Predicting sorption of groundwater bacteria from size distribution, surface area, and magnetic susceptibility of soil particles
Research output: Contribution to journal › Article
Spatial heterogeneity of hydraulic and sorptive processes may have a strong impact on the mobility of bacteria in porous soils. Whereas heterogeneity at the microscopic solid particle/interstitial pore scale has been conceptually addressed by dividing the porous material into two or more coexisting sorbing fractions, field-scale spatial variability has been approached by developing stochastic methodologies for coupling advective flow and sorption reactions. This calls for an elucidation of correlations between certain soil parameters and sorption of bacteria, with the purpose of substituting tedious and labourious sorption measurements with indirect methods allowing more rapid screening of field samples. Sorption of two H-3-labeled strains of groundwater bacteria was examined in test tubes with sandy aquifer material and filter sterilized groundwater from the same aquifer. The aquifer particles were sieved to different size fractions (0-63, 63-106, 106-212, 212-500, 500-1000, and 1000-1500 mum) and mixtures thereof, and their specific surface area, magnetic susceptibility, and mineral composition were determined. Columns filled with the sieved or original aquifer material were saturated with groundwater, and the breakthrough of the labeled bacteria at a flow rate of 0.2 mL min(-1) was measured by liquid scintillation. Cell sorption to soil particles in test tube experiments was controlled by magnetic susceptibility, which characterizes the mineral surfaces primarily according to their iron content. In contrast, there was a negative nonlinear regression between the sorption coefficient and the particle size and a positive linear regression with the specific surface area, each with an R-2 Of similar to0.85 in the column experiments. The work demonstrates that sorption of groundwater bacteria to aquifer material of low organic carbon content can be predicted from the particle size distribution or the specific surface area.