Sampling sites
The study area for this research included eight sampling sites along the lower SJR (Fig. 1). The sites were selected based on land use (e.g., residential, commercial, wastewater treatment facilities, superfund sites; Fig. 1), impairments near the sites, and previous research efforts in this area (Bielmyer-Fraser et al. 2020; 2022a). Water quality has been studied at these sites since 2017, using Jacksonville University research vessels, the Seabatical and the RV Larkin, with approximately six sampling trips per year. Data for this study were collected from 2019 to 2022, between 11:00 am and 4:00 pm. However, due to logistical limitations site 8 could not be sampled during every trip (five times during this study). Also, sampling was limited during the COVID-19 pandemic. A more detailed description of the land use and types of contamination at each site can be found in Bielmyer-Fraser et al. 2020. It is important to note that water quality has been shown to fluctuate spatially to some extent and more prominently over time (Bielmyer-Fraser et al. 2020). Additionally, metal (cadmium, copper, lead, nickel, silver) concentrations have been shown to fluctuate above water quality criteria and accumulate in submerged aquatic vegetation in the LSJR (Bielmyer-Fraser et al. 2020; 2022a).
Field collection
Depth and GPS latitude and longitude coordinates were recorded at each site. A subsurface water sample was collected at each site in a 1-L polypropylene bottle to be used for chemical analysis in the laboratory. Two YSI multiparameter meters (YSI®, Yellow Springs, OH) with select probes were used to measure temperature, salinity, pH, and DO in the upper 3 m of the surface water. Probes were calibrated daily. Plankton samples were collected using a 25 µm mesh net towed from the boat for approximately 2–3 minutes. Phytoplankton samples were stored in 1 L polypropylene containers and fixed with Lugol’s iodine. Samples were stored in a 3°C refrigerator. Approximately 10-mL of water was extracted from the sample bottle using a syringe, filtered through a 0.45 µm filter, acidified with trace-metal grade nitric acid and preserved in a 15-mL polypropylene centrifuge tube for metal analyses. All containers were pre-cleaned with detergent, rinsed thoroughly with tap water, and then with ultra-pure Milli-Q® water (Millipore Sigma, Burlington, MA). Containers used for metal analyses were washed as specified above with addition of a 10% nitric acid rinse prior to rinsing with ultra-pure Milli-Q® water.
Water chemistry analyses
Alkalinity and hardness of each water sample was measured in duplicate using titrimetric assays and standard methods (Greenberg 1985). Additionally, colorimetric assays (LaMotte® Colorimetric kit (Chestertwon, MD) were used to measure ammonia-N, nitrate-N, nitrite-N, and phosphate-P in the water samples, in duplicate.
Metal analyses
Filtered and acidified water samples were diluted with ultra-pure Milli-Q® water and measured in triplicate for cadmium, copper, nickel, lead, silver, and zinc concentrations using atomic absorption spectrophotometry (AAS; Perkin Elmer AAnalyst 800, Waltham, MA) with graphite furnace detection. Blanks and certified standards (Fisher Chemical, Fairlawn, NJ) were also analyzed in triplicate in every analysis. Recalibration of standards was performed every 40 samples and quality control samples were measured periodically throughout each analysis. Detection limits for each metal were as follows: 0.005 µg/L silver, 0.02 µg/L cadmium, 1.06 µg/L copper, 2.60 µg/L nickel, 0.70 µg/L lead, and 0.52 µg/L zinc.
Diatom identification
Preserved phytoplankton samples were concentrated by sedimentation. A portion of the concentrated material was cleaned with nitric acid/potassium dichromate and rinsed 10 times with deionized (DI) water. Aliquots of the cleaned material were air-dried onto square coverslips, mounted on microscope slides using Naphrax and observed in light microscopy (LM) using an Olympus BX60 microscope equipped with differential interference contrast (DIC) optics and a Canon EOS Rebel digital camera or air-dried onto circular glass cover-slips mounted on an SEM stub using double-sided carbon tape, sputter-coated with gold/palladium using a Denton Vacuum Desk IV sputter-coater, and examined using a JEOL 6480 LV scanning electron microscope.
Counts were made of at least 600 valves for each sample, separated according to morphological groups (taxa) distinguishable at 1000x times magnification; when one taxon showed overwhelming dominance, additional valves were counted so that at least ten groups reached counts of at least ten valves. Taxa were identified to the least practical taxonomic unit, sometimes with the aid of SEM observations, using standard references as a starting point (Bishop et al. 2017, Hasle & Syversten 1996, Spaulding et al. 2021).
Species richness for each sample was estimated using Chao 1:
$${\text{S}}_{Chao}= {S}_{obs} +\frac{{f}_{1}^{2}}{2{f}_{2}}$$
where SChao is the estimated number of taxa, Sobs is the number of taxa observed, f1 is the number of taxa represented by a single valve, and f2 is the number of the number of taxa represented by two valves (Chao 1984, Gotelli & Colwell 2011). Diversity was calculated using a) the Shannon-Weaver Diversity Index (Shannon 1948, Shannon & Weaver 1963) given by
\({H}^{{\prime }}= -\sum _{i=1}^{k}{p}_{i}\text{l}\text{n}({p}_{i}\))
where k is the total number taxa observed and pi is the relative abundance of species i; and b) the Simpson Index of Diversity (Simpson 1949) given by
$$S=1- \sum _{i=1}^{k}\frac{{n}_{i}^{2}}{{N}^{2}}$$
where k is the total number of taxa observed, ni is the number of valves belonging to taxon i, and N is the total number of valves counted (see Morris et al. 2014).
Data analyses
Sigma plot 15.0 was used to determine significant differences in a particular water quality parameter among sites, over time, and to determine relationships among parameters. The data were tested for normality and equality of variance using a Shapiro-Wilk and Brown-Forsythe test, respectively. Significant differences (p ≤ 0.05) in a particular variable among sites and over time were determined using one-way analysis of variance followed by pairwise multiple comparison procedures (p ≤ 0.05). Correlations among measured parameters were determined using Pearson Product Moment correlation test (p ≤ 0.05). A positive correlation between variables was determined by a positive correlation coefficient and p ≤ 0.05; and a negative correlation between variables was determined by a negative correlation coefficient and p ≤ 0.05.
Metal concentrations in the water samples were compared to the Florida ambient water quality standards, including EPA class III water quality criterion values for both freshwater and saltwater/marine (haline; surface chloride concentration ≥ 1500 mg/L) to determine the health of the river at a particular site. A hardness-based adjustment (water hardness of 100 mg CaCO3/L) was made in calculating the freshwater criteria for cadmium.