3.5. Environmental Factors
Effects on Trace Metal Concentrations. For all metals except arsenic, a significant and strong correlation between metal concentrations and TOC concentrations was found (P-values< 0.008) (Table S5 in the SI). TOC levels, as would be expected, showed a strong negative correlation (R=-0.69, P-value=0.028) with the sand percentage in bed sediment samples.
Effects on Microbial Communities. Correlation analysis showed no significant correlation among bed water quality parameters and alpha diversity indices. However, the results did show significant and negative correlations (R> 0.73, P-values< 0.038) among trace metal concentrations and the Simpson index. For the Shannon-Wiener index, all metals except arsenic and cadmium showed significant correlations. Copper, chromium, nickel, and lead concentrations at station S1 were identified as influential points.
The groups in the weighted Unifrac (Figure 4B) showed some similarities to the PCA plot developed using field water quality parameters, and grain size distributions as shown in Figure 9C. The PCA1 shown in the x-axis of Figure 4C accounted for 42.85% of variation in the dataset and was dominated by the level of sand, clay, silt, and TOC. The sampling stations were grouped into S1, S10, S2/S5/S6/S11, and S3/S4/S7/S8/S9. The similarities in groups between the weighted Unifrac and environmental PCAs (see Figure 9) and the fact that the x-axis in environmental PCA is describing sediment distribution may suggest that unlike rare bacteria, the distribution of common bacteria in samples could be a function of sediment composition in the system.
Figure 10 shows the results of the CCA for both scenarios described in Section 2.6, one with all environmental factors except metals and the other one with reduced size environmental factors including metals. For both scenarios, the results were very similar to the PCoA plot of the weighted Unifrac (Figure 4B). Extremophile phylum (rare) could be found on the right side of the CCA figures (Figure 5) with the closest station being S1. Among environmental factors, metals and TOC had the closest distance to S1, indicating the potential effect of both organic and inorganic pollutants on the microbial communities at S1. Station S7 through S11 all located in the side bays and Galveston Bay (see Figure 1B for locations) showed the shortest distance to salinity, DO, and pH while stations S1 through S5 located in the western and deeper parts of the system showed more proximity to total depth and sediment grain size variables. The permutation test for both scenarios showed P-values greater than 0.05 meaning failing to reject the null hypothesis that the observed correlation among the environmental factors and microbial diversities was due to chance. Overall, the lack of a significant correlation among metal concentrations and the field water parameters at the bed layer and among environmental factors and microbial diversities could be an indication that the system has not reached equilibrium in terms of chemical and microbial properties and partitioning among phases.