Hydrogeological modeling to improve remediation strategies for a drinking water aquifer contaminated by an aqueous phase liquid

• Main Author: Bergvall, Martin
• Format: Doctoral or Postdoctoral Thesis
• Language: Swedish; English
• Published: 2011
• Subjects: Other earth sciences; Water engineering; Oceanography; Hydrology; Water Resources; Other Earth and Related Environmental Sciences; Northern Sweden
• Online Access: https://pub.epsilon.slu.se/8065/; https://pub.epsilon.slu.se/8065/1/bergvall_m_110412.pdf

In many communities, groundwater is an important source of drinking water. Groundwater aquifers are, however, vulnerable to the widespread and increasing problem of contamination from anthropogenic sources. Once in the groundwater, contaminants are likely to remain there for a long time as the attenuation rate is slow. In this thesis, different tools for modeling subsurface transport were adapted and evaluated in order to improve remediation strategies for a contaminated esker aquifer. The work focuses on the entire transport process at a regional scale from the source at the soil surface, through the vadose zone, and in groundwater. Few comparable studies exist, especially for aquifer systems in glaciofluvial sediments. The studied aquifer supplies drinking water to the municipality of Umeå, which is a medium-sized city in northern Sweden. The aquifer is contaminated by the commonly found pesticide degradation product 2,6-dichlorobenzoamide (BAM). Hydrogeological and chemical field data were collected, and the contaminant migration analyzed by a stationary non-distributed model and a transient distributed model. To remediate the aquifer so that it meets the drinking water standard, it was necessary to combine extraction at two up-gradient wells, with an increased rate of artificial recharge via two infiltration ponds. Using only one of the techniques would either affect the water balance negatively, or would increase the risk of clogging the infiltrating surface. However, in order to reinstate the two up-gradient wells as producers of drinking water as soon as possible, it was necessary to establish the remediation wells in close proximity to the contaminant source. When the data quality is insufficient the simple mass-balance model was found to be most useful, since it reflects the uncertainty of the result. However, if it is essential for the contaminant transport to be calculated more accurately, a distributed model is required. To strengthen the credibility of such a model, it should be validated with independent data from various sources: in this study it was stable isotope oxygen-18 data, data on the BAM contamination, and time-variant hydraulic head data. The overall findings are expected to be relevant to many other sites in similar settings.