Investigation of NA processes at reactive fringes: Sampling bias introduced by high resolution multi-level monitoring

Monitored natural attenuation is a potentially valuable risk-based remediation strategy for contaminated groundwater. The most important mass-removal process for natural attenuation is biodegradation. Certain zones or fringes of a contaminant plume offer supporting conditions for biodegradation: microbes, nutrients, contaminants and electron donors / acceptors are not only found together but also in the required reaction ratios. Due to this fact these areas show a relative rapid degradation and provide a significant contribution to the overall reduction of mass within the plume. As can be shown by high resolution numerical simulations of reactive transport in groundwater, the spatial distribution of these highly reactive zones, compared to the volume of the whole plume, is quite small and characterized by steep concentration gradients, which can not be detected using standard monitoring procedures. High resolution multi-level sampling (MLS) in the order of decimeters or less is an essential prerequisite for the investigation of NA processes at the reactive fringes at field scale. Furthermore, in contrast to technical remediation techniques which most often deal with high contaminant concentration levels close to the source zone, MNA relies heavily on the accuracy of the low concentration levels (down to the legal limits) measured in the plume. Quite often these data are strongly biased due to the monitoring equipment. This contribution presents results from ongoing controlled laboratory material tests and research on high resolution MLS at six field sites in different European countries. The focus was on a optimized site-specific hydraulic design and contaminant - MLS-material interaction. Most acceptable solutions (which means MLS resolution in the order of 0.1m) were found using sampling tubes with a small inner diameter (3-4mm). This results in a small stagnant water volume prior to sampling, but is still not problematic with respect to the flow induced hydraulic losses within the tube. In batch experiments under well controlled conditions a variety of commonly used sampling equipment materials were exposed to three typical contaminants (PAH, BTEX and CHC) to determine the respective (material and contaminant specific) partitioning- and intra-polymeric diffusion coefficients. These parameters were subsequently used in a simple analytical model, which considers partitioning coefficients, intra-polymeric diffusion coefficients, contact time and tube-volume to tube-surface to ratios, to predict the bias due to sorptive losses and thus the quality of the groundwater samples under the actual sampling conditions in the field. A first validation of these data sets is obtained by controlled flow-through experiments under field-like conditions (i.e. equivalent tubing length and sampling rate). The major conclusions are: (a) due to sorptive losses in the dedicated, pristine tubing material MLS systems tend to underestimate the contaminant concentrations which leads to an overestimation of the NA potential, to avoid this false negative concentration bias the most inert tubing material should be used; (b) due to leaching of plasticisers (e.g. Phtalates) out of the pristine tubing material MLS systems tend to overestimate TOC sum parameters, to avoid this false positive concentration only polymers with no or low plasticiser contents shall be used. Acknowledgements: This work was funded by European Commission project 'CORONA', Contract EVK1-CT-2001-00087.