LARGE-SCALE SPATIAL AND TEMPORAL PATTERNS OF DEEP-SEA BENTHIC SPECIES DIVERSITY
297
the bivalves, gastropods, cumaceans and polychaetes.
The depth for peak diversity estimated from the
regression curves varies among taxa from around 2000
to 3000 meters. A similar pattern was found in the
megafauna (Rex, 1981, 1983).
Allen and Sanders (1996), using a larger database,
reported on depth patterns of bivalve diversity in
the western North Atlantic. Bivalves are abundant
in the deep sea, but have comparatively low species
diversity (Fig. 10.1). For example, thirty epibenthic sled
samples taken in the western North Atlantic yielded
nearly 38 000 individuals, but only 36 species. The
deep-sea bivalve fauna has been exceptionally well
characterized, in that diversity appears to have reached
an asymptote with continued sampling (Allen and
Sanders, 1996), and, therefore, is probably a reasonable
approximation to the actual regional species pool. The
analysis by Allen and Sanders (1996) also indicated
a unimodal pattern of diversity, with peak diversity at
lower bathyal depths.
It is important to note that these studies are based
on data from qualitative sampling gear – epibenthic
sleds (Hessler and Sanders, 1967) and anchor dredges
(Sanders et al., 1965). Thus, diversity was measured
over relatively large spatial scales; epibenthic sleds,
for example, are towed for about a kilometer and
sample on the order of 1000 m
2 of the bottom. These
sampling methods obscure the spatial dispersion of
individual species. Collecting devices that sample
on large scales could traverse and combine together
smaller patches with distinctive faunas. If smallscale patch structure varies with depth, the shape of
diversity–depth gradients would reflect this. Therefore,
unimodal trends, like those shown in Fig. 10.1, could be
an artifact of changes in dispersion patterns on smaller
scales.
Jumars (1976) first drew the attention of deep-sea
biologists to the problem of scale – later recognized
as the central problem in ecology (Levin, 1992). The
development of the box corer by Hessler and Jumars
(1974) was a major advance in the understanding of
scale in the deep-sea benthos. The box corer collects
a relatively undisturbed 0.25 m
2 area of sediment
with the animals still in position, making it possible
to compare diversity on small scales that are more
relevant to the minute organisms that inhabit deep-sea
sediments. It also provides an effective way to explore
patterns of community structure among subcores of
a single box corer (centimeter scales), and among
core samples spaced different distances apart (meters,
kilometers and so on) in a spatial array. Diversity
can be measured as actual species densities (number
of species per unit area) rather than estimated by
artificially normalizing large samples. The number
of coexisting species can be correlated directly with
animal abundance and with sediment characteristics
from the same core samples. Deployment of box corers
greatly increased understanding of spatial dispersion
on small scales (Jumars and Eckman, 1983; Gage
and Tyler, 1991). It also made it possible to study
bathymetric patterns of diversity by using sampling
scales that capture microhabitat patchiness.
The Atlantic Continental Slope and Rise Study off
the east coast of the United States was the most
intensive box-corer sampling study ever undertaken
(Blake et al., 1985, 1987; Maciolek et al., 1987a,b).
Sampling was carried out in three geographic regions:
south of New England, east of New Jersey, and
off the Carolinas. The two more northerly regions
showed similar unimodal patterns of diversity and
shared much of the same fauna (Etter and Grassle,
1992; Rex et al., 1997). The combined data for
these two regions representing the whole macrofaunal
community are plotted in Fig. 10.2a. Diversity peaks
at around 1500 meters, roughly 1000 meters shallower,
on average, than for separate groups collected by
qualitative sampling near the same geographic area
(Fig. 10.1). Hence, the general unimodal diversity–
depth pattern seems to hold, but maximum diversity is
shifted to shallower depths. No sampling was carried
out below mid-bathyal depths.
Why peak diversity should appear somewhat shallower in the box-corer study remains unclear. The
two sets of diversity curves (Figs. 10.1 and 10.2a)
are not directly comparable. The data from qualitative
sampling gears (Fig. 10.1) are for only a subset of
deep-sea benthic taxa, whereas the box-corer data
represent the entire benthic community. The sizes to
which samples are normalized differ between the two
types of samples. The shape of the fitted curve for the
box-corer study (Fig. 10.2a) represents only the upper
bathyal region. It is clear from Fig. 10.1 that abyssal
samples can strongly affect the overall diversity–depth
pattern. The most abundant and diverse group, the
polychaetes, has the shallowest peak in diversity of
the four groups shown in Fig. 10.1, perhaps either
because their maximum diversity actually is shallower,
or because the data are from anchor dredge samples,
which typically collect over a smaller area than do the
epibenthic sleds used to collect the other three groups.
297
the bivalves, gastropods, cumaceans and polychaetes.
The depth for peak diversity estimated from the
regression curves varies among taxa from around 2000
to 3000 meters. A similar pattern was found in the
megafauna (Rex, 1981, 1983).
Allen and Sanders (1996), using a larger database,
reported on depth patterns of bivalve diversity in
the western North Atlantic. Bivalves are abundant
in the deep sea, but have comparatively low species
diversity (Fig. 10.1). For example, thirty epibenthic sled
samples taken in the western North Atlantic yielded
nearly 38 000 individuals, but only 36 species. The
deep-sea bivalve fauna has been exceptionally well
characterized, in that diversity appears to have reached
an asymptote with continued sampling (Allen and
Sanders, 1996), and, therefore, is probably a reasonable
approximation to the actual regional species pool. The
analysis by Allen and Sanders (1996) also indicated
a unimodal pattern of diversity, with peak diversity at
lower bathyal depths.
It is important to note that these studies are based
on data from qualitative sampling gear – epibenthic
sleds (Hessler and Sanders, 1967) and anchor dredges
(Sanders et al., 1965). Thus, diversity was measured
over relatively large spatial scales; epibenthic sleds,
for example, are towed for about a kilometer and
sample on the order of 1000 m
2 of the bottom. These
sampling methods obscure the spatial dispersion of
individual species. Collecting devices that sample
on large scales could traverse and combine together
smaller patches with distinctive faunas. If smallscale patch structure varies with depth, the shape of
diversity–depth gradients would reflect this. Therefore,
unimodal trends, like those shown in Fig. 10.1, could be
an artifact of changes in dispersion patterns on smaller
scales.
Jumars (1976) first drew the attention of deep-sea
biologists to the problem of scale – later recognized
as the central problem in ecology (Levin, 1992). The
development of the box corer by Hessler and Jumars
(1974) was a major advance in the understanding of
scale in the deep-sea benthos. The box corer collects
a relatively undisturbed 0.25 m
2 area of sediment
with the animals still in position, making it possible
to compare diversity on small scales that are more
relevant to the minute organisms that inhabit deep-sea
sediments. It also provides an effective way to explore
patterns of community structure among subcores of
a single box corer (centimeter scales), and among
core samples spaced different distances apart (meters,
kilometers and so on) in a spatial array. Diversity
can be measured as actual species densities (number
of species per unit area) rather than estimated by
artificially normalizing large samples. The number
of coexisting species can be correlated directly with
animal abundance and with sediment characteristics
from the same core samples. Deployment of box corers
greatly increased understanding of spatial dispersion
on small scales (Jumars and Eckman, 1983; Gage
and Tyler, 1991). It also made it possible to study
bathymetric patterns of diversity by using sampling
scales that capture microhabitat patchiness.
The Atlantic Continental Slope and Rise Study off
the east coast of the United States was the most
intensive box-corer sampling study ever undertaken
(Blake et al., 1985, 1987; Maciolek et al., 1987a,b).
Sampling was carried out in three geographic regions:
south of New England, east of New Jersey, and
off the Carolinas. The two more northerly regions
showed similar unimodal patterns of diversity and
shared much of the same fauna (Etter and Grassle,
1992; Rex et al., 1997). The combined data for
these two regions representing the whole macrofaunal
community are plotted in Fig. 10.2a. Diversity peaks
at around 1500 meters, roughly 1000 meters shallower,
on average, than for separate groups collected by
qualitative sampling near the same geographic area
(Fig. 10.1). Hence, the general unimodal diversity–
depth pattern seems to hold, but maximum diversity is
shifted to shallower depths. No sampling was carried
out below mid-bathyal depths.
Why peak diversity should appear somewhat shallower in the box-corer study remains unclear. The
two sets of diversity curves (Figs. 10.1 and 10.2a)
are not directly comparable. The data from qualitative
sampling gears (Fig. 10.1) are for only a subset of
deep-sea benthic taxa, whereas the box-corer data
represent the entire benthic community. The sizes to
which samples are normalized differ between the two
types of samples. The shape of the fitted curve for the
box-corer study (Fig. 10.2a) represents only the upper
bathyal region. It is clear from Fig. 10.1 that abyssal
samples can strongly affect the overall diversity–depth
pattern. The most abundant and diverse group, the
polychaetes, has the shallowest peak in diversity of
the four groups shown in Fig. 10.1, perhaps either
because their maximum diversity actually is shallower,
or because the data are from anchor dredge samples,
which typically collect over a smaller area than do the
epibenthic sleds used to collect the other three groups.