Coupling among Microbial Communities, Biogeochemistry, and Mineralogy across Biogeochemical Facies

Abstract

Physical properties of sediments are commonly used to define subsurface lithofacies and these same physical properties influence subsurface microbial communities. This suggests an (unexploited) opportunity to use the spatial distribution of facies to predict spatial variation in biogeochemically relevant microbial attributes. Here, we characterize three biogeochemical facies-oxidized, reduced, and transition-within one lithofacies and elucidate relationships among facies features and microbial community biomass, richness, and composition. Consistent with previous observations of biogeochemical hotspots at environmental transition zones, we find elevated biomass within a biogeochemical facies that occurred at the transition between oxidized and reduced biogeochemical facies. Microbial richness-the number of microbial taxa-was lower within the reduced facies and was well-explained by a combination of pH and mineralogy. Null modeling revealed that microbial community composition was influenced by ecological selection imposed by redox state and mineralogy, possibly due to effects on nutrient availability or transport. As an illustrative case, we predict microbial biomass concentration across a three-dimensional spatial domain by coupling the spatial distribution of subsurface biogeochemical facies with biomass-facies relationships revealed here. We expect that merging such an approach with hydro-biogeochemical models will provide important constraints on simulated dynamics, thereby reducing uncertainty in model predictions.

Figure 1

(a) Spatial positions of sampling locations within the Hanford Site 300 Area, adjacent to the Integrated Field Research Challenge (IFRC) site, near the Columbia River, about 2 km upstream from Richland, WA. The figure was drafted using Adobe PhotoShop CS6 Version 13.0 × 64 using well-location survey data. (b–d) Photos of sampled sediments from each biogeochemical facies within well C7870. (e–i) Vertical structure of formations, facies, samples, and number of biological replicates for each data type. Only samples used for community composition profiling via 16S rRNA gene sequencing had more than one biological replicate per depth band; the number of biological replicates that successfully sequenced is indicated by the number of solid red circles at each depth. Microbial data were collected only from C7870 and C7867. The ‘other sediment data’ category included pH, organic carbon, AVS, Fe(II), pore structure, and mineralogy. Mineralogy was not available for elevation 101.6 m in C7869 and for elevation 94.2 m in C7867 (Table 2). Pore structure data were not available for elevation 101.6 m in C7869 (Table 2).

Figure 2. Mineral concentrations related to elevation.

Linear regression statistics are provided on each panel and dashed lines indicate the regression model; only minerals that were significantly related to elevation are shown.

Figure 3. CT reconstructions showing (Left) 3D volume reconstruction of reduced biogeochemical facies sediment; (Middle) isolated occurrences of biogenic iron sulfide minerals; (Right) a single-voxel slice reconstruction and its included iron sulfide framboid.

The position of that framboid in each representation is indicated.

Figure 4. qPCR-based estimates of 16S rRNA gene copies per gram of sediment—as a proxy of microbial biomass—in two wells, as a function of the vertical distance above the transition biogeochemical facies.

Multiple points at each elevation represent technical replicates. Solid lines connect mean gene abundances within each well.

Figure 5. Partial regression plots showing contribution of each variable retained in the OTU richness multiple-regression model, holding the other retained variable constant.

Both retained variables were significant (p ≪ 0.05) and the model R² was 0.65.

Figure 6. NMDS plot showing clustering of microbial communities based on Bray-Curtis dissimilarity.

Figure 7. Partial regression plots showing contribution of both variables retained within the βNTI multiple regression model, holding the other retained variable constant.

Figure 8

(Left) Spatial variation in the upper elevation of the oxidized biogeochemical facies across the IFRC field site and well C6209 (see Fig. 1). (Right) Spatial distribution of biogeochemical facies and predicted spatial variation in 16S rRNA gene copy abundance—a proxy for microbial biomass; facies-specific ranges in gene copy number (copies/g) are provided in Log₁₀ transformed units. The vertical elevation axis in both panels ranges from 96–100 m.