Variability Of Deep-Sea Meiofaunal Abundances On The

Florida State University Libraries
Electronic Theses, Treatises and Dissertations
The Graduate School
2012
Variability of Deep-Sea Meiofaunal
Abundances on the Continental Rise off the
West Coast of the United States
Melissa Rohal
Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected]
THE FLORIDA STATE UNIVERSITY
COLLEGE OF ARTS AND SCIENCES
VARIABILITY OF DEEP-SEA MEIOFAUNAL ABUNDANCES ON THE CONTINENTAL
RISE OFF THE WEST COAST OF THE UNITED STATES
By
MELISSA ROHAL
A Thesis submitted to the
Department of Earth, Ocean, and Atmospheric Science
in partial fulfillment of the
requirements for the degree of
Master of Science
Degree Awarded:
Fall Semester, 2012
Melissa Rohal defended this thesis on October 4th, 2012.
The members of the supervisory committee were:
David Thistle
Professor Directing Thesis
William M. Landing
Committee Member
Ian MacDonald
Committee Member
The Graduate School has verified and approved the above-named committee members, and
certifies that the thesis has been approved in accordance with university requirements.
ii
I dedicate this thesis to my family: Cindy (Mom), Jim (Dad), Kristen (Sister), Danielle (Sister),
and Kayli (Sister) and friends. Whom without their love and support none of this would have
been possible.
iii
ACKNOWLEDGEMENTS
The manuscript benefited from comments by my committee, which consisted of D.
Thistle (major professor), W. Landing, and I. MacDonald. C. Armstrong, S. Bode, S. Bourgoin,
M. Bublitz, E. Carroll, R. Carvalho, E. Darrow, S. Dorado, E. Easton, R. Rowland, A. Skinner
McInnes, F. Stephenson, and G. D. F. Wilson helped at sea or in the laboratory. J. Herring
advised on sediment particle-size distribution procedures. C. Mundoma, M. Santema, and M.
Seavy helped with the chlorophyll analyses. H. Bik and W. Decraemer helped with nematode
identification techniques. S. Ramsier helped with statistical methods. Cal-Atlas, Oregon Coastal
Atlas, and the USGS for providing GIS map files. This material is based upon work supported
by the National Science Foundation under Grant No. 0727243. I am grateful for all this kind
assistance.
iv
TABLE OF CONTENTS
List of Tables ................................................................................................................................. vi
List of Figures .............................................................................................................................. viii
Abstract .......................................................................................................................................... xi
1.
INTRODUCTION ...................................................................................................................1
2.
MATERIALS AND METHODS ............................................................................................3
2.1
2.2
2.3
2.4
2.5
3.
RESULTS ................................................................................................................................9
3.1
3.2
3.3
3.4
4.
Stations...........................................................................................................................3
Sampling and Processing of Meiofauna ........................................................................5
Sample and Processing for Environmental Variables ....................................................5
Faunal-Abundance Comparisons ...................................................................................7
Statistical Analyses ........................................................................................................7
The Fauna.......................................................................................................................9
Question 1 ......................................................................................................................9
Question 2a ..................................................................................................................20
Question 2b ..................................................................................................................26
DISCUSSION........................................................................................................................29
4.1
4.2
4.3
The Fauna.....................................................................................................................29
Question 1 ....................................................................................................................29
Question 2 ....................................................................................................................34
REFERENCES .............................................................................................................................37
BIOGRAPHICAL SKETCH .........................................................................................................41
v
LIST OF TABLES
Table 1. Station information. The positions and depths are the averages of the two multiplecorer lowerings that I used from a given station. Temperature, salinity, and oxygen were
measured ~50 m above bottom. No measurements (N/A) were made at station 3 (bad weather).
Temp = temperature. ........................................................................................................................4
Table 2. Absolute abundance in the 0-to-1-cm layer (surface area = 78.5 cm2) for each
meiofaunal major taxon and the nematode family Desmoscolecidae (= Desmos), showing the
variation among stations. Kinors = Kinorhynchs. Note that the sum of the Desmoscolecidae and
the non- Desmoscolecidae yields the total abundance of nematodes in the sample. .....................10
Table 3. Relative abundance in the 0-to-1-cm sediment layer (surface area = 78.5 cm2)
expressed as a percent for each meiofaunal major taxon and the nematode family
Desmoscolecidae (= Desmos), showing the variation among stations. Kinors = Kinorhynchs.
Note that the sum of the Desmoscolecidae and the non-Desmoscolecidae yields the abundance of
the nematodes in the sample. The reported error is the standard deviation. .................................12
Table 4. Reports of abundances of meiofaunal major taxa from 1000 to 5500 m depth that used
techniques sufficiently similar to mine to allow comparisons. All studies used a multiple corer.
The position for Radziejewska (2002) is the midpoint of the depth range she reported.. .............13
Table 5. Two-way (depth crossed with region) ANOVA results for each taxon based on
log10(X+1) transformed absolute abundance, showing the significant results... ...........................14
Table 6. Percent of sediment in Wentworth-scale size fractions by dry weight, showing that the
sediment in all samples was more than 70% mud (sediment particles smaller than 62 μm)... ......16
Table 7. Food-related variables. The carbon and nitrogen values are percent by sediment dry
weight, enzyme hydrolyzable amino acids (= EHAA) values are mg EHAA per gram dry weight
of sediment, and pigments are chlorophyll a concentrations (= Chl a) and phaeopigment
concentrations (= Phaeo) in ng g-1 sediment.... ..............................................................................17
Table 8. Significant p-values for Kendall-tau correlations based on log10(X+1)-transformed,
absolute-abundance data. %Mud = the percentage of sediment particles smaller than 62 μm by
dry weight. Chl a = chlorophyll a concentration in the sediment. %C = carbon as a percentage
of sediment dry weight. %N = nitrogen as a percentage of sediment dry weight. EHAA =
concentration of enzyme hydrolizable amino acids in the sediment..... ........................................18
Table 9. Two-way (depth crossed with region) ANOVA results for each taxon based on arcsinesquare-root transformed relative abundances, showing the significant results..............................18
Table 10. Significant p-values for Kendall-tau correlations based on arcsine-square-roottransformed, relative-abundance data. %Mud = the percentage by weight of sediment particles
smaller than 62 μm. Chl a = chlorophyll a concentrations in the sediment. %C = carbon as a
vi
percentage of sediment dry weight. %N = nitrogen as a percentage of sediment dry weight.
EHAA = concentration of enzyme hydrolizable amino acids in the sediment....... .......................20
Table 11. The primary differences responsible for the bifurcations in Fig. 8. Abundance =
number of individuals from the 0-to-1-cm layer of sediment from a core with a surface area of
78.5 cm2........ .................................................................................................................................23
Table 12. The primary differences responsible for the bifurcations in Fig. 10......... ...................26
Table 13. Significant LINKTREE bifurcations from Fig. 12 and the environmental difference
associated with each. Chl a = chlorophyll a in ng g-1 sediment dry weight.......... .......................27
Table 14. Significant LINKTREE bifurcations from Fig. 13 and the environmental difference(s)
associated with each. Chl a = chlorophyll a concentration in ng g-1 dry weight of sediment. %N
= nitrogen as a percentages of sediment dry weight........... ...........................................................28
vii
LIST OF FIGURES
Figure 1. Chart of the sea off the west coast of the United States, showing the locations of the
regions and stations. .........................................................................................................................3
Figure 2. Abundance per 10 cm2 in the top 1 cm of sediment versus depth. Note that y-axis
scales differ among graphs. Circles = this study. Squares = Radziejewska (2002). Diamonds =
Itoh et al. (2011). Triangles = Ingles et al. (2009). Radziejewska’s (2002) copepod values are
for harpacticoid copepods. .............................................................................................................14
Figure 3. Untransformed copepod absolute-abundance data in the 0-1-cm layer of sediment
versus depth by region, showing the decrease with depth for all regions. Region B (off northern
California) = circles. Region C (off central California) = triangles. Region D (off southern
California) = squares. Points are the averages of two samples at each depth. ..............................15
Figure 4. Untransformed kinorhynch absolute-abundance data versus region in the 0-1-cm layer
of sediment, showing that it is greatest in Region C. Open diamonds are the average for a region
(n = 4). Horizontal lines indicate minimum and maximum values. Region A is off Oregon.
Region B is off northern California. Region C is off central California. Region D is off southern
California.. .....................................................................................................................................15
Figure 5. Untransformed Desmoscolecidae absolute-abundance data versus depth by region,
showing the decrease with depth for Regions B (off northern California = circles) and D (off
southern California = squares) and an increase for Region C (central California = triangles).
Points are the averages of the two samples at each depth..............................................................16
Figure 6. Untransformed relative abundance-data of nematodes that do not belong to the family
Desmoscolecidae versus region in the 0-1-cm layer of sediment, showing that it is lowest in
Region C. Open diamonds are the average of the four samples in a region. Horizontal lines
indicate minimum and maximum values. Region A is off Oregon. Region B is off northern
California. Region C is off central California. Region D is off southern California.... ...............19
Figure 7. Untransformed relative-abundance data of kinorhynchs versus region in the 0-1-cm
layer of sediment, showing that it is greatest in Region C. Open diamonds are the average of the
four samples in a region. Horizontal lines indicate minimum and maximum values. Region A is
off Oregon. Region B is off northern California. Region C is off central California. Region D is
off southern California..... ..............................................................................................................19
Figure 8. The dendrogram from the cluster analysis of the log10(X+1)-transformed, absoluteabundance data, showing the five significant groupings of samples found by the PRIMER v6
SIMPROF procedure. Solid horizontal lines (labeled I to IV) indicate statistically significantly
bifurcations. The taxa included in the analysis were the nematode family Desmoscolecidae, nonDesmoscolecidae nematodes, copepods, kinorhynchs, and ostracods. The branch ends are
labeled by sample number, region of collection, and relative depth. Region A is off Oregon.
Region B is off northern California. Region C is off central California. Region D is off southern
California...... .................................................................................................................................21
viii
Figure 9. A nonmetric, multidimensional-scaling plot for the log10(X+1)-transformed, absoluteabundance data, showing the five groups of samples. Samples represented by the same symbol
do not differ significantly from each other but do differ significantly from all other samples. The
taxa included were non-Desmoscolecidae nematodes, the nematode family Desmoscolecidae,
copepods, kinorhynchs, and ostracods. The points are labeled by sample number, region of
collection, and relative depth. Region A is off Oregon. Region B is off northern California.
Region C is off central California. Region D is off southern California. The labeled circles
around the groups correspond to the bifurcations in Fig. 8....... ....................................................22
Figure 10. The dendrogram from the cluster analysis of the log10(X+1)-transformed, relativeabundance data, showing the four significant groupings of samples. Solid horizontal lines
(labeled I to III) indicate statistically significantly bifurcations. The taxa included in the analysis
were the nematode family Desmoscolecidae, non-Desmoscolecidae nematodes, copepods,
kinorhynchs, and ostracods. The branch ends are labeled by sample number, region of
collection, and relative depth. Region A is off Oregon. Region B is off northern California.
Region C is off central California. Region D is off southern California........ ..............................24
Figure 11. A nonmetric, multidimensional-scaling plot for the log10(x+1)-transformed, relativeabundance data, showing the four groups of samples. Samples represented by the same symbol
do not differ significantly from each other but do differ significantly from all other samples. The
taxa included were non-Desmoscolecidae nematodes, the nematode family Desmoscolecidae,
copepods, kinorhynchs, and ostracods. The points are labeled by sample number, region, and
relative depth. Region A is off Oregon. Region B is off northern California. Region C is off
central California. Region D is off southern California. The labeled circles around the groups
correspond to the bifurcations in Fig. 10........ ...............................................................................25
Figure 12. The dendrogram of samples produced by the PRIMER v6 (Clarke and Gorley, 2006)
procedure LINKTREE based on chl a and absolute abundance values. Bifurcations labeled with
Roman numerals were found to be significant by the PRIMER v6 procedure SIMPROF. The
ends of the branches are labeled by sample number, region of collection (A = off Oregon, B = off
northern California, C = off central California, D = off southern California), and relative depth (S
= shallow, D = deep). ........ ...........................................................................................................27
Figure 13. The dendrogram of samples produced by LINKTREE (PRIMER v6, Clarke and
Gorley, 2006) based on chlorophyll a, percent nitrogen, and relative meiofaunal abundance.
Bifurcations labeled with Roman numerals were found to be significant by the PRIMER v6
procedure SIMPROF. The ends of the branches are labeled by sample number, region of
collection (A = off Oregon, B = off northern California), C = off central California, D = off
southern California, and relative depth (S = shallow, D = Deep). ........ .......................................28
Figure 14. Chlorophyll a concentration (in ng g-1 sediment dry weight) versus depth, showing
that as depth increases in a region, chlorophyll a concentration in the sediment declines. Region
B (off northern California) = circles. Region C (off central California) = triangles. Region D
(off southern California) = squares........ ........................................................................................30
ix
Figure 15. Kinorhynch abundance versus chl a concentration (in ng g-1 dry weight of sediment),
showing that both variables are highest in Region C. Region B (off northern California) =
circles. Region C (off central California) = triangles. Region D (off southern California) =
squares............................................................................................................................................31
Figure 16. Ostracod abundance versus the percent of the sediment by weight that was less than
0.062 mm (= percent mud), showing that the abundance of ostracods is higher in less muddy
sediment. Region B (off northern California) = circles. Region C = triangles (off central
California). Region D (off southern California) = squares.......... .................................................32
Figure 17. Untransformed, relative abundance by region, showing the proportional increase of
kinorhynchs and the proportional decrease of non-Desmoscolecidae nematodes in Region. For
each taxon, the proportion in each of the 16 samples is indicated by a bar. For each taxon, the
bars are in numerical order by sample number. Black bars = samples from Region A off Oregon.
Dark-grey bars = samples from Region B off northern California. Light-grey bars = samples
from Region C off central California. Gridded bars = Region D off southern California........... .33
x
ABSTRACT
Deep-sea meiofauna along the continental rise off the west coast of North America are
nearly unstudied. To begin to remedy this situation, I sampled each of four regions (off Oregon
and off northern, central, and southern California) for meiofaunal major taxa (i.e., nematodes,
copepods, ostracods, and kinorhynchs) and for environmental factors in the band between 2700
and 3700 m depth. The abundances in each region fit well with those from previous reports from
the Pacific. For each meiofaunal major taxon, abundance was greatest off central California.
Food, as measured by chlorophyll a concentration in the sediment, was the best explanatory
variable except for the ostracods, for which abundance was best explained by the proportion of
mud in the sediment. Unexpectedly, only copepod abundance decreased significantly with depth
in the 1000-m wide band studied. My assessment of multivariate similarity revealed that the
Bray-Curtis similarity in meiofaunal major taxa between samples was no less than 89% despite
as much as 1500 km of separation, so the basic composition of the meiofauna along the rise in
this depth band varies little. Despite the overall similarity, I found groups composed of samples
that were not significantly different from each other but were significantly different from all
other samples. The groups were formed primarily on differences among samples in abundance
and were best explained by variations in the amount of food in the sediment.
xi
CHAPTER ONE
INTRODUCTION
Whatever depth one chooses to use to define the beginning of the deep sea, the habitat is
huge. Its surface area is greater than that of the continents, and mud covers most of it (see, e.g.,
Gage and Tyler, 1991). The extreme conditions that characterize this environment (e.g., high
hydrostatic pressure, near-freezing temperatures, constant darkness, and meager food supply) led
early workers to believe that the sediment-covered, deep-sea floor (hereafter the deep-sea floor)
contained few species (see Anderson and Rice, 2006; Ramirez-Llodra et al., 2010; Snelgrove and
Smith, 2002 for a review). In reality, it is one of the most species-rich habitats on the planet
(Grassle and Maciolek, 1992; Hessler and Sanders, 1967; see Snelgrove and Smith, 2002, for a
review). Further, it is of interest as an extreme environment (e.g., Young and Tyler, 1993) and is
an important provider of ecosystem services (Snelgrove et al., 1997; Snelgrove, 1999). The
work on the basic description of the distribution of taxa in the deep sea has a long history (see,
e.g., Murray, 1895) but is far from complete, particularly for the smaller animals.
Metazoan meiofauna (animals with body sizes of tens of micrometers to a few
millimeters) are ubiquitous in shallow-water sediments, where they play many important roles.
For example, they help move matter and energy up the food chain (see, Giere, 2008, for a
review), and they are food for the juveniles of many fish species (see, e.g., De Morais and
Bodiou, 1984; McCall and Fleeger, 1995). Metazoan meiofauna also inhabit the deep-sea floor
(see, e.g., Vincx et al., 1994; Wigley and Mcintyre, 1964). Because their abundance is
proportionately greater in the deep sea than in shallow water (Shirayama, 1983; Thiel, 1979),
their functional importance may be as well.
In this chapter, I will report on the distribution of the metazoan meiofauna (hereafter
meiofauna) in a depth band along the continental rise off the west coast of North America from
Oregon to southern California, a region where the abundance of meiofauna rarely has been
studied (see e.g., Carmen et al., 2004). In addition, I asked questions of ongoing interest to deepsea ecologists. (1) Do the abundances of the major meiofaunal taxa differ along the continental
rise (e.g., Danovaro et al., 1995; Dinet, 1979) or with depth (e.g., Baguley et al., 2006; de Bovée
et al., 1990; Shirayama, 1984; Soltwedel, 2000), and do any differences correlate with
differences in the environment (e.g., Thiel, 1979; Shirayama, 1984; Soltwedel, 2000)? To
1
supplement this univariate approach, I asked (2) whether samples can be grouped based on their
multivariate similarity in meiofaunal major taxa, and whether these groupings correlate with
differences in the environment.
2
CHAPTER TWO
MATERIALS AND METHODS
2.1. Stations
I planned to collect samples from a station near the 2700-m isobath and a station near the
3700-m isobath in four regions off the west coast of the United States (Fig. 1), but the depth
pattern was not followed in the northernmost region. At each station, the science team (hereafter
the team) made 4 to 7 successful lowerings with a version of the Barnett et al., (1984) multiple
corer (Ocean Instruments MC 800 Multi Core, San Diego, CA). For more information, see
Table 1.
Figure 1. Chart of the sea off the west coast of the United States, showing the locations of the
regions and stations.
3
125.8804°W
4
39.9965°N
125.4416°W
5
36.7993°N
123.6945°W
6
36.6824°N
122.8245°W
7
32.8752°N
120.6141°W
8
32.7974°N
120.3701°W
C
D
37.4
961
78.6
945
24.4
1145
4
[Oxygen] (mg L-1)
39.9925°N
Salinity
3
B
Temp (°C)
131.9200°W
Depth (m)
42.5599°N
357
0.87
3232
1.55
34.66
3.41
0.44
3589
1.56
34.67
3.44
1.87
3682
N/A
N/A
N/A
0.46
2721
1.69
34.65
2.91
0.39
3669
1.51
34.69
3.20
0.51
2724
1.66
34.66
3.09
1.17
3853
1.57
34.67
3.24
0.20
2708
1.80
34.64
2.80
lowerings (km)
2
202.3
Distance between
130.3953°W
between stations (m)
44.0031°N
Depth difference
1
stations(km)
Station
A
Position
Distance between
Region
Table 1. Station information. The positions and depths are the averages of the two multiple-corer lowerings that I used from a given
station. Temperature, salinity, and oxygen were measured ~50 m above bottom. No measurements (N/A) were made at station 3 (bad
weather). Temp = temperature.
2.2. Sampling and Processing of Meiofauna
The multiple corer had eight 0.1-m (inner diameter) core tubes. In a cold room (1316°C), the team processed each core from a lowering in a predetermined, random order. The
team imaged the side of the transparent core tube, noted any unusual features, collected the water
overlying the sediment, imaged the sediment surface, and collected the top 0.01 m of sediment.
The water and sediment samples were combined, preserved with cold, 95% ethyl alcohol, and
stored at -20ºC.
At home, the team randomly selected a station, a lowering within that station, and an
undisturbed core from that lowering. They repeated this process twice, yielding 16 samples. For
a given sample, the team measured its volume and separated the macrofauna from the meiofauna
using stacked 300-μm and 30-μm sieves. Specimens of meiofaunal taxa (i.e., nematodes,
copepods, kinorhynchs, and ostracods) retained in the 300-μm sieve were included in the
meiofaunal counts. The team used a silica sol-flotation technique inspired by Burgess (2001) to
separate most of the meiofauna from the sediment. That is, they added 10-15 ml of the sample
and 30 ml of Ludox® HS-40 (E.I. du Pont de Nemours & Co., Inc., Wilmington, Delaware) to a
50-ml centrifuge tube. This step was repeated until the entire sample was placed in tubes. The
team homogenized the contents of each tube on a vortexer for 5 min and centrifuged it at 900 × g
for 5 min. They poured the supernatant into a 30-μm sieve while retaining the concentrated
sediment (hereafter the pellet) in the centrifuge tube. The team saved the Ludox® from the first
extraction, measured its specific gravity, and repeated the extraction using this Ludox ®. After
the second extraction, the team combined the two supernatant fractions. The team stained the
supernatant sample, the pellet sample, and the 300-μm sample with rose bengal. I hand sorted
them under a stereomicroscope at 25×. A second person resorted the supernatant and 300-μm
samples. The sorter removed and counted nematodes, copepods, kinorhynchs, ostracods, and
noted which nematodes belonged to the family Desmoscolecidae. When I report total nematode
abundances, the Desmoscolecidae individuals are included.
2.3. Sampling and Sample Processing for Environmental Variables
2.3.1. Sample handling at sea. From a randomly selected core from each Multiple Core
lowering, the team removed and discarded the water overlying the sediment and inserted cut-off
syringes. All syringes were inserted into the core before any were removed. All the
environmental variables were measured in the top 1-cm of sediment. The particle-size5
distribution samples were preserved in a 10% solution of formalin and reagent-grade sodium
chloride in distilled water, which was buffered to neutrality with sodium bicarbonate. The
samples for enzyme-hydrolyzable amino acids (= EHAA; see Mayer et al., 1995) were frozen at
-20ºC. The pigment (chlorophyll a and phaeopigments) and carbon-nitrogen (percent C and
percent N) samples were frozen at -20ºC in new, foil-closed, scintillation vials that had been
combusted at 450°C for 4 h. The pigment samples were stored in the dark.
2.3.2. Particle-size distributions. To determine the particle-size distribution, the team first
separated the sediment into a course and a fine fraction by pouring the sample onto a 30-μm
sieve that was in a container filled with deionized water and gently moved the sieve up and down
for 5 min. To separate the sediment that passed through the 30-μm sieve into size fractions, the
team first dried nine Whatman GF/D filters over night at 60°C, let them cool, and weighed them.
The team then rinsed the sediment remaining on the 30-μm sieve through nested sieves of 500,
350, 250, 177, 125, 88, 62.5, 45, and 30-μm aperture mesh (see Folk, 1968). The contents of
each sieve were then vacuum filtered onto a filter, which was then dried over night at 60ºC,
cooled, and reweighed. For the sediment that passed through the 30-μm, the team first prepared
filter “stacks” by placing a Whatman GF/D filter on a 0.1-micron, Whatman polycarbonate
membrane filter. The stack was dried overnight at 60°C, cooled, and weighed. The team then
vacuum filtered the contents of the rinsing container onto the filter stack, which was then dried
over night at 60º, cooled, and reweighed.
2.3.3. Chlorophyll a and phaeopigments. For chlorophyll a and phaeopigment analysis, the
team transferred each sample to a tared microcentrifuge tube, weighed it, freeze-dried it, and
reweighed it. The team added 600µl of 90% acetone to the dried sample, vortexed the contents
of the tube for 15–30 s, and froze it overnight. The following day, the sample was vortexed long
enough to resuspend the sediment and was centrifuged for 2 min to remove particles from the
supernatant. A Varian Cary Eclipse Fluorescence Spectrophotometer was used to measure
dissolved pigments (5 nm excitation and emission slits, 600 v voltage, excitation at 410 nm, and
emission from 600–800 nm). After measurement, two drops of 10% HCl were added, and the
sample was remeasured to quantify phaeopigments.
2.3.4. Carbon and nitrogen. For these measurements, the sample was dried overnight at 60ºC,
ground until no grit remained, and ~200 milligrams were transferred to a preweighed,
microcentrifuge tube, which was then reweighed. To remove inorganic carbon, ~twice the
6
sediment volume of 5% HCl was added, allowed to sit for 2 h, and centrifuged for 30 s at 800 ×
g. After the supernatant was removed, one milliliter of Milli-Q water was added, centrifuged for
30 s at 800 × g, and the supernatant removed. The team repeated these two steps five times. The
sample was then centrifuged at 16,100 × g for 60 s and dried at 60°C. The measurements were
made at U. C. Davis (see, http://stableisotopefacility.ucdavis.edu).
2.3.5. Enzyme hydrolyzable amino acids. The samples were freeze dried and sent to Kathleen
R. Hardy (University of Maine) who analyzed them following Mayer et al. (1995), except that
the samples were measured on a microplate reader instead of with a spectrometer.
2.4. Faunal-Abundance Comparisons
For comparisons, I chose stations between 1000 and 5500 m depth from the Pacific
Ocean. Because meiofaunal abundance estimates are sensitive to sampling and processing
techniques, I used data from samples collected with those devices thought to be the most
accurate, i.e., multiple corers, remotely operated vehicles, and research submersibles (Bett et al.,
1994; Shirayama and Fukushima, 1995; see Thistle, 2003, for discussion) and the mesh size was
no larger than 0.062 mm. I used data from the first centimeter of sediment and estimated values
from published figures when necessary. For Radziejewska (2002), I used her intact-site data and
assumed that they came from the midpoint of the depth range. I used her values, but note that
she only counted harpacticoid copepods.
2.5. Statistical Analyses
2.5.1. General. I studied total nematodes, the nematode family Desmoscolecidae (hereafter
Desmoscolecidae), total nematodes excluding the Desmoscolecidae (hereafter nonDesmoscolecidae), copepods, kinorhynchs, and ostracods. For univariate analyses, I used the
SAS (SAS Institute Inc, Cary, NC) 9.2 Analyst program. For multivariate analyses, I used
PRIMER version 6 (Clarke and Gorley, 2006). Any method written in capital letters except
“ANOVA” is a PRIMER routine. To display relationships among data points, I used MDS to
perform nonmetric multidimensional scaling (hereafter nMDS). To verify its utility in an
application, I checked the stress level, examined the Sheppard’s diagram, and compared the
nMDS plot to a group-averaged cluster analysis of the same resemblance matrix. In all the
nMDS plots, the stress value was acceptable according to Clarke and Warwick’s (2001)
standard. The two-dimensional plots were comparable to three-dimensional plots, so for ease of
interpretation, I report the former.
7
I used the 5% significance level throughout and did not correct for multiple testing.
2.5.2. Question 1. Do the abundances of meiofaunal major taxa change with region or
depth, and do the changes correlate with changes in the environment? The four regions
were chosen approximately equally spaced in latitude along the west coast of the United States.
Station 1 in Region A was not at the appropriate depth, so Region A was not used in analyses
involving depth. Prior to analysis, the absolute-abundance data were log10(X+1) transformed,
and the relative-abundance data were arcsine-square-root transformed. I then used two-way
ANOVA (with region and depth as fixed factors) for each taxon followed by a Tukey’s honestly
significant difference test (Sokal and Rohlf, 1997). For each taxon, I checked whether the data
met the assumptions of ANOVA by examining residual-and normal-probability plots.
I tested for a correlation between each meiofaunal major taxon and each environmental
variable with Kendall’s tau (correcting for ties).
2.5.3. Question 2. Can samples be grouped based on their similarity in meiofaunal major
taxa, and do these groupings correlate with differences in the environment? For this
question, I transformed the abundance data with (log10(X+1)) and used RESEMBLANCE to
calculate all pair-wise resemblances between samples with the Bray-Curtis measure of similarity
(Bray and Curtis, 1957). I used MDS to produce an nMDS plot and CLUSTER to create a
dendrogram. To identify groups of samples that were significantly more similar to each other
than to any other sample, I used SIMPROF (options = group-average clustering, 999 simulation
permutations). To visualize the results, I superimposed the SIMPROF groupings on the nMDS
plot.
To test whether station groupings correlated with environmental variables, I first
determined that no environmental variable required transformation before statistical analysis.
Because Draftsman plots revealed that chlorophyll a and phaeopigment concentrations were 95%
correlated, I included only chlorophyll a concentration (hereafter chl a).
To determine which of the environmental variables best explained the groupings of
samples, I normalized each environmental variable and used BEST (options = BIOENV,
Euclidean distances, Kendall rank correlation, 999 permutations). To determine how the
environmental variables grouped the samples, I used the LINKTREE and SIMPROF procedures
on the unnormalized environmental data.
8
CHAPTER THREE
RESULTS
3.1 The Fauna
The absolute abundance data for each meiofaunal taxon are in Table 2. For abundance
summed across the taxa I studied, the highest were in Region C (stations 5 and 6) and the lowest
were in Region A (stations 1 and 2). The relative abundance data are in Table 3.
3.2. Question 1. Do the Abundances of Meiofaunal Major Taxa Change with Region or
Depth, and do the Changes Correlate with Changes in the Environment?
3.2.1. Absolute abundance. When I asked whether abundance of a taxon differed among
regions or depths (Table 5), I found that (1) copepod abundance decreased significantly with
depth in the three regions included in the ANOVA analysis (Fig. 3). (2) Kinorhynchs were
significantly more abundant in Region C than in Regions B or D (Fig. 4). Although I did not use
the Region A data in the ANOVA, I have included it in Fig. 4 to show that the abundance of
kinorhynchs in Region A is also lower than in Region C. (3) For the Desmoscolecidae, depth
and region interacted significantly, so the apparent difference among regions (Table 5) cannot be
interpreted. The interaction arises (Fig. 5) because in Regions B and D the abundance at the
shallow station was greater than at the deep station, as would be expected (see, e.g., de Bovée et
al., 1990; Shirayama, 1984), but in Region C, the abundance at the deeper station was greater
than at the shallow station.
When I asked if the abundance of individual meiofaunal major taxa correlated with
environmental variables (Tables 6 and 7), I found different results for each taxon (Table 8). (1)
As the sediment became increasingly muddy (i.e., the percentage of sediment particles smaller
than 62 μm), ostracod abundance decreased. (2) The other taxa correlated with two or more of
the measures of food availability. In particular, as chl a increased, the abundance of all taxa
except ostracods increased.
9
Table 2. Absolute abundance in the 0-to-1-cm layer (surface area = 78.5 cm2) for each meiofaunal major taxon and the nematode
family Desmoscolecidae (= Desmos), showing the variation among stations. Kinors = Kinorhynchs. Note that the sum of the
Desmoscolecidae and the non- Desmoscolecidae yields the total abundance of nematodes in the sample.
Sample
Non-
Region
Station
name
Desmos
Desmoscolecidae
A
1
1331
156
1425
72
10
17
1680
1421
192
821
131
12
24
1180
2341
158
1608
76
8
15
1865
2431
121
1669
129
7
9
1935
Average
156.8
1380.8
102.0
9.3
16.3
1665.0
Standard deviation
±29.0
±387.3
±32.4
±2.2
±6.2
±340.8
3331
232
1205
149
17
25
1628
3421
329
1560
157
26
25
2097
4111
451
1338
248
29
25
2091
4271
644
2569
208
37
25
3483
Average
414.0
1668.0
190.5
27.3
25.0
2324.8
Standard deviation
±177.6
±618.3
±46.4
±8.3
±0
±802.2
5231
958
2676
179
92
11
3916
5351
1033
2233
150
115
9
3540
6111
595
1974
311
68
23
2971
2
B
3
4
C
5
6
10
Copepods Kinors Ostracods
Total
Table 2 - continued
Sample
Non-
name
Desmos
Desmoscolecidae
6261
647
2123
273
109
13
3165
Average
808.3
2251.5
228.3
96
14.0
3398.0
Standard deviation
±219.4
±302.2
±76.2
±21.1
±6.2
±418.4
7161
203
1680
150
22
18
2073
7451
196
1621
147
15
15
1994
8151
557
4083
121
25
8
4794
8481
270
2470
220
31
29
3020
Average
306.5
2463.5
159.5
23.3
17.5
2970.3
Standard deviation
±170.3
±1147.0
±42.4
±6.7
±8.7
±1302.1
Region
D
Station
7
8
11
Copepods Kinors Ostracods
Total
Table 3. Relative abundance in the 0-to-1-cm sediment layer (surface area = 78.5 cm2)
expressed as a percent for each meiofaunal major taxon and the nematode family
Desmoscolecidae (= Desmos), showing the variation among stations. Kinors = Kinorhynchs.
Note that the sum of the Desmoscolecidae and the non-Desmoscolecidae yields the abundance of
the nematodes in the sample. The reported error is the standard deviation.
Sample
Non-
Region
Station
Name
A
1
1331
9.3
84.8
4.3
0.6
1.0
1421
16.3
69.6
11.1
1.0
2.0
2341
8.5
86.2
4.1
0.4
0.8
2431
6.3
86.3
6.7
0.4
0.5
10.1
81.7
6.5
0.6
1.1
(±4.3)
(±8.1)
(±3.3)
(±0.3)
(±0.7)
3331
14.3
74.0
9.2
1.0
1.5
3421
15.7
74.4
7.5
1.2
1.2
4111
21.6
64.0
11.9
1.4
1.2
4271
18.5
73.8
6.0
1.1
0.7
17.5
71.5
8.6
1.2
1.2
(±3.2)
(±5.0)
(±2.5)
(±0.2)
(±0.3)
5231
24.5
68.3
4.6
2.3
0.3
5351
29.2
63.1
4.2
3.2
0.3
6111
20.0
66.4
10.5
2.3
0.8
6261
20.4
67.1
8.6
3.4
0.4
23.5
66.2
7.0
2.8
0.4
(±4.3)
(±2.2)
(±3.1)
(±0.6)
(±0.2)
7161
9.8
81.0
7.2
1.1
0.9
7451
9.8
81.3
7.4
0.8
0.8
8151
11.6
85.2
2.5
0.5
0.2
2
Average
B
3
4
Average
C
5
6
Average
D
7
8
Desmos Desmoscolecidae Copepods Kinors
12
Ostracods
Table 3 - continued
Sample
Region
Station
Name
NonDesmos Desmoscolecidae Copepods Kinors
8481
Ostracods
8.9
81.8
7.3
1.0
1.0
10.0
82.3
6.1
0.8
0.7
(±1.1)
(±1.9)
(±2.4)
(±0.3)
(±0.4)
Average
When I compared abundances to those of other studies in the Pacific Ocean (Table 4), I
found that my values were comparable to those reported from similar depths (Fig. 2).
Table 4. Reports of abundances of meiofaunal major taxa from 1000 to 5500 m depth that used
techniques sufficiently similar to mine to allow comparisons. All studies used a multiple corer.
The position for Radziejewska (2002) is the midpoint of the depth range she reported.
Depth range
Sieve aperture
Source
Site
Position
(m)
(μm)
Ingels et al. (2009)
Slope Middle
40.5950°N
3400-3403
32
10.3680°W
Slope Deep
40.0730°N
4275-4277
10.3650°W
Itoh et al. (2011)
Kuril Trench region
41.8856°N
3320-5570
38
145.5047°E
Radziejewska (2002)
Ryukyu Trench
25.7015°N
region
128.5115°E
Clarion-Cliperton
11.0667°N
Fracture Zone
119.6667°W
13
1290-5330
4380-4430
32
Figure 2. Abundance per 10 cm2 in the top 1 cm of sediment versus depth. Note that y-axis
scales differ among graphs. Circles = this study. Squares = Radziejewska (2002). Diamonds =
Itoh et al. (2011). Triangles = Ingles et al. (2009). Radziejewska’s (2002) copepod values are
for harpacticoid copepods.
Table 5. Two-way (depth crossed with region) ANOVA results for each taxon based on
log10(X+1) transformed absolute abundance, showing the significant results.
Taxon
Depth
Region
Interaction
0.003
0.03
Total nematodes
Desmoscolecidae
Non-Desmoscolecidae
Copepods
0.02
Kinorhynchs
0.0003
Ostracods
14
Figure 3. Untransformed copepod absolute-abundance data in the 0-1-cm layer of sediment
versus depth by region, showing the decrease with depth for all regions. Region B (off northern
California) = circles. Region C (off central California) = triangles. Region D (off southern
California) = squares. Points are the averages of two samples at each depth.
Figure 4. Untransformed kinorhynch absolute-abundance data versus region in the 0-1-cm layer
of sediment, showing that it is greatest in Region C. Open diamonds are the average for a region
(n = 4). Horizontal lines indicate minimum and maximum values. Region A is off Oregon.
Region B is off northern California. Region C is off central California. Region D is off southern
California.
15
Figure 5. Untransformed Desmoscolecidae absolute-abundance data versus depth by region,
showing the decrease with depth for Regions B (off northern California = circles) and D (off
southern California = squares) and an increase for Region C (central California = triangles).
Points are the averages of the two samples at each depth.
Table 6. Percent of sediment in Wentworth-scale size fractions by dry weight, showing that the
sediment in all samples was more than 70% mud (sediment particles smaller than 62 μm).
Sample
Coarse sand
Medium
Fine
Very fine
Station
name
and larger
sand
sand
sand
Mud
1
1331
0.9
0.9
2.8
3.6
91.8
1421
0.7
1.3
2.5
4.4
91.1
2341
0.6
0.5
1.3
1.2
96.4
2431
0.7
0.3
0.5
1.0
97.5
3331
0.3
1.0
3.3
1.6
93.8
3421
0.4
0.7
1.8
2.8
94.4
4111
2.6
5.0
10.6
9.5
72.3
4271
0.7
3.7
10.8
14.0
70.8
5231
0.1
0.6
1.6
2.3
95.4
5351
0.2
0.8
1.5
2.4
95.1
6111
0.8
3.5
2.7
2.4
90.6
2
3
4
5
6
16
Table 6 - continued
Station
7
8
Sample
Coarse sand
Medium
Fine
Very fine
name
and larger
sand
sand
sand
Mud
6261
0.1
1.9
1.3
1.3
95.4
7161
0.3
0.3
3.4
2.3
93.7
7451
0.2
0.6
2.4
2.3
94.5
8151
0.0
1.0
3.9
8.4
86.7
8481
0.6
3.2
9.9
9.8
76.5
Table 7. Food-related variables. The carbon and nitrogen values are percent by sediment dry
weight, enzyme hydrolyzable amino acids (= EHAA) values are mg EHAA per gram dry weight
of sediment, and pigments are chlorophyll a concentrations (= Chl a) and phaeopigment
concentrations (= Phaeo) in ng g-1 sediment.
Station Lowering
1
2
3
4
5
6
7
8
Chl a
Phaeo %C %N EHAA
1-3
40
84
1.1
0.2
0.5
1-4
8
126
1.1
0.2
0.5
2-3
47
84
0.6
0.1
0.3
2-4
53
107
0.6
0.1
0.4
3-3
83
214
1.3
0.2
0.7
3-4
150
223
1.5
0.2
0.8
4-1
347
509
1.1
0.2
0.6
4-2
506
692
1.0
0.2
0.8
5-2
900
1142
2.8
0.4
0.7
5-3
858
1745
2.8
0.4
0.5
6-1
1391
1535
3.1
0.4
0.9
6-2
1478
1964
3.0
0.4
0.8
7-1
269
497
1.6
0.3
0.7
7-4
183
246
1.8
0.3
0.6
8-1
683
839
5.2
0.4
0.8
8-4
227
174
5.8
0.3
0.6
17
Table 8. Significant p-values for Kendall-tau correlations based on log10(X+1)-transformed,
absolute-abundance data. %Mud = the percentage of sediment particles smaller than 62 μm by
dry weight. Chl a = chlorophyll a concentration in the sediment. %C = carbon as a percentage
of sediment dry weight. %N = nitrogen as a percentage of sediment dry weight. EHAA =
concentration of enzyme hydrolizable amino acids in the sediment.
Taxon
%Mud
Chl a
%C
%N
EHAA
Total nematodes
0.0006
0.0117 0.0068 0.0384
Desmoscolecidae
0.0002
0.0241 0.0150
Non-Desmoscolecidae
0.0040
Copepods
0.0025
0.0132
Kinorhynchs
0.0002
0.0244 0.0115 0.0192
Ostracods
0.0117 0.0190
0.0157
3.2.2. Relative abundance. I asked whether the relative abundance of a taxon differed among
regions or depths (Table 9). I found that (1) the relative abundance of non-Desmoscolecidae was
significantly less in Region C than in Region B or D (Fig. 6). (2) The relative abundance of
kinorhynchs was significantly greater in Region C than in Region B or D (Fig. 7). Although I
did not use Region A in the ANOVAs, I included it in Figs. 6 and 7. For the Desmoscolecidae,
depth and region interacted significantly, so the apparent differences among regions cannot be
interpreted.
Table 9. Two-way (depth crossed with region) ANOVA results for each taxon based on arcsinesquare-root transformed relative abundances, showing the significant results.
Taxon
Depth
Region Interaction
Total nematodes
Desmoscolecidae
0.0001
Non-Desmoscolecidae
0.001
Copepods
Kinorhynchs
0.002
Ostracods
18
0.01
Figure 6. Untransformed relative abundance-data of nematodes that do not belong to the family
Desmoscolecidae versus region in the 0-1-cm layer of sediment, showing that it is lowest in
Region C. Open diamonds are the average of the four samples in a region. Horizontal lines
indicate minimum and maximum values. Region A is off Oregon. Region B is off northern
California. Region C is off central California. Region D is off southern California.
Figure 7. Untransformed relative-abundance data of kinorhynchs versus region in the 0-1-cm
layer of sediment, showing that it is greatest in Region C. Open diamonds are the average of the
four samples in a region. Horizontal lines indicate minimum and maximum values. Region A is
off Oregon. Region B is off northern California. Region C is off central California. Region D is
off southern California.
19
When I asked whether the relative abundances of a given major meiofaunal taxa
correlated with any environmental variables, I found significant correlations only with measures
of food in the sediment (Table 10). In particular, chl a was positively correlated with the
Desmoscolecidae, negatively correlated with the non-Desmoscolecidae , negatively correlated
with the ostracods , and positively correlated with the kinorhynchs. There were no significant
correlations for nematodes and copepods.
Table 10. Significant p-values for Kendall-tau correlations based on arcsine-square-roottransformed, relative-abundance data. %Mud = the percentage by weight of sediment particles
smaller than 62 μm. Chl a = chlorophyll a concentrations in the sediment. %C = carbon as a
percentage of sediment dry weight. %N = nitrogen as a percentage of sediment dry weight.
EHAA = concentration of enzyme hydrolizable amino acids in the sediment.
Taxon
%Mud
Chl a
%C
%N
EHAA
Total nematodes
Desmoscolecidae
0.0150
Non-Desmoscolecidae
0.0384
Copepods
Kinorhynchs
0.0025
Ostracods
0.0150
0.0272
0.0471
3.3. Question 2a. Can Samples be Grouped Based on their Multivariate Similarity in
Meiofaunal Major Taxa?
3.3.1. Absolute abundance. To answer this question for absolute abundances (Table 2), I used
two approaches. With cluster analysis (Fig. 8), I showed that all samples are at least 91% similar
to each other, even though I used a data transformation that reduced sample-to-sample similarity.
When I analyzed the resemblance matrix with SIMPROF, I found five groups of samples that
differed significantly from each other. I labeled the bifurcations that separate the samples into
groups I to IV (Fig. 8). The percent similarity at the bifurcation from the cluster analysis and the
probability level from the SIMPROF procedure for each follow: I (91.81%, p = 0.001), II
(93.89%, p = 0.009), III (94.23%, p = 0.002), and IV (96.3%, p = 0.005). Table 11 explains the
20
reasons for the bifurcations. The sample groupings can also be seen in an nMDS plot (Fig. 9),
which represents similarities in a continuous rather than a tree-like manner.
Figure 8. The dendrogram from the cluster analysis of the log10(X+1)-transformed, absoluteabundance data, showing the five significant groupings of samples found by the PRIMER v6
SIMPROF procedure. Solid horizontal lines (labeled I to IV) indicate statistically significantly
bifurcations. The taxa included in the analysis were the nematode family Desmoscolecidae, nonDesmoscolecidae nematodes, copepods, kinorhynchs, and ostracods. The branch ends are
labeled by sample number, region of collection, and relative depth. Region A is off Oregon.
Region B is off northern California. Region C is off central California. Region D is off southern
California.
21
Figure 9. A nonmetric, multidimensional-scaling plot for the log10(X+1)-transformed, absoluteabundance data, showing the five groups of samples. Samples represented by the same symbol
do not differ significantly from each other but do differ significantly from all other samples. The
taxa included were non-Desmoscolecidae nematodes, the nematode family Desmoscolecidae,
copepods, kinorhynchs, and ostracods. The points are labeled by sample number, region of
collection, and relative depth. Region A is off Oregon. Region B is off northern California.
Region C is off central California. Region D is off southern California. The labeled circles
around the groups correspond to the bifurcations in Fig. 8.
22
Table 11. The primary differences responsible for the bifurcations in Fig. 8. Abundance =
number of individuals from the 0-to-1-cm layer of sediment from a core with a surface area of
78.5 cm2.
Left side of
Right side of
Bifurcation
bifurcation
bifurcation
I
< 2080 total abundance
> 2080 total abundance
II
> 4000 total abundance
< 4000 total abundance
III
< 40 Kinorhynchs
> 40 Kinorhynchs
IV
> 700 Desmoscolecidae
< 700 Desmoscolecidae
and > 3200 total abundance
and < 3200 total abundance
3.3.2. Relative abundances. I followed the same procedures to analyze the relative-abundance
data (Table 3) and found that the samples were at least 89% similar (Fig. 10), even with the preanalysis transformation that decreased sample-to-sample similarity. Four groups of samples
differed significantly from one another. I report information on the bifurcations in the format
used above: I (89.17%, p = 0.001), II (89.5%, p = 0.001), and III (91.88%, p = 0.001). Table 12
gives an explanation of the bifurcations. I also show the sample groupings in an nMDS plot
(Fig. 11).
23
Figure 10. The dendrogram from the cluster analysis of the log10(X+1)-transformed, relativeabundance data, showing the four significant groupings of samples. Solid horizontal lines
(labeled I to III) indicate statistically significantly bifurcations. The taxa included in the analysis
were the nematode family Desmoscolecidae, non-Desmoscolecidae nematodes, copepods,
kinorhynchs, and ostracods. The branch ends are labeled by sample number, region of
collection, and relative depth. Region A is off Oregon. Region B is off northern California.
Region C is off central California. Region D is off southern California.
24
Figure 11. A nonmetric, multidimensional-scaling plot for the log10(x+1)-transformed, relativeabundance data, showing the four groups of samples. Samples represented by the same symbol
do not differ significantly from each other but do differ significantly from all other samples. The
taxa included were non-Desmoscolecidae nematodes, the nematode family Desmoscolecidae,
copepods, kinorhynchs, and ostracods. The points are labeled by sample number, region, and
relative depth. Region A is off Oregon. Region B is off northern California. Region C is off
central California. Region D is off southern California. The labeled circles around the groups
correspond to the bifurcations in Fig. 10.
25
Table 12. The primary differences responsible for the bifurcations in Fig. 10.
Left side of
Right side of
Bifurcation
bifurcation
bifurcation
I
Kinorhynchs > 2%
Kinorhynchs < 2%
II
Copepods < 3%
Copepods > 3%
Ostracods < 0.3%
Ostracods > 0.3%
Desmoscolecidae > 10%
Desmoscolecidae < 10%
III
Non-Desmoscolecidae
< 80%
Non-Desmoscolecidae
> 80%
3.4. Question 2b. Do these Groupings Correlate with Differences in the Environment?
I asked this question in two ways. In the first, I used BIOENV (rho = 0.46) to identify
the environmental variable(s) (Tables 6 and 7) that best explained the similarity among samples.
For groupings based on absolute abundance, chl a by itself did so.
In the second analysis, I used LINKTREE, which takes the variable(s) identified in the
first approach (in this case just chl a) and uses them to describe how best the samples can be split
into groups by successive binary division (Clarke and Gorley, 2006). SIMPROF showed that
two of the bifurcations were significant. They are labeled I (difference between groups = 91.2%,
p = 0.001) and II (difference between groups = 59.5%, p = 0.004) in Fig. 12. Table 13 lists the
reasons for the bifurcations.
26
Figure 12. The dendrogram of samples produced by the PRIMER v6 (Clarke and Gorley, 2006)
procedure LINKTREE based on chl a and absolute abundance values. Bifurcations labeled with
Roman numerals were found to be significant by the PRIMER v6 procedure SIMPROF. The
ends of the branches are labeled by sample number, region of collection (A = off Oregon, B = off
northern California, C = off central California, D = off southern California), and relative depth (S
= shallow, D = deep).
Table 13. Significant LINKTREE bifurcations from Fig. 12 and the environmental difference
associated with each. Chl a = chlorophyll a in ng g-1 sediment dry weight.
Left side of Right side of
Bifurcation
bifurcation
bifurcation
I
Chl a > 506
Chl a < 506
II
Chl a < 53
Chl a > 83
The relative abundances of meiofaunal major taxa are best explained by the amount of
chl a and percent nitrogen (BIOENV analysis rho = 0.47). When I tested (LINKTREE and the
corresponding SIMPROF analyses, Fig. 13) the ability of chl a and percent nitrogen to explain
the pattern of similarity among samples using the meiofaunal relative-abundance data, I found
four significant bifurcations labeled I (difference between groups = 85.3%, p = 0.001), II
(difference between groups = 83.7%, p = 0.002), III (difference between groups = 63.1%, p =
27
0.001), and IV (difference between groups = 42.6%, p = 0.004). Table 14 lists the reasons for
the bifurcations.
Figure 13. The dendrogram of samples produced by LINKTREE (PRIMER v6, Clarke and
Gorley, 2006) based on chlorophyll a, percent nitrogen, and relative meiofaunal abundance.
Bifurcations labeled with Roman numerals were found to be significant by the PRIMER v6
procedure SIMPROF. The ends of the branches are labeled by sample number, region of
collection (A = off Oregon, B = off northern California), C = off central California, D = off
southern California, and relative depth (S = shallow, D = Deep).
Table 14. Significant LINKTREE bifurcations from Fig. 13 and the environmental difference(s)
associated with each. Chl a = chlorophyll a concentration in ng g-1 dry weight of sediment. %N
= nitrogen as a percentages of sediment dry weight.
Bifurcation
Left side
Right side
I
chl a > 858
chl a < 683
chl a > 683
chl a < 506
II
%N > 0.44
%N < 0.30
III
%N > 0.15
%N < 0.11
IV
%N < 0.17
%N > 0.18
28
CHAPTER FOUR
DISCUSSION
4.1. The Fauna
One of ecology’s primary goals is to be able to predict variation of the abundance of
organisms in space. Deep-sea ecologists have begun to accumulate the necessary data to do so at
biogeographic scales (e.g., Netto et al., 2005; Vincx et al., 1994), but because of the immense
size of the habitat and the small number of studies that have been done, coverage remains sparse.
The data in Table 2 will begin to fill in the gap along the continental rise off the west coast of the
United States.
To put the abundances in context, I plotted them along with data from comparable studies
in the Pacific (Table 4). For nematodes, the scatter in reported abundances was substantial, but
my values were not unusually high or low. For copepods, ostracods, and kinorhynchs, the data
are more orderly in that abundance generally decreases with depth, and my observations fit the
trends. Thus, at least to this extent, my values appear to be reasonable.
4.2. Question 1. Do the Abundances of Meiofaunal Major Taxa Differ with Region or
Depth, and do the Differences Correlate with Differences in the Environment?
4.2.1. Absolute abundances. When I tested each taxon for differences in log10(X+1)transformed, absolute abundance with region and depth, I found four significant results. (1)
Copepod abundance decreased with depth in the three regions analyzed (B, C, and D). This
pattern was not surprising because, in general, the abundance of deep-sea life decreases with
depth and distance from the continents, with depth dominating (see, e.g., Rowe, 1983). This
regularity is thought to arise because the supply of food to the deep sea tends to decrease as these
two variables increase (see Levin et al., 2001, for a review). In my data, chl a in the sediment
decreased with depth in each region (Fig. 14), so the decline in food supply could underlie the
decrease in copepod abundance with depth.
29
Figure 14. Chlorophyll a concentration (in ng g-1 sediment dry weight) versus depth, showing
that as depth increases in a region, chlorophyll a concentration in the sediment declines. Region
B (off northern California) = circles. Region C (off central California) = triangles. Region D
(off southern California) = squares.
(2) Kinorhynch abundance (Fig. 4) was significantly greater in Region C than in Regions
B and D. Chl a in the sediment showed the same pattern (Fig. 15), which suggests that
differences in regional food supply underlie the regional difference in kinorhynch abundance.
My results for kinorhynchs extend those of workers who reported that variation in the abundance
of deep-sea meiofauna correlated positively with measures of food in the sediment (see e.g.,
Baguley et al., 2006; Grove et al., 2006; Pfannkuche, 1985; Soetaert et al., 1991; Soltwedel,
1997; but see Soltwedel, 2000).
30
Figure 15. Kinorhynch abundance versus chl a concentration (in ng g-1 dry weight of sediment),
showing that both variables are highest in Region C. Region B (off northern California) =
circles. Region C (off central California) = triangles. Region D (off southern California) =
squares.
(3) In contrast, the abundance of ostracods decreased as the mud content of the sediment
increased (Fig. 16). It is believed that food availability it possibly the most important limiting
factor in deep-sea ecology (Gage and Tyler, 1991), this result raises the question of why
sediment grain-size distribution dominated for ostracods. Given the small amount of natural
history information on deep-sea ostracods, many explanations are possible. For example, for
shallow-water ostracods, Danielopol et al. (1988) suggested the quality and texture of the
substrate can affect ostracod burrowing. The movement of deep-sea ostracods is also likely to
depend on sediment properties, so it might be worthwhile in future research to determine (1)
whether deep-sea ostracods move more easily in sediment that is ~70% rather than ~95% mud
and (2) whether ease of movement conveys an advantage to ostracods, such as more rapid access
to food patches.
31
Figure 16. Ostracod abundance versus the percent of the sediment by weight that was less than
0.062 mm (= percent mud), showing that the abundance of ostracods is higher in less muddy
sediment. Region B (off northern California) = circles. Region C = triangles (off central
California). Region D (off southern California) = squares.
(4) For the nematode family Desmoscolecidae, region and depth interacted significantly.
Inspection of Fig. 5 shows that rather than declining in abundance with depth as in Regions B
and D and as expected (see, e.g., de Bovée et al., 1990; Shirayama, 1984), the abundance of
Desmoscolecidae increased in Region C, thus the interaction. Food in the sediment decreased
with depth in Region C as it does for Regions B and D, and the sediment was muddy to
essentially the same degree (Table 6); so neither of these variables seems likely to have caused
the significant interaction. Although the literature on deep-sea Desmoscolecidae has begun to
accumulate (e.g., Radziejewska et al., 2002; Vanaverbeke et al., 2004; Vanhove et al., 2004), I
found no information that seems to explain the interaction. (For an earlier example of an
interaction involving meiofauna, see Baguley et al., 2006).
4.2.2. Relative abundance. When I tested each taxon for differences in transformed relative
abundance with region and depth, I found two significant results. (1) The relative abundance of
kinorhynchs was significantly greater in Region C (Figs. 7 and 17) than in Regions B and D.
This correspondence suggests that it might be worthwhile investigating whether kinorhynchs, as
a taxon, are relatively more successful than the other meiofaunal major taxa where food
32
concentration in the sediment is high. (2) The relative abundance of non-Desmoscolecidae
nematodes was significantly less in Region C than in Regions B and D (Figs. 6 and 17). The
decrease is unlikely to result from their displacement by kinorhynchs, because, in each sample,
non-Desmoscolecidae nematodes were an order of magnitude more abundant than kinorhynchs.
Radziejewska et al.’s (2001) results suggest a second possibility. They found that more
Desmoscolecidae individuals occurred when copious amounts of phytodetritus were present.
This is seen in my data with a positive correlation to chl a sediment concentration and the
relative abundance of Desmoscolecidae nematodes compared to a negative correlation for the
relative abundance of non-Desmoscolecidae nematodes. It might be worthwhile to test whether
Desmoscolecidae nematodes displace enough non-Desmoscolecidae nematodes from food-rich
sediment to account for the effect.
Percent of sample abundance
90
80
70
60
50
40
30
20
10
0
Figure 17. Untransformed, relative abundance by region, showing the proportional increase of
kinorhynchs and the proportional decrease of non-Desmoscolecidae nematodes in Region. For
each taxon, the proportion in each of the 16 samples is indicated by a bar. For each taxon, the
bars are in numerical order by sample number. Black bars = samples from Region A off Oregon.
Dark-grey bars = samples from Region B off northern California. Light-grey bars = samples
from Region C off central California. Gridded bars = Region D off southern California.
33
4.3. Question 2
4.3.1. Question 2a. Can Samples be Grouped Based on their Similarity in Absolute
Abundances of Meiofaunal Major Taxa? In this approach, I used the Bray-Curtis similarity
coefficient, so both the absolute abundances and the identities of the major taxa influenced the
pair-wise similarity among samples (Clarke and Warwick, 2001). Although the pre-analysis
transformation reduced the overwhelming dominance of the nematodes on similarity, the
similarity between any pair of samples was still greater than 90%, even for those separated by
more than 1500 km. That is, at the major-taxon level, the differences I found in similarity among
samples along the west coast of the United States were less than 10% in the transformed space
and would have been even less if the data had not been transformed before analysis.
Despite this overall similarity among samples, the cluster analysis followed by SIMPROF
revealed groups of samples that differed significantly from each other. The groupings did not
arise because the samples came from similar depths or were close to each other in space. Rather,
the initial bifurcation (Fig. 8 and Table 11) separated them into low- and high-abundance groups.
Ecologists often assume that faunal similarity decreases with increasing separation
distance. This rule of thumb provides no guidance here. That is, the samples from the two
Region C stations formed a group consisting of all and only samples from the region. The
samples from the other regions did not (Figs. 8 and 9). The separation between the two stations
in Region A was 202.3 km, in Region B was 37.4 km, in Region C was 78.6 km, and in Region
D was 24.4 km (Table 1). Thus, the region in which the samples grouped (C) was not the region
where the stations were closest.
The similarity results for depth are of the same type. As above, only for Region C did the
samples from the two stations form a group consisting of all and only samples from the region
(Figs. 8 and 9). The depth difference between the two stations in Region A was 357 m, in
Region B was 961 m, in Region C was 954 m, and in Region D was 1145 m (Table 1). The
depth difference between the stations in Region C was little different from that of Region B and
much greater than that of Region A, where the samples did not group. At least for Region C,
similarity in the abundance of meiofaunal major taxa cannot be explained by either proximity in
space or in depth.
34
Another counter-intuitive relationship involved the two replicate samples from the
shallow station in Region D (i.e., station 8). These samples were taken 0.19 km apart, but the
nMDS plot (Fig. 9) shows that they are more dissimilar than some samples that were separated
by more than 1000 km. I consider the implications below.
My analyses of similarity among samples based on relative abundances added little to the
results from my analysis of absolute abundance, so I do not discuss them further.
4.3.2. Question 2b. Do these Groupings Correlate with Differences in the Environment?
The BIOENV analysis showed that chl a by itself was the environmental variable that best
explained the sample groupings formed above. That is, the addition of other environmental
variables did not add any explanatory power.
I also looked for relationships with environmental variables by creating sample groupings
with LINKTREE followed by SIMPROF based on similarity among samples in the
environmental variable(s) identified by BIOENV, which in my case was just chl a. The initial
bifurcation separated samples with high chl a from those with low chl a. The groupings based
on chl a do not correspond one-to-one with those based on similarity in absolute abundance
(compare Figs. 8 and 12), so more must be involved than just the response of meiofaunal major
taxa to food. Still, some groupings were found by both analyses. In particular, the samples from
Region A were members of the low-abundance (first analysis bifurcation I left side, see Table
11) and low-chl-a (second analysis, bifurcation II left side, see Table 13) groups. In contrast, the
Region C samples were members of the high-abundance (first analysis, bifurcation III right side
see Table 11) and high-chl-a (second analysis, bifurcation I left side see Table 13) groups. These
results suggest that differences in the amount of food in the sediment caused the sample
groupings in both cases, despite the effects of separation in space and differences in depth among
the samples in the two regions.
The two samples from Station 8 were unusually dissimilar (see above). The cause is
unclear, but one sample had 683 ng g-1 of chl a and the other had only 227 ng g-1. To put these
values in context, these samples differed by 456 ng g-1, but the next largest difference between
samples at a single station was only 160 ng g-1. This difference in the amount of food is likely to
explain at least part of the difference between samples in the meiofauna and thus would be a
good focus for further research. One approach would be determine whether relatively high food
concentrations attracted them, such as has been observed for nematodes in shallow water
35
(Schratzberger et al., 2004). If so, the abundances of meiofaunal major taxa in the two types of
patches would diverge. A large enough divergence would cause the Bray-Curtis similarity
between the samples to decrease, as I observed. Further, the marked difference in food
concentration in the sediment suggests that it could be worthwhile to see whether topographic
differences could cause differences in the delivery of food for meiofauna at this scale.
36
REFERENCES
Anderson, T.R., Rice, T., 2006. Deserts on the sea floor: Edward Forbes and his azoic
hypothesis for a lifeless deep ocean. Endeavour 30, 131-137.
Baguley, J.G., Montagna, P.A., Hyde, L.J., Kalke, R.D., Rowe, G.T., 2006. Metazoan
meiofauna abundance in relation to environmental variables in the northern Gulf of
Mexico deep sea. Deep-Sea Research I 53, 1344-1362.
Barnett, P.R.O., Watson, J., Connelly, D., 1984. A multiple corer for taking virtually
undisturbed samples from shelf, bathyal and abyssal sediments. Oceanologica Acta 7,
399-408.
Bett, B.J., Vanreusel A., Vincx, M., Soltwedel, T., Pfannkuche, O., Lambshead, P.J.D., Gooday,
A.J., Ferrero, T., Dinet, A., 1994. Sampler bias in the quantitative study of
deep-sea meiobenthos. Marine ecology progress series 104, 197-203.
Bray, J.R., Curtis, K.R., 1957. An ordination of the upland forest communities of Southern
Wisconsin. Ecolology Monographs 27, 325-349.
Burgess, R., 2001. An improved protocol for separating meiofauna from sediments using
colloidal silica sols. Marine Ecology Progress Series 214, 161-165.
Carmen, K.R., Thistle, D., Fleeger, J.W., Barry, J.P., 2004. Influence of introduced CO2 on
deep-sea metazoan meiofauna. Journal of Oceanography 60, 767-772.
Clarke, K.R., Gorley, R.N., 2006. Primer v6: User manual/Tutorial. PRIMER-E, Plymouth.
Clarke, K.R., Warwick, R.M., 2001. Change in marine communities: An approach to statistical
analysis and interpretation, second ed. Primer-E, Plymouth.
Danielopol, D.L., Geiger, W., Tölderer-Farmer, M., Orellana, C.P., Terrat, M.N., 1988. In search
of Cypris and Cythere a report of the evolutionary ecological project on limnic Ostracoda
from the Mondsee (Austria), in: Hanai, T., Ikeya, N., Ishizaki, K. (Eds.), Evolutionary
Biology of Ostracoda, Its Fundamental Applications. Proceedings of the 9th International
Symposium on Ostracoda. Kodosha/Elsevier, Tokyo/Amsterdam, pp. 485–500.
Danovaro, R., DellaCroce, N., Eleftheriou, A., Fabiano, M., Papadoulou, N., Smith, C.,
Tselepides, A., 1995. Meiofauna of the deep Eastern Mediterranean Sea: distribution and
abundance in relation to bacterial biomass, organic matter composition and other
environmental factors. Prog Oceanogr 36, 329-341.
De Bovée, F., Guidi, L.D., Soyer, J., 1990. Quantitative distribution of deep-sea meiobenthos in
the northwestern Mediterranean (Gulf of Lions). Continental Shelf Research 10, 11231146.
37
De Morais, L.T., Bodiou, J.Y., 1984. Predation on meiofauna by juvenile fish in a Western
Mediterranean flatfish nursery ground. Marine Biology 82, 209-215.
Dinet, A., 1979. A quantitative survey of meiobenthos in the deep Norwegian Sea. Ambio
Special Report 6, 75-77.
Folk, R.L., 1968. Petrology of Sedimentary Rocks. Hemphills’s, Austin.
Gage, J.D., Tyler, P.A., 1991. Deep-sea Biology: A Natural History of Organisms at the Deepsea Floor. Cambridge University Press, Cambridge, UK.
Giere, O., 2008. Meiobenthology the microscopic motile fauna of aquatic sediments., second ed.
Springer-Verlag, Berlin.
Grassle, J.F., Maciolek, N.J., 1992. Deep-sea species richness: regional and local diversity
estimates from quantitative bottom samples. American Naturalist 139, 313-341.
Grove, S.L., Probert, P.K., Berkenbusch, K., Nodder, S.D., 2006. Distribution of bathyal
meiofauna in the region of the Subtropical Front, Chatham Rise, south-west Pacific.
Journal of Experimental Marine Biology and Ecology 330, 342-355.
Hessler, R.R., Sanders H.L., 1967. Faunal diversity in the deep-sea. Deep-Sea Research 14, 6578.
Ingles, J., Kiriakoulakis, K., Wolff, G.A., Vanreusel, A., 2009. Nematode diversity and its
relation to the quantity and quality of sedimentary organic matter in the deep Nazaré
Canyon, Western Iberian Margin. Deep-Sea Research I 56, 1521-1539.
Itoh, M., Kawamura, K., Kitahashi, T., Kojima, S., Katagiri, H., Shimanaga, M., 2011.
Bathymetric patterns of meiofaunal abundance and biomass associated with the Kuril and
Ryukyu trenches, western North Pacific Ocean. Deep-Sea Research I 58, 86-97.
Levin, L.A., Etter, R.J., Rex, M.A., Gooday, A.J., Smith, C.R., Pineda, J., Stuart, C.T., Hessler,
R.R., Pawson, D., 2001. Environmental influences on regional deep-sea species
diversity. Annual Review of Ecology and Systematics 32, 51-93.
Mayer, L.M., Schick, L.L., Sawyer, T., Plante, C.J., 1995. Bioavailable amino acids in
sediments: A biomimetic, kinetics-based approach. Limnology and Oceanography 40,
511-520.
McCall, J.N., Fleeger, J.W., 1995. Predation by juvenile fish on hyperbenthic meiofauna: a
review with data on post-larval Leiostomus xanthurus. Vie et Milieu 45, 61-73.
Murray, J., 1895. A summary of the scientific results, in: Report on the scientific results of the
voyage of HMS Challenger. Eyre and Spottiswoode, London.
38
Netto, S.A., Gallucci, F., Fonseca, G.F.C., 2005. Meiofauna communities of continental slope
and deep-sea sites off SE Brazil. Deep-Sea Research I 52, 845-859.
Pfannkuche, O., 1985. The deep-sea meiofauna of the Porcupine Seabight and abyssal plain (NE
Atlantic): population structure, distribution, standing stocks. Oceanologica Acta 8, 343–
353.
Radziejewska, T., 2002. Responses of deep-sea meiobenthic communities to sediment
disturbance simulating effects of polymetallic nodule mining. Internat. Rev. Hydrobiol.
87, 457-477.
Ramirez-Llodra, Brandt, A., Danovaro, R., De Mol, B., Escobar, E., German, C.R., Levin, L.A.,
Martinez Arbizu, P., Menot, L., Buhl-Mortensen, P., Narayanaswamy, B.E., Smith, C.R.,
Tittensor, D.P., Tyler, P.A., Vanreusel A., Vecchinoe, M., 2010. Deep, diverse and
definitely different: unique attributes of the world’s largest ecosystem. Biogeosciences 7,
2851-2899.
Rowe, G.T., 1983. Biomass and production of the deep-sea macrobenthos, in: Rowe, G.T. (Ed),
Deep-sea biology. Wiley, New York, pp 97-122.
Schratzberger, M., Whomersley, P., Warr, K., Bolam, S.G., Rees, H.L., 2004. Colonisation of
various types of sediment by estuarine nematodes via lateral infaunal migration: a
laboratory study. Marine Biology 145, 69-78.
Shirayama, Y., 1983. Size structure of deep-sea meio- and macrobenthos in the western pacific.
Internationale Revue Der Gesamten Hydrobiologie 68, 799-810.
Shirayama, Y., 1984. The abundance of deep-sea meiobenthos in the Western Pacific in relation
to environmental factors. Oceanologica Acta 7, 113-121.
Shirayama, Y., Fukushima, T., 1995. Comparisons of deep-sea sediments and overlaying water
collected using multiple corer and box corer. Journal of Oceanography 51, 75-82.
Snelgrove, P.V.R., Blackburn, T.H., Hutchings, P.A., Alongi, D.M., Grassle, J.F., Hummel, H.,
King, G., Koike, I., Lambshead, P.J.D., Ramsing, N.B., Solis-Weiss, V., 1997. The
importance of marine sediment biodiversity in ecosystem processes. AMBIO 26, 578583.
Snelgrove, P.V.R, 1999. Getting to the bottom of marine biodiversity: sedimentary habitats –
ocean bottoms are the most widespread habitat on earth and support high biodiversity and
key ecosystem services. BioScience 49, 129-138.
Snelgrove, P.V.R, Smith, C.R., 2002. A riot of species in an environmental calm: The paradox
of the species-rich deep-sea floor. Oceanography and Marine Biology: an Annual
Review 40, 311-342.
39
Soetaert, K., Heip, C., Vincx, M., 1991. The meiobenthos along a Mediterranean deep-sea
transect off Calvi (Corsica) and in an adjacent canyon. P.S.Z.N.I. Marine Ecology, 12,
227–242.
Sokal, R.R., Rohlf, F.J., 1997. Biometry., third ed. W.H. Freeman and Company, New York.
Soltwedel, T., 1997. Meiobenthos distribution pattern in the tropical East Atlantic: indication for
fractionated sedimentation of organic matter to the sea floor? Marine Biology 129, 747756.
Soltwedel, T., 2000. Metazoan meiobenthos along continental margins: a review. Progress in
Oceanography 46, 59-84.
Thiel, H., 1979. Structural aspects of the deep-sea benthos. Ambio Special Report 6, 25-31.
Thistle, D., 2003. Harpacticoid copepod emergence at a shelf site in summer and winter:
implications for hydrodynamic and mating hypotheses. Marine Ecology Progress Series
248, 177-185.
Vanaverbeke, J., Soetaert, K., Vincx, M., 2004. Changes in morphometric characteristics of
nematode communities during a spring phytoplankton bloom deposistion. Marine
Ecology Progress Series 273, 139-146.
Vanhove, S., Vermeeren, H., Vanreusel, A., 2004. Meiofauna towards the South Sandwich
Trench (750-6300 m), focus on nematodes. Deep-Sea Research II 51, 1665-1687.
Vincx, M., Bett, B.J., Dinet, A., Ferrero, T., Gooday, A.J., Lambshead, P.J.D., Pfannkuche, O.,
Soltwedel, T., Vanreusel, A., 1994. Meiobenthos of the deep northeast Atlantic.
Advances in Marine Biology 30, 1-88.
Wigley, R.L., McIntyre, A.D., 1964. Some quantitative comparisons of offshore meiobenthos
and macrobenthos south of Martha’s Vineyard. Limnology and Oceanography 9, 485493.
Young, C.M., Tyler, P.A., 1993. Embryos of the deep-sea echinoid Echinus affinis require high
pressure for development. Limnology and Oceanography 38, 178-181.
http://stableisotopefacility.ucdavis.edu. Last accessed in 2009.
40
BIOGRAPHICAL SKETCH
Education
M.S., Biological Oceanography, Florida State University, Tallahassee, Florida, 2012
B.S., Marine Science and Biology, Coastal Carolina University Conway, South Carolina, 2005
Thesis Topics
2005
Undergraduate honors senior thesis. Supervisor Dr. Paul Richardson.
“Examination of the immunoglobulin superfamily proteins in Strongylocentrotus
purpuratus.”
Research Cruises
2011
HURL Research Cruise
Purpose: Deep-seal coral collection.
Supervisor: Dr. Amy Baco-Taylor.
Days at sea: 3 weeks.
2011
BP Research Cruise
Purpose: BP oil spill assessment
Supervisor: Dr. Ian MacDonald.
Days at sea: 2 weeks.
41
2008
NSF Cruise
Purpose: Collected deep-sea meiofauna samples
Supervisor: Dr. David Thistle
Days at sea: 3 weeks.
2005
NOAA Cruise
Purpose: Identify ground water sources and old river channels.
Supervisor: Dr. Paul Gayes.
Days at sea: 2
Teaching experience
2010-Present Mentoring and training undergraduates on meiofauna research in the lab.
2009-2010
Teaching assistant for basic oceanography online course.
Presentations
2009-2012
Benthic Ecology Meeting Poster Presentations
Employment
2008-Present Florida State University graduate research assistant and teaching assistant.
2006-2008
RPS Energy Consultant- Marine Mammal Observer
Days at sea: 5-10 week rotations.
42
2006
Columbus Zoo and Aquarium, Shores Regional
2005
Lab Aid, Coastal Carolina University Marine Science Department
2005
Ripley’s Aquarium Internship, Myrtle Beach, SC
Awards and Certifications
2006
Marine Mammal Observer Certification
2005
Invocation Speaker at Commencement
2005
Nominated as Marine Science student of the year
2004
Advanced Scuba Certified
2002
ODK Undergraduate Leadership award
43