Geographic patterns in the demersal ichthyofauna of the Aleutian Islands E. A. LOGERWELL, 1, * K. AYDIN, 1 S. BARBEAUX, 1 E. BROWN, 1 M. E. CONNERS, 1 S. LOWE, 1 J. W. ORR, 1 I. ORTIZ, 2 R. REUTER 1 AND P. SPENCER 1 1 Alaska Fisheries Science Centre, 7600 Sand Point Way, Seattle, WA 98115, USA 2 School of Aquatic and Fisheries Science, University of Washington, Seattle, WA 98195, USA ABSTRACT The goals of this research were to investigate geo- graphic patterns in the Aleutian Island region’s de- mersal ichthyofauna and to determine whether they reflected the physical and biological oceanographic patterns documented by other authors in this volume. The analyses were structured according to the level of organization: at the community level, patterns in species occurrence and community structure were investigated; at the population level, distribution and abundance were examined; at the individual level, food habits and growth were studied. There were step- changes in species occurrence, diversity, population distribution and food habits at Samalga Pass and at sites farther west. These longitudinal trends indicated physical and biological variation along the length of the Aleutian Islands chain; however, depth-related patterns were as common as longitudinal patterns in demersal fish distribution. In addition, high catches of patchily distributed species occurred in areas expected to be biological ‘hot spots’ because of increased pro- ductivity and prey availability. These patterns suggest linkages between demersal fish ecology and the bio- physical processes described by other authors in this volume and indicate that inter-disciplinary research is needed to elucidate the underlying mechanisms. Key words: Aleutian Islands, cluster analysis, demersal fish, distribution, food habits, geographic patterns, groundfish, growth, species diversity INTRODUCTION The Aleutian Islands region is a unique ecosystem. Formed by the peaks of the Aleutian ridge, it is the world’s only longitudinally oriented, high-latitude island archipelago. It is also unique in its great length (1800 km). Befitting a partially submerged mountain range, the shelf along the Aleutian Islands is narrow and the continental slope is steep. The primary cur- rents in the region are the Alaska Coastal Current and the Alaskan Stream to the south, and the Aleutian North Slope Current to the north. The Aleutian Passes are conduits through which the North Pacific and the Bering Sea interact through tidal currents and mixing. The result of the interaction between cur- rents, passes and bathymetry is a spatially and tem- porally complex ocean environment (Coyle, 2005; Ladd et al., 2005a; Mordy et al., 2005; Stabeno et al., 2005). Much of the island chain is undeveloped and provides important habitat for fish, seabirds (Byrd et al., 2005; Jahncke et al., 2005) and marine mammals (Call and Loughlin, 2005; Sinclair et al., 2005). Demersal fishes of the Aleutian Islands are important economically and ecologically. The com- mercial catch of demersal fish in 2003 in the Bering Sea/Aleutian Islands totaled 1.9 million metric tonnes worth an ex-vessel value of $481 million (Hiatt et al., 2004). Walleye pollock (Theragra chalcogramma) is the dominant species of the commercial catch in the Bering Sea/Aleutian Islands (63% of the total ex- vessel value in 2003), followed by Pacific cod (Gadus macrocephalus), flatfishes (Pleuronectidae), sablefish (Anoplopoma fimbria), Atka mackerel (Pleurogrammus monopterygius), and rockfishes (Sebastes and Sebastolo- bus spp.) (Hiatt et al., 2004). In addition to their commercial value, demersal fishes play important ecological roles as both predators and prey in the Aleutian Islands region. Piscivorous species include Pacific cod and arrowtooth flounder (Atheresthes sto- mias), which consume other demersal fishes (primarily Atka mackerel and walleye pollock) and myctophids (Yang, 2003). Planktivores include Atka mackerel, walleye pollock and several species of rockfishes (Yang, 2003). Demersal fishes are prey for upper *Correspondence. e-mail: [email protected]Received 20 October 2003 Revised version accepted 5 January 2005 FISHERIES OCEANOGRAPHY Fish. Oceanogr. 14 (Suppl. 1), 93–112, 2005 ȑ 2005 Blackwell Publishing Ltd. 93
Transcript
Geographic patterns in the demersal ichthyofauna of theAleutian Islands
E. A. LOGERWELL,1,* K. AYDIN,1
S. BARBEAUX,1 E. BROWN,1 M. E. CONNERS,1
S. LOWE,1 J. W. ORR,1 I. ORTIZ,2 R. REUTER1
AND P. SPENCER1
1Alaska Fisheries Science Centre, 7600 Sand Point Way, Seattle,WA 98115, USA2School of Aquatic and Fisheries Science, University of
Washington, Seattle, WA 98195, USA
ABSTRACT
The goals of this research were to investigate geo-graphic patterns in the Aleutian Island region’s de-mersal ichthyofauna and to determine whether theyreflected the physical and biological oceanographicpatterns documented by other authors in this volume.The analyses were structured according to the level oforganization: at the community level, patterns inspecies occurrence and community structure wereinvestigated; at the population level, distribution andabundance were examined; at the individual level,food habits and growth were studied. There were step-changes in species occurrence, diversity, populationdistribution and food habits at Samalga Pass and atsites farther west. These longitudinal trends indicatedphysical and biological variation along the length ofthe Aleutian Islands chain; however, depth-relatedpatterns were as common as longitudinal patterns indemersal fish distribution. In addition, high catches ofpatchily distributed species occurred in areas expectedto be biological ‘hot spots’ because of increased pro-ductivity and prey availability. These patterns suggestlinkages between demersal fish ecology and the bio-physical processes described by other authors in thisvolume and indicate that inter-disciplinary research isneeded to elucidate the underlying mechanisms.
The Aleutian Islands region is a unique ecosystem.Formed by the peaks of the Aleutian ridge, it is theworld’s only longitudinally oriented, high-latitudeisland archipelago. It is also unique in its great length(1800 km). Befitting a partially submerged mountainrange, the shelf along the Aleutian Islands is narrowand the continental slope is steep. The primary cur-rents in the region are the Alaska Coastal Current andthe Alaskan Stream to the south, and the AleutianNorth Slope Current to the north. The AleutianPasses are conduits through which the North Pacificand the Bering Sea interact through tidal currents andmixing. The result of the interaction between cur-rents, passes and bathymetry is a spatially and tem-porally complex ocean environment (Coyle, 2005;Ladd et al., 2005a; Mordy et al., 2005; Stabeno et al.,2005). Much of the island chain is undeveloped andprovides important habitat for fish, seabirds (Byrdet al., 2005; Jahncke et al., 2005) and marinemammals (Call and Loughlin, 2005; Sinclair et al.,2005).
Demersal fishes of the Aleutian Islands areimportant economically and ecologically. The com-mercial catch of demersal fish in 2003 in the BeringSea/Aleutian Islands totaled 1.9 million metric tonnesworth an ex-vessel value of $481 million (Hiatt et al.,2004). Walleye pollock (Theragra chalcogramma) is thedominant species of the commercial catch in theBering Sea/Aleutian Islands (63% of the total ex-vessel value in 2003), followed by Pacific cod (Gadusmacrocephalus), flatfishes (Pleuronectidae), sablefish(Anoplopoma fimbria), Atka mackerel (Pleurogrammusmonopterygius), and rockfishes (Sebastes and Sebastolo-bus spp.) (Hiatt et al., 2004). In addition to theircommercial value, demersal fishes play importantecological roles as both predators and prey in theAleutian Islands region. Piscivorous species includePacific cod and arrowtooth flounder (Atheresthes sto-mias), which consume other demersal fishes (primarilyAtka mackerel and walleye pollock) and myctophids(Yang, 2003). Planktivores include Atka mackerel,walleye pollock and several species of rockfishes(Yang, 2003). Demersal fishes are prey for upper
trophic level predators such as marine mammals. Forexample, walleye pollock and Atka mackerel are themost common prey of Steller sea lions (Eumetopiasjubatus) in western Alaska, followed by salmonids andPacific cod (Sinclair and Zeppelin, 2002).
Geographic patterns in the demersal fish commu-nities of the Gulf of Alaska and the eastern Bering Seahave been well studied. For instance, geographic pat-terns in species occurrence and community structureof demersal fishes in the Gulf of Alaska have beendocumented (Mueter and Norcross, 1999, 2002).Relationships between fish distribution and environ-mental characteristics, such as depth, temperature,bathymetry and bottom substrate type, have beeninvestigated in both systems (OCSEAP, 1986; Baileyet al., 1999; Krieger and Ito, 1999; Wyllie-Echeverriaand Ohtani, 1999; McConnaughey and Smith, 2000;Mueter and Norcross, 2000; Abookire et al., 2001;Brodeur, 2001; Mueter and Norcross, 2002; Baileyet al., 2003; Duffy-Anderson et al., 2003). There hasalso been extensive study of the distribution, abun-dance, feeding and growth of juvenile fishes, primarilyage-0 walleye pollock, in the Gulf and Bering Sea(Mueter and Norcross, 1994; Brodeur et al., 2000;Wilson, 2000; Brodeur et al., 2002; Ciannelli et al.,2002). Finally, the food habits of Gulf of Alaska andeastern Bering Sea demersal fishes are well-documen-ted (Mito et al., 1999; Yang and Nelson, 2000; Langet al., 2003).
In contrast to other Alaska ecosystems, very little isknown about the ecology of the demersal fish com-munity of the Aleutian Islands region. The food habitsof demersal fishes have been summarized for the regionas a whole (Yang, 1999, 2003), and geographic trendsin Atka mackerel growth have been documented(Lowe et al., 1998). In contrast to the extensive lit-erature for the Gulf of Alaska and eastern Bering Sea,there are no published studies on geographic patternsin Aleutian demersal fish species occurrence, com-munity structure, distribution and abundance. Neitherhave spatial patterns in food habits within the Aleu-tian Islands been investigated.
The goals of this research were to investigate geo-graphic patterns in the demersal ichthyofauna of theAleutian Islands and to determine whether the pat-terns we found reflected those in the physical andbiological oceanography documented by other authorsin this volume (Coyle, 2005; Ladd et al., 2005a,b;Mordy et al., 2005; Stabeno et al., 2005). We struc-tured our inquiries according to the level of organ-ization from the community down to the individual.At the community level we investigated geographicpatterns in species occurrence and community
structure. At the population level, we examined fishdistribution and abundance. At the individual level, westudied geographic patterns in food habits and growth.
METHODS
Bottom trawl surveys
The Alaska Fisheries Science Centre (AFSC) of theNational Marine Fisheries Service (NMFS) conductedbottom trawl surveys to collect standardized data onthe distribution, abundance and biological conditionof Alaska groundfish stocks. The Aleutian Islandsbottom trawl surveys covered depths to 500 m alongthe north side of the island chain from Unimak Pass(165�W) westward to the Islands of Four Mountains/Samalga Pass (170�W) and on both sides of the chainfrom Islands of Four Mountains/Samalga Pass toStalemate Bank (170�E) (Harrison, 1993). The areasouth of the chain from Unimak Pass to the Islands ofFour Mountains/Samalga Pass was surveyed during theGulf of Alaska survey (Britt and Martin, 2001). TheGulf of Alaska survey extended east to DixonEntrance (132�40¢W), covering the continental shelfand upper continental slope to 1000-m depth. Surveysin the Aleutian Islands were conducted in 1980, 1983,1986, 1991, 1994, 1997, 2000, and 2002. Surveys inthe Gulf of Alaska were conducted in 1984, 1987,1990, 1993, 1996, 1999, 2001, and 2003. Surveys wereconducted during summer beginning as early as Mayand ending as late as September. Survey durationswere approximately 140 days.
A stratified random sampling design was employedduring both surveys (Harrison, 1993; Britt and Mar-tin, 2001). The Aleutian Islands survey region wasdivided by North Pacific Fishery ManagementCouncil (NPFMC) regulatory areas (Fig. 1) thatwere further divided into 45 area-depth strata to adepth of 500 m. The Gulf of Alaska survey regionwas divided into 59 strata categorized by water depth,type of geographical area and NPFMC regulatoryareas. The stratified random design greatly facilitatedthe collection of data on the distribution and abun-dance of groundfishes in the Aleutian Islands, but thedata are nonetheless restricted to trawlable areas suchthat there is limited information on the distributionand abundance of fishes in very rocky or steep hab-itat.
Data on catch weight by species were collectedfrom each haul. All fishes were identified to specieswhenever possible. Additional biological data werecollected on species selected because of theircommercial value or high abundance. A random
subsample of these fish was sorted by sex, and indi-vidual fork lengths (FL) and wet weights were meas-ured. Age structures (otoliths, scales, or both) werealso collected. Stomach samples from selected specieswere collected during the surveys.
Biomass estimates were calculated using the area-swept method (Alverson and Pereyra, 1969). Foreach species, catch-per-unit-effort (CPUE) was cal-culated for each tow by dividing catch weight (kg)by the area swept by the tow (km2). A mean CPUEfor each stratum was calculated as the mean of theindividual tow CPUE (including zero catches)within that stratum. Biomass estimates were calcu-lated by multiplying each stratum mean CPUE bythe stratum area (Harrison, 1993; Britt and Martin,2001).
Analyses described in this paper were restricted tosurvey data collected from 1990 to 2002 because mostsurvey effort prior to 1990 was conducted with non-standard survey gear and without rigorous fishing effort
measurements. Exceptions are the analyses of dietcomposition and growth. The diet composition ana-lysis utilized the entire data set because the non-standard trawl survey methodology was not expectedto affect groundfish stomach contents data. The dietcomposition analysis also utilized stomach samplesobtained by fishery observers during commercial fish-ing operations from 1982 to 1999. The growth analysiswas restricted to the 1997 and 2000 data, years duringwhich sufficient sample sizes of fish otoliths by subareawere collected.
Many of the results presented in this paper werebased on survey catch or biological data averaged overseveral years’ surveys, and thus interannual andseasonal patterns were not examined. The broadtemporal scale of our analyses also likely resulted inthe loss of information on fine-scale spatial patterns.This approach was appropriate because our goal was toexamine geographic trends in the Aleutian Islandsichthyofauna at a broad scale.
Amch
itka
Pass
Sam
alga
Pas
s
543
541
610542
514
550
620
524
521
523
517
513
518
509
530
512516508
0
0
0
0
0
0
519
0
0
0
0
00
0
0
0
0 0
0
0
0
176°E168°E160°E152°E148°E
180°W 176°W
176°W
172°W 168°W
168°W
164°W 160°W 156°W
156°W
38°N
42°N
46°N
46°N
50°N
50°N
54°N
54°N
58°N
58°N
62°N
62°N
66°N
70°N
74°N
160°E
152°E
Figure 1. Aleutian Islands with NorthPacific Fishery Management Councilregulatory areas referred to in text.
The distributions of demersal fish species in theAleutian Islands region were examined using pub-lished records (Allen and Smith, 1988; Sheiko andFederov, 2000; Mecklenburg et al., 2002) andvouchered records in the AFSC Aleutian Islandsand Gulf of Alaska bottom trawl survey database.Species with distributions that were continuousacross the Aleutian chain were eliminated, becausethese species would be less informative regarding theinfluence of Aleutian passes on demersal fish distri-butions at the community level. Many commerciallyimportant species, such as Pacific ocean perch (Se-bastes alutus), walleye pollock, Pacific cod and Atkamackerel were thus not included in this analysis, butwere considered individually in later sections of thispaper. Species that were not consistently identifiedor were inadequately sampled during AFSC bottomtrawl surveys because of habit or habitat (e.g. semi-demersal or deeper than 500 m) were also elimin-ated.
The Aleutian Islands area was divided into sixregions, defined by the following passes: Unimak,Samalga, Amukta, Tanaga, Amchitka and Buldir(the unnamed pass between Buldir and SemichiIslands; Fig. 2). Species that occurred in each regionwere then listed, and the total number of species ineach region was summed. All species were classifiedby biogeographic province (Allen and Smith, 1988)and considered to be members of the Aleutianprovince. Species were further grouped according totheir biogeographic affinities outside of the Aleutianprovince following the criteria of Allen and Smith
(1988). The goals of grouping species in this waywere to assess the broader scale distribution of speciesfound in the Aleutians and to gain some insight intopatterns of distribution at the evolutionary scale. Theprovinces of interest (outside the Aleutian province)were: Arctic-Kurile, which included species found inboth Arctic and Kurile provinces, ranging from northof the Bering Strait into the western Bering Sea;Kurile, species found in the western Pacific whoserange may include the western Bering Sea, KurileIslands and Commander Islands; and Oregonian,species found in Southeast Alaska and Canada(Fig. 2).
Community structure
Cluster analyses were conducted to identifyco-occurring groups of demersal fish species and theirgeographic and depth distributions. Analyses werecarried out at two different scales of resolution withregard to the numbers of species included: a commu-nity-wide resolution and a ‘rockfish-specific’ resolu-tion. The purpose of the community-wide analysis wasto identify broad-scale species associations and toexamine differences in distribution among clusters ofspecies. The original motivation for the ‘rockfish-spe-cific’ analysis was to assess the co-occurrence ofrockfish with commercially important species (Pacificcod, Atka mackerel and sablefish) so that the effects ofmanagement regulations to limit rockfish bycatchcould be better understood. Our purpose in presentingthe results here is to examine the geographic distri-bution of assemblages of species within which rockfishoccur.
Samalga Pass
Amchitka Pass
Unimak Pass
Amukta Pass
Tanaga Pass
"Buldir Pass"
Near Strait
Kamchatka Strait
140°E
140°E
160°E
160°E
180°E
180°E
160°W
160°W
140°W
140°W
40°N 40°N
60°N 60°N
Coastal zoogeographic provinces
Arctic
Oregonian
Aleutian
Kurile
Okhotsk
Oriental−Japan
Russia Alaska
Sea of Okhotsk
Bering Sea
Gulf of Alaska
Figure 2. Geographic locations men-tioned in text and zoogeographic prov-inces of the North Pacific (reproducedfrom Allen and Smith, 1988).
Abundance matrices were first assembled fromAFSC Aleutian Islands bottom trawl survey data. Thematrix for the community analysis was constructedfrom the computed mean biomass by strata for 43species that were selected based on criteria of goodcatchability by trawls and consistent identification bysurvey personnel. The community matrix was con-structed by strata to describe differences in the generalcommunity structure across the Aleutian Islands at agross scale. The matrix for the rockfish analysis wasbased on CPUE by haul for several rockfish species(plus Pacific cod, Atka mackerel and sablefish). Therockfish matrix was constructed by haul because theintent of this matrix was to examine rockfish com-munity structure at a finer scale and also to examinehow the three major commercial species in theAleutian Islands region associate with rockfish species.
Cluster analyses were then conducted on dissimi-larity matrices calculated from the two abundancematrices (Logen et al., 1980). The community clus-tering analysis employed an agglomerative clusteranalysis as per Van Tongeren (1995) using the S-Pluscomplete linkage clustering method (Insightful Cor-poration, 2001). Of the various cluster analysismethods, the complete linkage method is most usefulto highlight differences between clusters because it isbiased toward overestimation of differences betweenclusters. This method is appropriate for the commu-nity-level analysis because community structure mayhave been blurred due to averaging of species biomassover interannual variability. Initially (at step 0), eachstratum is considered as a separate cluster (InsightfulCorporation, 2001). The rest of the computationconsists of iteration of the following steps: first, mergethe two clusters with smallest between-cluster dissi-milarity (R), and second, compute the dissimilaritybetween the new cluster and all remaining clusters(Q). The between-cluster dissimilarity is defined as:
dðR;QÞ ¼ maxi2R;j2Q
dði; jÞ ð1Þ
where R and Q represent clusters of merged strata andwhere i and j are the strata within R and Q. Thus, thedissimilarity between two clusters d(Q,R) is given bythe maximum dissimilarity between any pair of stratad(i,j) of the clusters.
The rockfish analysis used the average linkageclustering method which is widely used in ecology(Van Tongeren, 1995). Average linkage clusteringuses the average dissimilarity measure distancesbetween all possible pairs of points within the twoclusters. Those clusters with the smallest average dis-tance between their points are then merged at each
step of the clustering analysis. This method wasselected for the rockfish community analysis becausethe goal of this analysis was to identify groups of haulsthat were on average most similar, as opposed to mostdifferent. In contrast to the complete linkage methodwhich groups species that are most dissimilar, theaverage linkage method groups species that are onaverage more similar to other members of that clusterthan to members of any other cluster. In the averagelinkage method, average dissimilarity d(r,s) is compu-ted as:
dðr; sÞ ¼ Trs
Nr �Nsð2Þ
where Trs is the sum of all pairwise distances betweencluster r and cluster s. Nr and Ns are the sizes of theclusters r and s, respectively. At each stage of hierar-chical clustering, the clusters r and s, for which d(r,s) isthe minimum, are merged.
Distribution and abundance
The distribution and abundance of four major com-mercial groundfish species in the Bering Sea andAleutian Islands were examined: walleye pollock,Pacific cod, Atka mackerel, and Pacific ocean perch.The survey area was broken into intervals of 1/4-degreelongitude and 100-m depth ranges. The number ofhauls in each interval and the mean CPUE for thosehauls was computed. These mean CPUEs were used asan indicator of fine-scale spatial patterns in abun-dance. This analysis pools data over several differentyears and does not account for interannual variabilityin fish density or distribution; the goal was to identifyspatial patterns that persist over time. Because of thistemporal pooling, the CPUE figures should be con-sidered as relative measures only. CPUE data for allfour species, and especially for Atka mackerel andPacific ocean perch, were strongly skewed with a largeproportion of zeroes. New statistical methods are cur-rently being developed at AFSC that will improveabundance estimates from these highly skewed data,but they are not available for the present analysis.
Food habits
Feeding habits of common demersal fish species wereassessed from stomach collections during AFSC trawlsurveys and from stomachs collected by fisheryobservers during commercial fishing operations during1982–99 (see Yang, 2003, for stomach content analysismethods). The degree of taxonomic resolution varieddepending on prey type. Most fish prey were identifiedto species with the exception of myctophids,which were identified only to family (Myctophidae).
Invertebrate prey species were rarely identified tospecies. Euphausiids and copepods were identified toorder (Euphausiacea and Calanoida, respectively).Data for this analysis were pooled into two-degreelongitude areas, from 164�W to 170�E. Sample sizesper species for each two-degree block ranged from 11to 3588. Diet composition (percent by weight) by two-degree area was calculated for walleye pollock, Pacificcod, Atka mackerel, and Pacific ocean perch. Percentcomposition of major prey items (approximately>80%) was examined in detail, the remainingapproximately 20% was summarized as ‘other species’.
Growth
Length and age data from the 2000 AFSC bottomtrawl survey were used to examine geographic patternsin growth of northern rockfish (Sebastes polyspinis) andPacific ocean perch, the two Aleutian rockfish speciesfor which age data have been collected. Data werestratified by NPFMC regulatory areas (Fig. 1), definedas ‘eastern’ Aleutians (541), ‘central’ Aleutians (542)and ‘western’ Aleutians (543). Note that area 541 isactually west of Samalga Pass and thus corresponds towhat has been termed ‘central’ Aleutians by otherauthors in this volume and in other sections of thispaper. Area 610 is defined as ‘western Gulf of Alaska(GOA)’, but because this area lies east of Samalga Passit is analogous to what has been called ‘eastern’Aleutians in other sections of this paper.
Growth curves for northern rockfish and Pacificocean perch were based on estimated populationnumbers at length from the 1997 and 2000 surveysand the length-at-age data from the samples of agedfish. These data provide a basis for an analysis ofdifferences in length at age between areas. The oto-liths collected in each year and subarea wereobtained with length-stratified sampling, and unbi-ased estimates of mean length at age were producedby multiplying the estimated numbers at length bythe age–length key. The use of age–length keys re-quires length composition estimates from the sameyear as the otolith collection, and thus separateanalyses were conducted for the 1997 and 2000 data.The results presented here were produced from the2000 data; the 1997 data produced nearly identicalpatterns. Mean length at age was calculated,accounting for the length-stratified sampling in thesurvey as described above. The standard deviation ofmean length at age was obtained from the deltamethod (Dorn, 1992). Sample sizes for aged northernrockfish were: 199 fish in area 541, 275 in area 542,and 228 in area 543, whereas samples sizes for Pacificocean perch in these area were 428, 319, and 357,
respectively. The von Bertalanffy model was used toestimate growth curves:
Lt ¼ L1ð1 � eð�k�ðt�t0ÞÞÞ ð3Þwhere Lt is length at age t (in yr), L¥ is the meanasymptotic length, t0 is the theoretical age at which afish would have been zero length, and k is a constant.Growth curves for each species were estimated fromdata in areas 541, 542, 543 and compared with modelparameters (L¥, k and t0) estimated by P. Malecha andJ. Heifetz (unpublished report) from data collected inarea 610.
RESULTS
Species occurrence
Approximately 245 fish species were identified insurveys in the Aleutian Islands. Of these, only 63 metthe criteria for inclusion in the analysis of speciesoccurrence patterns relative to Aleutian passes. Ofthese 63 species considered, there was a large per-centage decline (28%) in the number of demersal fishspecies between Unimak/Samalga and Amukta Passes(Table 1). The number of these 63 species occurringin the region between Samalga and Amchitka Passesremained relatively constant, declining by only 4%.There was another decline in number of species fur-ther west, beyond Buldir Island (20%). Patterns in thedistribution of species with respect to their presumedgeographic province were also evident. Approximatelyone-third of the species with Oregonian affinities werenot found further west than Unimak Pass, anotherthird were not found further west than Samalga Pass(Table 1). In contrast to the Oregonian species,approximately 70% of Kurile species are found acrossthe entire Aleutian Islands chain.
Demersal community structure
The demersal community cluster analysis yielded fivespecies assemblages (Table 2). Two of the speciesassemblages were dominated by Atka mackerel andtended to be stratified by depth: the <100 m ‘ShallowAtka cluster’ and the 100–200 m ‘Deep Atka cluster’(Table 2 and Fig. 3). The primary difference in thespecies composition of these two clusters was theabundance of Pacific ocean perch and walleye pollock,which were proportionally more abundant in the deepAtka cluster than in the shallow Atka cluster. The‘Pacific ocean perch cluster’ and ‘deep cluster’ likewiseshowed unique depth distributions (Table 2 andFig. 3). The Pacific ocean perch cluster was made up ofstrata at depths ranging from 200 to 300 m and wasdominated by Pacific ocean perch and walleye pollock.
Table 1. Demersal fishes with partial distributions in the Aleutian Islands, classified by zoogeographic provinces (Allen andSmith, 1988). Passes listed are the easternmost boundaries of areas in which species occurrence was tabulated. ‘X’ indicatesrecords of this species occurring.
Taxon Common name Unimak Samalga Amukta Tanaga Amchitka Buldir
Arctic-Kurile ProvinceAnisarchus medius Stout eelblenny XLumpenus fabricii Slender eelblenny XAnarhichas orientalis Bering wolffish XOcella dodecahedron Bering poacher XPlatichthys quadrituberculatus Alaska plaice X XEleginus gracilis Saffron cod X XAspidophoroides monopterygius Alligatorfish X X X X XLeptoclinus maculatus Daubed shanny X X X X XLumpenus sagitta Snake prickleback X X X X X
Kurile ProvinceLumpenella longirostris Longsnout prickleback XIcelus canaliculatus Blacknose sculpin* X X X X X XBathyraja lindbergi Commander skate X X X X X XBathyraja minispinosa Smallthorn skate X X X X X XBathyraja maculata White-blotched skate X X X X X XSebastolobus macrochir Broadbanded thornyhead X X X X X XCareproctus ostentum Microdisk snailfish X X X X X XCareproctus simus Proboscis snailfish X X X X X XElassodiscus tremebundus Dimdisk snailfish X X X X X XCareproctus zachirus Blacktip snailfish X X X X XHemilepidotus zapus Longfin Irish lord X X X X XThyriscus anoplus Sponge sculpin X X X X XAllocareproctus jordani Jordan’s allocareproct X X X X XBathyraja violacea Okhotsk skate X X X X XBathyraja taranetzi Mud skate X X X X XSigmistes smithi Arched sculpin X X X XIcelus uncinalis Uncinate sculpin X XPercis japonica Dragon poacher X
Oregonian ProvinceSynchirus gilli Manacled sculpin XIcelus spatula Spatulate sculpin XArtedius lateralis Smoothhead sculpin XLeptocottus armatus Pacific staghorn sculpin XBathyagonus alascanus Gray starsnout XHypomesus pretiosus Surf smelt XRaja rhina Longnose skate XSebastolobus altivelis Longspine thornyhead XSebastes brevispinis Silvergray rockfish XLycodes brevipes Shortfin eelpout XAnarrhichthys ocellatus Wolf-eel XEopsetta jordani Petrale sole XArtedius fenestralis Padded sculpin X XArtedius harringtoni Scalyhead sculpin X XSebastes flavidus Yellowtail rockfish X XSebastes ruberrimus Yelloweye rockfish X XCryptacanthodes giganteus Giant wrymouth X XCryptacanthodes aleutensis Dwarf wrymouth X XRaja binoculata Big skate X XBathyraja interrupta Sandpaper skate X XMicrogadus proximus Pacific tomcod X XBathyagonus infraspinatus Spinycheek starsnout X X
Taxon Common name Unimak Samalga Amukta Tanaga Amchitka Buldir
Psettichthys melanostictus Pacific sand sole X XLepidopsetta bilineata Rock sole X XIsopsetta isolepis Butter sole X XPsychrolutes paradoxus Tadpole sculpin X X X XSebastes crameri Darkblotched rockfish X X X XThaleichthys pacificus Eulachon X X X X XSebastes babcocki Redbanded rockfish X X X X XSebastes melanops Black rockfish X X X X XSebastes variegatus Harlequin rockfish X X X X XSebastes proriger Redstripe rockfish X X X X XPoroclinus rothrocki Whitebarred prickleback X X X X XRonquilus jordani Northern ronquil X X X X XTriglops macellus Roughspine sculpin X X X X X XAnoplagonus inermis Smooth alligatorfish X X X X X X
Total number of species 54 43 29 29 28 19
Table 2. Mean proportion of the average cluster biomass for each species. Clusters are: shallow Atka cluster (SAC), deep Atkacluster (DAC), Pacific ocean perch cluster (POP), deep cluster (DEEP) and northeast shallow shelf cluster (NESSC). Allproportions >0.10 are in italics.
The deep cluster was made up of strata at depthsranging from 300 to 500 m and was dominated bygiant grenadier (Albatrossia pectoralis), Kamchatka
flounder (Atheresthes evermanni) and shortraker rock-fish (Sebastes borealis) (Table 2). The ‘northeast shal-low shelf cluster’ was the only assemblage that showed
Figure 3. Cluster analysis of the Aleutian Islands survey data using a dissimilarity matrix and agglomerative cluster analysis.Three-number codes in the cluster tree represent National Marine Fisheries Service bottom trawl survey strata.
a longitudinal pattern. The strata containing this spe-cies assemblage were found exclusively north of theAleutian chain and east of Adak Strait (Fig. 3). Thisassemblage was also the most diverse – no single speciesdominated the cluster; instead arrowtooth flounder,Pacific cod, Pacific halibut (Hippoglossus stenolepis), rocksoles (Lepidopsetta spp.), and walleye pollock each madeup 12 to 22% of the average cluster biomass (Table 2).
Rockfish community structure
The rockfish cluster analysis yielded six assemblages:(1) northern rockfish–Atka mackerel, (2) Pacific cod,(3) northern rockfish–small Pacific ocean perch, (4)rougheye rockfish (Sebastes aleutianus)–shortrakerrockfish, (5) large Pacific ocean perch and (6) miscel-laneous rockfish–Pacific cod (Table 3). Similar to thedemersal community cluster analysis, this analysisidentified a Pacific ocean perch cluster and a clusterdominated by northern rockfish and Atka mackerel.Rockfish species that clustered together had uniquehabitat preferences based on depth or location along theAleutian Islands chain. The rougheye–shortrakerrockfish, large Pacific ocean perch and miscellaneousrockfish–Pacific cod clusters were all found in relativelydeep waters of the outer shelf and upper slope (101–500 m). Within these deep waters, each species clustershowed a distinct depth preference. The rougheye–shortraker rockfish cluster was found in the deepestwaters, followed by large Pacific ocean perch and mis-
cellaneous rockfish–Pacific cod (Fig. 4). The mediandepth (and the 25th to 75th percentile) of the miscel-laneous rockfish–Pacific cod cluster was greater than themedian (and 25th to 75th percentile) of the northernrockfish–Atka mackerel, Pacific cod, and northernrockfish–small Pacific ocean perch clusters. These latterthree clusters were all found at shallower depths on themiddle shelf (51–150 m) and there was no evidence ofunique depth preference among the three groups.However, differences in longitudinal distribution wereevident. The northern–small Pacific ocean perch clusteroccurred predominantly east of Amchitka Pass and thenorthern rockfish–Atka mackerel cluster occurred pre-dominantly to the west of the pass (Fig. 5).
Distribution and abundance
The number of survey hauls used to calculate CPUE ineach 1/4-degree longitude interval is summarized inFig. 6a. Atka mackerel and Pacific ocean perch werebinned into 0, <100 and >100 kg ha)1 categories.Walleye pollock CPUE was binned into 0, <50 and>50 kg ha)1 categories; and Pacific cod CPUE into 0,<25 and >25 kg ha)1 categories. Atka mackerel cat-ches >100 kg ha)1 occurred almost exclusively in thewestern and central Aleutians, west of Samalga Pass(Fig. 6b). High walleye pollock and Pacific cod cat-ches (>50 and 25 kg ha)1, respectively) were observedeast and west of Samalga Pass (Fig. 6c,d). High cat-ches of walleye pollock (>50 kg ha)1) were found in
Table 3. Species assemblages as identified by hierarchical cluster analysis. Individual species are listed in rows, speciesassemblages in columns. The numbers in each cell represent the proportion of catch-per-unit-effort (CPUE) comprised of eachspecies. Values in italics represent species that dominate the composition of each assemblage (>10% of CPUE).
the 50–300 m depth strata east of Samalga Pass, and inthe 100–400 m depth strata west of Samalga Pass(Fig. 6c). Walleye pollock catches were almostexclusively <50 kg ha)1 west of Buldir Island. Atkamackerel and Pacific ocean perch catches>100 kg ha)1 occurred east and west of Buldir Island(Fig. 6b,e). Pacific cod catches >25 kg ha)1 likewiseoccurred east and west of Buldir Island.
Differences in the spatial and depth distributions offish independent of longitude were also found. Exceptin the far west, high catches of Atka mackerel(>100 kg ha)1) were always in the two shallowestdepth strata (50–200 m) (Fig. 6b). In contrast, highPacific ocean perch catches (>100 kg ha)1) occurredalmost exclusively within the 100–400 m depthintervals (Fig. 6e).
Atka mackerel and Pacific ocean perch were morepatchily distributed than walleye and Pacific cod.
Hauls accounting for 70 and 52% had zero catches ofAtka mackerel and Pacific ocean perch, respectively,whereas 32 and 45% of hauls had zero catches ofPacific cod and walleye pollock. On average, however,Atka mackerel and Pacific ocean perch catches weremore than twice as large as those for walleye pollockand Pacific cod. Average CPUE of Atka mackerel andPacific ocean perch was 116 and 114 kg ha)1, whileaverage CPUE of walleye pollock and Pacific cod was47.4 and 24.8 kg ha)1. Several localized areas withhigh Atka mackerel and Pacific ocean perch catcheswere evident. Large Atka mackerel catches(>100 kg ha)1) were found at Samalga, Seguam andTanaga Passes; and at Petrel Bank and Buldir andTahoma Reefs (Fig. 6b). High Pacific ocean perchcatches (>100 kg ha)1) occurred south of Amlia Is-land (near Seguam Pass), Petrel Bank, Buldir andTahoma Reefs, and Stalemate Bank (Fig. 6e).
Figure 4. Depth distributions of six spe-cies clusters found in the Aleutian Islandssurvey data (1991–2000). Horizontal linesin the middle of each box represent themedian depth, the boxes encompass the25th to the 75th percentile, whiskersrepresent the 5th and 95th percentileand the points are the minimum andmaximum values. Sample sizes are (N ¼number of hauls): northern rockfish–Atkamackerel (N ¼ 38), Pacific cod (N ¼504), northern rockfish–small Pacificocean perch (POP) (N ¼ 59), rough-eye–shortraker rockfish (N ¼ 191), largePOP (N ¼ 114), miscellaneous rockfish–Pacific cod (N ¼ 468).
100 m1000 m
Amchitk
a Pass
Northern - small POP
Northern - Atka mackerel
175°E173°E171°E
176°E
177°E
178°E
179°E
180°W
179°W
178°W
177°W
176°W
175°W
174°W 172°W47°N
48°N
49°N
50°N
51°N
51°N
52°N
52°N
53°N
53°N
54°N
55°N
56°N
57°N
Figure 5. Geographic distribution oftwo species clusters found in the Aleu-tian Islands survey data (1991–2000),northern rockfish–small Pacific oceanperch (POP) and northern rockfish–Atka mackerel.
Figure 6. Number of Alaska Fisheries Science Center Aleutian Island survey hauls and catch-per-unit effort (CPUE) of selecteddemersal fish species aggregated in cells of 1/4-degree longitude and 100-m depth intervals. The top panel of each figurerepresents data from the Bering Sea side of the island chain, the bottom represents the North Pacific side. The depth intervalsare shown on the left and ride of each figure, the longitudinal divisions are shown in the center (with the location of selectedpasses and islands). Note the different CPUE scale range for each figure. (a) Number of survey hauls, (b) Atka mackerel CPUE,(c) walleye pollock CPUE, (d) Pacific cod CPUE, (e) Pacific ocean perch CPUE.
The size range of fish sampled for food habits was14–55 cm FL for Atka mackerel, 13–124 cm FL forPacific cod, 6–58 cm FL for Pacific ocean perch and13–72 cm FL for walleye pollock. Longitudinal pat-terns in diet were apparent in the data. Euphausiidsmade up 50–90% of the diets of Pacific ocean perch,walleye pollock, and Atka mackerel east of SamalgaPass (from 164 to 168�W) (Fig. 7a–c). In contrast,euphausiids generally made up <50% of the diets ofthese fishes west of Samalga Pass. Copepods andmyctophids dominated the remaining portion of thediets to the west. The declining trend in the propor-tion of euphausiids in the diets of Pacific ocean perchand Atka mackerel continued from Samalga Pass tothe far western Aleutian Islands. The diet compositionof Pacific cod likewise showed a shift near SamalgaPass (Fig. 7d). East of 170�W, Pacific cod diets weredominated by walleye pollock, whereas west of 170�W,
Atka mackerel became an increasingly abundantcomponent of the diet (beginning at around 174�W).The remainder of Pacific cod diet west of Samalga Passwas dominated by shrimp, squid and other fishes, dif-ferent from the region east of the pass where these taxacomprised a small component of the diet.
In addition to the shift in diet composition atSamalga Pass, there appeared to be changes in dietfurther west. The proportion of myctophids in thediets of Pacific ocean perch increased dramaticallywest of 176�E, near Buldir Island (Fig. 7a). In contrast,the proportion of myctophids in the diet of walleyepollock decreased toward the west (Fig. 7b). Theremainder of walleye pollock diet to the west wascomposed of copepods, squid and other invertebrates.
Growth
Longitudinal trends in fish growth were observed. Fornorthern rockfish, collected in 2000, for a given age(above approximately age 3), fish length increased from
west to east among the three Aleutian Islands NPFMCregulatory areas (541, 542, 543) (Fig. 8a). Data collec-ted in 1997 (not shown) produced nearly identicalpatterns. In contrast to the results for northern rockfish,the geographic changes in the estimated growth curveswere less dramatic for Pacific ocean perch (Fig. 8b).
DISCUSSION
As revealed in several other studies published in thisvolume, Samalga Pass is a major biophysical transitionzone in the Aleutian Islands region. Surface waterswest of Samalga Pass are oceanic (cold, salty andnutrient-rich), whereas surface waters east of the passare coastal (warm, fresh and nutrient-poor; Ladd et al.,2005a; Mordy et al., 2005). This transition is the resultof Aleutian-wide current patterns (Ladd et al., 2005a)and increased advection from the Bering Sea, com-bined with greater depth of mixing in the passes westof Samalga (Ladd et al., 2005a). Despite the fact thatnutrient levels are higher west of Samalga Pass, pri-mary production is lower, perhaps due to mixing ofphytoplankton below the euphotic zone and ironlimitation (Mordy et al., 2005). The species composi-tion of the zooplankton community reflects the watermass differences east and west of Samalga Pass. Neriticcopepod and euphausiid species are most abundant tothe east, and oceanic copepod and euphausiid species
Figure 7. Diet composition (percent by weight) of fourmajor demersal fish species in the Aleutian Islands by 2�longitudinal blocks. Longitude labels refer to the western-most longitude of the block. The approximate locations ofselected passes and Buldir Island are also shown. UnimakPass is at 165�W, Samalga Pass is at 169�W, Amchitka Passis at 180�W and Buldir Island is at 176�E. (a) Pacific oceanperch, (b) walleye pollock, (c) Atka mackerel, (d) Pacificcod.
Figure 8. Estimated growth curves for (a) northern rockfishand (b) Pacific ocean perch by National Marine FisheriesService management area based on trawl survey data from2000. Length is fork length.
are most abundant to the west (Coyle, 2005; Coyleand Pinchuk, 2005).
Many characteristics of the demersal fish commu-nity likewise change at Samalga Pass. There is a strongdecline in the number of fish species from east to westof the pass. This decline may be a reflection of therelatively poor primary production in the central andwestern Aleutian Islands region (Mordy et al., 2005).In addition, the passes west of Samalga are relativelydeep and lack zones of upward advection and surfaceconvergence that are important for foraging seabirds(Ladd et al., 2005b) and possibly other top levelpredators such as demersal fishes.
The distribution of species changes at Samalga Pass.Species with Oregonian affinities are more numerouseast of the pass, whereas species with Kurile affinitiesare distributed throughout the Aleutian Islands region.Although the species included in these analyses have avariety of life histories, nearly all have pelagic larvae.Thus, currents are expected to influence their dispersalpatterns and range boundaries (Palumbi, 1994; Briggs,1995; Gaylord and Gaines, 2000). Samalga Pass is thewesternmost extent of the west-flowing Alaska CoastalCurrent, which may explain why species with Orego-nian affinities are found primarily east of the pass. Thepasses west of Samalga are also relatively deep andincreasingly wide, such that they may serve as addi-tional barriers to westerly dispersal for some species.That Kurile species are found throughout the AleutianIslands region suggests that species that cross the dis-tance between the Commander Islands to the far westand Attu Island (the westernmost island in theAleutian region) are generally able to disperse widely,even across deep and broad passes such as the Kam-chatka Strait, Near Strait and Amchitka Pass. Exam-ination of genetic differentiation among AleutianIsland endemic populations and among species couldalso lend insight into dispersal patterns. In addition,newly described fish species (Orr and Busby, 2001; Orr,2004; Stevenson et al., 2004) that appear to beendemic to the east-central Aleutian Islands suggestthat vicariant geological events have also affected thespeciation and distribution of species in the area.
In addition to changes in species richness andgeographic affinities, we observed changes in the dis-tribution and abundance of fish populations atSamalga Pass. For instance, Atka mackerel are moreabundant west of Samalga Pass. This trend is oppositeto that of primary productivity (Mordy et al., 2005),but instead could be a reflection of the relativeincrease of oceanic zooplankton west of Samalga.Euphausiids and calanoid copepods are the most fre-quent food item of Atka mackerel in the Aleutian
Islands region (Yang, 2003). However, Yang (2003)did not identify zooplankton to species, so it is notknown whether Atka mackerel prefer oceanic to ne-ritic species of euphausiids and copepods. Further de-tailed study of the zooplankton composition of Atkamackerel diets is necessary to test the hypothesis thatthe increase in Atka mackerel abundance west ofSamalga Pass is because of the increase in oceanicspecies in the zooplankton community.
Fisheries scientists have long recognized that Gulfof Alaska and Aleutian Islands Atka mackerel showsignificant differences in population size, distribution,recruitment patterns, and resilience to fishing (Loweet al., 2003). Despite genetic similarities, there doappear to be different environmental factors affectingthese two populations giving rise to phenotypic dif-ferences in morphological traits such as weight-at-age(Kimura and Ronholt, 1988) and meristic traits suchas number of fin rays, vertebrae and gill rakers (T.P.Levada, unpublished data). However, the preciseboundary of these two populations is not known, norare mechanisms for differentiation proposed. Our workshows that the transition between the Gulf of Alaskaand Aleutian Island populations likely occurs at ornear Samalga Pass and indicates that future work onthe mechanisms of population differentiation (genet-ics, growth, fecundity, mortality, etc.) should focus onthe Samalga Pass area.
Finally, we found that the diet of demersal fisheschanges at Samalga Pass. Similar to changes observedin northern fulmar diets (Jahncke et al., 2005), eup-hausiids are more abundant in the diets of walleyepollock, Atka mackerel and Pacific ocean perch eastof Samalga Pass. West of this pass, copepods andmyctophids make up a relatively large proportion ofthe diets of these fishes. Although both copepods andeuphausiids are found throughout the Aleutian Is-lands region, a detailed comparison of the distribu-tion of zooplankton across a shallow eastern pass(Akutan) and a deep central pass (Seguam) showedthat euphausiids are relatively more abundant atAkutan Pass and are concentrated near the bottom(Coyle, 2005). Euphausiids are less abundant, lessdensely aggregated and shallower at Seguam Pass. Wehypothesize that the dominance of euphausiids in thediets of demersal fishes east of Samalga Pass is be-cause of the effects of pass geometry on the distri-bution of euphausiids, with the result thateuphausiids are more available to demersal fishes inthe shallow passes to the east. Further study of thedistribution of euphausiids among Aleutian passesand of the distribution of foraging demersal fishes isneeded to confirm this hypothesis.
In addition to the significant transition zone atSamalga Pass, there may be additional transition zonesin the western Aleutians that have not yet beenidentified oceanographically. Buldir Island is an areawhere several demersal fish characteristics change. Atthe community level, the number of fish speciesdeclines from east to west of Buldir Island and at thepopulation level, walleye pollock survey catches arelower west of Buldir Island. The passes west of Buldirare some of the deepest in the Aleutian Islands region.They provide most of the transport from the NorthPacific to the Bering Sea but likely contribute fewnutrients to the euphotic zone (Stabeno et al., 2005).This decline in productive potential may contribute tothe decline in species richness, in general, and walleyepollock abundance, in particular.
Another potential transition zone occurs atAmchitka Pass. Two rockfish community clusters showdistinct geographic distributions, with the northernrockfish – Atka mackerel cluster occurring almostexclusively west of Amchitka Pass, and the northernrockfish – Pacific ocean perch cluster occurring east ofthe pass. Amchitka Pass appears to be an area wherecurrents diverge, perhaps resulting in different watermass properties from east to west. Modelling resultssuggest that waters crossing Amchitka Pass from theNorth Pacific into the Bering Sea turn east, formingthe westernmost extent of the Aleutian North SlopeCurrent. The model also shows that a portion of theBering Sea current flow from the west towards Am-chitka Pass is diverted north and west at Bower’sRidge, located immediately west of the pass (W. Ma-slowski, personal communication). Further physicaland biological oceanographic study of current patterns,water-column properties and primary production isneeded to determine if transition zones exist west ofthe transition zone already documented at SamalgaPass (Ladd et al., 2005a).
In addition to step changes at Samalga Pass, BuldirIsland and Amchitka Pass, there are longitudinaltrends in demersal fish characteristics that indicatecontinuous physical and biological variation along thelength of the Aleutian Islands chain. The most stri-king is the longitudinal trend in northern rockfishgrowth. Size at age declines from east to west. Growthparameters from the westernmost Gulf of Alaskamanagement area (P. Malecha and J. Heifetz, unpub-lished report) show that northern rockfish farther eastare comparable with the eastern Aleutians (Table 4).Similarly, estimated growth curves for Atka mackerelshow that fish are largest in the western Gulf of Alaskaand smallest in the western Aleutians (Lowe et al.,1998); Fig. 9). Our current understanding of northern
rockfish and Atka mackerel genetics indicates thatmechanisms at the evolutionary scale are not in-volved. An analysis of 37 protein-coding gene lociprovide no support for Atka mackerel genetic stockstructure within the Aleutian Islands region (Loweet al., 1998), although further genetic studiesemploying microsatellite DNA are currently beingpursued at AFSC. Geographic growth differences thatexist despite the lack of genetic differentiation may bea result of a single stock that mixes during the pelagiclarval and juvenile stages but then aggregates duringthe demersal adult stage (Lowe et al., 1998). Thecombination of the westward-flowing Alaskan Stream,the eastward flowing Aleutian North Slope Currentand northward flow through the passes (Ladd et al.,2005a) would likely facilitate this mixing of pelagiclife-history stages. For this hypothesis to be justified,growth patterns would need to be established in theadult stage after fish have settled in a particular region.A preliminary study of northern rockfish geneticssimilarly showed no evidence of population structureamong samples collected at Kodiak Island (Gulf
Table 4. Parameters from the von Bertalanffy growthequation for northern rockfish in the three Aleutian IslandsNational Marine Fisheries Service management areas andthe western Gulf of Alaska management area (P. Malechaand J. Heifetz, unpublished report).
Management area L¥ (cm) k t0 (yr)
West (543) 33.28 0.18 )0.49Central (542) 35.48 0.16 )0.52East (541) 40.48 0.18 0.26W. Gulf of Alaska (610) 39.16 0.17 )0.64
Figure 9. Estimated growth curves for Atka mackerel byNational Marine Fisheries Service management area basedon trawl survey data from the Aleutian Islands and Gulf ofAlaska during 1993 and 1994, respectively (Lowe et al.,1998). Length is fork length.
of Alaska), Unimak Pass (eastern Aleutians) andStalemate Bank (western Aleutians) (A. Gharrett,University of Alaska Fairbanks, USA, personal com-munication). However, the sample sizes were small,and much larger sample sizes will be required toidentify possible subtle changes in genetic structure.
Atka mackerel and northern rockfish growth dif-ferences may not be the result of genetically isolatedstocks but may instead represent phenotypic expres-sion of environmental differences among areas withinthe Aleutian Islands region. The environmental fac-tors that determine growth of these fishes are notknown, but the decline in growth rate from east towest could be related to the decline in primary pro-ductivity toward the west (Mordy et al., 2005).Understanding growth patterns is important because,for practical fisheries management, populationparameters such as mortality, spawning behaviour andgrowth are particularly useful for recognition of stocks(Casselman et al., 1981; Ihssen et al., 1981). Thegrowth differences among regions within the AleutianIslands suggests that in addition to large-scale pheno-typic differentiation between Aleutian Islands andGulf of Alaska fishes (Lowe et al., 2003), there alsomay be important differences occurring at a smallerscale within the Aleutians.
In contrast to the results for northern rockfish andAtka mackerel, there is little geographic pattern in theestimated growth curves for Pacific ocean perch. Infor-mation on the early life history of Atka mackerel, nor-thern rockfish and Pacific ocean perch is sparse, exceptthat each has pelagic larval and early juvenile stages ofunknown duration. The difference in growth patterns ofthe three species would be consistent with a longerpelagic stage for Pacific ocean perch, such that growthpatterns are established before fish are distributed intoseparate spawning locations. In contrast, a relativelyshort pelagic stage for Atka mackerel and northernrockfish could result in less mixing of sub-adults andstronger geographic differentiation of growth patterns.Further information on the early life history of thesethree fish species is needed to evaluate this hypothesis.
In addition to biophysical transition zones andlongitudinal trends, analysis of demersal fish distribu-tions shows that some species are more patchily dis-tributed than others, and that high catches of thesepatchily distributed species occur in areas expected tobe biological ‘hot spots’ because of increased produc-tivity and prey availability. ‘Hot spots’ (areas ofhydrographically generated prey aggregations) areimportant for predators in many other systems (Olsonand Backus, 1985; Fiedler and Bernard, 1987;Schneider, 1990; Lang et al., 2000; Brodeur, 2001). In
the Aleutian Islands, Atka mackerel and Pacific oceanperch occur at exceptionally high densities at passessuch as Samalga, Seguam and Tanaga. These passes areof intermediate depth (120–200 m) and are efficient atmixing nutrients upwards because their sills are deeperthan the nutricline but shallow enough so that strongtidal currents can mix the water column vertically andintroduce nutrients into the euphotic zone (Stabenoet al., 2005). The interaction of currents with bathy-metry also creates upward advection and surface con-vergence that aggregate prey and are expected tofacilitate foraging by seabirds and other predators(Ladd et al., 2005b). Atka mackerel and Pacific oceanperch are also abundant at banks and reefs such asPetrel Bank, Buldir Reef and Tahoma Reef. Theshallowing of the water in these areas similarly mightbe expected to result in mixing of nutrients into theeuphotic zone, increased production and prey aggre-gation. Aleutian Island passes are important foragingareas for other predators, such as cetaceans and Stellersea lions (Sinclair et al., 2005).
Another important feature of the Aleutian ecosys-tem is change in demersal fish habitat with depth. Wefind that depth-related patterns are as common as lon-gitudinal patterns in fish distribution. For instance,cluster analyses of the demersal community in general,and the rockfish community in particular, show distinctdepth preferences for groups of species. Atka mackereland northern rockfish are distributed at relatively shal-low depths on the shelf, whereas pollock and Pacificocean perch distribution extends to the outer shelf.Giant grenadier, Kamchatka flounder, shortraker rock-fish and associated species are distributed over the outershelf and slope. Mueter and Norcross (2002) and Bro-deur (2001) report similar depth ranges for rougheyerockfish, shortraker rockfish, Pacific ocean perch andgiant grenadier in the Gulf of Alaska and Bering Sea.
Fish species that cluster together have similar habitatpreferences (Auster et al., 2001; Williams and Ralston,2002); however, very little is known about habitatrequirements of Aleutian Islands fishes. Studies ofSebastes rockfishes in other systems indicate a depend-ence on hard bottom substrates with high vertical relief(Pearcy et al., 1989; O’Connell and Carlile, 1993;Krieger and Ito, 1999), with the possible exception ofPacific ocean perch, which have been found to inhabitflat, pebble substrate in Southeast Alaska (Krieger,1992). Consistent with these studies, the shallowerwaters in the Aleutian Islands region where northernrockfish and Atka mackerel are found are thought to becharacterized by especially rocky bottom.
These patterns in demersal fish ecology are newfindings for the Aleutian Islands region. We find step-
changes at Samalga Pass and sites farther west,patchily distributed species at biological ‘hot spots’,longitudinal trends and depth-related variation. Ininterpreting these patterns, we suggest linkagesbetween demersal fish ecology and the biophysicalprocesses described by other authors in this volume.Because the mechanisms resulting in these linkages arelargely unknown, we recommend inter-disciplinaryresearch combining observations of the physical andbiological oceanographic properties of the water col-umn, bottom habitat, and demersal fish ecology (spe-cies occurrence, community structure, distribution andabundance, food habits, and growth).
ACKNOWLEDGEMENTS
We would like to thank S. McDermott, H. Zengerand G. Duker of AFSC and two anonymousreviewers for helpful reviews of the manuscript. Wewould also like to acknowledge the invaluable workof the vessel captains, crew and scientific personnelwho were responsible for the success of the AFSCbottom trawl surveys. Finally, thanks to G. Kruse(guest editor) and A. Macklin for coordinating thisspecial issue.
REFERENCES
Abookire, A., Piatt, J. and Norcross, B. (2001) Juvenilegroundfish habitat in Kachemak Bay, Alaska, during latesummer. Alaska Fish. Res. Bull. 8:45–56.
Allen, M. and Smith, G. (1988) Atlas and zoogeography ofcommon fishes in the Bering Sea and northeastern Pacific.NOAA Tech. Rep. NMFS 66:151pp.
Alverson, D. and Pereyra, W. (1969) Demersal fish exploration inthe northeastern Pacific Ocean – an evaluation of exploratoryfishing methods and analytical approaches to stock size andyield forecasts. J. Fish. Res. Bd. Can. 26:1985–2001.
Auster, P., Joy, K. and Valentine, P. (2001) Fish species andcommunity distributions as proxies for seafloor habitat dis-tributions: the Stellwagen Bank National Marine Sanctuaryexample. Environ. Biol. Fish. 60:331–346.
Bailey, K., Powers, D., Quattro, J. et al. (1999) Populationecology and structural dynamics of walleye pollock (Theragrachalcogramma). In: Dynamics of the Bering Sea. T. Loughlin &K. Ohtani (eds) Fairbanks, AK: University of Alaska SeaGrant, pp. 581–614.
Bailey, K., Brown, E. and Duffy-Anderson, J. (2003) Aspects ofdistribution, transport and recruitment of Alaska plaice(Pleuronectes quadrituberculatus) in the Gulf of Alaska andeastern Bering sea: comparison of marginal and centralpopulations. J. Sea Res. 50:87–95.
Briggs, J. (1995) Global Biogeography. Amsterdam: ElsevierScience BV.
Britt, L. and Martin, M. (2001) Data report: 1999 Gulf of Alaskabottom trawl survey. U.S. Dep. Commer. NOAA Tech.Memo. NMFS-AFSC-121: 249pp.
Brodeur, R. (2001) Habitat-specific distribution of Pacific oceanperch (Sebastes alutus) in Pribilof Canyon, Bering Sea. Cont.Shelf Res. 21:207–224.
Brodeur, R., Wilson, M. and Ciannelli, L. (2000) Spatial andtemporal variability in feeding and condition of age-0 wall-eye pollock (Theragra chalcogramma) in frontal regions of theBering Sea. ICES J. Mar. Sci. 57:256–264.
Brodeur, R., Wilson, M., Ciannelli, L., Doyle, M. and Napp, J.(2002) Interannual and regional variability in distributionand ecology of juvenile pollock and their prey in frontalstructures of the Bering Sea. Deep-Sea Res. II 49:6051–6067.
Byrd, G., Renner, H. and Renner, M. (2005) Distribution pat-terns and population trends of breeding seabirds in theAleutian Islands. Fish. Oceanogr. 14(Suppl. 1):139–159.
Call, K. and Loughlin, T. (2005) An ecological classification ofAlaska Steller sea lion (Eumetopias jubatus). Fish. Oceanogr.14(Suppl. 1):212–222.
Casselman, J., Collins, J., Crossman, E., Ihssen, P. andSpangler, G. (1981) Lake whitefish (Coregonus clupeaformis)stocks of the Ontario waters of Lake Huron. Can. J. Fish.Aquat. Sci. 38:1772–1789.
Ciannelli, L., Brodeur, R., Swartzman, G. and Salo, S. (2002)Physical and biological factors influencing the spatial dis-tribution of age-0 walleye pollock (Theragra chalcogramma)around the Pribilof Islands, Bering Sea. Deep-Sea Res. II49:6109–6126.
Coyle, K. (2005) Zooplankton distribution, abundance andbiomass relative to water masses in eastern and centralAleutian Island passes. Fish. Oceanogr. 14(Suppl. 1):77–92.
Coyle, R. and Pinchuk, A. (2005) Seasonal cross-shelf distri-bution of major zooplankton taxa on the northern Gulf ofAlaska shelf relative to water mass properties, species depthpreferences and vertical migration behaviour. Deep-Sea Res.II 52:217–245.
Dorn, M.W. (1992) Detecting environmental covariates ofPacific whiting Merluccius productus growth using a growth-increment model. Fish. Bull. 90:260–275.
Duffy-Anderson, J., Cianelli, L., Honkalehto, T. et al. (2003)Distribution of age-1 and age-2 walleye pollock in the Gulfof Alaska and eastern Bering Sea: sources of variation andimplications for higher trophic levels. In: The Big Fish Bang.Proceedings of the 26th Annual Larval Fish Conference.H. Browman & A. Skiftesvik (eds) Bergen, Norway: Instituteof Marine Research, pp. 381–394.
Fiedler, P. and Bernard, H. (1987) Tuna aggregation and feedingnear fronts observed in satellite imagery. Cont. Shelf Res.7:871–881.
Gaylord, B. and Gaines, S. (2000) Temperature of transport?Range limits in marine species mediated solely by flow. Am.Nat. 155:769–789.
Harrison, R. (1993) Data report: 1991 bottom trawl survey ofthe Aleutian Islands area. U.S. Dep. Commer. NOAA Tech.Memo. NMFS-AFSC-12:144pp.
Hiatt, T., Felthoven, R., Seung, C. and Terry, J. (2004) StockAssessment and Fishery Evaluation Report for the GroundfishFisheries of the Gulf of Alaska and Bering Sea/Aleutian IslandArea: Economic Status of the Groundfish Fisheries off Alaska,2003. Anchorage, AK: North Pacific Fishery ManagementCouncil, 144pp.
Ihssen, P., Booke, H., Casselman, J., McGlade, J., Payne, N. andUtter, F. (1981) Stock identification: Materials and methods.Can. J. Fish. Aquat. Sci. 38:1838–1855.
Insightful Corporation (2001) S-PLUS 6 for Windows Guide toStatistics, Vol. 2. Seattle, WA: Insightful Corporation.
Jahncke, J., Coyle, K. and Hunt, G.L. Jr (2005) Seabird distri-bution, abundance and diets in the central and easternAleutian Islands. Fish. Oceanogr. 14(Suppl. 1):160–177.
Kimura, D.K. and Ronholt, L.L. (1988) Atka mackerel. In:Condition of Groundfish Resources of the Eastern Bering Sea andAleutian Islands region in 1987. R.G. Bakkala (ed.) US. Dep.Commer., NOAA Tech. Memo. NMFS F/NWC-139, pp.141–171.
Krieger, K. (1992) Distribution and abundance of rockfishdetermined from a submersible and by bottom trawling. Fish.Bull. 91:87–96.
Krieger, K. and Ito, D. (1999) Distribution and abundance ofshortraker rockfish, Sebastes borealis, and rougheye rockfish,S. aleutianus, determined from a manned submersible. Fish.Bull. 97:264–272.
Ladd, C., Hunt, G.L. Jr, Mordy, C., Salo, S. and Stabeno, P.(2005a) Marine environment of the eastern and centralAleutian Islands. Fish. Oceanogr. 14(Suppl. 1):22–38.
Ladd, C., Jahncke, J., Hunt, G.L. , Coyle, K. and Stabeno, P.(2005b) Hydrographic features and seabird foraging inAleutian passes. Fish. Oceanogr. 14(Suppl. 1):178–195.
Lang, G., Brodeur, R., Napp, J. and Schabetsberger, R. (2000)Variation in groundfish predation on juvenile walleye pol-lock relative to hydrographic structure near the Pribilof Is-lands, Alaska. ICES J. Mar. Sci. 57:265–271.
Lang, G., Derrah, C. and Livingston, P. (2003) Groundfish foodhabits and predation on commercially important prey speciesin the eastern Bering Sea from 1993 through 1996. AFSCProcessed Rep. 2003–04:352pp.
Logan, D., Moy, S., Norris, R. and Sydor, W. (1980) Metho-dology for assessing population and ecosystem level effectsrelated to intake of cooling waters. Volume 2. Handbook ofMethods Community Analysis Techniques, One Blue Hill PlazaPearl River New York: Lawller, Matursky and Skelly Envi-ronmental Science and Engineering Consultants, 277 pp.
Lowe, S., Van Doornik, D. and Winans, G. (1998) Geographicvariation in genetic and growth patterns of Atka mackerel,Pleurogrammus monopterygius (Hexagrammidea), the Aleu-tian archipelago. Fish. Bull. 96:502–515.
Lowe, S., Ianelli, J., Zenger, H. and Lauth, R. (2003) Stockassessment of Aleutian Islands Atka mackerel. In: StockAssessment and Fishery Evaluation Report for the GroundfishResources of the Bering Sea/Aleutian Islands Regions. The PlanTeam for the Groundfish Fisheries of the Bering Sea andAleutian Islands (eds) Anchorage, AK: North Pacific FisheryManagement Council, pp. 711–776.
McConnaughey, R. and Smith, K. (2000) Associations betweenflatfish abundance and surficial sediments in the easternBering Sea. Can. J. Fish. Aquat. Sci. 57:2410–2419.
Mecklenburg, C., Mecklenburg, T. and Thorsteinson, L. (2002)Fishes of Alaska. Bethesda, MD: American Fisheries Society,1037pp.
Mito, K.-I., Nishimura, A. and Yanagimoto, T. (1999) Ecologyof groundfishes in the eastern Bering Sea, with emphasis onfood habits. In: Dynamics of the Bering Sea. T. Loughlin & K.Ohtani (eds) Fairbanks, AK: University of Alaska SeaGrant, pp. 537–580.
Mordy, C., Stabeno, P., Ladd, C., Zeeman, S., Wisegarver, D.and Hunt, G.L. Jr (2005) Nutrients and primary productionalong the eastern Aleutian Islands Archipelago. Fish. Ocea-nogr. This issue.
Mueter, F.-J. and Norcross, B. (1994) Distribution, abundance,and growth of larval walleye pollock, Theragra chalcogramma,in an Alaskan fjord. Fish. Bull. 92:579–590.
Mueter, F. and Norcross, B. (1999) Linking community struc-ture of small demersal fishes around Kodiak Island, Alaska, toenvironmental variables. Mar. Ecol. Prog. Ser. 190:37–51.
Mueter, F. and Norcross, B. (2000) Species composition andabundance of juvenile groundfishes around Steller sea lionEumetopias jubatus rookeries in the Gulf of Alaska. AlaskaFish. Res. Bull. 7:33–43.
Mueter, F. and Norcross, B. (2002) Spatial and temporal pat-terns in the demersal fish community on the shelf and upperslope regions of the Gulf of Alaska. Fish. Bull. 100:559–581.
O’Connell, V.M. and Carlile, S.W. (1993) Habitat-specificdensity of adult yelloweye rockfish Sebastes ruberrimus in theeastern Gulf of Alaska. Fish. Bull. 91:304–309.
OCSEAP (1986) Marine fisheries: resources and environments.In: The Gulf of Alaska: Physical Environment and BiologicalResources. D. Hood & S. Zimmerman (eds) Washington, DC:US Department of Commerce. pp. 417–460.
Olson, D. and Backus, R. (1985) The concentrating of organ-isms at fronts: a cold-water fish and a warm-core Gulf Streamring. J. Mar. Res. 43:113–137.
Orr, J.W. (2004) Lopholiparis flerxi, a new genus and species ofsnailfish (Scorpaeniformes: Liparidae) from the Aleutian Is-lands, Alaska. Copeia 2004:551–555.
Orr, J.W. and Busby, M.S. (2001) Prognatholiparis ptychoman-dibularis, a new genus and species of the fish family Liparidae(Teleostei: Scorpaeniformes) from the Aleutian Islands,Alaska. Proc. Biol. Soc. Wash. 114:51–57.
Palumbi, S.R. (1994) Genetic divergence, reproductive isola-tion, and marine speciation. Ann. Rev. Ecol. Syst. 25:547–572.
Pearcy, W., Stein, D., Hixon, M., Pikitch, E., Barss, W. andSteiner, R. (1989) Submersible observations of deep-reeffishes of Heceta Bank, Oregon. Fish. Bull. 87:955–965.
Schneider, D. (1990) Seabirds at fronts: a brief overview. PolarRes. 8:17–21.
Sheiko, B. and Federov, V. (2000) Chapter 1. Class Cephal-aspidomorphi – lampreys. Class Chondrichthyes – cartila-ginous fishes. Class Holocephali – chimaera. ClassOsteichthyes – bony fishes. In: Catalog of Vertebrates ofKamchatka and Adjacent Waters. R. Moiseev & A. Tokranov(eds) Petropavlovsk: Kamchatsky, pp. 7–69.
Sinclair, E. and Zeppelin, T. (2002) Seasonal and spatial dif-ferences in diet in the western stock of Steller sea lions(Eumetopias jubatus). J. Mamm. 83:973–990.
Sinclair, E., Moore, S., Friday, N., Zeppelin, T. and Waite, J.(2005) Do patterns of Steller sea lion (Eumetopias jubatus)diet, population trend and cetacean occurrence reflectoceanographic domains from the Alaska Peninsula to thecentral Aleution Islands? Fish. Oceanogr. 14(Suppl. 1):223–242.
Stabeno, P., Kachel, D., Kachel, N. and Sullivan, M. (2005)Observations from moorings in the Aleutian passes:temperature, salinity and transport. Fish. Oceanogr.14(Suppl. 1):39–54.
Stevenson, D.E., Orr, J.W., Hoff, G.R. and McEachran, J.D.(2004) Bathyraja mariposa, a new species of skate (Rajidae:Arhynchobatinae) from the Aleutian Islands. Copeia2004:305–314.
Van Tongeren, O. (1995) Cluster analysis. In: Data Analysis inCommunity and Landscape Ecology. R. Jongman, C. Ter
Braak & O. Van Tongeren (eds) New York: CambridgeUniversity Press, pp. 174–212.
Williams, E. and Ralston, S. (2002) Distribution andco-occurrence of rockfishes (family: Sebastidae) over traw-lable shelf and slope habitats of California and southernOregon. Fish. Bull. 100:836–855.
Wilson, M. (2000) Effects of year and region on the abundanceand size of age-0 walleye pollock, Theragra chalcogramma, inthe western Gulf of Alaska. Fish. Bull. 98:823–834.
Wyllie-Echeverria, T. and Ohtani, K. (1999) Seasonal sea icevariability and the Bering Sea ecosystem. In: Dynamics of theBering Sea. T. Loughlin & K. Ohtani (eds) Fairbanks, Alaska:University of Alaska Sea Grant, pp. 435–452.
Yang, M.-S. (1999) The trophic role of Atka mackerel, Pleuro-grammus monopterygius, in the Aleutian Islands area. Fish.Bull. 97:1047–1057.
Yang, M.-S. (2003) Food habits of important groundfishes in theAleutian Islands in 1994 and 1997. AFSC Processed Rep.2003–07: 233 pp.
Yang, M.-S. and Nelson, M. (2000) Food habits of the com-mercially important groundfishes in the Gulf of Alaska in1990, 1993, and 1996. U.S. Dep. Commer. NOAA Tech.Memo. NMFS-AFSC-112: 174 pp.
