An audiomagnetotelluric survey of the Carnmenellis granite

The electrical resistivity of competent/unfractured rock depends mainly on porosity and the conductivity of the pore fluid. In fractured/jointed rocks a further degree of dependence on the degree of fluid saturation and permeability is observed; The audiomagnetotelluric survey described in this report is primarily concerned with mapping resistivity variations (i.e. rock/fluid properties) through the critical depth range (2 to 8 km) of geothermal energy extraction within and across the Carnmenellis granite. This report describes the audiomagnetotelluric experiment that was carried out during 1988. Two field surveys were performed. A main survey to assess the presence of a refractor (a wide-angle reflector) at about 7 to 8 km (Brooks et al., 1984) and a secondary survey to assess the depth e_xtent of a surface lineament. The principal results of the field experiment are now summarised. A seven-site investigation of a NW-SE surface lineament was carried out at spacings of 50m. To our knowledge, this is the first audiomagnetotelluric experiment involving such small spacings. The survey data were partially marred by an additive noise component in one of the electric field channels. Subsequent detailed analysis of these data revealed electric fence noise as the cause. Although such hindsight detection is unfortunate, the uncontaminated data offer valid constraints in relation to the detection of a fluid-filled fracture zone. The results presented indicate that it is highly unlikely that the target lineament can be associated with a conductive zone possessing a reasonable resistivity contrast which intersects the surface or near-surface. The results of the main experiment extend this null interpretation to the majority of surface lineaments as described below. The main granite survey consisted of 17 soundings in total. An E-W profile across the granite outcrop comprised 12 soundings and two further soundings were conducted off the outcrop for control. The two off-granite, soundings provided very different 'dimensional' characteristics to soundings on the granite. They therefore confirm the broad homogeneity of the geoelectric anisotropy on the granite. The two off-granite soundings also provided information on the resistivities of the Devonian cover rocks and the depth to granite. A further 3 soundings were performed in the vicinity of the Hot Dry Rock (HDR) reservoir at Rosemanowes quarry. Overall the Carnmenellis granite, as defined by its broad resistivity characteristics, appears predominantly homogeneous. A very consistent set of resistivity values are found for the whole granite structure below depths of 2 km. The majority of resistivity values are in the range 1000 to 10,000 ohm.m. When laboratory analyses of thermal and pressure dependence are taken into account the granite is found to correspond to a 'wet' granite saturated with several weight-percent of free water down to at least IO km. · The consistency of the resistivity values below 2 km across a major portion of the granite indicates that lateral geoelectric effects are likely to be confined to the near-surface (i.e. < 2 km). An examination of the lateral anisotropic behaviour across the granite confirms this general conclusion. We conclude that with the exception of one location the surface lineations, within the area surveyed, do not appear to represent deep vertical or sub-vertical zones with rock/fluid properties that would distinguish them from 'background' properties. Judging by the hydrogeological models for the near-surface granite, a spatially-complex hydrothermal circulation system can be considered to operate at depth (Gregory and Durrance 1987). The spatial distribution of such a system is likely to be closely tied to the distribution of major water-conducting fractures. Although the scale and degree of resistivity contrasts should be considered, the main survey profile has detected only one such near-surface feature. Significantly this conductive zone is spatially localised and is directed NE-SW. The zone appears ·to be correlated with a lineament and a main arterial alluvial fan of the survey region. The E-W survey profile intersects the NE-SW trending zone 0.5 km directly south of the village of Porkellis. There is an indication of a conductive layer at a depth of I km in the vicinity of this lineament. The sites defining the SE portion of the granite display anisotropic features which are different from those to the west. It is suggested that the cause is associated with boundary or off-granite variations rather than with variations across the outcrop although this cannot be ruled out. The features observed in the SE appear to define a broad, large-scale effect. Resistivity values through the granite also appear slightly larger in this region of the outcrop. The lateral geoelectric anisotropy transfers (i.e. azimuths rotate) from local and near-surface penetrations (e.g. 1 to 2 km) to a regional scale anisotropy at large-volume penetrations (e.g. in excess of 25 km). The rotation pattern is consistent at the majority of survey locations beginning NE/ENE at depths of about l km to NINE at depths of 6 to 10 km and thence to NW/N at depths in excess of 25 km. The information on geoelectric anisotropy has been compared with the principal joint and stress directions of the granite in order to identify the mechanism controlling the resistivity variations. The main conclusion is that the directions of resistivity anisotropy do not display any persistent alignment with the principal horizontal stress directions. Such a conclusion assumes that the present indicators of stress directions are representative of the in situ stress at depths in excess of 2.5 km. On this basis then the mechanism of aligned microcracks does not appear to control the observed geoelectric anisotropy. The results indicate that within the upper 1.5 km (at least), resistivity is controlled by one of the two principal joint systems of the granite. The results identify the fracture system parallel to the NW-SE master joints as being preferentially 'open' and containing enhanced concentrations of fluids. A definitive interpretation (from geoelectric anisotropy) at greater depths appears to be restricted by the observed 'intermediate' rotations as other larger-scale (regional) effects become more dominant. The vertical field is, however, influenced by a major resistivity contrast striking NW-SE beyond, and to the SW, of the Carnmenellis outcrop. Below a depth of 2 km resistivity values increase slowly with depth attaining maximum values by about 6 km. The anticipated linear decrease of resistivity with increasing temperature is not observed and a more dominant pressure/stress dependence must control the spatially-consistent. depth dependence observed. Laboratory measurements on a wide range of granitic rocks indicate that transfer from crack-dominated behaviour to pore-dominated behaviour will be complete by an applied pressure of 200 MPa. The extrapolated overburden and stress magnitudes, from the HDR borehole measurements. suggest that this will be achieved within the Carnmenellis by about 6 km. The resistivity profiles are therefore consistent with the completion of crack closure by a depth of 6 km and a transfer to a pore-dominated resistivity mechanism below this depth. Thus, in simple terms, if a 'joint' can be defined as a feature that is capable of 'closing' (and closed here means an inability to support ionic conduction of interstitial fluids) the observations suggest the absence of such joints below 6 km. A comparison of the vertical geoelectric profiles across the granite with the boundary reflector (RI) of the deep low-velocity zone modelled by Brooks et al. (1984) is possible. The resistivity profiles however do not reveal any spatially-consistent major discontinuities in the upper 12 km. The depth interval of the low-velocity zone appears merely to be associated with an interval of approximately constant and maximum resistivity. We conclude that no detectable geoelectric variations in rock/fluid properties can be identified at a depth associated with the Rl reflector. Thus RI does not appear to represent the upper surface of a fractured zone with an associated enhancement of conducting fluids. The three sounding sites above and around the HDR reservoir provide interesting results in the upper 5 km. The identical resistivity profiles at two of the sites display a more conductive profile when compared with a third site some 1.1 km away. The two sites, 500 m apart, appear to define a 'reservoir-influenced' section down to depths of 4 to 5 km. This depth appears consistent with the termination depth of the microseismic zone defined during hydrofracturing. The finite lateral extent of the low-resistivity reservoir volume is also apparent in the azimuthal (anisotropic) behaviour across the three locations. Thus the zone of enhanced fluid concentration (i.e. the reservoir) has been shown to have finite dimensions both laterally and vertically.