Geoscience Reference
In-Depth Information
exposed on a wave-cut platform) clump size can be assessed by
laying out a grid of contiguous quadrats. Increasing 'quadrat'
sizes are then built up by blocking adjacent quadrats in twos,
fours, eights, etc. A graph can then be constructed of mean
squares (variance) against block size. Peaks in the graph
correspond to the mean clump sizes. An example is shown in
Figure 5.12d in which there are two clump sizes: one at 4 block
(quadrat) size and one at 32. In practice there may be severe
constraints on laying out a contiguous grid of quadrats and the
use of this technique is likely to be limited. Nevertheless
pattern analysis of this kind can prove particularly rewarding
in autochthonous assemblages.
5
Edge effects
Inevitably some specimens will intersect quadrat boundaries. The
smaller the quadrat size the more important these edge effects
become because a greater proportion of the sampled material
will be represented by intersected specimens. Similarly the
perimeter length to area (or volume) ratio of the quadrat will
also be important. Quadrats approximating to circles or spheres
would reduce edge effects to a minimum but these shapes are
impractical. Squares or cubes (or approximations to them
controlled by bed thickness and lithological factors) are the
conventional choice. A possible exception here might be the
sampling of elongate specimens that are fl ow oriented. An
elongate, rectangular quadrat arranged so that its longest axis
was parallel to the fl ow alignment would minimize edge effects.
A common practice is to include within the sample any
individual (specimen) that has more than half its area within
the quadrat and exclude the others. This approach biases the
sample because it is infl uenced strongly by the size of the
specimens (small specimens will intersect the boundary less
frequently than large specimens). The actual area of the
specimen falling inside the quadrat could be included in cover
estimates (Section 5.5.3), but this would preclude any statistics
such as mean specimen size that might be related to transport
distance and differential degradation studies.
An alternative strategy is to include 50% of specimens of a
particular taxon that intersect the quadrat boundaries and
measure the whole area of those specimens included. The 50%
measured should be a random sample.
5.5 Estimating abundance
Abundance can be determined both qualitatively and
quantitatively. The ease of measures of abundance, their accuracy
and their relationship to sedimentological and ecological
phenomena varies greatly. Abundance can be estimated or
measured in the fi eld but an alternative method is to
