Automatic Resistivity Profiling (ARP) survey was conducted in October 2013 within the study site, with a device equivalent to the MuCEP system (Panissod et al., 1997; Cousin et al., 2009). For each measurement point, three apparent resistivity values (namely ARP1, ARP2 and ARP3) were computed, corresponding to three depths of investigation linked to the inter-electrode spacings (about 0–0.5 m, 0–1 m and 0–1.7 m, respectively). Data were collected continuously along regular lines all over the study site and georeferenced using a DGPS positioning system. The average interval between two resistivity lines was about 6 m. For each location determined by the systematic random sampling (see Section 3.1), we calculated the median of ARP2 values of all measurement points within a 5 m distance. This median value for each location will be called ‘‘estimated SER” hereafter. Note that we used ARP2 because it corresponds to the apparent resistivity values at a depth of investigation equal to the limit depth used for AWC estimates.