Structure and zonation of demersal fish assemblages off the Azores Archipelago (mid-Atlantic)

20 Pages·2006·1.8 MB·English

MARINE ECOLOGY PROGRESS SERIES
Vol. 324: 241–260, 2006 Published October 23
Mar Ecol Prog Ser
Structure and zonation of demersal fish assemblages
off the Azores Archipelago (mid-Atlantic)
Gui M. Menezes1,*, Michael F. Sigler2, Helder M. Silva1, Mario R. Pinho1
1Departamento de Oceanografia e Pescas, Universidade dos Açores, Cais de Santa Cruz, 9901-862 Horta, Portugal
2US National Marine Fisheries Service, Alaska Fisheries Science Center, Auke Bay Laboratory,
11305 Glacier Highway, Juneau, Alaska 99801, USA
ABSTRACT: The assemblages of the demersal fish fauna of the Azores Archipelago are described
from longline surveys that extended from the coastline to 1200 m water depth. A total of 104 fish spe-
cies from 47 different families were caught, and despite the changes of biogeographic affinities with
depth, most species caught are of subtropical origin (mainly from the Eastern Atlantic/Mediterranean
areas) or have a broad geographic distribution. Four large-scale fish assemblages following a depth-
aligned structure were found: a shallow-shelf/shelf-break assemblage at depths < 200 m, an upper-
slope assemblage at 200–600 m, a mid-slope assemblage at 600–800 m and a deep mid-slope assem-
blage at 800–1200 m. Within the main shallow assemblage, 4 small-scale fish assemblages were
found: an inner-shelf-island assemblage, an outer-shelf-island assemblage, a seamount/island-
shelf/shelf-break assemblage and a transitional shelf/break assemblage. The bathymetric delin-
eation of the mid-slope assemblages coincides with the known distributions of the North Atlantic
Central Water (NACW), Mediterranean Water (MW) and the upper influence of the intermediate
waters in the region: the northern sub-polar waters (Subarctic Intermediate Water [SAIW], the
Labrador Sea Water [LSW]) and the Antarctic Intermediate Water (AAIW). The delineation of the
shallow small-scale fish assemblages appears to be determined by small-scale environmental factors
(e.g. bottom characteristics, seamounts or island areas).
KEY WORDS: Azores Archipelago · Demersal fish assemblages · Islands · Seamounts · Fish
distribution· Zonation· Water masses
Resale or republication not permitted without written consent of the publisher
INTRODUCTION have been studied for several coastal habitats in the
Azores (Azevedo 1997, Afonso 2001).
The ichthyofauna of the Azores islands has been Description, structuring factors and persistence of
studied since the end of the nineteenth century (Santos demersal fish assemblages have been documented for
et al. 1995), and more than 460 marine fish species have several continental-shelf and -slope habitats (e.g.
been listed for the area (Santos et al. 1997). During the Colvocoresses & Musick 1984, Overholtz & Tyler 1985,
1970s and 1980, fishing surveys were conducted by the Mahon & Smith 1989, Bianchi 1992a,b,c, Gomes et al.
former USSR (Vinnichenko 2002), Norway (Dias et al. 1992, 1995, 2001, McKenna 1993, Rogers et al. 1998,
1976), and Portugal (Dias & Cascalho 1991). In 1995,the Gaertner 2000, Williams et al. 2001). Knowledge of
Department of Oceanography and Fisheries (of the the deep-demersal fish fauna in the Atlantic also has
University of the Azores – DOP/UAç) began regular improved considerably in recent years (Mauchline &
longline surveys (Menezes 2003), providing data for Gordon 1985, Haedrich & Merrett 1988, Merrett et al.
new descriptions of the structure and spatial distribu- 1991, Stefanescu et al. 1993, Gordon et al. 1996,
tion of demersal and deep-water fish assemblages in Haedrich 1997, Merrett & Haedrich 1997, Bergstad
the Azores. In recent years, shelf fish assemblages also etal. 1999, D’Onghia et al. 2004).
*Email: [email protected] © Inter-Research 2006 · www.int-res.com
242 Mar Ecol Prog Ser 324: 241–260, 2006
Benthic and benthopelagic fish communities typi- identify major fish-assemblage boundaries, elucidat-
cally are studied using bottom trawls. On the Mid- ing, whenever possible, their environmental deter-
Atlantic Ridge (MAR), commercial trawling has been minants.
conducted, but only a few exploratory fishing surveys
have used trawls (Hareide & Garnes 2001, Vin-
nichenko, 2002). Recently, exploratory fishing target- STUDY AREA
ing orange roughy Hoplostethus atlanticus took place
within the Azores EEZ below 800 m depth, and more The archipelago of the Azores (36 to 40°N, 24 to
than 20 new fish species were recorded for the area 32°W) is the most isolated and extensive island group
(Melo & Menezes 2002). The use of trawls in certain in the northeastern Atlantic (Fig. 1). It is composed of 9
areas, such as the Atlantic islands, is difficult due to the volcanic islands distributed in 3 groups along a tec-
steep, rocky and irregular bottom. Because of this, tonic zone extending about 600 km WNW–ESE and
longlines have been used in most studies of the demer- several small islets, with a coastline that totals about
sal and deep-water fish species assemblages in these 790 km. As a volcanic archipelago of recent origin, the
archipelagos (Uiblein et al. 1996, Menezes 2003). Azores islands are characterized by very narrow
In this study, we characterize the structure and spa- shelves and steep slopes. The seafloor around the
tial distribution of the shelf and slope demersal fish Azores is very irregular and rocky, with numerous sub-
fauna of the Azores islands and identify the main fish marine elevations of very different size, shape, depth
assemblages from a few meters to 1200 m water depth. and degrees of isolation. The size of the Azores EEZ is
To our knowledge, this is the first time that longline about 1000000 km2, with an average depth of about
gear has been used to study the ecological features of 3000 m, but the exploitable fish habitat is relatively
the demersal ichthyofauna of the Azores over such a small, with the depths less than 600 m covering only
wide depth range. The objectives of this work were to about 1% of the entire EEZ (7000 km2; Menezes 2003).
(1) characterize the species composition and general During winter a deep mixed layer is present around
structure of the catches, (2) determine major demersal 150m, and in summer a seasonal thermocline develops
fish assemblages of the Azores waters, as well as around 40 to 100 m. Average sea surface temperature
describe their general composition, spatial distribu- varies between 15 and 20°C during winter and be-
tion, and characteristic species, and (3) examine and tween 20 and 25°C during summer (Santos et al. 1995).
40°N
Corvo 100 km
Flores –2000 –2000 Graciosa –2000 600 m depth
39°N –2000
Faial São Jorge –2000 Terceira
DJC Bank
AÇO BankPico –2000–2000
38°N São Miguel
PAL Bank –2000
Azores –2000
Madeira
Canaries –2000
37°N VCearpdee AWfreicsat –2000 MPR Ban–k2000
Atlantic Sta. Maria
Ocean
32°W 31°W 30°W 29°W 28°W 27°W 26°W 25°W 24°W
Fig. 1. The Azores archipelago. Location of fishing stations sampled during bottom longline surveys covered by the RV
‘Arquipélago’ in 1995, 1996, 1997, 1999, and 2000. Stations are indicated with circles. Abbeviations for the 4 banks covered
during the cruises are as follows: PAL – Princesa Alice; AÇO – Açores; DJC – Dom João de Castro; MPR – Mar da Prata
Menezes et al.: Demersal fish assemblages off the Azores 243
Currents. The Azores region is located at the the demersal ichthyofauna. Most sampling covered
northern edge of the North Atlantic Subtropical Gyre depths up to 600 m, but was extended to 1200 m depth
(SG), characterized by a high horizontal temperature at some stations. All surveys followed a stratified ran-
gradient, with the oceanographic conditions being dom design; stratification was based on geography and
strongly influenced by the Gulf Stream western cur- depth (50 m depth intervals). Areas and sub-areas
rent, which transports warm water of equatorial and were defined based on their geographical proximity
tropical origin into colder northern water. The cur- (e.g. central group of islands) and type (e.g. islands,
rent system that surrounds the Azores is complex seamounts). The average number of 30 stations per
(Alves & Verdière 1999, Johnson & Stevens 2000, annual cruise was proportionally allocated according
Bashmachnikov et al. 2004). The system is largely to the size of each area or sub-area (from 0 to 600 m
dominated in the south by the warm Azores depth). The station locations were randomly selected
Front/Current (AFC or AzC) system, that crosses the within each sub-area.
MAR at 34 to 36°N, and in the north by the cold In all surveys we used the ‘long-chain shots’ strategy
southern branch of the North Atlantic Current (NAC; (a long transect), where a continuous length of gear
Alves & Verdière 1999, Bashmachnikov et al. 2004). was deployed at 0 to 600 m (or 0 to 1200 m at some sta-
One additional flow lies between the NAC and the tions). The sample units were the depth stratum obser-
AzC and is found between 39 to 42°N (Bashmach- vations within each station. In most years, the data
nikov et al. 2004). In addition, mesoscale eddies (100 were stratified by 50 m increments. Different depth
to 200 km) on both sides of the AzC have been stratification was applied for the 1995 survey: 50 m
reported (Alves & Verdière 1999). The small and depth interval at 0 to 200 m depth; 100 m depth inter-
mesoscale oceanographic features of the region are val at 200 to 600 m depth; and 200 m depth interval at
poorly known. Upwelling phenomena are frequently 600 to 1000 m depth. For this reason, the data from the
observed around the Azores islands through high- 1995 survey were excluded from some of the analysis
resolution SeaWiFS satellite imagery (A. Martins presented in this work, and in other cases only the
pers. comm.). The complex bottom topography also comparable first 4 depth strata (50 m strata) were used.
causes other oceanographically important phenom- A total of about 620000 hooks were deployed and
ena such as jets and/or trapping currents (e.g. Taylor 145stations were sampled during the annual longline
columns) around seamounts. surveys.
Water masses. Several water masses occur around Samplinggear:A standardized longline gear baited
the Azores. The North Atlantic Central Water (NACW) with ‘chopped salted sardine’ was used for all surveys.
is located at depths shallower than 700 m and is out- The longline gear used is identical to the one typically
lined by the permanent thermocline. Beneath the used in the Azorean commercial fishery and is locally
permanent thermocline lie the intermediate waters known as ‘stone/buoy longline’. This gear minimizes
formed at high latitudes. In the region, northern sub- the risk of gear loss in the typically rugged bottoms;
polar waters (Subarctic Intermediate Water [SAIW], buoys spaced 74 m apart float the gear off the bottom
the Labrador Sea Water [LSW]) and Antarctic Inter- and alternate with stones that weight the gear. The
mediate Water (AAIW) predominate at intermediate longline units (approximately 37 m long) are termed
depths. Mediterranean Water (MW) can penetrate into ‘quarter-skate’ and contain 25 to 30 J-type hooks no. 9
the region at about 800 and 1200 m (Santos et al. 1995, (Mustad size equivalent with a gape of 12 mm)
Johnson & Stevens 2000). (Menezes 2003).
SAIW and LSW water masses make a major contri- Biological and fishing data: All fishes caught were
bution to the characteristics and export of North sorted and identified to the lowest taxonomic level
Atlantic Deep Water (NADW) (Dinter 2001). This possible (usually species), counted, weighed, and
water, occurring between about 2000 and 4500 m measured. More detailed biological sampling was
(Mann & Lazier 1996), is the dominant water mass conducted on some species. Species catch data, in
below intermediate depths. number and weight, were standardized by effective
fishing effort (per 1000 hooks). The effective fishing
effort (those hooks that fish effectively) by station and
MATERIALS AND METHODS depth stratum was calculated by subtracting the
estimated missing hooks (lost or tangled hooks) from
Surveys and data.We analyzed the data from long- the estimated total number of hooks deployed. Esti-
line surveys conducted annually off the Azores Archi- mates of the number of lost hooks were obtained
pelago between 1995 and 2000 (except 1998) by the through sampling and classification of the hooks at
RV ‘Arquipélago’. The sampling design was primarily thetime of gear retrieval (for more details, see Mene-
implemented to estimate annual abundance indices for zes et al. 2001).
244 Mar Ecol Prog Ser 324: 241–260, 2006
In this study, nomenclature, species identification, samples and identitying species using 2 multivariate
and auxiliary information on species distribution, zoo- techniques of classification and ordination: the 2-way
geographic affinities, and prey habitats, follow Fish- indicator species analysis (TWINSPAN, Hill et al. 1975)
Base (Froese & Pauly 2005), Compagno (1984a,b), and non-metric multidimensional scaling (MDS), intro-
Whitehead et al. (1984–1986) and Santos et al. (1997). duced by Shepard (1962, in Clarke & Warwick 1994)
Environmental data: Seven abiotic measurements and Kruskal (1964, in Clarke & Warwick 1994).
were recorded: longitude, latitude, mean depth, mean To complete these analyses, a pooled 2-way data
temperature, steepness (local bottom slope), irregular- matrix (station/depth stratum vs species) was built
ity (depth variability) and depth stratum size (horizon- from all 4 survey years (1996, 1997, 1999 and 2000).
tal distance between the start and end positions of Species were excluded if they occurred in less than 1%
each depth stratum). Temperature profiles were col- of the stations. The pelagic species Scomberjaponicus
lected at some stations using expendable bathyther- was excluded because they are likely caught in the
mograph probes (XBT). Average temperature depth water column during line deployment or retrieval. The
profiles (50 m interval) were calculated and assumed final data matrix used in the multivariate analysis
equal for all stations sampled that year. resulted in 1364 sampling units ×29 species.
Steepness, irregularity and depth stratum size were The TWINSPAN classification analysis was completed
calculated from data collected during gear deployment. by applying the routine implemented in the software
Geographic position and depth were recorded at the PC-ORD, ver. 4.25 (developed by McCune & Mefford
beginning of each quarter-skate longline section (QS). 1999). This classification method is a divisive method
This data was then used to estimate the station/depth that classifies sites and species and produces a list of
stratum steepness as follows (in degrees): sorted species by station. This allows the identification of
groups of stations showing faunal similarities in terms of
⎛50⎞
steepness =arctan⎜ ⎟ with L =n ×L species composition and relative abundances. Despite
ij ⎝L ⎠ ij ij QS
ij some criticism of this method (Clarke & Warwick 1994,
where 50 m is the constant depth stratum height, L is Legendre & Legendre 2000), TWINSPAN has the ad-
ij
the total estimated horizontal length of the gear de- vantage of handling a large number of samples without
ployed at stratum jof station i, n is the total number of needing to group samples into a manageable number.
— ij
QSs deployed, and L is the estimated mean length of Final results are displayed in an easy-to-read 2-way
QS
each QS of the longline used, which was assumed to be table arranged according to station and species simi-
about 36.5 m length. larities. Station clusters can be characterized by their
The irregularity index of each station/depth stratum differential species compositions, diversity, total biomass
was calculated according to the following expression: orother ecological features.
The ordination analysis was completed using the
n∑–1
(dijm+1–dijm)2 S2 MDS routine implemented in the software PRIMER
S2 = m=1 irregularity = ij (Plymouth Routines in Multivariate Ecological Re-
ij n –1 ij (d –d )
ij ijmax ijmin search) for Windows ver. 5.2.2 (2001). MDS positions
where S2is the variance of the successive differences samples in a 2-dimensional space such that the relative
ij
ofthe depth soundings dof the QSs m, divided by the to- distance of the samples reflects their relative similar-
tal depth-sounding range (d – d ) for each stratum j ity/dissimilarity, with samples most similar occurring
max min
at station i. In those cases where only 1 or 2 QS werede- closer together. With more than 100 samples, the MDS
ployed in a given depth stratum, the irregularity index analysis increases the complexity of the sample
value was assumed equal to the value of the preceding relationships and a 2-dimensional representation is
stratum of that particular station. The type of substrata unlikely to be adequate (Clarke & Warwick 1994). In
(e.g. rock, sand, mud) is an important structuring factor this study, 2 MDS ordinations were performed. As
that affects the spatial distribution and abundance of the suggested by Clarke & Warwick (1994), the initial
demersal fish fauna. In the absence of this type of auxil- large data sets were sub-divided a priori on the basis
iary information, the steepness and irregularity indices ofthe subsets obtained from the TWINSPAN classifica-
can be regarded as proxies for those sea-bed topo- tion analysis. The samples within each group (clusters
graphic features. Empirically, higher values of the irreg- of sample sites at the 4th TWINSPAN dichotomy level)
ularity index are expected for rocky areas and lower val- were averaged and then input to the MDS to display
ues of the steepness index are expected for soft-bottom ‘large-scale’ structures. In a second ordination, all
areas, where sedimentation is expected to be greater. observations (sampling units) within each of the 4 main
Analytical methods. Multivariate analysis: We ex- groups obtained at the 2nd TWINSPAN dichotomy
amined the large-scale structure of the demersal fish level were analysed separately (without averaging) to
assemblages by evaluating the similarities between display any ‘small-scale’ structures.
Menezes et al.: Demersal fish assemblages off the Azores 245
All biotic station similarity matrices used in the MDS RESULTS
were constructed using the Bray–Curtis similarity
index (Clifford & Stephenson 1975). Because the long- Taxonomy and abundance
line catches and the sampling fishing efforts allocated
in each station/depth-stratum combination were, in A total of 44881 fish (weighing a total of 41968 kg) and
general, highly variable, the data were standardized representing 104 species belonging to 49 families were
(each species abundance divided by the total sample) caught in surveys from 1995 to 2000. Teleost fishes were
and log(x+ 1) transformed. caught in large numbers comprising 78 species from 39
Interpretation techniques: Fish assemblages (site families, while elasmobranchs comprised only 26 species
groups) emerging from the TWINSPAN analysis are belonging to 10 families (Appendix 1, available at
characterized by homogeneous faunal composition. www.int-res.com/articles/suppl/m324p241_app.pdfin
Each group can be distinguished not only by their dif- MEPS Supplementary Material). Among the 39 families
ferential species composition but also by their typify- identified, the Scorpaenidae and the Moridae families
ing species. The Indicator Values (INDVAL) method were represented by the largest number of species, with
(Dufrene & Legendre 1997) implemented in PC-ORD 6 species each. Among the 10 elasmobranch families, the
ver. 4.25 was used to identify those typifying species Dalatiidae was the best represented, with 10 species, fol-
for each group at the 4th TWINSPAN dichotomy level. lowed by the Centrophoridae and the Rajidae, with 4
The relationship between the biotic MDS configura- species each. About 51% of the species were benthic,
tion and the environmental datawas analyzed with the 39% benthopelagic and only 10% pelagic (Fig. 2,
BIO-ENV procedure (Clarke & Ainsworth 1993) imple- Appendix 1). The number of benthic and pelagic species
mented in PRIMER. Two analyses were performed at 2 generally decreased from shallow to deep waters,
different levels of data aggregation and, therefore, at whereas the number of benthopelagic species increased
2 different levels of the assemblage structure. One in the deeper strata, dominating below 800 m depth.
analysis was conducted on the average similarity Most species were classified as subtropical (51%) and
matrix at the 4th TWINSPAN dichotomy level groups mainly from the Eastern Atlantic and Mediterranean
(‘large-scale’ structure), and the second was per- areas (33%) (Figs. 3 & 4). Those species with broad
formed separately for each of the 4 main clusters (at geographical distributions (i.e. a ‘wide distribution’,
the 2nd TWINSPAN dichotomy level), using all within- Appendix 1) dominated the deepest strata, while the
group similarities (‘small-scale’ structure). This analy- proportion of subtropical species was higher in shallower
sis consists of choosing a subset of environmental vari- waters, greatly decreasing with depth.
ables that provides the best match with the biotic MDS According to the assumed distribution patterns of
ordination of the samples. The strength of agreement their preys (Fig. 5, Appendix 1), most of the species
between the biotic and abiotic samples dis/similarity caught feed on benthopelagic species (45%). The
matrices was determined by the Spearman rank corre- importance of the epibenthos preys decreased in the
lation coefficient (Clarke & Warwick 1994). The envi- deepest strata, whereas the proportion of fishes that
ronmental dissimilarity matrices used in the BIO-ENV feed on pelagic and benthopelagic preys increased in
analysis were calculated with the normalized Euclid- the deepest strata.
ean distance, which seems more appropriate for envi- In addition, the species were classified according to
ronmental data (Clarke & Warwick 1994). their relative abundance and relative frequency of
100
Pelagic
%) Benthic 10%
y ( 80 51%
c
n
e 60
u
q
e
r
e f 40
v
ati
el 20
R
0
Benthopelagic
25 125 225 325 425 525 625 725 825 925 1025 1125
39%
Depth strata - midpoints (m)
Fig. 2. Species lifestyles in each depth strata and in the total catch
246 Mar Ecol Prog Ser 324: 241–260, 2006
100 CIRCG CIRC-
%) WIDE 9% TROP 1%
80 27% EA
y ( 10%
c
n MACR
e 60
u 1%
q
e
r
e f 40
v
ati
el 20
R AMPHIA-
MED4% EA-MED
0
25 125 225 325 425 525 625 725 825 925 1025 1125 AMPHIA END 33%
Depth strata - midpoints (m) 14% 1%
Fig. 3. Species distribution patterns in each depth strata and in the total catch. Areas indicated: CIRCG – circumglobal;
CIRC-TROP – circumtropical; EA – Eastern Atlantic; MED – Mediterranean; END – endemic; AMPHIA – amphiAtlantic;
MACR– Macaronesian; WIDE – wide distribution
100 TROP
3%
SUB-TROP
%) 80 TEMP-TROP 3%
y ( 15%
c
n
e 60
u
q
e
e fr 40
v
ati
el 20
R
TEMP
0 14% SUB
TEMP-SUB
25 125 225 325 425 525 625 725 825 925 1025 1125
14% 51%
Depth strata - midpoints (m)
Fig. 4. Species climate origin in each depth strata and in the total catch. Climates indicated: TROP – tropical; SUB – subtropical;
TEMP – temperate
100
P-BP
%) 80 EP-BP 6% P
y ( 17% 14%
c
n
e 60
u
q
e
r
e f 40
v
ati
el 20
R
0 EP BP
25 125 225 325 425 525 625 725 825 925 1025 1125 18% 45%
Depth strata - midpoints (m)
Fig. 5. Species major prey groups in each depth strata and in the total catch. Prey groups indicated: EP – epibenthic; BP –
benthopelagic; P – pelagic
occurrence (Appendix 1). Dominant species (Group A) average frequency of occurrence higher than 25%.
include only 3 species—Moramoro, Helicolenus dac- Frequent species (Group B) present moderate abun-
tylopterus dactylopterushereafter called Helicolenusd. dances and include 34 species with relative abundance
dactylopterus)and Pagellusbogaraveo—each account- values between 0.1 and 10% and with an average
ing for more than 10% of the total abundance, and an frequency of occurrence between 1 and 10%. Com-
Menezes et al.: Demersal fish assemblages off the Azores 247
mon species (Group C) includes 32 species with aver- Species depth distribution patterns
age relative abundance values between 0.1 and
0.01%. A fourth group of rare species (Group D) was In the shallowest depth stratum, many species have
also considered, and includes about 32 species with similar abundances, while in deep waters 1 or 2 spe-
relative abundances smaller than 0.01%. cies clearly dominate the catch (Fig. 6). The contribu-
In general, more abundant species have higher fre- tion of all the other species (‘Others’) comprises about
quency of occurrence and wider distributions. Dominant 30% in the first 4 depth strata and then strongly
species appeared in all 6 areas surveyed, frequent spe- decreases along a depth gradient. At intermediate
cies usually appeared in more than 3 of the 6 areas sur- depths, Helicolenusd.dactylopterusdominated, while
veyed (Appendix 1) and the common and rare species in the deepest strata Mora moro strongly dominated
occurred only in 1 or 2 of the areas surveyed. Some of the the catch. The clinalchange in the species composition
most abundant species, however, were restricted to spe- shows a clear vertical zonation over the surveyed
cific areas. Pagellusacarneand Diplodussarguswere depths (Fig. 7). At the very first depth strata the species
never caught in the seamount areas (Areas 1, 4 and 5), Diplodussarguscadenatiwas caught in larger propor-
while Pagruspagruswas never caught in the Princesa tions. The species Pagrus pagrus assumes its greatest
Alice bank (Area 1, Sub-area: PAL). The deep-water contribution in the interval 50 to 100 m. Below this
species Phycisblennoidesand Etmopterusspinaxalso shallower strata, the species that most contributed to
were never caught in Area 6 (Western island group; the total catch were Pagellus bogaraveo (between 50
Appendix 1). About 25 species were caught only once, and 200 m depth), H. d. dactylopterus (between 275
and the highest number of unique records was observed and 675 m depth) and M. moro (between 700 and
in the southern area (Area 6, S. Miguel andSanta Maria 1200 m depth). Shelf species occupy narrower depth
islands), with 10 unique records (Appendix 1). ranges than slope species, and most species have
Fig. 6.Total relative number
of the top 4 species in each
depth strata. The group
‘Others’ includes all other
species caught during the
Azores surveys
248 Mar Ecol Prog Ser 324: 241–260, 2006
Gymnothorax unicolor Ruvettus pretiosus
Pseudocaranx dentex Polyprion americanus
Muraena augusti Lophius piscatorios
Seriola dumerili Dipturus batis
Gnathophis mystax Beryx splendens
Diplodus sargus cadenati Heptranchias perlo
Scorpaena azorica Lepidorhombus whiffiagonis
Balistes carolinensis Torpedo nobiliana
Epinephelus marginatus Leucoraja fullonica
Pomatomus saltatrix Malacocephalus laevis
Sarda sarda Helicolenus d. dactylopterus
Bodianus scrofa Phycis blennoides
Scorpaena maderensis Gadella maraldi
Pagrus pagrus Caelorinchus caelorhincus
Boops boops Beryx decadactylus
Pteroplatytrygon violacea Dalatias licha
Bothus podas maderensis Etmopterus spinax
Serranus atricauda Molva macrophtalma
Enchelycore anatina Chlorophthalmus agassizii
Dasyatis pastinaca Centrolophus niger
Galeorhinus galeus Physiculus dalwigki
Scorpaena notata Benthodesmus elongatus simonyi
Pagellus acarne Pterycombus brama
Sphyrna zygaena Micromesistius poutassou
Sphyraena viridensis Dipturus oxyrinchus
Scomber japonicus Squaliolus laticaudus
Raja clavata Etmopterus pusillus
Anthias anthias Deania profundorum
Zeus faber Epigonus telescopus
Scorpaena scrofa Mora moro
Synodus saurus Centrophorus squamosus
Serranus cabrilla Somniosus rostratus
Sphoeroides pachygaster Lepidion eques
Muraena helena Lepidion guentheri
Labrus mixtus Synaphobranchus kaupi
Aspitrigla cuculus Deania cf. calcea
Prionace glauca
Centrophorus granulosus
Phycis phycis
Galeus murinus
Acantholabrus palloni
Aphanopus carbo
Aulopus filamentosus
Centroscymnus crepidater
Macroramphosus scolopax
Pseudoscopelus altipinnis
Trachurus picturatus
Etmopterus princeps
Ariosoma balearicum
Alepocephalus rostratus
Zenopsis conchifer
Centroscymnus coelolepis
Capros aper
Centroscymnus cryptacanthus
Pontinus kuhlii
Trachyscorpia cristulata echinata
Lepidopus caudatus
Chiasmodon niger
Xiphias gladius
0 0 0 0 0 0 0
Pagellus bogaraveo 0 0 0 0 0 0 0
1 3 5 7 9 1 3
1 1
Schedophilus ovalis
Paraconger macrops Soundings (m)
Conger conger
0 0 0 0 0 0 0
0 0 0 0 0 0
1 2 3 4 5 6
Soundings (m)
Fig. 7. Individual bathymetric distribution of all fish species caught. Median-weighted depth distribution (±80% percentile)
obtained from the Azores cruise surveys. Median values are weighted by species-abundance data. The 10 most abundant species
are represented by larger circles
Menezes et al.: Demersal fish assemblages off the Azores 249
Biomass Abundance
skewed depth distributions (Fig. 7). For those species 80 40 0 40 80
occurring in deeper strata, the maximum depths of 25
their distributions are not necessarily the observed val-
ues, due to the maximum survey depth (1200 m).
125
The depth profiles of all species abundance and bio-
mass indices (Fig. 8) follow very similar distribution
225
patterns and have significant decreasing trends with
increasing depth (abundance: r = –0.87, p < 0.001; bio-
s
325
mass: r = –0.48, p < 0.05). The highest abundance val-
s
ues were observed in the 25 to 600 m interval, followed m)
by a decrease in the 600 to 925 m interval, and a small nt ( 425
peak at the 950 to 1000 m depth stratum. The relative oi
p
values of biomass below 950 m depth were relatively d 525
mi
larger than those of abundance, reflecting an increase -
a
in the mean individual weight of the deeper fish at 625
species caught. str
h
pt 725
e
D
Fish assemblage structure and zonation
825
The TWINSPAN classification analysis primarily
925
classified samples according to depth (Table 1, Fig. 9).
The first dichotomy level separates all samples shal-
lower than 500 m (Groups S13 to IS20) from other 1025
deeper samples (Groups ID21 to D28). At the second
dichotomy level, the shelf and shelf-break samples 1125
(<250 m depth, Groups S13 to S16) are separated from
the upper-slope samples (250 to 650 m depth, Groups kg/1000 hooks n/1000 hooks
IS17 to IS20). At the same dichotomy level samples of
Fig. 8. Total median species abundance (no. per 1000 hooks)
the mid-slope zone (between 650 to 800 m, Groups and biomass (kg per 1000 hooks) indices by depth strata, from
ID21 to ID24) are separated from the deep mid-slope the Azores longline cruise surveys (1995 to 2000)
samples below the 800 m depth (Groups D25 to D28).
Three of the 4 main assemblages are well delineated temperature) that best matches the biotic ordination
along the depth gradient. The deep intermediate structure shows a correlation value of 0.906. The aver-
assemblage (mid-slope ID groups) is less distinct and age similarity between all samples within each main
largely overlaid with the upper-slope and deep mid- assemblage shows an increasing trend from the shal-
slope groups, with a peak of observations
at 700 to 800 m depth (Fig. 9).
Shelf / Shelf-break (S)
The MDS ordination analysis (Fig. 10) 100
Shelf-break / Upper-slope (IS)
revealed a large-scale structure corre-
sponding to that found in the classifica-
tion analysis. The clear separation of the %) 80
main assemblages displayed in the hori- y (
c
zontal gradient represented in the first n 60
e
axis of the MDS plot is by far the most im- qu Mid-slope (ID)
e
p(noortt asnhto. wRne)s uclotsn foirf mth teh BatI Od-eEpNthV a annda tleymsis- e fr 40 Deep mid-slope (D)
v
perature variables explain well the main ati
biotic assemblage ordination (along the Rel 20
MDS horizontal axis), both with Spear-
man rank correlations of about 0.90. The
0
Spearman rank correlations obtained for
25 225 425 625 825 1025 1225
the stratum size, bottom irregularity and Depth strata - midpoints (m)
steepness vary between 0.33 and 0.16.
Fig. 9. Relative frequency of samples within the main 4 assemblages at the
The abiotic variable combination (depth/ 2nd TWINSPAN dichotomy level by depth strata
250 Mar Ecol Prog Ser 324: 241–260, 2006
d catchsample D28 11 21 21 810132.510.7156 3.77 8.58 48.535.66
ecies. Values shown represent average standaramples per area, species diversity, evenness, diate shallow; ID: intermediate deep; D: deep INTERMEDIATE / DEEP Deep mid-slopeMid-slope ID22ID23ID24D25D26D27 3114183112128221447111242552613734511832642742216271108103322856494582180077089797510232.231.131.631.222.462.330.760.670.890.790.800.7851819435149 11.4131.538.13 9.8023.577.6112.0111.0816.396.233.3211.426.7615.519.026.3112.6816.6710.113.9510.9611.3613.1642.5533.0551.6136.6139.177.1612.89
×29 sper of snterme ID21 792 566129476862.770.7977 1.49 5.27 50.554.167.7516.435.1312.804.968.774.4032.811.49
mples numbw; IS: i IS20 3041151835141316931945081.800.7862 5.267.14 6.9016.627.4210.557.573.585.437.7258.898.1216.928.225.857.315.656.223.6711.67
on analysis from an initial survey data matrix of 1384 sah group at the 4th dichotomy classification level. The average depth per site group are also indicated. S: shallo SHALLOW / INTERMEDIATE Upper-slopeInner Shelf / Outer shelf / Shelf-break S13S14S15S16IS17IS18IS19 30140915172133177209214841351410911725142419191381713161024115186972476204201314121461786361016426142585511518514017113898771102182393594195193.953.553.431.712.602.462.980.740.760.770.820.780.780.7673629351777994 11.502.3712.346.2422.372.286.679.5711.554.278.5713.573.1114.123.692.782.802.843.4120.2810.6119.654.1017.241.592.992.774.721.494.082.614.695.555.468.045.023.442.021.815.2912.797.099.4211.0510.799.012.633.9740.2722.4318.366.523.182.295.268.131.6914.716.4015.0512.0118.7412.042.505.496.286.428.466.263.4018.7315.8623.3034.0529.7245.3221.255.6610.5810.4629.9321.017.789.536.905.621.731.477.549.1030.8928.5510.7812.221.801.135.512.969.0423.493.0834.2831.6153.992.525.765.7115.815.0312.796.023.4415.5416.932.338.3327.036.276.558.364.6610.9811.112.53 16.3030.712.21
ble 1.Results of the TWINSPAN classificatispecies (fish per 1000 hooks) within eacaverage standard numbers, and the 1st Dichotomy 2nd DichotomyAssemblage levels3rd Dichotomy 4th DichotomyGroup identifier sAçores BankknPrincesa Alice BankaBMar-da-Prata BankSão Jorge islandr sub-areaTerceira islandsFaial and Pico islandsdnGraciosa islandsalSão Miguel islandsISanta Maria islandFlores and Corvo islandsTotal number of samplesAverage depth (m)Average Hill's diversity N1Average Hill's equitability E5Average standard catch (fish/hook *1000) Serranus atricaudaTAXADiplodus sargus cadenatiPagellus acarneGaleorhinus galeusPagrus pagrusZeus faberAspitrigla cuculusRaja clavataPhycis phycisTrachurus picturatusMuraena helenaPontinus kuhliiConger congerPagellus bogaraveoBeryx splendensPolyprion americanusLepidopus caudatusMalacocephalus laevisHelicolenus dactylopterus dactylopterusPhycis blennoidesEtmopterus spinaxBeryx decadactylusMolva macrophtalmaCaelorhinchus caelorhinchusEtmopterus pusillusDeania profundorumSynaphobranchus kaupiMora moroDeania cf. calcea
Tafor s pe
e
pl
m
a
S