ber of species of which only Chondrus crispus is distributed over almost all clusters. The concentration patterns of the species in CIIA2 are the main factor for the ...
Marine Biology 92, 255-265 (1986)
Marine ==BiOlOgy 9 Springer-Verlag1986
Rocky substrate communities of the infralittoral fringe of the Boulonnais coast, NW France: a quantitative survey J. A. Kaandorp Institute of Taxonomic Zoology, University of Amsterdam; P.O. Box 20125, NL-1000 HC Amsterdam, The Netherlands
Abstract
Infralittoral fringe communities were investigated at 26 stations, in all four seasons (summer 1982, spring, autumn, winter 1983), in different habitats. The investigated habitats can be divided into undersurfaces of boulders, cavelike habitats, crevices between rocks, horizontal and vertical surfaces on rocks. The stations were distributed over four locations, all in the district Pas de Calais, arranged in a sequence of increasing exposure to wave action: Boulogne, Audresselles, Ambleteuse and Cap Gris Nez. Data on quantitative distribution of the biota were collected using quadrat sampling techniques; the data were processed by using cluster analysis. Several communities could be distinguished as separate units, viz. those of undersurfaces of boulders, caves, medium-exposed/exposed vertical/horizontal surfaces. A comparison with previous, preliminary quantitative investigations carried out in 1980 revealed that communities from shaded habitats are apparently stable enough to yield completely reproduceable results. When all habitats are compared, 78% of the samples of 1980 were clustered together with corresponding samples of 1982/1983.
Introduction
A reproduceable description of marine communities is an essential prerequisite for understanding the ecological effects of marine pollution and other human activities. Only when identical sampling procedures of different investigators deliver reproduceable results, does it become possible to use these methods for ecological monitoring. A large number of studies on marine hard substrate communities only deal with a particular taxonomic group, such as Algae (den Hartog, 1959; Boudouresque, 1971; Nienhuis, 1976); Porifera (Pansini et al., 1977); Octocorallia (Weinberg, 1978). Studies on marine hard substrate corn-
munities as a whole are often qualitative studies (e.g. Stephenson and Stephenson, 1972; KOnnecker and Keegan, 1983) With the quantitative approach it is possible to define communities by using species that are restricted to one community (characteristic species) and also by using species that are present everywhere but are dominant in just a few communities (dominant species). These quantitatively defined communities are more recognizable in the field and will yield more reproduceable results compared with the qualitative approach. The infralittoral fringe, the lowest part of the eulittoral from the upper limit of laminaria to L L W S , is the part of the eulittoral least influenced by desiccation during the period of low tide and with the highest diversity, making the infralittoral fringe communities the most suitable eulittoral communities for ecological monitoring. Compared to the sublittoral, it is possible to visit the infralittoral fringe more easily, since it is not necessary to dive and a lot of data can be collected in a relatively short time. Two examples of quantitative investigations on infralittoral fringe communities were carried out by van Soest and Weinberg (1981) and Bibiloni et al. (1982). The factors which are the primary causes of the distribution of the intertidal organisms on rocky shores have been discussed by, for example Crisp and Southward (1958); Lewis (1964); McQuaid and Branch (1984). The factors which may cause the differences between intertidal rocky shore communities mentioned by Lewis (1964) are wave action, topography and substrate, influence of sunlight and shade, climatic conditions, biological factors, and the time of low water of spring tides. The stations investigated are all situated in the same area (Fig. 1), and a number of factors such as climatic conditions, the time of low water and springtides, salinity and nutrient salts of the seawater (discussed by Crisp and Southward, 1958) may be assumed to remain constant within this relatively small area. Nearly all stations were on the same substrate, viz. Portland Sandstone (see Table 1; and for a general geo-
256
J.A. Kaandorp: A quantitative survey of rocky substrate communities
Table 1. A description of the sampled stations. + : thin layer of sediment of the polychaete Polydora ciliata (< 89cm); + + : a layer of P. ciliata sediment of medium thickness (89 cm); + + + : a thick layer ofP. ciliata sediment (> 1 cm). The substrate in Stations 14, 15, 16, 18, 20 was marble chalk, in all other stations the substrate consisted of Portland sandstone Undersurfaces of boulders
Cave-like habitats
Crevices
19 Erosion material
15 Caves between boulders
+
Vertical surfaces
Horizontal surfaces
17 Crevices in a 14 Surfaces directed rock formation towards the seaside
2 Surfaces on the south side of a pier-like rock formation
22 Surfaces on boulders
++
+
++
+
+
23 Erosion material
21 Caves between boulders
8 Crevices in a rock formation
16 Surfaces directed towards the pier side
3 Surfaces on the middle of a pier-like rock formation
6 Surfaces on the middle of a pier-like rock formation
+
+
+
++
+
+
12 Plate-like boulders, erosion material
4 Caves at the north side of a pier-like rock formation
18 Surfaces directed towards the pier side
7 Surfaces on a pier-like rock formation
11 puddles ("cuvettes") on a rock plateau
+
+++
+
+
5 Caves in the middle of a pier like rock formation
20 Surfaces directed towards the sea side
10 Surfaces on boulders directed towards the beach side
13 Surfaces on a rock plateau
+
+++
+
+
9 Caves between boulders
1 Surfaces on the north side of a pier-like rock formation
25 Surfaces on huge spherical boulders
24 Surfaces on huge spherical boulders
+
+
+
++
+
Vertical surfaces
26 Caves between huge spherical boulders +
CAP GRISN E Z ~ 2425 26 k~~
AUDRESSELLES 10111213 789
1234 5623
Capdela C r e c t ?_~ 12
DigueN o r d
~
DigueCarnot
BOULOGNE
logical description of the substrate see Bourn6rias et al., 1983). The r e m a i n i n g factors, which m a y cause the differences between the infralittoral fringe communities, are the influence of sunlight a n d shade, biological factors and (a factor most difficult to define and measure but which correlated strongly with the factor topography) exposure to wave action. The "classical" approach of estimating exposure to wave action at different sites is the comparison of the composition of eulittoral zonation at the different sites (Ballantine, 1961; Dalby, 1978). A more objective method of estimating the exposure to wave action has always been a great problem. Southward (1953) measured the wave wash above the predicted tide level. This method fails when coasts of different structure, causing different effects on wave wash, are compared. A more objective assessment of wave action was described by Jones a n d Demetropoulos (1968), who used dynamometers. Another method for assessing the wave action is by measuring the erosion of gypsum blocks. Muus (1968) introduced this method. A n important difference between this method a n d the dynamometers is that it is an integral measurement: wave' action is assessed during a certain period. In this paper, exposure to wave action is estimated by using gypsum blocks. Fig. 1. Situation of the stations over the four locations: Boulogne, Audresselles, Ambleteuse and Cap Gris Nez
J. A. Kaandorp: A quantitative survey of rocky substrate communities Materials and methods
Infralittoral fringe communities were investigated in all four seasons in order to find seasonal effects and to define communities, which are recognizeable in all seasons. The communities were investigated in habitats exposed to different light intensities and in locations with differing exposures to wave action. Cap Gris Nez (Fig. 1) is a cape protruding into the sea, and is the location with the greatest exposure to wave action. Boulogne (Cap de la Creche) is situated in the shelter of a pier (Digue Nord) and is the most sheltered location (Fig. 1). The two locations Audresselles and Ambleteuse are intermediate with respect to wave action. The habitats investigated may be divided into five main groups: undersurfaces of boulders, cave-like habitats, crevices between rocks, and horizontal and vertical surfaces. Other investigations, in the same area, have been performed by Giard (1913); den Hartog (1965); Glacon (1975). These three studies were all qualitative studies. A preliminary, unpublished, quantitative investigation was made by J. Pompe in 1980 using the same locations and stations as during the present study. The results of both studies were compared, in order to determine whether the sampling procedures of rocky substrate yield reproduceable results. The data collected were analysed by means of numerical classification using the Cyber 750 computers at SARA (University of Amsterdam). The advantage of numerical classification by computer, compared with a conceptual treatment, is that large data tables can be analysed very efficiently and reproduceably. The processing of the data was done partially with self-designed programs. All selfdesigned programs are PASCAL programs, suitable for the PAS3 compiler (Strait et al., 1981) and are available for common use. A detailed documentation manual for these programs may be received on request from the author. The cluster analysis and inverse analysis of data tables will be discussed in another paper (Kaandorp, in preparation). A comparison and description of the numerous numerical classification methods have been made by Sneath and Sokal (1973); Clifford and Stephenson (1975); Boesch (1977); Meurs (1978); Gauch (1982). The curves in Fig. 2 and Fig. 3 and the bar chart in Fig. 4 were plotted by using routines out of the DISSPLA graphics software package (Anonymous, 1981).
257 as sessile organisms were estimated as percentage cover and not counted, in order to obtain a data set in which all data are comparable with each other. One completely sampled station consisted of ten quadrats (a sampled area of 4 000 cm2). The percentage cover in one sampled station range from 0.1% (the chosen minimum value for a species present) to 100.0%. The date and ELWS (Extreme Low Water at Springtide) level, as measured by the mareograph of the Service Maritime des Ports de Boulogne s/mer et de Calais and the predicted ELWS value from the tide table (all ELWS levels are given in meters) were appended to the samples.
Minimal area The minimal area was calculated with the program MINAR. The algorithm used in this program has been described by Weinberg (1978b). In this, the similarity is calculated between each possible combination of subsets of m elements out of the ten quadrats, where the same element is not used twice in one combination (1 _-< m =< 5). The (dis)similarity coefficient used was the same as in the cluster analysis: the Bray-Curtis coefficient (the dissimilarity form, Equation 1); in which S(X,Y) is the dissimilarity between the samples X and Y, M is the total number of species, Xk and Yk are the transformed percentage cover for species k in ~he samples X and Y. M
S ( X , Y ) - i=I M
'
(1)
E (Xj -F Yj) j=l The percentage cove,r were logarithmically transformed using Eq. (2); in which x is the percentage cover and T (x) the transformed value, s. x=0.0
~ T(x) = 0
0. I _-
