Alluvial aquifers can provide ecosystem services and drinking water, but much remains unknown about human effects on aquifer microbiomes. Therefore, we used amplicon sequencing and hydrochemical characterization to pair microbial communities with environmental conditions across 37 alluvial aquifer wells. The study region spanned eastern Iowa and southern Minnesota (USA) and contained a combination of drinking water and monitoring wells. In terms of microbial ecology, dominant phyla across the wells included Proteobacteria, Bacteroidota, Patescibacteria, Planctomycetota, and Nitrospirota. Tritium, an indicator of infiltration and surface water influence, was the highest correlated variable with the Shannon index (α-diversity) by the Spearman rank sum (ρ = 0.60) and one of only four significant environmental variables in the constrained correspondence analysis. We built random forest regression models to predict tritium concentrations from microbial family relative abundance (held-out testing coefficient of determination (R²) = 0.77 and mean absolute percentage error = 7%) and interpreted the models with Shapley additive explanation values. The most important families for predicting tritium concentrations were Nitrosopumilaceae and Methylomirabilaceae. Upwelling methane could contribute to the unusual coupling of ammonia oxidation by Nitrosopumilaceae with simultaneous nitrite-dependent methane oxidation by Methylomirabilaceae. Taken together, we illuminate the relationship among hydrochemistry, hydraulic connectivity, and alluvial aquifer microbiomes.