A modelling experimental activity was developed to characterise the hydraulic behaviour of water-bearing fractures in crystalline rocks.
Three conservative tracers were injected into two packed-off sections of the same well, 69 m deep in a granite formation, and recovered by pumping from an isolated section of a second borehole, 46 m deep, 14 m apart. The concept of this design is to characterise separately an isolated fracture zone intersecting the lower parts of both wells from the fracture network intersecting the bulk of the rock.
Before using the tracers in the field, their behaviour was studied in the laboratory under controlled conditions. Fluorescein, eosin and iodide were finally chosen as the best spikes.
The field experiments were developed under strictly controlled conditions, such as (1) checking the hydraulic pressure in the packed-off sections and in other parts of circuits; (2) mixing the tracer solutions during the injection and checking their homogeneity; (3) performing a continuous and automatic monitoring of tracer concentration in the arrival well.
Iodide and fluorescein were injected in one section, and eosin in the other section. The pumping rate was maintained at 2 1 min. The test lasted 27 d after which from 40 to 60% of the injected masses were recovered.
Breakthrough curve analysis considered two conceptual models: radial advective-dispersive transport with and without matrix diffusion: the first model returns thickness-porosity values around 0.2 × 10 m and dispersivity around 4 m. Parameters are remarkably consistent for iodide and fluorescein, although fittings can be improved. The matrix diffusion model provides much better fittings by decreasing thickness porosity to 0.8 X 10-Z m. Dispersivities range from 0.5 to 0.9 m and the molecular diffusion term differentiates the behaviour of conservative tracers such as fluorescein and iodide.