Self-potential chenges associated with volcanic activity: Short-term signals associated with March 9, 1998 eruption on La Fournaise volcano (Reunion Island)

After six years of quietness La Fournaise volcano entered into activity on March 9, 1998. Fissures opened gradually downwards on the northern flank of the cone. Two cones, Kapor and Krafft built, from which lava poured until September 1998. Several other vents opened during this eruption. Mappings, surveys, and continuous recordings of the Self-Potential have been performed on the volcano for twenty years. SP mappings disclose the variability of large scale SP anomalies due to the modification of the hydrothermal system over some ten years. Most of the eruptions take place along a Main Fracture Zone (MFZ) in which ground water flows prevail. SP measurements have also regularly been made on the northern flank of the cone, on a west-east profile crossing the MFZ. Between 1981 and 1992 an enlargement and a shift of the MFZ to the east are evidenced. In particular, the eastern fissural axis trending N35°E could be related to the possible collapse of the east flank of the volcano. After a decrease between 1992 and 1997, the SP anomaly was enhanced again by the 1998 eruption. Short scale, about 250 m wide, 750 mV amplitude anomalies were superimposed on a large scale one, 2500 m wide, and about 250 mV in amplitude. For several years, continuous stations have been measuring the electric field along two directions, with a 20 s sampling, in order to record the genesis of SP signals associated with the volcanic activity. Oscillations belonging to the ULF band were evidenced several days before the 1988 eruption, some of them at 9 km from the summit. Their amplitude reached several tens mV/km. These oscillations sometimes present a phase lag from one station to another; they progressively shift towards the location of the future effusive vents. The polarisation of the oscillations is similar to the polarisation of longer SP variations (1 h period or more) and are correlated with the structural anisotropy. Finally, during the last hours preceding the effusive activity, huge SP signals, up to a few Volts/km, appeared at the stations located on the MFZ, and especially on the branch where the magma migrated. We interpret these SP signals as due to electrokinetic effects generated by fluid flow in cracks opened by the stress field changes