Fróðskaparrit - 31.12.2000, Side 114

118
DOMINANT SPECIES ABUNDANCE RELATED TO ENVIRONMENTAL
FACTORS ON ROCKY SHORES IN THE FAROEISLANDS
Ejfects of environmental factors at sites
with tidcil amplitude 0.4 m or less
In the DCA of the nine stations with tidal
amplitude 0.4 m or less (Fig. 5 and Table
5), axes 1 and 2 explained 33% and 10%,
respectively, of the species data (eigenval-
ues 0.29 and 0.09). The sequence of species
along the first axis was similar to that along
the first DCA axis of the other stations
(Spearman’s Rank Correlation Coefficient,
rs=-0.85, p<0.01). This axis was not corre-
lated, however, with the wave exposure in-
dex (correlation coefficient -0.14), but it
did correlate with the fjord index (correla-
tion coefficient 0.92), aspect (correlation
coefficient 0.69), and substrate (correlation
coefficient 0.57). This may have been the
reason why these stations did not fit in the
canonical ordination with the others,
wherein wave exposure was revealed as the
most important variable. The second axis
was correlated with slope (correlation coef-
ficient 0.58) and wave exposure index (cor-
relation coefficient 0.49).
Discussion
Effects of environmental factors at sites
with tidal amplitude larger than 0.4 m
The interpretation of some variables other
than wave exposure should be done with
some caution. The location of the various
stations was based on a stratified strategy
by which the entire geographical area was
covered and includeđ different environ-
mental factors, one of which was exposure
to wave action. If too many environmental
variables had been included in the stratifi-
cation determination, however, the stations
would be no longer representative of the
environmental conditions in the Faroes,
and the effects of the major factors might
have been wrongly estimated. As a conse-
quence, some variables did not have a bal-
anced distribution across stations (Table 2),
which led to some arbitrariness when at-
tributing effects to these factors. This ap-
plied particularly to the substrate variable
where only 12 of the 159 stations had sub-
strate other than bedrock. Most of the non-
bedrock stations were situated in relatively
sheltered locations, as reflected in the neg-
ative coiTelation between substrate and
wave exposure (Table 4), which further
complicated the interpretation of the sub-
strate variable.
Wave exposure is well known as a poten-
tial structuring factor for rocky shore com-
munities (e.g. Lewis, 1964). Table 6
demonstrates that the first axis in Detrend-
ed Correspondence Analysis (DCA) for
stations with tidal amplitude larger than 0.4
m reflected to a considerable degree the
species responses to wave exposure de-
scribed by other authors. The sequence of
species along the axis was correlated to the
species occurrence in relation to wave ex-
posure in the Faroe Islands following
Borgesen's (1902; 1905) description
(Spearman's Rank Correlation Coefficient,
rs =0.79, p<0.01), and to their responses to
wave exposure in the British Isles follow-
ing the classification (rs =0.79, p<0.01) of
Connor et al. (1996). Borgesen (1902;
1905) made a thorough description of the
algal flora in the Faroe Islands, but the
quantitative interpretation is ours. The tab-
ulateđ distribution in the British Isles is
based on the Connor et al. (1996) classifi-