Requirement of ecological replication with independent parallel analysis of each replicate plot to support soil remediation

Consequences of pseudoreplication or the sole use of plot averages even within one ecosystem on the reliability of laboratory remediation tests have barely been studied even though they may mask field variability. Soils from three field plots were tested separately to reveal the consequences of inadequate use of replicates. Parallel remediation experiments consisting of treatments control (non-contaminated soil), natural attenuation (gasoline-contaminated soil) and biostimulation (gasoline-contaminated and methylene urea fertilized soil) were
conducted to each plot soil. Three destructive sampling days and three pseudoreplicates per treatment, per plot and per sampling day were designed. The concentrations of total and individual gasoline aromatics were followed
using both within-plot pseudoreplicates and plot averages. Results showed that natural attenuation of total or individual gasoline aromatics depended on field plots and that fertilization showed enhancing, inhibiting and negligible effects, respectively, also depending on field plots. Additionally, plot averages masked withintreatment variation between field plots. Conclusively, within-plot pseudoreplication using a single field plot or interpretations based solely on plot averages cannot reveal remediation patchiness observed in practice. Due to
the high soil heterogeneity, ecological between-plot replication together with the parallel analysis of each replicate is encouraged in laboratory studies aiming to design in situ remediation strategies.