Zooplankton sampling was carried out in the oceanic province of the southern Gulf of Mexico aboard the R/V “Justo Sierra” in two contrasting seasons, winter 2013 (24 oceanographic stations from January 25 to February 3) and summer 2014 (31stations from June 4 to 14) ( Figure 1 ; Supplementary Table S1 ). At each station, temperature and salinity were recorded before the zooplankton sampling with a CTD sonde (Sea-Bird SBE 9) from the surface to 1000 m depth. Zooplankton samples were obtained using a stratified cylinder-conical zooplankton multinet system (nylon, 75 cm mouth, 2 m length, and 500 µm mesh size); this system consisted of a Double -Trip Mechanism (General Oceanics 1000-DT), whose purpose is the opening and closure of plankton nets. At each oceanographic station, samples were obtained at five depth levels of the water column: 0–200 m; 200–400 m; 400–600 m; 600–800 m; and 800–1000 m, depending on bottom depth. Sampling was performed for approximately 25 minutes following double oblique tows at a speed of approximately 3 knots; each net was equipped with a standard mechanical flowmeter (General Oceanics 2030R) to calculate the filtered water. The oceanographic stations were arranged in five longitudinal transects for each season (transects Bw to Fw in winter, and As to Es in summer), but their latitudinal positions slightly differed ( Figure 1 ). The time of day (day/night) at which each sampling was carried out was recorded. A total of 255 samples were obtained from which 104 were taken in winter and 151 in summer. Zooplankton samples were fixed in a 4% formalin-seawater solution and, posteriorly, preserved in 70% ethanol because of the objectives of the project, concerning one, the study of fish larvae growth throughout the analysis of otoliths, which are best preserved in alcohol (Campana, 1989).

In the laboratory, the zooplankton biomass of each sample was estimated by the displacement volume method and data were expressed as mL/1000 m³ (Postel et al., 2000). Besides, all the amphipods were sorted from samples identified by performing microdissections of the structures of taxonomical interest under a stereoscopic microscope (Leica M80), following specialized literature (Vinogradov et al., 1996; Vinogradov, 1999; Zeidler, 2003a, 2003b, 2004a, 2004b, 2006, 2009, 2016; Hughes and Lowry, 2015; Violante-Huerta et al., 2020). The abundance data of each sample were standardized to 1000 m³ of filtered water (ind/1000 m³).

The hydrological conditions in each sampling season were characterized by the use of θ-S_A diagrams. To create these diagrams, conservative temperature (θ) and absolute salinity (S_A) values were calculated following the TEOS-10 equation (McDougall and Barker, 2011). The water masses were identified in the θ-S_A diagrams according to the latest classification carried out for the Gulf of Mexico (Cervantes-Díaz et al., 2022). Water masses were located in the vertical plane by plotting longitudinal transects of the density (Sigma) to make comparisons with the temperature variations. Because greater variability was found at the first two levels, only values from 0 to 400 m depth were plotted. This analysis was done using Ocean Data View v5.6.5 software. We also calculated the mean integrated value of each parameter at each sampling level to explore their association with the biological data.

For each sampling season, the species/stations matrix was transformed by applying a square root to smooth the bias of dominant species, and after that, the Bray-Curtis similarity index was applied. Then, to probe the differences in the vertical plane, the amphipod community was analyzed from two points of view: first, considering two assemblages defined a priori, the ‘epipelagic’ and the ‘mesopelagic’; and second, considering each of the five sampling levels as separate groups.

The ‘epipelagic’ assemblage was represented by the first sampling level of 0-200 m depth, and the ‘mesopelagic’ assemblage was composed by the four sampling levels between 200-1000 m depth. For each season, a nMDS analysis was applied to the similarity matrix, and the environmental variables (zooplankton biomass, temperature, salinity) were also plotted as bubbles to represent their influence on the structure of the assemblages. The correlation between the biological and environmental matrices was evaluated using a BEST-BIOENV test based on Spearman’s rank correlation coefficient (Clarke and Ainsworth, 1993). Besides, to test the difference between the two assemblages, an ANOSIM hypothesis test (9999 permutations) was performed. The R ANOSIM test is a non-parametric method based on ranks used to determine if two or more groups are statistically different. The R values near 1 indicate differences among groups, whereas values near zero suggest more similarity among them. Furthermore, a SIMPER analysis was also used to determine the taxa with the greatest contribution to the separation of assemblages. These analyses were carried out with PRIMER v7 software (Clarke et al., 2014). The amphipod abundance data were plotted on vertical transects of temperature using the SURFER v15 software.

The two assemblages were also compared in terms of diversity. In both seasons, we used species accumulation curves of alpha diversity with interpolation-extrapolation of Hill Numbers and the incident-based estimator Chao2 (Chao and Jost, 2012). In addition, we estimated the sample completeness to infer the representativeness of the sampling effort of each assemblage (Chao and Jost, 2012), considering a common sampling coverage value of 0.95 to compare the alpha diversity. This integrated methodology yielded less biased comparisons between a set of communities by the use of an equal sample coverage (completeness) value (Chao and Jost, 2012). Diversity analysis was performed with the statistical software R using the iNEXT package (Hsieh et al., 2016).

In the second approach, considering the five sampling levels as a factor, the similarity matrix of each season was subjected to a metric MDS using Bootstrap to reduce the stress in the two-dimensional plane. This analysis generates average values and regions of each factor (levels) to identify trends in data sets and improve their graphic representation in the ordination plane, through randomization of the data with replacements (Jacoby and Armstrong, 2014). Furthermore, the average abundance of the entire community, as well as the three main groups of species (infraorder Physocephalata, infraorder Phyososomata, and non-hyperiid species) was plotted by sampling level to visualize the differences between day- and nighttime. An ANOSIM hypothesis test (9999 permutations) was used to determine whether the five sampling levels and the day- and nighttime average abundance had statistically significant differences. Particularly, for this finer analysis, we compared the diversity (by completeness) between levels using the common sampling coverage value of 0.85 to interpolate the alpha diversity.

The θ-S_A diagrams allowed us to recognize the five main water masses in the Gulf: the Caribbean Surface Water (CSW), the Subtropical Underwater (SUW), the Gulf Common Water (GCW), the Tropical Atlantic Central Water (TACW), and the Antarctic Intermediate Water (AAIW) ( Figure 2 ). The greatest variations of temperature and salinity occurred above 200 m depth and the main difference between the seasons was due to temperature (θ): values around 29°C were recorded in most surface waters during summer, while in winter they did not exceed 25°C. Salinity (S_A) had a variation of 2 units from the surface to the depth ( Figure 2 ).

The vertical transects showed that the characteristic water masses of the epipelagic zone (CSW and GCW) occurred above 200 m depth, with some depth variations along each transect during the winter and the summer ( Figures 3 , 4 ). Regarding the temperature, the values between 15 and 18°C, associated with the TACW water mass, occurred deeper in some areas, especially during the summer in the northernmost transect, where these values were observed up to 300 m depth ( Figure 4 , Transect As).

The metric MDS analysis with bootstrap performed taking the sampling levels as factors, separated the 0-200 m level from those located below 200 m depth for both seasons ( Figure 9 ). During the winter, levels of the mesopelagic zone were highly homogeneous (ANOSIM test, p > 0.05) ( Figure 9 ). In contrast, during the summer, the pair-R values of the 200-400 m level with the other levels of the mesopelagic zone are close to zero indicating certain affinity; however, the p-value showed significant differences of the 200-400 m level with two mesopelagic levels ( Table 4 ). Thus, the mesopelagic zone is more heterogeneous in summer than in winter. SIMPER analysis indicated that Stenopleura atlantica was the species with the greatest contribution to the separation between the 200-400 m level and the other deep levels of the ‘mesopelagic’ assemblage during this season ( Table 5 ).

Records of zooplankton biomass and mean amphipod density were higher in summer, showing a decreasing pattern with depth ( Table 6 ). Considering day and night sampling time, no differences were observed in the whole amphipod density between the stations sampled in the upper 200 m layer in both seasons ( Figure 10A ), and this was confirmed by the ANOSIM test (winter: global R = -0.08, p = 0.927; summer: global R = 0.068, p = 0.072). However, looking the amphipod community as separate groups (Physocephalata, Phyososomata, and non-hyperiids) we found that, during the summer, the infraorder Physosomata showed significant differences (R = 0.598, p = 0.001) in the 0-200 level between the night- (~25 ind/1000 m³) and daytime (~3 ind/1000 m³) ( Figure 10C ). In contrast, the infraorder Physocephalata and the non-hyperiids had no significant differences (ANOSIM test, p > 0.05) ( Figures 10B, D ) in any of the seasons.

The species accumulation curves of the alpha diversity among sampling levels indicated that species richness was highest in the 0-200 m level (69 in winter; 88 in summer) and gradually decreased with depth during the two seasons ( Figures 11A, B ). In general, sampling coverage values of the levels were ≥ 0.85 indicating a good representativeness of their sampling effort ( Figures 11C, D ). Interpolation of the species richness to a completeness value of 0.85 indicated that the highest diversity was in the 400-600 m level during winter with 50 species ( Figure 11E ), and in the 200-400 m level during summer with 60 species ( Figure 11F ).

The mesopelagic zone of the oceans has been hardly explored. Ecological and diversity information of zooplankton in this zone is very scarce. In this study, the coarse-scale vertical analysis demonstrated that the assemblages defined a priori ‘epipelagic’ and ‘mesopelagic’ showed significant differences (ANOSIM test, p < 0.05), being stronger in winter. This result contradicts the hypothesis that the summer would have the greatest difference due to stronger gradients in environmental conditions, especially temperature. Paradoxically, the deepening of 15-18°C temperature values in summer, related to the TACW water mass, makes the upper mesopelagic zone (up to 300 m depth) more similar to the epipelagic zone ( Figure 4 , transect As) than in winter. In turn, this could induce a deepening of the epipelagic assemblage during the summer, resulting in less heterogeneity between the epi- and the mesopelagic zones.

As expected, the BEST-BIOENV test signaled the zooplankton biomass (as a measure of food availability) as an important factor affecting the distribution of the amphipods in the water column, especially during the summer ( Table 2 ). In winter, the temperature and salinity conditions gained importance, but the best three models involved the zooplankton biomass ( Table 2 ). Despite being surrounded by land masses, the Gulf of Mexico exhibits similar features to oligotrophic regions of the great oceans in terms of zooplankton biomass (Landry and Swalethorp, 2021). While seasonal changes in secondary productivity over the shelf are marked and strongly dependent on continental water discharges, in the oceanic area of the southern Gulf, productivity is lower and seasonal changes less pronounced than over the shelf; even though, in both zones, the summer exhibits the highest zooplankton biomass (Zavala-García et al., 2016; Färber-Lorda et al., 2019). As stated, pelagic amphipods tend to inhabit areas of high productivity (Gasca, 2004; Hereu et al., 2020), where they can easily find their food. In this study, amphipod density in the epipelagic zone was at least ten times higher than the mesopelagic zone in summer, and four times in winter ( Table 6 ). The zooplankton biomass showed a similar pattern among the seasons. Following our results, other studies found the greatest abundance of pelagic amphipods during the summer. Besides the food availability, during this warm season occurs a strong reproductive activity of some amphipods. For instance, species of Lestrigonus and Primno have recorded a high abundance of juveniles and adults in this warm period in several world oceans (Gasca, 2004; Gasca and Suárez-Morales, 2004; Zhang et al., 2014; Hereu et al., 2020).

According to the SIMPER analysis, the species with the greatest contribution to the separation of the assemblages ‘epipelagic’ and ‘mesopelagic’ were: L. bengalensis during winter, and juveniles of the genus Primno during summer ( Table 3 ). Lestrigonus bengalensis is widely recognized as the dominant species in the Gulf of Mexico (Gasca, 2003a, 2003b, 2004; Gasca et al., 2009; Hereu et al., 2020; Sanvicente-Añorve et al., 2023) and this is probably the reason why it had the greatest contribution to the separation of both assemblages during winter. This species has a wide vertical range (>1000 m depth) (García-Madrigal, 2007), which allows the individuals to inhabit the mesopelagic zone; however, its dominance is restricted to the epipelagic zone as observed in the current survey ( Table 1 ) and other regions of the world (Thurston, 1976b; Shulenberger, 1978; Siegel-Causey, 1982; Vinogradov et al., 1996; Vinogradov, 1999). During summer, L. bengalensis was not the dominant species, nevertheless, its abundance was more than double that recorded in winter ( Table 1 ), confirming the affinity of the species with the warm season, as previously observed in the Gulf (Gasca, 2004; Hereu et al., 2020).

The species of the genus Primno are commonly found in the Gulf (LeCroy et al., 2009), and their vertical distribution range reaches 1000 m depth (Vinogradov et al., 1996; Vinogradov, 1999). In the Gulf, juveniles of the genus Primno have been observed as a dominant taxon during the summer, suggesting greater reproductive activity during the warm season (Gasca, 2004; Hereu et al., 2020). This could be the reason why the taxon dominated in the summer season ( Tables 1 , 3 ).

The analysis performed taking the sampling levels as a factor showed differentiation of the 200-400 m level in the mesopelagic zone during the summer ( Figure 9 ; Table 4 ). This could be explained by the deepening of the TACW water mass ( Figure 4 , transect As), which separates the second level (200-400 m) from the other deep ones ( Figure 9 ; Table 4 ). In a recent study, Cervantes-Díaz et al. (2022) indicated that the deepening of water masses in the southern Gulf could be related to the noticeable presence of the CSW water mass during this warm season. However, the effect of this hydrological feature on the vertical structure of zooplankton communities had not been previously explored. According to the SIMPER analysis, a non-hyperiid species (Stenopleura atlantica) had the greatest contribution to the separation of the levels of the mesopelagic zone during the summer ( Table 5 ), due to its dominance in the upper mesopelagic zone (200-400 m). In the vertical plane, S. atlantica is found from 0 to 1000 m depth, however, the highest abundance of adults, juveniles, and ovigerous females occurs above 200 m depth, which confirms the affinity of the species to inhabit the warm waters of the epipelagic zone (Thurston, 1976a; Vinogradov, 1999; Violante-Huerta et al., 2020). In ecological terms, pelagic amphipods of the suborder Hyperiidea are the most studied due to their high abundance and diversity; however, our results suggest that a better understanding of the vertical distribution patterns of amphipods could be attained by including non-hyperiid species, such as S. atlantica.

Daytime differences in total amphipod abundance were not observed in any season ( Figure 10 ), and this is consistent with the last survey in the Gulf (Hereu et al., 2020). The diurnal vertical migration range of the pelagic amphipods is generally less than 100 m (Thurston, 1976b; Shulenberger, 1977), so it can only be observed at finer sampling scales (Cornet and Gili, 1993; Pai et al., 2010; Domínguez-Nava et al., 2021; Schaafsma et al., 2024). Vertical migration has been related to some preferences for occupy a specific depth range during the day- or nighttime or even as a response to physiological requirements (Elder and Seibel, 2015a, 2015b; Taniguchi et al., 2023). In this study, the broader interval of sampled levels likely makes it difficult to observe this general migratory pattern. However, in the finer analysis made by groups, we found that during the summer, the abundance of members of the infraorder Physosomata showed significant differences (ANOSIM test, p < 0.05) in the 0-200 level between the night- and daytime samplings ( Figure 10C ). These organisms are mainly inhabitants of the meso- and bathypelagic zones (Vinogradov et al., 1996; Vinogradov, 1999), so their greater presence in the upper layer (0-200 m) during the night could be indicative of a migratory process, as other field observations found (Thurston, 1976b; Granata et al., 2020; Domínguez-Nava et al., 2021; Véliz et al., 2021). The daily migration range of the Physosomata species is not well known, but most probably our findings correspond to species that inhabit the upper mesopelagic zone (e.g. Scina, Acanthoscina) (Thurston, 1976b), which can reach the adjacent epipelagic zone during their daily vertical movements.

This study recorded the greatest number of species in the Gulf of Mexico to date ( Table 1 ). Previous studies in the Gulf have only examined the epipelagic community of amphipods, recording an average of 61 species, with a maximum value of 79 species (Gasca, 2003a, 2003b, 2004; Gasca et al., 2009; Hereu et al., 2020; Sanvicente-Añorve et al., 2023). The large number of species recorded here was due to the examination of the previously unexplored mesopelagic zone; indeed, this region recorded the greatest interpolated values in species ( Figures 11E, F ). The high diversity in the mesopelagic zone was related to the occurrence of some exclusive species from the deep waters, especially those belonging to the infraorder Physosomata ( Table 1 ). Species of the infraorder Physosomata (Hyperiidea) inhabit mainly the mesopelagic waters (Thurston, 1976b; Vinogradov et al., 1996; Vinogradov, 1999), which explains the observed pattern in this study.

Seasonally, interpolations to a common sampling coverage value showed that the highest species richness was found during summer ( Figure 8F ). Our findings partially agree with a previous study conducted in the oceanic zone of the southern Gulf in winter and summer (Gasca et al., 2009), which found the highest species richness in summer, but associated with the lower amphipod density. This implies that more research is required because the seasonal cycles of abundance and diversity in the study area are not well understood.

When exploring the diversity by sampling level, the interpolation indicated that the greatest diversity was found in the ‘mesopelagic’ assemblage: in the 400-600 m level in winter, and the 200-400 m level in summer ( Figures 11E, F ). A deep-sea diversity peak was previously observed in other zooplankton groups and has been related to the diurnal vertical migration of species (Angel, 1989; Andersen et al., 2001; Steinberg et al., 2008). A recent study revealed that copepods increased their diversity in the deeper mesopelagic zone, probably due to less competition in that environment with limited food resources (Stefanoudis et al., 2019). Other environmental features related to the greater diversity in the deep sea could be a lower risk of predation, more stable hydrological conditions, or even the occurrence of different water masses (Kosobokova and Hopcroft, 2010; Kosobokova et al., 2011). However, these statements are difficult to probe due to the complexity and methodological replication to obtain the sampling representativeness in this extreme environment (McClain and Schlacher, 2015; Paulus, 2021).

In the Gulf of Mexico, the mesopelagic zone is considered an “ecotone” of tropical, subtropical, and temperate species (Sutton et al., 2017), as evidenced by the presence of water masses of different origins ( Figure 2 ). Our results demonstrated that the greatest diversity lies in this scarcely explored area, suggesting the need of further studies to better understand the ecology of the deep-sea pelagic amphipods of the Gulf.