The 2018-2020 seismo-volcanic crisis, east of Mayotte, Comoros islands: in-depth study of poorly instrumented first months of crisis

Abstract:

On May 10th of 2018, a seismic crisis began 50km east of Mayotte, followed by a magnitude Mw 5.9 event on May 15th and the eruption of an underwater volcano in the following months, as it was discovered in May 2019 (Feuillet et al., in revision). Although Mayotte is the easternmost island of the Comoros volcanic archipelago located in the northern Mozambique Channel between Tanzania and Madagascar, these seismo-volcanic events were unexpected. Before February 2019 and the on- and off-shore instrumentation survey campaigns (within the framework of INSU-TELLUS, MAYOBS, REVOSIMA), the local seismic monitoring network was poor. The detection and localization of many events associated to magma migration phases at the beginning of the crisis were limited (Lemoine et al., 2020). Several stations from several international networks allowed to obtain a disposable monitoring network, before the ReVoSiMa installed better tools.

The four first months correspond to 76% of the cumulated seismic moment of the crisis, and to the formation of some recent structures, including the new 800m high submarine volcano. We ought to better define the actual known phases, which are well illustrated on the new catalog. A migration phase during first weeks (until June 2018, epicenters move south and east) with largest events preceded a phase of strong activity (01 to 09 June 2018) until the beginning of the eruption (early July 2018, Lemoine et al., 2020). It was associated to an unprecedented eruption and then the initiation of another swarm westward from end of August 2018 that is still ongoing after two years (Hoste Colomer, abstract AGU 2020).

Because the frequency and moment magnitude of events were the strongest during May and September 2018, and because this period is associated with unrest phases leading to the eruption, the catalog of events has been completed and reprocessed. In particular, we tested new velocity models, added data from unused regional onshore stations, hence improving the quality and quantity of events and their localization. The catalog of the first 4 months went from around 300 events to almost 1200 seismic events localized and allow to better describe the unrest phases of the first months of this crisis.