Water-sand contrasts and site-associated variation in bacterial and eukaryotic communities along the Korean coast

Microbial assemblages can differ between coastal water and adjacent Sea-sand, yet bacteria and eukaryotes are seldom evaluated together within a paired sampling design. We profiled 16S and 18S rRNA gene amplicons from the same 75 samples collected at seven South Korean coastal Sites in Spring and Autumn. Sea-sand and Water were sampled at corresponding positions where available, and fixed positions were revisited when possible. Descriptive profiles and observed-data effect sizes showed Sample type-associated, among-Site, and Season-associated variation. Under stricter Site-synchronized exact permutations, however, none of the four predefined beta-diversity contrasts per marker had q < 0.05, limiting inference beyond the sampled Sites. For 16S, Shannon, Simpson, and inverse Simpson diversity were higher in Sea-sand after Benjamini-Hochberg correction across five predefined contrasts per metric (q = 0.0391), but not after correction across all 30 tests (six metrics × five contrasts; q = 0.0781). Richness metrics and all 18S alpha-diversity contrasts were nonsignificant after their corresponding corrections. The marker-specific multivariate structures were strongly concordant. Environmental analyses were exploratory and limited to seven Spring Sea-sand Site centroids. These findings support treating Water and Sea-sand as separate sampling strata while requiring broader Site replication before generalization.IMPORTANCECoastal biodiversity surveys often examine water or nearby sand separately and may also analyze bacterial and eukaryotic communities independently. Using co-collected Sea-sand-Water pairs and a common 75-sample 16S/18S data set, we observed Sample type-associated and among-Site variation in both markers, together with strongly concordant multivariate distance patterns. However, Site-synchronized analyses provided limited support for generalization beyond the seven sampled Sites: only Shannon, Simpson, and inverse Simpson diversity met the Benjamini-Hochberg criterion across five predefined contrasts per metric, and none of the predefined beta-diversity contrasts per marker had q < 0.05. These findings support sampling Water and Sea-sand as separate strata, retaining spatial replication, and revisiting fixed locations over time. More broadly, they emphasize that statistical inference in coastal biodiversity surveys should reflect the number of independently sampled Sites and avoid unsupported attribution to environmental causes, ecological mechanisms, or broad geological categories.