Hidden Consumers in Wave-exposed Boulder Beaches: Implications to Trophic Studies in Marine-terrestrial Ecotones

Invertebrate communities and ecological processes are well understood in high-energy sandy beaches, where beach-cast wrack (macrophyte detritus) often accumulates and forms hotspots of nutrient cycling as well as enhancing diversity and driving food webs. Grazing invertebrates play a key role in recycling wrack and facilitating the transfer of nutrients for spatial subsidies across this marine-terrestrial ecotone. Cobble and boulder beaches can also form a prominent feature of wave-exposed coasts and accumulate wrack, yet we know far less about the invertebrates in these beaches and their possible role for recycling wrack. Here, we determine the biomass of detrital macrophytes on, and embedded in, the boulder matrix, as well as the density and biomass of macroinvertebrates in high-energy, boulder beaches in south-eastern New Zealand. We use these data to compare densities and biomass of wrack and invertebrates with published data for sandy beaches to examine the possible importance of these understudied coastal habitats in recycling wrack and facilitating trophic subsidies. The biomass of beach-cast macroalgae exceeded 100 g DW 0.1m -2 on boulder beaches, where the kelps Durvillaea spp. and Macrocystis pyrifera were the main forms of wrack on both types of beaches. This was comparable to many other sandy beaches across the globe. However, the total biomass of invertebrates on boulder beaches in our study was higher than that reported for sandy beaches in the region and across the globe, while densities were similar or higher than those found on sandy beaches. Like sandy beaches across the globe, amphipods were abundant on boulder beaches, however, the relatively large gastropod Diloma nigerrimum was particularly dominant in this habitat. With its known high grazing rates, combined with high densities and biomass, this grazer is likely to play a disproportionately important role in the processing of beach-cast kelp and a key role in transferring nutrients back into the ocean or to adjacent terrestrial food webs. from of kelp. treatment included in the if of beach-cast kelp represented the of kelp detritus on the beach. Large cobbles or boulders (>10 cm were then individually removed from the quadrat, and biological material was picked from each boulder and transferred to a plastic bag. and all organic material were removed and transferred to the plastic bag until a base of rm sand was reached (typically ~ 10-30 cm depth).


Introduction
Shoreline ecotones provide a transition zone between land and sea, where ecological functions support biodiversity and productivity in both terrestrial and marine environments. Organic matter is often deposited in large quantities along shorelines (e.g. Gri ths et al. 1983; Orr et al. 2005;Mateo, 2010), where it can enhance biodiversity and secondary production through the provision of nutrients and food resources in this marine-terrestrial ecotone (Polis et al. 1997), as well as return bioavailable nutrients to adjacent marine waters (Colombini et al. 2002). Our understanding of the ecological functions of beach ecotones is mainly based on sandy-beach systems (see Schlacher et al. 2008). The deposition of macroalgal and/or seagrass detritus (known as "wrack") in these systems can support a high diversity and density of invertebrates comprising a mixture of marine and terrestrial species (e.g. Colombini Crawley and Hyndes 2007). These grazers also often exhibit a preference for kelp (species from the orders Laminariales and Fucales) as a food source (Pennings et al. 2000; Dugan et al. 2003;Crawley and Hyndes 2007), indicating that kelp in particular plays a pivotal role in driving secondary production in this ecotone. These consumers, along with insects such as ies and beetles, provide important ecological roles through breaking down detritus for nutrient recycling and forming vectors for the transfer of organic material to adjacent ecosystems (Hyndes et al. 2014).
The complex three-dimensional structure provided by cobble or boulder habitats has been shown to contain highly diverse and abundant fauna (Chapman 2002(Chapman , 2005 Grzelak and Kuklinski 2010). However, the focus on these habitats has been on broad assemblage structure in relatively wave-sheltered intertidal or subtidal environments, rather than wave-exposed shoreline habitats where wrack accumulates. Disturbance through wave energy and sand deposition reduces the biodiversity of boulder elds (McGuinness 1987a, b), suggesting that cobbles and boulders on high-energy beaches may contain a lower diversity of invertebrates than similar habitats in more sheltered environments. However, the high retention of detrital kelp within the structure of cobble beaches (Orr et al. 2005) could promote a greater diversity and density/biomass of grazers, and therefore form an important vector for the transfer of nutrients derived from kelp to terrestrial or adjacent marine ecosystems. Our understanding of the ecology of these systems is poor, despite their presence across most, if not all, continents (Oak 1984 In this study, we aimed to determine the biomass of beach-cast wrack and the densities and biomass of macroinvertebrates in cobble/boulder beaches in a high wave-energy environment where kelp detritus is known to accumulate, and compare the ndings to published studies of sandy beaches in similar settings. We examined this in the coastal region of southern New Zealand, where boulder beaches are a prominent feature of the high wave-energy coastline (Morton 2004). We focused on wrack and macroinvertebrates in areas of the beach where wrack is visible and absent to determine whether material is bound in the boulder matrix. We then discuss our results in the context of existing published data from the region and elsewhere across the globe for both sandy and cobble/boulder beaches.
Since detritivores (often termed "mesograzers" in beach systems) dominate the macroinvertebrates (see above), we have focused on this functional group. This has allowed us to discuss the role of mesograzers/detritivores in processing beach-cast wrack and forming vectors for the transfer of nutrients from stranded drift material to other ecosystems.

Methods
Our study was undertaken along the wave-exposed coastline of north Otago in south-eastern New Zealand, where cobble or boulder beaches are common. The kelp Macrocystis pyrifera [Laminariales] and Durvillaea spp.
[Fucales] form a major component of the wrack along sandy-beach shorelines (Suarez et al. 2015) and detrital wrack has been observed on cobble or boulder beaches in the region. Furthermore, a range of gastropod species has been observed in cobble or boulder beaches, but they have not been quanti ed (Morton, 2004). Three beach sites comprising a mixture of cobble and/or boulders ( Figure 1) were selected for this study (Seconds 45° 54'54.91"S, 170° 29'13.23"E; Puketeraki 45° 39'17.72"S, 170° 39'10.95"E; Boulder 45° 38'23.61"S, 170° 40'18.89"E). We use the term "boulder beaches" to simplify the text. All sites had a predominantly easterly aspect and were therefore exposed to ocean swells tending SE to NE, with a mean swell height ranging between 0.64 and 1.03 m in October and 0.55 and 1.21 m in November 2010 (SOM Table 1). Maximum swell height ranged from 3.4 and 4.5 m in October 2010 and 1.4 to 1.7 m in November 2010. The site "Seconds" tended to have higher mean wave height than the other sites (SOM Table 1).
Sampling was conducted from the 16 th to 19 th of November, 2010, on beaches that comprised beach-cast wrack that was dominated by kelp. At each site, 0.1 m 2 quadrats were positioned haphazardly along the mean high-water mark, with ve quadrats positioned in patches where large kelp (Durvillaea spp. and/or Macrocystis pyrifera) were visibly stranded, and ve quadrats positioned in areas visibly free of kelp wrack that were at least 5 m away from patches of kelp. "Kelp" treatment was included in the design to determine if the visible presence of beach-cast kelp represented the amount of kelp detritus on the beach. Large cobbles or boulders (>10 cm diameter) were then individually removed from the quadrat, and biological material was picked from each boulder and transferred to a plastic bag. Gravel and all organic material were removed and transferred to the plastic bag until a base of rm sand was reached (typically ~ 10-30 cm depth).
All samples were transported to the laboratory where they were stored frozen (-20 °C) until processing.
In the laboratory, samples were thawed then individually picked by hand to separate macroinvertebrates from sediment and wrack. Once fauna and ora were identi ed to the lowest possible taxonomic level, individuals of each faunal taxon were counted, and then blotted wet mass for samples of each taxon of fauna and ora were measured. Samples were then dried at 60°C until a constant weight was reached, which was recorded. Since juvenile amphipods were numerous and di cult to identify, the total number and biomass of amphipods were recorded and the proportion of species within Hyalidae and Talitridae was estimated.
Total densities and biomass of grazers, as well as densities and biomass of the most dominant taxa, were tested across kelp treatment (2 levels; xed factor) and sites (3 levels; random factor). Data were tested for homogeneity of variance using Levene's test, and normality using Shapriro-Wilk test. Where homogeneity or normality did not result, data were log-transformed.
To compare biomass of kelp and biomass and density of grazers/detritivores in our boulder beaches to those in other similar beaches and sandy beaches, we examined and extracted published data based on searches in Clarivate Web of Knowledge and Elsevier Scopus using the terms ("beach*" OR "shore") AND ("wrack" OR "beach-cast"OR "allochthonous"). For consistency with the metrics used in the current study, data extraction was restricted to those studies in which data were either provided or could be converted to biomass and/or abundances per m 2 . Since most studies provided wrack biomass as wet weight, we converted our dry weight data to wet weight based on dry weight to wet weight ratios measured during the study.

Results
Where wrack was visible at the surface, the kelps Durvillaea spp. were either the only or dominant form of beach-cast material, contributing 78-100% of the wrack ( Table 1). The kelp Macrocystis pyrifera was the next most dominant form of wrack, contributing 0-11% to the biomass of wrack. The remaining wrack mainly comprised other brown algae, along with some red algae (Table 1). In comparison, Durvillaea spp. were almost absent from areas where wrack was absent on the surface, while M. pyrifera and other brown algae were present but in smaller amounts. Red algae were generally present in small quantities in areas where wrack was visible and absent on the surface (Table 1), and was typically found at the base of the boulders. Total biomass of beach-cast wrack was approximately 100 g DW 0.1 m -2 when beach-cast wrack was visibly present in the surface, and 1-2 orders of magnitude less when wrack was not visible on the surface (0.2-28 g DW 0.1 m -2 ; Figure 2A).
The macroinvertebrates found in the boulder beaches were dominated by grazers or detritivores (Table 3). Focusing on grazers/detritivores (here on termed grazers), total grazer density and biomass differed when macroalgae was visibly present or absent on the beach surface (p<0.05, Table 2). However, mean densities and biomass exceeded 100 ind. and 100 g DW 0.1 m -2 , respectively, regardless of whether or not wrack was visible on the beach surface ( Figure 2B, E). In patches where wrack was visible, total grazer densities exceeded 1,000 ind. 0.1 m -2 while biomass was around 1,000 g DW 0.1 m -2 .
Amphipods were the most numerous macroinvertebrate present in the boulder beaches (Table 3), and were typically found at the base of the boulders. Approximately 80% belonged to the family Hyalidae. The trochid gastropod Diloma nigerrimum was also highly abundant and was present in high biomass, while the ellobiid gastropod Marinula lholi was also relatively abundant. The densities and biomass of amphipods did not differ between areas with the visible presence or absence of wrack ( Table 2). The mean density of amphipods ranged from 31 to 6,343 ind. 0.1 m -2 , while biomass ranged from 0.07 to 13.6 g DW 0.1 m -2 ( Figure 2C, F). In comparison, the densities and biomass of D. nigerrimum differed when beach-cast macroalgae was visibly present or absence, with mean densities and biomass ranging between 1,037 and 9,240 ind. 0.1 m -2 , and 688 and 2,721 g DW 0.1 m -2 , respectively, where wrack was visible, and ranging between 192 and 3,137 ind. 0.1 m -2 , and 119 and 1,372 g DW 0.1 m -2 , respectively, where wrack was visibly absent (Figure 2D, G). Since the soft tissue to total biomass ratio for this gastropod was 0.95, biomass of soft tissue ranged from 113 to 2,585 g DW 0.1 m -2 .

Discussion
Our ecological understanding of wave-exposed beaches is typically based on sandy beaches ( is surprising that our understanding of these beaches remains poorly understood in the framework of western science. The high amount of drift kelps that become stranded in cobble or boulder beaches (current study; Orr et al. 2005) suggests that the grazing assemblages in this habitat are likely to play a major role in processing beach-cast kelp and transferring its nutrients to terrestrial systems or returning bioavailable nutrients to marine ecosystems. While we recognize that our study has limited temporal coverage, we argue that this shoreline ecotone appears to provide a hotspot of nutrient recycling via grazers or detrivores and requires far greater research effort to understand its role in these processes.
The combined densities of mesograzers in boulder beaches in our study exceeded those reported for mesograzers on sandy beaches in the region, and far exceeded densities of mesograzers reported on sandy beaches elsewhere across the globe (Table 4). Supporting our results, amphipod densities are also high in gravel/cobble beaches in north-eastern USA (Behbehani and Croker 1982). However, the full extent of the value of boulder beaches as a habitat for mesograzers is demonstrated by their biomass, which was far higher on boulder beaches sampled in this study, compared to those reported for sandy beaches in the region (>10kg vs 12 g DW m -2 ; Table 4). This largely re ects the presence of the gastropod D. nigerrimum, which is likely to bene t from the greater heterogeneity of the boulder beaches through larger interstitial spaces and presence of hard substrata (Le Hir and Hily 2005), and the high biomass of stranded kelp, its preferred food source (Quintanilla-Ahumada et al. 2018).
The high biomass of particularly kelp on the boulder beach surface would re ect the sporadic inputs of this detrital material over the tidal and storm cycles. While our study was restricted to a short timeframe, it coincided with a period of high biomass on some nearby sandy beaches (Suárez et al. 2015). Unlike sandy beaches, the greater interstitial spaces among the larger grain size of cobbles and boulders are more effective at trapping and retaining large wrack particles (Behbehani and Croker 1982;Orr et al. 2005). This is supported by studies examining the long-term retention of oil on beaches exposed to oil spills from the crash of the Exxon Valdez in 1989. Monitoring of boulder beaches affected by the spill found that oil was surprisingly well retained in the matrix of the beaches despite the intermittently high-energy nature of the environment and apparent mobility of boulders and gravels (Hayes and Michel 1999;Irvine et al. 2006). This pattern of retention was explained by surface boulders acting to "armour" underlying sediment from wave forces, thus leading to a high level of oil retention in the sediment matrix. This "boundary layer" would allow detritus particles to be retained, and at least partly explain the presence of kelp and other macroalgae within the cobble/boulder matrix, as shown by the presence of macroalgae in areas where wrack was absent at the surface. This would allow detritus to be retained for longer periods than on high-energy, sandy beaches, but further research is required to con rm this. Greater retention, and the movement of boulders during strong surges (pers. obs.), could enhance the physical and microbial breakdown of macroalgal particles within the boulder matrix, and facilitate consumption by grazers. . This difference partly re ects the increase in feeding rates of marine mesograzers with increasing body size (Ruesink, 2000). The biomass we observed for D. nigerrimum in the boulder beaches was three orders of magnitude greater than those recorded for B. quoyana on sandy beaches in the region (see Table 4). Even at maximum densities for B. quoyana on sandy beaches, the biomass we recorded for D. nigerrimum in boulder beaches far exceed those of amphipods on sandy beaches (see Table 4). The difference could be even higher if amphipods were included, but their role in consuming kelp is unknown. It is worth noting that a high proportion of the amphipods found in the boulders belonged to Hyalidae. Since these amphipods have been shown to display preferences for green or red algae (Taylor and Steinberg 2005), and they occurred at the base of the boulder matrix where red algae accumulated (Pers. obs), it is possible that they were consuming red algae.
Our study highlights the need to consider cobble and boulder beaches when determining the capacity of beaches to process beach-cast macrophytes. The high biomass of mesograzers in wave-exposed boulder beaches in our study is consistent with the results of a metaanalysis by Ricciardi and Bourget (1999), who showed that boulder habitats contain 10-100 times the biomass of macroinvertebrates than sandy and rocky shore habitats in sheltered environments. Our results suggest that the differences may be even greater between boulder and sandy beaches in wave-exposed coasts. Macroalgae can be cast in large quantities onto cobble and boulder beaches on wave-exposed coasts in others regions (Orr et al. 2005  Invertebrates inhabiting wave-exposed cobble or boulder beaches tend to be inconspicuous, which may partly explain the paucity of studies on those communities, but our study suggests they could play a disproportionate role in the processing of beach-cast macroalgae compared to sandy beaches, and are likely to play an important role in the transfer of that material through marine (Crawley et al. 2009) and terrestrial (Polis and Hurd 1996;Mellbrand et al. 2011) food webs as well returning nutrients back into the ocean. A broader and more comprehensive understanding of their role is particularly relevant given their susceptibility to long-term impacts from oil spills (Irvine et al. 2006) and the vulnerability of shoreline systems to coastal armouring ) and rising sea levels (Vitousek et al. 2017