Geolympics 8: brackets, logic trees and the earthquake cycle

Every sport-addicted is familiar with group stage, round robins and knockout stages. And each of us likes to forecast the result of a competition. That mean that we are all applying our knowledge (real or presumed), to guess how things will go.

Forecasting is the holy grail for hazard management: if we know where and when a hazardous event is going to occur, we can get prepared. In many instances, scientists can make probabilistic forecasts about the likelihood of a damaging event, but we are far from making deterministic predictions (see here for more details on earthquake forecasts by the UK Geological Society).

One of the most used tools in hazard assessment are logic trees, which are a way to account for uncertainty associated with our limited knowledge or with the intrinsic variability of natural phenomena.

Logic trees are somehow like sport tournaments: each team (or model) plays against the others in a group, and a rank is established. No model is knocked out, but the “strongest” (most probable) gives higher weight to the final result.

Some of the most seismic events show a high degree of complexity, teaching us something of earthquake recurrence. For instance, several faults can rupture in a single event (multi-fault rupture, like the 2016 Kaikoura, New Zealand event) or they can rupture in separate events close in time (like the 2019 Ridgecrest or the 2016 Central Italy sequences).

If we assume that earthquake occurrence is time-independent, as was commonly done up to few years ago, the probability of the next earthquake is constant through time. Instead, if earthquake occurrence is time-dependent, the probability of having a large earthquake is small after a big earthquake and then increases. For instance, Seth Stein, Professor at the Northwestern University, used an analogy with hot streaks and slumps to explain the complexity of earthquake recurrence.