Iezzi, Roberts & Faure Walker (2020) Throw-rate variations within linkage zones during the growth of normal faults: Case studies from the Western Volcanic Zone, Iceland, J. Struct. Geol., 133, 103977

Fault throw-rates and slip-rates are a fundamental input into fault-based seismic hazard assessments (SHA) i.e. how likely are earthquakes to occur….

Francesco Iezzi, Gerald Roberts and Joanna Faure Walker studied active faults in the Western Volcanic Zone, Iceland, to determine whether changes in fault throw-rates across fault bends, as identified in previous works in central Italy, are present in other tectonic settings.

This study shows that fault throw-rate increases within fault bends in response to non-planar fault geometry are present at a range of stages of maturity of the bend and extends examples of this phenomenon to mid-ocean ridge settings. This suggests that extrapolating fault slip-rates and slip during past earthquakes from individual sites along a fault must consider the location of data collection in relation to the geometry of the fault.

Why is this so important? Because if we use individual measurements of how fast a fault is moving, we need to understand whether this measurement is representative of the fault as a whole or whether it is underestimating or overestimating the slip. If we do not do this, we will overestimate or underestimate earthquake hazard.

Dr Zoe Mildon, former IRDR PhD student and now lecturer at University of Plymouth, together with Dr Joanna Faure Walker  (UCL IRDR), Prof Gerald Roberts (Birkbeck) and Prof Shinji Toda (Tohoku University IRIDeS), have published a paper in Nature Communications showing we are a step closer in understanding which faults could rupture in the next earthquake:

Coulomb pre-stress and fault bends are ignored yet vital factors for earthquake triggering and hazard

In this paper, we use long-term stress loading on faults in the central Apennines, Italy, together with stress loading from historical earthquakes in the region to test whether we can identify faults which have a positive stress and hence are ripe for rupture.  We found that 97% large earthquakes within the central Italian Apennines from 1703-2006 occurred on positively stressed faults. Therefore, we can use our modelling to calculate which faults are currently positively stressed and hence help us to determine which faults could rupture in the future. This is not the same as earthquake prediction – saying exactly when and where an earthquake will occur, but it is a step closer to better seismic hazard assessments and understanding why, how and when earthquakes occur.

Dr Joanna Faure Walker standing by a limestone fault scarp in the central Italian Apennines

The paper is available through open access: Mildon et al. (2019)

An article was written about the paper in the Daily Mail

The original press release is available here.

This work is part of the IRDR’s continuing collaboration with Tohoku University, IRIDeS (International Research Institute for Disaster Science). Our collaboration has led to papers including topics such as earthquake stress transfer (Mildon et al., 2016), disaster fatalities (Suppasri et al., 2016), and temporary housing (e.g. Naylor et al., 2018).

Publication of: Occurrence of partial and total coseismic ruptures of segmented normal fault systems: Insights from the Central Apennines, Italy by Iezzi et al. (2019)

Francesco Iezzi (PhD student, Birkbeck) together with Prof Gerald Roberts (Birkbeck), Dr Joanna Faure Walker (IRDR) and Ioannis Papanikolaou (Agricultural University of Athens) have published a detailed study of the long-term displacements across the fault responsible for the 2009 L’Aquila Earthquake, Italy, and the surrounding faults. This reveals that the different faults are behaving together so that the displacement across the system of faults looks similar to if it were one larger fault on ten thousand and million year timescales. This finding can help provide clues regarding the relative local seismic hazard between the different fault segments. The study also provides evidence that the vertical displacement (throw) across a fault increases across fault bends, a result that has been demonstrated in previous papers by the research group (e.g. Faure Walker et al., 2009; Wilkinson et al., 2015, Iezzi et al., 2018). The Iezzi et al. (2019) paper discusses the synchronised and geometrically controlled activity rates on the studied faults in terms of the propensity for floating earthquakes, multi-fault earthquakes, and seismic hazard.

Photograph of damage following the 2009 L’Aquila earthquake, taken by Joanna Faure Walker while accompanying the EEFIT mission.

In 1908 a Mw7.1 earthquake struck the town of Messina in Sicily, Italy.  This earthquake killed over 80,000 people making it the most deadly earthquake in Europe since 1900. Despite causing great losses and prompting research into earthquake environmental effects worldwide, the fault responsible for this earthquake has before now remained a source of contention.

However, new research has identified the fault responsible for this event. This was done using elastic half-space modelling and levelling data from 1907–1909. This research has highlighted the importance of studying mapped faults to locate past events.

This work was led by PhD student Marco Meschis (Birkbeck College) in collaboration with researchers from UCL IRDR, Birkbeck College, University of Plymouth and Università degli Studi dell’Insubria.

Meschis, Roberts, Mildon, Robertson, Michetti and Faure Walker (2019) Slip on a mapped normal fault for the 28^thDecember 1908 Messina earthquake (Mw 7.1) in Italy, Scientific Reports, doi:10.1038/s41598-019-42915-2

Recent IRDR research on Italian earthquakes includes:

Iezzi,  Mildon, Faure Walker, Roberts, Wilkinson, Robertson, (2018) Coseismic Throw Variation Across Along-Strike Bends on Active Normal Faults: Implications for Displacement Versus Length Scaling of Earthquake Ruptures, Journal of Geophysical Research: Solid Earth 

Faure Walker J.P., Visini F., Roberts G., Galasso C., McCaffrey K., and Mildon Z., (2018) Variable Fault Geometry Suggests Detailed Fault-Slip-Rate Profiles and Geometries Are Needed for Fault-Based Probabilistic Seismic Hazard Assessment (PSHA), BSSA 109 (1), 110-123

PhD student Francesco Iezzi (Birkbeck College), supervised by Prof Gerald Roberts (Birkbeck College) and Dr Joanna Faure Walker (UCL IRDR), has published a paper that could revolutionalise how geologists and seismic hazard modellers use long established scaling relationships between fault lengths and surface rupture parameters.

The paper is freely available to all and can be found here.

What new observations have been made?

For five earthquakes studied, the surface fault slip (the amount the fault surface moved during the earthquake) and the throw (the vertical component of the slip) was higher where there was a bend along the length of the fault.

Following the central Italy August and October 2017 earthquakes that ruptured the ground surface, we made detailed high spatial resolution measurements of surface fault displacement along the length of the surface fault ruptures. A study of the amount of vertical and horizontal displacement that occurred along the length of the fault revealed that the throw and slip that occurred during the earthquakes increased where there are bends in the fault. This result is critical and has not been identified before for individual earthquakes.

Damage in Amatrice from the August 2016 Earthquake. Photograph take during EEFIT fieldwork by Dr Joanna Faure Walker.

Why does this occur?

We hypothesis that this occurs in order to maintain the horizontal strain (change in length relative to the original length) across a fault during an earthquake and the long-term horizontal strain-rate that accumulates from multiple earthquakes over thousands of years.

Are there other examples of this?

We then went back and studied other examples earthquakes where there was enough information to determine whether a similar pattern of higher throw and slip could be seen across bends in the fault. In the three further events studied in USA Basin and Range, Greece, and Mexico, we found the same relationship. So it seems this phenomenon occurs worldwide in normal (extensional) faults.

This was the first time that the change in vertical component of slip during an earthquake has been shown to be predictable. However, the observed relationship of increased throw across fault bends has been identified previously in long-term displacements that have accumulated over 15 thousand years as a result of multiple earthquakes in Italy (Faure Walker et al., 2009, Wilkinson et al., 2015). Before now, it was not known whether this increase was caused by there being more earthquakes across the bends or more movement during individual events.  We now know that there can be more slip during individual events, however we do not know whether this is the only mechanism for creating a long-term higher throw-rate across the bends.

What does this mean for earthquake science?

This paper suggests that slip during an earthquake will change where there is a bend along the length of the fault and this change can be quantified and predicted using the proposed theory. This means that close to the fault, earthquakes may be more damaging near a bend in the fault. This finding suggests that we cannot use fault scaling relationships between fault length and expected slip in earthquakes without consideration of fault geometry. This paper can also explain much of the scatter seen in existing plots of maximum surface slip against fault length because when collecting the data as input for such relationships, consideration was not given about whether the measurements were taken across fault bends or not.

These changes in slip along faults in individual earthquakes related to the fault geometry should be included in probabilistic seismic hazard assessments (PSHA).

What other research in the IRDR relates to this?

This work contributes to the IRDR and colleagues’ work on investigating fault behaviour to improve our understanding of earthquake hazard. Recent papers have demonstrated the importance of including detailed fault geometry and slip-rates in seismic hazard calculations (Faure Walker et al., 2018) and Coulomb stress transfer calculations (Mildon et al., 2016, 2017).

Iezzi et al (2018), Coseismic Throw Variation Across Along‐Strike Bends on Active Normal Faults: Implications for Displacement Versus Length Scaling of Earthquake Ruptures, Journal of Geophysical Research, https://doi.org/10.1029/2018JB016732 

New research demonstrates the importance of having detailed measurements at multiple sites along a fault of how fast the fault is moving and how the surface orientation of the fault changes. To access the full paper click here.

Why do we need fault measurements?

Measurements of fault slip rate and the geometry of the fault (it’s 3d orientation) can be used to calculate earthquake recurrence intervals to give probabilities of how likely earthquakes of different magnitudes are to occur. We also need these measurements to model how much ground shaking there will be at given locations. Hazard maps of expected ground shaking can be used to inform building codes and identify where buildings including homes and schools might need retrofitting to improve their resistance to earthquake shaking.

There are other methods available for creating earthquake hazard maps, such as using historical records of earthquake shaking. However, these records unlikely go back far enough in time to capture all faults capable of hosting large earthquakes because some faults will not have hosted earthquake within the time period covered by such records. Therefore, the hazard from some faults would be missing in hazard maps based solely on historical seismicity leading to underestimations in earthquake hazard.

What new insights have been revealed in the research publication?

The paper, entitled “Variable fault geometry suggests detailed fault slip rate profiles and geometries are needed for fault-based probabilistic seismic hazard assessment (PSHA)” demonstrates that relying on only one or a few measurements of how fast the fault is moving along a fault and projecting these measurements along the entire fault may lead to underestimating the uncertainty in the earthquake hazard calculations. Crucially, there may be locations where the hazard is underestimated, meaning people could be at more risk than suggested by simpler models (the converse is also possible). Therefore, earthquake hazard assessments based on fault parameters need to either use detailed measurements including measurements of how fast the fault is moving at multiple sites along the fault or to incorporate how the lack of such data increases the uncertainty in calculated earthquake hazard assessments.

Why are detailed measurements not being already used?

In many regions it is difficult to constrain the fault slip rate (how much the fault has moved in a given time) or throw rate (vertical component of slip rate) along a fault at even one location, let alone several. However, there are regions where this is possible so as more data is collected, this detail should help to improve earthquake hazard assessments both in those regions and worldwide.

Where can I find out more?

Faure Walker J., Visini F., Roberts G., Galasso C., McCaffrey K., and Mildon Z., (2018) Variable fault geometry suggests detailed fault slip rate profiles and geometries are needed for fault-based probabilistic seismic hazard assessment (PSHA), Bulletin of the Seismological Society of America, doi: 10.1785/0120180137

The Fault2SHA Working Group is an ESC (European Seismological Commission) group of researchers in both universities and civil protection authorities collaborating to increase incorporation of fault data in seismic hazard assessments and to improve our understanding of how such data should be used.