The discrete fracture network (DFN) concept assumes flow through a fractured rock is predominantly through an inter-connected network of flow-conductive fractures with groundwater moving from one fracture to another at the intersections between them. The properties of the network are usually characterised in terms of:

• Spatial distribution (e.g. Poissonian, fractal, clustered around points or hydrozones/lineaments);
• Fracture intensity (and its spatial variation);
• Fracture sets distinguished by orientation;
• Fracture size (e.g. log-normal, power-law distributions);
• Transmissivity-size relationships.

For individual fractures, either deterministic or stochastic, their properties are primarily:

• Size;
• Orientation (strike and dip);
• Transmissivity (and possibly spatial variability within the fracture);
• Transport aperture;
• Storativity.

In ConnectFlow, fractures are usually rectangular, but may be right-angled triangles where a complex surface has been triangulated into many pieces. For stochastic fractures, the properties are sampled from probability distribution functions (PDFs) specified for each fracture set. The properties may be sampled independently or correlated with other properties.

The DFN concept is very useful since it naturally reflects the individual flow conduits in fractured rock, and the available field data. However, DFN models can sometimes become computationally challenging, and to understand flow and transport on a larger regional-scale it is often necessary to consider the larger-scale bulk properties in the context of an Equivalent Continuous Porous Medium (ECPM) concept. This requires methods to:

• (i) convert the properties of a network of discrete fractures of lengths less than the continuum blocks into ECPM block properties, known as upscaling, and
• (ii) represent larger scale features such as fracture zones by appropriate properties in a series of continuum blocks, i.e. a downscaling method.

The characterisation of fractures in rocks and their role as conduits for fluid has significance to a wide range of applications including underground disposal of radioactive waste, oil and gas production, storage of carbon dioxde, geothermal energy and underground construction.

DFN models have found use in many areas, including:

• Geological radioactive waste disposal, for site-characterisation and safety assessments.
• Oil and gas industry to aid well planning, simulation of well tests, and the parameterisation of oil simulation software by calculation of up-scaled equivalent continuum parameters.
• Civil engineering projects concerned with construction or groundwater remediation in fractured rock. This includes the estimation of water ingress due to excavation of tunnels, studies of underground oil caverns, dam construction in fractured rocks, remediation or containment of contaminated fractured sites.