Earthquake hazard from 36-Cl exposure dating of elapsed time and Coulomb stress transfer

Overview: We request funds to make measurements of the elapsed time since major earthquakes on active faults in central Italy using 36-Cl cosmogenic dating, and calculate stress transfer from historical/palaeoseismic earthquakes. This will allow (1) knowledge transfer to at-risk communities in the region so they can prepare for future earthquakes if a fault with a long earthquake elapsed time has had stress transferred onto it by a neighboring earthquake(s), and (2) communication of this process to other regions with similar earthquake hazard.

Technical Summary: Active faults experience earthquake rupture due to stress transfer from neighboring earthquakes only if the fault in question is close to its failure stress. We lack knowledge of which faults are close to their failure stress and thus cannot interpret calculations of stress transfer in terms of the probability of impending earthquakes. We propose, for an active normal fault system in central Italy, to measure the elapsed time since the last earthquake normalised to fault slip-rates using in situ 36-Cl cosmogenic isotope dating, because this is a proxy for how close a fault is to its failure stress. We will combine this with calculations of stress transfer from historical and palaeoseismic earthquakes in order to calculate which faults have the highest probability of rupture.

Background: When an earthquake ruptures an active fault, stress is transferred onto neighboring active faults. This transfer of stress may cause a neighboring active fault to rupture in a subsequent earthquake. For example, the 2004 Boxing day earthquake on the subduction plate boundary near Sumatra caused severe loss of life on that day, but also triggered subsequent earthquakes in 2005, 2007, 2009 and 2010, each of which caused major loss of life. Such triggered earthquakes also occur on active faults within plates, such as the three 9th September > Mw 6 earthquakes in 1349 A.D. in central Italy, which occurred on the same day, but on different active faults; this has increased concern for the possibility of a future mainshock to follow the 2009 L'Aquila earthquake (Mw 6.3) whose ongoing aftershocks have transferred onto a neighboring fault (Fig. 1). A key point is that, despite the above examples, earthquakes do not always trigger subsequent earthquakes. Subsequent earthquakes only occur if the neighboring fault(s) are already close to failure due to long-term loading from motions in the crust or between plates. Identification of such faults could inform local populations and civil protection agencies in advance of a future earthquake allowing location-prioritised mitigation efforts. However, unfortunately, we cannot directly measure stress on a fault at 12-15 km depth where intra-plate mainshocks nucleate and so cannot identify such faults. However, we can measure a proxy for stress-through-time, that is elapsed time since the last earthquake, using cosmogenic isotopes (36-Cl). In the sub-surface, 36-Cl concentrations accumulate through time mainly due to hits on calcium atoms by cosmic particles. With 1-2 m slip in each earthquake on active normal faults, and with knowledge of 36-Cl production rates at depth, 36-Cl concentrations measured at 1-2 metres depth quantify elapsed time since the last earthquake. We can dig trenches to expose the fault plane to 1-2 metres depth and measure 36-Cl concentrations on the fault planes. If a neighboring earthquake has loaded/stressed a location with a high 36-Cl concentration, and hence a long elapsed time, we will be able to inform civil protection agencies responsible for planning mitigation; no such data are available at present. We can make such measurements, and have ongoing links with government civil protection project partners who make the seismic hazard maps for central Italy, and who are involved in communicating seismic hazard worldwide.

Critical information arising from earthquake hazard studies needs to be issued through proper government channels in order to (1) minimise panic amongst at-risk populations, and (2) ensure the information is used correctly. We have enlisted key project partners who work for government civil protection agencies in Italy (ISPRA, INGV, OGS, University of Chieti - see Letters of Support), to satisfy (1) and (2). This is the strongest possible guarantee of impact.

Our project partner at the OGS (Peruzza) works with project partners at the University of Chieti (Pace, Visini), to carry out earthquake probability calculations for Italy and construct earthquake hazard maps. They are the PIs on these projects in Italy and are hence the key people to work with. We have worked with them for several years (they are project partners on NERC Standard Grant NE/E01545X/1 - see Track Record) and we are already preparing publications with them. Peruzza, Visini and Pace highlighted the importance of elapsed time to us, so this guarantees the impact of our measurements, since these end-users essentially requested the measurements. They are also interested in implementing Coulomb stress transfer values into seismic hazard calculations as this has not been done for detailed fault geometries in Italy, again ensuring the impact of our study. Peruzza, Visini and Pace report their results to the INGV (our project partner Barba - see Letters of Support).

Project partner Barba at the INGV is the PI who releases seismic hazard maps to at-risk populations in Italy. He performs the quality control on seismic hazard mapping and earthquake probability calculations, and he too informed us that elapsed time is the key unknown in such calculations. Again, this guarantees the impact of our measurements. Barba is also the PI for geological and geophysical emergency response to earthquakes in Italy. In the days after the 2009 L'Aquila earthquake (Mw 6.3), we communicated every few hours with Barba whilst we were in the field searching for the earthquake rupture (see NERC Urgency Grant NE/H003266/1A in Track Record). Our estimates of slip-rates on faults (see refs. 1 and 2 in Track Record) were used by Barba to guide his team of field investigators concerning the rupture location in 2009. Barba has indicated that our elapsed time measurements would be used in the same way following future earthquakes, as will our calculations of Coulomb stress transfer on detailed fault geometries, guaranteeing the impact of our measurements and calculations.

Project partner Vittori at ISPRA coordinates field mapping of active faults in Italy and is the PI for this in the Italian Geological Survey. He has indicated that he is particularly interested in faults with long elapsed times, and he wishes to prioritise field studies on such faults. Vittori also coordinates post-earthquake site effect investigations for Italian earthquakes, and we have published papers with him on this topic (1998 Lauria earthquake Mw 5.8) and have another submitted on the 2009 Mw 6.3 L'Aquila earthquake. He has indicated that knowledge of elapsed time would prioritise post-earthquake field studies, again guaranteeing the impact of our measurements. He is particularly interested in the geographic variation of combined Coulomb stress transfer and long-elapsed time values, guaranteeing both the impact of our stress transfer calculations and 36Cl measurements.

All of the above project partners and the various local PIs and Co-Is on this project are committed to disseminating the measurements/findings of this project to the wider, worldwide seismic hazard community and are best placed to do this as Italy is becoming a benchmark for such work. The impact of this benchmark is being disseminated worldwide through internal and external newsletters and websites from the OGS, ISPRA, INGV, and the UCL Institute for Risk and Disaster Reduction and the Aon Benfield UCL Hazard Centre.