Water and phytoplankton sampling

Sample collection was performed monthly during North winds and dry seasons (November 2011 to April 2012). Five sampling sites were considered (Table 1) among which, four oyster banks and a site near the sea were located, as a reference, because they are marine organisms. Physicochemical parameters, such as dissolved oxygen, temperature, and salinity, were recorded in situ using an YSI 6600 Multi-parameter, Water Quality Probe. During the process, water was collected using a 3L Van Dorn bottle at a depth of 30 cm. The volume of each sample was 1L and was used for quantification of phytoplankton, which was fixed in situ using 0.5 to 1 mL of lugol acetate solution per 100 mL (^{Ferrario et al., 1995}). Samples for analysis of phosphates (PO₄ ^-3), nitrates (NO₃-), silicates (SiO₂) and chlorophyll-a were collected in 500 mlL polyethylene bottles.

Phytoplankton samples were collected to identify complex species by towing a 60-µm mesh net and 250 mL collection tube behind a boat with an outboard motor for 5 minutes as it maneuvered slowly in circles. Samples were fixed with a final solution of 4 % formalin for later identification (^{Ferrario et al., 1995}; ^{Band-Schmidt et al., 2011}) and transported to the Laboratorio de Investigación de Recursos Acuáticos (LIRA), at the Instituto Tecnológico de Boca del Río.

Laboratory analysis

Chlorophyll-a determination was performed using the techniques of ^{Jeffrey and Humphrey (1975)}, water samples were filtered in the laboratory, using Whatman GF/C Glass microfiber filters, Grade GF/C of 47 mm, placed in 90% acetone and refrigerated at 4° C for 24 hours for extraction, after this, samples were read in a Thermo Scientific Genesys 20 spectrophotometer. The nutrients PO₄ ^-3, NO₃ ^- and SiO₂ were determined using the colorimetric method CHEMets® Water Test Kit: K-8510, K-6904 and K-9010 respectively. Phytoplankton sample processing was performed directly by observation of a subsample and preliminary identification mainly of genus of diatoms and dinoflagellates was made using rapid recognition of taxonomic characteristics. Identification of phytoplankton was based on published keys (^{Hallegraeff, 1995}; ^{Esqueda-Lara & Hernández-Becerril, 2010}). The observations were made using a binocular optical microscope Carl Zeiss Axiostar 47190 and a 10x objective. The total cell count for each phytoplankton sample was performed using a Sedgewick Rafter Counting Chamber (1 mL). Phytoplankton abundance was reported, according to the number of cells per liter (cel L^-1). The analysis focused primarily on the diatoms and dinoflagellates as the predominant groups contributing to primary productivity.

Statistical analysis

To achieve normality, all the data were transformed using a logarithmic function (log₂ X + 1) (^{Aké-Castillo, 2015}), and the Levene test was used to check for homoscedasticity of variances. The software program STATISTICA version 7 (^{StatSoft, 2016}), a Kruskal-Wallis nonparametric analysis of variance, a parametric F-test, a parametric analysis of variance (ANOVA), and the Tukey test were used to search for significant differences among physicochemical parameters of the water during sampling months, November 2011 to April 2012, in sampling sites. The relationship between water physicochemical parameters and phytoplankton were evaluated using a multivariate Canonical Correspondence Analysis in an Addinsoft ® XLSTAT version 2015.

Monthly mean values of the physicochemical parameters

Maximum mean temperatures were recorded in March (29.1 ± 0.8 °C) and April, 2012 (28.0 ± 0.5 °C), and minimum mean temperatures were observed in November, 2011 (21.9 ± 0.7°C). The analysis of variance showed temporal variation between months. The temperature recorded in January, March and April were different from the rest of the samplings (F = 76.9, p ≤ 0.05) (Table 2). The results obtained show the maximum temperatures recorded are similar to those reported by ^{Martínez Del Rosario (2011)} during the dry season (March and April), as well as those reported by ^{Zaballa-Carranza (1982)} during spring (April) (28.12 °C), and winter (February) (25.27 °C).

Maximum mean salinity was recorded in March and April, 2012 (30.8 ± 1.1 and 31.0 ± 2.2 psu, respectively), while minimum mean salinity was recorded in November, 2012 (20.6 ± 3.6 psu). Salinity showed differences in the first three months of sampling with respect to the rest, the latter presented the highest (F = 15.20, p ≤ 0.05) (Table 2) The variation of temperature and salinity during the last months of sampling is associated with seasonal change, and that the lagoons system of Mandinga is shallow (1-8 m in depth), is permanently connected to the sea, and is influenced by freshwater inputs from the Jamapa and Arroyo Hondo rivers (^{Ruíz-Barreiro, 2012}).

Maximum mean dissolved oxygen levels occurred in November 2011, and February 2012 (9.3 ± 1.3 mg L^-1 and 8.8 ± 0.7 mg L^-1, respectively), and a minimum mean of 6.7 ± 0.5 mg L^-1 in April 2012. Analysis of variance (ANOVA) allowed observing significant differences during the month of April with respect to November and February F = 5.29, p = 0.05 (Table 2).

Values of dissolved oxygen obtained during the study are similar to the average value of 4.5 mg L^-1 reported by ^{Contreras & Castañeda (2004)}. These variations are due to North winds and evaporation, characteristics of the seasonal climate in the area (^{Zaballa-Carranza, 1982}; ^{Martínez Del Rosario, 2011}). ^{Bornn & Ruíz-Zamites (1982)} reported an even lower mean concentration of 1.73 mg L^-1 in April, 1982; due to the increase of organic matter in suspension and biochemical demand caused by dredging in the system.

A maximum mean concentration of chlorophyll-a was 14.6 ± 5.6 mg m-³ in April 2012 and a minimum level of 7.9 ± 4.1 mg m-³ occurred in December 2011. Analysis of variance showed variations in the concentrations of silicates and chlorophyll-a during November, February and April (F = 6.83, p ≤ 0.05; F=6.78, p ≤ 0.05) (Table 2). The chlorophyll-a in the lagoons system was consistent with ^{Barreiro-Güemes & Aguirre-León (1999)}, who reported mean values of 8.7 and 14 mg m-3 of chlorophyll-a in areas with greater influence of the tide, greater turbulence and less transparency. While, in November 2011 and April 2012 (Table 2), a minimum mean for silicates was 119.8 ± 24.6 µM and the maximum mean was 256.2 ± 25.2 µM. The silicate concentration is inversely proportional to that of chlorophyll-a, where values obtained were not considered limiting within the system (^{Bornn & Ruíz-Zamites, 1982}).

Maximum monthly means for nitrates were recorded in February and March, 2012, with 4.0±0.0 and 20.3±44.9 µM, and a minimum mean of 0.2 ± 0.4 µM in April, 2012. Nitrogen compounds showed significant differences in January, February and April with respect to the rest of the samples (W = 12.18, p ≤ 0.05, Table 2).

The variability in nitrate concentrations that were recorded during the study may be due to climatic or extemporaneous factors, which caused turbulence in the bottom, mobilizing and incorporating to the water column the nitrogen accumulated in the sediment. For ^{Kennish & Perl (2010)}, nitrogen and other nutrients precipitate and tend to accumulate in the bottom attached to the sediment where the biogeochemical interactions take place.

Maximum mean of phosphate was 2.9 ± 4.2 µM in February, 2012, and a minimum mean (0.8 ± 0.3) µM was observed in January, 2012. The analysis of variance allowed recognizing significant differences in physicochemical parameters, with the exception of phosphates (F = 1.09, p = 0.39) (Table 2). November, 2011, and March, 2012, were transition periods between the windy and rainy seasons, where many nutrients were transported (^{Contreras et al., 1996}). ^{Rodríguez (2008)} found that nitrogen compounds are inversely proportional to temperature, and total phosphate is greater during the warmer months. ^{Bornn & Ruíz-Zamites (1982)} determined that nutrients are not a limiting factor in phytoplankton growth. Such compounds settle out of the water column to eventually be incorporated by benthic organisms, and have their greatest concentrations in coastal sites.

Monthly averages of water physicochemical parameters in sampling sites

A maximum mean temperature was 29.9 °C during March 2012, at site S4. A minimum mean occurred in November 2011, at site S2 (20.8 °C) (Figure 2). A maximum mean salinity predominantly occurred in site S1, as it was the closest site to the ocean. However, in April 2012, a maximum mean occurred at site S2 (33.9 psu), and a minimum mean (17.9 psu) at site S5 in November 2011 (Figure 3). The maximum level of dissolved oxygen was 10.9 mg L^-1 at site S4 during March 2012, and a minimum during April 2012, (5.8 mg L^-1) at site S3 (Figure 4). A maximum mean for chlorophyll-a was 35.1 mg m-³, which occurred at site S5 during January 2012, and a minimum mean (0.63 mg m-³) occurred during December 2011, at site S1 (Figure 5).

Figure 2 Mean monthly water temperatures, from November 2011 to April 2012, in the lagoon system of Mandinga, Veracruz, Mexico. 

Figure 3 Mean monthly water salinity, from November 2011 to April 2012, in the lagoon system of Mandinga, Veracruz, Mexico. 

Figure 4 Mean monthly values for dissolved oxygen in the water column, from November 2011 to April 2012, in the lagoon system of Mandinga, Veracruz, Mexico. 

Figure 5 Mean monthly for chlorophyll-a in the water column, from system of Mandinga, Veracruz, Mexico. 

According to ^{La Barre et al. (2014)}, the variability of physicochemical characteristics of water in coastal systems is due to climate change, rainfall events, off-season winds, and contributions from industrial and household nutrients. The latter can contribute to eutrophication of the environment, resulting in reduced dissolved oxygen and high concentrations of nitrates, phosphates and silicates (^{Álvarez-Gongóra et al., 2012}) (e.g. site S1 in February, 2012). In general, the shallow coastal systems have short hydraulic residence times, which allow a renovation in a few days, entering well-oxygenated seawater. In addition, other factors such as primary production, the mechanical action of wind and the lack of barriers that prevent its action, also favor the incorporation of oxygen from the atmosphere (^{López-Monroy et al., 2017}).

Maximum mean concentrations of nitrates and phosphates were 100.7 and 10.5 µM during March (S5) and February 2012 (S1), respectively (Table 3). Concentrations observed in site S5 can be explained by increased tourist and commercial activities in the area. There is also influence from sewage discharge from the population of Mandinga, which has no drainage system. The lowest means for nitrates were detected during December, 2011, March and April, 2012. However, phosphates were at their lowest concentrations during January 2012, followed by consecutive increases during February, March and April 2012. A maximum mean for silicates was recorded during January 2012 (332.8 µM) at site S1, and the minimum mean (83.2 µM) at site S4 during April 2012. Chlorophyll-a concentrations increased during March and April, 2012.

In the phytoplankton analysis, thirty-one genera of microalgae and sixteen species were identified, classifying four dinoflagellates and twenty-seven diatoms. The keys that correspond to these genera and species are in the last column of Table 4.

These types of dinoflagellates/diatoms associated with algal blooms have been recorded in different coastal systems, such as: Tripos furca (Ehrenberg) F. Gómez; Prorocentrum mican Ehrenberg; Asterionellopsis gracialis (Castracane) Round; Cylindrotheca closterium (Ehrenberg) Reimann & J.C. Lewin; Skeletonema costatum (Greville) Cleve; Thalassiosira spp. and Chaetoceros spp. (^{Guerra-Martínez & Lara-Villa, 1996}; ^{Orellana-Cepeda et al., 2002}; ^{Gárate-Lizarraga et al., 2007}; ^{Nuñez-Vázquez et al., 2008}; ^{Muciño-Márquez et al., 2011}b; ^{Aké-Castillo et al., 2014}; ^{Calvo-Vargas et al., 2016}).

Phytoplankton samples from the lagoons system of Mandinga, during the study period, consisted mainly of benthic and planktonic diatoms and dinoflagellates taxa. This is consistent with ^{Okolodkov & Blanco-Pérez (2011)}, who stated that the microalgal flora in lagoon systems of Veracruz are dominated by benthic diatoms in shallow areas and planktonic forms in coastal lagoons having strong influence from the ocean. The lagoons system of Mandinga has a semi-closed and shallow mouth to the ocean, directly influences primary productivity thus, hydrological characteristics are euryhaline, especially during the rainy season (^{Castán-Aquino, 2013}; ^{Gómez-Sánchez, 2013}). As reported by ^{Álvarez-Gongóra et al. (2012)}, who observed a greater variability in total phytoplankton abundance during the rainy season.

The diversity and ecology of phytoplankton along the coastal zone of Veracruz is influenced by higher densities of diatoms and dinoflagellates, although in the present study the concentrations of these microalgae were not greater than 6.7 x10⁴ cells L^-1 (Table 4) (^{Muciño-Márquez et al., 2011b}; ^{Rodríguez-Gómez et al., 2015}). Their reduced concentration can be explained by the reportedly greater zooplankton abundance which directly influences primary productivity (^{Orduña-Medrano, 2012}). According ^{Contreras & Castañeda (2004)}, lagoons and estuaries along the Gulf of Mexico are characterized by the dominance of copepods. According to ^{Contreras (2016)}, the community structure of zooplankton at the mouth of the Jamapa River is made up of eleven families of copepods and two families of cladocerans that dominate in North winds and rainy seasons, respectively.

Canonical Correspondence Analysis

Canonical Correspondence Analysis (CCA) revealed that the abundance and distribution of phytoplankton in the lagoons system of Mandinga accounted for 62.67 % of the variance (p = 0.0001) in physicochemical parameters (Table 5). Salinity (r = -0.553) best defined axis 1 followed by temperature (r = -0.455). These parameters alone explained 44.90 % of the variance. ^{Muciño-Márquez et al. (2011a)}, also considered salinity as an important factor (r = 0.90) that promoted increases in total phytoplankton abundance and HAB forming species. The parameters defining axis 2 were silicates (r = -0.733) and dissolved oxygen (r = -0.598), and explained 17.76 % of the variation (Figure 6). For ^{Álvarez-Gongóra et al. (2012)}, the axes that best explained 83 % of the variability were: Axis I, related to dissolved oxygen, silicates and chlorophyll-a and Axis II, to temperature, salinity, nitrates and phosphates.

Figure 6 Canonical Correspondence Analysis (CAC): Ordination diagram for microalgae taxa (▲) and physicoche-mical parameters (vectors) in the lagoon system of Mandinga, Veracruz, Mexico. 

High salinity and temperature were related to the presence of dinoflagellates and Diatoms, such as: Prorocentrum micans, Protoperidinium oceanicum, Scrippsiella sp., Lithodesmium undulatum and Grammatophora oceanica (Figure 6).

Other microalgae such as Tripos furca, Navicula sp., Amphora sp. and Skeletonema costatum prefer environments having medium to high concentrations of dissolved oxygen, phosphates and nitrates. The diatom taxa Coscinodiscus sp., Tabellaria sp., Nitzschia sp. and Melossira moniliformis was present in 66 % of samples. The diatoms Pseudo-nitzschia sp., Asterionellopsis gracialis, Entomoneis sp., Guirnardia striata, Leptocylindrus danicus and Licmophora ehrenbergi predominantly occurred at medium to high concentrations of silicates and dissolved oxygen, and at lower temperatures. Other taxa such as Diploneis sp., Cylindrotheca closterium, Rhizosolenia setigera, Gyrosigma sp., Chaetocero sp., Thalassionema nitzschioides, Bacteriastrum sp. and Surirella sp. preferred environments with less salinity, dissolved oxygen, and nutrients, and lower temperatures.

According to Ayala (2008) Prorocentrum micans, Protoperidinium sp., occur under conditions of salinity and high temperatures. ^{Álvarez-Gongóra et al. (2012)}, found that the dinoflagellate Tripos furca is related to the presence of salinity, temperature and nutrients (nitrates and phosphates). ^{García et al. (2015)}, mention that diatoms, such as Leptocylindrus danicus and Pseudonitzschia seriata complex are favored by having availability of nutrients and low temperatures.

^{Bucheli (2016)} refers that Cylindrotheca closterium is a species that has a greater relation with nutrients. On the other hand (^{García et al., 2015}) related its growth to low temperatures. Unlike ^{Ayala (2008)} in a study conducted in two lagoons, different phytoplankton species such as Prorocentrum micans, Protoperidinium sp., Scrippsiella sp., Cylindrotheca closterium, Asterionellopsis gracialis, Rhizosolenia setigera have a successful correlation with values closer to the N:P ratio and high concentrations of oxidized nutrients. While, Lithodesmium undulatum, Protoperidinium sp., Pseudo-nitzschia sp., tends to predominate under conditions of high temperatures and concentrations of silicic acid.

According to ^{Okolodkov (2008)}, dinoflagellates in the genus Protoperidinium proliferate at high salinities and are represented by 46 species in the Parque Nacional Sistema Veracruzano (PNSV). According ^{Muciño-Márquez et al. (2011a)}, Prorocentrum spp. is widely distributed along the Gulf Coast of Mexico. In the Sontecomapan Lagoon, Veracruz, five species were identified including P. micans. ^{Muciño-Marquez et al. (2011a)}, reported that salinity is a determining factor for the abundance and distribution of the genus Prorocentrum, which also was tested in the present study (Figure 6). However, we found no relationship between the presence of dinoflagellates and nutrient concentrations except for Tripos furca. ^{Dortch (1990)} states that there is significant variation between species, both in the inhibition and the preference of nutrients such as nitrogen and ammonium without any apparent pattern, which is influenced by environmental conditions. ^{Gárate-Lizárraga et al. (2008)}, reported a proliferation of Prorocentrum micans during tidal flows in Magdalena Bay. This suggests that Prorocentrum occurs in shallow sampling sites, approximately 5 to 10 m in depth, and within a narrow temperature range. This information coincides with the results in the present study.

Reductions in freshwater input lead to increased salinity, and changes in the structure and composition of phytoplankton assemblages (^{Ferreira et al., 2005}). This could lead to proliferation of opportunistic marine species, which can produce HABs, such as Tripos furca in the lagoons system of Mandinga (^{Guerra-Martínez & Lara-Villa, 1996}). The presence of Tripos furca in the lagoons system of Mandinga was recorded during the windy season (January-February, 2012), and reported as a common species in the Gulf of Mexico and Pacific Ocean (^{Okolodkov, 2010}). This dinoflagellate produces hypoxia and anoxia in these regions, causing death by suffocation of aquatic species (e.g. Puerto Escondido and Baja California in 2002).

Dinoflagellate in the genera Prorocentrum and Ceratium, two dominant HAB species found along the Pacific coast of Guerrero (^{Gárate-Lizárraga et al., 2008}), Prorocentrum produces toxins that, when consumed, cause diarrheal shellfish poisoning (^{Hernández-Becerril et al., 2007}). According to ^{Calvo-Vargas et al. (2016)}, T. furca is associated with HABs such as P. micans at maximum concentrations of 20x10³ cel L^-1 during warm months (April), at temperatures near 27 °C, dissolved oxygen levels of 5.8 mg L^-1, and salinities of 34 psu. The presence of S. costatum, also associated with HABs has been observed during the rainy season, at maximum abundances of 6.5x106 cel L^-1, temperatures near 27 °C, 30 psu salinity and dissolved oxygen concentrations near 7 mg L^-1. Such conditions also were recorded in Costa Rica during El Niño events from 2008 to 2010.

According to ^{Wells et al. (2015)}, temperature is a parameter that affects various metabolic processes of organisms that produce algal blooms. This parameter has a strong influence on the presence and trajectory of the phytoplankton community. However, there is no significant evidence linking temperature changes to toxin production. In this investigation, the highest temperatures were recorded during March and April, identifying the presence of Coscinodiscu sp., Thalassiosira sp., Tripos furca and Prorocentrum micans, phytoplankton reported by other authors as algal blooms. However, no changes in water color or other characteristics indicating algal blooms were observed.

In this study, Pseudo-nitzschia had maximum abundance in November (6.7x10⁴ cel L^-1) and December 2011 (5x104 cel L^-1) (Table 4). This genus was associated with low temperatures and high concentrations of dissolved oxygen and silicates, even though it has optimum growth at salinities of 32.5 to 34.4 psu and temperatures of 19 to 23 °C (^{Méndez and Ferrario, 2009}). According to ^{Aké-Castillo et al. (2014)}, the availability of high concentrations of orthophosphates and nitrates, and increases in water temperature are some factors that have triggered HABs in the coastal zone of Veracruz, Mexico. Thus, and in accordance with ^{Rodríguez-Gómez et al. (2015)}, continuous monitoring of phytoplankton in the coastal zones of Veracruz is necessary for recognition of the diversity, abundance and seasonality of phytoplankton, especially with regard for those species involved in toxic algal blooms. This resistance to environmental variability may be influenced by the presence of non-endemic species that have been introduced into coastal areas by the ballast water of ships that land in the state of Veracruz (^{Okolodkov et al., 2007}; ^{Aké-Castillo et al., 2014}). Water and sediments may contain dinoflagellate cysts capable of resisting for a long time (^{Meave Del Castillo, 2014}). The strategic plan for biological diversity carried out by different foreign government agencies puts pressure on the Mexican legal framework, which requires decision makers to act more effectively in aspects related to invasive species already introduced and potential bio invasions through maritime transport (^{Okolodkov & García-Escobar, 2014}).

According to ^{Okolodkov et al. (2007)}, it is essential to develop legislation for emptying ballast water in Veracruz marine waters from ships and other vessels. In addition to strict regulatory oversight of the maximum permissible limits on discharges of wastewater into natural bodies of water. The construction of a drainage system in the areas surrounding coastal systems, could reduce the availability of nutrients and the formation of algal blooms.