Groundwater contamination is receiving increasing attention in the UK. It will be highlighted this year with the introduction of the Groundwater Regulations (1998), and the expected introduction of part II(a) of the Environmental Protection Act (1990) and works notices under section 161A of the Water Resources Act (1991) this summer. This legislation gives new powers and responsibilities to regulatory bodies to prevent groundwater pollution and to manage contaminated land and groundwater.
Of the many types of contamination, organic contaminants are often the most troublesome as they can be environmentally significant at low concentrations and very difficult to remove. A particularly problematic group of contaminants are the dense non-aqueous phase liquids (DNAPLs) which include chlorinated solvents (commonly used as degreasers and cleaners), PCB transformer oils, some pesticides and coal tars. Some chlorinated hydrocarbon DNAPLs and/or their degradation products are both suspected human carcinogens and highly recalcitrant, with the potential to contaminate aquifers for decades or even centuries.
Chlorinated solvents are typically characterised by densities greater than water, low solubilities and low degradabilities (Pankow & Cherry, 1996). They represent a significant threat to aquifer systems where they can 'sink' as a discrete, immiscible liquid. Historical processes which would today be unacceptable have resulted in the majority of urban groundwater being contaminated with chlorinated hydrocarbons, much of it with concentrations above the recommended drinking water standard of about 10mg/l (Burston et al).
With this issue in mind the Engineering and Physical Sciences Research Council (EPSRC) last year began funding a three-year research project led by Professor David Lerner and the Groundwater Protection and Restoration Group (GPRG) at the University of Sheffield, with support from the BGS, the Environment Agency and ICI Halochemicals. The project is investigating methods to characterise DNAPL contamination of the Permo-Triassic sandstone, one of the two major aquifers in the UK. The aim of the project is not to investigate remediation technologies, but to learn about the processes and pathways of DNAPL migration in fractures, in order to better understand penetration depths as an aid to risk assessment.
The approach to the complex problem of determining DNAPL penetration depths is one of multi-scaling. At the field scale, the team is trying new ways of characterising fracture networks. At the borehole scale, variation in fracture aperture is being studied. In the laboratory, changes in physiochemical properties are being examined. All of this will contribute to a Geographic Information System (GIS) based risk assessment package of aquifer vulnerability to groundwater contamination by DNAPLs. This 'risk' component is being undertaken by Nigel Tait and is supported by the Environment Agency's National Groundwater and Contaminated Land Centre.
Rock mass characterisation
The migration, and resulting spatial distribution, of DNAPLs in a fractured rock mass is primarily a function of the orientation and extent to which fracture sets intersect and link together (Figure 1). It is generally accepted (Renshaw, 1996; Bolt et al, 1995) that the geometry of complex fracture networks cannot be described deterministically; instead, a stochastic modelling approach may be adopted, where fracture networks are generated with statistical fracture properties that can be compared directly with those observed in the field. A discrete fracture network modelling approach is being used in which:
the potential DNAPL migration pathways are represented by a 3D fracture network;
the two-phase flow component is not modelled directly; and
fracture statistics are used to predict fracture-network connectivity.
A summary of the modelling approach, the applicability to immiscible phase flow, and its relevance in the Triassic sandstone is summarised in Figure 2. This part of the study will also assess the validity of upscaling fracture parameters measured at the borehole scale (cm) to those observed at the outcrop or regional scale (100s of m to km), with the aim of constraining future studies of DNAPL migration in dual porosity aquifers.
Fracture aperture studies
DNAPLs are typically immiscible with groundwater. As a result, their subsurface migration is controlled by capillary pressure rather than conventional groundwater transport mechanisms. Consequently, DNAPLs migrate downwards following the path of least capillary resistance, which can often mean backwards or sideways against the prevailing groundwater gradient. The pressure required for a DNAPL to enter and therefore pass through a fracture is a function of the fluid properties and the fracture aperture. To understand the migration pathways and penetration depths of DNAPLs in fractures, a range of fracture apertures should be determined.
This aspect of the project will attempt to determine fracture aperture distributions insitu, by calculating aperture from a novel field test involving injection of a DNAPL surrogate into fractured Permo-Triassic sandstone.
DNAPL fluid properties
The physical properties of DNAPLs have a considerable influence on their transport and eventual fate, where migration is dominated by the driving forces of density and viscosity. In addition, both interfacial tension and contact angle - used as a measure of aquifer 'wettability' (its preference for one liquid over another) - are key in determining the liquids' entry into the fracture network or matrix. While considerable information exists for 'pure' DNAPLs, there is little knowledge on the properties of 'field' or 'waste' samples. Similarly the influence of previous exposure to contaminants has largely been ignored, although work by Powers et al (1995) and Thakur et al (1995) point to their significance.
The aim is to further the understanding of the movement of these liquids by measuring the physical properties of field or waste samples, and linking this in with their chemistry. Implicit within this work is the construction of a predictive properties model based on the results. Second, the influence of prior exposure is being examined. Provisional results suggest that the physical properties of 'used' solvents can differ considerably from those of pure solvents, because of the presence of other materials within the DNAPLs. Additionally, prior exposure to the immiscible phase can drastically change aquifer wettability whereby the capillary network will take in these liquids rather than representing a barrier to migration.
Analysis and modelling of the complex processes which determine the transport and fate of DNAPLs in the Permo-Triassic sandstone aquifer requires the handling of a large variety of data, ranging from physical to socio-economic parameters, all of which have a spatial reference. The use of GIS technology is therefore straightforward, and will provide a framework for risk assessment and decision support activities.
The core of the framework will be a risk matrix, adapted for the analysis of groundwater contamination. It will allow the consideration of all relevant factors influencing risk situations. Emphasis will be put on spatial components and high-resolution spatial analysis of DNAPL aquifer contamination. GIS technology will be applied as an instrument for spatial analysis and for integrating the various models and highly heterogeneous data (geology, hydrology etc). Overall, the anticipated framework will provide spatial decision support facilities to the Environment Agency as different scenarios can be postulated and the influence of the different models and data can be evaluated.
The integrative character of the framework is expected to open new perspectives and stimulate further research in areas found to be particularly important in governing the fate of DNAPLs in the subsurface.
Bolt, JE, Bourke, PJ, Jefferies, NL, Kingdon, RD, Pascoe, DM and Watkins, VMB, 1995. The application of fracture-network modelling to the prediction of groundwater flow through highly-fractured rock. Nirex Radioactive Waste Disposal Safety Studies Report NSS/R281, United Kingdom Nirex Limited.
Burston, MW, Nazari, MM, Bishop, PK & Lerner, DN 1993. Pollution of groundwater in the Coventry region (UK) by chlorinated hydrocarbon solvents. Journal of Hydrology, 149, pp137 - 161.
Pankow, JF & Cherry, JA (eds). Dense Chlorinated Solvents and other DNAPLs in Groundwater. Portland, Oregon. Waterloo Press, 1996. 522pp.
Powers, SE, et al (1995). Wettability of Porous Media After Exposure to Synthetic Gasolines, Journal of Contaminant Hydrology, 19(2), 105-125.
Renshaw, CE, 1996. Influence of subcritical fracture growth on the connectivity of fracture networks. Water Resources Research, 32 (6), 1519-1530.
Thakur, S, et al (1995). Flow Characteristics of DNAPLs in Sand Packs: Water Influx Displacement, Journal Env Sci Health, 30(5), 1105-1118.