Introduction

Habitat complexity, which refers to the different types, numbers and arrangements of structural elements in a given area (McCoy and Bell 1991; Tews et al. 2004; Tokeshi and Arakaki 2012; Loke and Chisholm 2022), is an important driver of biodiversity (Pianka 2000; Loke and Todd 2016; Strain et al. 2021). The prevalent hypothesis is that topographically complex habitats have a higher number of distinct resources and variations in environmental conditions, therefore allowing the coexistence of different species (MacArthur and MacArthur 1961). However, human activities, such as the replacement or fragmentation of natural habitats by artificial structures, are leading to a global physical homogenisation (i.e., loss of topographically complex habitats) of natural habitats (e.g., Jongman 2002; McKinney 2006; Thrush et al. 2006). This uniformity of land and seascapes is likely to lead to declines in the diversity of resources, resulting in significant biodiversity loss across habitats at global scales. To reverse this biodiversity loss, management interventions aimed at increasing the complexity of urban infrastructure are being attempted (Strain et al. 2018; Lengyel et al. 2020). Manipulations of complexity and their effect on organisms can vary across different scales (Archambault and Bourget 1996; Matias et al. 2010). To make interventions most effective, the complexity of habitat needs to match the requirements of the species it is designed to support.

Rocky shores are highly diverse habitats and have varying levels of topographical complexity. Topographic complexity, hereafter referred to as complexity, in these habitats is provided by microhabitats such as rock pools, pits, grooves or biogenic habitats (e.g., macro-algae, oysters) (McGuinness and Underwood 1986; Sebens 1991). Microhabitats can mediate, among other things, environmental stressors, and species interactions, such as competition and predation (Garrity 1984; McGuinness and Underwood 1986; Klein et al. 2011; Aguilera et al. 2014). Rock pools are ubiquitous features of natural rocky shores and add complexity by providing a constantly submerged habitat on the otherwise dry rock bed. They can support greater species richness and species abundances compared to emergent rock substrata (Firth et al. 2013, 2014; Aguilera et al. 2022, but see Bugnot et al. 2018), and can provide a protective habitat for a variety of species not found on emergent rock (Chapman 2003; Bugnot et al. 2018). Mobile taxa may also be largely limited to rock pools on natural rocky shores. Rock pools are therefore important in influencing the structure (i.e., composition and relative abundance of species) and functioning of rocky shore communities. The urbanisation of coastal habitats has, however, caused the fragmentation and/or replacement of many natural rocky shores by artificial structures such as seawalls and pilings (Goodsell et al. 2007; Bishop et al. 2017), which have reduced the availability of or entirely lack these important habitats. Such structures usually support different assemblages regarding species composition and abundance, often with less species overall and a lower number of species and abundance of mobile fauna (Chapman 2003; Bulleri and Chapman 2010; Firth et al. 2014; Mayer-Pinto et al. 2018). Therefore, the addition of rock pool mimics to artificial structures, such as flowerpots”, “vertipools’ (Browne and Chapman 2014; Hall et al. 2019), or drill-holes (Evans et al. 2016), has the potential to mitigate, or partly mitigate, this loss.

Rock pools come in various shapes and sizes, and previous research has shown that a variety of rock pool characteristics can influence mobile diversity (Metaxas and Scheibling 1993; Bugnot et al. 2018; Schaefer et al. 2019). For example, deeper and larger pools support an increased diversity of fish (Bugnot et al. 2018). For benthic invertebrates, the effect of rock pool size characteristics varied. For example, rock pool diameter had no effect on mobile species (Underwood and Skilleter 1996), whereas rock pool depth had no consistent effect on mobile species (Astles 1993; Firth et al. 2014; Schaefer et al. 2019), which may be linked to species-dependent effects (Astles 1993; Bugnot et al. 2018), or the local environmental conditions (Schaefer et al. 2019). To date, artificial rock pool designs have been of simple shape and based on variations in these characteristics (e.g., total size and/or depth), but have not considered smaller-scale topographical complexity features (microhabitats) within rock pools. Given the small size of mobile species in the pools (usually millimetres to a few centimetres) relative to the entire pool size (usually tens of centimetres), it is important to assess the effect of complexity at the scale relevant for the organism (Matias et al. 2010). Smaller scale features such as overhangs and pits within rock pools can provide thermal refuges even in submerged conditions. Pits within rock pools for example can vary in temperature compared to the rest of the rock pool (Morris and Taylor 1983). Similarly, overhangs within the rock can provide shaded habitat (Waltham and Sheaves 2018), whereas vertical features are also more likely to be shaded throughout part of the day, reducing the duration of solar exposure (Firth et al. 2014). These smaller-scale topographical features (microhabitats) within pools, can also influence the diversity of pools. For instance, rock pools with more algal cover and a greater percentage of overhanging rock ledges have been linked to greater fish abundances and potential predator-avoidance strategies (White et al. 2015). Additionally, some fish species were only found in pools with specific characteristics (e.g. loose shells), which suggested a level of habitat selectivity (White et al. 2015). Observations done during surveys of the size, depth, and position of rockpools on intertidal rocky shores in Sydney (Schaefer et al. 2019) suggest that features of pools may also influence the distribution of mobile macro-invertebrates. Assessing microhabitats within rock pools may therefore provide an important link for detailed habitat-diversity relationships that can be used to build more sophisticated rock pool designs that target mobile macro-invertebrates. Advancements in technologies, such as 3D laser scanning and 3D printing, enable engineers to design and build highly structured and more complex structures replicating natural topographic features (Evans et al. 2021), which can also be applied to artificial rock pool designs. This, however, requires knowledge of features present in natural rock pools to inform these designs. (Evans et al. 2021).

Here, we surveyed natural rock pools to assess the types and abundance of different topographical features (microhabitats, here: overhangs or pits) that are commonly present at 3 zones in Sydney: along the open coast, and in two areas within Sydney Harbour: the outer zone, which has greater oceanic flushing, and in the more sheltered inner zone of the Harbour. At a single zone (open coast) we assessed whether pits and overhangs within rock pools influence the richness of mobile taxa and abundance of mobile macro-invertebrates, and whether this varies between seasons. We hypothesized that rock pools with overhangs or pits present would have greater diversity and abundance of mobile macro-invertebrates compared to rock pools without overhangs or pits. We also expected this pattern to vary across zones and to be stronger in summer than in winter, due to the protection provided by the two microhabitats from temperature stress.

Materials and methods

Study area

Intertidal rock pools were sampled at seven sites across three zones (as described above) in Sydney, New South Wales, Australia to assess the generality of results at larger spatial scales. Rocky shores in this area are composed of sandstone and are mainly horizontal or gently sloped (Johnston et al. 2015). Three sites (Coastal: Bondi, Freshwater and Curl Curl) were located along the coastline of Sydney (open coast) and four sites were located within Sydney Harbour, one of the largest urbanised harbours in the world (33° 51′ S, 126 151° 14′ E). Of those, two sites (Outer Harbour: Bradleys Head and Delwood) were located in the outer zone, closer to the mouth of the estuary, and two sites (Inner Harbour: Balmain and Berry Island Reserve) were located in the inner zone of the estuary (Fig. 1).

Fig. 1
figure 1

Map of Sydney showing the location of the estuary on the coast of NSW, Australia. Rock pools were sampled at two sites located in the inner zone of Sydney Harbour [Balmain (B), Berry Island Reserve (BIR)], two sites in the outer zone of Sydney Harbour [Bradleys Head (BH), Delwood (D), and three sites along the open coast (Bondi (NB), Freshwater (F), Curl Curl (CC)]. Inner and outer zones are indicated and redrawn by Dafforn et al. (2015)

Full size image

At each site, 16 natural rock pools of a range of sizes (i.e., volumes) were haphazardly selected (Table S1). Selected pools were distributed across the intertidal (from 0.2 MLLW to 0.8 MLLW). For this study, rock pools were defined as depressions in the rock shelf that retain water during the entire low tide and included all rock pool features permanently covered by water and along the interface between the pool and the horizontal exposed rock (Table 1, Fig. S2.1–2.9).

Table 1 Microhabitats assessed in each rock pool and the definitions used for each feature
Full size table

Rock pool characteristics

The volume of each pool was measured once in situ by emptying water from the pools into a measuring cylinder (1 L jar or 50 mL syringe, depending on the size of the pool). Different physical features, including vertical and sloping areas within the pools, as well as the types of microhabitats, were identified (Table 1, Fig. S2.1–2.10). Here, the focus was on two types of microhabitats that change the topographic complexity within rock pools: overhangs and pits. We estimated the size of each microhabitat as the proportion (in percentage) of the area ‘made up’ by the microhabitat in relation to the total rock pool area when viewed directly from above. The total rock pool surface area (= 100%) included all emerged and submerged parts of pools (Supplementary Material 2). Submerged features were classified as any part of the rock pool permanently covered by water, while emergent features included the interface (defined as the area between rock pool water and horizontal exposed rock) between pools and exposed rock (Fig. S.10) Therefore, overhangs estimated to account for a quarter of the total area of the pool, were considered to ‘make up’ 25% of the pool. Canopy cover, if present, was estimated as stated above and moved to look for hidden pits and overhangs.

Rock pool biota and habitat mapping

The abundance of mobile macro-invertebrates within rock pools was determined twice in Austral Winter 2016 and twice in Austral Summer 2016/2017. Surveys were done 6–8 weeks apart within each season. During each sampling, rock pools were emptied of all water and videos or photos of the entire rock pool were then taken with a handheld iPhone 5S camera positioned ~20 cm above each pool (camera: 8 megapixels with 1.5µ pixels, video recording 1080p HD (30fps)). The type of assessment (video or photo) did not affect the number or types of species found (for comparison of methodologies see Supplementary Material 1). Additional close-ups of overhangs were taken as these microhabitats were not visually accessible when looking from above. From the videos and photos, mobile organisms were identified, assigned to a physical feature within the pool (Table 1) and counted. This allowed us to determine the total abundance of mobile organisms per physical feature within the pool as well as the whole rock pool. Organisms were identified to the lowest possible taxonomic level (usually species or genus, phylum for flatworms). Taxa that could not be identified to genus level were classified into broad functional/taxonomic groups (e.g., limpets, barnacles, etc.) as “unidentified” (e.g., unidentified limpet). Due to the limited resolution of images and videos, organisms smaller than ~0.5 cm were excluded from the analyses. Faster-moving mobile organisms (fish and crustaceans) occurred in low abundances in the surveyed pools and were excluded from analyses, as they could not be assigned to a particular physical feature.

Statistical analyses

Due to the overall low number of pits and overhangs within pools at the Inner and Outer Harbour, the abundance of mobile macro-invertebrates within rock pools with and without pits and overhangs could only be quantitively compared in the three sites at the open coast zone. Biogenic features were not considered microhabitats since they can be of temporary nature. Extreme weather and/or algal canopy prevented accurate assessment of abundances of mobile taxa in 8 (out of 48) pools, so those were excluded from statistical analyses. Rock pools included in analyses had a maximum algal canopy cover of 20%. Three rock pools had both pits and overhangs and were excluded as the effects of individual microhabitats could not be distinguished. Eleven pools had only overhangs, 8 pools had only pits, and 18 pools had no pits/overhangs. As such, there were still sufficient replicates to address the hypotheses at one of the zones. To compare pools with and without any type of microhabitat of similar sizes, only the 8 (for analyses with pits) or 11(for analyses with overhangs) largest rock pools without any microhabitats were included in analyses. Therefore, equal numbers of pools (n = 8 for analyses of pits, 11 for analyses of overhangs) with and without pits or overhangs were compared. Nevertheless, the sizes of analysed pools ranged from 0.26 to 12.560 L, therefore rock pool volume (as a proxy of size) was used as a covariate in the model to account for differences in size (see below).

Richness and abundance of mobile macro-invertebrates within rock pools with and without microhabitats

To assess whether the richness and abundance of mobile macro-invertebrates varied between rock pools with and without pits or overhangs, we used generalised linear mixed effects models (GLMM) in the package ‘glmmTMB’ (Magnusson et al. 2017). The model included an interaction of the categorical factors overhang or pit (present, absent) and season (winter, summer). Site, sampling time and an individual rock pool identifier were included as random factors in the model to account for the repeated sampling of the same rock pools over time and potential spatial autocorrelation of rock pools within sites. The volume of each rock pool was used as a proxy for the rock pool size and was included as a covariate to account for differences in the size of rock pools with and without pits or overhangs. Residual and qqplots of all models were checked using the ‘DHARMa’ package (Hartig 2017). Data were overdispersed, so data were transformed or distributions adjusted to account for this. A gaussian distribution of log-transformed data (log + 1) was used for taxon richness, and a negative binomial distribution (nbinom1) was used in all analyses for abundance. While Poisson distribution is often more appropriate to use for count data, such as taxon richness, the residuals fit model assumptions of heterogeneity and normality better with a normal distribution, which was therefore chosen for taxon richness only. There were some minor violations of assumptions, but generalised linear mixed effect models are robust again these violations (Schielzeth et al. 2020). For analyses, we did not differentiate between pits of different sizes or overhangs above and below the waterline due to overall low abundances within each microhabitat category. To ensure enough replication of each level of the type of pool (i.e. with pits, with overhangs and without pits or overhangs), analyses comparing rock pools with overhangs or pits with pools without these microhabitats were done separately; i.e. overhang vs no overhang and pits vs no pits. P-values were obtained using type III Qui-Square (Wald) tests using the package ‘car’ (Fox et al. 2012). The package ‘emmeans’ (Lenth et al. 2019) was used to obtain estimated marginal means and SEs for plotting results. Analyses of the abundance of the three most abundant species were done as described above to test whether results were driven by particular species. All statistical analyses were conducted in R (Version 4.1.0).

Additionally, we calculated species frequency of occurrence (i.e. number of times a species was present in a particular pool across all sampling times) in pools with and without pits or overhangs.

Results

Characterisation of rook pools among zones

Volumes of rock pools varied greatly among zones, with pools at the coastal site of Freshwater having, on average, the highest volume, followed by the Inner Harbour site, Berry Island Reserve (Table 2). Rock pools with the lowest volumes, on average, were found at the coastal site, Curl Curl (Table 2). Average temperatures of rock pools at each site varied between sites and sampling times and were 2–12 °C colder in winter than in summer.

Table 2 Mean, minimum and maximum values of volumes of rock pools (mL) are presented for each zone and site and mean rock pool water temperature at the time of sampling (low tide) across all pools within each site at each sampling time (1–4)
Full size table

The number of different types of microhabitats (pits or overhangs) found concurrently within any given pool and the abundance of rock pools with microhabitats within each zone decreased with increasing distance upstream towards the inner parts of Sydney Harbour (Fig. 2). Fifty percent of rock pools at coastal sites had at least one type of microhabitat, while microhabitats were only found in ~ 28% and ~ 16% of the outer and inner harbour rock pools, respectively (Fig. 2). Pits were the most common type of physical microhabitats within pools, with 39 pits found in a total of 24 pools (out of 112 pools sampled) across the three zones (27 pits in 14 pools at coastal sites; 12 pits within 10 pools in the Outer Harbour). Of the 39 pits, small pits were the most common type found (26), followed by medium (10) and deep pits (3).

Fig. 2
figure 2

Proportion of rock pools that had none, one or two or more types of microhabitats (pits or overhangs) within a single rock pool in each zone. No. of pools = 48 (Coastal), no. of pools = 32 (Outer Harbour), no. of pools = 32 (Inner Harbour). N none; OH overhang; DP deep pit; MP medium pit; SP shallow pit

Full size image

Similarly, the total percentage area of rock pool ‘made up’ by pits and/or overhangs (relative to the overall area of the pool) decreased from the coastal to the Inner Harbour zone, ranging from 2% to 75% at the coastal site, 5–40% in the Outer Harbour and 5–20% in the Inner Harbour (Fig. S1.1).

Relationships between the presence of physical microhabitats within rock pools and the abundance of mobile organisms

We found that the presence of overhangs or pits increased the richness of mobile taxa (Fig. 3a,  b, Table 3), whereas season had no effect on the richness of mobile taxa. Based on the frequency of occurrence of individual taxa, a few species could explain this trend (Table S1.2). For example, the gastropod Nodilittorina pyramidalis, the seastar Meridiastra calcar and Ischnochitons were only found in very low abundances (3, 1 and 1 individuals, respectively), but were only present in rock pools with overhang or pit. The seastar Parvulastra exigua occurred in similar numbers when comparing rock pools with and without pits, but its occurrence more than doubled in rock pools with overhangs compared to those without overhangs. The gastropods Austrocochlea spp. occurred more frequently in rock pools with overhangs or pits. The predatory whelks Tenguella marginalba and Dicathais orbita occurred two to three times more often in rock pools with an overhang compared to those without overhangs. Furthermore, chitons, particularly the most abundant chiton Sipharochiton pelliserpentis, occurred more frequently in rock pools with overhangs or pits compared to those without pits or overhangs. The most common limpet found in our surveys, Cellana tramoserica occurred at similar frequencies in rock pools with and without pits or overhangs.

Fig. 3
figure 3

Richness of mobile taxa (a, b), abundance of mobile macro-invertebrates (c) and Nerita melanotragus (d) found in rock pools with pits and without microhabitats (pits or overhangs) (a) and with overhangs and without pits or overhangs (b, c). Bars are model means and standard errors. *P < 0.05, **P < 0.01, ***P < 0.001

Full size image
Table 3 Summary table of type III Wald chi-square tests for comparisons of richness and abundance of mobile macro-invertebrates and of the most abundant species between rockpools with and without pits or overhangs
Full size table

The total abundance (number of organisms) of all mobile macro-invertebrates only changed with the presence of overhangs, but not with the season (Table 3). Specifically, rock pools with overhangs had approximately double the abundance of mobile macro-invertebrates than those without overhangs (Fig. 3c, Table 3).

When investigating the three most abundant species separately (Nerita melanotragus, Bembicium nanum, Austrolittorina unifasciata), analyses showed that abundances of the grazing gastropod N. melanotragus increased with the presence of overhangs in rock pools, but also not with the season (Fig. 3d, Table 3). Nerita was approximately six times more abundant in rock pools with overhangs than those without overhangs (Fig. 3d, Table 3). Neither the abundance of B. nanum nor A. unifasciata varied with the presence of overhangs or pits or with the season (Table 3).

Discussion

It is widely accepted that habitat complexity is an important driver of biodiversity, but the influence of complexity is often not measured at a scale that is relevant to the organisms being studied (Matias et al. 2010). We found that “microhabitats” (pits or overhangs) within intertidal rock pools—which are often considered a microhabitat themselves—can affect the richness of mobile macro-invertebrates in these habitats. Additionally, the importance of microhabitats on abundances of mobile taxa in rock pools varied with the type of microhabitat present (i.e., “pit” or “overhang”), and was often species-specific. These results suggest that the relationship between diversity and habitat structure may be complex and that, on intertidal rocky shores, microhabitats within rock pools could be equally or more important for biodiversity than the size and shape of the rock pools themselves. However, there was a lack of microhabitats within the two zones within the harbour, which prevented the assessment of the generality of the results. This study adds to a growing body of evidence that complexity on the scale of centimetres can influence intertidal communities (Strain et al. 2021; Clifton et al. 2022), and can vary from positive to neutral based on the identity of species (Strain et al. 2021) and across different spatial scales.

Rock pools with pits or overhangs supported more species than those without them, and rock pools with overhangs supported higher abundances of mobile taxa, which was likely driven by the grazing gastropod N. atramentosa. Additionally, several species, including the whelk T. marginalba and the chiton S. pelliserpentis, occurred more frequently in rock pools with overhangs. Chitons are posited to have specific requirements for habitat at a small spatial scale (Grayson and Chapman 2004). On artificial structures, chitons were found to be more abundant in crevices than on exposed surfaces (Moreira et al. 2007). Additionally, on natural rocky shores, whelks move away from areas without shelter towards areas with shelters (Fairweather 1988). These results confirm that specific, tailored requirements of future management interventions, e.g., eco-engineering, in intertidal coastal areas, could be extremely beneficial to these organisms, which are important components of rocky shore assemblages.

Microhabitats can reduce environmental disturbances by facilitating the development of a range of microclimates (Sebens 1991), with the potential to buffer even extreme environmental conditions (Scheffers et al. 2014). Besides providing refuge from environmental stressors, microhabitats may also provide protection from visual predators such as birds (Cantin et al. 1974). The aggregation of Nerita under overhangs could be linked to the predator refuge and thermal buffer these habitats can provide. Overhangs also provide shade and thus shelter from direct sunlight, which can reduce heat stress. Gastropods have been found to move to habitats of reduced heat and desiccation stress when not foraging (Garrity 1984; Fairweather 1988; Williams and Morritt 1995; Moreira et al. 2007). In the present study, overhangs were either submerged within the rock pool water or emerged along the interface between pools and emerged rock. However, studies done on only emergent intertidal areas found similar relationships between species distribution and microhabitats. For example, the gastropods Nerita funiculata, Nerita sabricosta and Tenguella marginalba aggregate in crevices (Levings and Garrity 1983; Fairweather 1988), suggesting therefore that it is the overhang structure itself that provides the sought-after shelter rather than the submerged nature of the habitat. The effect of overhangs did not, however, differ among seasons, suggesting that other factors, such as shelter from predators, may have been more important than shelter from heat stress.

Pits increased the richness of mobile taxa, which may be relayed to differences in water temperatures. Water in rock pools can vary locally, with deeper parts and rock depressions to be of different temperatures than water closer to the surface (Morris and Taylor 1983). Surprisingly, there was no effect of the presence of pits on mobile abundances, which may be related to richness being a more sensitive measure than abundance. In our study, only a single rock pool had pits deeper than 10 cm, and thus the relationship between differently sized pits on the distribution of mobiles could not be tested. This may have hidden potential effects of pits on abundances.

Due to an unbalanced number of rock pools with and without pits or overhangs and their difference in sizes, we could not differentiate between all microhabitats, such as overhangs above and underwater, or pits of different depths. While rock pools with pits or overhangs were slightly bigger than those without pits or overhangs, comparisons of taxon richness across all rock pools without these microhabitats showed that taxon richness did not increase with increasing volume (Fig. S1.3). Therefore, differences in volume should have had limited effects on the patterns found. Sessile cover in rockpools was not assessed in this study and may have affected mobile abundances. To control for these differences, distinguish the effects of specific microhabitats (e.g., emergent and submerged overhangs), and establish cause-effect relationships, manipulative studies are necessary, such as testing artificial rock pools in which these features are incorporated, tested independently of surface area and from another, and sessile cover is absent (see Loke and Todd 2016 for an example).

The low number of rock pools without microhabitats at the inner and outer zones of the harbour prevented the testing of the generality of the patterns found in the coastal zone. The lack of microhabitats in the harbour zones may be a function of the environment itself, as it is a more protected environment with lower wave action than coastal sites, which may have limited the erosion of rock and thus the development of microhabitats. This may suggest that diversity within pools is driven by other factors, not considered here, rather than microhabitats. Alternatively, the low numbers of rock pools with overhangs or pits observed in the harbour zones may partly explain the overall low number of species of mobile taxa in rock pools at these sites (Schaefer et al. 2019), since there is compelling evidence that microhabitats increase niche availability (Menge and Sutherland 1976), which generally positively affects diversity.

Conclusion and implications

We showed that the relationship between diversity and microhabitats (here: pits and overhangs) within rock pools is complex and varies with the species as well as the type of microhabitat present. Findings from this study highlight the importance of studying assemblages and potential drivers of biodiversity at relevant spatial scales and that such issues should be further investigated with manipulative experiments. We found a greater richness of mobile taxa and a greater abundance of mobile macro-invertebrates (driven by Nerita melanotragus) and a frequency of occurrence of certain taxa (whelks and chitons) in rock pools with overhangs. Rocky reefs across the globe have become increasingly fragmented and replaced by homogeneous built artificial structures, particularly in urban environments (Goodsell et al. 2007; Dafforn et al. 2015). Recent research suggests that the provision of microhabitats can benefit associated communities (for a review see Strain et al. 2018), but we need to understand the habitat requirements of the relevant species if we want to successfully manage, conserve and restore disturbed habitats. Although attaching artificial rock pools onto artificial habitats, such as seawalls, has proven to be a successful way to increase the diversity of these habitats (Chapman and Blockley 2009; Browne and Chapman 2014), our findings suggest that pits and overhangs within rock pools play an important role in supporting biodiversity. Advances in technologies, such as 3D laser scanning and 3D printing, enable engineers to design and build highly structured and more sophisticated artificial rock pools informed by ecologists to provide more targeted outcomes. Where the objective is to increase the abundances of species diversity or abundance, rock pools should be constructed with overhangs and pits. However, further experimental testing using rock pools of similar size, with and without pits or overhangs and with different combinations of pits and overhangs, is needed to test the generality of our findings at multiple zones.