Environmental parameters
Salinity and water temperature showed strong seasonal variations in each pond (Fig. 2). The highest temperature reached a little over 30 oC in August 2022 and was slightly below 10 oC during winter. In mid-March 2023, a cold front resulted in a temperature decrease to around 5 oC. Spatially, the temperature was very similar among the ponds most of the year and slightly variable, < 3 oC, in summer and winter (Fig. 2a). Salinity was weakly seasonal and ranged 25–45 ppt in summer and 25–31 ppt in winter. The extreme high salinity of ~ 45 ppt was detected in OMWM-2 in August 2022. The differences among the ponds were usually < 5 ppt (Fig. 2b). The REF ponds showed less fluctuations throughout the year compared to others. Higher salinity in summer (~ 35 ppt) may have resulted from high heat-induced evaporation and relatively small water volumes in the ponds.
The pH values showed slight seasonality with generally higher summer and lower winter values and seemingly fluctuated more in warm months from April to September in all the ponds. Typically, pH ranged between 7.3 to 8.6, but extremely high values of 9-9.7, were detected in OMWM-2 and OMWM-1 during summer and very low values, 6.9-7.0, in PGD-1 in May 2022 and April 2023 (Fig. 2c). The DO concentration fluctuated among months without strong seasonality. Typically, DO measurements varied between 6 to 14 mg/mL, but some extremely high values (> 18–20 mg/L) were detected in OMWM-1 and 2, and low concentrations ~ 2 mg/L in PGD-1 in May 2022 and REF-1 in April 2023 (Fig. 2d).
The environmental variables were homogeneous within each pond as indicated by the standard deviation and coefficient of variation from the replicate measurements taken in December, February, and June (Table S2 in Supplementary Material). Significant linear correlations were obtained between salinity and conductivity (Fig. S1, a), and dissolved oxygen (DO) concentration and DO saturation (Fig. S1, b).
Algal compositions and harmful algal blooms
In total, we recorded 126 taxa (species, genera, or taxonomic groups) from seven classes including Bacillariophyceae (diatoms), Dinophyceae (dinoflagellates), Cyanophyceae (cyanobacteria), Chlorophyceae (chlorophytes, or green algae), Cryptophyceae (cryptophytes), Raphidophyceae (raphidophytes), and Haptophyceae (haptophytes). Over 60 taxa were detected as ‘common’, ‘abundant’ or ‘highly abundant’ in one or more samples (Table S1 in Supplementary Material). Diatoms were the most diverse group in terms of the number of taxa recorded, followed by dinoflagellates and cyanobacteria. More than a dozen potentially toxic and harmful species were ranked as highly abundant and dominant, including dinoflagellates Akashiwo sanguinea, Kryptoperidinium foliaceum, Levanderina fissa (syn. Gymnodinium uncatenum), Gymnodinium aureolum (Karlodinium sp.), Amphidinium carterae, and Prorocentrum lima, Cyanobacteria Planktothrix agardhii, Phormidium autumnale and Lyngyba aestuarii, Raphidophytes Chattonella subsalsa and Heterosigma akashiwo, and Haptophyte, Prymnesium parvum (commonly referred to as “golden alga”) (Table 3).
Table 3
Major harmful algal bloom (HAB) species in SMPs, the months and SMPs that HAB species rated ‘common’ and higher, maximum cell density, and potential toxins and toxicity. Months in bold are the time(s) when the HAB species were detected abundant or highly abundant. SMPs in bold are the ponds where the HAB species were observed to be abundant or highly abundant.
Species | Months | SMPs | Maximum cell density (cells/mL) | Toxins & Toxicity |
Dinoflagellates |
Prorocentrum lima | Dec-Jan, Apr- May, Jun-Aug, Sept | REF-2, OMWM-1 OMWM-2 | 6.7×104 | Okadaic acid; Diarrheic shellfish poisons (DSPs) |
Amphidinium carterae | Dec-Mar, Jun-Sept | OMWM-1, PGD-2 | 0.8 ×103 | Hemolysins; Ciguatera fish poisons (CFP) |
Akashiwo sanguinea | May, Jun, Jul, Nov | REF-1 &2 PGD-1 & 2 | 1.04 ×104 | Ichthyotoxic, cause gill-clogging; suffocation due to low oxygen |
Gymnodinium aureolum (Karlodinium sp.) | April-Aug, Sept | REF-1, PGD-1&2 | 5.7 ×103 | Karlotoxins, polyketide toxins |
Levanderina fissa (syn. Gymnodinium uncatenum) | June, July, Aug | REF-1, OMWM-2, PGD-1&2 | 1.8 ×104 | Toxin unknown; Ichthyotoxic effects; blooms related shrimp kills |
Kryptoperidinium foliaceum (syn. Peridinium foliaceum) | May-Aug | PGD-1 & 2 OMWM-2 | 2.7 ×104 | Cause metabolic stresses |
Cyanobacteria |
Planktothrix agardhii | Jan, Apr, Jun, Nov | REF-1 & 2, PGD-2, OMWM-1& 2 | 0.2 ×103 (as # of filaments) | Microcystins; Hepatotoxins; possible human carcinogen |
Phormidium autumnale | Feb | OMWM-2 | 0.6 ×103 (as # of filaments) | Anatoxin-a; Neurotoxic shellfish poisons (NSP) |
Lyngbya aestuarii | May, Jun | PGD-1 &2, OMWM-1&2 | 8 ×103 (as # of filaments) | Multiple toxins: lyngbyatoxin A and debromoaplysiatoxin; cause irritation on skin, eyes and respiratory system |
Haptophyte |
Prymnesium parvum (“golden alga”) | Apr, May Jun | REF-1 & 2, PGD-2 OMWM-1 & 2 | 8.5×105 | Prymnesins; Cytotoxic, hemolytic, neurotoxic and ichthyotoxic poisons |
Raphidophytes |
Chattonella subsalsa | Jun, Aug, Nov | REF-2, PGD-2, OMWM-2 | 8.8×103 | Brevetoxins and hemolytic substance, NSP |
Heterosigma akashiwo | Jul, Aug | REF-1, PGD-1 & 2 | 1.4 ×104 | Brevetoxins and hemolytic substance, NSP |
The temporal and spatial variations of the algal compositions in the ponds were highly dynamic during the study period. In each pond, the dominance of algal taxonomic groups, particularly diatoms vs. non-diatoms, changed successively (Fig. 3). Diatoms dominated fall assemblages from September to November of 2022 in most of the ponds. In the following months, the succession of algal groups was fast within each pond and differed between ponds. In REF-1, diatom percentage decreased in December 2022-January 2023 and increased again from February until May-June 2023, accounting for > 60–97% of the total community during that time. In comparison, diatom percentage in REF-2 decreased to 20% in February due to green algae growth and back to > 50% from March through June of 2023. In PGD and OMWM ponds, diatoms contributed most to the winter and spring assemblages, especially in PGD-1 and OMWM-1, but green algae and dinoflagellates became abundant and dominant in summer and fall. In OMWM-2, dinoflagellates were dominant most of the time. In summer, particularly July-August 2022, dinoflagellates, raphidophytes, or cyanobacteria dominated these SMPs.
Within each taxonomic group, species succession occurred and differed among ponds (Table S1). For example, diatoms were dominant in REF-1 in October-November (fall) and March-April (spring) as shown in Fig. 3, however, fall diatoms were composed of Gyrosigma, Pleurosigm, and Naviculoids whereas the spring diatoms were abundant of Haslea ostrearia. With few exceptions, most diatoms are nontoxic. The toxin-producer Pseudo-nitzschia species were detected ‘occasionally’ in December and January in REF-1, PGD-1 and OMWM-1. A harmful diatom Cylindrotheca closterium was abundant from April to June of 2023 in all SMPs. Most dinoflagellate species are toxic and harmful. A. sanguinea, K. foliaceum, L. fissa, G. aureolum, and P. lima dominated the summer assemblages in REF-1, PGD-1, PGD-2, and OMWM-2 (Table 3), accounting for 60–93% of total biomass. In winter, although algal biomass was generally lower than summer, dinoflagellates, mainly represented by P. lima (Table 3 and Table S1), could still be abundant and dominant in December - January, making up to 60% and 98% of total biomass in PGD-2 and OMWM-2, respectively. Moreover, P. lima made up 40–90% of the algal community in OMWM-2 in March-May 2023. Two toxin-producing raphidophyte species, C. subsalsa and H. akashiwo, were highly abundant in PGD-2 in July and OMWM-2 in August in 2022, respectively. Co-existence of both species was observed in high abundance in PGD-2 in August 2022. Cyanobacteria were present frequently in various ponds throughout the year. The highest percentage of cyanobacteria was detected in OMWM-1 in early June 2022, accounting for 65% of total algae, and composed of L. aestuarii and P. agardhii, both are potential toxin producers. Other cyanobacteria Chroococcus and pico-coccoical Aphanocapsa species were common but rarely abundant or dominant. Another harmful cyanobacterium, Ph. autumnale was abundant in OMWM-2 in February 2023 (Table S2). The golden alga P. parvum was abundant in PGD-2 in April and dominant in OMWM-1 in May 2023 (Fig. 3), reaching extremely high cell density > 8 x 105 cells/ml (Table 3).
Differences in algal communities between two ponds located in the same type of treatment stratum were noticeable (Fig. 3). In August 2022, REF-1 was dominated by dinoflagellates, whereas REF-2 was characterized by highly abundant raphidophytes. Differences were also detected between PGD-1 and PGD-2 in July 2022 and April 2023. In May 2023, the golden alga was dominant in OMWM-1 while dinoflagellates dominated OMWM-2 (Fig. 3). The differences were also detected at the species/genus level. In April 2023, diatoms were dominant in both REF ponds, however, REF-1 was dominated by Haslea ostrearia (90% of total biomass estimate), a diatom with blue pigment, whereas REF-2 was abundant with several other diatoms including Striatella unipunctata and Melosira moniliformis. At the end of July 2022, dinoflagellates became abundant in REF-1&2, PGD-1 and OMWM-2, however, each pond held different dinoflagellate assemblages. In REF-1, A. sanguinea was dominant and accounted for 85% of the assemblage, while REF-2 was characterized by a mix of abundant A. carterae, P. lima and A. sanguinea. Meanwhile, PGD-1 was occupied by K. foliaceum (55%), and OMWM-2 by G. aureolum. Community differences were also indicated in the ordination diagram of samples derived from NMDS analysis (Fig. 4).
Seasonal variations were evident in most ponds, as indicated by the distances between samples in the ordination. In general, summer assemblages, from June to August, were distinct from the cold months. Abundant taxa could be comparable among all ponds but dominant at different times, creating a spatial-temporal mosaic that left ponds similar overall but dissimilar at any particular time. In an exception, the OMWM-1 sample from May 2023 (labeled as O10523 in Fig. 4) was highly distinct from other samples attributed to the absolute dominance of P. parvum.
Water temperature and pH were the major corresponding (of the examined) factors on the changes of algal communities (p < 0.01), followed by salinity and DO, as shown by CCA analysis (Fig. 5). These four parameters explained about 6.9% of the total variation of the algal community. The associations of species occurrences and successions were correlated to the environmental parameters as indicated by the distances between species and those of species to the environmental arrows in Species-Environment biplot (Fig. 6). Diatoms, such as Melosira, Gyrosigma, Skeletonema and tube-dwelling Navicula sp. were more popular in cold seasons, whereas summer diatoms included Entomoneis, Fragilaria, and Rhopalodia. The HAB species, A. sanguinea, K. foliaceum, and G. uncatenum (syn. L. fissa), and C. subsalsa and H. akashiwo were correlated to temperature gradient and highly abundant in summer. P. parvum was less correlated with temperature but more to pH (Fig. 6).