Hyperiid amphipods from the Gulf of Ulloa and offshore region, Baja California: The possible role of the gelatinous zooplankton as a transport vector into the coastal shelf waters

Hyperiid amphipod species from the Gulf of Ulloa, Baja California, and the adjacent region (from the shelf break to 200 km offshore) were analyzed to evaluate diversity and abundances. This productive area supports small-scale commercial fisheries, including sand bass (Paralabrax nebulifer), California spiny lobster (Panulirus interruptus), abalones, clams, and others. Strong coastal upwelling events were observed during summer seasons of the period 2002–2008 between Punta Eugenia and Punta Abreojos. The upwelling plumes at Punta Abreojos are transported southward in slope waters bordering the coastal shelf of the Gulf of Ulloa, contributing to the separation of coastal and oceanic regions, and explain differences in amphipod diversity and abundances between both regions. In the offshore region, the most abundant species were Vibilia armata, Lestrigonus schizogeneios, Primno brevidens, and Eupronoe minuta, similar to previous findings in northern regions of Baja California and southern California. However, abundances of these species were lower (10–30 individuals/1000 m3), only reaching 20–50% of abundance levels reported off northern Baja California. In the coastal shelf of the Gulf of Ulloa, amphipods were virtually absent during 2002, 2003 and 2006. However, during 2004 and 2005, abundances of P. brevidens increased (54 and 20 ind/1000 m3, respectively). Moreover, during the late summer of 2007, abundances of L. schizogeneios, P. brevidens, Lycaea nasuta, Lycaea pulex, and Simorhynchotus antennarius increased considerably (261, 39, 31, 68, 416 ind/1000 m3, respectively), indicating occasional utilization of the coastal shelf by pelagic amphipods. Changes in gelatinous populations (medusae, siphonophores, ctenophores, doliolids, and salps) paralleled changes in hyperiid populations, with highest abundances in 2005–2008 in the coastal shelf. Significant correlations of 17 amphipod species with gelatinous taxa, which are often used as host organisms by hyperiid amphipods, suggest that gelatinous presence enhanced hyperiid abundance and promoted the progression of hyperiid amphipods onto the coastal shelf during parts of the 2002–2008 period.


Introduction
The zooplankton community has been intensively studied in the northern regions of the California Current System (CCS), but off Baja California it has received less attention, particularly in terms of taxa such as the hyperiid amphipods that comprise only a small proportion of the community and are therefore assumed to have minimal ecological importance. However, this perception may be incorrect for oceanic waters, where they are relatively abundant [1][2][3][4] and may represent attractive forage food for predators in the California Current such as small pelagic and myctophid fishes and seabirds [5][6][7]. Similar observations have been made in other eastern boundary upwelling ecosystems (EBUS), including the trophic roles of the amphipod Themisto gaudichaudi in the Benguela Current [8] and Hyperia curticephala in the Humboldt Current [9]. In the eastern tropical Pacific, some flying fishes and myctophids prey on hyperiids [10,11]. Hyperiid amphipods have maximum abundances in high latitudes [12][13][14][15], while in tropical and subtropical regions they have low abundance but higher species diversity [16,17]. However, more studies are required to quantify these crustaceans and their roles more precisely in trophic webs in tropical and subtropical seas.
Seasonal changes in hyperiid amphipod species assemblages have shown a strong coupling with upwelling dynamics off Oregon [18] and the southern sector of the CCS (southern California [19] and Baja California [1][2]). Lowest abundances occur in winter, followed by increases in spring-summer and maximum levels in autumn. However, hyperiids appear to avoid the zone of highest upwelling activity, as suggested by their relative scarcity in inshore waters compared to higher concentrations in the core and offshore part of the California Current [1,18]. The oceanic abundances of hyperiids (<1 to 100 ind 1000 m-3 per species) are 50% higher than in the coastal shelf region [1]. The difference is striking because the main zooplankton species responsible for secondary production in Vizcaino Bay increase in neritic waters. The copepods Calanus pacificus and Acartia tonsa reach up to 138,000 and 60,000 ind 1000 m -3 respectively [20], and the euphausiid Nyctiphanes simplex reaches 35,000-87,000 ind 1000 m -3 [21,22]. Mesoscale structures, particularly eddies, may contribute to concentrating the high phytoplankton productivity in the Gulf of Ulloa [23], where grazer organisms as copepods and euphausiids feed while amphipods are more carnivorous [24,25]. The region also displays seasonal changes in circulation, with enhanced nearshore poleward flow from July to October [26]. This pattern is also observed in other EBUS [27] and is responsible for transporting oceanic fauna onto the shelf areas. As a consequence of such dynamic circulation, productivity is high, similar to other EBUS [28,29], making the region a hotspot for whales [30], sea turtles [31,32], tuna [33], and swordfish [34], which currently supports a successful artisanal coastal fishery out of Punta Abreojos [35,36]. Understanding the zooplankton community dynamics in this area can help improve the management of these small-scale fisheries.
In order to expand our understanding of zooplankton community and mechanisms influencing their variability in upwelling ecosystems of subtropical latitudes comparable to the Gulf of Ulloa and offshore region, the present research describes interannual variability in summer species composition of hyperiid amphipods during 2002-2008. The study specifically explores (1) the distribution and abundance of hyperiid amphipods, by assessing interannual variability in species assemblages; and (2) investigating the correlation between hyperiids and gelatinous zooplankton. This study presents evidence of intermittent occupation by hyperiids of the coastal shelf in the Gulf of Ulloa, Baja California, and possible mechanisms underlying this behavior.

Study area
The study area is located off southwest Baja California between Punta Eugenia and Cabo San Lazaro and is characterized by a relatively wide coastal shelf (Fig 1). The coastline is oriented NW-SE, favorable for upwelling [37], which is further reinforced by the narrow step-shelf between Punta Eugenia and Punta Abreojos. Farther south, the coastal shelf broadens to form the embayment named the Gulf of Ulloa (GU). Upwelling events induced by trade winds (Easterlies) are particularly frequent during spring and summer [38]. The lowest sea surface temperature (SST) occurs in February-April (18˚C) while the maximum is recorded in late summer (25˚C), associated with a seasonally enhanced poleward current [26]. Poleward flow produces a decrease in chlorophyll during summer months, contrasting with the high

Sampling
The sampling stations were arranged in four transect-lines perpendicular to the coast (Fig 1) from seven cruises performed every summer from 2002-2008 by the IMECOCAL program (Spanish acronym of: Mexican Investigations of the California Current). At each station, hydrographic data were recorded with a CTD (Seabird Electronics Inc., 9/11). Zooplankton sampling was carried out by oblique tows of a bongo net (71 cm-diameter, 505 μm mesh width) with a digital mechanical flowmeter. In the epipelagic oceanic region, tows were conducted from 0-210 m, while at stations located in the coastal shelf the tow depth was from the surface to 10-15 m above to the bottom. The samples were preserved with 4% formalin.
Only the nighttime samples were selected to perform taxonomic analysis of hyperiid species in the offshore region to reduce variability due to vertical migration, as has been reported for some hyperiid species [see 39,40]. However, for the coastal shelf (bottom depth < 200 m) daytime samples were also included due to the low number of stations. The cruise dates and number of samples analyzed are shown in Table 1 (details of sampling stations in S1 Table).
In the present study, hyperiid species occurring in >55% of samples are considered dominant, those in 31-55% are common, those in 10-30% are sparse, and those in <10% are rare, following definitions from [46]. The geometric mean (GM) is used instead of numerical mean because it is less affected by particularly small or large values and data distribution skew.

Statistical analysis
To provide a context of seasonal changes in amphipod abundance and distribution, data from all seasons in 2005 were included to produce a full annual cycle. For interannual analyses (summer hyperiids during 2002-2008), zooplankton stations were separated into onshore or Table 1. IMECOCAL cruises with dates and number of zooplankton samples used in taxonomic identification. Cruise dates correspond to the area considered in the present study (see Fig 1). offshore regions, as demarcated by the 200 m isobath (Fig 1). For each region, tnterannual comparisons of the most abundant species were done with the Kruskal-Wallis test. To determine the specific years with differences, a posteriori comparison with the Tukey test was made. Interannual comparisons were also made for gelatinous zooplankton groups. Multivariate cluster analysis was used to address community structure across years. Summer 2007 was excluded in this analysis due to sampling bias (September cruise, compared to Jul-Aug cruises in other years). To explore amphipod community grouping by stations, hierarchical cluster analysis was performed with STATISTICA 7.1. Euclidian distances were calculated using transformed data (log10 [x-+1]) to weight the contribution of abundant and rare species. The data matrix consisted of 70 species × 83 samples, after exclusion of 19 rare species (occurring in only one sample) and 7 samples from coastal stations without amphipods. Clusters were defined with the Ward´s linkage hierarchical method [47,48]. Further, analysis of similarities (ANOSIM) was done to test the hypothesis of differences between clusters based on a resemblance matrix of Bray-Curtis index with the software PRIMER-7. The resemblance matrix was also used to estimate percentages of similarity (SIMPER) to determine the contribution to similarity of each species in the clusters [49].

Cruise
The potentially symbiotic relationship between hyperiid species and gelatinous zooplankton groups was analyzed using Spearman correlation analysis of all summer samples (N = 101).

Physical environment
Vertical sections of temperature along Line 130, extending from Punta Abreojos to 250 km offshore (Fig 2a-2g), showed strong stratification during summer. Sea surface temperatures were slightly lower in 2002, 2005, and 2008, ranging from 19-22˚C offshore and 18-20˚C inshore.
Maximum SST values were recorded in 2007 (22-23˚C) because sampling was conducted in September (Fig 2f), when SST increases compared to mid-summer, the period when all other years were sampled. The upwelling footprint throughout the study period was more evident in station 130.35, slightly farther offshore, than in the most coastal station 130.30. Despite low SST at Station 130.35, temperature at 100-200 m depth was consistently higher than the rest of stations in Line 130 due to the influence of the California Undercurrent, transporting warm Pacific Equatorial Water poleward.
A high salinity core in stations 130.35 and 130.40 also indicates the influence of the California Undercurrent (Fig 2h-2n). The tilting of the isohalines is pronounced closest to the coast, indicating intense upwelling activity. In July 2002 and again in summers 2007 and 2008, highsalinity upwelled water (>34 psu) reached the coastal shelf, but from 2003-2006 low-salinity water in the upper layer masked upwelling.

Distribution and abundance of hyperiid amphipods
First examining seasonal changes in hyperiid amphipods in 2005, abundances were low in February (< 200 ind/1000 m 3 ) but homogeneously distributed across the GU and offshore region (Fig 4a). Low abundances persisted in spring, apart from three stations with 326-541 ind/1000 m 3 , located in the coastal shelf, the shelf break, and oceanic region (Stns. 127.35, 127.36, 133.55, respectively). A regionwide increase in abundance occurred in oceanic waters in July, while low abundances persisted in the neritic region. By October 2005, high abundance (> 500 ind/1000 m 3 ) was observed near the coast and slope waters (Fig 4a). Therefore, hyperiid amphipods showed a gradual increase from winter to fall in the offshore region (Fig 5a) while the increase in the GU was limited to autumn (Fig 5b). However, the number of samples from the coastal shelf during May and October 2005 was too low to establish a more robust conclusion.
Turning to interannual comparisons of summer abundances, total hyperiid amphipods varied widely in samples, from 0 to 3,732 ind/1000 m 3 . Except for 2007, hyperiids collected at GU had low abundance (< 150 ind/1000 m 3 ) and were absent in seven samples ( (Fig 4b).
The Kruskal-Wallis test comparing summer abundances across years in the offshore region was significant, and summer 2003 was notably lower than all other years (H = 27.6, p < 0.001).  Interannual variability of hyperiid species. Hyperiid amphipod diversity was high in the study region: 91 species were recorded in the 2002-2008 period (S2 Table). Eighty-eight of these species occurred in the oceanic region, while 56 were in neritic waters. There were five dominant species in the whole region (frequency 58-80%; abundance GM 4-15 ind/1000 m 3 ). All dominant species pertained to the infraorder Physocephalata, with one in the superfamily Vibiloidea (Vibilia armata), two in the superfamily Phronimoidea (Lestrigonus schizogeneios and Primno brevidens), and two in the superfamily Platysceloidea (Eupronoe minuta and Simorhynchotus antennarius).
Sparse species (frequency 10-30%) occurred with GMs of 0.1-0.9 ind/1000 m 3 . They included one species from Physosomata (Scina borealis) and 32 from Physocephalata, spread over the superfamilies Vibiloidea (12%), Phronimoidea (28%), and Platysceloidea (59%). The complete list of species is shown in S2 Table. Almost all sparse species were missing in one or A high number of species were categorized as rare (<10% in samples): 7 from Physosomata and 36 from Physocephalata. Their global GM was below 0.1 ind/1000 m 3 (S2 Table). The low abundance of some rare species is probably due to their meso-and bathypelagic distributions, meaning they are rarely captured in the upper layer. Such is the case for Lanceola clausi, Scypholanceola aestiva, and Scina curvidactyla [16]. Most of the rare species only occurred in a single year (51%) and some occurred in only a single sample (42%). The year with the lowest number of rare species was 2003, and the highest was 2007, at 5 and 51% respectively. The remaining years showed values between 23-35%. Most of the rare species were found in the offshore region (72%); only 7% occurred exclusively inside the GU, and the remaining 21% occurred in both regions.
The differences among the six clusters, including groups and subgroups, were confirmed with ANOSIM, which had a global R = 0.654 (p = 0.001). The pairwise tests between clusters also showed significant differences in all cases (p < 0.007).
The clusters considered coastal showed a strong contrast in species composition between Cluster A2 and the other two clusters (B1 and B2, Fig 9). Besides different species composition, Cluster A2 had low amphipod abundance. This cluster was dominated by samples from 2006 (Fig 8). Clusters A2 and B2 showed the highest contribution to similarity by P. brevidens, with 95.9 and 51.3% respectively (Fig 9b). However, the high percentage of this species in cluster A2 was due to a low diversity (22 species with abundances <0.6 ind/1000 m 3 ); P. brevidens also had low abundance (GM = 6 ind/1000 m 3 ) compared to the other coastal clusters. In contrast, the assemblage B2 had abundant P. brevidens (GM = 53 ind/1000 m 3 ) and included33 other species, though only six (E. minuta, H. medusarum, L. schizogeneios, V. armata, Phronima sedentaria, and S. borealis) were moderately abundant (1-7 ind/1000 m 3 ). The assemblages of hyperiid species in clusters A2 and B2 represent typical neritic communities since they consist of data from various years, even though B2 was more characteristic of 2004 and 2005 (Fig 8).
In the offshore region, the interannual separation within different clusters was also clear but slightly different than for the neritic region (Figs 7-9). The lowest amphipod abundance occurred in cluster A1, characteristic of 2003, with dominance by V. armata, Platyscelus serratulus, L. bengalensis, and E. minuta (GM of 12, 3, 2, and 2 ind/1000 m 3 respectively). The main contribution to similarity was from V. armata (45.6%). P. brevidens, important in other clusters, was absent in cluster A1, and L. schizogeneios had low contribution to similarity (4.6%).
The offshore community from 2004, represented in cluster C1, was distinguished from other clusters by its high abundance of A. blossevillei (GM = 17 ind/1000 m 3 ), which together with L. schizogeneios, E. minuta, and V. armata contributed 38.8% to similarity (Fig 9). Although diversity was high in cluster C1, cluster C2 had a higher number of species (44 in C1 In contrast, during September 2007 certain species not observed or scarce in the summer cruises analyzed above (e.g., Lycaea pulex, L. nasuta Lestrigonus bengalensis, V. stebbingi) were abundant in neritic waters (Fig 10). Dominant species such as S. antennarius, and L. schizogeneios also reached high abundances (GM 416 and 261 ind/1000 m 3 respectively). In the offshore region, S. antennarius and L. schizogenios were also dominant but an order of magnitude lower (GM of 59 and 87 ind/1000 m 3 respectively).

Correlation of hyperiids and gelatinous plankton
Interannual changes in gelatinous organism abundances paralleled changes in hyperiid populations ( Table 3). The low hyperiid abundances observed during 2003 corresponded with low abundances of medusae, ctenophores, and doliolids in the offshore region. Abundances of medusae and doliolids were significantly different in the coastal shelf than offshore (Fig 11). Differences in abundances of siphonophores and salps are also significant at p < 0.05. Low siphonophore abundance occurred in 2002 but remained at consistent levels from 2003-2008  Table 3). In the offshore region, ctenophores were the least abundant gelatinous group throughout the study period but increased in 2005, 2007, and 2008 compared to 2002, 2003 and 2006 (Table 4). No significant differences were observed for ctenophores in the GU.
Tunicates presented significant interannual differences in the GU, but tendencies were different for doliolids and salps (Fig 11b, Table 3) Correlation analysis between gelatinous groups and the 50 most frequent hyperiid species returned significant for 17 species (p < 0.001, Table 4; 24% with dominant, 35% common, and 41% sparse species). Notably, L. vespuliformis presented significant correlations with all five gelatinous groups. Four other species (L. schizogeneios, O. clausi, Platyscelus ovoides, and S. antennarius) correlated positively with three gelatinous groups, but the remaining species only correlated with one or two gelatinous groups (Table 4). The highest number of significant correlations (p<0.001) was with medusae and ctenophores (56% of the total). These included 4 species correlated with medusae (V. stebbingi, L. nasuta, L. pulex, and P. ovoides), 3 species with ctenophores (S. tullbergi, E. minuta, and T. malmi), and 7 species correlated with both medusae and ctenophores (Hyperoche medusarum, L. vespuliformis, L. schizogeneios, P. brevidens, O. clausi, and S. antennarius). There was no apparent taxonomic preference for one of the two groups, and all superfamilies had one or more species correlated with both medusae and ctenophores (Table 4).

PLOS ONE
Onshore-offshore changes in hyperiid amphipods from the Gulf of Ulloa Scinoidea had some correlation with tunicates ( Table 4). The siphonophores only presented one significant correlation, with Laxohyperia vespuliformis.

Interannual variability and the weak El Niño events
The region located south of Punta Eugenia is the CCS region most strongly influenced by tropical biota. Some tropical species are resident, and some appear in the region during El Niño   32˚N), [3]. The low amphipod abundance found in the GU could be related to El Niño events, which appeared to have a higher impact south of Punta Eugenia, based on changes in euphausiid populations [52]. Comparing the interannual variability observed off north Baja California [3], changes at the species level were out of phase between the northern and southern Baja California regions. For example, the most abundant species, L. schizogeneios, had a GM of <1 ind/1000 m 3 in the present study for the oceanic region during summer 2003, while off northern Baja California its GM was 75 ind/1000 m 3 [3]. The same situation occurred for P. gracilis, with high abundance (GM = 77 ind/1000 m 3 ) in the northern region during 2003 [3] but low abundances off the GU (0.6 ind/1000 m 3 ). The abundance of P. gracilis is in general low off the GU compared to northern Baja California (<3 ind/1000 m 3 across the period 2002-2008). The low abundance during 2003 off the GU was also observed in eleven other species (Table 2), suggesting a more severe effect of El Niño 2002Niño -2003 in the southernmost region of the CCS [50,51].
It is interesting to note that the low hyperiid abundance recorded in 2003 south of Punta Eugenia was similar to the decrease that occurred during 2005 off northern Baja California [3]. In both cases, decreases were due largely to decreased abundance of L. schizogeneios and to a lesser extent P. brevidens. The question of why these decreases occurred in different years from region to region may be answered by the occurrence of two short El Niño events, one in Jun 2002-Feb 2003 [53] and other in Jul 2004-Apr 2005 [54]. Both El Niño events were weak, but they evolved differently. The development of El Niño 2002Niño -2003 was combined with a subarctic water intrusion coming from the north, which cooled the northern region off Baja California but not the southern region. The presence of a cyclonic eddy in the latitude of 27-29˚N [50,55] retained the cool subarctic water, helping to maintain population densities of typical California Current species north of Punta Eugenia [51].
In contrast, El Niño 2004Niño -2005 had no simultaneous subarctic water intrusion. El Niño 2004-2005 appears to have affected all regions of the CCS mainly in autumn and winter [54,56], with different regions experiencing these impacts at different times within those seasons. El Niño 2004Niño -2005 first affected the GU during summer 2004. Species such as A. blossevillei and L. vespuliformis, with modest abundances in northern Baja California [3], have relatively more importance in the hyperiid community off the GU. A. blossevillei was particularly abundant in July 2004, comprising 11% of total hyperiids in that year. This species was also found in high abundance in the northern Baja California region but in January 2005 [1], associated with oligotrophic oceanic waters. Because the warm anomalies associated with El Niño spread from south to north, there was more evidence of this event during July 2004 in the oceanic region off the GU (present study) and in January 2005 off northern Baja [1]. El Niño 2004Niño -2005 was not a predicted event [54] and is still under discussion for its occurrence and magnitude of surface water advection [57], but the hyperiid findings from the present and previous studies [1,3] suggest a south-to-north advection of A. blossevillei.
There are no studies documenting the effects of El Niño on hyperiids or other zooplankton during 2004-2005. However, off Oregon, high proportion of warm-water neritic copepods were noted in the summer 2004, attributed to northward advection of warm water [58]. El Niño effects on higher trophic levels in the eastern Pacific are better studied and appear to relate to availability of food. For example, California sea lions (Zalopus californicus) performed longer foraging trips into oceanic waters during summer 2004 [59]; the Galapagos fur seal (Arctocephalus galapagoensis) migrated to the coast of Ecuador in September 2004 [60]; and the intertidal black sea urchin (Tetrapygus niger) recorded dietary changes between August 2004 and January 2005 [61].

Cross-shelf differences in species assemblages and La Niña
Hyperiid amphipods present clear cross-shelf differences, as was established first for the California Current off Oregon [18], and further confirmed off Baja California [1]. In the present study, the influence of coastal upwelling off Punta Abreojos was consistently observed during all summers between 2002-2008 (Figs 2 and 7). Upwelling plumes spread to the south in the oceanic region and, due to the topography of the area, produce an inshore-offshore split, isolating the GU, which maintains warmer temperature. Summer enhancement of the poleward coastal current also increases temperatures inside the GU [26]. The upwelling front appears to be an effective barrier against the entry of amphipods from the open ocean to the coastal shelf.
The However, the increase of amphipods in the GU and the offshore region during summer 2007 could also result from a seasonal effect due to sampling in later summer (one month later than the rest of summers analyzed in the present study). Samples from 2007 likely show processes more typical of autumn or late summer, when there is a seasonal increase in hyperiid abundance and diversity off Baja California [1,2]. The trend toward increasing abundance during autumn was first observed in inshore waters off Oregon for some species (Hyperoche medusarum, Themisto pacifica, and Hyperia medusarum) in 1963, and for P. gracilis in 1965 and 1966 [18]. However, Hyperia medusarum increased during winter 1967 [18]. In Sagami Bay, Japan, maximum abundance during the year occurred in September for 15 of 25 species, while six species increased in November [39]. Alternatively, both the seasonal and interannual influences could be the cause of increased hyperiid abundance during 2007 in the GU.
Hyperoche medusarum and L. schizogeneios, as well as other species of the families Hyperiidae and Lestrigonidae, have many records of hyperiid-host associations, mainly involving medusae and ctenophores [63]. This is consistent with the positive correlations of H. medusarum, L. vespuliformis, and L. schizogeneios with medusae and ctenophores in the present study. L. vespuliformis is the only one of these three species without previous records of symbiotic associations, because it is a new species recently described [16] from the northeast Pacific near the study region. L. vespuliformis has been found in a few other locations, all in tropical-subtropical latitudes [1,3,17,41]. This species was considered rare [41], but in the present study it occurred frequently and showed correlations with all five gelatinous organisms. Therefore, it is highly probable that L. vespuliformis, which is morphologically like H. medusarum, may have a symbiotic association with one or more gelatinous species, as suggested by the results of the current study.
Other abundant species in the GU during 2007 were L. pulex and S. antennarius, which both showed significant correlation with medusae and salps; S. antennarius also showed correlation with ctenophores. L. pulex has been reported in association with diverse salp species [71]. S. antennarius has only two symbiotic records with hydromedusae, one with Geryonia proboscidalis in the Mediterranean [42], and the other with Liriope tetraphylla in Monterey Bay, California [64]. Both medusae species were present in the GU during the summer of 2007 (S2 Table).

Conclusions
In conclusion, the study region is less populated with hyperiid amphipods compared to northern Baja California. The transition zone species present in northern Baja California and other northern sectors of the CCS were also present in the GU and offshore region, but their abundances declined, particularly for P. brevidens and P. gracilis, while S. antenarius increased. During summer, contrasting cross-shelf differences in species assemblages were observed with high interannual variability. Active upwelling off Punta Abreojos forms a plume separating the inshore and offshore regions, preventing entry of amphipods to the coastal shelf. Changes in temperature and the proliferation of gelatinous organisms during La Niña 2007-2008 promoted the occupancy of the GU by hyperiids.