Intertidal fishes of Mauritius with special reference to shallow tidepools

Abstract Intertidal fishes are found in large numbers and play an important role in their ecosystems, but knowledge of their ecology is still very limited in many tropical regions. Within this context, data from intertidal fishes in Mauritius were compiled from different sources and intertidal resident species were examined in Mauritian tidepools. A total of 292 fish species occurring in Mauritius were reported from intertidal habitats, of which 62 species represent permanent intertidal residents. The species number in the studied pools increased, not only with the proportion of stones and rock covering the pool bottom, but also with pool facilities, for example, the supply of boulders and a high coverage of macro-algae. All examined pools were dominated by two species, Bathygobius coalitus and Istiblennius edentulus. Their abundance increased with decreasing pool size, peaking in pools with a surface area between 1-2 m2 during the lowest level of ebb tide. This 'overcrowding effect' may be linked to the absence of predators in these very small pools. The comparison of present data with results of a survey made in the same area in 1995 suggested a decrease of resident species occurred during the last decades, probably linked to human influences, such as eutrophication and water pollution.


Introduction
Intertidal species occupy the narrow band of near-shore habitats between the tidemarks of seas and oceans (Horn et al. 1999). Fishes occurring in the intertidal zone represent numerous different families and constitute a rather inhomogeneous group. Some of these fishes visit the near-shore habitats only briefly, mainly to forage. Others inhabit the intertidal zone during part of their life cycle, for example, the juvenile phase, whereas a third group of fishes, having highly specific adaptations in terms of behaviour and physiology, spends their entire life in intertidal habitats (Horn et al. 1999, Griffiths 2003a.
In a first overview of intertidal fishes,  recorded 702 species representing 110 genera in one chondrichthyan and 20 teleost families. This large number of species is based mainly on the North Atlantic, North-eastern Pacific, South African and New Zealand regions. Chotkowski et al.'s review included only one study each from the Western Central Atlantic, Indian Ocean, subtropical Australia, Central Pacific and Southeastern Pacific and three studies from the tropical Eastern Pacific. The level of knowledge about intertidal fish communities in many tropical regions, for example, in the tropical Eastern Atlantic and Western Indian Ocean, has hardly changed since then. Possible reasons for the limited knowledge are the low importance of intertidal fishes for fisheries, their occurrence in extremely shallow waters, their small size and their sometimes secluded way of live. The species in question are rarely recognised by divers and hardly considered by shore visitors.
On the other hand, intertidal fishes are found in large numbers of species and may reach high abundances. Based on their quantitative occurrence, it can be inferred that they make an important contribution to the intertidal food web by functioning in their roles as herbivores or predators. Though quantitative analyses of food webs including intertidal fishes are still scarce, we know that the diversity of herbivores is higher in tropical zones than in temperate regions and that intertidal fishes may have a strong influence on algal diversity and algal abundance in temperate as well as tropical areas (see Horn and Ojeda 1999 for a summary). For example, herbivorous fishes accounted for 20-30% of the fish communities in temperate rocky reefs of Australia (Jones and Andrew 1990) and 20% of total intertidal fish abundance (or 51% of total fish biomass) in temperate Chile (Stepien 1990), while 81.6% of the fish biomass at a subtropical intertidal reef in South Africa consisted of herbivorous fishes and omnivorous fishes with a notable herbivorous portion of their diet (Berry et al. 1982). Thus, intertidal fishes play an important role in material turnover in their ecosystems (Bennett et al. 1983). More detailed knowledge on these communities would provide a deeper insight into the ecology of marine shallow water environments. Since tidepools, estuaries and other intertidal habitats are located nearshore, they are vulnerable to human impacts such as eutrophication, pollution, disturbance or destruction. Therefore, better knowledge and monitoring of intertidal fishes would open new opportunities in applied ecology, for example, by using these fishes as ecological indicators for environmental pollution and climate change.
Against this background, one aim of the present study is to make a first compilation of data on fish species occurring in the intertidal zone in Mauritius. A second aim is the examination of intertidal resident species in Mauritian tidepools. The results of this study will increase our knowledge on intertidal communities in the Western Indian Ocean and could inspire further research towards use of intertidal species for monitoring and applied approaches in environmental research.

Study area
The island of Mauritius is located in the Western Indian Ocean (Fig. 1), 800 km east of Madagascar between 19.58°S and 20.31°S, as well as between 57.18°E and 57.46°E. Together with Réunion, Rodrigues and several smaller islands, it forms the Mascarene islands. Mauritius has a coastline of 322 km (Duvat 2009) and is surrounded by a submarine platform, allowing the development of a fringing coral reef with a broad lagoon that protects much of the coastline in the north, east, south-west and north-west (Pichon 1971). The fringing reef is disrupted by rocky shores and cliffs in the south and west. Sea surface temperatures vary seasonally, with a minimum of 22°C in August and reaching up to 29°C in February. The dominant current patterns are from east to west, influenced by the south-eastern trade winds, which are stronger during the winter months than during the summer (McClanahan et al. 2005, Ragoonaden 1997. Tides occur semi-diurnally, with a mean tidal range of 0.9 m during springtime and 0.1 m during neap time (Montaggioni and Faure 1997). estimate their ground cover. One of these pools was isolated except during the highest point of flood and two were continuously streamed through by seawater.
The five pools with highest diversity were all located in the vicinity of Albion (Suppl. material 2).

Pointe aux Caves, below lighthouse (LH).
Pointe aux Caves is characterised by a high rock cliff with two platforms lying one metre above the mean water level (Bhikajee 1996), i.e. a few decimetres above the middle high water level. The northern platform contains three rockpools, two of which were examined (Fig. 2b). The pool surfaces were 28 and 30 m , respectively, the depth 100 and 60 cm. Green macroalgae covered about 30% of the bottom surface in the extended shallow parts of both pools. Due to the high platform, the pools did not have any seawater exchange or connection to the water in the open sea, except at the highest level of flood during spring tides. Nevertheless, a high number of fish (permanent pool residents) inhabited these pools.

Blue Bay (BB).
A triangular area of flat rocks, consisting of lava tuff, extends at the southern end of Blue Bay, close to the 'Le Peninsula Bay' Beach Resort and Spa. Depending on seasonal water level, the lower parts of this rock flat constitute a number of tidepools. Seven of these tidepools were examined. The surface area of these pools 2 Figure 2.
Examples of examined tidepools in Mauritius. a -ALD6 in Albion, exposed to the open sea; b -LH1, located on a rock platfom below the lighthouse at Pointe aux Caves; c -BB1 in Blue Bay, located inside the lagoon; d -PE1 near Péreybère, located inside the lagoon. ranged between 1-15 m during the lowest level of ebb tide and 4-36 m during flood tide. The maximum depth ranged between 10-25 cm at lowest level of ebb tide. The bottom surface was mostly characterised by sand (50-80% bottom cover) and stones (10-40% bottom cover), while living brown algae occurred in four of the pools (usually with 10-15% bottom cover, but covering 50% of the bottom surface in one of the smaller pools). The largest and westernmost pool is shown in Fig. 2c.
Péreybère (PE). Several shallow tidepools are located 1000 m north-east of Péreybère public beach. Three pools were examined; these had a surface area of 36-54 m and a maximum depth of 15-25 cm during the lowest level of ebb tide. The bottom was covered with sand (70-90% of the bottom surface) and stones and a notable amount of dead brown algae floated in two of the pools. The pools were isolated for less than 2-6 hours during a single tidal cycle. The largest and westernmost pool is shown in Fig. 2d.

Compilation of data
In order to get a complete list of intertidal fish species in Mauritius, all available literature was searched, extracting data on fishes recorded from intertidal habitats and their localities. Besides the published records from Mauritius (Fricke 1999), data from Réunion (Fricke et al. 2009), Europa Island (Fricke et al. 2013, Glorieuses Islands ) and several further places were also considered for those species that are known from Mauritius but for which published intertidal records from Mauritius were lacking so far. Additionally, material from the following museum collections was included: USNM, Smithsonian Collection, Washington (records by P.C. Heemstra, A.C. Gill, D.G. Smith and M.J. Smale from April/May 1995, see Fricke 1999), SMNS (Staatliches Museum für Naturkunde, Stuttgart, Germany) and BMNH (The Natural History Museum, London). The complete list of species and sources is given in the Results chapter. Family authorships were cited according to van der Laan et al. (2014); genus and species classification follows Fricke et al. (2019).
Particular attention was given to fishes reported from tidal pools, because pools are welldefined intertidal habitats that isolate their communities at least during ebb tides, making observations and examination of fishes relatively easy. In contrast, the border between the tidal and subtidal zones appears to be a smooth transition that is difficult to identify in many other near-shore habitats such as sand flats, rock cliffs or lagoons. Tidepools were also the subject of field examinations in the scope of the present work.
Fishes may occur in the intertidal zone for quite different reasons or in various periods of their life. Therefore, it is necessary to classify different 'types' of intertidal fishes. The definition of these terms follows Thomson and Lehner (1976) and Griffiths (2003a) R -permanent residents, spending their entire life (juvenile to adult) within tidepools or the intertidal zone. They are often highly adapted for intertidal life by possessing specialised behavioural or physiological adaptations. O -opportunists (also secondary or temporary or partial residents), living in the intertidal zone / tidepools during specific life history stages or seasons. They are also widely distributed in the subtidal zone.
T -transients (or tidal visitors), using the intertidal zone including tidepools transiently for foraging. They may end up accidentally trapped in pools as the tide goes out.
The assignment to one of these categories was established on the basis of literature references such as Murase (2013) and Sindorf et al. (2015), our own observations and information about size as well as developmental stage of the examined fishes in different studies, such as Ntiba et al. (1993), Durville and Chabanet (2009), Ghanbarifardi and Malek (2009) and González-Murcia et al. (2016). It is worth mentioning that there are no generally used abbreviations of the aforementioned categories in the existing literature, for example, PR means partial resident in Cox et al. (2011), but permanent resident in Sindorf et al. (2015).

Field examination of fishes in rockpools
The fishes of selected tidepools (see description above) were examined in September/ October 2017 and 2018 by underwater visual census, i.e. fishes in each pool were observed and photographed. The main objectives during field work in 2017 were the selection of pool sites and the identification of fish species in these pools. Individuals were not counted during the field work in 2017. In 2018, the pools were examined systematically, i.e. fishes in each pool were observed and registered both near to the lowest level of ebb tide and highest level of flood by day. One additional examination took place in each pool by night. Photos for later analyses were taken of any fish for which a definite field identification was not possible. During the surveys at low tide-level, the number of permanent resident species was counted to estimate their abundance (cf. Gibson 1999). Counting as well as taking photos was feasible, because the pools were shallow and not too large. The permanent resident species showed territorial behaviour or used a limited area of their pools only, all of them being diurnal. Despite a detailed search, seclusive or nocturnal species were not found. During night, the permanent resident species rested motionless on the pool bottom or in small crevices and allowed close-up shots of morphological details. The species could be identified on the basis of photographs taken by day and night. Other species, representing secondary residents or transients, were simply photographed and registered without counting.

Statistical analyses
Besides pool resident and transient fishes, several pool parameters were recorded in order to assess whether pools with particular characteristics support a higher number of fish species or fish individuals. The pool parameters are:

Data resources
All primary data collected for this study are available as supplementary files. Suppl. material 1 contains parameters and characteristics of the examined tidepools; suppl. material 2 contains the recorded permament intertidal resident fish species in Mauritian shallow tidepools.

Intertidal permanent residents in examined tidepools
Eight permanent residents were recorded from the 23 tidepools examined in September/ October 2018 ( Bathygobius coalitus and Istiblennius edentulus were the most widespread species in the tidepools. The abundance of B. coalitus decreased with increasing depth of the pools (p = 0.04) and with increasing pool surface (p < 0.001). B. coalitus occupied very shallow areas of pools, often closely crowded during lowest level of ebb tide. A similar 'overcrowdingeffect' was seen in I. edentulus with a significantly negative correlation between its abundance and pool surface area during ebb tide (p < 0.001). The influence of these parameters on the occurrence of B. coalitus and I. edentulus was also confirmed by the RDA. Furthermore, the ordination showed a close relationship between the abundance of I. edentulus and the coverage of algae on one side and a relationship between B. coalitus and pools with a long period of isolation during the tidal cycle on the other side (Fig. 3). The second RDA calculation, including the rockpools around Albion and Pointe aux Caves, revealed that the surface area and bottom structures of pools, as well as the duration of isolation from in-and out-streaming water during the tidal cycle, may be important parameters that influence the settlement of species (Fig. 4). The starry moray (Echidna nebulosa) and the ebony gregory (Stegastes limbatus) occur in large pools with a supply of boulders. Species of Istiblennius are abundant in pools with high coverage of macro-algae. Whether some of these macro-algae serve as food for these fishes could not be verified. Rather, they were observed feeding on tiny algae, the coverage of which was not estimated because these algae were too small.    Redundancy analysis (RDA) of the intertidal residents in the examined rockpools at the west coast of Mauritius (AL, LH). Axes 1 and 2 explain 50.0% of the variance of species data and 97.9% that of environmental data. A species fit was set by 15% (meaning that the three species with lowest explanatory value are not shown).

Discussion
Besides the present study, a few further publications specifically dealing with intertidal fishes in the Western Indian Ocean do exist. Durville and Chabanet (2009) found a total of 32 different fish species in intertidal rockpools on the Glorieuses Islands. They denominated 19 of these species as 'typical population' also occurring during the adult stage in these habitats, whereas the remaining 13 species were observed during their juvenile stage only. The examined pools were located high in the infra-littoral zone, thus more or less isolated during long periods of the tidal cycle and they had a surface area of approximately 2 m during ebb tide. For these reasons, the results by Durville and Chabanet (2009) (1986) and Burger (1990) confirm that eel blennies may represent permanent intertidal residents and may occur in tidepools in large numbers. Sindorf et al. (2015)recorded 55 species during an examination of intertidal fishes in Watramu Marine National Park (Kenya), 21 of which were permanent residents. Twentyfive of these species also occur in Mauritian tidepools; however, we assigned several of them to different resident categories (see remarks in Table 1). Five wrasses and seven damselfishes coincided in Kenyan and Mauritian tidepools, but not a single permanent resident of either gobies or blennies was found in both studies(?).
Two tidepool studies took place in the northern Indian Ocean. Ghanbarifardi and Malek (2009)  In the present field study, the total number of permanent intertidal residents was low compared with the results of the field trip by Heemstra and co-workers in the year 1995 (in Fricke 1999), Durville and Chabanet (2009) or Sindorf et al. (2015). This low number of residents could be due to the low depth and small size of most of the examined pools. In particular, Heemstra et al. (in Fricke 1999) recorded a larger number of permanent residents in tidepools in Albion and at the lighthouse in Pointe aux Caves, including two longfins of the genus Plesiops, the gobies Eviota prasina and Hetereleotris zonata and the triplefin Enneapterygius philippinus. Some of these pools were re-examined during the present study, but neither the longfins, nor the triplefins or the mentioned genera of gobies were confirmed, even though these species are distinct and not particularly cryptic. It seems likely that the apparent decrease in intertidal species around Albion between 1995 and 2018 has fundamental reasons, for example, environmental change. Eutrophication and industrial wastes including, in particular, metal pollution in the Port Louis area, effects of eutrophication due to a high input of nitrate probably of agricultural origin in the Flic en Flac coastal area and effects due to increasing tourism have been recorded at the Mauritian west coast next to Albion (Ramessur 2013). Seawater pollution had already caused a decline in coverage of live corals by 10-30% in coastal lagoons around Mauritius by 2012 (Ramessur 2013). It is conceivable that pollution affected the diversity of intertidal fish as well. We know that I. edentulus, one of the most abundant species in our field study, may tolerate poor water quality (P. Bourjon, verb. commun.).
A comparably low number of resident species does not necessarily mean low abundances of fish. On one hand, a number of opportunists and transients appeared especially in larger pools (cf. species in Table 1). On the other hand, the abundances of some resident species were not at all low. The total abundance of permanent residents averaged 5.31 per m (SD ± 2.1) in pools ≤ 6 m and 1.24 (SD ± 1.54) in pools ≥ 10 m , the maximum value being 9 individuals/m . Lundquist and Pinkerton (2008), who examined tidal pools in New Zealand, estimated an abundance of 10 fish/m across their intertidal study area, Bennett and Griffiths (1984) counted 7.42 individuals per m in South Africa, while González-Murcia et al. (2016) found mean total abundances of between 5 fish/m in high shore pools and 12 fish/m in pools at lower shore sections in El Salvador, but opportunist and transients were included in these studies.
It is not surprising that larger tidepools, containing more algal and rock ledge cover, host a larger and more diverse population of fish. Pool depth, volume and also the variety of microhabitats such as presence of shells, pebbles and rock ledges influence richness and total abundance of fish strongly (Mahon andMahon 1994, White et al. 2015). White et al.
(l.c.) examined very small pools with an area of 20 cm to 8 m in New South Wales (Australia) and included all 27 recorded fish species in their analysis. In the present study, we used another approach focusing on permanent residents. The species number increased with pool size as well, but did not yield a significant result due to the much lower number of species in question. However, the results show that specific pool microhabitats are associated with the occurrence and abundance of particular species. After exclusion of 2 2 2 2 2 2 2 2 2 2 fundamental oceanographic parameters, because lagoon or wave-exposed open sea sites influence stone and sand coverage of examined pools differently, facilities like coverage of algae and presence of boulders contributed mainly to the occurrence of certain permanent residents (cf. Fig. 4). The abundance of the two most common species, B. coalitus and I. edentulus, correlated negatively with the pool surface area (Fig. 3), that of B. coalitus also with the depth of the pools. As deep and large pools tend to contain predators (White et al. 2015), shallow and small pools with a long isolation period during the tidal cycle obviously may be beneficial for these permanent residents.
It is well known that a large number of intertidal fishes stay in intertidal habitats only temporarily, most of which use the shelter in these narrow and more or less isolated habitats as juveniles, but move to deeper water once they reach the adult stage. This can be observed in mangrove areas (Nagelkerken et al. 2000, Laegdsgaard and Johnson 2001, Ikejima et al. 2003, Mumby et al. 2004, Jaxion-Harm et al. 2012) and estuaries (Miller et al. 1985, Able 2005, Vasconcelos et al. 2008, Figueiredo and Pessanha 2016, in which vegetated habitats within estuaries tend to harbour higher densities of many fish species than unvegetated substrates (Sogard 1992).
Tidepools also offer shelter for juvenile temporary residents (Mahon and Mahon 1994, Gibson and Yoshiyama 1999, Ghanbarifardi and Malek 2009, Murase 2013, Sindorf et al. 2015, González-Murcia et al. 2016. In Mauritius, we identified 175 temporary residents (60% of all intertidal species in the study area, cf. Table 1) and in our field study, 16 of those species represented juvenile temporary residents that correspond to 66.7% of recorded species in shallow tidepools. Moray eels (Muraenidae), gobies (Gobiidae), damselfishes (Pomacentridae), groupers (Serranidae) and surgeonfishes (Acanthuridae) comprise temporary residents that can be found most frequently in the tidepools of Mauritius.

Conclusions
A considerable portion of littoral fishes occurs in the intertidal environment and the present study yielded a large number of intertidal species in Mauritian waters. However, knowledge about local distribution and ecology of these species is still much more fragmentary than that of intertidal species in the North-eastern Pacific or North Atlantic. Intertidal habitats are prone to human influences. The comparison of past and present data from Mauritius suggests a decline of intertidal residents over the last decades. A more detailed knowledge of intertidal communities and more long-term data could enable us to use intertidal fishes as indicators of environmental change and human impact.
For this reason, a future monitoring of tidepool communities and more detailed analyses, for example, with respect to the distribution of feeding types in the communities and the linking of community parameters to parameters of water quality, is highly desirable, not only in Mauritius but also elsewhere.