Insights of the September 2007 Cerralvo Earthquake–Hurricane Henriette Crisis in La Paz, Mexico: Aftershocks Detection with Artificial Neural Networks

Abstract

Abstract Disastrous earthquakes and hurricanes, in general, are not contemporaneous. Here, we report the 2007 Cerralvo earthquake that had stricken the Gulf of California (GoC) when Hurricane Henriette landed during the aftershock activity. The greatest aftershock struck almost at the same time with Hurricane Henriette inducing panic among the inhabitants of La Paz, Baja California, peninsula. After the earthquake’s main event, the disaster preparedness professionals’ effort focused more on hurricane protocols than on the effects that the earthquake’s aftershocks could have been produced. This decision was probably related to the fact that the southern region of the GoC did not experience an earthquake greater than Mw 6.0 since 1995. The Cerralvo earthquake was unusual for its high number of aftershocks. After a visual inspection of the 24 hr heliplots, we noticed that many aftershocks were missing when we used the short-term average/long-term average method. For this reason, we developed an artificial neural network to recognize earthquakes more efficiently. We could not count a precise number of aftershocks during the crisis, but we detected ∼800 earthquakes using standard techniques from September to December. Eventually, a total of 1475 aftershocks were detected with the artificial neural network, contrasting with the low number of aftershocks usually reported in the middle GoC, which on average is about a dozen. Accurate information about the development of the seismic and hurricane phenomena is the most critical issue for authorities and people in general during this unusual experience. We learned that a robust algorithm to detect a huge number of aftershocks and better communication between researchers and authorities is a successful key in these situations.