Fróðskaparrit - 01.01.2007, Page 123

GRÓÐURIN Á FLAGTEKJUM í FØROYUM OG Á GRASLENDINUM UTTANUM
- EIN KANNING ÚR SANDOYNNI
121
DCA analysis of roof, infield and outfield vegetation
Figurc 6. DCA scatterplot of roof, infield and outfield sanrples and taxa. Key: samples - open circles; taxa -
filled circles. See Figure 4 legend for taxon abbreviations. A: infield samplcs; B: turf roof samples: C:
outfield samples.
considerable amount of the variation in the
turf-roof data set is explained by axis 1 (in
Fig. 5), and its length of 3.5 S.D. suggests
that there is not a complete ‘tumover’ of
species (which would be indicated by a
value of >4) within the turf roofs, and
therefore, that despite differences in pre-
sumed condition, some species are com-
mon to all roofs.
The second axis in Figure 5 explains a
smaller amount of variation in the turf-roof
data set (eigenvalue of 0.372), and with no
trend apparent, it will be discussed no fur-
ther.
Very clear separation of three vegeta-
tion types is apparent in Figure 6, which
ordinates turf-roof, infield and outfíeld
vegetation. To the right, the samples from
the outfield vegetation are clearly grouped
together and characterized by mire taxa
such as Calluna, Nardus and Sphagnum
spp. Appearing more centrally on the first
axis are Prunella vulgaris and Viola spp.,
indicative of better-drained outfíeld sam-
ples where grazing has resulted in the
prevalence of grasses. The high eigenvalue
for the first axis (0.797) indicates that it
represents much of the variation in the data
and its length (4.644 SD) shows that there
is a complete turnover of species between
the mire outfíeld samples and those from
the turf roofs (i.e. the floristic composition
at the extremes of these two habitats is
completely different). The infíeld samples,