EVALUATING CONSISTENCY OF GEOCHEMICAL INDICATORS FOR REPRESENTATIVENESS OF GROUNDWATER SAMPLES COLLECTED FROM WELL SCREENS IMPACTED BY RESIDUAL DRILLING FLUIDS

Wells have been installed in and around Los Alamos National Laboratory to characterize its hydrogeologic setting and to understand potential contaminant pathways. In some of these wells, groundwater data quality has been compromised due to incomplete removal of drilling fluids during well development (Simmons et al., “Obtaining Representative Groundwater Quality Data from Heterogeneous Aquifer Systems,” these proceedings). Native groundwater is characterized as oxic: low but measurable NO3 and SO4; negligible dissolved Fe and Mn. However, biodegradation of residual organic drilling fluids by in situ microbes initiates a sequence of geochemical events: dissolved O2(g) is reduced to water, NO3 is reduced to N2(g), Mn oxide is reduced to dissolved Mn(II), ferric (oxy)hydroxide is reduced to dissolved Fe(II), and finally, SO4 is reduced to dissolved sulfide. Initial efforts to use geochemical indicators to identify historic and current water samples in which these drilling-related impacts are present are documented in the Well Screen Analysis Report (LANL 2005, Report LA-UR-05-8615). Current efforts build upon this initial approach by fine-tuning details of the geochemical tests used to flag such non-representative water samples. Our study focuses on evaluating internal consistency among indicators used to characterize redox conditions in these water samples. Application of the original test criteria (from the Report) to 350 relevant samples yields apparently thermodynamically-inconsistent outcomes for 31% of the cases (e.g., measurable NO3 apparently coexisting with SO4-reducing conditions). These inconsistencies are traced to the following primary causes: inappropriately-set threshold values, complications due to concurrent drilling-fluid related geochemical impacts, and inability of the original protocol to accommodate site-specific geochemical factors such as localized contaminant plumes. We address these inconsistencies by augmenting the original testing protocol with additional indicators, and by re-evaluating threshold values through a combination of improved statistical characterization of background groundwater geochemistry and more refined geochemical speciation modeling.