Electrical Resistivity Tomography (ERT) surveys were undertaken at the Hydrogeological Experimental Site (HES) of Poitiers, France to identify karst intervals at depth of 35-40 m and 85-87 m host in the Dogger Limestone fractured-karst aquifer. Dogger Limestone occurs at a depth ranging between 30 and 120 m and is overlain by argillaceous limestone. Three-dimensional resistivity imaging was obtained from full inversion of combined 2D ERT data collected along five parallel 470 m long profiles with 50 m line spacing. Data were collected using both Wenner-Schlumberger (WS) and Pole-Dipole (PD) arrays. A 3D block measuring 481 × 203 m in size with a maximum depth of 100 m was surveyed. The hybrid array sequence (the combination of WS and PD array sequences) showed better imaging, due to the combination of the large vertical resolution of WS, large lateral resolution and penetration depth of PD. First, resistivity soundings have been sorted from ERT apparent resistivity data in order to verify the ERT detectability when identifying karst intervals. Then, synthetic dataset modelling was performed to assess the detectability of the ERT resistivity block when identifying karst intervals. Different starting models were implemented to perform forward modelling. The use of a very heterogeneous and anisotropic starting model implemented by using all information available from resistivity and stratigraphic logs and imagery of borehole walls provided unrealistic results. This is due to the too high level of detail of the starting model respect to the ERT system detectability and resolving ability. Then, a simpler, layered starting model comprising argillaceous limestone, low porosity limestone layers and karst intervals was implemented. Inversion of the synthetic dataset obtained by forward modelling was performed. Modelling allowed addressing different issues limiting the karst intervals detectability: an excessively low thickness with respect to the system's ability; an excessively large resistivity gap when compared to the shallow argillaceous limestone that also reduced the investigation depth and led to a severe underestimate of the limestone resistivity; variation of the layers resistivity with depth.
THE USE OF SYNTHETIC DATASET MODELLING TO ASSESS THE DETECTABILITY OF ERT SURVEYS WITH DIFFERENT ARRAYS WHEN IDENTIFYING KARST INTERVALS IN THE HES OF POITIERS (FRANCE)
TORRESE, PATRIZIO
2015-01-01
Abstract
Electrical Resistivity Tomography (ERT) surveys were undertaken at the Hydrogeological Experimental Site (HES) of Poitiers, France to identify karst intervals at depth of 35-40 m and 85-87 m host in the Dogger Limestone fractured-karst aquifer. Dogger Limestone occurs at a depth ranging between 30 and 120 m and is overlain by argillaceous limestone. Three-dimensional resistivity imaging was obtained from full inversion of combined 2D ERT data collected along five parallel 470 m long profiles with 50 m line spacing. Data were collected using both Wenner-Schlumberger (WS) and Pole-Dipole (PD) arrays. A 3D block measuring 481 × 203 m in size with a maximum depth of 100 m was surveyed. The hybrid array sequence (the combination of WS and PD array sequences) showed better imaging, due to the combination of the large vertical resolution of WS, large lateral resolution and penetration depth of PD. First, resistivity soundings have been sorted from ERT apparent resistivity data in order to verify the ERT detectability when identifying karst intervals. Then, synthetic dataset modelling was performed to assess the detectability of the ERT resistivity block when identifying karst intervals. Different starting models were implemented to perform forward modelling. The use of a very heterogeneous and anisotropic starting model implemented by using all information available from resistivity and stratigraphic logs and imagery of borehole walls provided unrealistic results. This is due to the too high level of detail of the starting model respect to the ERT system detectability and resolving ability. Then, a simpler, layered starting model comprising argillaceous limestone, low porosity limestone layers and karst intervals was implemented. Inversion of the synthetic dataset obtained by forward modelling was performed. Modelling allowed addressing different issues limiting the karst intervals detectability: an excessively low thickness with respect to the system's ability; an excessively large resistivity gap when compared to the shallow argillaceous limestone that also reduced the investigation depth and led to a severe underestimate of the limestone resistivity; variation of the layers resistivity with depth.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
