Partitioning of the GoM based on a multi-locus assessment of the zooplankton community
Both the results of the multivariate regression and the PCA suggest a clear spatial and temporal segregation of zooplankton that was mainly explained by oxygen, temperature, and longitude gradients. The importance of longitude in the model suggests that additional unexplored predictors related to this variable may also play key roles in shaping the composition of the zooplankton community. In this sense, we did not consider the longitudinal spatial gradient as a source of ecological patterns per se but as a proxy of still unrevealed environmental predictors that may explain a portion of the variability in the zooplankton community (Gluchowska et al., 2017; Hawkins & Diniz-filho, 2004). Our results suggest that the physical characteristics of the water column may support the occurrence of at least three heterogeneous ecoregions within the studied area. Such ecoregions comprise the north, south, and an eastern area around the Yucatan peninsula. Furthermore, the characteristics of the water profiles of each region may determine the distribution of at least some zooplankton species, resulting in the aforementioned structural segregation. Accordingly, the main empiric boundaries inferred in this study may be located around 86 °W and 22 °N, which coincide with the presence of the LC (eastern boundary) and with the southern edge of quasi-permanent cyclonic and anticyclonic eddies (north-south boundary), respectively.
So far, few efforts have described the pelagic communities across the entire GoM; however, our proposed partitioning strongly supports earlier insights based on both chlorophyll concentrations and water mass dynamics. In this context, Damien et al. (2018) proposed the first partitioning of the GoM along 22 °N based on chlorophyll concentrations, while Sheinbaum et al. (2016) estimated that among 60–80% of the horizontal variance of the Yucatan Channel is directly and/or indirectly related with the LC. The proposed GoM partitioning was clearly evident during the summer expeditions (cruises XIXIMI-04 and XIXIMI -06), whereas an alternative oxygen-temperature partitioning was observed and related with seasonal water dynamics during XIXIMI-05. The water analysis of this cruise revealed the occurrence of two anticyclonic eddies that crossed 22 °N, creating a north-south boundary and willing the GoM into a diagonal south-west/north-east thermal configuration (Fig. 6). As this boundary was detected once, we cannot state if this water organization may be considered a usual season-specific configuration or an anomaly, and more research is needed to resolve this issue.
Finally, the zooplankton in the GoM seems to form a more stable community in the southern region (south of 22 °N) compared to that of the northern region. Indeed, stations from sampling Lines F, G, H, and J showed low structural variability over the 3 years of observations. This is likely explained by the higher productivity observed in this section of the GoM compared to that of the southern region. Indeed, the northern region is under the influence of a quasi-permanent cyclonic gyre, which has been associated with higher nutrient concentrations, fluorescence, and productivity compared to that of the southern region (Färber Lorda, Athié, Camacho Ibar, Daessle, & Molina, 2019; Linacre et al., 2015; Pérez-Brunius, García-Carrillo, Dubranna, Sheinbaum, & Candela, 2013). The northern region is also influenced by upwelled waters from the Bank of Campeche (Salmerón-García et al., 2011).
Temporal zooplankton variability may be also related with oxygen, temperature, and longitude gradients. Although low environmental variability was observed during the study, the abundance of the same taxa, such as Mysidae, Hormathiidae, Euphausiidae, and Calanidae, likely varied according to the co-occurrence of certain environmental conditions. An analysis of the distributions of single taxa was not directly addressed in this paper since our goal was to study the temporal effects of abiotic factors on the entire zooplankton community. However, the integration of single taxa into the analysis is essential to fully understand the functioning of the GoM ecosystem. Their representation in the model is critical and requires a set of taxon-specific DNA libraries that were not available when this study was conducted; however, the development a finer taxonomic scale is currently underway.
Detecting the spatio-temporal variability in the zooplankton community in this study relied on the information of a multi-locus metabarcoding approach. Notably, despite the aforementioned observed taxonomic discrepancies among the two loci employed in this study, similar zooplankton distributions were obtained for both 18S and COI. This suggests that the results obtained strongly reflect the natural variability of the zooplankton community, and methodological bias (e.g., implemented molecular markers) is likely to have only marginally affected the results. It is probable that the main limitation associated with using the combined information of COI or 18S rRNA markers lies in the incompleteness of reference databases and that not all zooplankton species have reference sequences deposited for either the COI or 18S rRNA markers (Larke et al., 2017; Stefanni et al., 2018). However, in contrast to previous studies that recommended 18S rRNA as the most suitable marker for surveying zooplankton communities (Zhan, Bailey, Heath, & Macisaac, 2014), our insights suggests that COI provide similar taxa coverage of zooplankton phyla with higher taxonomic resolution, as suggested by Machida et al. (2017). Finally, our results support that COI may provide better taxa classifications for lower taxonomic levels compared to those of 18S rRNA (Larke et al., 2017; Stefanni et al., 2018). In this study, we observed that 18S rRNA presented a relatively high affinity to Calanoida and Euphausiacea while COI provided relatively uniform taxonomic coverage (L. J. Clarke, Soubrier, Weyrich, & Cooper, 2014; Deagle, Jarman, Coissac, Pompanon, & Taberlet, 2014). Due to the taxonomic complementarity of those loci, we suggest that most comprehensive assessments of zooplankton biodiversity may be conducted using a synergistic multi-locus approach.