New paper on potential errors in seismic hazard calculations revealed from structural geology field data
By Claudia Sgambato, on 15 April 2020
A new paper by Claudia Sgambato (PhD student at IRDR), Dr Joanna Faure Walker (IRDR) and Prof Gerald Roberts (Birkbeck) has been published providing insight into uncertainties in earthquake probability calculations. Specifically, this work shows that calculations in earthquake recurrence that do not consider changes in slip-rate measurements along a fault could have significant uncertainties. This problem is displayed using geological data collected during fieldword for an example fault section from Southern Italy.
Detailed data of the geometry, kinematics and rates of slip have been collected across a normal fault in the Southern Apennines, Italy, and the results show variations in slip vector (direction of fault movement) and throw (vertical displacement) along such a fault. The throw gradually decreases towards the tip of the fault, but variations are observed along the fault. In particular, we see changes in the throw across areas of structural complexity such as in an along-strike bend in the fault plane where the fault dip is greater, i.e. where the fault has a local change in direction and becomes steeper.
It has been previously demonstrated that such variability affects seismic hazard calculations (Faure Walker et al., 2018). This was shown using a fault with high resolution geological field data. In the newly published study we ask: How would the uncertainty in seismic hazard calculations be affected for faults where detailed slip-rate data are not used?
To investigate how incomplete data produce greater uncertainty in seismic hazard calculations, we calculated strain-rate values across the fault (as strain-rate can be used as a proxy for seismic moment released during earthquakes). First this was done using all our measurements, and then degrading the dataset, removing one measurement at a time. We found that excluding measurements in such a way produces a high variability in strain-rate and if such variations are unnoticed, different values of strain-rate would be produced, and hence different values would result in seismic hazard calculations.
Therefore, this study suggests that: (1) field structural data are fundamental to understand variations in slip-rate; (2) using only one measurement of slip-rate along a fault for calculating hazard is not advisable; (3) the potential error associated to the slip-rate variability should be implemented in seismic hazard calculations.