Fróðskaparrit - 01.01.2007, Qupperneq 120

118
THE VEGETATION OF GRASS ROOFS IN THE FAROE ISLANDS AND
THE SURROUNDING GRASSLAND VEGETATION - A STUDY FROM SANDOY
by Hansen and Jóhansen (1982) and Jó-
hansen (1985)).
Hypothesis two: the decline in species di-
versity is a result of freedom from grazing
on the roof-top turfs (which then allows a
small number of more competitive species
to predominate).
Hypothesis three: the low species diversity
arises from the selection of species-poor
turfs for roofing.
Materials and Methods
The vegetation of 13 grass roofs displaying
a variety of dominant species, was asses-
sed in 76 randomly placed, 1 m2 quadrats.
Multiple quadrats were used for each roof
segment in order to ensure that the direc-
tion faced by each part of the roof (aspect)
was included.
The abundance of each plant taxon (in-
cluding cryptogams) was detennined on a
percentage scale. Measurements were also
made of the aspect of each roof segment,
and its slope, but soil samples were not re-
moved to prevent damage to the fabric of
the roofs. All of the samples were derived
from buildings in Húsavík (6 buildings)
and Sandur (7 buildings) on the island of
Sandoy (Figs 1-3), where a reasonable
number of turf-roofed buildings still exists.
The vegetation of the infield (bøur) and
outfield (hagi) areas in the vicinity of the
turf-roofed buildings was sampled (39
quadrats) in the same manner, for compari-
son with that of the roofs, and the same en-
vironmental variables (aspect and slope)
were also recorded.
The presence of grazing in field and
roof samples was determined from the
presence of dung and a score of 1 was
given to those samples were grazing was
evident (by sheep) and a score of 0 given
where grazing animals were unlikely to
gain access and/or dung was absent. Vas-
cular plant nomenclature follows Jóhansen
(2000) and Smith (1978) for the mosses.
Fieldwork took place in August 2003. A
major problem met with by the authors
was that householders (where resident at
the time of surveying), often had an impre-
cise idea of when the roofs - or segments
of the roofs - had last been replaced. Fur-
thermore, they had no idea of where the in-
dividual roofing specialists had obtained
their turf. The difficulty of locating indi-
vidual householders with a view to in-
creasing the number of roofs examined and
thus increasing the level of independence
of the data, was exacerbated by the fact
that a number of houses were not perma-
nently inhabited.
Statistical methods
A wide range of measures of diversity are
now in existence that measure both the
number of species present and taxon ho-
mogeneity. Separation of these two com-
ponents in the number expressed by the
method chosen is difficult (Hayek and
Buzas, 1997; Magurran, 1988) and in this
instance, a more straightforward compari-
son of species richness (a simple count of
the number of species) is used to assess
any difference in diversity between the
roof and ground vegetation.