Fish Larval Assemblage Associated to an Eastern Tropical Pacific Coral Reef: Seasonal and Interannual Variability

1. Introduction

The Eastern Tropical Pacific (ETP) has unique environmental characteristics for marine life. Specifically, the abiotic conditions in the region were modulated by the combined effects of seasonal and regional fluctuations in local environmental and oceanographic factors across a range of spatial and temporal scales [1,2,3] The Colombian Pacific Ocean (CPO) is located in the easternmost sector of the ETP, and there are different types of marine-coastal ecosystems, including rocky shores, sandy beaches, estuaries, mangrove forests, and coral reefs, which support a wide variety of fish with a wide range of life history strategies [4,5].

Coral reefs are highly biodiverse ecosystems and host the richest communities of fish species on the planet [6,7,8]. In general, fish have a life cycle with two different phases: the first pelagic larval phase, which allows them to disperse, and the second demersal or pelagic adult phase [9,10]. Fish larvae are a phase of great interest in marine ecology research because of the replenishment of populations, the connectivity or exchange of individuals between them, the understanding of fish dispersal patterns, and in general, a wide range of ecological processes of these populations depend largely on the understanding of the dynamics of the larval stages [11,12]. Therefore, studies on variations in the abundance and structure of fish larval assemblages are necessary to understand the dynamics of the fish populations present in coral reefs.

In coastal marine environments, variations in environmental conditions at different time scales can modulate the attributes of fish larval assemblages [13], such that during an annual cycle, which incorporates different hydrographic conditions, changes in their composition and abundance have been reported [14,15,16,17,18,19]. In general, the spatiotemporal variation of physical factors generally contributes the most to the biological responses of marine communities. For example, fluctuations in circulation, water column stability, turbidity, salinity, oxygen, and temperature can determine the viability of a particular species and consequently modulate the composition of a set or assemblage [20,21,22,23].

To date, research efforts on reef fishes in the ETP have focused on both the adult and larval stages [15,24,25,26,27,28,29,30]; however, in the CPO, efforts have focused mainly on adults [31,32,33,34,35,36,37]. In this region of the ETP, only two studies have analyzed larval fish assemblages, the first of which was carried out by Calle-Bonilla et al. [38], who evaluated the role of surface circulation in the movement of fish larvae associated with coral reef environments on Gorgona Island during two contrasting oceanographic periods, and the other one carried out by Ramírez et al. [39], who evaluated the temporal variability of the assembly of fish larvae associated with a coral reef formation on Gorgona Island on a daily (day-night) and monthly scale.

In this study, we assessed the response of ichthyoplankton assemblages associated with well-developed coral reef formation into intra-annual and inter-annual variability of hydrographic conditions. Specifically, we addressed the following questions: (1) Does ichthyoplankton abundance and community structure differ among the intra-annual contrasting oceanographic periods described for this region? (2) Is the structure of the fish larval assemblage similar on an inter-annual scale? (3) What are the main environmental factors affecting the abundance and community structure of fish larvae in this coral reef formation?.

2. Materials and Methods

The Colombian Pacific Ocean is an integral part of the eastern region of the ETP. In this region, the most developed coral reefs are located on the eastern side of Gorgona Island (2°58'N - 78°11'W) (Figure 1), on the continental shelf of Colombia's southern Pacific region, 36 km from the mainland [40]. This area, declared National Natural Park since 1984, has an average air temperature of 27°C (varying between 24°C and 30°C) and annual rainfall ranges between 4000 mm and 8000 mm distributed in two rainy periods, the most intense between April and November, and a second less intense period between December and March [41]. Two contrasting oceanographic periods have been described for this locality, the first from May to December with low surface salinity values and a deep thermocline (c.a. 47 m) and the second from January to April with high surface salinity and a shallow thermocline (c.a. 7.5 m), with a semi-diurnal tide that can reach a maximum record of 5.7 m [42].

Gorgona Island has a great variety of coastal habitats [43], among which the coral formations of La Azufrada, Playa Blanca, El Muelle, La Ventana, and Farallones stand out as habitats for diverse marine fauna [21,26,44,45,46]. All these coral formations have a low impact from anthropogenic activities, and the La Azufrada coral reef is considered one of the largest and well-developed coral reefs in the ETP [46]. The coral reef is 780 m long and 80-180 m wide, and it is formed mainly by coral species of the Pocilloporidae family, although there are also coral species from Psammocoridae, Agariciidae and Poritidae families represented by genera such as Psammocora, Pavona, Porites, and Gardineroseris [45,46]

Since 2007, the pelagic environment of the La Azufrada coral reef has been monitored twice a year as part of the protected area's oceanographic monitoring program, which consists of a sampling grid of nine equidistant stations over the reef (Figure 1). For this study, only samples collected during March and September 2017, 2018, and 2019 were used during the cold and warm oceanographic periods, which have been described for this location of ETP [38,42,47,48,49,50].

Discrete water samples were collected at a depth of 1 m using a Niskin bottle, and temperature, salinity, chlorophyll-a, and dissolved oxygen were recorded using a YSIpro® multiparameter probe and water column transparency using a Secchi disk. Fish larvae were obtained by day and night surface trawls of 5 min duration, using a bongo net of 30 cm diameter and 180 cm length, equipped with 300 and 500 μm pore meshes, with Hydrobios® flowmeters calibrated to quantify the volume of filtered water. Samples were fixed in a formalin solution (formalin-seawater) at a final concentration of 4% for transport to the Oceanographic Sciences Laboratory of the Universidad del Valle. All samples were reviewed to separate fish larvae from each zooplankton sample, and taxonomic identification was carried out to the closest possible species level, following the guidelines [51,52]. The identified species were deposited in the Zoology Teaching Collection of the Universidad del Valle.

To quantify the zooplankton biomass, half of the sample collected with the 300 μm mesh was concentrated on cellulose filters that were previously dried and weighed. The samples were dried at 60°C for 24 h and weighed on an analytical balance with a precision of 0.0001 g, determining the dry biomass by weight difference.

To reduce the effect of day and night variation in the structuring of the data matrix of fish larvae in the water column of the La Azufrada coral reef, the day and night records made during each sampling campaign were taken together. In addition, the information obtained with the 300 μm and 500 μm nets was consolidated, assuming that they were complementary sources of taxonomic information [53].

Temperature, salinity, dissolved oxygen, chlorophyll-a, zooplankton biomass, and fish larval abundance were compared between periods and years of study (summing the cold and warm periods for each year) using a non-parametric Kruskall Wallis test (KW) with Bonferroni correction and a post-hoc Tukey rank test. The representativeness of the sampling effort was established from the construction of rarefaction curves for the study period and the expected richness was defined from Clench's non-linear estimator, which predicts that the probability of finding a new species will increase (up to a maximum) as the sampling effort increases, that is, the probability of adding new species eventually decreases, but field work increases it [54]. The community attribute indices Shannon and Simpson were used to describe the diversity of the assemblages, and a comparison was made between periods by resampling 1000 replicates using the bootstrap technique.

The structure of the assemblage was described based on the Bray Curtis similarity index, defined from the abundance matrix of species with a frequency of occurrence equal to or greater than 5%. Transformation was performed using the log (x+1) function to reduce the abundance effect of dominant species. The average clustering algorithm was used to construct a similarity dendrogram, and the similarity profile routine (SIMPROF) was used to establish the significance of the clusters. Moreover, the contribution of species within the clusters generated in the similarity analysis was established using the similarity percentages (SIMPER) routine in the PRIMER v6.0® software. Finally, to determine whether there were differences in fish larval assemblage structure between study periods, a one-way abundance matrix similarity analysis (ANOSIM) was used to evaluate the differences in the structure of fish larval assemblages between study periods (totally different assemblages when R values are between 0.75 and 1) [55,56].

Finally, the relationship between the established assemblage structure and local environmental conditions (temperature, salinity, chlorophyll-a, and dissolved oxygen) was explored using a canonical correspondence analysis (CCA) carried out with the free ecological software Past® v3.12. CCA is a multivariate association analysis that maximizes the degree of correlation between fish larval abundance and physicochemical parameters.

3. Results

3.1. Fish Larval Assemblage Composition and Structure

A total of 4779 fish larvae were captured, of which 4556 (95% of the total), belonging to 88 species and 46 families, were identified during the six study periods. In our study, 41 species were first recorded in the La Azufrada coral reef, increasing the number of larvae of fish species associated with the coral formations of this locality to 162 taxa (Table 1, Table A1).

3.2. Ambient Conditions

Mean sea surface temperature ranged between 28.4 ± 0.13 °C (March 2017), 28.8 ± 0.13 °C (September 2017), 27.4 ± 0.24 °C (March 2018), 27.7 ± 0.09 °C (September 2018), 25.8 ± 0.29 °C (March 2019) and 27.5 ± 0.13 °C (September 2019), no significant differences were found for SST between oceanographic periods (KW, P>0.05). Mean sea surface salinity ranged from 29.6 ± 0.03 (March 2017), 25.2 ± 0.45 (September 2017), 26.7 ± 0.18 (March 2018), 26.1 ± 1.14 (September 2018), 30.7 ± 0.44 (March 2019) and 30.3 ± 0.13 (September 2019), there were significant differences between March and September 2017 (KW, P>0.05). The maximum temperature and salinity values were recorded in September 2017 and March 2019, respectively (Table 4).

3.3. Relationship Between Fish Larval Assemblages and Environmental Variables

The first two ordination axes of the canonical correspondence analysis (CCA) explained 67% of the total variance of the fish larval assemblage present in the water column associated with the La Azufrada coral reef. Salinity, chlorophyll-a, zooplankton biomass, pH and dissolved oxygen were inversely related to axis 1, whereas chlorophyll-a, temperature and pH were inversely related to axis 2 (Table 5, Figure 5).

The Bregmaceros bathymaster, Ophichthus zophochir, Bathycongrus macrurus, Antennarius sanguineus, Gobiesox sp., Lythrypnus sp., Achirus sp., and Coryphaena hippurus were directly related to chlorophyll-a and pH, whereas Anchoa sp., Anisotremus sp., Auxis sp., and Opistognathus panamensis were directly correlated with temperature. Finally, most of the species identified in the study periods and years evaluated were directly correlated with salinity, dissolved oxygen and zooplankton biomass, such as Cetengraulis mysticetus, Gobiidae sp., Oligoplites saurus, Caranx sp., Chromis sp., and Lampanyctus parvicauda (Figure 5).

% (Clench's nonlinear estimator) and the species richness and diversity varied significantly among the study periods and years, being higher in March 2018 and September 2019 than in the rest of the periods and years evaluated (Table 2).

Larval density (ind/100 m3) between the study periods did not show significant differences, except between March and September 2018. Zooplankton biomass (g/100 m3) was significantly higher in March 2019 than in the other periods evaluated (Figure 2). A non-significant positive correlation was established between larval density and zooplankton biomass (R= 0.14, p<0.05; Figure 3) in the study area, suggesting that zooplankton biomass did not have a dominant effect on larval density.

Based on the analysis of similarity between periods and years, six significant groups (I, II, III, IV, V, and VI) were established (SIMPROF, p < 0.05), with dissimilarity percentages between groups above than 65% (Figure 4). Each established group consisted of nine sampling stations in each study period. Group I (March-2018) was established due to the contribution of C. mysticetus, D. pacificus, Anchoa sp., and Anisotremus sp. (SIMPER, 57.36% similarity), which showed a tendency to cluster with those of group II. Group II (March-2017) was formed by the contribution of species such as: Anchoa sp., C. mysticetus, B. bathymaster, and O. saurus (SIMPER, 72% similarity). Group III (March-2019) was formed owing to the contribution of species such as: C. mysticetus, Caranx sp. Seriola sp., and Anisotremus sp. (SIMPER, 70% similarity) (Figure 4).

Group IV (September-2018) was formed due to the contribution of B. bathymaster, Gobiidae sp. 6, Anisotremus sp., and Gobiidae sp (SIMPER, 62.65% similarity). On the other hand, group V (September-2017) was established due to the contribution of species such as: Anisotremus sp., C. mysticetus, Microgobius sp., and S. evermanni (SIMPER, 58.72% of similarity) and group VI (September 2019) was formed due to the contribution of C. mysticetus, Anisotremus sp., Gobiidae sp. and Caranx sp. (SIMPER, 49.0% similarity). Finally, the ANOSIM similarity analysis established a significance level of 0.1% for the periods, with an overall R of 0.764, indicating that there was a difference between assemblages from all periods (Table 3).

4. Discussion

In the water column of the La Azufrada coral reef, there were abundant larvae of fish species that are typical of different marine environments. This characteristic is frequently reported in scientific literature, where species have been recorded from oceanic areas, sandy bottoms, and even mangrove environments in rocky or coral reef areas [9,58,59]. Planktonic organisms, such as fish larvae, are modulated by physical processes of transport (such as advection), behavioral (vertical migrations), and life history characteristics, and they tend to exhibit a high dispersal capacity [60]. This means that they can be present in environments that are atypical for adults. Furthermore, on the reef coral structures can function as originators of accumulation areas by mechanically disrupting currents, waves, and tidal flows [61,62,63].

The species richness distribution during the sampling periods was variable and did not show a clear temporal or annual trend. However, there were significant differences in larval density between the periods (March and September 2018), which coincided with a temporal order consistent with the climatic and oceanographic periods of the area (cold and warm seasons). Similarly, it was observed that there is a higher total larval density in the months of March in all the years evaluated, this may be due to the fact that in general there was a greater availability of zooplankton biomass in the months of March, finding significant differences between March and September of the years 2017 and 2018, thus favoring the fish species to synchronize their spawning periods for these dates. Furthermore, although the cluster analysis suggested that each period evaluated was a separate group from the others, in a general way, a tendency was observed for the groups to be ordered concordantly with the cold and warm seasons. Similarly, the analysis of similarity (ANOSIM), where the overall R determined that all groups (periods evaluated) were different, but there was a tendency for the fish larval assemblages from September to be more like each other.

Over the last few decades, marine fish larval assemblages have been described to present high spatial and temporal dynamics, where variations in composition and abundance occur on time scales ranging from hours to annual seasons [64]. In general terms, this temporal variability would modulate the simultaneous effect of multiple hydrographic and environmental factors that may come to determine larval survival [15,65,66,67], so it could come to determine the distribution, abundance and, in general, the structure of fish larval assemblages [68,69,70,71,72].

Another determining factor in the distribution, abundance and structure of fish larval assemblages is food availability, which is largely responsible for their survival and growth in the early stages of fish development [73]. Generally, fish larval density is closely related to zooplankton biomass and correlates with an increase in local food availability [74,75,76]. This coincides with the results obtained in this study, in which most fish larval species found on the coral reef were directly correlated with zooplankton biomass. Moreover, it has been documented that marine surface circulation patterns have a strong effect on the spatial distribution of fish larvae [77,78,79], giving rise to retention or advective processes that can decrease or increase the abundance of fish larvae in a locality [80,81]. During September, on the eastern side of Gorgona Island, the instantaneous surface circulation pattern is towards the sea, favoring the drifting movement of fish larvae, while during February the trend is towards the coast, which increases retention of larvae above the La Azufrada coral reef [38]. In our study, a trend towards a higher larval density in March compared to September was evident in all years evaluated. This result would be associated with the formation during the first months of the year of a cyclonic gyre in the vicinity of the la Azufrada coral reef [38], which may be favoring the retention of fish larvae and directly increasing the larval density in the cold period of the year.

The structural complexity of coral reefs is considered one of the most important factors in the organization of fish assemblages [82]. These habitats provide refuge areas against predators for many fish species and generate a food source consisting mainly of benthic invertebrates and macroalgae [83,84,85,86]. They also constitute natural sites where fish larvae can find the protection and food necessary for their rapid growth [87]. However, although the La Azufrada coral reef may serve a function as a nursery area for different fish species, most of these species probably exhibit life history strategies that reduce dependence on coral environments, and are likely to exploiting different habitat types during vulnerable life stages [15], this is reflected in the few species of fish larvae found in this study that in their adult stage are related to coral reefs (Antennarius sanguineus, Pseudogramma thaumasium, Oligoplites saurus, and Chaetodipterus zonatus), in addition to the fact that all were found exclusively in the month of March with the exception of Oligoplites saurus.

Another aspect to consider is the seasonal fluctuations in larval production, which are strongly associated with the reproductive events of the species and seem to be synchronized with the requirements of the larvae to increase recruitment success, defining the spatiotemporal pattern of distribution [65,88,89,90]. Thus, species such as C. mysticetus, B. bathymaster, Caranx sp., and Anisotremus sp. reproduce several times a year. Species collected only during a local oceanographic period, could be larvae of species that present only one reproductive event per year, such as Dormitator sp., and Menticirrhus sp. that were recorded in the month of March and Microgobius sp., and Stelifer sp. that were present in September.

5. Conclusions

The composition and structure of the fish larval assemblage in all evaluated periods were different and specific for each period. However, a tendency was observed for them to group together or be more similar between the climatic and oceanographic periods described for the study area (cold and warm seasons). This suggests that the intra-annual variation has a greater effect on the fish larval assemblages, where it is evident that the cold period (March) presents better conditions for the fish species to carry out their spawning season; and subsequently there is a higher larval density in the la Azufrada coral reef.

The results of this study increase the available knowledge on the assemblage of fish larvae associated with the ETP coral reef and how it varies during the year and between the years, showing that intra-annual variation, and in particular the cold season, have a greater effect on the assemblage of fish larvae associated whit the La Azufrada coral reef. Some species were frequent in the years and seasons evaluated and others in part of these, probably reflecting different spawning strategies or ontogenetic changes in distribution. These differences resulted in temporal variations in the species assemblage. However, considering the method implemented to capture fish larvae in our study (zooplankton trawls), it is necessary to implement complementary capture methods, such as light traps, to obtain information on more advanced stages of development and thus obtain information on population recruitment processes, which allows a complete overview of the temporal dynamics of the different fish species that are related to the coral reefs of the Eastern Tropical Pacific.