Distribution and biodiversity patterns of hyperiid amphipods across the coastal-offshore gradient of the sub-tropical Southeast Pacific

Abstract

The subtropical region of the Southeast Pacific exhibits a strong onshore-offshore gradient in hydrographic conditions and biological production from the eutrophic upwelling zone to the ultra-oligotrophic oceanic area in the central South Pacific gyre (SPG). Across this gradient, zooplankton must cope with either gradual or abrupt changes in environmental conditions. Here, the distribution and diversity of hyperiid amphipods were assessed over this gradient in the upper 1000 m in relation to temperature, salinity, oxygen, chlorophyll-a, sea level anomalies, the bulk of mesozooplankton biomass and the biomass of salps, siphonophores, and other hydrozoans during October/November 2015 from the coastal zone off Chile (27° 00‵ S, 70° 52‵ W) to near Easter Island within the SPG (27° 10‵ S, 109° 20‵ W). The most frequent and abundant species were Hyperioides longipes Chevreux, 1900, Eupronoe minuta Claus, 1879, and Hyperioides sibaginis (Stebbing, 1888). Significant changes in abundance and community structure across the gradient with respect to the evaluated environmental variables and significant correlations of 17 hyperiid species with the gelatinous category other hydrozoans were found. These changes were closely linked to previously defined zonation patterns, which contained distinct species assemblages and a unique dominant species per zone. These zones represented ecoregions based on diversity patterns of hyperiids and other shared species among such ecoregions suggesting a possible ecological connectivity among the zones, promoted by mesoscale eddies travelling westward from the coastal upwelling zone to offshore waters. Environmentally forced zonation and the interactions with mesoscale features are thus suggested as the driving processes maintaining spatial patterns of diversity of the hyperiid community in the Southeast Pacific.

Introduction

The Southeast Pacific (SEP) region is one of the largest, although poorly explored, ecosystems of the world’s oceans. In this region, a remarkable zonal gradient in productivity and oceanographic conditions can be observed from the eutrophic highly productive coastal zone, in the Humboldt Current system (HCS) (Strub et al. 1998; Schneider et al. 2007), to the oligotrophic/ultra-oligotrophic South Pacific central gyre (Raimbault and Garcia 2008). Along this gradient, the HCS stands out with an oxygen-deficient subsurface layer, known as the oxygen minimum zone (OMZ) (Paulmier and Ruiz-Pino 2009), which is a remarkable feature associated with high biological production. Westward, a mesotrophic coastal transition zone (CTZ) can be distinguished, exhibiting an intense mesoscale activity characterized by a variety of eddies mostly travelling northwest (Hormazabal et al. 2004) and further offshore the ultra-oligotrophic South Pacific anticyclonic gyre (SPG) with an extremely low primary production and phytoplankton biomass (Claustre and Maritorena 2003; Mcclain et al. 2004; Morel et al. 2007; Raimbault et al. 2008; Von Dassow and Collado-Fabbri 2014).

The contrasting properties of the water masses across the zonal gradient of the SEP may exert a strong influence on the distribution patterns of plankton and other organisms. For instance, based on remote sensing data and some in situ measurements along this gradient, it has been demonstrated that phytoplankton biomass may have values of chlorophyll-a (Chl-a) > 20 mg m⁻³ in the HCS in the upwelling season with a predominance of chain-forming diatoms (Vargas et al. 2007) whereas the presence of mesoscale eddies can contribute to patches containing ca. 0.25 to 1 mg Chl-a m⁻³ in the CTZ (Morales et al. 2010), while values of < 0.02 mg Chl-a m⁻³ can be found in the centre of the SPG with the dominance of pico- and nano-phytoplankton (Ras et al. 2008) although some slight increases in phytoplankton biomass with values of 0.03 to 0.05 mg Chl-a m⁻³ have been reported near Easter Island (110°W) (Pizarro et al. 2006).

Strong effects of the physical-chemical and biological gradients on the spatial distribution and structure of pelagic communities have been shown in micro-organisms (Masquelier and Vaulot 2008), epipelagic siphonophores (Palma and Silva 2006), and zooplankton size-classes (González et al. 2019). For some groups, the species and their distribution along this gradient have been described, such as for copepods within the HCS (Escribano et al. 2007) and the CTZ (Morales et al. 2010) although this group is poorly studied in the South Pacific central gyre (Yáñez et al. 2009; Von Dassow and Collado-Fabbri 2014).

A less-studied zooplankton group in the Southeast Pacific is hyperiid amphipods. In this region, some researchers studying this group have focused on particular zones, such as the South Pacific central gyre near Nazca and Salas and Gómez Islands (17° 35‵ to 26° 01‵ S, 80° 10‵ to 100° 50‵ E) (Vinogradov 1991) and the surface waters (0–100 m) of the coastal zone of Chile (Meruane 1980, 1982). However, large-scale distribution and abundance patterns of hyperiids in the SEP are unknown as in other oceans as well (Bowman and McGuiness 1982; Burridge et al. 2016).

The hyperiid amphipods are exclusively marine crustaceans with a distribution from the surface to abyssal depths (Vinogradov et al. 1996; Martin and Davis 2001). A rare exception seems to be Hyperia galba (Montagu, 1813), which may shift to a benthic mode of life when their host is scarce (Fleming et al. 2014). The ecological role of this diverse group is related to their predation impact over zooplankton (Auel et al. 2002; Dalsgaard et al. 2003; Yamada and Ikeda 2006; Pinchuk et al. 2013), and therefore acts as a link to higher trophic levels in polar and temperate regions (Dauby et al. 2003; Dalpadado et al. 2008; Weil et al. 2019). They are also known to be closely related to gelatinous zooplankton (Laval 1978, 1980; Phleger et al. 2000; Nelson et al. 2001; Nishikawa et al. 2005; Aoki et al. 2013; Gasca et al. 2015; Riascos et al. 2015). Hyperiids have been considered a suitable indicator for environmental changes linked to climatic variability at different temporal scales: e.g. interdecadal (Lavaniegos and Ohman 1999), interannual (Gasca et al. 2012; Valencia et al. 2013) and seasonal (Espinosa-Leal and Lavaniegos 2016). However, little is understood about how they interact with environmental variables in the ocean.

In the case of oceanographic conditions, primary production (e.g. spring phytoplankton bloom) (Abe et al. 2016), water mass characteristics and distribution (Shulenberger 1977a; Siegel-Causey 1982; Lavaniegos and Hereu 2009; Gorbatenko et al. 2017), circulation processes, including mesoscale activity (Shulenberger 1977b; Gasca 2003a, 2004; Lavaniegos and Hereu 2009), the presence of a strong thermocline (Cornet and Gili 1993), and an OMZ (Elder and Seibel 2015a, b) have been shown as the environmental factors that influence the hyperiid community.

In this work, the diversity and spatial distribution of hyperiid amphipods are assessed, covering a vast subtropical productivity gradient from the highly productive coastal upwelling zone off Chile up to the oligotrophic waters around Easter Island in the South Pacific subtropical gyre. Therefore, the study aims at testing the influence of a large-scale oceanographic gradient on the hyperiid community and how hyperiids respond to a changing ocean over such a large region. The expectation is that this analysis can provide insights on the controlling factors of species abundance, diversity, and hyperiid amphipods distribution in this ecologically important, but poorly known marine basin.

Material and methods

Environmental data

The CIMAR 21-Oceanic islands cruise was carried out between October 12 and November 11 of 2015 onboard the R/V AGS 61 “Cabo de Hornos” of the Chilean navy. The survey area was located between Caldera at the coastal zone of Chile (27° 00‵ S, 70° 52‵ W) and near Easter Island within the SPG (27° 10‵ S, 109° 20‵ W). The survey covered 33 oceanographic stations, as illustrated in Fig. 1. At each station, an oceanographic CTDO-Rosette was deployed to a maximum depth of 1500 m. From these casts, vertical profiles of temperature, salinity, and dissolved oxygen were obtained. These profiles were then utilized to build section profiles across the zonal transect using a weighted average gridding algorithm from the Ocean Data View software (Schlitzer 2018). The mixed layer depth (MLD) was estimated for each CTD station using a threshold value in temperature (ΔT = 0.2 °C) and density (Δρ = 0.03 kg m ⁻³) from a near-surface depth of 10 m, following the procedures of Kara et al. (2000) and de Boyer Montégut et al. (2004).

Fig. 1

The southeast Pacific region and the location of sampling stations during the CIMAR-21 cruise in the austral spring 2015

Monthly delayed-time sea level anomaly (SLA) from satellite data was used to represent the geostrophic velocity field for October and November 2015. These data were obtained from the Ssalto/Duacs AVISO altimetry product (http://www.aviso.altimetry.fr) in a spatial grid of 0.25° (~ 25 km) off central Chile region (20–30°S, 70–110°W). In addition, monthly satellite-derived surface Chl-a for the study region and for the same period were obtained from version 3.0 of the Ocean Colour Climate Change Initiative (OC-CCI, http://www.oceancolour.org/), at processing level 3 and a spatial resolution of 4 km. Using these surface Chl-a satellite data, and as suggested by González et al. (2019), the entire zonal gradient was divided into four major zones in terms of productivity: (a) eutrophic zone, representing the coastal upwelling zone (Chl-a > 0.5 mg m⁻³), (2) mesotrophic zone, corresponding to the CTZ (Chl-a = 0.1–0.5 mg m⁻³), (3) oligotrophic zone, representing a large offshore region (Chl-a = 0.05–0.1 mgm⁻³), and (4) ultra-oligotrophic zone which represented the South Pacific central gyre (Chl-a < 0.05 mg m⁻³). These four zones were thus characterized in terms of their oceanographic conditions, such that their hyperiid assemblages could be assigned and compared.

Zooplankton sampling and taxonomic analysis

Zooplankton samples were collected at 18 stations. Sampling was performed by means of oblique tows of Tucker Trawl net of either 8 m² (300 μm mesh-size) or 1 m² (200 μm mesh-size) opening diameter. The Tucker trawl gear consisted of three nets for stratified sampling: the first net was deployed to the maximal sampling depth so as to integrate the water column, the second net sampled the middle layer, and the third net covered the upper layer (Table S1). The volume of filtered water was estimated with a calibrated flowmeter (General Oceanic). Zooplankton samples were preserved with 4% formaldehyde buffered with sodium borate (Smith and Richardson 1977).

All individual amphipods were counted and sorted from the entire samples and identified using the taxonomic keys provided by Bowman and Gruner (1973), Shih (1991), and Zeidler (2004a, b, 2016). Because of the morphological similarities between Phronima stebbingi Vosseler, 1901 and Phronima dunbari Shih, 1991, and Lestrigonus schizogeneios (Stebbing, 1888) and Lestrigonus shoemakeri Bowman, 1973 caused difficulty of distinguishing species accurately, and thus data for these species were pooled for avoid erroneous identification (Lavaniegos 2017). Juvenile stages (individuals < 1 mm, and/or no secondary sexual characters evident according to Yamada and Ikeda [2001]) were excluded from this analysis. The hyperiid abundance was standardized as individuals per 1000 m³. In addition to sorting and identification of hyperiids, mesozooplankton biomass (ZB) and particularly the biomass of gelatinous zooplankton groups (salps, siphonophores, and other hydrozoans) for the upper 200 m layer were estimated utilizing digitized images (2400 dpi) obtained with a ZooScan Hydroptic and after applying the volumetric method, as described in Carlotti et al. (2008). Details for this analysis and results were reported in González et al. (2019).

Statistical analysis

In order to assess community changes across the zonal gradient, a multivariate analysis was performed with the species matrix (all species included) at all stations. For this, the Bray-Curtis similarity index was estimated, and species abundances were transformed to the fourth root prior to that. SIMPER analyses were performed to determine the percentage of contribution of species to the derived clusters. A further multivariable analysis was performed to test the effect of the predefined zones on the species matrix. For this, a similarity cluster analysis was conducted on the species abundance matrix including the zones as a factor. The level of similarity between stations was represented with a non-metric multi-dimensional scaling (NMDS) plot (Clarke and Gorley 2015). Differences between depth strata and zones for hyperiid abundance were tested with PERMANOVA test with a significance level of p < 0.05 (PRIMER-E v7) (Table S2).

The correlation between the hyperiid community and environmental conditions was tested using values of hydrographic variables at discrete depths (10, 30, 50, and 200 m) from the vertical profiles of temperature, salinity, and oxygen, and the MLD value. For SLA and satellite Chl-a, the closest pixel to the location of the sample was used. In stations with missing data, mean values from the two nearest stations were used. The relationships of these environmental variables with hyperiids were tested with the distance based in linear model (DistLM) and distance-based redundancy analysis (dbRDA) by zone. Before analysis, the environmental variables were normalized, and collinearity was tested using principal component analysis (PCA) in Primer-E software. The correlation between hyperiid species abundance and gelatinous group biomass (salps, siphonophores and other hydrozoans) was tested using Spearman correlation analysis (SYSTAT V-12.02).

Results

Oceanographic setting

Surface Chl-a revealed a strong zonal gradient in the study region, with a fall in Chl-a from the coastal zone towards the central gyre. In the Chilean coast, the Chl-a values ranged from 0.3 to 1.0 mg m⁻³ between 72°W to 74°W (stations 2 and 4), whereas Chl-a levels were lower (< 0.5 mg m⁻³) from 78°W to 82°W (stations 9 and 11). At ~ 90°W and westward, Chl-a values were < 0.1 mg m⁻³ reflecting an extreme oligotrophic condition and an ultra-oligotrophic condition at about 110°W (from station 23 to 76). Both months (October and November) exhibited similar Chl-a patterns (Fig. 2a).

Fig. 2

Monthly satellite data of a surface Chl-a (mg.m⁻³); and b sea level anomaly (SLA) across the four study zones: upwelling zone (Upwe), the mesotrophic zone (Meso), the oligotrophic zone (Oligo), and the ultra-oligotrophic zone (U-Oligo). Black dots denote the position of zooplankton sampling stations

The mean circulation pattern for both months was assessed by the geostrophic current field and the sea-level anomaly (SLA) (Fig. 2b). An intense mesoscale eddy activity was observed in the upwelling and the mesotrophic zones, between 74°W and 86°W, with a decreasing trend towards the most oceanic region. There was a remarkable presence of anticyclonic eddies from near the coast up to ~ 86°W with positive values of SLA (Fig. 2b).

The intensity of the zonal gradient of temperature, salinity, and dissolved oxygen strongly varied depending on depth (Fig. 3). For instance, temperature between the surface and 1500 m depth varied from 2.6 to 21.6 °C across the whole range, but the sea surface temperature (SST) gradient changed from 14.9 to 21.4 °C in the upwelling and ultra-oligotrophic zones, respectively (Fig. 3a). Sea surface salinity (SSS) showed values from 34.7 near the coast to 35.9 in the open ocean area. A salinity minimum (< 34.2) was observed in subsurface water, (ca. 200 m) offshore (station 4) and the upper 100 m in the mesotrophic zone. On the other hand, waters with 34.4 to 34.6 appear to prevail in the surface layer of the upwelling and mesotrophic zones, and high salinity water (> 35.5) dominated the upper 200 m in the oligotrophic and ultra-oligotrophic zones (Fig. 3b).

Fig. 3

Vertical and zonal gradient of a temperature (°C); b salinity (PSU); and c dissolved oxygen (mLL⁻¹) between Caldera and Eastern Island. Data corresponds to sampling CTD stations shown in Fig. 1

The surface layer along the zonal gradient was well oxygenated (> 5 mL L⁻¹) although the depth of the well-oxygenated layer varied drastically across the gradient, from a range of ~ 35 m near the coast to ~ 215 m in the offshore region. In the upwelling zone, the presence of the OMZ in the subsurface layer was a remarkable feature, almost reaching the near-surface layer in the nearshore while remaining below 200 m in the oligotrophic zone and horizontally extending up to near 84°W in the mesotrophic zone. At about 95°W (station 20), an apparent intra-thermocline eddy was observed with oxygen < 2.5 mL L⁻¹. Below 500 m depth from 110°W to 80°W, an intrusion of highly oxygenated water was observed from the oceanic area towards the coastal zone (Fig. 3c).

Regarding the variation in the mixed layer depth (MLD), it was found that in stations situated in the upwelling zone, the MLD had an average of 21 m depth, deepening to ~ 59 m depth in the mesotrophic zone, and even deeper in the oligotrophic zone (98 m). However, this layer showed a high variation among stations in the ultra-oligotrophic zone where it grew shallower (mean ~ 54 m depth).

When characterizing and comparing conditions among the predefined zones, it was found that the coastal upwelling zone (represented by station 2) had much cooler conditions than the other zones, with SST near 15 °C, SSS of about 34.7, surface oxygen ca. 5.6 mL L⁻¹, and a surface Chl-a of 0.56 mg m⁻³ in average. The mesotrophic zone, represented by stations 4, 9, 11, and 14 had an SST in the range of 16.3 to18.6 °C, SSS in the range of 34.7 to 34.8, surface oxygen range of 5.4 to 5.7 mL L⁻¹, and Chl-a in the range of 0.08 to 0.3 mg m⁻³. SST in the oligotrophic zone was in the range of 18.9 to 19.6 °C, SSS in the range of 35.2 to 35.5, surface oxygen range was 5.2 to 5.3 mL L⁻¹, and surface Chl-a was between 0.02 and 0.04 mg m⁻³. Finally, the ultra-oligotrophic zone was characterized by an SST in the range of 19.8 to 21.3 °C, SSS with values > 35.5, well-oxygenated surface water (5.1 to 5.3 mL L⁻¹), and surface Chl-a in the range of 0.01 to 0.02 mg m⁻³. One-way ANOVA resulted in highly significant differences (P < 0.01) among zones in SST, SSS, subsurface oxygen (200 m), MLD, and surface Chl-a, whereas non-significant differences among zones were found for surface oxygen, SLA, and current velocities (P > 0.05). Mesozooplankton biomass in the upper 200 m was in the range of 12.6 to 1691.6 mg C m⁻² and as described in González et al. (2019), it reached a maximum in the upwelling zone and then fell rapidly towards the offshore region with extremely low values detected in the oligo and ultra-oligotrophic zones. The one-way ANOVA and Tukey pairwise test showed significant differences in mesozooplankton biomass among zones (P < 0.05), except between the oligotrophic vs the ultra-oligotrophic zones which had similar levels (P > 0.05).

Species composition

A total number of 2132 specimens of hyperiid amphipods were counted and identified from all the samples from 17 families, 35 genera, and 69 species (Table 1). All of them belonged to the infra-order Physocephalata. The best represented superfamily was Platysceloidea with nine families, 20 genera and 35 species, and the least abundant was the superfamily Lycaeopsoidea with one family, one genus and two species. When considering the total of samples (17), 11 species were the most frequent (10–17 samples), whereas 49 species were considered rare (1–5 samples). On the other hand, on considering the predefined zones, 58 species were found in the mesotrophic zone, contrasting with 44, 22, and 18 in the ultra-oligotrophic, oligotrophic, and the upwelling zone, respectively (Table 1).

Table 1 List of species of hyperiid amphipods found between Caldera and Eastern Island off Chile. Density (Ind. 1000 m⁻³) and frequency with respect to total samples (17)

Spatial distribution

The predefined zones displayed significant differences in the abundance of hyperiids (2-way PERMANOVA P < 0.05) with no differences between depth strata (Table S2). According to this, the abundance (ind. 1000 m⁻³) of different depths strata was thus pooled by stations. Changes in total abundance of hyperiids by stations and zones, across the study gradient, are illustrated in Fig. 4a. The highest number of individuals (1278 ind. 1000 m⁻³) was found at station 2 in the upwelling zone, contrasting with the lowest abundance (0.10 ind. 1000 m⁻³) found in station 11 of the mesotrophic zone. When comparing the zones, the highest abundance was found in the ultra-oligotrophic zone (2517.27 ind. 1000 m⁻³).

Fig. 4

Total abundance (ind. 1000 m⁻³) of a hyperiids by stations and predefined zones; and b the 10 most abundant species across the zonal gradient. The abundance of the pooled species is shown as Lestrigonus schizogeneios/L. shoemakeri (Lestrigonus schi/shoe), and Phronima stebbingi/ P. dunbari (Phronima stebb/dun)

The ten most abundant species across the zones were represented in Fig. 4b. In this case, Hyperioides longipes Chevreux, 1900 had an abundance of 1159. 25 ind.1000 m⁻³, followed by Eupronoe minuta Claus, 1879 (895.64 ind.1000 m⁻³), Hyperioides sibaginis (Stebbing, 1888) (434.68 ind.1000 m⁻³), Primno latreillei Stebbing, 1888 (382.92 ind.1000 m⁻³), Primno brevidens Bowman, 1978 (369.20 ind.1000 m⁻³), and Eupronoe armata Claus, 1879 (359.78 ind.1000 m⁻³). Eight species showed an abundance from 103.65 to 294.44 ind.1000 m⁻³. On the other hand, low values were found in 45 species (1.82 to 93.49 ind.1000 m⁻³), while eight species demonstrated abundance of < 1 ind.1000 m⁻³.

To compare the changes in community structure among zones, the relative abundance (%) of the most abundant species by zones was utilized. For this, the finding was that in all cases, 5 to 10 species could account for more than 70% of total abundance at each zone (Fig. 5). It is important to note that species assemblages as represented by dominant species varied substantially between zones. For instance, in the upwelling zone Eupronoe minuta, Primno brevidens, and Vibilia armata Bovallius, 1887 dominated with 75% of total abundance. In the mesotrophic zone, the dominant species were Primno latreillei, Lestrigonus schizogeneios/L. shoemakeri, and Hyperioides longipes with 44% of total abundance. In the oligotrophic zone, Eupronoe armata, Eupronoe minuta, and Primno latreillei dominated with 39%. Finally, the ultra-oligotrophic zone was dominated by about 50% with only two species, Hyperioides longipes and Hyperioides sibaginis. Moreover, it was found that a single dominant and distinct species could characterize each zone: Eupronoe minuta (42%), Primno latreillei (26%), Eupronoe armata (19%), and Hyperioides longipes (38%) for the upwelling, mesotrophic, oligotrophic, and ultra-oligotrophic zones, respectively.

Fig. 5

Relative abundances (%) of the most abundant species by pre-defined zones. The selected species are a combination of the five species with the highest jrelative abundance in each zone. The abundance of the pooled species is shown as in Fig. 4

Community descriptors

Based on the species matrix, the highest species richness (37 species) and diversity (Shannon-Winer index, H′ = 2.65) were found at station 14 located in the mesotrophic zone. Eupronoe laticarpa Stephensen, 1925, Amphithyrus bispinosus Claus, 1879, Lycaea pulex Marion, 1874, Glossocephalus milneedwardsi Bovallius, 1887, and Oxycephalus clausi Bovallius, 1887, were found exclusively at this station. Additionally, another eight species showed the highest abundance for the study area. In contrast, the lower diversity (H′ = 0.94) of the station 67 (ultra-oligotrophic zone) was underlined by the presence of only four species (Themistella fusca Bovallius 1887, Hyperioides longipes, Hyperioides sibaginis and Hyperietta stephenseni Bowman, 1973), being important considering that this station had the shallowest sampling depth (0 to 80 m) of the study. On the other hand, the highest dominance (Simpson index, D = 0.90) was found in the stations 18 (oligotrophic zone) and 53 (ultra-oligotrophic zone) (Fig. 6). In these stations, a total of 23 species were found of which 12 were shared species among zones. One-way ANOVA, however, indicated that none of these descriptors changed significantly across zones (F_3,14 < 1.5, P > 0.05), indicating the absence of a zonal pattern in diversity according to these community descriptors.

Fig. 6

Total hyperiid amphipod species (R), amphipod diversity (Shannon-Winer index (H′), and dominance (Simpson index (D) by station across the onshore-offshore gradient

Similarity analysis

Similarity analysis based on species abundance (ind.1000 m⁻³) resulted in two significant clusters (Simprof test, p < 0.05), indicating considerable mixing of species between zones even though a significant zonation can be assumed. Species contribution (70%) in terms of abundance to the clustering indicated that cluster A (Average similarity 51.27%) was accounted by the station of the upwelling zone and two stations of the mesotrophic zone (Fig. 7). Vibilia armata (11.18%), Eupronoe minuta (10.95%), Primno brevidens (10.37%), Lestrigonus shizogeneios/L. shoemakeri (9.01%), Themistella fusca (8.63%), Tryphana malmi Boeck, 1870 (8.29%), Lycaeopsis zamboangae (Stebbing, 1888) (7.82%), and Hyperietta stephenseni (7.64%) were the species that contributed with most similarity in this group. The cluster B (Average similarity 54.33%) mixed the station of the oligotrophic and the ultra-oligotrophic zones, and its similarity was represented by Hyperioides longipes (15.5%), Eupronoe minuta (11.74%), Phrosina semilunata Risso, 1822 (11.15%), Hyperioides sibaginis (8.49%), Hyperietta stephenseni (7.06%), Phronima stebbingi/P. dunbari (6.81%), Hyperietta luzoni (Stebbing, 1888) (6.79%), and Primno brevidens (6.04%).

Fig. 7

Shade plot of abundance for hyperiid amphipods resulting from SIMPER analysis. Linear-scale intensity is proportional to the fourth root transformation of abundance. Stations have been grouped using the Bray-Curtis similarities derived from transformed data by the hierarchical % cluster mode average group, and the SIMPROF tests to identify the A and B groups, whereas the cluster C is used for grouping the non-selected stations by the SIMPROF test. Species are also clustered, with the standard agglomerative method, based on the ‘index of association’. The abundance of the pooled species is shown as in Fig. 4

Finally, the stations excluded by the analysis (grouped in Fig. 7 as cluster C) showed different characteristics. In the station 11, a high number of species (21), but low abundances were noted (total abundance 0.1 ind.1000 m⁻³). At station 67, only four species were found but the abundance of Hyperioides longipes (107.46 ind.1000 m⁻³) was remarkable, contrasting with station 14 which showed a high number of species (23) and total abundance (538.62 ind.1000 m⁻³).

In order to assess whether the pre-defined zones could indeed represent significantly different species assemblages, the effect of the zones on the similarity analysis was evaluated by grouping the stations for each zone (Fig. 8). The dissimilarity among zones (SIMPER test) resulted in upwelling vs mesotrophic = 54.04%, upwelling vs oligotrophic = 57.94%, upwelling vs ultra-oligotrophic = 59.53%, mesotrophic vs oligotrophic = 43.91%, mesotrophic vs ultra-oligotrophic = 41.06, and oligotrophic vs ultra-oligotrophic = 37.46%.

Fig. 8

Non-metric multidimensional scaling graph (NMDS) based on the Bray-Curtis similarity index on transformed abundance data (fourth root) associated with the four pre-defined zones: upwelling zone, the mesotrophic zone, the oligotrophic zone, and the ultra-oligotrophic zone

Hyperiid assemblages and environmental conditions

According to a principal components analysis (PCA), the temperature at 30 m depth and salinity at 30 m and 50 m depth showed significant collinearity with their surface values (R² = 0.93; 0.99; 0.92, respectively) and were thus excluded from the analyses. Thereafter, on assessing the influence of environmental variables on the hyperiid community structure across the entire gradient, the DistLM test (marginal test) revealed that every individual variable, except for oxygen (30 and 50 m depth), temperature (50 m), salinity (10 m), and the MLD had a significant relationship (adjusted R² = 0.38) (Table S3; Fig. 9a).

Fig. 9

Distance-based redundancy analysis (dbRDA) plot of the selected environmental variables best explaining the variation of the hyperiid abundance: a across the study area; b the mesotrophic zone (Meso); c the oligotrophic zone (Oligo); and d the ultra-oligotrophic zone (U-Oligo). The dbRDA was obtained from the best-fit explanatory variables after a multivariate multiple regression analysis (DistLM, marginal test) shown in Table S3

In contrast, the same analysis by zone showed that in any case, the environmental variables had a significant correlation with the hyperiid abundance (Table S3). However, the values that explain the greatest variance and the best selection of the model were different in each zone. Even though the upwelling zone was not tested, it was observed that ZB and oxygen at 10 m depth could be important sources of variance for the hyperiid in this zone (Fig. 9a). In the mesotrophic zone, the oxygen at 30 m in depth and the MLD were selected, while the SLA was selected in the oligotrophic zone, and finally, in the ultra-oligotrophic zone the oxygen (at 30 and 50 m depth), temperature (50 and 200 m depth), the ZB and the MLD were excluded from the analysis. The adjusted R² for each zone were 0.59, 0.32, and 0.30, respectively (Fig. 9b–d).

The correlation analysis among the total abundance of hyperiids and gelatinous groups, across the trophic gradient, was not significant (p > 0.005). However, the same analysis by species showed significant correlations (p < 0.005) between the abundance of 22 hyperiid species and at least one of the gelatinous groups. Most of the significant correlations were with the category other hydrozoans. Themistella fusca and Vibilia armata showed a significant correlation with the three groups. Brachyscelus crusculum, Simorhynchotus antennarius, and Lycaea lilia were correlated positively with salps and siphonophores (Table 2).

Table 2 Spearman correlation between the abundance of hyperiid amphipod species and gelatinous groups biomass

Discussion

Diversity and spatial distribution of hyperiid amphipods were assessed over an extensive zonal gradient of trophic and oceanographic conditions in the southeast Pacific. Regarding the hyperiid community structure, all the encountered species in this study have a tropical and subtropical distribution (Vinogradov et al. 1996; Zeidler 2004b), and they were previously reported from different areas of the South Pacific basin (Meruane 1980, 1982; Vinogradov 1991). The number of species found was comparable to that reported by Burridge et al. (2016) along the Atlantic Ocean with 37 shared species. The species, Lestrigonus crucipes Bovallius, 1889, reported from the study zone by Meruane (1980, 1982) and Eupronoe maculata Claus, 1879, reported by Vinogradov (1991) were not found in this study. These two species exhibit a tropical affinity (Vinogradov et al. 1996) and are more frequently observed in the Atlantic ocean (Cornet and Gili 1993; Burridge et al. 2016).

When deepening the analysis of community structure, it became clear that species composition changed through the zonal gradient, and that the most remarkable variations were the substitution of dominant species, as well as the replacement of various species from one zone to another. However, the lack of significant variation in species abundance and the community descriptors (species richness, diversity and dominance) across this gradient, with no significant differences among zones, may deserve further discussion. Community descriptors, such as diversity indexes, cannot account for species replacements, as long as species richness and diversity remain stable. In fact, the community analysis showed that each of the pre-defined zones contained different species assemblages, also represented by a unique numerically dominant species, Eupronoe minuta, Primno latreillei, Eupronoe armata and Hyperioides longipes for the upwelling, mesotrophic, oligotrophic and ultra-oligotrophic zones, respectively. Also, these species (except Eupronoe armata) were the most abundant and frequent species of hyperiid amphipods identified from the SEP. High abundances for Eupronoe minuta and Primno latreillei coincided with previous report from subtropical gyres of the Atlantic and North Pacific ocean (Shulenberger 1977a; Burridge et al. 2016); Hyperioides longipes is an abundant species in the eastern Indian Ocean (Bowman and McGuiness 1982), and Eupronoe armata is a common species of oceanic, tropical and subtropical waters (Tranter 1977; Zeidler 1984).

It was further demonstrated that the structure of the hyperiid assemblages changed across the zones. Particularly, in the upwelling zone, the assemblage comprising Eupronoe minuta, Primno brevidens, and Vibilia armata has been associated with coastal upwelling waters of the California Current in the zone off Baja California (Lavaniegos 2009; Espinosa-Leal and Lavaniegos 2016). Vibilia armata is also a highly abundant species in other eastern boundary upwelling ecosystems (Thurston 1976; Cornet and Gili 1993) and in the Indian Ocean upwelling system (Tranter 1977). By contrast, the assemblage of the ultra-oligotrophic zone (Hyperioides longipes and Hyperioides sibaginis) seems similar to that in other studies, and it has been associated with tropical and oligotrophic waters (Shulenberger 1977b; Vinogradov 1991; Gasca 2007; Burridge et al. 2016). However, contrary to the tendency of the studies in other zones, in the SEP Hyperioides longipes was more abundant than Hyperioides sibaginis. According to Shulenberger (1978), when Hyperioides sibaginis is absent, Hyperioides longipes can extend their vertical range to the upper layer and expand their population.

A comparison of the assemblage of species found in the oligotrophic and the ultra-oligotrophic zones with that reported by Vinogradov (1991) for the SPG resulted in a similar species list, except for Phronimopsis spinifera Claus, 1879, Vibilia australis Stebbing, 1888, and Primno macropa Guérin-Méneville, 1836, which were found in our study and should now be considered new records for the study region. In the SPG, the most abundant species were considered scarce in the zones in which, Phronima atlantica Guérin-Méneville, 1836, had an abundance of 0.17 and 0.07% in the oligo and ultra-oligotrophic zones, respectively, while Phronimella elongata was absent in the oligotrophic zone and showed an abundance of 2.88% in the ultra-oligotrophic zone. In contrast, Anchylomera blossevillei Milne-Edwards, 1830, whose abundance was low in the oligotrophic zone (1.90%), showed a similar abundance in the ultra-oligotrophic zone and SPG (6% in both cases). Particularly, the abundant species Hyperioides longipes and Hyperioiodes sibaginis were characterized as rare in SPG.

Hyperiid assemblages and oceanographic conditions

The hyperiid community distributing across the trophic gradient and within each of the pre-defined zones may result from the impact of contrasting changes in oceanographic conditions across the area, as previously shown for other zooplankton groups (e.g. Palma and Silva 2006; Riquelme-Bugueño et al. 2012). For instance, positive correlations between temperature and salinity (for the entire gradient and the ultra-oligotrophic zone) indicated the changes in the composition and dominant species associated with subantarctic water (SAAW) in the upwelling zone (Reid 1973; Silva et al. 2009) and subtropical water (STW) (Schneider et al. 2007; Fuenzalida et al. 2008) in the oceanic areas of the SEP. This association had been described for euphausiids, copepods, and siphonophores (Palma and Silva 2006; Morales et al. 2010; Riquelme-Bugueño et al. 2012). In this study, the presence and high abundance of Eupronoe minuta, Primno brevidens, and Vibilia armata suggest the influence of SAAW, while those of Primno latreillei, Hyperioides longipes, Hyperioides sibaginis, and Eupronoe armata indicated a shift to influence of STW, possibly associated with mixing processes.

Despite the changes in species composition among zones, the presence of shared species among zones appears as an important feature when comparing these areas. The positive correlation of the hyperiid community with SLA across the entire gradient (Table S3; Fig. 9a) and the oligotrophic zone (Fig. 9c), where a propagation of cyclonic eddies was observed (Fig. 2), suggested that the mesoscale activity in the SEP is an important process favouring the dispersion, advection, mixing and a possible ecological connectivity of hyperiids among zones. The importance of mesoscale processes influencing the transport and mixing of the population has been demonstrated for phytoplankton and other zooplankton groups (Hormazabal et al. 2004; Correa-Ramirez et al. 2007; Morales et al. 2010; Riquelme-Bugueño et al. 2015), and it may thus explain the notable presence of Eupronoe minuta, Primno brevidens, and Lestrigonus schizogeneios/L. shoemakeri across all zones.

Other studies have also proven that distributional patterns of hyperiids are related to mesoscale circulation (Young and Anderson 1987), as they can form species assemblages linked to mesoscale eddies (Gasca 2003a, b; Lavaniegos 2009). Indeed, in this study, near the presence of a cyclonic eddy (Fig. 2) located in the mesotrophic zone, a second peak in richness and abundance of the community was observed (Fig. 6). This peak was also associated with the highest abundance of the whole study for Primno latreillei, and the presence of a group of exclusive species (Amphityrus bispinosus Claus, 1879, Glossocephalus milneedwardsi, Lycaeopsis themistoides, Oxycephalus clausi and Paratyphis maculatus Claus, 1879).

An additional oceanographic feature to consider is the oxygen minimum zone (OMZ). In the HCS, the influence and significant role of the OMZ for the community structure and vertical distribution have been described for copepods (Hidalgo et al. 2005; Hidalgo and Escribano 2008; Escribano et al. 2009), euphausiids (Escribano et al. 2000; Antezana 2009), and chaetognaths (Giesecke and González 2004); therefore, the positive correlation among the hyperiids and oxygen concentration in the upwelling zone and mesotrophic zone was an expected result. In hyperiids, few studies had reported the correlation between the OMZ and abundance (Zhang et al. 2014), and the influence of this low-oxygen layer has been shown for the species Phronima sedentaria in the Eastern Tropical North Pacific (ETNP) (Elder and Seibel 2015a, b).

Among biological environmental factors influencing the community structure of hyperiids, positive relationships were found with zooplankton biomass (ZB) and Chl-a, across the zones, and particularly ZB in the upwelling zone. ZB can indicate that the food resource is an important variable influencing the community structure of hyperiids across the trophic gradient in the SEP. In this context, it has been demonstrated that contrasting biogeochemical environments found in these zones can affect the supply of nutrients and induce changes in the microbial community (Masquelier and Vaulot 2008; Valdés et al. 2017) and ultimately the zooplankton composition (González et al. 2019).

Although hyperiids are considered mostly carnivorous (Bowman and Gruner 1973), which can feed on prey ingested by their host or on their tissues (Phleger et al. 1998, 2000; Nelson et al. 2001), they may change their food diet according to ontogeny (Sugisaki et al. 1991; Phleger et al. 1998; Richoux 2011), or alternate their preys depending on their availability (Pakhomov and Perissinotto 1996; Froneman et al. 2000; Yamada and Ikeda 2006; Kraft et al. 2015). This behaviour is expected, considering that changes in the productivity can affect the zooplankton biomass and the potential prey for the hyperiids across the zones in the SEP. Food variation should thus be considered a critical component of the environmental conditions determining the ecological zonation and the composition of the assemblages of hyperiids.

Another biological component associated with hyperiids is the presence of gelatinous zooplankton, since the distribution of both groups seems closely related (Laval 1978, 1980). However, in this study, no significant correlations between total hyperiid abundance and the biomass of the gelatinous zooplankton were found. Nevertheless, the abundance of several hyperiid species was significantly correlated with the biomass of other hydrozoans in the study area (Table 2). In other studies, Lestrigonus bengalensis have been reported in association with the leptomedusa Eirene pyramidalis (Agassiz, 1862) (Madin and Harbison 1977) and also with salps and siphonophores (Lima and Valentin 2001), and Brachyscelus crusculum had been associated with the leptomedusa Aequorea coerulescens (Gasca and Haddock 2004), the salp Metcalfina hexagona (Quoy & Gaimard, 1824) and the siphonophore Rosacea cymbiformis (Delle Chiaje, 1830) (Gasca and Browne 2018). Simorhynchotus antennarius has been reported in association with the siphonophore Sulculeolaria quadrivalvis Blainville, 1830 (Lima and Valentin 2001). Vibilia armata has been found in association with only different salps species (Laval 1980; Gasca et al. 2015) and a siphonophore (Lavaniegos and Ohman 1999). The analysis in this study showed a positive correlation between some species and a gelatinous group previously not reported. For example, the usual host for the genus Phronima are salps (Gasca et al. 2015), but in this study, the significant correlation was with other hydrozoans; whereas the association record for Hyperioides longipes had been with siphonophores (Laval 1980), and these species were associated with salps. Also, other positives correlations were found for some hyperiid with gelatinous groups, which had not been previously described, e.g. Themistella fusca, and Hyperioides sibaginis. However, these spatial correlations may not necessarily mean a functional relationship, because of the natural aggregation of the entire zooplankton community, and therefore trophic or symbiotic studies are required to assess true interactions between hyperiids and specific groups of gelatinous organisms.

Conclusion

This study is a contribution to the knowledge of biodiversity and spatial distribution of hyperiid amphipods in the subtropical Southeast Pacific onshore-offshore gradient of productivity, from the eutrophic upwelling zone off Chile and Peru to the extremely oligotrophic region of the Southeast Pacific central gyre. Community variation across the gradient showed zonation patterns, characterized by distinct species assemblages and a dominant species, coinciding with the predefined zones: (1) Eupronoe minuta, Primno brevidens, and Vibilia armata in the eutrophic coastal upwelling zone, (2) Primno latreillei, Lestrigonus schizogeneios/L. shoemakeri, and Hyperioides longipes in the mesotrophic coastal transition zone, (3) Eupronoe armata, Eupronoe minuta, and Primno latreillei in the oceanic-oligotrophic zone, and (4) Hyperioides longipes and Hyperioides sibaginis in the South Pacific gyre or ultra-oligotrophic zone. These zones are suggested to represent ecoregions based on changing species assemblages of hyperiids. The shared species between the ecoregions reflected a possible level of ecological connectivity between zones forced by mesoscale activity. Environmentally forced zonation, including prey field variation, and interaction with mesoscale circulation are thus suggested as the driving processes maintaining spatial patterns and shaping up the zooplankton community structure in this large Southeast Pacific basin.