Geographical Information System coverages were provided by Ezemvelo KZN Wildlife. For assistance in the field, we thank Mr Ian du Plessis, Mr Kallyn Gunkel, Ms Corrina Naidoo, Mr Kian Moodley, Dr Cormac Price, Mr Albert Wilken, and Mr Mesuli Zondi. Ms Bimall Naidoo kindly produced Figure 1. Two anonymous reviewers improved the manuscript considerably. The work was conducted under Ezemvelo KZN Wildlife permits OP 2301/2021 and OP 2402/2021 and eThekwini Municipality LOS 6 August 2021.

Disclosure statement

No potential conflict of interest was reported by the author(s).