10 Oct 2001 16:3 AR ar142-03.tex ar142-03.

sgm ARv2(2001/05/10) P1: GJB

Annu. Rev. Ecol. Syst. 2001. 32:51–93

ENVIRONMENTAL INFLUENCES ON REGIONAL

DEEP-SEA SPECIES DIVERSITY∗
Lisa A. Levin,1 Ron J. Etter,2 Michael A. Rex,2 Andrew
J. Gooday,3 Craig R. Smith,4 Jesús Pineda,5 Carol
T. Stuart,2 Robert R. Hessler,1 and David Pawson6
1
Scripps Institution of Oceanography, University of California, San Diego, La Jolla,
California 92093-0218; e-mail: [email protected]; [email protected]
2
Department of Biology, University of Massachusetts, Boston, Massachusetts 02125;
e-mail: [email protected]; [email protected]; [email protected]
3
Southampton Oceanography Centre, European Way, Southampton SO14 3ZH United
Kingdom; e-mail: [email protected]
4
Department of Oceanography, University of Hawaii, Honolulu, Hawaii 96822;
e-mail: [email protected]
5
Department of Biology, MS 34, Woods Hole Oceanographic Institution, Woods Hole,
Massachusetts 02543; e-mail: [email protected]
6
National Museum of Natural History, Smithsonian Institution, MRC 106, Washington,
DC 20560; e-mail: [email protected]

Key Words Biodiversity, benthos, environmental gradients, depth gradients,

diversity measures, bathyal, abyssal, sediments
■ Abstract Most of our knowledge of biodiversity and its causes in the deep-sea
benthos derives from regional-scale sampling studies of the macrofauna. Improved
sampling methods and the expansion of investigations into a wide variety of habitats
have revolutionized our understanding of the deep sea. Local species diversity shows
clear geographic variation on spatial scales of 100–1000 km. Recent sampling programs
have revealed unexpected complexity in community structure at the landscape level that
is associated with large-scale oceanographic processes and their environmental conse-
quences. We review the relationships between variation in local species diversity and the
regional-scale phenomena of boundary constraints, gradients of productivity, sediment
heterogeneity, oxygen availability, hydrodynamic regimes, and catastrophic physical
disturbance. We present a conceptual model of how these interdependent environmen-
tal factors shape regional-scale variation in local diversity. Local communities in the
deep sea may be composed of species that exist as metapopulations whose regional
distribution depends on a balance among global-scale, landscape-scale, and small-scale
dynamics. Environmental gradients may form geographic patterns of diversity by influ-
encing local processes such as predation, resource partitioning, competitive exclusion,

∗
The US Government has the right to retain a nonexclusive, royalty-free license in and to
any copyright covering this paper.

51
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

52 LEVIN ET AL.

and facilitation that determine species coexistence. The measurement of deep-sea

species diversity remains a vital issue in comparing geographic patterns and evaluat-
ing their potential causes. Recent assessments of diversity using species accumulation
curves with randomly pooled samples confirm the often-disputed claim that the deep
sea supports higher diversity than the continental shelf. However, more intensive quan-
titative sampling is required to fully characterize the diversity of deep-sea sediments,
the most extensive habitat on Earth. Once considered to be constant, spatially uniform,
and isolated, deep-sea sediments are now recognized as a dynamic, richly textured en-
vironment that is inextricably linked to the global biosphere. Regional studies of the last
two decades provide the empirical background necessary to formulate and test specific
hypotheses of causality by controlled sampling designs and experimental approaches.

INTRODUCTION

Regional-scale sampling studies have provided the primary evidence for spatial
patterns of deep-sea community structure and their causes. There has been no syn-
optic review of regional diversity patterns published since the early 1980s (Rex
1981, 1983). During the last two decades, our knowledge has expanded dramati-
cally from descriptions of geographic gradients in diversity and indirect inferences
about the biotic and abiotic factors that may shape them, to more precise under-
standing of associations between diversity and specific environmental conditions.
These findings have transformed our understanding of the deep-sea ecosystem.
The deep-sea, soft-sediment environment is highly complex at the landscape level
of biodiversity as well as at the local community level, and is dynamically linked
to oceanographic processes at the surface.
Here we summarize regional-scale geographic trends of deep-sea species di-
versity and propose a conceptual model to explain their causes. We are primarily
concerned with ecological structuring agents that function on generational rather
than evolutionary time scales. The stochasticity or patchiness observed at very
small scales in deep-sea communities (Jumars & Eckman 1983, Grassle & Morse-
Porteous 1987, Grassle & Maciolek 1992) becomes resolved at larger scales into
measurable geographic patterns. The key to understanding deep-sea species diver-
sity lies in documenting these patterns and in discerning the scales at which various
causes operate to generate them. Patterns of diversity in the deep sea are much
more complicated than previously thought. Early attention centered on bathymetric
gradients in the western North Atlantic, the most intensively sampled region of the
deep sea. Qualitative (Rex 1981) and quantitative (Etter & Grassle 1992) sampling
studies indicated that diversity-depth patterns in the deep sea are unimodal with a
peak at intermediate depths and depressed diversity at upper bathyal and abyssal
depths. However, unimodal patterns do not appear to be universal (Rex et al. 1997,
Stuart et al. 2001), and where they do occur in other basins have been attributed
to varied environmental gradients (Paterson & Lambshead 1995, Cosson-Sarradin
et al. 1998). Also, a variety of unanticipated oceanographic conditions at specific
depths interrupt and modify bathymetric horizontal diversity trends (Gage 1997,
Levin & Gage 1998, Vetter & Dayton 1998).
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 53

We relate diversity patterns to the regional-scale phenomena of geogra-

phic boundary effects, variation in sediment grain size, productivity gradients,
oxygen minimum zones, current regimes, and catastrophic disturbances. Regional
gradients of diversity represent the combined effects of these ecological factors in
governing the rates of local processes. Regional-scale processes are, in turn, a con-
sequence of global oceanography and climate. Compared to other major ecosys-
tems, deep-sea communities appear to be controlled by the same basic mechanisms
of energy availability, biological interactions, disturbance, and heterogeneity, but
exerted through a very distinctive set of environmental circumstances that is new
to ecology. We then consider the important issue of measuring diversity in the
deep sea—a basic and persistent problem in comparisons among deep-sea studies
and between deep-sea and surface environments (Gage & May 1993, Rex et al.
2000). Finally, a conceptual model is offered to explain how environmental factors
interact with biotic processes to generate observed regional patterns.
It is difficult to define, in general terms, a physically or biologically meaningful
regional scale in the deep sea. A region is often envisioned topographically as a
major deep basin, such as the North American Basin of the North Atlantic that
is bounded by the North American Continent and the mid-Atlantic Ridge system
(Stuart & Rex 1994). However, deep basins are confluent at considerable depths
and connected by the deep thermohaline circulation. Their faunas are not entirely
or even largely endemic, suggesting that they neither have distinctive ecologies
nor act as isolated theaters of evolution. For our purposes, regions represent
areas of roughly 100s to 1000s km2 that have been well sampled and encompass
potentially significant and measurable environmental gradients or ecotones. Most
regional studies have taken place at bathyal depths (200–4000 m) on continen-
tal margins; much less is known about large-scale spatial variation in abyssal
(>4000 m) communities.
We focus on the deep-sea macrofauna of soft sediments because it is the
most diverse and well-studied component of the benthos. The macrofauna is
composed of animals retained on a 300 µm sieve. It includes most familiar in-
vertebrate phyla, and is dominated particularly by polychaete worms, peracarid
crustaceans, and mollusks (Gage & Tyler 1991). Where there are sufficient data,
we also include the smaller meiofaunal elements. We do not consider reducing
environments (hydrothermal vents and seeps); Van Dover (2000) provides an ex-
cellent and extensive treatment of these remarkable deep-sea habitats and their
ecology.
This review developed from discussion by the Working Group on Deep-Sea
Biodiversity supported by the National Center of Ecological Analysis and Syn-
thesis (NCEAS). The Working Group’s main objectives are to integrate pattern
and scale of biodiversity in deep-sea benthic communities and to apply this to the
design of future research programs. Mechanisms of species coexistence on local
scales (<1 m2) are addressed for the Working Group by P. V. Snelgrove & C. R.
Smith (unpublished manuscript). Later reviews will cover the historical develop-
ment and global spread of deep-sea faunas and propose strategies for conservation
based on our current understanding of pattern and scale.
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

54 LEVIN ET AL.

PATTERNS AND CAUSES OF DIVERSITY

Boundary Constraints and Species Ranges
Deep-sea ecology has focused primarily on the role of local processes in regulating
community structure. Explaining species coexistence at small scales in the deep
sea remains a major challenge to ecological theory (Gage 1996; P. V. Snelgrove &
C. R. Smith, unpublished manuscript). It has proven difficult to extend mechanisms
of local community structure to regional patterns in any ecological system (Ricklefs
1987). An alternative approach is to view diversity within regions as a compos-
ite of overlapping species distributional ranges (MacArthur 1972, Stevens 1989).
Recently, there has been considerable interest in how physiographic and physi-
ological boundaries affect species ranges and consequently patterns of diversity
within geographic domains (Colwell & Hurtt 1994, Pineda & Caswell 1998, Willig
& Lyons 1998, Lees et al. 1999, Colwell & Lees 2000). Interestingly, a stochas-
tic placement of geographic ranges between boundaries will produce within the
bounded region a unimodal pattern of diversity (Colwell & Hurtt 1994) similar to
bathymetric gradients found in the deep sea (Figure 1). It is therefore possible to
construct null models to test whether observed species diversity gradients depart
significantly from those generated randomly by boundary constraints alone. If not,
there is little justification for invoking an environmental gradient as the cause.
Pineda (1993) and Pineda & Caswell (1998) applied the geometric constraints
model to patterns of deep-sea species diversity along depth gradients in the west-
ern North Atlantic. As noted earlier, the macrofauna of this region tend to show
unimodal patterns of diversity with depth. The frequency distribution of bathy-
metric ranges of marine species resembles latitudinal ranges found in terrestrial
taxa (Pineda 1993, Brown et al. 1996). Are unimodal diversity depth trends caused
solely by boundary constraints imposed by the sea-air interface and the abyssal
plain? Pineda & Caswell (1998) tested this possibility for gastropods and poly-
chaetes by using a null model that randomly placed species ranges between the
presumed upper and lower boundaries. Actual and randomly simulated diversities
were compared for three components of the unimodal pattern: position of peak
diversity along the depth gradients, peakedness of the diversity curve, and height
of the peak. The null model explained only one aspect of the diversity pattern for
each taxon, suggesting that whereas boundaries may affect the general shape of
diversity-depth patterns, many important features of the patterns were decidedly
nonrandom and attributable to other processes associated with depth.

Sediment Heterogeneity
The remarkable number of species that coexist within deep-sea assemblages is
surprising when one realizes that the overwhelming majority of these species are
deposit feeders that rely on organic detritus for food (Jumars & Eckman 1983).
How can so many species coexist while exploiting the same limited resource?
Because the primary food resource is detritus, the nature of the sediments ought
 10 Oct 2001
16:3
AR
ar142-03.tex
ar142-03.sgm
ARv2(2001/05/10)
P1: GJB

Figure 1 Schematic representation of the potential effects of boundaries on bathymetric patterns in

species diversity. The gray panels show ten species evenly and randomly distributed with depth, from
0 to 6000 m (to the left and to the right, respectively). In this representation, all species have a vertical
REGIONAL DIVERSITY IN THE DEEP SEA

range of 2400 m. Sampling these distributions every 400 m (gray circles) yields bathymetric patterns
in species diversity peaking at intermediate depths.
55
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

56 LEVIN ET AL.

to play an important role in structuring deep-sea communities. Numerous studies

have shown that the structure and composition of soft-sediment communities are
related to sediment characteristics (e.g., Petersen 1913, Sanders 1968, Rhoads
1974, Gray 1981), but the explanations for these relationships are varied and
remain controversial (Snelgrove & Butman 1994).
On the northwest Atlantic slope, spatiotemporal variation in species diversity
is correlated with the heterogeneity of sediment grain size across a wide variety of
spatial scales (Etter & Grassle 1992). Where sediment grain size is more varied,
more species coexist. This is consistent with the hypothesis that species partition
the sediments with respect to size. There is abundant evidence that deposit feeders
selectively ingest sediments of particular size classes (Taghon 1982, Whitlatch
1980, Wheatcroft & Jumars 1987, Self & Jumars 1988, Wheatcroft 1992). In
addition, several species exhibit interspecific differences in particle size preference
(Fenchel et al. 1975, Fenchel & Kofoed 1976, Whitlatch 1980), suggesting that
the sediments may be partitioned by size in some shallow-water communities.
However, this has not been tested for deep-sea species.
The strong correlation between species diversity and sediment heterogeneity
does not imply causality. The relationship may be spurious or reflect more im-
portant proximal factors. For example, the activities of more diverse communities
might actually increase sediment heterogeneity. Also, the size ranges used by Etter
& Grassle (1992) in the correlations were of disaggregated grains, and thus may
bear little resemblance to food diversity or to the aggregated sediments organisms
experience in situ. Definitive answers will require manipulative experiments that
tease apart the various potential mediating processes.

Productivity and Food Supply

Spatial gradients in productivity are widely believed to influence species diversity
(Waide et al. 1999). In terrestrial systems, the number of animal and plant species
often appear to vary unimodally with productivity or nutrient availability (e.g.,
Tilman 1982, Rosenzweig & Abramsky 1993, Rosenzweig 1995). Diversity within
a functional group or taxon increases from regions of low to moderate productivity,
and then declines toward regions of higher productivity. Similarly, the unimodal
species diversity-depth gradient that attends the exponential decrease in benthic
standing stock with depth has been attributed to productivity and its potential
mediation of biological interactions (Rex 1973, 1976, 1981).
The deep sea lacks in situ primary production, apart from chemoautotrophic
production in reducing environments such as hydrothermal vent and seep habitats
(Van Dover 2000). Most food material sinks from the euphotic zone to the ben-
thos in the form of small particles. Particulate organic-carbon (POC) flux can be
measured directly with sediment traps, and records integrating annual time scales
now exist for at least 37 sites in the open ocean (Lampitt & Antia 1997). Energy
availability in deep-sea benthic habitats is also positively correlated with, in order
of decreasing strength: 1. sediment-community respiration (Jahnke 1996, Berelson
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 57

et al. 1997), 2. rate of organic carbon burial within the sediment (Jahnke 1996),
3. benthic biomass and abundance (Rowe et al. 1991, Smith et al. 1997, Cosson
et al. 1997), and 4. overlying primary productivity (Deuser et al. 1990, Watts et al.
1992, Lampitt & Antia 1998). It is negatively correlated with depth of water through
which phytodetrital food sinks (Suess 1980, Martin et al. 1987). In some cases, the
concentrations of organic carbon and chlorophyll a in surface sediments may be
directly related to the flux of particulate organic carbon to the seafloor (Emerson
1985, Stephens et al. 1997). But these positive relationships are not necessarily
maintained, and may even reverse, over large spatial scales (Jahnke 1996).
A number of productivity gradients have been characterized in the deep sea.
There is a general decrease in POC flux (as well as sediment-community respira-
tion and benthic standing crop) from shelf depths to the abyssal plain (Smith &
Hinga 1983, Rowe et al. 1991). This depth-related decrease in productivity is well
substantiated for gradually sloping and well oxygenated margins such as in the
northwest Atlantic (Rowe et al. 1991), but becomes more complicated on irregular
slopes or where oxygen minimum zones intersect continental margin (Reimers
et al. 1992) and seamounts (Levin et al. 1991). A second productivity gradient is
the decrease in seafloor POC flux from the productive coastal zone to the open
ocean (Smith & Hinga 1983, Cosson et al. 1997). A third is the reduction in POC
flux, sediment-community respiration, and standing crop observed at the abyssal
seafloor from the Pacific equatorial zone and its associated nutrient upwelling
northward or southward into the oligotrophic central gyres (Smith et al. 1997).
Of these three productivity gradients, only that in the abyssal equatorial Pacific
varies independently of many other variables that may influence diversity such
as the hydrodynamic regime, bottom-water oxygen concentration, and physical
environmental stability.
As in other environments (Waide et al. 1999), the relationships between POC
flux or proxy variables for productivity and diversity are complicated and scale-
dependent. POC flux and sediment-community respiration drop roughly fivefold
at the abyssal seafloor from 0◦ N to 23◦ N latitude in the central Pacific Ocean (Smith
et al. 1997). Mean local nematode diversity, expressed as the number of species
rarefied to 51 individuals, exhibits a significant monotonic decline with decreasing
POC flux along this gradient [P. J. D. Lambshead, C. J. Brown et al., unpublished
manuscript) (Figure 2)]. Within this same general equatorial region, polychaete
species richness, normalized to 163 individuals, was weakly positively correlated
with total polychaete density, and by inference, POC flux (A. Glover, C. Smith,
et al., unpublished manuscript, Figure 2). Neither taxon exhibits a unimodal diver-
sity pattern comparable to those documented for terrestrial productivity gradients
(Rosenzweig 1995). However, it is unclear from a comparative standpoint whether
we are examining diversity along the ascending or descending part of the diversity-
productivity curve, and diversity has been assessed for only one meiofaunal taxon
and one macrofaunal group at the family level. Whereas productivity levels be-
neath central-gyre waters (e.g., 23◦ N) are extremely low, the productivity levels
attained beneath equatorial upwelling are still moderate by deep-sea standards, so
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

58 LEVIN ET AL.

Figure 2 Top: Mean local rarefaction species richness (normalized to 51 individu-

als) as a function of POC flux for nematode samples collected along the equatorial
Pacific POC flux gradient. Stations were located at water depths ranging from 4300
to 5000 m, 0◦ N 140◦ W; 2◦ N 140◦ W; 5◦ N 140◦ W; 9◦ N 140◦ W; and 23◦ N 158◦ W. The
correlation between POC flux and rarefaction diversity is high (r = 0.83) but not
statistically significant (P = 0.09) most likely due to small sample size (n = 5).
Data from Lambshead et al. (2001a, in press). Bottom: Rarefaction diversity for
macrofaunal polychaetes (normalized to 163 individuals) as a function of poly-
chaete numerical density in the equatorial Pacific. Polychaete density is used as a
proxy for POC flux and benthic productivity. Data are for pooled box-core samples
(3 to 47 per station) collected from depths of 4300–5000 m at the following locations:
0◦ N 140◦ W; 2◦ N 140◦ W; 5◦ N 140◦ W; 8◦ 270 N 150◦ 470 W; 12◦ 570 N 128◦ 190 W; and
14◦ 400 N 126◦ 250 W. For these data, r = 0.616, and P = 0.193. Data from A. Glover,
C. Smith, et al. (unpublished manuscript).
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 59

it is quite conceivable that we are examining diversity along the ascending portion
of a more general unimodal diversity-productivity curve.
Cosson-Sarradin et al. (1998) studied polychaete species diversity along a tran-
sect of three stations from the continental shelf (<100 m depth) to the abyss (4600 m)
off the west coast of central Africa. POC flux varied more than 16-fold from
the shelf site to the oligotrophic abyssal station. Local polychaete diversity, based
on the Shannon-Wiener index, exhibited a unimodal pattern as a function of POC
flux, with a peak in diversity occurring at the mid-slope site (1700 m). However,
this transect also represents a strong gradient in physical disturbance resulting
from current scour, and possibly physiological stress from organic loading on the
continental shelf. Paterson et al. (1998) studied local polychaete diversity (using
rarefaction) at six abyssal stations in the Atlantic and Pacific with presumed (but
unmeasured) differences in seafloor POC flux. Whereas polychaete abundance
covaried with putative POC flux, species diversity showed no obvious relation-
ship. Tietjen (1984, 1989) evaluated local nematode species diversity at six sites in
the deep northwest Atlantic. The highest nematode diversity occurred at the sites
presumed to have the highest POC flux and productivity.
On the North Carolina slope, the burial rate of organic carbon varies roughly
80-fold along the 850-m contour (Schaff et al. 1992, Blair et al. 1994, Levin et al.
1999). Schaff et al. (1992) and Levin et al. (1994b) examined local polychaete
diversity at three stations spaced equidistantly along this productivity gradient and
found reduced macrofaunal rarefaction diversity at the station with the highest
carbon burial rate. Similar patterns were observed for macrofaunal foraminifera
(>300 µm) (Gooday et al. 2001). Levin & Gage (1998) used sedimentary organic-
matter content as a proxy of food availability to examine diversity-productivity
relationships within existing data sets from the deep Indo-Pacific. Their study
included a broad range in sedimentary organic carbon (<0.5% to >6%), and,
presumably, habitat productivity. They found negative correlations between sedi-
mentary organic-carbon concentrations and the local diversity of total macrofauna
and polychaetes. Dominance in particular, was positively correlated with sediment
POC concentrations, suggesting that competitive interactions may shift along POC
gradients.
Sampling studies in the western North Atlantic provide, by far, the largest and
most geographically extensive database available to examine diversity-productivity
relationships. The unimodal diversity-depth patterns revealed by 1. rarefying large
qualitative samples (Rex 1981) and 2. tabulations of species number from intensive
quantitative sampling (Etter & Grassle 1992) parallel an exponential decline in
benthic standing stock with increased depth. As mentioned above, benthic biomass
and abundance are assumed to reflect the rate of nutrient input to the seafloor, but
this has not been directly measured (Smith et al. 1997). Rex (1973) suggested
that depressed diversity in the abyss was imposed by extremely low population
densities, essentially as a chronic Allee Effect. There seems to be general agreement
among ecologists that this kind of mechanism is responsible for low diversity at
very low productivity (Rosenzweig & Abramsky 1993). The drop in diversity
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

60 LEVIN ET AL.

in the upper bathyal zone, close to high surface production and terrestrial runoff,
could be driven by pulsed nutrient inputs and its potential to accelerate competitive
exclusion (Rex 1976, Huston 1979).
Diversity trends in deep-sea microfossils over geological time scales mirror
some of the spatial patterns described for modern productivity gradients and cor-
roborate the inference that energy supply can regulate species diversity. Fluc-
tuations in Shannon Wiener (H0 ) diversity of benthic ostracods correspond to
orbitally driven glaciation cycles over a 550,000 year period in the late Pliocene
in the North Atlantic (Cronin & Raymo 1997). Diversity was highest during inter-
glacial periods and declined during glacial advances, when surface production may
have been lower in the North Atlantic. Thomas & Gooday (1996) proposed that the
establishment of a Southern Hemisphere latitudinal gradient in foraminiferan di-
versity in the late Eocene around 38 my ago, which persists today (Culver & Buzas
1998, 2000), was created by an increase in the seasonally pulsed organic flux that
accompanied high latitude, Southern Hemisphere cooling, and the buildup of ice on
the Antarctic continent. This temporal diversity pattern was driven by fluctuating
dominance of opportunistic foraminifera that consume sinking phytodetritus.
The recent studies discussed above that use direct measures of productivity,
such as POC flux, sediment organic content, and carbon burial rates, appear to
give conflicting results; diversity varies positively, negatively, or unimodally with
productivity. However, these findings are not necessarily contradictory. The pro-
ductivity/input gradients examined may represent ranges of food availability that
occupy different segments of a unimodal diversity-productivity relationship. The
increase in diversity from areas of low to high POC flux in the abyssal equatorial
Pacific may correspond to the ascending part of the diversity-productivity curve,
much as proposed for the elevation of diversity from abyssal to the bathyal regions
in the western North Atlantic. The pattern reported for the eastern tropical At-
lantic spans a large vertical and horizontal geographic range that may experience
productivity inputs broad enough to realize the full unimodal trend. Regions that
include exceptionally high nutrient input, such as the North Carolina slope, may
fall on the descending part of the curve where diversity and productivity are neg-
atively correlated. Other observations suggest that depressed diversity in the deep
sea is associated with periodic high organic loading such as might be associated
with areas subject to intense upwelling (Sanders 1969), deposition from lateral
transport of nutrients (Blake & Hilbig 1994), benthic storms that expose reactive
sediments (Aller 1997), or bottom topography that concentrates food (Jumars &
Hessler 1976, Vetter & Dayton 1998). However, all involve multiple factors that
could affect diversity, and because the level and variation of production may co-
vary, it is difficult to separate their influence. In general, a unimodal relationship
between diversity and productivity at large scales in the deep sea is plausible,
but is not well substantiated. Just as in other environments (Waide et al. 1999), it
has proven difficult to accurately place available studies on a continuous produc-
tivity gradient and to identify clearly the underlying mechanisms through which
productivity influences species diversity.
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 61

At low levels of productivity, food limitation is thought to constrain the number

of species that can survive. Declines in diversity at higher productivity levels may
result from four possible causes. 1. Differential numerical responses among species
to nutrient loading (varying population growth rates) such that a small number of
opportunistic species take over. This will elevate dominance and lower diversity
measures that incorporate evenness. In many cases local species richness may
not change much (see Levin et al. 2000). 2. Faster rates of competitive exclusion
as envisioned by Huston (1979). 3. Increased variability in productivity, which
is often correlated with amount of productivity. This variability may bring about
declines due to demographic stochasticity. 4. Excess oxygen demand creating
hypoxia and leading to declines in both richness and evenness due to physiological
stress.

Bottom-Water Oxygen
Bottom-water oxygen concentrations in the deep oceans vary from near 0 to over
7 ml/l (Tyler 1995). Although much of the ocean has oxygen values near satura-
tion, there are extensive midwater regions where oxygen is depleted; these typically
occur between 100 and 1200 m depth. They are usually formed beneath highly pro-
ductive, upwelled waters by degradation of organic matter. The resulting hypoxic
zones, referred to as oxygen minimum zones (OMZs) and operationally defined as
areas where O2 < 0.5 ml/l, persist over geologic time. OMZs occur in much of the
eastern Pacific Ocean, in the Arabian Sea, and off West Africa (Kamykowski &
Zentara 1990). Certain deep basins (e.g., off southern California) and fjords also
contain permanently hypoxic or anoxic waters. Where these low oxygen regions
intercept the continental seabed, the benthos experiences either permanent hypoxia
or an oxygen gradient, which may fluctuate daily with internal tides (e.g., Levin
et al. 1991), interannually (e.g., with ENSO events, Arntz et al. 1991, Gallardo
1985), or over geologic time (den Dulk et al. 1998, Rogers 2000).
Sediments having oxygen-depleted overlying bottom water typically exhibit
substantially reduced macrofaunal diversity. Within OMZs the macrofauna exhibit
low species richness and very high dominance (Table 1). This pattern was first
reported on the West African margin off Walvis Bay by Sanders (1969), and has
since been observed in the eastern Pacific on a seamount off Mexico (Levin et al.
1991), on the Peru and Chile margins (Levin et al. unpublished data), and in the NW
Arabian Sea off Oman (Levin et al. 1997b, 2000). Among the macrofauna, many
molluscs, crustaceans, echinoderms, and cnidarians appear less tolerant of hypoxia
than other taxa (Diaz & Rosenberg 1995), although there are exceptions (Levin &
Gage 1998). No single taxon dominates the macrofauna of low oxygen settings,
although annelid species are often prevalent. At upper slope depths, within the least
oxygenated portions of OMZs, dominance of the most abundant species typically
ranges from 40% to 85% (Levin & Gage 1998) (Table 1). Less information is
available concerning the diversity responses to reduced oxygen concentrations of
bacteria, small protists (nanofauna), meiofauna, or megafauna. Smaller organisms
 10 Oct 2001

62
16:3
AR

TABLE 1 Community structure of macrofauna and foraminifera at hypoxic sites within oxygen minimum zones of the eastern Pacific and Indian
Oceans

Depth Bottom-water Dominant R1D

LEVIN ET AL.

Location (m) oxygen (ml/l) taxon (%) Es (100) H0 Density Reference

ar142-03.tex

Oman margin 400 0.13 Spionidae 63 5.6 1.45 (log2) 12,362 m−2 Levin et al. 2000
(Polychaeta)
Oman margin 412 0.13 Foraminifera 27 23.9 1.23 (log10) 2533 10 cm−2 Gooday et al.
>125 µm, 0–1 cm 2000
ar142-03.sgm

Santa Barbara Basin 550 0.06 Tubificidae 44 N/A 1.77 (log2) 1,691 m−2 Levin et al.
(Oligochaeta) unpublished
Santa Barbara Basin 590 0.05 Foraminifera 49 8.9 0.72 (log10) N/A Gooday et al.
610 0.15 >63 µm, 0–1 cm 25 21.4 1.10 (log10) 2000
Volcano 7, Mexico 750 0.08 Aplacophora 47 10.8 2.59 (log2) 1,854 m−2 Levin et al. 1991
(Mollusca)
ARv2(2001/05/10)

Peru margin 300 0.02 Tubificidae 83 4.7 0.84 (log2) 13,539 m−2 Levin et al.
(Oligochaeta) unpublished
N. Chile margin 300 0.26 Dorvilleidae 73 N/A N/A 1,834 m−2 Levin et al.
P1: GJB

(Iquique) (Polychaeta) unpublished

Central Chile 364 0.52 Amphinomidae 31 17.6 2.46 (log2) 14,206 m−2 Gallardo et al.
margin (Polychaeta) unpublished
(Concepcion)
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 63

living entirely within the sediments and with no access to the surface may be
confined to hypoxic or even anoxic pore-waters, even when the overlying bottom
water is oxic (Corliss & Emerson 1990, Jorissen et al. 1998). Yet foraminifera and
a variety of larger metazoans (polychaetes, crustaceans, molluscs, echinoderms)
all display abundance peaks close to OMZ boundaries (Mullins et al. 1985, Levin
et al. 1991).
Where a range of faunal size groups have been compared, larger taxa (megafauna
and macrofauna) exhibit density reductions within the most hypoxic portions of
OMZs (Wishner et al. 1990, Levin et al. 1991, 2000) that are not evident in bacteria
and metazoan meiofauna (Levin et al. 1991, Cook et al. 2000, Neira et al. in press).
Metazoan meiofauna of the eastern Pacific OMZ exhibit loss of harpacticoid cope-
pods, and dominance by nematodes (Levin et al. 1991, Neira et al. 2001), but
diversities have not been recorded. Foraminiferal assemblages characterized by
high dominance and few species are reported from modern, low-oxygen basins in
the eastern Pacific and elsewhere (Phleger & Soutar 1973, Douglas et al. 1980,
Hermelin & Schimmield 1990, SenGupta & Machain-Castillo 1993, Bernhard et al.
1997). The fossil record yields evidence for similar responses among ancient deep-
sea faunas (e.g., den Dulk et al. 1998, Jorissen 1999). However, precise diversity
data are rarely reported. At a site within the NW Arabian Sea OMZ, foraminifera
exhibited reduced species richness and elevated dominance compared with assem-
blages from a deep site below the OMZ (Gooday et al. 2000). These low-oxygen
faunas consisted largely of small calcareous forms, a trend observed in the OMZ
off Peru as well (A. Rathburn, unpublished manuscript). In contrast to fully oxic
deep-sea sites, monothalamous and other delicate agglutinated and allogromiid
taxa are rare compared to calcareous forms within OMZs (Gooday et al. 2000).
These limited deep-sea observations are consistent with field (Josefson &
Widbom 1988, Murrell & Fleeger 1989, Radziejewska & Maslowski 1997, Luth &
Luth 1997) and experimental (Moodley et al. 1997) studies conducted in shallow
water that suggest that meiofauna, particularly nematodes and foraminifera, are
less affected by hypoxia than the macrofauna, at least at higher taxonomic lev-
els. There are, however, shallow-water examples in which nematode diversity is
reduced by severe hypoxia (Keller 1986, Austin &Widbom 1991).
Historical records support a role for oxygenation in control of deep-sea diversity.
In the northern Arabian Sea, foraminiferal diversity appears to have been strongly
influenced during the Quaternary by changes in the flux and quality of organic
matter to the seafloor, which led to variations in the thickness and intensity of
the oxygen minimum zone (Hermelin & Shimmield 1995, den Dulk 2000). den
Dulk et al. (1998) studied a 120,000 years-long Quaternary record from the
Pakistan margin in the Northern Arabian Sea. Two foraminiferal assemblages
were recognized. A low diversity assemblage with high dominance recurred every
23,000 years, possibly as a result of enhanced summer surface productivity and
therefore intensified OMZ development, linked to the precessional component of
orbital forcing. A more sustained period of low diversity occurred under glacial
conditions, perhaps reflecting a strengthening of the NE monsoon, leading to higher
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

64 LEVIN ET AL.

winter productivity. In a detailed multiproxy study of shorter cores (spanning the

last 30,000 y) from the same margin, von Rad et al. (1999) detected a switch from
low to high foraminiferal diversity on the Pakistan margin during brief, late Qua-
ternary to early Holocene climatic oscillations (Younger Dryas, Heinrich events
1 and 2) when surface productivity was unusually low and presumably the OMZ
intensity diminished. Evidence for a close coupling between foraminiferal ben-
thic community structure (including dominance) and bottom-water oxygenation
is seen in the Santa Barbara Basin on even shorter time scales. There, fluctuations
on decadal to millennial time scales are associated with major climate oscillations
that change thermolhine circulation and ventilation (Cannariato et al. 1999).
During the Late Cretaceous, anoxia sometimes occurred on a much larger scale
than in modern oceans, and caused widespread benthic mortality (Rogers 2000).
Most of these events are far too ancient to have any residual effect on modern di-
versity patterns. A possible exception occurs in the eastern Mediterranean, which
was subject to repeated episodes of basin-wide, deep-water anoxia during the Late
Quaternary and most recently around 6000 years ago. Anoxic episodes, repre-
sented in the sedimentary record by clearly defined, dark horizons (sapropels),
had a profound effect on the benthic foraminiferal faunas (Jorissen 1999). Dis-
tinct regional-scale differences in foraminiferal species richness and composition
between the eastern and western Mediterranean basins may reflect these major
disturbances in the recent geological past (de Rijk et al. 1999).
There is a difficulty with interpreting diversity responses to bottom-water oxy-
gen conditions that derives from the tight linkage between bottom-water oxygen
concentration and organic matter inputs in the deep sea (Levin & Gage 1998).
Multiple regression analyses of a large macrofaunal data set from the Indian and
eastern Pacific Oceans suggest that oxygen exerts a strong effect on species rich-
ness, although organic matter availability (evaluated from sediment POC) has a
greater influence on dominance (Levin & Gage 1998). Together, these factors
lower diversity within OMZs. Although food availability may affect diversity over
a broad range of oxygen values, significant reduction of macrofaunal species rich-
ness by low oxygen may not occur until concentrations fall below 0.4 or 0.3 ml l−1
(Figure 3); this value may be even lower for annelids, the taxon most tolerant to
hypoxia. Sulfides, which are toxic at high concentrations to most marine organisms
(Bagarinao 1992), are associated with high inputs of labile organic matter, and also
vary inversely with oxygen concentration. Their role in reducing diversity within
OMZs has yet to be explored.
A major challenge is unraveling the relative importance of physiological stress
and biotic interactions in creating deep-sea responses to low oxygen. Although
stress may cause loss of species and dominate diversity responses at the lowest
oxygen levels, facilitation may also be important. Animal structures such as tubes
or burrows that conduct seawater into the sediments, or irrigation activities that
mix solutes may enhance oxygen availability and dilute sulfides, facilitating colo-
nization by other taxa. At higher oxygen levels, competitive abilities and predation
may regulate diversity.
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 65

Figure 3 Macrofaunal Rank 1 dominance (% of the total accounted for by the top ranked
species) and species richness (E[S100]) plotted as a function of bottom-water oxygen concen-
tration for bathyal stations within and beneath oxygen minimum zones in the eastern Pacific
and northern Indian Oceans.

Deep-Sea Currents
Near-bottom flow rates in the deep ocean are typically a few cm/sec, too weak
to erode the seabed (Munk 1970, Tyler 1995). The ocean floor is not uniformly
quiescent, however, and numerous areas are subject to currents strong enough
to erode and transport sediments and disturb soft-bottom communities on scales
of 10s or 100s of km, (Heezen & Hollister 1971, Hollister et al. 1984). Episodic
benthic storms characterize areas beneath western boundary currents where surface
kinetic energy is transmitted through the water column to the sea floor; near-bottom
currents reach speeds of 15–40 cm/sec and persist for several days (Gross &
Williams 1991, Hollister & Nowell 1991, Aller 1997, Weatherley & Kelley 1985).
First described in the HEBBLE area on the Nova Scotian continental rise, benthic
storms occur along the western margins of the North and South Atlantic, around
South Africa, and in regions around the Antarctic continent (Rowe & Menzies
1968, Flood & Shor 1988, Hollister & Nowell 1991, Richardson et al. 1993).
Similar transient, high-energy episodes have been reported in abyssal areas distant
from continental margins, for example in the NE Atlantic (Klein 1988) and NE
Tropical Pacific (Koutar & Sokov 1994).
There are many other sources of strong current activity in the deep sea. Thermo-
haline-driven bottom currents transport vast amounts of sediment around the deep
ocean, depositing them as huge sediment drifts (contourites), for example, west
of Scotland (Hollister et al. 1984, Stow & Holbrook 1984, Kidd & Hill 1986,
Stow & Faugères 1993, Viana et al. 1998). In the Rockall Trough, current speeds
>15 cm/sec are frequently recorded by current meters moored on the upper slope
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

66 LEVIN ET AL.

above 1000 m (Paterson & Lambshead 1995) with peak flows reaching 48 cm/sec
(Viana et al. 1998). Along continental margins, more localized hydrographic phe-
nomena such as internal tides, water column instability, and storm-driven eddies
may create strong, erosional currents on the upper slope (Dickson & McCave 1986,
Pingree & New 1989, Rice et al. 1990, Gage 1997). Complex interactions between
steeply sloping topographic features and local hydrography create regions of in-
tensified near-bottom flow on seamounts (Noble & Mullineaux 1989), as well as
in canyons (Shepard et al. 1979, Gage et al. 1995, Gage 1997) where turbidity
currents also may be active (Jorissen et al. 1994).
Near-bottom currents are among the agents of disturbance that can modify the
structure and composition of benthic faunas (Hall 1994). As Levin et al. (1994a)
emphasize, however, the critical factor may be sediment mobility, which depends
on sedimentary characteristics as well as the current flow itself. Gage et al. (1995)
and Gage (1997) demonstrate with rarefaction curves, rank abundance plots, and
univariate diversity indices that polychaete diversity is highest at tranquil sites on
the Tagus Abyssal Plain and in the central North Pacific, lower in the Rockall
Trough and in the hydrodynamically active Sebutal Canyon, and much reduced
at the HEBBLE site. Bivalve diversity exhibits a similar trend, although the dif-
ferences are much less pronounced. Foraminifera at the main (Mid) HEBBLE
site were significantly less diverse (H0 ) than at the relatively tranquil shallow site
(Kaminski 1985). Allen & Sanders (1996) attribute anomalously low protobranch
bivalve mollusc diversity in the North American Basin to possible benthic storms.
Paterson & Lambshead (1995) observed a strong linear relationship between the
frequency of current velocities >15 m/sec in the Rockall Trough and the equitabil-
ity statistic V for polychaetes. Negative values of V, indicating high dominance,
were associated with high current speeds on the upper slope in the Rockall Trough.
As at HEBBLE, polychaete taxa generally considered to be opportunists predom-
inated in parts of the Rockall Trough where physical disturbance was high. On
Fieberling Guyot, Shannon-Wiener (H0 ) values for macrofauna were very simi-
lar at contrasting sites with daily and infrequent, episodic sediment transport, but
species richness measured by rarefaction was somewhat higher at the more stable
SPR site (Levin et al. 1994a, see also Levin & DiBacco 1995).
In the case of meiofaunal taxa, strong erosive currents do not necessarily depress
diversity. Thistle (1983) found no difference in harpacticoid copepod diversity in
samples from the energetic HEBBLE site and the tranquil San Diego Trough. Con-
trary to his initial expectation, Thistle (1998) observed that harpacticoid copepod
diversity was actually higher at the Fieberling Guyot WSS site, where the Globige-
rina sands are mobile on a daily basis, than at the more stable SPR site. Nematode
diversity was very similar at three stations (545 m, 835 m, 1474 m depth) in the
Rockall Trough and at 1050 m in the San Diego Trough, despite differences in
current velocities both between these two basins and with depth within the Rock-
all Trough (Lambshead et al. 1994). Severe benthic storms do appear, however, to
depress nematode species richness [measured as E(S51)] at the HEBBLE site com-
pared with values from relatively tranquil localities on the Porcupine and Hatteras
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 67

Abyssal Plains (Lambshead et al. 2001). This may reflect a greater intensity of
current disturbance at HEBBLE than in the Rockall Trough.
The observations reviewed above suggest that currents can modify benthic di-
versity both locally and regionally, although the mechanisms involved are not
well understood. The effects can be either positive or negative, and there are in-
dications that macrofauna are more strongly impacted than meiofauna. Possible
mechanisms can be divided into those having a direct impact and those having
an indirect impact on diversity. Strong currents may depress diversity directly by
eroding surficial sediments and carrying away the animals living in them (Aller
1997). Observations at the HEBBLE site (Thistle et al. 1985, 1991, Lambshead
et al. 2001) and the WSS site subject to daily sediment transport on Fieberling
Guyot (Levin & DiBacco 1995) suggest that episodic disturbance by erosive flow
creates repeated opportunities for recolonization. Constant reworking of the sedi-
ment ensures that the benthic fauna remains in an early successional state, favoring
opportunists, and keeping diversity low.
Near-bottom currents have the potential to impact benthic faunal diversity in a
variety of indirect ways. Moderate currents can enhance the food supply by deliv-
ering organic matter and stimulating bacterial production (Thistle et al. 1985, Aller
1989). As long as these inputs are not excessive, they should lead to an increase in
both abundance and diversity of macro- and meiofauna. Currents may also entrain
larval and subadult organisms, allowing animals to colonize disturbed patches of
sediment and enhance local diversity. Moderate flow conditions may potentially
increase sediment heterogeneity by creating sedimentary structures and by con-
centrating organic matter in localized patches. At the White Sand Swale (WSS)
site on Fieberling Guyot, active ripples generate heterogeneity by successively
burying and exposing organic matter accumulated in the troughs. This process
may explain why harpacticoid copepod diversity was higher at WSS than at the
more stable Sea Pen Rim site (Thistle 1998). When flow velocities were enhanced
experimentally at WSS for a 6-week period, however, there was no detectable
change in harpacticoid or total macrofaunal diversity or equitability, possibly be-
cause the assemblages were already adapted to strong erosive flow (Levin et al.
1994a, Thistle & Levin 1998).
As indicated above, near-bottom flows in excess of 20–25 cm/s can poten-
tially depress diversity by eroding epifaunal species, but they may also impact
diversity indirectly by smoothing out and reducing physical heterogeneity. On a
regional scale, erosive bottom currents will tend to homogenize the fauna by dis-
persing juveniles and subadults. Given the large areas of seafloor swept by erosive
currents (Hollister et al. 1994, Hollister & Nowell 1991), the impact of hydrodyna-
mics on regional deep-sea species diversity and biogeography may be considerable.
Hydrodynamic processes are probably involved in the creation of larger habitat
patches that increase seafloor heterogeneity. On the upper slope around the NW
European continental margin, interactions between the seafloor and internal tides
and waves are apparently linked to conspicuous concentrations of hexactinellid
sponges (Rice et al. 1990) and corals (Fredrickson et al. 1992). In the Porcupine
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

68 LEVIN ET AL.

Seabight (1000–1300 m water depth), sponges and spicule mats derived from them
enhance the abundance and modify the taxonomic composition of the macrofauna
(Bett & Rice 1992). Both corals and sponges have numerous associated organisms
(Klitgaard 1995, Jensen & Frederiksen 1992, Bartel & Gutt 1992). However, at
least in the NE Atlantic, most of the associated organisms are also present in the
background community (Klitgaard 1995, Jensen & Frederiksen 1992), suggesting
that regional diversity may not be increased substantially by the presence of these
large, habitat-creating organisms.

Catastrophic Disturbance
Over geological time scales, continental margin sediments have been disrupted by
gravity-driven mass movements, including slumps, slides, debris flows, and tur-
bidity currents (Masson et al. 1994, 1996). At their distal extremities, disturbance
by turbidity currents is probably similar to that caused by severe erosive currents
such as benthic storms. In the Atlantic Ocean, mass movements are well docu-
mented off NW Europe, NW Africa, southern Africa, the United States between
New York and Cape Hatteras, and Brazil (Emery & Uchupi 1984). Off NW Africa
debris flows, particularly the unusually large Canary and Saharan flows (Jacobi &
Hayes 1982, Simms et al. 1991, Embley 1976, Masson et al. 1994, Masson 1996),
have displaced 600 km3 of sediment on the upper continental rise. Elsewhere in the
NE Atlantic, the mid-Norwegian margin (the “Storegga” area; 62◦ N) is notable for
a series of enormous slides, the most recent of which occurred about 7000 years
ago (Bugge et al. 1987). These slides, and associated debris flows and turbid-
ity currents, have transported 6000 km3 of sediment from the shelf to depths of
3500 m over horizontal distances of 800 km. The flanks of volcanic islands such
as the Hawaiian Islands in the Pacific and the Canary Islands in the Atlantic are
prone to catastrophic collapses that give rise to massive avalanche deposits on the
adjacent deep-sea floor (Lipman et al. 1988, Moore et al. 1989, Cochonat et al.
1990, Masson 1996). These mass movements of sediments are often associated
with turbidite deposition, an important mechanism in the formation of abyssal
plains adjacent to continents. The best known modern turbidity flow occurred
in 1929 when an earthquake on Grand Banks, Newfoundland, triggered a sed-
iment slump that developed into a turbidity current extending 800 km from its
source across the abyssal plain. More recently, Thunnell et al. (1999) provided the
first direct, real-time documentation of earthquake-generated suspended sediment
flows.
Although major mass movements of sediment must devastate the benthic fauna,
most of these events occurred thousands of years ago and are unlikely to affect
modern faunas directly. Indirect effects, however, may persist for much longer.
Debris avalanches and similar chaotic deposits will introduce long-lasting phys-
ical heterogeneity on spatial scales up to kilometers (Masson 1996). Turbidites
have granulometric characteristics and total organic carbon (TOC) values that dif-
fer from those of pelagic sediments (Huggett 1987) and can potentially influence
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 69

diversity. There is evidence for such an effect at a location on the Madeira Abyssal
Plain (MAP; 4950 m water depth) that was swept by a turbidite about 1000 years
ago. Here, polychaetes exhibit lower abundance, much lower species richness [ex-
pressed as E(Sn) and per unit area], and higher dominance compared to other
abyssal NE Atlantic sites, including the equally oligotrophic EUMELI site, where
sedimentation is entirely pelagic (Glover et al. 2001). The sedimentary character-
istics of the turbidite deposit, which is overlain by only a thin veneer of pelagic
sediment, may have confined recolonization to a relatively small suite of oppor-
tunistic polychaete species. Nematodes, which lack a dispersive larval phase, also
exhibit lower diversity [E(S51)] at the MAP than at nonturbidite sites on the Porcu-
pine and Hatteras Abyssal Plains (Lambshead et al. 2001b in press). Foraminiferal
diversity, on the other hand, is not noticeably depressed at the MAP (Gooday 1996).
There is no evidence for low nematode diversity at a site in the Venezuela Basin
(5054 m depth) subject to periodic turbidite impacts. This is probably because the
sedimentation rate (7.2 cm · ky−1) is much higher and the turbidites are older in the
Venezuela Basin than in the MAP (Lambshead et al. 2001b in press).
Submarine and subaerial volcanic eruptions may impact deep-sea benthic fau-
nas directly, through deposition of lava or ash layers (Cita & Podenzani 1980), and
indirectly through changes in climate and water column stratification (Genin et al.
1995). Fossil foraminiferal evidence from the eastern Mediterranean suggests that
the benthic fauna was obliterated by an ash deposit 35,000 years ago but that a very
similar assemblage reestablished rapidly (Cita & Podenzani 1980). We are aware
of only one species-level study of the effect of a modern ashfall in the deep sea
on a benthic taxon (Hess & Kuhnt 1996). During 1991, an ash layer >2 cm thick
was deposited over 36,000 km2 in the South China Sea following the eruption of
Mt. Pinatubo, Philippines. Rarefaction curves based on data from Hess & Kuhnt
(1996) indicate that foraminiferal assemblages were severely affected by a 6 cm
ash deposit at 2503–2506 m; most foraminifera died, although epifaunal speci-
mens found below the ash layer may have survived for some time in a starved,
quiescent state. After almost three years, the ash layer had been recolonized by
a low-diversity assemblage of infaunal opportunists. In a 2-cm deep ash deposit
at 4226 m, however, many mobile infaunal taxa survived and species richness
decreased only slightly.

MEASURING THE LEVEL OF DEEP-SEA SPECIES DIVERSITY

Very basic questions remain about how diversity should be measured and inter-
preted. A wide variety of diversity measures exist that incorporate both richness
and evenness but differ in how these components of diversity influence the mag-
nitude of the index (Magurran 1988). Because most diversity indices are sensitive
to both evenness and richness, differences can reflect changes in either or both;
changes in evenness should not be interpreted as changes in richness. The choice
of an index depends on the nature of the question and the type of data available. In
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

70 LEVIN ET AL.

most cases, it will be useful to use a variety of indices that together provide greater
insight into how and why diversity varies than does a single index.
The most widely used measure of diversity in deep-sea ecology is E(Sn), the Ex-
pected Number of Species (Sanders 1968, Hurlbert 1971). This index estimates the
number of species in samples normalized (rarefied) to successively smaller sample
sizes, providing an interpolated curve of the relationship between the number of
species (S) and the number of individuals (N) based on resampling the relative
abundance distribution of an actual sample. It is sensitive to both evenness and
richness (Gage & May 1993), and was originally designed to allow comparisons of
nonquantitative samples taken by epibenthic sleds. The sleds are towed from a sur-
face vessel for a variable distance of about a kilometer and are apt to bounce along
the bottom. The only way to compare diversity among such samples is by normaliz-
ing them to the same number of individuals, typically the lowest common number of
individuals. Quantitative coring devices (Hessler & Jumars 1974) sample a precise
area of bottom and accurately measure faunal density. Should quantitative samples
be rarefied? This is problematic if the habitats sampled differ greatly in density,
which they often do. For example, a boxcore may yield 10,000 individuals from a
continental shelf habitat, but fewer than 100 at greater bathyal depths. Normalizing
the two samples to 100 individuals effectively compares the number of species at
9 cm2 on the shelf to 900 cm2 in the deep sea. This procedure will mask the differ-
ences in density between sites, which may be critical for understanding why diver-
sity differs. Because no shallow-water samples may yield as few as 100 individuals,
the number of species is predicted from the lower portion of the species-individuals
curve where the slope is steepest and the error for estimating S is largest. Finally,
because rarefaction assumes a random distribution between S and N, replicate sam-
ples from a site are often pooled (e.g., Grassle & Maciolek 1992, Etter & Grassle
1992). This obscures the actual relationship between S and N at the sampling
scale, and how differences in heterogeneity among sites may contribute to varia-
tion in diversity. Rarefaction has been an important and useful tool for quantifying
and comparing community structure, but for a comprehensive understanding of
regional patterns of diversity it should be used in combination with other measures.
A better approach for comparing regional species richness with quantitative
samples is species accumulation curves with randomly pooled samples (Etter &
Mullineaux 2000). This requires replicate samples at each location, but maintains
the relationship between S and N, which preserves differences in heterogeneity
and allows one to compare diversity based on the number of individuals or area
sampled. Most importantly, it can be used to estimate the asymptotic S (the num-
ber of species in the community if it were completely sampled) and how well the
community has been sampled (the position on the species accumulation curve),
both of which are crucial for comparing and interpreting differences in S. Random-
ized species accumulation curves will typically be lower than rarefaction curves
(Figure 4) because they maintain the nonrandom distribution of S and N at the scale
of the sampling device. The difference between the two curves reflects the amount
of heterogeneity at the scale of the sampler. Two programs that allow one easily to
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 71

Figure 4 Randomized species accumulation and rarefaction curves for Station 3 on

Georges Bank (GB3, 100 m, 40◦ 53.70 N, 66◦ 46.50 W) and Station 3 from the North
ACSAR (Atlantic Continental Slope and Rise) (North 3, 1350 m, 41◦ 1.400 N, 66◦ 20.200 W).
The randomized species accumulation curves are the average cumulative number of species
from 50 randomizations of pooling successively larger numbers of replicates at each station.
Data are from Maciolek et al. (1985) and Maciolek et al. (1987b).

compute randomized species accumulation curves are Estimate S (Colwell 1997)

and Rosenzweig’s (1995).
Gray (1994) and Gray et al. (1997) provide a recent example of the prob-
lems encountered when comparing diversity among habitats. Benthic samples
from a variety of shallow-water habitats (<200 m) were compared to those col-
lected from the deep northwest Atlantic by Grassle & Maciolek (1992) to argue
that shallow-water and deep-sea communities may be quite similar in diversity.
However, these comparisons were confounded by the shallow-water samples being
collected from broader geographic areas (Gray 1994) at different latitudes and by
using different sieve sizes (1 mm rather than 300 µm), both of which can influence
diversity (Bachelet 1990, Warwick & Clarke 1996, Rex et al. 1993). More impor-
tantly, because estimates of species richness are highly sensitive to sampling effort
(May 1975, Colwell & Coddington 1994, Rozenzweig 1995), comparisons should
be made only at the asymptote of a species accumulation curve. Although the
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

72 LEVIN ET AL.

shallow-water samples of Gray et al. (1997) approach an asymptote, the deep-sea

samples show no sign of leveling off (Grassle & Maciolek 1992). The problem
of assessing diversity when no asymptote has been reached is often overlooked
in spatial and temporal comparative studies of richness. Rarefying the samples
to a common number of individuals does not eliminate the sensitivity of richness
estimates to sampling effort. For example, samples of 100 individuals in one com-
munity may contain 90% of the species, although another may contain only 10%.
The only way to overcome this potential artifact is to restrict comparisons of S to
the asymptotic S.
To reexamine the question of whether deep-sea habitats support more species
than shallow-water habitats, we can compare diversity between a series of deep-
water samples collected off the coast of Massachusetts [the North data collected as
part of the Atlantic Continental Slope and Rise Study (ACSAR)] (Maciolek et al.
1987b) to a very similar set of samples collected from the nearby shallow waters of
Georges Bank (Maciolek et al. 1985). This comparison was selected because the
shallow and deep samples are geographically adjacent (both just off the coast of
Massachusetts), were collected in a similar way, were sorted by sieves of the same
size, have consistent taxonomy (done by the same individuals), and represent the
entire macrofaunal community. The Georges Bank samples were collected with
0.04 m2 Ekman grabs from 38 to 167 m. The ACSAR samples were collected
with 0.25 m2 boxcores from 250 to 2500 m, but only the center 0.09 m2 was
used for community analysis. No other shallow- and deep-water data bases are as
comparable.
For the deep-water samples (those >200 m), on average 278 species coexist
in an area of 1 m2 and the expected number of species was 156 when normalized
to 1000 individuals (ES1000). In contrast, the shallow-water samples produced an
average of only 165 species m−2 and 68.8 species per 1000 individuals (Table 2).

TABLE 2 Average measures of diversity and density for the Georges Bank
and the ACSAR samples. E(S1000) are Hurlbert’s (1971) expected number of
species normalized to 1000 individuals. ACSAR North (130 m) are averages
based only on samples collected between 1220–1350 m from the North. Data
are from Maciolek et al. (1987b) and Maciolek-Blake et al. (1985)

Georges ASCAR ASCAR

Bank North North (130 m)
Samples 1149 191 63
Depth Range (m) 38–167 250–2180 1220–1350
Individuals 680,600 95,140 27,906
Species m−2 165 278 319
E(S1000) 68.8 156 188
Shannon Wiener Index (log2) 4.09 5.59 6.70
Species Richness 680 952 599
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 73

This comparison of averages is conservative because the deep-sea samples span

a 2000-m depth gradient and include relatively shallow depths that are lower in
diversity. If the deep-sea samples are restricted to a 130-m depth interval (similar
to the range of Georges Bank samples) centered at 1300 m, the differences are
even more pronounced (Table 2). Randomized species accumulation curves based
on either individuals or area also indicate that diversity is greater in the deep-sea
samples (Figure 5). Asymptotic S values estimated from the species accumulation
curves at each site using Chao1 (Colwell & Coddington 1994) indicate that richness
varies unimodally with depth (Figure 6), as suggested previously based on rarefied
qualitative (Rex 1981, 1983) and quantitative samples (Etter & Grassle 1992).
Pooling the ACSAR samples and Georges Bank samples by depth produces
estimates of species richness that are quite similar (Table 2), as Gray [(1994), Gray
et al. (1997)] suggested for his studies. In fact, when pooled over 130 m, rich-
ness appears to be greater on Georges Bank, but the species accumulation curves
(Figure 7) clearly show that the shallow-water estimates are near an asymptote

Figure 5 Randomized species accumulation curves for all North (ACSAR) and Georges
Bank stations as a function of (A) area sampled and (B) numbers of individuals. The Georges
Bank curves lack symbols while the North stations have symbols. Data are from Maciolek
et al. (1985) and Maciolek et al. (1987b).
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

74 LEVIN ET AL.

Figure 5 (Continued )

although the deep-sea samples are not. This demonstrates the fallacy of comparing
richness among locations while ignoring how well each community has been
sampled (location on species accumulation curve). When comparisons are re-
stricted to contiguous geographic regions, sampled in similar ways, as are the
Georges Bank and North (ACSAR) data, diversity (richness, species/area, species/
individual) is much lower on the shelf than at bathyal depths. Although the diver-
sity of shallow-water communities in some parts of the World Ocean may equal
or exceed that of the deep sea, if comparisons are not controlled geographically,
taxonomically, by habitat, and by sampling methods, they can tell us little about
what forces are shaping species diversity in marine ecosystems.

INTERRELATIONSHIPS OF LOCAL
AND REGIONAL DIVERSITY
Spatiotemporal variation in deep-sea species diversity represents an integration of
ecological and evolutionary processes that operate at different spatial and tempo-
ral scales (Figure 8). Smaller-scale processes are embedded hierarchically within
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 75

Figure 6 Asymptotic species richness estimated by Chao1 (Estimate S [Colwell 1997])

as a function of depth for Georges Bank and ACSAR North stations. Error bars are 95%
confidence limits. Data are from Maciolek et al. (1985) and Maciolek et al. (1987b).

larger-scale processes, and tend to occur at faster rates. Species within a local
assemblage (1–10 m2) are controlled by small-scale processes involving resource
partitioning, competitive exclusion, predation, facilitation, physical disturbance,
recruitment, and physiological tolerances, all of which are mediated by the nature
and degree of heterogeneity. How these processes might regulate local diversity has
been reviewed elsewhere (Etter & Mullineaux 2000, P. V. Snelgrove & C. R. Smith,
unpublished manuscript), but they will tend to occur on much shorter times scales
than landscape or regional-level changes examined here. At regional scales of 100s
to 1000s of m, several environmental gradients, dispersal, metapopulation dynam-
ics, and gradients in habitat heterogeneity are likely to be important (Figure 8).
We suggest that the environmental gradients discussed in this paper essentially
control variation in local diversity by accelerating or decelerating local processes
(e.g., patch dynamics). We describe in more detail below how each gradient may
influence these local processes. Dispersal among patches will also be important at
landscape or regional scales because it determines the potential for membership
in local assemblages and plays a critical role in metapopulation dynamics and
species persistence at local and larger scales (Caswell & Cohen 1993, Caswell &
Etter 1999). For instance, low dispersal rates can decrease the intensity of biolog-
ical interactions (Menge & Sutherland 1987) and potentially foster coexistence
through recruitment limitation (Tilman 1994, Hubbell et al. 1999). At regional
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

76 LEVIN ET AL.

Figure 7 Species accumulation curves for pooled samples over a depth range of
130 m for Georges Bank and ACSAR North studies. The North samples were pooled from
stations between 1220–1350 m to compare to a similar depth span in the Georges Bank study.
All Georges Bank samples were used (38–167 m). See Table 2 for other diversity metrics
comparing these pooled samples.

scales, many species will exist as metapopulations; their presence locally or re-
gionally will be determined by local dynamics, dispersal among patches, and the
degree of reproductive asynchrony among populations (e.g., Hanski 1998).Varia-
tion in local diversity within regions is also likely to be influenced by the number
and different types of habitats at intermediate scales (e.g., 1–10 km) through a
mass effect (Shmida & Wilson 1985). We know little about how mass effects
contribute to local diversity, but the large number of singletons (species repre-
sented by a single individual) within natural deep-sea assemblages suggests that
it may be extensive.
It is clear from the foregoing discussion that local diversity varies regionally
along environmental gradients in the deep sea. How the numerous environmental
factors act to control local diversity is not well understood, but we have developed
a conceptual model of how they may directly and indirectly regulate the number of
species within communities (Figure 9). These environmental factors do not work
in isolation; they are often very interdependent such that changes in one precipitate
changes in others, producing cascade effects, all of which modify the rates of local
processes.
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 77

Figure 8 Processes regulating species diversity at local, regional, and global scales. Each
box represents one scale and is embedded within the larger scales.
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

78 LEVIN ET AL.

Figure 9 Conceptual model indicating direct and indirect effects of various envi-
ronmental factors on species richness of local communities.

Near-bottom flow has pervasive effects; it can influence the number of species
directly as well as indirectly through modification of other environmental fac-
tors (Gage 1997). Near-bottom current speed can directly control species number
through larval supply, and at sufficiently high speeds will erode the sediments and
many of the organisms on or within them. As current speeds vary, they can also
indirectly influence diversity by changing food supply (POM flux and bacterial
production), sediment characteristics, oxygen levels above and within the seabed,
spatial heterogeneity, and levels of biotic and abiotic disturbance (Figure 9). Each
of these factors can also affect the number of species within communities directly
as well as indirectly.
Hydrodynamic influences on diversity can be extensive and involve numerous
paths of different length and intensity. A unimodal relationship between diversity
and flow strength is predicted (Figure 10), but has not been well documented.
Low flows may be associated with depositional regimes and reduced sediment
heterogeneity. Very strong flows are expected to reduce effectiveness of predators,
and eliminate heterogeneity in particle size and surface features such as ripples.
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 79

Figure 10 Patterns of diversity change along environmental gradients in the deep sea
and biotic interactions hypothesized to be responsible for generating diversity patterns.
(A) Productivity, (B) Flow, (C) Bottom-water oxygen concentration, (D) Sediment
heterogeneity, (E) Biotic disturbance.
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

80 LEVIN ET AL.

Disturbance resulting from substrate instability or the stress of maintaining feeding

activities and dwelling position are likely to reduce species richness and evenness
when currents are sufficiently strong to resuspend sediments, although these ef-
fects may be ameliorated by large or stabilizing biogenic structures. Intermediate
or variable flows are expected to promote maximal diversity through resource par-
titioning that is maximized by interactions of food and larval supply (particle flux)
and sediment heterogeneity.
POM flux or food supply may ultimately play the most significant role in reg-
ulating the number of species. As with flow, effects may be direct or be mediated
through currents, sediments, disturbance, or hydrologic properties (Figure 9). At
low food levels, diversity will be low because there are insufficient resources to sup-
port viable populations of many species. As food supply increases, diversity will
increase because more species can maintain viable populations. As food availabil-
ity continues to increase, diversity may decline, producing a unimodal relationship
between diversity and food inputs (Figure 10). Why diversity declines at high lev-
els of food supply is not well understood but may reflect a decrease in habitat
heterogeneity, differential numerical responses leading to increased dominance by
a few species (Rosenzweig & Abramsky 1993, Waide et al. 1999), and/or increased
physiological stress due to oxygen limitation or sulfide toxicity. A strong negative
correlation of evenness with sediment POC (Levin & Gage 1998) may provide
some clues about the importance of species interactions in reducing diversity. Be-
cause increases in food supply are often coupled with variability (seasonality)
in supply, decline may also reflect increased extinction rates of more specialized
forms, due to either demographic instabilities or accelerating rates of competitive
exclusion (Rex 1976).
If deposit feeders select sediment particles according to characteristics such as
their size, shape, composition, and quality, as has been argued by Wheatcroft et al.
(1990), Etter & Grassle (1992), and Smith et al. (1993), diversity will respond
positively to increased sediment heterogeneity (Figure 10). Any environmental or
biotic factor that alters sediment structure (hydrodynamics, bioturbation, distur-
bance) has the potential to regulate diversity indirectly. The distribution of particle
sizes within the sediments also influences organic content (Milliman 1994); thus
changes in diversity due to alterations of sediment heterogeneity may be mediated
through food resources.
Oxygen is a requirement for most metazoan life; reduced oxygen availability
can exert both direct and indirect negative effects on diversity (Figures 9, 10). At
low oxygen concentrations, species richness may be reduced directly through dif-
ferential tolerance to hypoxia or to the sulfides that build up in organic-rich hypoxic
sediments (Diaz & Rosenberg 1995, Bagarinao 1992). Hypoxic conditions may
impose habitat homogeneity, contributing to spatially uniform assemblages of low
species richness (Levin et al. 2000). Such effects probably occur mainly at bottom-
water oxygen concentrations ¿0.5 ml/l (Levin & Gage 1998). Differential nu-
merical responses and competitive exclusion may explain increased dominance
and reduced species richness where oxygen is low but food availability remains
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 81

high. Tolerance to hypoxia varies with body size; meiofauna are much more toler-
ant to low oxygen than most macrofauna or megafauna. Organisms that oxygenate
sediments or reduce sulfide concentrations through feeding, dwelling structures,
and burrowing may indirectly facilitate other organisms, ameliorating physiologi-
cal stress and potentially elevating diversity (e.g., Reise 1982, Aller & Aller 1986).
Predation pressure and its potential diversifying effects (Hall et al. 1994, Dayton
& Hessler 1972) are expected to increase with increasing oxygen saturation.
The processes of disturbance and positive interactions (facilitation) pervade the
environmental factors considered above and merit further consideration. Distur-
bance, a well-known mechanism regulating community structure in other ecosys-
tems (Connell 1978, Petraitis et al. 1989, Pickett & White 1985), undoubtedly plays
a similar role in the deep sea. Disturbance to the seafloor is manifested through a
variety of biotic and abiotic factors that can affect diversity by regulating levels
of competition, predation, and physiological stress. Indirectly, disturbance can in-
crease heterogeneity at low levels and reduce it at high levels by smoothing out
biogenic or topographic structures or by inducing stress. Disturbance can also alter
the structure of the sediments and thus can influence diversity indirectly through
the generation or removal of habitat heterogeneity (Figure 9).
Facilitative interactions, like disturbance, can be important over a range of con-
ditions and are predicted to have strong diversifying effects where conditions are
most extreme (Bertness & Callaway 1994). We hypothesize that these effects are
especially important in marine sediments, where animal feeding, dwelling, and
respiratory activities alter the surrounding milieu. When food is scarce, animal
activities that enhance or redistribute food may elevate species richness (Bianchi
et al. 1989, Levin et al. 1997a). Stress or disturbance associated with low redox
conditions may be ameliorated by infaunal activities (solute pumping, burrowing)
that oxygenate sediments (Aller 1982), thereby allowing more species to persist
(Reise 1982). Structures formed by animals can provide shelter and a stable sub-
strate, and may serve as hotspots of diversity under strong flow conditions (e.g.,
tests of xenophyophores on seamounts, Levin et al. 1988).

CONCLUSIONS
Two important conclusions result from our survey of diversity patterns at regional
scales. First, the deep-sea ecosystem, long known to support high local species
diversity, also shows high biodiversity at the landscape level. High diversity is
related to environmental gradients and habitat shifts. Second, there appear to be
multiple forces that shape patterns of diversity on regional scales. Variation in
species diversity is associated with large-scale variation in sediment grain size di-
versity, nutrient input, and productivity as well as oxygen availability, hydrologic
conditions, and catastrophic events. Disturbance and facilitation are predicted to
be agents of particular importance in mediating environmental effects on diver-
sity in the deep sea. These environmental factors are highly interdependent and
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

82 LEVIN ET AL.

ultimately reflect global-scale oceanographic processes and climate. It seems likely

that regional-scale phenomena interact in complex ways to form biogeographic
patterns of local diversity by controlling gradients of ecological opportunity that
affect population dynamics at small scales. Exactly how this occurs is uncertain,
but, as shown here, the basic features of the regional ecological system are becom-
ing clearer. Importantly, the recognition of conspicuous variation in biodiversity at
the landscape level and the discovery of how relevant processes are geographically
distributed now make it possible to design controlled sampling studies and ma-
nipulative, experimental approaches to better understand the causality underlying
regional diversity trends.
Although several large data sets on the composition of deep-sea communities
are available, taxonomic and geographic coverage is still very limited. Most large-
scale studies of deep-sea species diversity have taken place in the Atlantic. It is
reasonable to expect that new biogeographic patterns and ecological forces of fun-
damental importance will continue to be discovered. The vast majority of diversity
studies center on the macrofauna and are often restricted to one or a few taxa.
Much less is known about the meiofauna and megafauna. Meiofauna and annelids
often demonstrate parallel trends along gradients of oxygen and organic matter
(e.g., Cook et al. 2000, Gooday et al. 2000, Gooday et al. 2001), although other
groups may respond to environmental pressures in different ways (Lambshead et al.
2001a in press). Thus, whereas recent regional studies provide important direction
for mechanistic studies of diversity in the deep sea, documenting the structure
and function of whole deep-sea communities on a planet-wide basis remains an
important priority.
Sampling methodology and the measurement of species diversity have impor-
tant implications for conceptualizing patterns of deep-sea diversity and how they
compare to those of other environments. Rarefaction, when critically applied, can
be a useful way to estimate relative diversity from samples collected by large
qualitative trawls and epibenthic sleds. Because the deep-sea fauna is sparsely
distributed, large samples are vital for taxonomic discrimination of species and for
assessing regional species pools. However, quantitative sampling with sufficient
replication to adequately characterize local communities is necessary to measure
the actual number of species, standing stock, and the environmental parameters
that appear to influence diversity directly. The first controlled comparison of this
kind, between the continental shelf and upper bathyal zone, confirms that the deep
sea in the Northwest Atlantic region supports considerably higher species richness
than in shelf communities. Although the shelf fauna is fairly well sampled, the
level of replication at bathyal depths is still inadequate to reveal an asymptote in
the species accumulation curve. An important objective for future sampling de-
signs is to determine the level and spatial scale of replication necessary to represent
local diversity. Also, because rarefaction of large qualitative samples suggests a
marked decline in diversity at abyssal depths (e.g., in the western North Atlantic),
quantitative sampling should be extended to the lower bathyal zone and abyssal
plain to firmly establish a diversity-depth gradient.
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 83

Contemporary ecological theory is based exclusively on the history of obser-

vations made in terrestrial, aquatic, and surface marine systems. Most of Earth’s
largest environment, the deep sea, remains unexplored. It is a vast world of ex-
treme physical circumstances, novel adaptations, and very basic permutations of
the carbon cycle that support all life (Gage & Tyler 1991, Rowe & Pariente 1992,
Van Dover 2000). Explaining the extraordinarily high level of species coexistence
encountered at very small scales in the deep sea remains a fundamental problem
(P. V. Snelgrove & C. R. Smith, unpublished manuscript; Etter & Mullineaux
2000). Although the regional studies reviewed here represent scattered areas of
the deep sea, they significantly enrich our knowledge of global biodiversity at the
landscape level, and should ultimately contribute to a comprehensive theoretical
understanding of community structure and function.

ACKNOWLEDGMENTS
The Deep-Sea Biodiversity Working Group is supported by the National Center
for Ecological Analysis and Synthesis, a Center funded by the National Science
Foundation (DEB-94–21535), the University of California, Santa Barbara, and
the State of California. The authors’ research is supported by the National Sci-
ence Foundation grants OCE-9811925 (RJE and MAR), OCE 98-03861 (LAL),
OCE 99–86627 (JP), OCE-9811925 (RJE), OPP 98–15823 (CRS), and the NOAA
National Undersea Research Program grant UAF 00–0050 (LAL). We thank
D. Morrow for assistance with manuscript preparation.

Visit the Annual Reviews home page at www.AnnualReviews.org

LITERATURE CITED
Allen JA, Sanders HL. 1996. The zoogeog- boundary region. Deep-Sea Res. Part I -
raphy, diversity and origin of the deep-sea Oceanog. Res. 44:39
protobranch bivalves of the Atlantic: the epi- Aller RC. 1982. The effects of macrobenthos on
logue. Prog. Oceanog. 38:95–153 chemical properties of marine sediment and
Aller J, Aller R. 1986. Evidence for localized overlying water. In Animal-Sediment Rela-
enhancement of biological activity associ- tions, ed. PL McCall, MJ Tevesz, pp. 53–102.
ated with tube and burrow structures in New York: Plenum
deep-sea sediments at the Hebble site; west- Arntz WE, Tarazona J, Gallardo VA, Flores
ern North Atlantic. Deep-Sea Res. 33:755– LA, Salzwedel H. 1991. Benthos commu-
90 nities in oxygen deficient shelf and upper
Aller JY. 1989. Quantifying sediment distur- slope areas of the Peruvian and Chilean Pa-
bances by bottom currents and its effect on cific coast, and changes caused by El Niño.
benthic communities in a deep-sea western In Modern and Ancient Continental Shelf
boundary zone. Deep-Sea Res. 36:901–34 Anoxia, ed. RV Tyson, TH Pearson, pp. 131–
Aller JY. 1997. Benthic community response 54. Tulsa, OK: Geol. Soc. Spec. Publ. No.
to temporal and spatial gradients in phys- 58
ical disturbance within a deep-sea western Austin MC, Widbom B. 1991. Changes in and
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

84 LEVIN ET AL.

slow recovery of a meiobenthic nematode Brown JH, Stevens GC, Kaufman DM. 1996.
assemblage following a hypoxic period in The geographic range: Size, shape, bound-
the Gullmar Fjord basin, Sweden. Mar. Biol. aries, and internal structure. Annu. Rev. Ecol.
111:139–45 Syst. 27:597–623
Bachelet G. 1990. The choice of sieving mesh Brown TN, Kulasiri D. 1996. Validating mod-
size in quantitative assessment of marine els of complex, stochastic, biological sys-
macrobenthos: a necessary compromise be- tems. Ecol. Model 86:129–34
tween aims and constraints. Mar. Environ. Bugge T, Befring S, Belderson RH, Eidvin T,
Res. 30:21–35 Jansen E, et al. 1987. A giant three-stage sub-
Bagarinao T. 1992. Sulfide as an environmental marine slide off Norway. Geo-Marine Lett.
factor and toxicant: tolerance and adaptations 7:191–98
in aquatic organisms. Aquat. Toxicol. 24:21– Cannariato KG, Kennett JJP, Behl RJ. 1999.
62 Biotic response to late Quaternary rapid cli-
Bartel D, Gutt J. 1992. Sponge associations in mate switches in Santa Barbara Basin: Eco-
the eastern Weddell Sea. Antarct. Sci. 4:137– logical and evolutionary implications. Geol-
50 ogy 27:63–66
Berelson WM, Anderson RF, Dymond J, Dem- Caswell H, Cohen JE. 1993. Local and re-
aster D, Hammond DE, et al. 1997. Bioge- gional regulation of species-area relations: a
nic budgets of particle rain, benthic rem- patch-occupancy model. In Species Diversity
ineralization and sediment accumulation in in Ecological Communities. ed. RE Ricklefs,
the equatorial Pacific. Deep-Sea Res. II ED Schluter, pp. 99–107. Chicago: Univ.
44:2251–82 Chicago Press
Bernhard JM, Sen Gupta BK, Bourne PF. Caswell H, Etter R. 1999. Cellular automaton
1997. Benthic foraminiferal proxy to es- models for competition in patchy environ-
timate dysoxic bottom-water oxygen con- ments: Facilitation, inhibition, and tolerance.
centrations: Santa Barbara Basin, U.S. Pa- Bull. Math. Biol. 61:625–49
cific continental margin. J. Foraminifer. Res. Charles C. 1998. The ends of an era. Nature
27:301–10 394:422–23
Bertness MD, Callaway R. 1994. Positive Cita MB, Podenzani M. 1980. Destructive ef-
interactions in communities. TREE 9:191– fects of oxygen starvation and ash falls on
93 benthic life: a pilot study. Quat. Res. 13:230–
Bett BJ, Rice AL. 1992. The influence of hex- 41
actinellid sponge (Pheronema carpenteri) Clarke KR, Warwick RM. 1994. Change in ma-
spicules on the patchy distribution of mac- rine communities, an approach to statisti-
robenthos in the Porcupine Seabight (bathyal cal analysis and interpretation. Natl Environ.
NE Atlantic) Ophelia 36:217–26 Res. Counc. UK. 144 pp.
Bianchi TS, Jones CG, Shachak M. 1989. Pos- Cochonat P, Lenat JF, Bachelery P, Boiwin P,
itive feedback of consumer populations on Cornaglia B. 1990. Gravity events as a pri-
resource supply Trends Ecol. Evol. 4:234– mary process in the construction of a subma-
38 rine volcano-sedimentary system (Fournaise
Blair NE, Plaia GR, Boehme SE, DeMaster Volcano, Reunion Island). C. R. Acad. Sci.
DJ, Levin LA. 1994. The remineralization Ser. II- 311:679–86
of organic carbon on the North Carolina con- Colwell RK. 1997. Estimate S: Statistical
tinental slope. Deep-Sea Res. 41:755–66 estimation of species richness and shared
Blake JA, Hilbig B. 1994. Dense infaunal species from samples. Version 5. Users Gui-
assemblages on the continental slope off de and Appl. http://viceroy.eeb.uconn.edu/
Cape Hatteras, North Carolina. Deep-Sea estimates
Res. 41:875–900 Colwell RK, Coddington JA. 1994. Estimating
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 85

terrestrial biodiversity through extrapola- the northern Arabian Sea. Geol. Ultraiectina,
tion. Philos. Trans. R. Soc. London Ser. B No. 188. 205 pp.
345:101–18 den Dulk M, Reichart GJ, Memon GM, Ro-
Colwell RK, Hurtt GC. 1994. Nonbiological elofs EMB, Zachariasse WJ, van der Zwaan
gradients in species richness and a spurious GJ. 1998. Benthic foraminiferal response to
Rappoport effect. Am. Nat. 144:570–95 variations in surface water productivity and
Colwell RK, Lees DC. 2000. The mid-domain oxygenation in the northern Arabian Sea.
effect: geometric constraints on the geogra- Mar. Micropaleontol. 35:43–66
phy of species richness. Trends Ecol. Evol. de Rijk S, Troelstra SR, Rohling EJ. 1999. Ben-
15:70–76 thic foraminiferal distribution in the Mediter-
Connell JH. 1978. Diversity in tropical rain ranean Sea. J. Foraminifer. Res. 29:93–
forests and coral reefs. Science 199:1302–10 103
Cook AA, Lambshead PJD, Hawkins LE, Deuser W, Muller-Karger F, Evans R, Brown
Mitchell N, Levin LA. 2000. Nematode O, Esaias W, Feldman G. 1990. Surface-
abundance at the oxygen minimum zone in ocean color and deep-ocean carbon flux: how
the Arabian Sea. Deep-Sea Res. 47:75–85 close a connection? Deep-Sea Res. 37:1331–
Corliss BH, Emerson S. 1990. Distribution of 43
rose bengal stained deep-sea benthic fora- Diaz RJ, Rosenberg R. 1995. Marine ben-
minifera from the Nova Scotian continental thic hypoxia: a review of its ecological ef-
margin and Gulf of Maine. Deep-Sea Res. fects and the behavioural responses of ben-
Part A- Oceanog. Res. Pap. 37:381–400 thic macrofauna. Oceanogr. Mar. Biol. Annu.
Cosson N, Sibuet M, Galeron J. 1997. Com- Rev. 33:245–303
munity structure and spatial heterogeneity of Dickson RR, McCave IN. 1986. Nepheloid lay-
the deep-sea macrofauna at three contrast- ers on the continental slope west of Porcupine
ing stations in the tropical northeast Atlantic. Bank. Deep-Sea Res. 33:791–818
Deep-Sea Res. 44:247–69 Douglas RG, Liestman J, Walch C, Blake G,
Cosson-Sarradin N, Sibuet M, Paterson GLJ, Cotton ML. 1980. The transition from live to
Vangriesheim A. 1998. Polychaete diver- sediment assemblage in benthic foraminifera
sity at tropical Atlantic deep-sea sites: en- from the Southern California Borderland.
vironmental effects. Mar. Ecol. Prog. Ser. In Quaternary Depositional Environments of
165:173–85 the Pacific Coast, ed. ME Field, AH Bouma,
Cronin TM, Raymo ME. 1997. Orbital forcing IP Colburn, RG Douglas, JC Dingle, pp.
of deep-sea benthic species diversity. Nature 257–80. Los Angeles: Pac. Coast Paleogeogr.
385:624–27 Symp. 4, Pac. Sect., Soc. Econ. Paleontol.
Culver SJ, Buzas MA. 1998. Patterns of oc- Mineral.
currence of benthic foraminifera in time and Embley RW. 1976. New evidence for occur-
space. In The Adequacy of the Fossil Record, rence of debris flow deposits in the deep sea.
ed. SK Donovan, CRC Paul, pp. 207–26. Geology 4:371–74
New York: Wiley Emerson S. 1985. Organic carbon preserva-
Culver SJ, Buzas MA. 2000. Global latitudinal tion in marine sediments. In The Carbon Cy-
species diversity gradient in deep-sea benthic cle and Atmospheric CO2: Natural Varia-
foraminifera. Deep-Sea Res. I 47:259–75 tions Archean to Present, ed. I Sundquist, W
Dayton PK, Hessler RR. 1972. Role of biolog- Broecker, 32:78–87. Washington, DC: Geo-
ical disturbance in maintaining diversity in phys. Monogr., AGU
the deep sea. Deep-Sea Res. 19:199–208 Emery KO, Uchupi E. 1984. The Geology of the
den Dulk M. 2000. Benthic foraminiferal re- Atlantic Ocean. New York/Berlin: Springer-
sponse to Late Quaternary variations in sur- Verlag. Charts I–XI. 2 Vols. 1050 pp.
face water productivity and oxygenation in Etter RJ, Grassle JF. 1992. Patterns of species
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

86 LEVIN ET AL.

diversity in the deep sea as a function of sedi- Tarazona, pp. 79–85. Inst. del Mar del Perú:
ment particle size diversity. Nature 360:576– Lima, Peru: Bol. Extraordin.
78 Genin A, Lazar B, Brenner S. 1995. Vertical
Etter RJ, Mullineaux L. 2000. Deep-Sea com- mixing and coral death in the Red Sea follow-
munities. In Marine Community Ecology ed. ing the eruption of Mount Pinatubo. Nature
MD Bertness, S Gaines, M Hay, pp. 367–93, 377:507–10
Chapter 14. Sunderland, MA: Sinauer Glover A, Paterson G, Bett B, Gage J, Sib-
Fenchel T, Kofoed LH. 1976. Evidence for ex- uet M, et al. 2001. Patterns in polychaete
ploitative interspecific competition in mud abundance and diversity from the Madeira
snails (Hydrobiidae). Oikos 27:367–76 Abyssal Plain, northeast Atlantic. Deep-Sea
Fenchel T, Kofoed H, Lappalainen A. 1975. Res. I 48:217–36
Particle size selection of two deposit feed- Gooday AJ. 1996. Epifaunal and shallow in-
ers: the amphipod Corophium volutator and faunal foraminiferal communities at three
the prosobranch Hydrobia ulvae. Mar. Biol. abyssal NE Atlantic sites subject to differ-
30:119–28 ing phytodetritus regimes. Deep-Sea Res.
Flood RD, Shor AN. 1988. Mud waves in the 43:1395–431
Argentine Basin and their relationship to re- Gooday AJ, Bernhard JM, Levin LA, Suhr S.
gional bottom circulation patterns. Deep-Sea 2000. Foraminifera in the Arabian Sea OMZ
Res. 35:943–71 and other oxygen deficient settings: taxo-
Frederiksen RA, Jensen A, Westerberg H. nomic composition, diversity and relation to
1992. The distribution of the scleractinian metazoan faunas. Deep-Sea Res. 47:54–73
coral Lophelia pertusa around the Faroe Is- Gooday AJ, Hughes JA, Levin LA. 2001. The
lands and the relation to internal mixing. Sar- foraminiferal macrofauna from three North
sia 77:157–71 Carolina (U.S.A.) slope sites with contrasting
Gage JD. 1996. Why are there so many species carbon flux: a comparison with the metazoan
in deep-sea sediments? J. Exp. Mar. Biol. macrofauna. Deep-Sea Res. I. 48:1709–39
Ecol. 200:257–86 Grassle JF, Maciolek NJ. 1992. Deep-sea
Gage JD. 1997. High benthic species diver- species richness: regional and local diversity
sity in deep-sea sediments: The importance estimates from quantitative bottom samples.
of hydrodynamics. In Marine Biodiversity, Am. Nat. 139:313–41
ed. RFG Ormond, JD Gage, MV Angel, pp. Grassle JF, Morse-Porteous LS. 1987. Macro-
148–77. Cambridge, UK:Cambridge Univ. faunal colonization of disturbed deep-sea
Press environments and the structure of deep-
Gage JD, Lamont PA, Tyler PA. 1995. Deep- sea benthic communities. Deep-Sea Res. 34:
sea macrobenthic communities at contrasting 1911–50
sites off Portugal, preliminary results: I intro- Gray JS. 1981. The Ecology of Marine Sed-
duction and diversity comparisons. Int. Rev. iments. Cambridge, UK: Cambridge Univ.
Hydrobiol. 80:235–50 Press
Gage JD, May RM. 1993. Biodiversity—a dip Gray JS. 1994. Is deep-sea species diversity re-
into the deep seas. Nature 365:609–10 ally so high? Species diversity of the Norwe-
Gage JD, Tyler PA. 1991. Deep-Sea Biology: gian continental shelf. Mar. Ecol. Prog. Ser.
A Natural History of Organisms at the 112:205–9
Deep-Sea Floor. Cambridge, UK: Cam- Gray JS, Poore GCB, Ugland KI, Wilson RS,
bridge Univ. Press. 504 pp. Olsgard F, Johannessen O. 1997. Coastal and
Gallardo VA. 1985. Efectos del fenómeno de El deep-sea benthic diversities compared. Mar.
Niño sobre el bentos sublitoral frente a Con- Ecol. Prog. Ser. 159:97–103
cepción, Chile. In El Niño y su impacto en Gross TF, Williams AJ. 1991. Characterization
la fauna marina, ed. W Arntz, A Landa, J of deep-sea storms. Mar. Geol. 99:281–301
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 87

Hall SJ. 1994. Physical disturbance and marine and Geochemistry of Abyssal Plains, ed. PPE
benthic communities: life in unconsolidated Weaver, J Thomson, pp. 105–12. Geol. Soc.
sediments. Oceanog. Mar. Biol.: An Annu. Spec. Publ. No. 31
Rev. 32:179–239 Hurlbert SM. 1971. The non-concept of species
Hall SJ, Raffaelli D, Thrush SF. 1994. Patchi- diversity, A critique and alternative parame-
ness and disturbance in shallow water benthic ters. Ecology 52:577–86
assemblages. In Aquatic Ecology: Scale, Pat- Huston M. 1979. A general hypothesis of
tern and Process, ed. PS Giller, AG Hildrew, species diversity. Am. Nat. 113:81–101
DG Raffaelli, pp. 333–75. London: Black- Jacobi RD, Hayes DE. 1982. Bathymetry, mi-
well Sci. crophysiography and reflectivity character-
Hanski I. 1998. Metapopulation dynamics. Na- istics of the West African margin between
ture 396:41–49 Sierra Leone and Mauritania. In Geology of
Heezen BC, Hollister CD. 1971. The Face of the Northwest African Margin, ed. U von
the Deep. New York: Oxford Univ. Press. 659 Rad, K Hinz, M Sarnthein, E Siebold, pp.
pp. 182–212. Heidelberg: Springer-Verlag
Hermelin JOR, Schimmield GB. 1990. The im- Jahnke R. 1996. The global ocean flux of par-
portance of the oxygen minimum zone and ticulate organic carbon, areal distribution
sediment geochemistry in the distribution of and magnitude. Global Biogeochem. Cycles
benthic foraminifera in the Northwestern In- 10:71–88
dian Ocean. Mar. Geol. 91:1–29 Jensen A, Frederickson R. 1992. The fauna
Hermelin JOR, Schimmield GB. 1995. Im- associated with the bank-forming deepwa-
pact of productivity events on benthic ter coral Lophelia pertusa (Scleractiniaria)
foraminiferal faunas in the Arabian Sea over on the Faroe shelf. Sarsia 77:53–69
the last 150,000 years. Paleoceanography Jorissen FJ. 1999. Benthic foraminiferal suc-
10:85–116 cessions across Late Quaternary Medi-
Hess S, Kuhnt W. 1996. Deep-sea benthic terranean sapropels. Mar. Geol. 153:91–
foraminiferal recolonization of the 1991 Mt. 101
Pinatubo ash layer in the South China Sea. Jorissen FJ, Buzas M, Culver S, Kuehl S.
Mar. Micropaleontol. 28:171–97 1994. Vertical distribution of living benthic
Hessler RR, Jumars PA. 1974. Abyssal commu- foraminifera in submarine canyons off New
nity analysis from replicate box cores in the Jersey. J. foraminiferal Res. 24:28–36
central North Pacific. Deep-Sea Res. 21:185– Jorissen FJ, Wittling I, Peypouquet JP, Rabo-
209 uille C, Relexans JC. 1998. Live forami-
Hollister CD, Nowell ARM. 1991. HEBBLE niferal faunas off Cap Blanc, NW Africa:
epilogue. Mar. Geol. 99:445–60 Community structure and microhabitats.
Hollister CD, McCave IN. 1984. Sedimenta- Deep-Sea Res. I 45:2157–88
tion under deep-sea storms. Nature 309:220– Josefson AB, Widbom B. 1988. Differential re-
25 sponse of benthic macrofauna and meiofauna
Hollister CD, Nowell ARM, Jumars PA. 1984. to hypoxia in the Gullmar Fjord basin. Mar.
The dynamic abyss. Sci. Am. 250:42–53 Biol. 100:31–40
Hubbell SP, Foster RB, O’Brien ST, Harms Jumars PA, Eckman JE. 1983. Spatial structure
KE, Condit R, et al. 1999. Light-cap dis- within deep-sea benthic communities. In The
turbances, recruitment limitation, and tree Sea, ed. GT Rowe, pp. 399–452. New York:
diversity in a neotropical forest. Science Wiley
283:554–57 Jumars PA, Hessler RR. 1976. Hadal commu-
Huggett QJ. 1987. Mapping of hemipelagic nity structure: implications from the Aleutian
versus turbidite muds by feeding traces ob- Trench. J. Mar. Res. 34:547–60
served in deep-sea photographs. In Geology Kaminski M. 1985. Evidence for control of
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

88 LEVIN ET AL.

abyssal agglutinated foraminiferal commu- the diversity of deep-sea North Atlantic ne-
nity structure by substrate disturbance: re- matodes. Mar. Ecol. Prog. Ser. 214:121–26
sults from the HEBBLE area. Mar. Geol. Lambshead PJ, Tietjen J, Moncrieff C, Ferrero
66:113–31 T. 2001b. North Atlantic latitudinal diversity
Kamykowski D, Zentara SJ. 1990. Hypoxia in patterns in deep-sea marine nematode data.
the world ocean as recorded in the historical Mar. Ecol. Prog. Ser. In press
data set. Deep-Sea Res. 37:1861–74 Lampitt R, Antia R, 1997. Particle flux in the
Keller M. 1986. Structure des peuplements deep seas: regional characteristics and tem-
méiobenthiques dans la secteur pollué par la poral variability. Deep-Sea Res. I 44:1377–
rejet en mer de l’égout de Marseille. Annales 73
Inst. océanog. 62:13–36 Lees DC, Kremen C, Andriamampianina L.
Kidd RB, Hill PR. 1986. Sedimentation on 1999. A null model for species richness gra-
mid-ocean sediment drifts. In North Atlantic dients: bounded range overlap of butterflies
Palaeoceanography, ed. CP Summerhayes, and other rainforest endemics in Madagascar.
NJ Shackleton, pp. 87–102. Geol. Soc. Spec. Biol. J. Linnean Soc. 67:529–84
Pub., No. 21 Levin LA. Blair N, DeMaster DJ, Plaia G,
Klein H. 1988. Benthic storms, vortices, and Fornes W, et al. 1997a. Rapid subduction of
particle dispersion in the deep western Eu- organic matter by maldanid polychaetes on
ropean Basin. Deutsches hydrographishes the North Carolina slope. J. Mar. Res. 55:
Zeitung 40:87–102 595–611
Klitgaard AB. 1995. The fauna associated Levin L, Blair N, Martin C, DeMaster D,
with outer shelf and upper slope sponges Plaia G, Thomas C. 1999. Macrofaunal pro-
(Porifera, Demospongiae) at the Faroe Is- cessing of phytodetritus at two sites on the
lands, northeastern Atlantic. Sarsia 80:1– Carolina margin: In situ experiments using
13
22 C-labeled diatoms. Mar. Ecol. Progr. Ser.
Kontar EA, Sokov AV. 1994. A benthic storm 182:37–54
in the northeastern tropical Pacific over the Levin LA, DiBacco C. 1995. The influence of
fields of manganese nodules. Deep-Sea Res. sediment transport on short-term recoloniza-
I 41:1069–89 tion by seamount infauna. Mar. Ecol. Progr.
Lambshead PJD, Brown CJ, Ferrero TJ, Jensen Ser. 123:163–75
J, Smith CR, Hawkins LE, Tietjen J. 2001a. Levin LA, Gage JD. 1998. Relationships be-
Latitudinal diversity patterns for deep sea tween oxygen, organic matter and the diver-
marine nematodes and organic fluxes—a test sity of bathyal macrofauna. Deep-Sea Res.
from the central equatorial Pacific. Mar. Ecol. 45:129–63
Prog. Ser. In press Levin LA, Gage J, Lamont P, Cammidge L,
Lambshead PJD, Elce BJ, Thistle D, Eckman Martin C, et al. 1997b. Infaunal community
JE, Barnett PRO. 1994. A comparison of bio- structure in a low-oxygen, organic rich habi-
diversity of deep-sea marine nematodes from tat on the Oman continental slope, NW Ara-
three stations in the Rockall Trough, north- bian Sea. In Responses of marine organ-
east Pacific. Biodiv. Lett. 2:95–107 isms to their environments, ed. L Hawkins,
Lambshead PJD, Tietjen J, Ferrero T, Jensen S Hutchinson, pp. 223–30. Proc. 30th Eur.
J. 2000. Latitudinal diversity gradients in Mar. Biol. Symp., Univ. Southampton
the deep sea with special reference to North Levin LA, Gage JD, Martin C, Lamont
Atlantic nematodes. Mar. Ecol. Prog. Ser. PA. 2000. Macrobenthic community struc-
194:159–67 ture within and beneath the oxygen mini-
Lambshead PJD, Tietjen J, Glover A, Ferrero mum zone, NW Arabian Sea. Deep-Sea Res.
T, Thistle D, Gooday AJ. 2001. The impact 47:189–226
of large-scale natural physical disturbance on Levin LA, Huggett CL, Wishner KF. 1991.
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 89

Control of deep-sea benthic community and rise. Washington, DC: US Dept. Int.,
structure by oxygen and organic-matter gra- Minerals Mgmt. Ser.
dients in the eastern Pacific Ocean. J. Mar. Magurran A. 1988. Ecological Diversity and
Res. 49:763–800 Its Measurement. Princeton, NJ: Princeton
Levin LA, Leithold EL, Gross TF, Huggett CL, Univ. Press, 179 pp.
DiBacco C. 1994a. Contrasting effects of Martin JH, Knauer GA, Karl DM, Broenkow
substrate mobility on infaunal assemblages WW. 1987. VERTEX: Carbon cycling in the
inhabiting two high-energy settings on Fie- northeast Pacific. Deep-Sea Res. 34:267–85
berling Guyot. J. Mar. Res. 52:489–522 Masson DG. 1996. Catastrophic collapse of the
Levin LA, Plaia GR, Huggett CL. 1994b. The volcanic island of Hierro 15 ka ago and the
influence of natural organic enhancement on history of landslides in the Canary Islands.
life histories and community structure of Geology 24:231–34
bathyal polychaetes. In Reproduction, lar- Masson DG, Kidd RB, Gardner JV, Huggett
val biology, and recruitment of the deep- QJ, Weaver PPE. 1994. Saharan continental
sea benthos, ed. CM Young, KJ Eckelbarger, rise: facies distribution and sediment slides.
pp. 261–83. Columbia, SC: Columbia Univ. In Geological Evolution of Atlantic Conti-
Press nental Rises, ed. VW Poag, PC Degrocianski,
Levin LA, Thomas CL, 1988. The ecology of pp. 3–10. New York: Van Nostrand Reinhold
xenophyophores (Protista) on eastern Pacific Masson DG, Kenyon NH, Weaver PPE. 1996.
seamounts. Deep-Sea Res. 35:2003–27 Slides, debris flows, and turbidity currents,
Lipman PW, Normark WR, Moore JG, Wilson In Oceanography: An Illustrated Guide, ed.
JB, Gutmacher CE. 1988. The giant Alika CP Summerhayes, SA Thorpe, pp. 136–51.
debris slide, Mauna Loa, Hawaii. J. Geophys. London: Manson
Res. 93:4279–99 May RM. 1975. Patterns of species abun-
Luth U, Luth CM. 1997. A benthic approach dance and diversity. In Ecology and Evolu-
to determine long-term changes in the water tion of Communities. ed. ML Cody, M Dia-
column of the Black Sea. In: The Responses mond.Cambridge, MA: Harvard Univ. Press.
of Marine Organisms to their Environment. pp. 81–120
Southampton, UK: Proc. 30th Eur. Mar. Biol. Menge BA, Sutherland JP. 1987. Community
Symp., Sept. 1995. Southampton Oceanogr. regulation: variation in disturbance, competi-
Centre. pp. 223–30 tion, and predation in relation to environmen-
MacArthur RH. 1972. Geographical Ecol- tal stress and recruitment. Am Nat. 130:730–
ogy. Patterns in the Distribution of Spe- 57
cies, Princeton, NJ: Princeton Univ. Press. Milliman JD. 1994. Organic matter content in
269 pp. U.S. Atlantic continental slope sediments:
Maciolek-Blake NJ, Grassle JF, Blake JA, Neff decoupling the grain size factor. Deep-Sea
JM. 1985. Georges Bank Infauna Monitor- Res. 41:797–808
ing Program:Final report for the third year Moodley L, van der Zwaan GJ, Herman PMJ,
of sampling. Washington, DC: US Dept. Int., Kempers L, van Breugel P. 1997. Differen-
Minerals Mgmt. Ser. tial response of benthic meiofauna to anoxia
Maciolek NJ, Grassle JF, Hecker B, Boehm with special reference to the Foraminifera
PD, Brown B et al. 1987a. Study of biologi- (Protista: Sarcodina). Mar. Ecol. Progr. Ser.
cal processes on the U.S. mid-Atlantic slope 158:151–63
and rise. Phase 2. Washington, DC: US Dept. Mullins HT, Thompson JB, McDougall
Int., Minerals Mgmt. Ser. K, Vercoutere TL. 1985. Oxygen-minimum
Maciolek NJ, Grassle JF, Hecker B, Brown zone edge effects, evidence from the central
B, Blake JA et al. 1987b. Study of biological California coastal upwelling system. Geol-
processes on the U.S. North Atlantic slope ogy 13:491–94
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

90 LEVIN ET AL.

Munk W. 1970. The circulation of the oceans. and meiobenthos of the Arkona Basin (west-
In Adventures in Earth History, ed. P Cloud, ern Baltic Sea): differential recovery follow-
p. 235–40. San Francisco: Freeman ing hypoxic events. In Responses of Marine
Murrell MC, Fleeger JW. 1989. Meiofaunal Organisms to Their Environments, ed. L.E.
abundances on the Gulf of Mexico continen- Hawkins, S. Hutchinson, pp. 251–262. Proc.
tal shelf affected by hypoxia. Cont. Shelf Res. 30th Eur. Mar. Biol. Symp., Univ. Southamp-
9:1049–62 ton
Neira C, Sellanes J, Levin LA, Arntz WA. Reimers CE, Jahnke RA, McCorkle DC. 1992.
2001. Meiofaunal distributions on the Peru Carbon fluxes and burial rates over the con-
margin: relationship to oxygen and organic tinental slope and rise off central California
matter availability. Deep-Sea Res. 48:2453– with implications for the global carbon cycle.
72 Global Biogeochemical Cycles 6:199–224
Noble M, Mullineaux LS. 1989. Internal tidal Reise K. 1985. Tidal Flat Ecology. Berlin:
currents over the summit of Cross Seamount. Springer-Verlag 191 pp.
Deep-Sea Res. 36:1791–1802 Rex MA. 1973. Deep-sea species diversity: de-
Paterson GLJ, Wilson GDF, Cosson N, La- creased gastropod diversity at abyssal depths.
mont PA. 1998. Hessler and Jumars (1974) Science 181:1051–53
revisited: abyssal polychaete assemblages Rex MA. 1976. Biological accommodation in
from the Atlantic and Pacific. Deep-Sea Res. the deep-sea benthos: comparative evidence
II 45:225–51 on the importance of predation and produc-
Paterson GLJ, Lambshead PJD. 1995. Bathy- tivity. Deep-Sea Res. 23:975–87
metric patterns of polychaete diversity in the Rex MA. 1981. Community structure in the
Rockall Trough, northeast Atlantic. Deep- deep-sea benthos. Annu. Rev. Ecol. Syst.
Sea Res. I 42:1199–1214 12:331–53
Petersen C. 1913. Valuation of the sea. II. The Rex MA. 1983. Geographic patterns of species
animal communities of the sea bottom and diversity in deep-sea benthos. In The Sea, ed.
their importance for marine zoogeography. GT Rowe, Vol. 8, pp. 453–72. New York: Wi-
Rep. Danish Biol. Stn. to Board Ag. 21:1–44 ley
Petraitis PS, Latham RE, Niesenbaum RA. Rex MA, Stuart CT, Coyne G. 2000. Latitudi-
1989. The maintenance of species diversity nal gradients of species richness in the deep-
by disturbance. Quar. Rev. Biol. 64:393–418 sea benthos of the North Atlantic. Proc. Natl.
Phleger PB, Soutar A. 1973. Production of ben- Acad. Sci. USA 97:4082–85
thic foraminifera in three east Pacific oxygen Rex MA, Stuart CT, Etter R. 1997. Large-
minima. Micropaleontology 19:110–15 scale patterns of species diversity in the deep-
Pickett STA, White PS. 1984. Natural Dis- sea benthos. In Marine Biodiversity: Patterns
turbance: An Evolutionary Perspective. New and Processes, ed. R Ormond, J D Gage, MV
York: Academic Angel, pp. 94–121. Cambridge, UK: Cam-
Pineda J. 1993. Boundary effects on the vertical bridge Univ. Press
ranges of deep-sea benthic species. Deep-Sea Rex MA, Stuart CT, Hessler RR, Allen JA,
Res. I 40:2179–92 Sanders HL, Wilson GDF. 1993. Global-
Pineda J, Caswell H. 1998. Bathymetric spe- scale latitudinal patterns of species diversity
cies-diversity patterns and boundary con- in the deep-sea benthos. Nature 365:636–39
straints on vertical range distribution. Deep- Rhoads DC. 1974. Organism sediment relations
Sea Res. II, 45:83–101 on the muddy sea floor. Oceanogr. Mar. Biol.
Pingree RD, New AL. 1989. Downward prop- Ann. Rev. 12:263–300
agation of internal tidal energy in the Bay of Rice A, Thurston M, New A. 1990. Dense
Biscay. Deep-Sea Res. 36:735–58 aggregations of a hexactinellid sponge,
Radziejewska T, Maslowski J. 1997. Macro- Pheronema carpenteri, in the Porcupine
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 91

Seabight (northeast Atlantic Ocean), and Shepard FP, Marshall NF, McLoughlin PA,
possible causes. Prog. Oceanog. 24:179–96 Sullivan GG. 1979. Currents in submarine
Richardson MJ, Weatherly GL, Gardner WD. canyons and other seavalleys. Tulsa, OK:
1993. Benthic storms in the Argentine Basin. AAPG Stud. Geol., Vol. 8
Deep-Sea Res. II, 40:975–87 Shmida A, Wilson MV. 1985. Biological de-
Ricklefs RE. 1987. Community diversity: rel- terminants of species diversity. J. Biogeogr.
ative roles of local and regional processes. 12:1–20
Science 235:167–71 Simms RW, Weaver PPE, Kidd RB, Jones
Rogers A. 2000. The role of the oxygen min- EJW. 1991. Late Quaternary mass movement
ima in generating biodiversity in the deep sea. on the lower continental rise and abyssal
Deep-Sea Res. II 47:119–48 plain off western Sahara. Sedimentology
Rosenzweig ML. 1995. Species Diversity in 38:27–40
Space and Time. Cambridge: Cambridge Smith CR, Berelson W, Demaster DJ, Dobbs
Univ. Press. 436 pp. FC, Hammond D, et al. 1997. Latitudinal
Rosenzweig ML, Abramsky H. 1993. How are variations in benthic processes in the abyssal
diversity and productivity related? In Species equatorial Pacific: control by biogenic parti-
Diversity in Ecological Communities: His- cle flux. Deep-Sea Res. II:2295–317
torical and Geographical Perspectives, ed. Smith CR, Pope RH, DeMaster DJ, Maga-
RE Ricklefs, D Schluter, pp. 53–65. Chicago: ard L. 1993. Age-dependent mixing of deep-
Univ. Chicago Press sea sediments. Geochim. Cosmochim. Acta
Rowe GT, Menzies RJ. 1968. Deep bottom cur- 57:1473–88
rents off the coast of North Carolina. Deep- Smith KL, Carlucci AF, Jahnke RA, Craven
Sea Res. 15:711–19 DB. 1987. Organic carbon mineralization in
Rowe G, Pariente V. 1992. Deep-Sea Food the Santa Catalina Basin: benthic boundary
Chains and the Global Carbon Cycle, pp. layer metabolism. Deep-Sea Res. 34:185–
560. Dordrecht, Netherlands: Kluwer 211
Rowe GT, Sibuet M, Deming J, Khripounoff Smith KL, Hinga KR. 1983. Sediment com-
A, Tietjen J, et al. 1991. ‘Total’ sediment munity respiration in the deep sea. In Deep
biomass and preliminary estimates of organic Sea Biology, ed. GT Rowe, pp. 331–70. New
carbon residence time in deep-sea benthos. York: Wiley
Mar. Ecol. Prog. Ser. 79:99–114 Snelgrove PV, Butman CA. 1994. Animal-
Sanders HL. 1968. Benthic studies in Buzzards sediment relationships revisited: cause vs. ef-
Bay. I: Animal-sediment relationships. Lim- fect. Oceanogr. Mar. Biol. Ann. Rev. 32:111–
nol. Oceanog. 3:245–58 177
Sanders HL. 1969. Benthic marine diversity and Stephens MP, Kadko DC, Smith CR, Latasa
the time-stability hypothesis. Brookhaven M. 1997. Chlorophyll-a and pheopigments
Symp. Biol. 22:71–81 as tracers of labile organic carbon at the
Schaff T, Levin L, Blair N, DeMaster D, Pope central equatorial Pacific seafloor. Geochim.
R, Boehme S. 1992. Spatial heterogeneity of et Cosmochim. Acta 61:4605–19
benthos on the Carolina continental slope: Stevens GC. 1989. The latitudinal gradient in
large (100-km)-scale variation. Mar. Ecol. geographical range: how so many species
Prog. Ser. 88:143–60 coexist in the tropics. Am. Natur. 133:240–
Self RFL, Jumars PA. 1988. Cross-phyletic pat- 56
terns of particle selection by deposit feeders. Stow DAV, Faugères JCF, eds. 1993. Con-
J. Mar. Res. 46:119–143 tourites and Bottom Currents. Sed. Geol.
Sen Gupta BK, Machain-Castillo ML. 1993. 82:310 pp.
Benthic foraminifera in oxygen-poor habi- Stow DAV, Holbrook JA. 1984. North Atlantic
tats. Mar. Micropaleontology 20:183–201 contourites: an overview. In Fine-Grained
10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

92 LEVIN ET AL.

Sediments: Deep-Water Processes and Fa- tor Y. 1999. Increased marine sediment sus-
cies, ed. DAV Stow, DJW Piper, pp. 245–56. pension and fluxes following an earthquake.
Oxford: Blackwell Sci. Nature 398:233–36
Stuart CT, Rex MA. 1994. The relationship Tietjen JH. 1984. Distribution and species
between developmental pattern and species diversity of deep-sea nematodes in the
diversity in deep-sea prosobranch snails. In Venezuela Basin. Deep-Sea Res. 31:119–32
Reproduction, Larval Biology, and Recru- Tietjen JH. 1989. Ecology of deep-sea nema-
itment of the Deep-sea Benthos, ed. CM todes from the Puerto Rico Trench area and
Young, KJ Eckelbarger, pp. 118–36. New the Hatteras Abyssal Plain. Deep-Sea Res.
York: Columbia Univ. Press 36:1579–94
Stuart CT, Rex MA, Etter RJ. 2001. Large Tilman D. 1982. Resource Competition and
scale spatial and temporal patterns of deep- Community Structure. Princeton, NJ: Prince-
sea benthic species diversity. In Ecosystems ton Univ. Press
of the World: Ecosystems of Deep Oceans, Tillman D. 1994. Competition and biodiver-
ed. PA Tyler. Amsterdam: Elsevier. In press sity in spatially structured habitats. Ecology
Suess E. 1980. Particulate organic carbon flux 75:2–16
in the oceans—surface productivity and oxy- Tyler PA. 1995. Conditions for the existence
gen utilization. Nature 288:260–263 of life at the deep-sea floor: an update.
Taghon GL. 1982. Optimal foraging by deposit- Oceanogr. Mar. Biol. Annu. Rev. 33:221–44
feeding invertebrates: roles of particle size Van Dover C. 2000. The Ecology of Deep-sea
and organic coating. Oecologia 52:295– Hydrothermal Vents, Princeton, NJ: Prince-
304 ton Univ. Press. 424 pp.
Thistle D. 1998. Harpacticoid copepod diver- Vetter EW, Dayton PK. 1998. Macrofaunal
sity at two physically reworked sites in the communities within and adjacent to a
deep sea. Deep-Sea Res. II 45:13–24 detritus-rich submarine canyon system.
Thistle D. 1983. The role of biologically Deep-Sea Res. 45:25–54
produced habitat heterogeneity in deep- Viana AR, Faugères J-C, Stow DAV. 1998.
sea diversity maintenance. Deep-Sea Res. Bottom-current-controlled sand deposits—a
30:1235–45 review of modern shallow- to deep-water en-
Thistle D, Ertman SC, Fauchald K. 1991. The vironments. Sed. Geol. 115:53–80
fauna of the HEBBLE site: patterns in stand- Von Rad U, Schulz H, Riech V, den Dulk
ing stock and sediment-dynamic effects. Mar. M, Berner U, Sirocko F. 1999. Multiple
Geol. 99:413–22 monsoon-controlled breakdown of oxygen-
Thistle D, Levin L. 1998. The effect of ex- minimum conditions during the past 30,000
perimentally increased near-bottom flow on years documented in laminated sediments
metazoan meiofauna at a deep-sea site, with off Pakistan. Palaeogeog., Palaeoclim.,
comparison data on macrofauna. Deep-Sea Palaeoecol., 152:129–61
Res. 45:625–38 Waide RB, Willig MR, Steiner CF, Mittelbach
Thistle D, Yingst UY, Fauchald K. 1985. A G, Gough L. 1999. The relationship between
deep-sea benthic community exposed to productivity and species richness. Annu. Rev.
strong near-bottom currents on the Scotian Ecol. Syst. 30:257–300
Rise (western Atlantic). Mar. Geol. 66:91– Warwick RM, Clarke KR. 1996. Relationships
112 between body-size, species abundance and
Thomas E, Gooday AJ. 1996. Cenozoic deep- diversity in marine benthic assemblages:
sea benthic foraminifers: tracers for changes facts or artifacts? J. Exp. Mar. Biol. Ecol.
in oceanic productivity? Geology 24:355– 202:63–71
58 Watts MC, Etter RJ, Rex MA. 1992. Effects of
Thunnell R, Tappa E, Varela R, Llano M, As- spatial and temporal scale on the relationship
 10 Oct 2001 16:3 AR ar142-03.tex ar142-03.sgm ARv2(2001/05/10) P1: GJB

REGIONAL DIVERSITY IN THE DEEP SEA 93

of surface pigment biomass to community Wheatcroft RA, Jumars PA. 1987. Statistical
structure in the deep-sea benthos. In Deep- reanalysis for size dependent bioturbation
sea Food Chains and the Global Carbon Cy- in the deep ocean. Limnol. and Oceanogr.
cle, ed. GT Rowe, V Pariente, pp. 245–54. 37:90–104
Dordrecht, The Netherlands: Kluwer Whitlatch RB. 1980. Patterns of resource uti-
Weatherley GL, Kelley EA. 1985. Storms and lization and coexistence in marine intertidal
flow reversals at the HEBBLE site. Mar. deposit-feeding communities. J. Mar. Res.
Geol. 66:205–18 38:743–65
Wheatcroft RA. 1992. Experimental tests for Willig MR, Lyons SK. 1998. An analytical
particle size-dependent bioturbation in the model of latitudinal gradients of species rich-
deep ocean. Limn. Oceanogr. 37:90–104 ness with an empirical test for marsupials and
Wheatcroft RA, Jumars PA, Smith CR, Nowell bats in the New World. Oikos 81:93–98
ARM. 1990. A mechanistic view of the par- Wishner K, Levin L, Gowing M, Mullineaux
ticulate biodiffusion coefficient-step lengths, L. 1990. Involvement of the oxygen min-
rest periods and transport directions J. Mar. imum in benthic zonation on a deep sea
Res. 48:177–208 mount. Nature 346:57–59

View publication stats