Sucessão ecológica em parcelas orgânicas de madeira, macroalgas e em carcassas de baleia no mar... por Angelo Fraga Bernardino - Versão HTML

ATENÇÃO: Esta é apenas uma visualização em HTML e alguns elementos como links e números de página podem estar incorretos.
Faça o download do livro em PDF, ePub para obter uma versão completa.

individual parcels spanning a broad range of sizes (1 - 1000 kg) to the seafloor (Wolff

1979, Smith 1983, C. Smith personal observations during >100 submersible and

remotely operated vehicle dives off California and Hawaii).

It is long recognized that food falls at the deep-sea floor can contribute to beta

diversity in deep-sea by creating patches of organic enrichment and chemical or

physical disturbance (Stockton & DeLaca 1982, Smith & Hamilton 1983, Smith 1985,

1986, Grassle & Morse-Porteous 1987, Snelgrove & Smith 2002). However, patterns

of succession resulting from the arrival of kelp and wood falls, and the identity of

their characteristic species remain essentially unstudied along the bathyal Californian

slope. Such information is essential to understanding the recycling of different types


of organic parcels (e.g., sunken kelp paddies, logs) on the continental slope and their

roles in maintaining biodiversity in the deep northeast Pacific (Snelgrove et al. 1992,

Snelgrove et al. 1996, Snelgrove & Smith 2002).

The effects of small-scale, low intensity disturbances on deep-sea infaunal

communities, resulting from algal enrichment and scavenger disturbance, have been

evaluated in a number of deep-experiments; these disturbances produce modest

increases on species diversity (Levin & Smith 1984, Snelgrove & Smith 2002).

However, substantial organic enrichment of deep-sea sediments (e.g., from

macroalgae, wood falls, diatom detritus, and whale falls) can cause dramatic

population enhancement of opportunistic and sulfophilic species (e.g. capitellid and

dorvilleid polychaetes, leptostracans and cumaceans; (Turner 1977, Smith 1986,

Grassle & Morse-Porteous 1987, Snelgrove et al. 1994, Smith et al. 2002, Smith &

Baco 2003). Community patterns in organic-rich sediments around whale falls

resemble the classic spatial succession described along organic-enrichment gradients

in shallow-water [e.g., (Pearson & Rosenberg 1978)] with very low diversity and

extremely high densities of opportunistic polychaetes species adjacent to the whale

fall, moderate macrofaunal enhancement and species diversity at intermediate

distances, and low abundances and high diversity in the background community

(Smith et al. 2002, Smith & Baco 2003). During later stages of succession, deep-sea

whale falls also support substantial assemblages of animals with chemoautotrophic

endosymbionts, producing reducing habitats ecologically similar to vents and seeps

(Smith & Baco 2003, Treude et al. 2009).

Large wood and kelp parcels also have the potential to produce intense

organic enrichment and reducing habitats in deep-sea sediments. Deep-sea wood falls

are initially processed by wood-boring xylophagainae bivalves (Turner 1977, Wolff


1979, Distel & Roberts 1997, Turner 2002, Voight 2007b) which may broadcast

organic-rich fecal material onto surrounding sediments (Turner 1977), potentially

creating organic and sulfide-rich sediments. The breakdown of large kelp

accumulations at the deep-sea floor also has the potential to create organic and

sulfide-rich sediment patches against an oligotrophic background (Vetter 1994, 1996,

Vetter & Dayton 1998).

While deep-sea wood and kelp falls are common on regional scales, temporal

patterns of infaunal community succession around large wood and kelp parcels

remain very poorly studied. Thus, the contribution of wood- and kelp-fall

communities to beta diversity in the deep sea, and their relationship to deep-sea

reducing assemblages (e.g., vent, seep and whale-fall communities) cannot be

evaluated. In this study, we investigated macroinfaunal structure for over 5 years

around experimentally implanted kelp and wood parcels in the bathyal NE Pacific to

test the following hypotheses:

1) Large wood and kelp falls exhibit distinct patterns of infaunal macrobenthic

community succession, promoting beta diversity in the deep sea;

2) There is little species overlap between sediment communities around large

wood and kelp parcels and other reducing habitats (whale falls, seeps and

vents) at similar depths in the deep sea.

2.2. Materials and Methods

Study site, deployments and sampling

Experimental implantations of kelp and wood parcels were conducted near

well-studied whale falls off southern California (Smith & Baco 2003) (Fig. 1.1). In



particular, wood and kelp parcels were implanted on the flat sediment-covered floor

of Santa Cruz Basin (SCrB), California (~33° 27’ N, 119° 22’W, 1670 m depth,

bottom temperature ~4°C, bottom-water oxygen concentration 260 µM (Treude et al.


Four parcels of wood and kelp were experimentally deployed at a mean water

depth of 1670 m in the SCrB (Fig. 1.1, point A; Table 2.1), approximately 100 m

from an experimental 30-ton whale carcass (Smith & Baco 2003, Treude et al. 2009).

Wood parcels were deployed in October 1999 and May 2002 (Table 2.1). Kelp

parcels were deployed in May 2002 and July 2002 (Table 2.1). Each kelp parcel

consisted of ~100 kg (4-5 whole plants) of fresh Macrocystis pyrifera contained in a

5-cm stretch mesh nylon net bag with one kelp holdfast protruding from the bag to

facilitate sampling of kelp associated fauna; Fig. 2.1). The wood parcels consisted of

~200 kg of untreated Douglas fir ( Pseudotsuga menziesii) with four small, 2-kg wood

packages inside a 2.5-cm stretch mesh nylon net bags attached to the upper surface

(Fig. 2.1).

Figure 2.1 A. Kelp parcel deployed at Santa Cruz Basin after 0.25 y on the sea floor. Notice

the presence of a whitish bacterial mat over the sediment neaby the parcel. Dark sediments

around it represent sulfide rich sediments. B. 200Kg wood parcel that has been deployed at

SCrB for two years. Small wood packs inside 2.5cm mesh net nylon bags are attached on the

top of the parcel. C. Detail showing the presence of bacterial mats nearby the kelp falls at 0.25 y.


Kelp and wood parcels were sampled using the Remotely Operated Vehicle

(ROV) Tiburon during three cruises on board of the RV Western Flyer (Table 2.1).

Kelp parcels were sampled at 0.25 and 0.5 years after deployment, while wood

parcels were sampled at 0.5, 1.8, 3 and 5.5 y after deployment (Table 2.1). Sampling

occurred around distinct parcels for each time interval; because parcels were located

within several hundreds meters of one another on the homogenous basin floor, we

assumed that all fall types were bathed in the same larval pool, i.e., location effects

were modest compared to treatment effects. At all sites, three randomly located

transects radiating outward from the treatment were sampled, collecting tube cores (7

cm diameter, 10 cm deep) at 0 m, 0.5 m, 1.0 m and 2.0 m from the parcel.

Background samples (17 replicate tube cores) were additionally collected in 2002,

2004 and 2005 at random locations of 9-100 meters away from any parcel, allowing

averaging of temporal patterns in the background Santa Cruz Basin. Tube cores were

immediately sectioned into 0-1, 1-5 and 5-10 cm horizons and fixed in 4% buffered

formaldehyde-seawater solution until laboratory analysis, or frozen at -20oC for

organic carbon analysis in laboratory.

Table 2.1. Drop and sampling date from organic parcels artificially deployed at SCrB. Note

that each wood and kelp parcels were sampled on distinct cruises, thus representing

community succession on several time scales. Background cores were collected at three

different cruises (see text).

Parcel type /



Bottom position

Sampling date/ ROV dive

Age at

deployment number





Oct. 25, 2002/TD 494

3 years

Wood (CRS 397)

Oct. 15,




Mar. 1, 2005/TD 827

5.5 years


Oct. 24, 2002/TD 492

0.5 years

Wood (CRS 800)

May 1,




Mar. 1, 2004/TD 652

1.8 years

Kelp (CRS 799)

May 1,



Oct. 25, 2002/TD 493

0.5 years




Kelp (CRS 806)

July 16,



Oct. 22, 2002/TD 489

0.25 years



2002/ TD 490, 502, 505

SCr Background


2004/ TD 655

2005/ TD822, 823, 824

Laboratory and data analysis

In the laboratory, macrobenthic samples were sieved on a 300-µm sieve,

stained with Rose Bengal, sorted and identified to the lowest attainable taxonomic

level. Metazoan trophic group classification was made according to (Kukert & Smith

1992). Sediment samples for organic carbon and nitrogen analysis were acidified to

remove carbonates by repeated additions of sulfurous acid (8 % v/v) until no

effervescence was observed (Verardo et al. 1990) and then analyzed using a Perkin-

Elmer 2400 CHN Elemental Analyzer, with a limit of detection of 1 mg and 1.2 mg

for C and N, respectively, while the precision was 0.3 % and 0.4 % for C and N,

respectively. CHN standards were made with acetanilide and blanks were made of

both non-acidified and acidified cups, both giving minor signals of TOC and TN.

A few top-ranked macrofaunal organisms at representative stages of

community succession on kelp and wood parcels were selected for stable isotope

analysis. Individuals were sorted using methanol-cleaned forceps and specimens were

rinsed and cleaned in DI water. Calcareous shelled organisms were decalcified with

phosphoric acid and placed in pre-weighed tin cups for overnight drying (35-40 ºC).

Multiple individuals of single species were combined in one sample to make up the

necessary dry weight of 0.5-2 mg for analysis. Samples were combusted in a

Eurovector elemental analyzer and resulting N2 and CO2 gases were separated by gas

chromatography and admitted into an IRMS mass spectrometer for determination of


15N/14N and 13C/12C ratios (reproducibility: ±0.5 ‰ for δ15N and ±0.2 ‰ for δ 13C).

Macrofaunal C-isotopic ratios were measured against a Pee Dee Belemnite (PDB)

standard for δ13C and atmospheric nitrogen for δ15N. Results are expressed as delta (δ)

notation representing the relative difference between sample and standard, where δX

(‰) = [(Rsample/Rstandard)-1] x 103, where R=15N/14N or R=14C/13C.

Macroinfaunal organisms were fixed in 4% formaldehyde solution, potentially

introducing artifacts in stable isotope values, although shifts in carbon isotope ratio

are usually small compared to the wide natural C variability in marine food sources

(Fry & Sherr 1989, Edwards et al. 2002, Sarakinos et al. 2002). In this study, we

corrected for preservation artifacts by adding 1‰ to δ 13C (Baco-Taylor 2002,

Sarakinos et al. 2002, Demopoulos et al. 2007). Trophic shift boxes from potential

organic matter sources at kelp and wood parcels helped to determine potential food

sources (Fry 2006). Trophic changes of 3 ‰ were added to the range of δ15N values

obtained for each organic matter source (DeNiro & Epstein 1978, Minagawa & Wada

1984, Fisher et al. 1994). Species that exclusively use a particular source for nutrition

are expected to fall within the appropriate trophic-shift box. Isotope values from kelp

plants were obtained from the published literature (Page et al. 2008). A multi-source

mixing model was used to calculate proportional contributions of each primary

organic matter source (i.e.: kelp, wood, sedimentary organic carbon and bacterial mats

growing on sulphidic sediments around treatments) to the benthic food web

[IsoSource software; (Phillips & Gregg 2003)]. This model examines all possible

combinations of each source contribution (0-100%) in small increments (e.g., 1%),

where the combinations that best fit to the observed species isotope signature are

considered feasible (Phillips & Gregg 2003). A mixing polygon was drawn around

fauna samples, with polygon apices representing end members for each organic


matter source (Demopoulos et al. 2007). Mixing contributions were calculated for

consumers that fit within the mixing polygon, using a priori faunal corrected (i.e.

trophic fractionation) isotope measurements for each trophic group (Kukert & Smith

1992, Phillips & Gregg 2003, Demopoulos et al. 2007). We assumed a δ15N

fractionation of 3‰ for omnivores, deposit feeders and “other” trophic levels (Fry

1988, 2006, Demopoulos et al. 2007). Statistical tests to detect differences on the

contribution of organic sources to the fauna were performed using the IsoError

software (Phillips & Gregg 2001).

Statistical analyzes on sediment organic content were examined with Student t

test of means or Kruskal-Wallis, when the number of replicates allowed. Patterns of

faunal abundance, faunal similarity and diversity were compared among distances and

time for each treatment type. Total densities around treatments and in ambient

sediments were examined with One-Way ANOVA if normality of variances was

present. If homogeneity of variances was not attained, a non-parametric Kruskal-

Wallis test was performed. For significant ANOVA and Kruskal-Wallis results, post-

hoc tests were used to examine difference in means [statistical package BioEstat

(Zar 1996)]. Species diversity was evaluated for pooled replicate cores at each

distance sampled (n=1-4) due to the low density of metazoans. Hulbert’s rarefaction

curves (ESn) were used to compare species diversity between treatments. ESn gives an

estimative of the number of species that would be found in a given number of

individuals, interpolated from the number of species collected in each sample

assuming a random distribution of individuals within samples (Hulbert 1971).

Although this index tends to overestimate within-sample diversity, it is widely use to

compare deep-sea samples of different sizes (Levin et al. 2001). Values of ESn were

compared at n=10, n=25 and n=50, where possible. Near treatment samples were


compared at higher n values, because these samples contained higher abundances and

were particularly interesting as they were directly influenced by the treatments.

Background replicate cores (n=17) from 2002, 2004 and 2005 were combined to

calculate a composite diversity from the background community; with confidence

limits calculated from pooled cores. From rarefaction “knots” obtained in each pooled

sample, a one-tailed 95% confidence interval was calculated using the T distribution.

Diversity curves were then compared to the background confidence curve in order to

test for statistical differences on diversity (Smith 1986). Additionally, Pielou’s

evenness values ( J’) is given to provide further information on community structure

(Clarke & Warwick 2001). The relative abundance of trophic guilds at each period

sampled was integrated by distance in order to increase the power of statistical

analyzes. Samples from 0 m and 0.5 m were compared against the background in

order to evaluate changes in trophic structure nearby our treatments; and examined

with ANOVA or Kruskal-Wallis.

Cluster analyses and non-metric multi-dimensional scaling (MDS) based on

species-abundance data from standardized quantitative samples (PRIMER v6, (Clarke

& Gorley 2006); were used to compare community structure across distance and time.

Square-root transformations were used prior to multivariate analysis in order to

balance the importance of common and rare species (Clarke & Warwick 2001).

Clustering and ordination analyses were often combined to verify mutual consistency.

Analysis of similarities (ANOSIM) was performed on standardized quantitative

samples to determine significant differences between groups (distance and time) and

dissimilarity percentages identified species contributions to these patterns (SIMPER

analysis, (Clarke & Warwick 2001).


In order to determine the level of species overlap between kelp, wood and

other reducing habitats, we restricted our direct comparisons to relatively abundant

species (i.e., over 1% rel. ab. at any treatment type) collected nearby our treatments

(i.e. 0 to 0.5 meters) and that were not sampled in ambient sediments (i.e. total of 17

background cores from 2002, 2004 and 2004). We excluded rare species (less than

1% rel. abundance) from our comparisons in order to account for difficulties in

taxonomic identification and undersampling biases. Comparisons of species overlap

with other published studies were carried out at the genus and species level in order to

account for differences in taxonomic identification between studies.

2.3. Results

Sedimentary organic carbon and nitrogen at kelp and wood parcels

After 0.25 years, percent organic carbon (TOC) in surface sediments near kelp

falls was only 0.5% above background levels, with a spatial decrease in sedimentary

TOC from 0 m to 1m (Fig. 2.2). After 0.5 years, the TOC levels at 0 meters was

significantly higher than 0.5 meters away (t=-11.045, d.f.=2, p=0.008), although no

significant temporal increase in TOC was observed at 0 meters (Fig. 2.2). The

combined TOC content in nearest 0-0.5 meters from the kelp falls at 0.5 y was

significantly higher than 0.25 y and background sediments (Kruskal-Wallis, H=7.833,

d.f.=2, p=0.02). Sedimentary percent organic nitrogen (TON) was also slightly

elevated near kelp parcels at 0.25 y, with significant differences from 0.5 meters and

background sediments (t=6.05, d.f.=5, p=0.0018; Fig. 2.2). Sedimentary C/N ratios

confirm the presence of relatively labile (i.e., nitrogen-rich) material at 0 m after 0.25


y. C/N ratios were significantly lower at 0.25 y at 0 meters than all other distances

and background sediments (t=-15.94, d.f.=4, p<0.01; Fig. 2.2).

Sedimentary TOC and TON around wood parcels resembled those in

background levels at 0.5 and 1.8 y after deployment (Fig. 2.2). However, by the 3 y,

surface sediments adjacent to wood parcels had become massively enriched in organic

enrichment, with TOC 3-6 times higher than background values (t=17.37, d.f.=1,

p<0.001; Fig. 2.2). In addition, C/N ratios were extremely high (up to 82.9 C:N;

t=52.49,d.f.=1, p<0.001) compared to all other distances sampled, suggesting

enrichment from relatively refractory woody material. Mean TOC and C/N ratios in

surface sediments near wood falls were still well above ambient levels after 5.5 y

(Fig. 2.2). In addition, percent organic nitrogen was depressed in surface sediments

near the wood parcel to distances of 2 m after 5.5 y (ANOVA, F=61.64, d.f.=1,

p=0.0005), suggesting that low C/N ratio woody organic material had spread some

distance from the wood parcel (Fig. 2.2).



Figure 2.2. Upper panels: Mean (+ 1 SE) TOC and TON in surface sediments (0-2 cm)

around kelp parcels. C/N ratios are shown in text format, with upper and lower values representing 0.25- and 0.5 y averages, respectively. Horizontal parallel lines demark average

%TOC/TON (± 1 SE) of background sediments and C/N ratio (in Italics). Lower panels: Temporal variability on TOC and TON in surface sediments (0-2 cm) around wood parcels

(± 1 SE when n >1). C/N ratios are shown for 0 m and background samples only (in text format).

Macrofaunal density, composition and patterns of succession

Kelp parcels

Spatial and temporal patterns of species composition and trophic structure

indicate that kelp falls dramatically influenced infaunal communities over short

spatial scales (< 1 m) for at least 0.5 y. At 0.25 y, macroinfaunal abundance within

0.5 m of kelp parcel was significantly enhanced by ~5–fold relative to background


sediments (Fig. 2.3A). This pattern persisted relative unchanged after 0.5 y (Kruskal-

Wallis, H =24.05; d.f.=8, p=0.002; Fig. 2.3A - Inset). At distances of over 0.5 meters

from the kelp, macroinfaunal abundance remained near ambient levels at both 0.25

and 0.5 y (Fig. 2.3A).

Dorvilleid polychaetes and cumaceans dominated sediments within 0.5 m of

kelp falls after 0.25 y and 0.5 y, but were rare in ambient sediments (Fig. 2.3B). The

four top-ranked species at 0.25 y constituted over 88% of all individuals at 0 m, and

dorvilleids and cumaceans represented over 36 and 34 % of the total macrofauna,

respectively (Table 2.2, Fig. 2.4). At 0.5m, cumaceans still responded for over 55%

of all macrofaunal individuals at 0.25 y (Fig. 2.3B, Fig. 2.4). At distances of ≥ 1 m

from the kelp parcels at 0.25 y, background species dominated the infaunal

community. At 0.5 y, cumaceans and dorvilleids still dominated the macrofaunal

community at a distance of 0 m, with background species becoming abundant at

distances ≥ 0.5 m form the kelp (Fig. 2.4, Table 2.2). The rapid increase in abundance

of the two species of cumaceans ( Cumella sp. A and Cumacean sp. K) and the

dorvilleid Ophryotrocha sp. A within 0.5 m from the kelp treatments suggests this

species are opportunistically responding to enrichment and disturbance conditions

(Grassle & Grassle 1974, Pearson & Rosenberg 1978). The cumacean species that

responded to the kelp enrichment in this study were not typically encountered in our

background communities, but they colonize enrichment experiments in deep-sea

regions, including fish and whale-falls in the NE Pacific (Smith 1986, Snelgrove et al.

1996, Smith et al. in prep).



Figure 2.3. A. Mean macrofauna abundance at 0.25- and 0.5-y kelp parcels deployed in SCr

Basin. The solid black line denotes the average background density (± 1SE, dotted line).

Inset: A posteriori multiple comparisons based on Kruskal-Wallis test (H=24.05, p<0.001).

Groups not underscored by a common line are statistically different at p<0.05. No significant

difference was found between 0.25- and 0.5-y at the same distance. B. Abundance of top-

ranked species at 0.25 y (top panel) and 0.5 y (bottom panel) kelp parcels.

The dominance of a Ophyotrocha sp. A and the two cumacean species,

lowered species evenness ( J’) and diversity nearby kelp treatments, at both times

sampled (Table 2.3). At 0 m, species dominance was significantly higher than

background sediments during both periods sampled (Kruskal-Wallis, H= 12.96, d.f.=

4, p=0.01). Species diversity (ESn) increased from 0.25 y to 0.5 y within 0.5 meters of

kelp parcels (Table 2.3, Fig. 2.5). At 0.25 y, rarefaction diversity was significantly

lower within 0.5 m of the kelp treatments, if compared to background samples (Fig.

2.5). At 0.5 y, diversity increased towards background sediments at distance of over

0.5 m, but diversity was still significantly depressed at the 0 m macroinfaunal

community (Fig. 2.5). The diversity changes around kelp treatments suggest that

although areas nearby the parcels were being negatively affected by the enrichment

and disturbance conditions after 0.5 y, the macrofauna in sediments over 0.5 m had

returned to background conditions.




Table 2.2. Mean density per core (38.5 cm-2) and relative abundances of top 5 ranked

macrofaunal species at 0 m and 0.5 m from kelp parcels, and background sediments in SCr

Basin. (P) Polychaeta, (Cr) Crustacea, (M) Mollusca.

Figure 2.4. Macro-infauna composition around kelp parcels at 0.25 y (a) and 0.5 y (b) in SCr

Basin. Bkgd - Background fauna from SCr Basin in 2002, 2004 and 2005 (n=17 cores).

Table 2.3. Diversity indices from kelp parcels and background sediments at SCrB. Mean

number of individuals (M) per core (38.5cm-2 ± 1SE). J’-Pielou eveness index, ESn- Expected number of species at 10, 25 and 50 individuals (inside brackets) per sample.



ES 10

ES 25 (50)

0.25 y

0.5 y

0.25 y

0.5 y

0.25 y

0.5 y

0.25 y

0.5 y










































3.3 (0.1)




Figure 2.5. Rarefaction diversity for Kelp macrofauna 0.25 y and 0.5 y.

Trophic-group analysis revealed large changes adjacent to kelp parcels (i.e., ≤

0.5 m) after 0.25 – 0.5 y. At 0.25 y, the high abundance of dorvilleids and cumaceans

dramatically increased the proportion of omnivores (OMNI, Kruskal-Wallis,

H=6.836, d.f.=2, p=0.02) and the “other” trophic group (Kruskal-Wallis, H=16.624,

d.f.=2, p<0.01) within 0.5 meters from the treatments (Fig. 2.6). The augmented

number of “other” trophic group at 0 m was still significant after 0.5 y (p<0.01; Fig.

2.6). Surface deposit feeders (SDF) were slightly depressed nearby kelp parcels

compared to background sediments, but no significant trend was observed (Fig. 2.6).

In addition, sub-surface deposit feeders (SSDF) were wholly absent adjacent to kelp

parcels (0 m) after 0.25 -0.5 yr (Kruskal-Wallis, H=5.70, d.f.=1, p=0.018; Fig. 2.6).

This suggests that sub-surface deposit feeders were particularly affected by the

physical and organic disturbances in sediments nearby the kelp parcels. Trophic

differences between 0.25 and 0.5 y at 0.5 m suggests recovery towards background

conditions at 0.5 y, with an increase in the importance of SDF accompanied by a



strong decrease in the “other” trophic group (Fig. 2.6). The elevated number of

omnivores and “other” trophic guilds are consistent with the opportunist species that

colonized sediments and possibly with the augmented amount of detritus within 0.5 m

from the treatments. Complete abundance and relative contribution of species at kelp

parcels are presented in Appendix 2A.

Figure 2.6. Percentage of total abundance in trophic groups at 0.25 y and 0.5 y kelp parcels

at SCr Basin. OMNI-Omnivores, Carnivores and Scavengers; SDF-Surface deposit feeders;

SSDF-Sub-surface deposit feeders, Other- Trophic groups not determined (see methods).

Wood parcels

Wood parcels also exhibited dramatic effects on macrofaunal community

structure in space and time. The sediment near the wood parcels accumulated a

massive number of Xylophaga washingtonia recruits during the first 0.5 – 3 years

(Fig. 2.7A), presumably resulting from settling response to available wood substrate

nearby. Brooder species of Xylophaga present on the wood parcels may have been the

source of juveniles that rained down to the sediments after 0.5 y (Voight 2007a). The

Xylophaga bivalves never attained adult size within the sediments, although adults

were present on the wood parcels from 0.5 y onward. Thus, Xylophaga represents a

sink population resulting from mass effects of the wood parcel (Leibold et al. 2004).


Because Xylophaga appear to represent sink populations in the sediment community

derived from mass effects of the wood parcels, we also evaluated macrofaunal

community patterns with Xylophaga removed from the analyses. At 0.5 y, mean

macrofaunal abundance, without Xylophaga, was significantly higher at 0 m than in

background sediments; at further distances macrofaunal abundances resembled

background levels (Fig. 8B; Kruskal-Wallis test, H=27.98, d.f.=13, p=0.03).

Macrofaunal densities around the wood parcels were significantly elevated at 0 m

after 1.8 y (ANOVA, F=12.906, d.f.=3, p=0.003), dropping to background levels by

0.5 m from the parcels (Fig. 2.7B). At subsequent sampling times (3.0 and 5.5 y),

macrofaunal abundance progressively increased adjacent to the parcels, reaching very

high numbers (15,100 ± 2400 ind.m-2) at 0 m after 5.5 y (p<0.001; Fig. 2.7B).

However, macrofaunal enhancement was limited in a spatial scale, with slight

increases after 1.8 – 5.5 y at 0.5 m, at all time points, and background levels of

abundance at > 1 m from parcels even at 5.5 y, when abundances at 0 m were 15

times background levels (Fig. 2.7B).

At the species level, Cumacean sp. K dominated macrofaunal assemblages

adjacent to wood parcels after 0.5 y, although three background species ( Monticellina

sp. A, Chaetozone sp. D and Cossura rostrata) remained among the 5 dominant

species (Fig. 2.8, Table 2.4). By 1.8 y, Ophryotrocha sp. A (a dorvilleid) and

ampharetid polychaetes dominated assemblages at 0 - 0.5 m, with cumaceans also

abundant at 0.5 m (Table 2.4); at further distances, background polychaetes continued

to dominate (Fig. 2.8). By 3.0 y, dorvilleids (two species of Ophrytotrocha) had

achieved very high densities at 0 m, representing over 40% of the macrofauna (Fig.

2.8, Table 2.4) and the chemoautotrophic-symbiont-containing mytilid Idas

washingtonia was first encountered at the wood parcels (Table 2.4). At 5.5 y,



dorvilleids ( Ophryotrocha and Parougia) and two species of ampharetids dominated

the macrofauna at 0 m, with very high abundances (Table 2.4, Fig. 2.8); Cumacean

sp. K, was also abundant. None of the dorvilleids, ampharetids and cumaceans were

collected in the background community, or even off distances of 2 m from the wood

falls. This, plus their larger response to kelp parcels suggest that they are

opportunistic species, or may be wood associated species enhanced by mass effects

(Leibold et al. 2004) from the adjacent wood parcels.

Figure 2.7. Temporal variability on macroinfaunal abundance at the wood parcels (Average

± 1SE). A. Xylophaga recruits included; B. Xylophaga not included. The solid black line and

shadowing represent the average background (100m) abundances (± 1SE).




Figure 2.8. Sediment macrofaunal composition at the wood parcels deployed at SCr Basin. a.

0.25 y, b. 1.8 y, c. 3 y, d. 5.5 y.

Table 2.4. Mean density per core (38.5 cm-2) and relative abundances of top 5 ranked

macrofaunal species at 0m and 0.5m at wood parcels and background fauna at SCr Basin.

The total percentage of top ranked species is given below each list. (P) Polychaeta, (Cr) Crustacea, (M) Mollusca.




The increase in macrofaunal density at 0 m from 0.5 y to 5.5 y was followed

by a significant decrease on community evenness in the latter period ( J’, ANOVA,

F=8.384, d.f.=3, p=0.001, Table 2.5), resulting from higher densities of opportunist

species. At the other periods, J’ was always within background levels (Table 2.5).

Rarefaction diversity demonstrates a clear temporal pattern adjacent to wood parcels.

At 0.5 y, rarefaction diversity adjacent to the wood parcel falls above background

levels (Table 2.5, Fig. 2.9); diversity then progressively declines at 0 m from 1.8 to

5.5 y (Table 2.5, Fig. 2.9). The diversity curve at 5.5 y falls well below all other

curves and is significantly lower than background (Fig. 2.9).

Table 2.5. Average number of individuals (M) per core (38.5cm-2 ± 1SE), Pielou’s species evenness (J’) and expected diversity around wood parcels deployed at SCr Basin. Expected

number of species E(Sn) estimated from 10 and 25 individuals (ES25 inside brackets).


Figure 2.9. Hulbert’s rarefaction plot detailing the temporal variability on macroinfaunal

diversity at the wood parcels (0 meters only).

Functional group analysis also indicated dramatic wood parcel effects on the

adjacent macrofaunal community, increasing in intensity from 0.5 – 5.5 y. At 0.5 y,

the dominant trophic groups were largely typical of ambient sediments, although at 0

m the importance of the “others” group slightly increased due to the presence of the

apparently opportunistic Cumacean sp. K (Fig. 2.10). By 1.8 y, omnivores at 0 m had

increased to 26%, and rose to 50% of the macrofauna at 3.0 - 5.5 y (Fig. 2.10). After

3 y, the omnivores within 0.5 m from wood falls were significantly more numerous

than the 0.5 y community and the background sediments (Kruskal-Wallis, H=12.19,

d.f.=4, p=0.016). SDF’s were relatively abundant during the whole period analyzed

(over 30%), with ampharetids being dominant nearby the parcels and cirratulids in

background sediments. Subsurface deposit feeders usually represented less than 1% of

the infauna at 0 m during all periods (Fig. 2.10). Complete abundance and distribution

of species at wood parcels are presented in Appendix 2B.



Figure 2.10. Trophic group patterns at the wood parcels implanted at SCr Basin. OMNI-

Omnivores, Carnivores and Scavengers; SDF-Surface deposit feeders; SSDF-Sub-surface

deposit feeders, Other-Trophic groups not determined.

MDS patterns of faunal distribution at Kelp and Wood parcels

Nonmetric Multidimensional Scaling (MDS) analyses further elucidate

successional patterns in space and time around the kelp and wood parcels. At kelp

falls, 0 m samples from 0.25 and one replicate from 0.5 y separated strongly from all

other distances (Fig. 2.11, Two-way ANOSIM R= 0.358, P=0.001). The top-ranked

species at 0 m accounted for over 70% of the dissimilarity from the other samples at

both time intervals (SIMPER). All other samples from 0.5 m to background, clustered

with each other at 0.25 and 0.5 y, with no distinct difference between them (Fig. 2.11,

ANOSIM). Overall, highest dissimilarities in sediments nearby kelp falls and

background communities resulted from the dominance of apparently opportunistic



species adjacent to the kelp fall at 3 months, and the low background densities of

local species.

Figure 2.11. Multidimensional scaling plot (MDS) of macrofaunal communities (>300µm) at

Kelp (top panel) and Wood (bottom panel) parcels on SCr Basin. Grouped symbols denote

stronger similarity among samples (see text).

At wood parcels, the MDS also revealed marked spatial and temporal patterns

(Two-way ANOSIM, p<0.01). At 0.5 y and 1.8 y, all samples generally grouped

together with no evident difference between treatments and background communities

(Fig. 2.11, ANOSIM). A dramatic shift in macrofaunal community structure was

indicated at 0 m at 3 y and persisted to 5.5 y (Fig. 2.11, ANOSIM, R= 0.319, P=

0.001). The top ranked macrofaunal species (including apparently opportunistic

dorvilleids in the genus Ophryotrocha) at these two later periods accounted for over

20% of the spatial dissimilarity among samples (SIMPER). Pairwise comparisons


revealed a lower degree of dissimilarity between 5.5 y and the two first periods

sampled, suggesting that macrofaunal community structure by 5.5 y may have been

recovering towards background conditions (ANOSIM).

These results confirm our first hypothesis, that kelp and wood falls promote

very distinct patterns of macrobenthic community succession. The dynamics of

dominant species were distinct between kelps and wood, with kelps promoting early

changes (scales of 1-3 months) on community composition, which lowered local

diversity in the short scale as a result of a massive colonization by a few omnivore

species (cumaceans and dorvilleids). On the other hand, dominant species altered

community structure around wood parcels only after 1 – 2 years, but the duration of

the organic enrichment promoted significant changes for over 5 years. A higher

number of species colonized sediments around wood falls, likely as a consequence of

increased food availability, but the fauna became less diverse with time as densities of

opportunist species increased.

Stable isotope analysis at Kelp and Wood parcels

Potential primary organic matter sources at kelp and wood falls include kelp

and wood biomass, sedimentary matter from deposited phytoplankton, and biomass

from chemoautotrophic sulfur-oxidizing bacteria utilizing sulfide emitted from kelp or

wood parcels. Kelp biomass (Macrocystis pyrifera) had the highest δ 13C and δ15N

isotope values from all primary organic sources (Table 2.6). Sediment organic carbon

and wood fragments had very similar isotope signatures (Table 2.6), which were

consistent with previous studies in the basin (Baco-Taylor 2002). Mats of sulfur-

oxidizing chemoautotrophic bacteria growing on whale skeletons off southern

California had δ 13C signatures similar to the other organic sources but with the lowest


δ15N values compared to the other sources (Table 2.6). Isotope signatures from sulfur-

oxidizing bacteria can have a much wider isotopic range due to variable inorganic C

sources (Ruby et al. 1987).

Table 2.6. Stable isotope values from primary organic matter sources used in this study. End

members values were used in the mixing model calculations in order to account for maximum

environmental variability. Values in delta notation.



End members


Average (min/max)

13C(Min/Max) 15N(Min/Max)







6.7 (6.6/6.8)



This study


Page et al.,

Kelp biomass



9.1 (8.5/9.7)





Wood biomass



5.5 (4.9/6.1)