Scientists have developed a new earthquake forecasting model called the Long-Term Fault Memory (LTFM) model. By examining how fault ruptures behaved during previous earthquakes, the LTFM model reflects partial strain release on a fault and the specific timing of past earthquakes. According to the lead author of the study, James Neely, this new model is more representative of what is happening on faults and may produce more realistic earthquake forecasts.

The LTFM model considers more realistic fault behaviors, including how faults behave like rechargeable batteries. Faults can discharge all the energy they have in one rupture event or release part of their strain through a cluster of multiple, smaller earthquakes. This new model is a step in the right direction for earthquake forecasting and can be applied to other settings such as the Cascadia Subduction Zone.

Although current earthquake forecasting models are efficient in producing a simple model for earthquake probability, these probabilities do not always match the geologic record of past earthquakes. The LTFM model may produce more realistic earthquake forecasts due to its consideration of more realistic fault behaviors. As more historical and long-term earthquake data get documented, earthquake forecasting models can be improved and become more realistic.