Thermal imaging of sparse permeable fractures embedded in intact granite

Subsurface processes in rocks are often sensitive to the presence of fractures and their ability to transport fluids, solutes and heat, which depends in turn on their geometry. Characterizing parameters like fracture length and orientation is therefore an important step towards making realistic flow and transport simulations. Despite advances in computational power, state-of-the-art approaches such as discrete fracture network (DFN) models still tend to be conditioned to conventional data sources, generally hydraulic tests or solute tracer tests, providing little constrains on 3-D fracture attributes (e.g. orientation). This highlights the need for novel experimental frameworks to better resolve complex flow patterns in fractured rocks with DFN models. Here we show how borehole thermal anomalies, caused by natural flow or by heat injection experiment, can be used to detect and characterize the orientation of permeable fractures. Heat has been increasingly used as a tracer in recent years owing to the commercialization of Raman scatter-based fiber optics distributed temperature sensing (DTS) systems tailored to environmental applications. We thus present a simple framework based on analytical and/or numerical methods to extract structural information on fractures intersecting or located nearby boreholes equipped with DTS systems. Our model assumes a single fracture, embedded in an impermeable rock mass, and is validated against a series of cross-hole thermal tracer tests performed in crystalline rock at the Grimsel Rock Laboratory, in Switzerland. Active heat injection was carried out by heating water using an electrical flow-through heater up to 45°C for a duration 40 days. Fluid injection took place across a discrete, 2-m long interval packing off a single flowing fracture. A continuous fiber optics loop was deployed along three fully-grouted boreholes, which managed to record thermal breakthroughs of 1-2°C up to 6-7 meters from the injection point.  We find that orientations constrained from thermal anomalies do not necessarily correspond to structural orientations of borehole fracture traces. This orientation is defined instead by the borehole axis and maximum thermal gradient along heat-carrying fractures. Such parameter provides information on the spatial organization of discrete flow paths and may offer an alternative calibration parameter to constrain flow and transport simulations on DFNs.