Fish community structure and food chain dynamics in the surf-zone of sandy beaches: The role of detached macrophyte detritus

J. Exp. Mar. Biol. Ecol., 1984, Vol. 84, pp. 265-283 Elsevier

265

JEM 395

FISH COMMUNITY

STRUCI’URE AND FOOD CHAIN DYNAMICS IN THE

SURF-ZONE OF SANDY BEACHES: TI-IE ROLE OF DETACHED MACROPHYTE DETRITUS

A. I. ROBERTSON’ Division of Fisheries Research, CSIRO Marine Research Laboratories, P.O.Box 20. North Beach, W.A. 6020, Australia

R. C. J. LENANTON Western Australian Marine Research Laboratories, P. 0. Box 20. North Beach, W. A. 6020, Australia Abstract: The surf-zones of sandy beaches near Perth, Western Australia often harbour huge accumulations of detached macrophyte detritus. During 2.5 yr sampling, 29 species of fishes were captured over two sandy beaches in this region and the fish community was dominated by juveniles. There was a highly significant positive relationship between the number of fishes and the quantity of detached macrophytes taken in each surf-zone netting. Comparisons of total fish abundance on beaches with and without surf-zone accumulations of detached plants, showed that fishes were two to 10 times more abundant on the beach with weed accumulations, depending on the time of day, and date of sampling. However, despite the overall lower abundance of fishes on the open sandy beach, there was a significant increase in the number of fishes captured over the sandy beach at night. There were also two to five times the number of species over the beach with weed during the day, as opposed to equal numbers of species at night. Seven fish species made up >95% of the total catch and these species fell into two groups with regard to diurnal distribution patterns; those that were equally abundant in weed dominated or open surf-zones, and those that were weed-associated. Analyses of the diets of these fishes and the daytime distribution of an important avian piscivore in the surf-zone suggested that the large quantities of weed in the surf-zone of sandy beaches in this region provide both a rich feeding site for fishes, as well as a remge from diurnal predators. At night, when visual feeding predators are absent, some fish species move to open sandy areas to feed. Because the majority of fishes in this surf-zone community feed on weed-associated prey, and the input of macrophyte detritus is the major source of primary production in the surf-zone, we argue that the food chain dynamics in the surf-zone in this region are fundamentally different to those of sandy beaches that have been studied previously. Key words: surf-zone; sandy beach; fish community; detached macrophytes; food chain

The surf-zones environments for beach surf-zones planktivores and

of exposed sandy beaches are harsh, structurally homogeneous nektonic organisms and previous work on the fish faunas of sandy has shown that they are low diversity faunas, dominated by small benthic feeding fishes and their larger but rarer fish predators (e.g.

’ Present address: Australian Institute of Marine Science, P.M.B. No. 3, Townsville, M.C., Qld. 4810, Australia. 0022-0981/84/$03.00 0 1984 Elsevier Science Publishers B.V.

266

A. I. ROBERTSON

AND R. C. J. LENANTON

Gunter, 1958; McFarland, 1963; Edwards, 1973a,b; Modde & Ross, 1981; Lasiak, 1983; McDermott, 1983). A majority of surf-zone fish species are represented solely by immature individuals (McFarland, 1963; Modde, 1980) and year-to-year variations in abundance and species composition of these fish communities (e.g. Hillman et al. 1977; Modde & Ross, 1981) are probably tied to fluctuations in recruitment success. The trophic spectrum observed within any surf-zone fish community appears to be controlled primarily by three factors, the form of primary production input to the surf-zone, the pattern of water movement and the geomorphology of the sandy beaches. For instance, phytoplankton blooms occur regularly on some beaches with broad, flat surf-zones and these beaches usually have rich infaunal and zooplankton assemblages available to fishes (e.g. Lewin, 1977; McLachlan, 1980b; McLachlan et al., 1981; Lasiak, 1983). By contrast, where surf-zones are narrow and beaches are subject to greater sand movement, or where input of macroalgal debris occurs in large quantities, there may be little in the way of infaunal prey available to fishes (McLachlan, 1980a). In habitats other than sandy beaches, including coral reefs, seagrass meadows, kelp forests and the open sea, increased structural heterogeneity provided by corals, and attached or drifting macrophytes has a positive influence on the abundance and species richness of fish communities (Mitchell & Hunter, 1970; Luckhurst & Luckhurst, 1978; Orth & Heck, 1980; Stoner & Livingston, 1980; Kulczycki et al., 198 1; Moreno & Jara, 1984; Schulman, 1984). The results of some studies suggest that detached plant debris, or algae growing on man-made structures may also influence the structure of sandy beach fish communities by providing shelter and/or food resources not otherwise available to fishes in this stark environment (e.g. Edwards, 1973b; Wheeler, 1980). In temperate Western Australia, storms and associated wave-action remove macroalgae and seagrass material from the extensive limestone reefs and seagrass meadows which run parallel to the shore, and large quantities of detached vegetation accumulate in the surf-zone of many sandy beaches (Robertson & Hansen, 1982). During certain periods of the year up to 1200 m3 of the surf-zone per kilometre of coast may be occupied by detached macrophyte detritus (Lenanton et al., 1982). Earlier work on the fishes of estuaries and embayments on the open coast of south-western Western Australia showed that these extensive deposits of macrophyte detritus harbour large numbers of fishes and may serve as important nursery grounds for several fish species which were previously thought to be estuarine dependent (Lenanton, 1982; Lenanton et al., 1982). Through comparisons of the species composition, abundances, habitat preferences and feeding habits of the fish community on sandy beaches with and without drifting macrophytes, we show in this paper how the presence of detached macrophyte accumulations causes a major shift in the structure of the surf-zone fish fauna of south-western Western Australia. On the basis of data presented here and previous work on food chain dynamics in the surf-zone (e.g. Robertson & Hansen, 1982; Robertson & Lucas, 1983) we also argue that the structure of the food chain supporting the fish fauna differs significantly from that for other sandy beaches that have been intensively studied.

SURF-ZONEFISHESANDMACROPHYTEDETRITUS

261

MATERIAL AND METHODS SITE AND

FIELD METHODS

All sampling of fishes was done within the surf-zone of Mullaloo and Sorrento beaches near Perth, Western Australia (3 lo 5 1’ S : 115 “45’ E). The two beaches are 1 km apart and both are semiprotected behind a limestone reef which runs parallel to the shore, usually >2 km offshore. The beaches are bordered on their seaward side by extensive, multispecies seagrass meadows which are usually between 100 and 200 m from the surf-zone (see Kirkman, 1981 for a map of the site). Modal wave period on the beaches is 6.5 s and mean significant wave height just off the beach is 0.4 m while modal maximum height is 1.0-1.5 m (McLachlan & Hesp, in press a) as opposed to 1.4 m and 2.0-2.5 m respectively outside the reef (Steedman et al., 1977). The coast in this region has a small tidal range and meteorological effects can override tides (Hodgkin & di Lollo, 1958; Elliot et al., 1983). The intertidal faces of both beaches are fairly steep (6”) and the surf-zone is usually narrow (lo-20 m). On a 20-point scale for rating the exposure of the intertidal zones of sandy beaches (McLachlan, 1980a) both beaches rate 11.5 ( = exposed) (McLachlan 8z Hesp, in press a). The surf-zone of Sorrento beach is usually free of large accumulations of macrophyte detritus, while the surf-zone of Mullaloo beach often harbours significant quantities of detached plants. A small seine net (15 m long, 1.6 m deep with a 3 m pocket and 1 cm mesh throughout) was used to sample fish. The net, which was set by hand in 1.0-1.5 m of water and dragged perpendicular to the beach, swept an area of x 60 m2. During the period April 1981 to February 1982, 18 daytime (1500-1800) nettings were taken to provide specimens for gut analyses. These samples, and our previous work on the open coast in this region (Lenanton, 1982; Lenanton et al., 1982) suggested that the presence or absence of many species of fish in the surf-zone of sandy beaches was related to the quantities of macrophyte detritus in the water. Therefore, we designed a sampling regime to investigate the habitat preferences of the major species in the community. In 20 days between 30 March and 19 April 1982 we performed 11 replicate daytime nettings in each of three habitat types in the surf-zone, viz; within dense accumulations of detached macrophyte detritus (WEED habitat), at the edge of detached macrophyte patches (EDGE habitat) and over areas of beach clear of plant detritus (SAND habitat). WEED and EDGE samples were taken on Mullaloo beach while SAND samples were taken 0.5 and 1 km to the south on Sorrento beach. For all nettings the species composition and abundance were recorded in the field and subsamples of the most abundant species were retained for gut analyses. The total volume of plant material in each netting was also recorded using a 25-l plastic bucket as the measuring device. To determine if the patterns of habitat preferences we observed in the above sampling program were maintained on a die1 and seasonal time scale we also sampled the fish community in SAND and WEED habitats at day and night in June and October 1983. Fishes were identified and counted in the field, the volumes of weed recorded, and in June 1983 samples of the most abundant species were retained for examination of possible die1 dietary shifts.

268

A. I. ROBERTSON

AND R. C. J. LENANTON

In using beaches with and without accumulations of detached macrophytes to investigate the influence of plant detritus on surf-zone fish community structure we faced the problem of most “natural experiments”, the lack of a suitable control for other factors which may act on the fish fauna at each site. However we found a significant relationship between the volume of weed and the abundance of fishes at both test beaches (see p. 270). This occurred even though the range of weed volumes recorded from nettings at Sorrento beach was very small (O-12.5 1)compared to that at Mullaloo beach (6.3-575 1). These results suggest that except for the presence or absence of detached plants, the two beaches offer similar habitats for fishes. The relationship between fish abundance and species richness and drift algal abundance may in part be related to the shelter from predators provided by drift algae (e.g. Kulczycki et al., 1981). In Western Australia one obvious and potentially important predator on surf-zone fishes is the pied cormorant, Phalacrocorax varius (Gmelin), which is known to feed on several of the major fish species recorded in nettings from the surf-zone, including cobbler Cnidoglanis macrocephalus and yellow-eyed mullet AZdrichettaforsteri(Serventy, 1938; Serventy & Whittell, 1976). To examine the relationship between cormorants and patches of detached plant material in the surf-zone, we counted the number of actively hunting cormorants engaged in predatory activity within 50 m of the shore along 19.7 km of sandy beaches north of Perth, Western Australia. We censused this area on three occasions between September 198 1 and January 1982, on each occasion recording the length of the surf-zone filled with detached macrophyte material and the number and feeding site (open sand or detached plant accumulations) of cormorants. FISH

GUT

ANALYSES

In the laboratory, fishes were measured and weighed wet and their stomachs (or the first third of the gut for species without true stomachs) were removed and the contents examined under a dissecting microscope. Prey were identified to as low a taxonomic level as possible. For each sample of a fish species we recorded the percent occurrence of prey (0) (i.e. the percentage of fish in a sample which contained each prey type) and the percentage of the total volume of food which was made up by each prey (V). Prey volume estimates followed the method described in Robertson (1977). A prey ranking index (K = 0. V) modified from Hobson (1974) was calculated for each prey item in the diet of a sample of fish (e.g. Harmelin-Vivien & Bouchon-Navaro, 1983). By combining measures of the frequency of occurrence and the bulk of each prey in the diet, the index gives a clearer indication of the most important prey to the “average” fish in a sample, than would either of the two measures used alone (Hobson, 1974).

SURF-ZONE

FISHES AND MACROPHYTE

DETRITUS

269

STATISTICAL ANALYSES

Fishes

Where possible, we used analysis of variance (ANOVA) to compare the means for abundance and distribution data. For the data from the intensive diurnal sampling of SAND, WEED and EDGE habitats in March and April 1982 we used one-way ANOVA to test the null hypotheses that equal total numbers of fishes and species of fish were captured per netting in the three habitats. In all cases raw data were heteroscedastic (F,,,, test), but transformed data [log,, (x + l)] satisfied the assumptions of ANOVA. For analyses of single species abundances from the above data set, transformed data for Pelsartia humeralis (Ogilby), Sillago bassensis Cuvier and Crapatalus arenarius McCulloch were also heteroscedastic (F,,, tests; F,,,, = 5.05, 5.39 and 4.52 respectively; 0.05 > P > 0.01) but ANOVA was used because the analysis is robust for such small deviations from homoscedaticity, especially when sample sizes are equal (Box, 1953). For Sillago bassensis and Crapatalus arenarius there was no problem in making a Type-l error (Sokal & Rohlf, 1969) in interpreting the ANOVAs, because the analyses showed that the null hypotheses of equal mean catches should be accepted. However, for Pelsartia humeralis ANOVA gave a highly significant result (see p. 272). So as to decrease the probability of rejecting the null hypothesis when it may have been true, we reanalysed the P. humeralis data using a modified ANOVA in which the means were weighted according to the reciprocal of the variance of the sample (Sokal & Rohlf, 1969). For two fishes, Cnidoglanis macrocephalus (Valenciennes) and Haletta semifasciata (Valenciennes), 11 daytime nettings from the SAND habitat in March and April 1982 yielded only 1 and 0 specimens respectively, indicating that this was not an important habitat for either species. Comparisons of mean catches of these species were therefore confined to EDGE and WEED habitats using t-tests on log,, (X + 1)-transformed data. For H. semijksciata an approximate t-test, with reduced degrees of freedom (Sokal & Rohlf, 1969, p. 374) was used for the analysis because transformed data were still heteroscedastic. We used a three-way ANOVA for unequal sample sizes, with date, time of day and habitat as factors, to analyse data on total catch per netting from the die1 sampling program in June and October 1983. Each of the habitat-time-date replicate samples was expressed as a proportion of the total catch taken during the two dates (CN = 1423 fishes) and the analysis was performed on arcsin &transformed data (Fm,, test P > 0.05), since the raw data were heteroscedastic. Raw data for the number of species per netting from the June and October sampling periods were homoscedastic, and were analysed using a three-way ANOVA for unequal sample sizes.

270

A.I. ROBERTSONAND R.C.J. LENANTON

Birds

Because the sample size of actively feeding cormorants was small on each of the three sampling dates (N$25) we used the binomial distribution to calculate the exact probabilities of finding birds feeding over sandy beaches or near patches of detached macrophytes (Sokal & Rohlf, 1969, p. 82). The expected probabilities of birds feeding in either of the two habitats were calculated by using the relative proportions of the total length of beaches that were taken up by either habitat. RESULTS FISH COMMUNITYCOMPOSITION A total of 29 species from 26 families were recorded during the study and seven species made up > 95 y0 of the total number of fishes captured (Table I). Amongst these major seven species, Pelsartia humeralis (Teraponidae), Sillago bassensis (Sillaginidae), Cnidoglanis macrocephalus (Plotosidae) and Crapatalus arenarius (Leptoscopidae) were captured in > 50% of all nettings, while Aldrichetta forsteri (Mugilidae), Haletta semifasciata (Odacidae) and Torquigenerpleurogamma (Tetraodontidae) were taken in fewer nettings (Table I). In numerical terms, the community was dominated by juvenile fishes, with five of the top seven species being represented almost entirely by immature individuals (Table I). DETACHEDPLANTSAND FISH DISTRIBUTION For all nettings (CN = 63) for which the volume of detached plants was recorded, i.e. after February 1982, there was a highly significant, positive correlation between the total number of fishes and the volume of weed captured in each netting (r, = 0.76, P < 0.01; Fig. 1). The ranges of weed volumes and total fish abundance were O-575 1and O-376 1, respectively. When day and night samples were considered separately they also exhibited significant correlations between weed volume and fish abundance (day r, = 0.74, P < 0.01; night r, = 0.81, P < 0.01). Although the surf-zone of Sorrento beach was nearly always clear of plant detritus, weed volumes and fish catches were significantly correlated at both Sorrento (rs = 0.55, P < 0.02) and Mullaloo (r, = 0.34, P < 0.05) beaches. To determine whether the abundances of all species reacted in a similar fashion to the presence of detached macrophytes in the surf-zone, we compared the mean day-time abundances of fishes captured over open sandy areas (SAND) and at the edge of (EDGE) and within accumulations of detached macrophytes (WEED). There were significantly fewer fishes and less species per netting in the SAND habitat than either the EDGE or WEEDareas (Table II). Much of the variance associated with the mean catch per netting in the EDGE samples was due to the presence or absence of schools ofAldrichettaforsteri. If the data for A. forsteti were omitted from the analysis, then the abundance of fishes

SURF-ZONE FISHES AND MACROPHYTE DETRITUS

271

in each habitat increased in the order SAND EDGE WEED. Comparisons of the mean abundances of individual species in each of the three areas showed that, for the five species for which sufficient data were available, there existed two groups of fishes which

TABLEI

Sandy beach surf-zone fish community composition: data are the total number of specimens of each species captured in 81 nettings, the percentage of those nettings in which each species occurred and the representation of life history stages for those species making up more than 1 percent of the total catch; .I, z 95% of catch were juveniles; B, juveniles and adults present in significant numbers. Total number of fish

Percent occurrence

Life history stage(s)

72.8

J J J J B J B

Brachaluteres jacksonianus (Quoy & Gaimard) Parequula melboumensis (Castemau) Gymnapistes marmoratus (Cuvier) Apogon rueppellii Gunther Muraenichthys breviceps Gunther Bhabdosargus sarba (Forsskal) Gonorhynchus greyi (Richardson)

1785 815 809 752 195 64 61 45 45 26 22 16 11 11 10 9 7 2 2 1 1 1 1 1 1 1 1 1 1

Total number

4697

Species Pelsartia humeralis (Ogilby) Aldrichetta forsteri (Valenciennes) Sillago bassensti (Cuvier) Cnidoglanis macrocephalus (Valenciennes) Crapatalus arenarius McCulloch Haletta semifasciata (Valenciennes) Torquigener pleurogramma (Regan) Pranesus ogilbyi Whitley Sc0rpi.r georgianus Valenciennes Schuettea woodwardi (Waite) Hyperlophus vittatus (Castelnau) Paraplagusia unicolor (Macleay) Sillago schomburghii Peters Siphamia cephalotes (Castelnau) Enoplosus armatus (Shaw) Cristiceps australis Valenciennes Callionymus goodladi (Whitley) Platycephalus 1ongispinLsMacleay

Syngnathidae (1 sp.) C0ntusu.r brevicaudus Hardy Pomatomus saltator (Linnaeus) Pseudolabti parilus (Richardson)

12.3 61.7 55.6 71.6 13.6 24.7 4.9 11.1 11.1 9.9 16.0 4.9 3.7 7.4 7.4 6.2 2.5 2.5 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2 1.2

could be separated on the basis of their diurnal distribution patterns (Table II). Sillago bussensis and Crapa~ulus arenarius exhibited no significant difference in mean densities across habitats, while Pelsartia humeralis, Cnidoglanis macrocephalus and Haletta semifusciutu were all most abundant in WEED areas (Table II). Aldricheta firsten’ form large schools which present an obvious problem for com-

:

r, = O-76**

.

. .. l

.

.

0 . 0. . .

.

.

. .

0

0

0 .

.

.

0 0

.

.

.

.

0

’

0 .

.

l

.

.

.

0

. 0

.

0: ”

03

cm w

maa . I

I

10

I

I

20

30

40

I

50

I

60

RANK OF WEED VOLUME Fig. 1. The relationship between the ranked abundance of fish per netting: data are for volumes of detached plants were recorded data;

volume of detached macrophytes in each netting and the ranked the 47 daytime (0) and 16 night-time (0) nettings in which the (see p. 267); rp, Spearman rank correlation coefficient for pooled **, significant at P = 0.01.

TABLE

II

Comparison of variables for day-time nettings in the SAND, EDGE and WEED habitats: all nettings (n = 11 for each habitat) performed between 30 March and 19 April 1982; the mean total number of fish and species per netting, as well as five single species comparisons are compared using one-way ANOVA on log,,, (x + I)-transformed data; for two species only one individual was captured in the SAND, habitat and r-tests were used to compare means for EDGE and WEED habitats (see p. 269); all means shown are for untransformed data and underlined means are not significantly different at P = 0.05 (SNK test); ***, significant at P = 0.001; NS, not significant. Habitat Variable

WEED

F ratio or t value

SAND

EDGE

17.9 17.9

103.8 38.0

121.1 120.8

F 2,30 = 16.12*** F 2,3” = 18.52***

2.5

4.9

5.6

F 2.30 = 13.45***

14.8 1.9 0.5 0.1 0

22.9 6.8 3.5 2.1 0.4

22.2 2.0 55.1 35.0 4.5

F 2,30 = 1.72 NS F 2,3,, = 2.67 NS F 2,3,, = 45.63*** f20 = 11.98*** r10 = 8.02***

COMMUNITY

Total Total for Total

fish/netting (ah species) fish/netting (minus data Aldrichetta forsteri) species/netting

SINGLE SPECIES

Sillago bassensis Crapatalus arenarius Pelsartia humeralis Cnidoglanis macrocephalus Neoodax semijbsciata

213

SURF-ZONE FISHES AND MACROPHYTE DETRITUS

parison of mean abundances between habitats using ANOVA. For A. forsteri the 33 nettings during the March-April sampling yielded 735 ~di~du~s, 732 of which were taken in two nettings in the EDGE habitat. Signikantly more fishes were also captured in diurnal nettings in WEED areas in June and October 1983, but the proportion of the total catch taken in each habitat at day or night changed significantly on both dates (Fig. 2 and Table III). At night there were TABLEIII Three-way ANOVA of the proportion of the total catch of fish captured at day and night in the two habitat types at two dates: F,, -test showed no significant heterogeneity of variance for arcsin @-transformed data at P = 0.05; d.E, degrees of freedom; MS, mean square; *, significant at P = 0.025; ***, significant at P = 0.005; NS, not significant. Source of variation A. Tie of day B. Habitat C. Date AxB AxC BxC AxBxC Within cell

d.f. 1 1

1 1 1

1 1 22

WEED

SAND

JUNE

MS

F ratio

0.179 488.806 15.146 122.512 11.685 2.266 13.486 10.501

0.02 NS 46.55*** 7.21* 11.67*** 1.11 NS 0.22 NS 1.28 NS

SAND

OCTOBER

WEE0

Fig. 2. The mean (+ 1 SE) proportions of the total catch of surf-zone fish (N = 1423) taken in SAND and WEEDhabitats at day and night on two sampling dates in 1983: the number of nettings in each time-site-date combination is shown in parentheses; also shown are the relative proportions of the major fish species captured in each time-habitat-date sampling; P, Pelsartia humeralis; C, Cnidoglanis macrocephalus; S, Sillago bassensis; 0, all other species.

274

A. I.

ROBERTSONAND R.C.J. LENANTON

fewer fishes captured in WEEDhabitats than during the day, while there were increased catches in the SAND habitat at night (Fig. 2 and Table III). Although there was no seasonal difference in the pattern of die1 habitat use by the total fish community on each date (Table III, three-way interaction term, P > OS), overall, significantly more fishes were captured in the June sampling period (Table III and Fig. 2). The die1 samplings also revealed that the patterns of habitat choice exhibited by individual fish species changed with the time of day (Fig. 2). Pelsartia humeralis was less abundant in night-time WEED samples than in the equivalent daytime nettings in both June and October (Fig. 2), while in June P. humeralis was also abundant in night-SAND samples. Individuals of Cnidoglanis macrocephalus increased in abundance after dark in the WEED areas on both sampling dates, and like Pelsartia humeralis this fish was also more abundant in night SAND samples (Fig. 2). On both dates individuals of Sillago bassensis were absent from daytime WEED samples, but were present in night WEED nettings (Fig. 2). In June S. bassensis was equally abundant in day and night samples in the sand habitat, but occurred only in night-SAND samples in October (Fig. 2). Analyses of the mean number of fish species per netting from the sampling in June and October showed that the main factors, time of day, habitat and date had a significant independent influence on the variance in the number of species (Table IV). Overall, less species were taken per netting in October, and there were more species per netting in WEED habitats during the daytime on both dates (Fig. 3 and Table IV). However, significantly more species were captured in the SAND habitat at night than during the day (Table IV and Fig. 3), and there was no difference in the number of species captured per netting in the WEED and SAND habitats at night. There was also virtually no difference in the number of species caught in the weed habitats between day and night (Fig. 3).

DAYi3J NIGHTB) SAND

JUNE

DAY (3) NIGHTBJ WEED

DAY(4)NIGHT(51 SAND

OAY(4INIGHT(5) WEED

OCTOBER

Fig. 3. The mean ( + 1 SE) number of fish species captured per netting in SAND and WEED and night on two sampling dates in 1983: the number of nettings in each time-habitat-date is shown in parentheses.

habitats at day combination

SURF-ZONE FISHES AND MACROPHYTE DETRITUS

215

TABLEIV Three-way ANOVA of the number of fish species per netting: F,,- test on raw data showed no significant heterogeneity of variance at P = 0.05; d.f., degrees of freedom; MS, mean square; ***, significant at P = 0.005; NS, not signifkant. Source of variation

d.f.

MS

F ratio

A. Time of day B. Habitat C. Date AxB AxC BxC AxBxC Within cell

1 1 1 1 1 1 1 22

23.198 17.202 24.505 5.445 0.045 0.080 0.180 1.625

14.28*** 10.59*** 15.08*** 3.35 NS 0.03 NS 0.05 NS 0.11 NS

FISH DIETS

The four fish species which were most abundant in the daytime EDGE or WEED nettings fed mainly on the weed-associated amphipods Allorchestes compressa Dana and Atylus sp. (Lenanton et al., 1982) (Fig. 4). Weed-associated sphaeromid isopods were also important in the diet of Haletta semifasciata, while both Cnidoglanis macrocephalus and Aldrichetta forsten’ exhibited ontogenetic dietary differences. Bivalves and infaunal polychaetes formed a large portion of the diet of the larger size group of Cnidoglanis macrocephalus, while larger Aldrichetta forsteri individuals fed almost entirely on green algae, mainly Ulva sp. and Enteromolpha sp. as has been observed elsewhere (Robertson, 1984). Torquigener pleurogamma also fed predominantly on the two weed-associated amphipod species (Fig. 4). Sillago bassensis and Crapatalus arenarius, which were equally abundant in SAND and WEED habitats during the day (Table II), had markedly different diets to all diurnally WEED-associated fish species (Fig. 5). The stomachs of individuals of Sillago bassensis that were captured in the WEED habitat in the surf-zone contained mainly infaunal and weed-associated polychaete species, as well as significant quantities of amphipods. Individuals of S. bassensis that were captured in the SAND habitat contained mainly bivalve siphons and some polychaetes. Infaunal polychaetes and sand were the major constituents of the gut contents of sandfish Crapatalus arenarius in both the WEED and SAND habitats (Fig. 5). Analysis of the diets of the three most common fishes captured during die1 sampling in June 1983, showed that there were often dietary shifts with both habitat and time of day (Fig. 6). For Pelsartia humeralis, day- and night-time diets were similar in the WEED habitat, with the weed-associated amphipods Allorchestes compressa and Atylus sp. and detritus being the main food items. The stomachs of fishes captured at day and night in the SAND habitat also contained large quantities of Allorchestes compressa (indicating that these fishes had only recently left the WEED habitat) as well as polychaetes, detritus and fishes (Fig. 6).

276

A. I. ROBERTSON AND R. C. J. LENANTON

Cnidoglanis macrocephalus also fed on weed-associated

amphipods in the WEED habitat at day and night, but shifted their diet to mainly bivalves when feeding in the SAND habitat at night (Fig. 6). SilZagobassensis, fed mainly on infaunal prey, including Pelsartia

Neoodax

N=ss (2.9 -14-2 Cm)

N =21 (6.0 - 11-9 cm)

Cnidoglanis N q69 (4.2 - 16.7cm)

N q17 (19.o-62.2cm)

@oQQ@ Aldrichetta N = 23 (2.8-5*2Cm)

N =20 (8.6 - 12.6cm)

T Torquigener N=20 (9-3 -13.5cm)

Q(i>Q

At

0

Fig. 4. The diets of day-time WEED-associated fish species in the surf-zone: all fish were captured during the day between April 1981 and April 1982; pie diagrams show the relative importance of prey items in the diet based on the ranking index K = 0. V, where 0 is the percent occurrence and V is the percent volume of each prey in the diet (see p. 268); size ranges of analysed fish (cm) and sample sizes are given for each species; A, Allorchestes compressa; At, Atylus sp.; Am, other amphipods; B, bivalves; D, detritus; G, green algae; H, harpacticoid copepods; P, polychaetes; S, sphaeromid isopods; T, tanaids; 0, other prey.

polychaetes and bivalve siphons, during the day in the SAND habitat. At night, bivalve siphons and mysids were the major prey of S. bassensis in the SAND habitat, while in the WEED habitat at this time fish contained a variety of foods including weed associated amphipods as well as polychaetes, detritus, holothurians, fishes and bivalve siphons (Fig. 6). DISTRIBUTION

OF FEEDING CORMORANTS

On all three dates when birds were counted, there was a highly significant association of actively-feeding cormorants Phalacrocorax varius, with patches of detached plants in the surf-zone (Table V).

SURF-ZONE FISHES AND MACROPHYTE DETRITUS

Sillago

Weed N q42 (4=3 - 8-9 cm)

Weed

N

Sand N =14 (6m8- 13n4

Crapatalus

211

cm)

Sand N =14 (5-2 - 7m5Cm)

=15

(5m6- 8-2 cm)

Fig. 5. The diets of the two fish species that were captured in equal numbers in day-time nettings in WEED and SAND habitats: all fish were captured during the day between April 1981 and April 1982; pie diagrams and symbols as for Fig. 4., with the following additions; Bs, bivalve siphons; Fo, foraminiferans; N, natantians; S, sand.

TABLEV

The distribution of actively-feeding cormorants, Phalacrocorax varius, in the surf-zone of sandy beaches: the data are the number ofbirds observed feeding over open sandy beaches (SAND) or in and near accumulations of detached plants (WEED); figuresin parentheses give the expected number of birds if feeding occurred randomly along the shore (see p. 270); ***, significant at P = 0.001. Date September 1981 October 1981 January 1982

SAND

2 (14.4) 10 (21.1) 5 (14.6)

WEED

13 (0.6)*** 15 (3.9)*** 10 (0.4)***

278

A. I. ROBERTSON AND R. C. J. LENANTON DAY,WEED

DAY,SANO

NIGHT,WEED

N=20, (!X-10.9cm)

NdD,(7.2-11.3cm)

N=20,(7.2-11.7cm)

N=21,(7.0-11.3cm)

NIGHT.SAND

N ~23, (14.8-45.2cm)

N = 15, (13.9-50.5cm)

N= 10,&5-13,Ocm)

N- 9, (51-9.7cm)

Pelsartia

N -14, (l&3-32.9cm)

Cnidoglanis

N=13,(4.5-12.1 cm)

Sillago

Fig. 6. Dietary changes with time of day and habitat (WEED or SAND) for the three major fish species captured in June 1983: Pie diagrams and symbols as for Fig. 4., with the following additions; F, fish; H, holothurians; M, mysids; T, Tethygeneia sp.; NA, no specimens available for analysis.

DISCUSSION FISH COMMUNITY STRUCTURE

Large accumulations of macrophyte detritus which remain in the surf-zones of sandy beaches after storms in south-western Western Australia (Robertson & Hansen, 1982; Lenanton et al. 1982; J. A. Hansen, unpubl. data) have a significant influence on the abundance and species composition of the surf-zone fish community in this region. Our data show that there was a significant relationship between the quantity of macrophyte detritus and the abundance of fishes for all nettings taken on the two study beaches. The relationship between these two variables is striking when it is considered that the data are pooled for day and night samples and for different sampling times during the year, including the periods of recruitment of several of the fish species (Lenanton et al., 1982). The fish community also includes several mobile, schooling species such as Aldrichetta forsteri, which a priori would be expected to obscure such a relationship. A further source of variance in the relationship between fishes and weed abundance is the macrophyte species composition and its influence on the fishes’ prey densities. All detached macrophyte detritus is not of equal value to amphipods as food or shelter (Robertson & Lucas, 1983) and variation in the plant species composition ofmacrophyte patches leads to variation in the abundance of amphipods (A. Robertson,

SURF-ZONEFISHESANDMACROPHYTEDETRITUS

279

unpubl. data). This may in turn influence the abundance of fishes associated with weed patches. The comparisons of the abundance of fishes and the number of species in the surf-zone of beaches with and without detached plants showed that the number of fishes in the surf-zone of the beach with weed accumulations was two to 10 times that on the open sandy beach, depending on the time of day and date of sampling. Although plant accumulations harboured the greatest densities of fishes during both day and night the number of fishes over the sand beach increased significantly at night. The analyses also showed that during the day there were two to five times more species captured per netting over the beach with plant accumulations, but that at night there were equal numbers of species per netting over both the WEED and SAND beaches. These increases in the density and number of species of fish in the sand habitat at night probably result from movements of fishes out of adjacent accumulations of detached plants in the surf-zone as well as from nearby seagrass meadows which are known to have species in common with the surf-zone fish community (Dybdahl, 1979; pers. obs.). Analyses of the distribution and abundance of individual fish species revealed that the numerically dominant species belonged to one of two groups. One group of species, which included Pelsartia humeralis and Cnidoglanis macrocephalus on all sampling dates as well as Haletta semijksciata on some occasions, were significantly more abundant in WEED patches during the day. The two major food items for all of these fishes were the weed-associated amphipods Allorchestes compressa and Atylus sp. Previous work has shown that both amphipods are only found associated with detached plants in the surf-zone (Lenanton et al., 1982; Robertson & Hansen, 1982; A. Robertson, unpubl. data). At night Pelsartia humeralis and Cnidoglanis macrocephalus were still more abundant in the WEED area, but some individuals of both species were found over SAND areas. Individuals of Pelsartia humeralis captured at night over both WEED and SAND areas contained mainly weed-associated amphipods and plant detritus, indicating that these fish had not fed in the SAND habitat. Cnidoglanis macrocephalus captured in the WEED habitat at night had a similar diet to daytime, WEED-inhabiting individuals, but at night in the SAND area C. macrocephalus switched its diet to infaunal bivalves. Because some individuals of both Pelsartia humeralts and Cnidoglanis macrocephalus leave patches of weed and move over adjacent sandy areas at night, it appears that in addition to the high availability of food in weed patches (Lenanton et al., 1982) other factors may be responsible for the day-time distribution pattern of Pelsartia humeralis and Cnidoglanis macrocephalus. Both species are eaten by pied cormorants (Serventy, 1938; Serventy & Whittell, 1976) and during the day cormorants feed in close association with patches of detached plants in the surf-zone (Table V). Other known piscivores such as Pomatomus saltator (Pomatomidae) also feed along sandy beaches in Western Australia (Thompson, 1958), and it is probable, therefore, that patches of weed act as daytime refuges from predators as well as being important feeding sites for individuals of Pelsartia humeralis and Cnidoglanis macrocephalus. In contrast, the distribution and feeding habits of two other major species in this fish

280

A. I. ROBERTSON AND R. C. J. LENANTON

community did not appear to be tied to patches of plant detritus. Although the die1 distribution pattern for Sillago bassensis changed seasonally, at no date was the abundance of S. bassensis greater in WEED than in the SAND habitat. And on the date when sufficient numbers of Crapatalus arenarius were available for proper analysis, the mean density of this species did not differ across habitats. Both species fed mainly on infaunal polychaetes and bivalve siphons, and were not dependent on the patches of weed for food. However, some weed-associated amphipods were present in the guts of Sillago bassensis. While the distribution of both species within the surf-zone does not appear to be influenced by diurnal predation, their behaviour may be modified by predator activity, as is evidenced by the ability of Crapatalus arena&s to bury in the sand during the day (Scott et al., 1974). Because previous work on other sections of the coast in south-western Western Australia has also shown that increased catches of fishes in the nearshore marine environment occur in the presence of detached plant material (Lenanton, 1982) we feel that the conclusions regarding the influence of macrophyte detritus on the abundance and species richness of the surf-zone fish fauna in this study can cautiously be applied to the whole temperate west coast region of Western Australia. The accumulation of detached macrophytes in the surf-zones of beaches in Western Australia is patchy in space and time, and we expect this to be a major influence on the structure of the fish community at any particular surf-zone site. Some beaches harbour deposits of macrophyte debris year-round, others accumulate detritus only after storms (which occur mainly in winter), and some beaches remain relatively clean year-round (J. Hansen, unpubl. data). Because the abundance and species composition of the fish fauna is so dependent on the presence or absence of detached plants, descriptions of the structure of the surf-zone fish community will be highly dependent on the scale of observation. Our previous work in the surf-zone has shown that when summed over 50 km of coast, the quantity of plant detritus in the surf-zone shows a definite seasonal fluctuation with most detritus present in the winter months, June to October (Lenanton et al., 1982; Robertson & Hansen, 1982) indicating that even on such a large scale there are likely to be seasonal changes in the abundance and composition of the surf-zone fish fauna. SURF-ZONE

FOOD CHAINS

Like other surf-zone fish faunas (e.g. McFarland, 1963 ; Modde, 1980; Lasiak, 1983) the present fish community is dominated by juvenile fishes. However, a striking feature of the fish fauna in the present study is the dominance of benthic feeding fishes and the relative scarcity of planktivores, which form a major trophic group within other surfzone fish communities (e.g. see Table V in Modde & Ross, 1981; McLachlan et al., 1981; Las& 1983; McDermott, 1983). Large schools of two planktivorous clupeids Spratelloides robustus Ogilby and Hyperlophus vittatus have been captured in other surf-zone sampling in Western Australia (Lenanton, 1982; McLachlan & Hesp, in

SURF-ZONE FISHES AND MACROPHYTE DETRITUS

281

press a), but these were either absent (e.g. Spmtelloides robustus) or rare in the present study. The reason(s) for the absence of planktivores in the surf-zones of Sorrento and M&loo beaches is not known, but the large numbers of benthic feeding fishes is obviously linked to the presence of extensive patches of macrophytes in the surf, which appear to provide both food and protection from predation for fishes. In Venezuela, Edwards (1973a, b) also observed that relative to an open sandy beach, there was a higher proportion of benthic feeders on a beach which had large quantities of detached Enteromolpha spp., and Wheeler (1980) has observed that algae attached to man-made structures on open sandy beaches in England caused a similar shift in the trophic habits of the fish community. The structure of the food chain supporting surf-zone fishes at the study sites, therefore, appears to be fundamentally different to that in the surf-zones of sandy beaches elsewhere in the world. In those sandy beach systems that have been extensively studied to date, it appears that phytoplankton blooms occur regularly in the surf-zone, providing the major energy and nutrient input to rich zooplankton and infaunal assemblages (Lewin, 1977; McLachlan, et al., 198 1) which are in turn consumed by planktivorous and benthic feeding fishes (McLachlan et al., 1981; Lasiak, 1983). Nutrients excreted by the dense infaunal populations appear to be suflicient to sustain the phytoplankton blooms, and the food chain is almost self-contained within the surf-zone (Lewin, 1977; Lewin etal., 1979; McLachlan, 1980b; McLachlan et al., 1981). By contrast the major input of primary production to beaches in south-western Western Australia is macrophyte detritus which is swept into the surf-zones by storms (Robertson & Hansen, 1982). Estimates of the turnover time of this material in the surf-zone and on beaches near Perth, Western Australia reveal that x 20 percent of the nearshore benthic primary production, or 240 tonnes dry wt * km- ’ coastline. yr - ‘, of detached plants are processed through the sandy beach system (Hansen, 1983 & her unpubl. data). Phytoplankton production is low in the inshore waters near Perth (D. Smith, pers. comm.) and phytoplankton blooms are rarely seen in the surf-zone along the lower west coast of Western Australia, although they have been observed on the southern coast (McLachlan & Hesp, in press b). As the macrophyte material decomposes in the surf-zone in Western Australia, it is fed upon by large populations of the amphipod Allorchestes compressa, which has a major influence on detrital turnover in this system (Robertson & Lucas, 1983). Since A. compressa is the major constituent of the diets of the majority of fish species occupying the surf-zone, the macrophyte detritus-amphipod-fish pathway appears to be the major route of energy flow in this system.

282

A. I. ROBERTSON AND R. C. J. LENANTON ACKNOWLEDGEMENTS

R. E. Johannes, C. J. Crossland and D.A. Hancock read and commented on earlier drafts of the manuscript. During the project A. I. Robertson was supported by a CSIRO Postdoctoral Fellowship and an MST Research Fellowship.

REFERENCES Box, G. E. P., 1953. Non-normality and tests on variances. Biometrika, Vol. 40, pp. 3 18-335. DYBDAHL,R.E., 1979. Technical report on fish productivity. An assessment of the marine fauna1 resources of Cockburn Sound. Department of Conservation and Environment, Western Australia, Report No. 4, 96 PP. EDWARDS,R.R.C., 1973a. Production ecology of two Caribbean marine ecosystems. I. Physical environment and fauna. Estuarine Coastal Mar. Sci., Vol. 1, pp. 303-3 18. EDWARDS,R. R. C., 1973b. Production ecology of two Caribbean marine ecosystems II. Metabolism and energy flow. Estuantie Coastal Mar. Sci., Vol. 1, pp. 319-333. ELLIOT,I. G., D. J. CLARKE& A. RHODES, 1982. Beach-width variation at Scarborough, Western Australia. J. R. Sot. West. Aust., Vol. 65, pp. 153-158. GUNTER, G., 1958. Population studies of the shallow water fishes of an outer beach in south Texas. Pub. Inst. Mar. Sci.. Univ. Tex., Vol. 5, pp. 186-193. HANSEN, J.A., 1983. Macrophyte decomposition in beach wrack and surf-zone accumulations along the Western Australian coast. Bull. Aust. Mar. Sci., No. 83, p. 19. HARMELIN-VIVIEN,M. L. & Y. BOUCHON-NAVARO,1983. Feeding diets and significance of coral feeding among chaetodontid fishes in Moorea (French Polynesia). Coral Reefs, Vol. 2, pp. 119-127. HILLMAN,R.E., N. W. DAVIS& J. WENNEMER,1977. Abundance, diversity, and stability in shore-zone fish communities in an area of Long Island Sound affected by the thermal discharge of a nuclear power station. Estuarine Coastal Mar. Sci., Vol. 5, pp. 355-381. HOBSON,E. S., 1974. Feeding relationships of teleostean fishes on coral reefs in Kona, Hawaii. Fish. BUN. NOAA, Vol. 72, pp. 915-1031. HODGKIN,E.P. & V. DI LOLLO, 1958. The tides of Southwestern Australia. J. R. Sot. West. Aust., Vol. 41, pp. 42-54. KIRKMAN,H., 1981. The first year in the life-history and the survival of the juvenile marine macrophyte, Ecklonia rudiatu (Turn.) J. Agardh. J. Exp. Mar. Biol. Ecol., Vol. 55, pp. 243-254. KULCZYCKI,G.R., R. W. VIRNSTEIN& W. F. NELSON, 1981. The relationship between fish abundance and algal biomass in a seagrass-drift algae community. Estuarine Coastal ShelfSci., Vol. 12, pp. 341-347. LASIAK,T. A., 1983. The impact of surf-zone fish communities on fauna1 assemblages associated with sandy beaches. In, Sandy beaches 4s ecosystems, edited by A. McLachlan & T. Erasmus, Dr. W. Junk Publishers, The Hague, pp. 501-506. LENANTON,R. C. J., 1982.Alternative non-estuarine nursery habitats for some commercially and recreationally important fish species of South-western Australia. Aust. J. Mar. FreshwaterRes., Vol. 33, pp. 881-900. LENANTON,R. C. J., A.I. ROBERTSON& J.A. HANSEN, 1982. Nearshore accumulations of detached macrophytes as nursery areas for fish. Mar. Ecol. Progr. Ser., Vol. 9, pp. 51-57. LEWIN,J., 1977. Persistent blooms of surf-diatoms along the northwest coast. In, The marineplant biomass of the Pacific northwest coast, edited by R. Krauss, Oregon State University Press, Corvallis, pp. 81-92. LEWIN,J., J. E. ECKMAN& G. N. WARE, 1979. Blooms of surf-zone diatoms along the coast of the Olympic Peninsula, Washington. XI. Regeneration of ammonium in the surf environment by the Pacific razor clam Siliqua pat&. Mar. Biol., Vol. 52, pp. l-9. LUCKHURST,B. E. & K. LUCKHURST,1978. Analysis of the influence of substrate variables on coral reef fish communities. Mar. Biol., Vol. 49, pp. 317-323. MCDERMOTT,J. J., 1983. Food web in the surf-zone on an exposed sandy beach along the mid-Atlantic coast ofthe United States. In, Sandy beaches as ecosysfems, edited by A. McLachlan & T. Erasmus, Dr. W. Junk Publishers, The Hague, pp. 529-538.

SURF-ZONE

FISHES AND MACROPHYTE

DETRITUS

283

MCFARLAND,W.N., 1963. Seasonal change in the number and the biomass of fishes From the surf at Mustang Island, Texas. Publ. Inst. Mar. Sci., Univ. Tex., Vol. 9, pp. 91-105. MCLACHLAN,A., 1980a. The definition of sandy beaches in relation to exposure: a simple rating system. S. Afr. J. Sci., Vol. 16, pp. 137-138. MCLACHLAN,A., 1980b. Exposed sandy beaches as semi-closed ecosystems. Mar. Environ. Res., Vol. 4, pp. 59-63. MCLACHLAN,A., T. ERASMUS,A.H. DYE, T. WOOLDRIDGE,G. VAN DER HORST, G. Rossouw, T.A. LASIAK & L. MCGYNNE, 1981. Sand beach energetics: an ecosystem approach towards a high energy interface. Estuarine Coastal ShelfSci., Vol. 13, pp. 1l-25. MCLACHLAN,A. & P.A. HESP, in press a. Faunal response to the morphodynamics of a reflective microtidal beach with cusps. Mar. Ecol. Progr. Ser. MCLACHLAN,A. & P. A. HESP, in press b. Surf zone diatom accumulations on the Australian coast. Search. MITCHELL,C.T. & J. R. HUNTER, 1970. Fishes associated with drifting kelp, Macrocystts pyriira, off the coast of southern California and northern Baja California. Caltf Fish Game, Vol. 56, pp. 288-297. MODDE,T., 1980. Growth and residency ofjuvenile fishes within a surf-zone habitat in the Gulf of Mexico. Gulf: Res. Rep., Vol. 6, pp. 377-385. MODDE, T. & ST. Ross, 1981. Seasonality of fishes occupying a surf zone habitat in the northern Gulf of Mexico. Fish. Bull. NOAA, Vol. 78, pp. 911-922. MORENO,C. A. & H. F. JARA, 1984. Ecological studies on fish fauna associated with Macrocystis pyrtifera belts in the south of Fueguian Islands, Chile. Mar. Ecol. Progr. Ser., Vol. 15, pp. 99-107. ORTH, R. J. & K. L. HECK, 1980. Structural components of eelgrass (Zostera marina) meadows in the lower Chesapeake Bay-fishes. Estuaries, Vol. 3, pp. 278-288. ROBERTSON,A.I., 1977. Ecology of juvenile King George whiting Sillaginodes punctatus (Cuvier and Valenciennes) (Pisces :Perciformes) in Westemport. Aust. J. Mar. Freshwater Res., Vol. 28, pp. 35-43. ROBERTSON,A. I., 1984. Trophic interactions between the fish and macrobenthos of an eelgrass community in Western Port, Victoria. Aquat. Bot., Vol. 18, pp. 135-153. ROBERTSON,A. I. & J. A. HANSEN, 1982. Decomposing seaweed; a nuisance or a vital link in coastal food chains? CSIRO Division of Fisheries Research Report 1980-1981, pp. 75-83. ROBERTSON,A. I. & J. S. LUCAS, 1983. Food choice, feeding rates and the turnover of macrophyte biomass by a surf-zone inhabiting amphipod. J. Exp. Mar. Biol. Ecol., Vol. 72, pp. 99-124. SCHULMAN,M. J., 1984. Resource limitation and recruitment patterns in a coral reeffish assemblage. J. Exp. Mar. Biol. Ecol., Vol. 74, pp. 85-109. SCOTT,T. D., C. J. M. GLOVER& R. V. SOUTHCO~, 1974. The marine andjieshwaterfihes of South Australia. Government Printer, South Australia, 392 pp. SERVENTY,D.L., 1938. The feeding habits of cormorants in South-western Australia. Emu, Vol. 38, pp. 293-316. SERVENTY,D.L. & H.M. WHI~ELL, 1976. Birds of Western Australia. University of Western Australia Press, Perth, 481 pp. SOKAL, R.R. & F.J. ROHLF, 1969. Biometry. W.H. Freeman and Co., San Francisco, 776 pp. STEEDMAN,R. K. & ASSOCIATES,1977. Mullaloo marina environmental investigations. Part 1. Physical oceanographic studies. Report to Wamieroo Shire Council, Perth, West. Aust., pp. l-210. STONER,A. W. & R. J. LIVINGSTON,1980. Distributional ecology and food habits of the banded blenny Paraclinusfasciatus (Clinidae): a resident in a mobile habitat. Mar. Biol., Vol. 56, pp. 239-246. THOMSON,J. M., 1958. The food of Western Australian estuarine fish. West. Aust. Fish. Fauna. Dept. Fish. Bull., No. 7, pp. 1-13. WHEELER,A., 1980. Fish-algal relations in temperate waters. In, The shore environment. Vol. 2. Ecosystems edited by J.H. Price, D.E.G. Irvine & W.F. Famham, Academic Press, London, pp. 677-698.