Introduction

The limits and production of ampeliscid amphipod populations are major concerns for whale conservation because they are critical prey of eastern and western North Pacific gray whales, Eschrichtius robustus (Lilljeborg, 1861) [1–5]. Recovery of the western gray whales (WGW from here on), since the ending of commercial whaling in the 1970s, has been slow relative to eastern gray whales (EGW). Eastern gray whales are no longer listed as threatened or endangered while the WGWs remain in the Red Book list [6]. We examined the life history and production of Ampelisca eschrichtii Krøyer, 1842, a critical prey species of WGWs in their Offshore feeding area on the northeastern Sakhalin Shelf of the Okhotsk Sea [7] to partially test whether WGWs could be food limited.

Ampelisca macrocephala Liljeborg, 1852 and other less abundant ampeliscid species are the primary prey of the eastern gray whales in their main feeding area, the Chirikov Basin of the Bering Sea [8]. Previous to the 1980s, Chirikov Basin ampeliscid populations were limited primarily by the epi-benthic flow of organic matter, intra- and inter-specific competition for space and reproductive success [9]. However, EGW numbers doubled between 1967 and the late 1980s [10–12]. The estimated 26,916 EGWs of 1987/1988 declined to 18,178 by the 1993/1994 census [11] and the EGWs have remained around 20,000 since 2002 [11–13]. Coyle et al. [8] estimated that the 2002/2003 EGWs consumed a significant fraction of A. macrocephala production in the Chirikov Basin and thus have limited ampeliscid populations since the late 1990s [14]. Punt and Wade [15] estimate that EGW are presently at around 85% of their prey carrying capacity. Coyle et al. [8] additionally proposed that the slow growth rates and the relatively long (4–5 year) life span of A. macrocephala could require decades for their populations to recover to densities observed previous to the 1980s. Climate-change related affects in the Bering Sea are also proposed as mechanisms that could limit gray whale food sources [16–18].

A potential for limited WGW food has similarly emerged as a particular concern in the western North Pacific. Weller et al. [19] identified 69 WGWs by 1998. The overall estimated WGW number remains at less than 200 to date and the annual increase in their numbers has declined each year since 2001 [13,20–23].

Ninety-seven of the 122 known western gray whales foraging on the Northeastern Sakhalin Shelf of the Okhotsk Sea in 2008 were sighted exclusively within the Nearshore area adjacent to the Piltun and Chayvo Lagoons (Fig 1) [20,24]. More than 40% of these whales occurred in the Piltun area, adjacent to Piltun Lagoon (52.7–53.4°N and 143.1–143.4°E) at less than 20 m depths and within 5 km of shore during the 2011 survey season [25]. Also in 2011, 37 to 83 whales were observed in the Offshore feeding area (Fig 1) [25] (approximately, 51.8–52.4°N and 143.4–143.9°E, southeast of Chayvo Bay, northeast of Niyskiy Bay and 30–45 km off of the Sakhalin coast in 40–60 m depths) [3,4,7,24,26–29]. The dimensions of these foraging areas are defined by the distributions of feeding whales observed primarily during aerial surveys [27]. The benthic communities of these areas are dominated by dense aggregations of gammaridean amphipods [30–32]. The restricted general position of the entire Offshore feeding area has been apparent in all sampling efforts since 2002 [7].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The Offshore feeding area.

NE Sakhalin Island (solid red rectangle), 2011 western gray whale foraging perimeter (blue dashed line), Complex I-III sample sites (small blue dots), 2002–2013 Complex IV Ampelisca eschrichtii (photograph) area (red dashed line) including station designations followed by their year of sampling.

https://doi.org/10.1371/journal.pone.0147304.g001

The Nearshore and Offshore feeding areas are within the region of highest primary and secondary production in the Okhotsk Sea [31,33–40]. Strong winds, tidal mixing, upwelling, and Amur River discharges over the northeastern Sakhalin Shelf induce intense diatom blooms in these areas during the ice free May-November period [37,41–43].

Fadeev [4,7] used cluster analysis of the macrobenthos abundance in the Offshore feeding area to identify four benthic complexes dominated by: sand dollars (Complex I), cumaceans and non-ampeliscid amphipods (Complex II), ampeliscid amphipods mixed with bivalves and sea anemones (Complex III) and densely dominant ampeliscids (Complex IV). Complex IV consists mainly of Ampelisca eschrichtii populations and covers the largest eastern and southeastern part of the study area (Fig 1 herein and in [7]). The Offshore area is assumed to be critically important for WGW feeding [7,24].

Ampeliscids construct mucous cemented sac-like tubes in dense aggregations on muddy, medium and fine sands in areas dominated by strong bottom currents [7,44–46]. Ampeliscids feed from their tubes on epibenthic particles including “phyto-detritus” [8] that they capture from the epibenthos and sediment surface with their second antennae [44,46]. Stable isotope ratios and fatty acid compositions of A. eschrichtii of the Sakhalin Shelf indicate that planktonic diatoms are the most important and ultimate A. eschrichtii food [47]. High Arctic shallow benthic food webs nevertheless, do not all vary directly with seasonal changes in primary production [48] and many benthic species rely on detrital pathways rather than direct primary production [49] which can dampen the impacts of large seasonal fluctuations in primary production.

In the Offshore Complex IV area, Ampelisca eschrichtii comprise >95% of the amphipod biomass and can exceed densities of 16,000 m⁻² and reach biomasses of up to 1 kg m⁻² [4,7,24,28]. Amphipods brood their young until they are released as fully formed juveniles and thus lack specialized larval dispersal stages. Quantitative benthic grab sampling is therefore sufficient for measuring entire amphipod populations. As in other high Arctic systems however, (e.g., A. macrocephala in the Chirikov Basin [50] and in Svalbard [51]), access to the Sakhalin shelf benthic communities has been possible only within the late June to early November ice free period [4,7,24,28]. The lack of access to these populations in winter months has confounded efforts to measure annual A. eschrichtii growth and production or to resolve the relations between A. eschrichtii production, primary production and detrital food pathways. The year to year contributions of these amphipods to the energy requirements of their WGWs predators has therefore also remained unclear. Expanded comparisons of all data from six years of sampling during summer months between 2002 and 2013, nevertheless, has permitted estimates of A. eschrichtii life history, population dynamics, production and biomass. These new results also allow comparisons of WGW prey abundance and production with EGW prey abundance and production.

Special permits were not required for the benthic invertebrate samples collected for this research. All information on WGW for this research is from previously published data. No interactions with endangered or threatened species occurred in the course of this research.

Materials and Methods

We measured the population dynamics and production of A. eschrichtii from 45 samples collected at 24 stations in 44–60 m depths within the Sakhalin Offshore area between June and October in six sampling years spanning 2002 to 2013 (Fig 1 and S1 Table). We compared male/female ratios, the occurrence and frequencies of female reproductive conditions, the presence of terminal phase breeding males, the presence of newly hatched juveniles and the relative weights of similar sized reproductive males and females. We expected high proportions of brooding females, major changes in the frequencies of brooding females between early and late summer months and significant changes in size frequencies among different sampling months in our samples if production and growth occurs primarily in summer.

Bottom water temperatures on stations ranged between -0.8 and 9.2°C, salinities ranged between 30.8 and 32.3 psu, opacity (in Formazin Turbidity Units [FTUs]) ranged between 1.0 and 8.9 and oxygen saturation levels ranged between 83% and 90%. Sediments analyzed from 14 of the 24 stations and ranged between medium, fine and silty sands (S1 and S2 Tables). Sampling was completed within a few days of each year except in 2007, when sampling extended over a 75 day period between July and October (Table 1 and S1 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Ampelisca eschrichtii density, biomass and production.

Day, month and year (Date), juvenile, adult and total density (respectively, J m^-2, A m^-2 and N m^-2), juvenile, adult and total grams biomass (respectively, J g m^-2, A g m^-2 and B g m^-2), grams production m^-2 per sample (P_s), production per biomass per sample (P/B_s) estimated from juvenile and adult cohorts within sample dates (P_s and P/B_s) and production per biomass estimated from juvenile and adult cohorts among consecutive years (P_yr and P/B_yr) of the Offshore A. eschrichtii populations between 2002 and 2013 and their mean values from the normalized averages among the 6 survey years (see also S1 Table).

https://doi.org/10.1371/journal.pone.0147304.t001

Samples were collected using an 0.25 m² Okean grab from the R/V Bukhoro in the 2002 TINRO-Center expedition and using an 0.2 m² van Veen grab in the IMB expeditions from the R/V Igor Maksimov (in 2011) and R/V Akademik Oparin in all other years. These samples were washed through a series of sieves ending at 1.0 mm mesh size in 2002 and a mesh size of 0.5 mm in all following years. The retained organisms were fixed in 4% formalin and later, sorted out, identified, weighted, measured and then transferred to 75% ethanol. We compared density and biomass of A. eschrichtii among the 45 samples (S1 Table) with sediment composition at the 14 stations where sediment data were available (S1 and S2 Tables).

Amphipods from 20 samples and 17 of the 24 stations were selected randomly for morphometric analyses (S1 Table). Quantitative subsamples of A. eschrichtii from each of the major samples were taken to reduce processing efforts. We noted sex and reproductive development in 17 samples from 17 sites (S1 Table and Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Reproductive development among months and years.

Female reproductive stages (F0-FIV) and female/male ratio (F/M) of the Offshore A. eschrichtii.

https://doi.org/10.1371/journal.pone.0147304.t002

All A. eschrichtii in each subsample were weighed together to the nearest 5 mg and then subsets of these samples were counted, sexed, measured, blotted dry on filter paper and individually weighed to the nearest milligram. Body lengths were measured to the nearest 0.1 mm from the tip of the rostrum to the basis of the telson of outstretched A. eschrichtii using a calibrated ocular micrometer under a stereomicroscope. Males were distinguished by the presence of penile papillae and secondarily, by their relatively longer antennae, powerful pleopods and transversely pleated coxal gills. Females were distinguished by the presence of oöstegites and secondarily, by unpleated coxal gills.

We verified morphological classifications of juvenile, male and female reproductive development by histological analyses of 5 juveniles, 14 morphological males, 10 morphological females and a possible intersex specimen collected in October, 2013 (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Reproductive development by sex and length.

Sex (assessed by morphology), estimated cohorts (C3-C11, see below), body length (mm), number of specimens (N), and reproductive development (assessed from morphology and histology of juveniles, males, females and an intersex specimen) of A. eschrichtii collected in October 2013.

https://doi.org/10.1371/journal.pone.0147304.t003

Each specimen used for histological analyses was soaked in fresh water for 24 h, dehydrated, cleared in xylene and then infiltrated with melted paraffin. The paraffin was cooled into blocks that were cut into 10 μm thick sections for mounting on slides for microscopy. Sections containing gonad tissue were stained using hematoxylin and eosin and permanently mounted on glass slides. The histology slides and whole dissected specimens for these analyses are deposited in the museum collections of the A.V. Zhirmunsky Institute of Marine Biology of the Far Eastern Branch of the Russian Academy of Sciences.

We assessed the sequence of reproductive development with size and with time (Tables 2 and 3) from the variation in reproductive development among male and female lengths and collection years. We followed Tzvetkova classification [52] of female reproductive condition:

1. F0 –rudimentary oöstegites lacking egg retention setae;
2. FI–brooding uncleaved eggs;
3. FII–brooding cleaved eggs;
4. FIII–brooding fully formed juveniles;
5. FIV–fully developed oöstegites bearing egg retention setae and empty brood pouch.

We counted eggs, where egg loss during handling was not apparent.

We estimated A. eschrichtii weight increase per length using the power function: (1) where, W is individual wet weight (g), L is body length (mm) and a and b are estimated coefficients. Weights of all specimens were estimated by this model where weight was not recorded. We compared deviations from exponential weight increase with length to assess potential growth limits and assumed that Ampelisca less than the estimated weight were underweight and Ampelisca greater than the estimated weight were above normal weight. We also used this relation to estimate total weight among length classes and biomass (B) of overall populations.

We identified sequential modes among 1 mm length-frequencies by the Bhattacharya and Normsep procedures in the FiSAT II software package [53,54]. The FiSAT II procedure distinguishes size modes by a separation index (S.I.) of greater than 2. We assumed that common FiSAT II identified length cohorts correspond to relative cohort ages and interpreted increasingly larger size frequency modes as sequentially older cohorts. We estimated changes in intra-annual growth, survival and biomass of the 2007 A. eschrichtii populations sampled in a 75 day period (July-October, S1 Table) from the changes in their length frequencies, densities and biomasses over time. We distinguished variations of population size structures among samples and years by cluster analysis (S1 Table, S2 Table and S1 Fig) in PRIMER v.6 statistical software [55]. We normalized these absolute frequencies by subtracting the mean and taking their square roots and constructing a similarity matrix (rows–length classes, columns–samples) based on the Bray Curtis coefficient for this analysis [56].

We estimated average annual A. eschrichtii production (P) over the n sampling years (n = 6) by cohort summation [57,58] where: (2) in which, N_1,i is the number of 0+ juveniles m^-2 in year i, N_2,i is the number of 1+ (one year older than the juveniles), adults m^-2 in year i, W_1,i is the average wet weight of a juvenile, and W_2,i is the average weight of an adult in year i. We estimated production to biomass P/B from the average annual production divided by the average summer biomass m^-2, B. We estimated production from the juvenile cohort of the first year relative to the adult cohort of the following year within sample years (2002, 2007, 2008, 2011, 2012 and 2013), among consecutive sampling years (2007–2008, 2011–2012 and 2012–2013) and by the normalized means of the six sample years (Table 1).

Sampling entirely within WGW feeding areas, defined by vessel based sightings, each year, is not possible and year to year positions of maximum Ampelisca beds also cannot be predicted. We therefore estimated the Offshore area overlap of 2007–2013 A. eschrichtii Complex IV sampling sites and WGW foraging (Fig 1) by the field perimeters 2011 WGW sightings in 2011 [25]. We estimated the overlapping areas of these perimeters by polygon triangles summation using their geographical coordinates [59]. We produced all maps using open source software QGIS v. 2.12.1-Lyon (http://qgis.org/ru/site/).

Results

Reproductive development

Less than 12 mm length A. eschrichtii lack gonads and do not have suitable external morphologies for distinguishing sexes. We therefore classified all less than 12 mm length individuals as juveniles. Nearly all 12 to 17 mm length A. eschrichtii bear penile papillae that we initially misinterpreted as male characters. These individuals however, lack gonads (Table 3, rows 1–3) and thus cannot be functional males. Eight morphologically identified males greater than 18 mm in length (Table 3, rows 4–8) contained gonads that held spermatocytes (Fig 2a), spermatids (Fig 2b) and spermatozoa (Fig 2c). No males contained ova. Testes of the 18 mm male contained only spermatocytes (Table 3, Fig 2a) while the twenty mm length male testes contained spermatocytes and spermatids (Fig 2b). Spermatozoa were most abundant in the testes of 24 and 26 mm length males (Fig 2c). The advancing reproductive condition with size among the eight 18 mm and larger males indicates that only the largest males are reproductive (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Histology of A. eschrichtii males and female gonads.

a) –18 mm male testis; b)– 21 mm male testis; c)– 26 mm male testis; d)–ovary of 15 mm female at stage F0; e)– 22 mm female at stage F0 and; f)– 32 mm female at stage FIV; Spc–spermatocytes, Spt–spermatids, Spz–spermatozoa in spermatophore, Pvo–previtellogenic oocytes, Vo–vitellogenic oocytes (scale = 100 μm).

https://doi.org/10.1371/journal.pone.0147304.g002

Results of our histological analyses of female and male gonads are consistent with a gonochoristic life history. The ovaries of 15 mm F0 females contained previtellogenous oocytes (Fig 2d). The 22–24 mm F0 female ovaries contained mostly vitellogenic oocytes that were completely formed and ready for oviposition and some previtellogenic oocytes (Fig 2e). The ovaries of 23–25 mm FII females (brooding cleaved eggs) were either empty or contained possibly resorptive vitellogenic oocytes and relatively few previtellgenous oocytes (Table 3, row 11). The large 32–33 mm F0 females ovaries were filled either by previtellogenic oocytes (Table 3, Fig 2f) or by vitellogenic oocytes with a new generation of the previtellogenic oocytes (Table 3, row 13) which indicates that A. eschrichtii can produce at least 2 broods. Two morphological intersexes, 19 mm and 30 mm in length, among 428 A. eschrichtii specimens examined from the 18 October, 2013 samples, bore small oöstegites and male genital papillae. However, the 30 mm specimen lacked ovaries and its testes contained only spermatozoa (Table 3, row 14) and thus it did not appear to be an intersex or to have changed sex.

Variations in length to weight among males and females also did not indicate sex change (Fig 3). The ranges of A. eschrichtii male and female lengths are similar among samples (Student’s test, p < 0.05) and the largest A. eschrichtii female by length (33 mm) in our samples, was collected at the same location 18 October, 2013 as the largest male (27 mm) that we found.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Ampelisca eschrichtii weight deviation with length.

a) Estimated (red dashed line) and observed female, male, and juvenile wet weights by length (blue diamonds, black circles and orange crosses, respectively) (r² = 0.99, df = 398, p < 0.001) collected in 2002, 2007, 2012 and 2013 with the red box enclosing greater than or equal to 16 mm males and females that; b) deviate from the expected weight (red dashes, Fig 3a and 3b) with increasing length (black line, r² = 0.187, df = 167, p < 0.001).

https://doi.org/10.1371/journal.pone.0147304.g003

The smallest F0 female confirmed by histology was 12 mm in length and bore small, bare oöstegites. F0 females (with oöstegites lacking setae and lacking developing ova) ranged between 12 and 33 mm and occurred in all samples. We found 21 to 27 mm length FI females brooding uncleaved eggs in all August and September samples (Fig 4) and FII females ranging between 23 and 28 mm in length, brooding cleaved eggs (Fig 4) in the latest seasonal collections, mainly in mid-October, 2013 (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Female reproductive stages among length frequencies.

a) 10–28 mm body length classes of stages F0, FI, FII and FIV from 2002, 2007, 2011, 2012 and 2013 combined with a vertical red line marking the greater than 16 mm size ranges included in; b) the percentages of reproductive stages F0, FI, FII and FIV among 16 mm and greater length females.

https://doi.org/10.1371/journal.pone.0147304.g004

Reproductive development of females is thus size dependent. All brooding females were greater than 20 mm in length and FIII females bearing juveniles were absent (Fig 4). The average female reproductive stages F0—FIV among all years were, respectively, 80%, 5%, 6%, 0% and 9% (Table 2). Thus, the combined average of stage FI and FII females among all months and years sampled was only 11%. Moreover, 20% of all females from all samples and years were of reproductive stages FI or greater (Table 2). Females brooding juveniles (stage FIII) were absent and all FIV females, that had released juveniles, were greater than 24 mm in length (Fig 4). The low frequency of egg bearing females (FI and FII) (Table 2) is not consistent with prolific reproduction during the June to October months sampled during this survey (S1 Table).

The ratio of morphological A. eschrichtii male per female among samples was 1:1.6 (Males = 1.02*Females– 6.3; r² = 0.73, df = 15, p < 0.01, Table 2). The low male frequencies can result from greater male mortality or to undercounting of small males. We did not find terminal-molt A. eschrichtii males in our samples, similar to those reported by Highsmith and Coyle [50] among A. macrocephala in the Chirikov Basin or by Mills [44] among A. abdita Mills, 1964 on the northwestern Atlantic shelf. Such males are expected during maximum reproductive periods. Increasing frequencies of lower than expected weights of large males and females among all sample months and years (Fig 3b) and the lack of terminal phase males are inconsistent with rapid growth or ongoing summer reproduction in these populations.

Length frequencies

Six to eight minor length class modes were apparent in all samples and years (Fig 5a–5j) that converged into two major length class modes. The two major modes separated within the 15 to 17 mm length intervals were apparent in all yeas (Fig 5a–5j). The juvenile-adult division is particularly apparent when all years are averaged in proportions of weight (Fig 5j). The small size mode consisted of juveniles and the larger size mode consisted of mature males and females (Fig 5j). The 16 mm division is consistent only with length classes being separate age classes in contrast to instars. The smallest A. eschrichtii in our samples (3.8 mm in length) was collected 29 September, 2008. The low frequencies of such newly hatched juvenile length classes additionally indicate that our summer-fall sample populations (Fig 5a–5i) were not actively reproducing. The constant sex ratio we found over the sampling season was also inconsistent with active reproductive activities and high male mortality observed among other ampeliscids [46,60].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Relative cohort densities or biomass among years and months.

Including (C1-C9) densities m^-2 among 1 mm length frequencies of the 2002–2013 sampling dates (a-i) and the proportion of average biomass m^-2 for all years (j) with number of specimens (N), observed frequencies (vertical bars), predicted frequencies (dashed lines), regressions of predicted and observed frequencies and example of juvenile and adult reproductive development and sex ratios in (i).

https://doi.org/10.1371/journal.pone.0147304.g005

The proportions of distinguishable juvenile and adult size modes varied greatly among samples and years (Fig 5a–5i). A greater than 30 mm length cohort occurred only in 2013 (Fig 5i). All other cohorts in all other years were less than 28 mm in length (Fig 5a–5h). The low variation in length frequencies within 2007 samples (Fig 5b–5e) and greater size variation within samples from other years (Fig 5a and 5f–5i) indicates that variations in population structure occur mainly among years rather than among sample locations. A low frequency of 15 to 17 mm length classes divides juveniles from reproductive males and females, in all samples (Fig 5a–5j). The 15–17 mm division is also consistent with reproductively mature length classes being one year older than the juveniles. Alternative potential growth and development, including longer ampeliscid life spans (> 4 years) [45], would obscure this separation.

The six to eight minor length frequency modes among the July-October samples of 2007 (Fig 5b–5e) were apparent also in other years (Fig 5a and 5f–5i) and are also likely to consist of individual cohorts. Average lengths of the individual cohorts and of the composite juvenile and adult cohorts over the 75 day sampling period of 2007 did not vary significantly (Fig 5b–5e) and thus, individual growth was not apparent in the 2007 sampling period. Moreover, the densities and biomasses of the 2007 the A. eschrichtii populations declined (Fig 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Summer declines of Ampelisca eschrichtii.

a) density and b) biomass over the 75 day sampling period (23 July-5 October, 2007).

https://doi.org/10.1371/journal.pone.0147304.g006

The two major length modes we found in all years indicate that the lack growth among the 2007 cohorts was not due to random sample variation or to exceptional growth conditions in 2007. The bimodal population structure, separating juveniles and reproductive adults therefore is a likely result of a two-year A. eschrichtii life span and juvenile release outside of summer months. We therefore assumed a two-year cycle for our estimates of annual production.

Production

We estimated production based on the 0+ year juvenile (≤16 mm length) and 1+ year adults (> 16 mm length) cohorts identified above (Fig 5) among samples (Table 1, P_s), from 0+ year juveniles of previous years relative to 1+ adults in consecutive following years (Table 1, P_yr) and from the averages of average year estimates (Table 1, Means). Although our estimates of biomass, production and energetics (Tables 1 and 4) include conservative parameters for the Offshore feeding area relative to the Chirikov Basin A. macrocephala and to other ampeliscids [8], the energetics of the two feeding areas appear dramatically different (Table 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Gray whale and ampeliscid energetics.

In the Offshore (Okhotsk Sea) and Chirikov Basin (Bering Sea) feeding areas.

https://doi.org/10.1371/journal.pone.0147304.t004

Thirty-nine percent of the 564 km² 2011 WGWs foraging ground in the Offshore was overlapped by the 665 km² 2002–2013 ampeliscid sampling area (Fig 1). A. eschrichtii occurred in 95% of all samples in the benthic “Complex IV” of the Offshore [28]. However, sea urchins Echinarachnius parma (Lamarck, 1816) present in 17% of samples from the whole Offshore feeding area and the predominance of gravel in 17% of all samples there [28] were likely to limit the value of these sites and locations for WGW foraging. Thus, up to 34% of the Offshore area could be unsuited for WGWs foraging. We therefore estimate, conservatively again, that the effective WGWs foraging ground of 2011 was about 372 km² (Table 4).

The Offshore WGW feeding area (that includes ampeliscid biomasses >300 g m^-2) is about 4% of the EGW Chirikov feeding area (Table 4). The maximum density of WGWs observed in the Offshore between 2002 and 2013 was less than half as great as EGW densities in the Chirikov area and the A. eschrichtii biomass per WGW in the Offshore area was 6 times the ampeliscid biomass per EGW in the Chirikov feeding area (Table 4). WGWs also had 8 times the ampeliscid production per whale than was available to EGWs in the Chirikov Basin (Table 4). Thus, ampeliscid biomass and production is greater in WGW feeding areas than in EGW feeding areas. These data nevertheless, do not challenge previous conclusions of food limitation for EGW [8,10,14].

Discussion

Northeastern Sakhalin Shelf Ampelisca eschrichtii have a predominantly two-year gonochoristic, iteroparous life cycle in which growth occurs primarily under ice cover. These populations therefore must require detritus or other winter based primary production food sources. Moreover, these abundant A. eschrichtii populations are unlikely to be limited by WGW predation.

Although the proportions of A. eschrichtii individuals within particular length modes varied dramatically among years and samples, adjacent length modes invariably clustered, with 3–4 juvenile length classes and 3–4 adult length classes divided by a low frequency gap at around 16 mm. We infer that the 16 mm division was overgrown by the adult cohorts of our summer samples in the previous winter but not reached by the juvenile cohorts produced in the previous winter. The 3–4 reproductively adult size classes in our summer populations were thus most likely to have been new juveniles two winters previously. Asynchronous and unequal growth rates over increasingly diverse ages force individual length classes of long lived species to overlap and merge. The widely separated length modes of A. eschrichtii are consistent only with a relatively short (2 year) life span and synchronous (winter-spring) growth among size classes.

Our observations of underweight reproductive sized males and females, a lack of newly hatched (less than 4 mm length) juveniles, the absence of FIII brooding females and the absence of secondary phase reproductive males are consistent with a lack of summer growth. The declines of abundances and biomasses of the Offshore A. eschrichtii between early summer and late autumn of 2007 are also consistent with winter or early spring reproduction and biomass production outside of our sample periods.

The increasing sizes of consecutive female reproductive stages identified by histology are consistent with no more than two broods in any winter and 3 broods (rarely) over a life span. Stage FI and FII females that we found are likely to be the first to reproduce in winter. Stage F0 females, having greater than average weights per length, are likely to produce the second winter cohort of juveniles. The largest and underweight stage FIV females of summer are likely to produce the third or fourth winter/spring juvenile cohorts. Survival of reproductive sized A. eschrichtii over two summers that we possibly observed in the 10^th (30 mm) cohort of 2013, appears to be rare. Thus, nearly all individuals in our samples were less than two years old. A. eschrichtii is thus gonochoristic and reproduces once or, rarely, twice during each extended winter and spring period of ice cover.

Our inferred life history of A. eschrichtii is consistent with previous observations of high latitude ampeliscid biology. Fifteen of the 18 benthic amphipod species on the west Spitsbergen coast, including A. eschrichtii, incubate eggs during the polar night and release their offspring in early April [51]. Broods of A. eschrichtii collected in summer could have been deposited before sampling commenced. The lack of early stage juveniles in the 5 month sampling period also indicates that the eggs we found were not hatching over the sampling period. These observations are consistent with the life history of A. macrocephala in the Øresund in which breeding occurs after October and is followed by the complete absence of mature males in later months [46]. The absence of less than 5 mm A. macrocephala juveniles in the Chirikov Basin [9] similarly indicates that these populations also carry developing eggs over summer and fall that hatch in winter. Our estimates of P/B are within the range of previous estimates for ampeliscids [60].

Although diatoms must be the ultimate food source of A. eschrichtii [47], winter growth of A. eschrichtii on the northeastern Sakhalin Shelf occurs after most diatom production has ceased and settled to the benthos as phytodetritus [61]. Large populations of A. eschrichtii occur throughout the Okhotsk Sea, including the western Kamchatka Shelf [62], the eastern Sakhalin Shelf [30,31] and Sakhalinskiy Bay [63] of the northernmost Sakhalin Island. Pasternak [63] proposed that A. eschrichtii aggregations in fine and silty sands, adjacent to silty muds at the mouth of the Sakhalinskiy Bay form in response to optimal currents and detritus accumulation. A winter dependence on detrital food is common among other high latitude benthic amphipods [48,49]. WGWs that forage on A. eschrichtii are therefore likely to depend on detrital pathways from primary production of previous summers.

The production and biomass of the Offshore area A. eschrichtii ([7,28] and herein) exceeds the previous record for ampeliscids [64] and may be the highest for any amphipod population in the world [60]. More than 99% of the summer A. eschrichtii biomass is among individuals that are greater than the critical 8 mm in length required to be retained by WGW baleen [2,65]. The Offshore area is not the most extensive WGW feeding area in Far Eastern Russia [20] and alternative WGW feeding areas remain in use each year. If WGW forage optimally, their continuous use of all feeding areas (including Offshore) indicates that they are of similar value. Our estimated high WGW food stocks correspond with previous observations [7,24]. The high production rates of the Offshore area ampeliscid populations provide significant potential to quickly recover from intense predation. Since other feeding areas are likely to contain similar prey sources as Offshore, the limited growth of the WGW feeding aggregate, numbering less than 200, off northeastern Sakhalin [20] is unexpected when 1,500–10,000 WGW are estimated to have been in the region in the 19th century [66].

All regions of the North Pacific are likely to be accessible to gray whales [13]. However, massive immigration by EGWs to the seemingly abundant Western Pacific food sources or unusually rapid growth of the WGW numbers has not occurred. Unless Allee effects [67] or mortality sources, such as predation, unsanctioned whaling, diseases and parasites are greater among WGWs than among EGWs, the potential mechanisms for food limitation of all Pacific gray whales warrants reconsideration. Prey abundance is necessary but not sufficient to preclude food limitation. Not all sites in the Offshore area where the highest Ampelisca biomasses occur appear to be used by WGW (Fig 1). Ineffective, naïve or otherwise inefficient foraging behaviors resulting from genetic or social bottlenecks [13] or prey refuges could limit WGW access to these prey [28]. Ampelisca populations in coarse sediments, for instance, could be less accessible to foraging whales [28]. Intense fishing could also alter benthic community composition on the Sakhalin Shelf and expand the distributions of inedible species that reduce WGW access to benthic prey [28]. Climate changes may also limit WGW foraging by altering primary production, timing and extent of ice cover, detritus distribution and the consequent benthic communities and sediments. More information on the conditions of benthic communities that affect WGW foraging success and winter sampling to resolve the winter life history and ecology of these amphipods would greatly benefit WGW conservation efforts.

Supporting Information

Acknowledgments

This research was initiated by Dr. V.I. Fadeev^† (24.10.1952–21.06.2014)—Head of the Laboratory of the Ecology of Shelf Communities of the A.V. Zhirmunsky Institute of Marine Biology, IMB. Dr. V.I. Fadeev participated in most expeditions to the Sakhalin Shelf and sampling of the amphipod material for this report, read and approved of an early draft of this manuscript but passed away before the final version was complete. The surviving authors (NLD, JWC and VBD) accept all responsibility for the integrity and validity of the data collected and analyzed. We thank E.P. Dats (Vladivostok State University of Economics and Service, VSUES) for assistance with software, editing formulas and calculations and support at all stages as a friend and a colleague. We are grateful to L.N. Khorol'skaya (IMB) for her assistance in measuring amphipods. We thank Dr. L.L. Budnikova and R.G. Bezrukov (TINRO-Center) for June 2002 collection material on A. eschrichtii from the Offshore area. We thank Dr. V.V. Ivin for help with statistical analysis in software PRIMER (IMB). We thank Prof. A.I. Buyanovsky (VNIRO) for critical reading and discussion of the results and Dr. Dave Rugh (National Marine Mammal Laboratory, NOAA, Seattle, Washington) for critically reviewing and assisting with our summaries of EGWs. We thank Dr. Vladimir Efremov (Exxon Neftegas Limited) for his encouragement and review of the manuscript. We also thank Dr. James Sumich (Hatfield Marine Science Center, Newport, Oregon), the PLOS ONE editor and an anonymous reviewer for their timely, helpful reviews.