As a part of the long-term marine ecosystem research project, we studied microbial communities from 9 locations of two bay areas (Garolim and Gyunggi) of Yellow Sea, South Korea for one year. In Garolim area, there is active water exchange in the open sea without much influx of freshwater from land. In contrast, the Gyunggi area receives large quantities of nutrient-rich freshwater input from the Han River. The objective of the study was to provide baseline description of seasonal variations of microbial communities and relationships with environmental factors. Bacterial and archaeal communities were analyzed by targeting V4-V8 region of 16S rRNA gene using 454 GS-FLX Titanium pyrosequencing technology. Sequence reads were first processed through QIIME pipeline, and OTU table was prepared based on Greengenes 13_8 database with 0.97 similarity cutoff. Data analysis and statistical modeling were carried out with packages and scripts in R (v. 3.3.1). Total of 16 environmental factors were measured including salinity, pH, DO, nutrients and chl-a concentration ([chl-a]).
While archaeal diversity increased from February to October, bacterial diversity was higher in April and October and lower in July. Archaeal diversity was well modeled using pH, DOC, POC and PO43- with adjusted R2 of 0.751, but bacterial diversity was less well predicted with the measured environmental parameters (R2adj = 0.326). Overall community structures were similar between archaea and bacteria, and they were well correlated with the environmental factors. Seasonal patterns were clear for both archaeal and bacterial communities in that their community structures were distinctively clustered per season (P < 0.001 by PERMANOVA). However, two bay areas did not harbor distinctive archaeal and bacterial communities, and environmental factors. The RDA ordinations between archaeal and bacterial communities were similar (t = 0.641, P < 0.001 by Procrustes test), but the RDA models with environmental factors were distinctive. Archaeal communities were well explained by salinity, temperature, DO, NH4+, PO43-, SiO2 and [chl-a] with 56.8% explainable variance by the first 2 axes. Bacterial community model was built with salinity, temperature, pH, DO, NO2-, PO43- and [chl-a] with 39.7% explainable variance by the first 2 axes. Overall, we were able to identify clear seasonal patterns in both bacterial and archaeal communities, and distinctive diversity and associated environmental parameters between bacterial and archaeal communities, in which archaeal communities were better predicted by environmental conditions of Yellow Sea salt marsh.