Modelling the influence of solution-enhanced conduits on catchment-scale contaminant transport in the Hertfordshire Chalk aquifer

Abstract

Catchment-scale (>40 km2) contamination of the Chalk aquifer of Hertfordshire by bromate, emanating from a disused industrial site north of St Albans, represents the largest occurrence of point-source groundwater contamination in the UK. The influences of ‘double porosity’ diffusive exchange and rapid transport along solution-enhanced conduits complicate predictive modelling of contaminant transport to threatened public supply wells. Tracer testing indicates that solution-enhanced flow routes exist beyond the surface distribution of dissolution features in Hertfordshire, more extensively than previously thought. A quantitative conceptual understanding of this flow system has been incorporated into a spatially distributed equivalent porous media representation in MODFLOW and MT3D-MS. The calibrated model reproduces essential features of the aquifer system, including heads and flows, seasonal responses, and the timing and spatial distribution of observed tracer breakthroughs in the solution-enhanced aquifer, but does not fully capture the magnitude and form of tracer and bromate advance. Due to the influence of local solution enhancement and matrix effects, detailed breakthroughs at receptors cannot be resolved at the coarse grid scale. However, the model is able to simulate general trends.

Diagnostic tests for conceptualizing transport in bedrock aquifers

“…In the Lee Valley north-east of London, source protection zones for three public-supply wells gave 50-day time-of-travel zones that extended about 1 km upgradient from the wells. However, tracer studies showed that the travel time for a distance of 16 km to the wells was a mere four days, demonstrating that the modeled source protection zones were far too small (Cook et al, 2012;Worthington, 2015).…”

“…Total porosity in the aquifer is 38%, but travel times of artificial tracers from sinking streams and of a long-persistence bromate plume were matched using MODFLOW and MT3DMS by using an effective porosity of 0.01% along flowpaths from sinking streams, and an effective porosity of 1-5% in the remainder of the model (Cook et al, 2012). A second model used a dual-continuum approach, with an effective porosity of 2% to match seasonal variation in water levels in one continuum, and an effective porosity of 5-10% in the second continuum to match drainage of the matrix over multi-year droughts (Taylor et al, 2012).…”

“…Artificial tracers flowed 16 km through the aquifer to springs and wells, with a mean residence time of 11 days (Cook et al, 2012). However, at some of the same springs and wells there is still substantial bromate contamination with near-peak concentrations from a spill that occurred more than thirty years ago.…”

Introduction: groundwater resources modelling: a case study from the UK

“…Two of these papers consider major pollution incidents on the Chalk aquifer and the simulation of contaminant transport at a regional scale (Cook et al 2012;Watson et al 2012). These papers form a useful contrast with the papers written by the practitioners, as they give some insight into the ways in which the regional models of the Environment Agency can be used, improved and/or adapted to assess future trends in water quality.…”

“…These papers form a useful contrast with the papers written by the practitioners, as they give some insight into the ways in which the regional models of the Environment Agency can be used, improved and/or adapted to assess future trends in water quality. In the case of Cook et al (2012), the contaminant transport model has been partly based on the Environment Agency regional groundwater model. Some of the academic papers also provide an insight into the information exchange between practitioners and academics.…”