Seismological inversion methods are extremely useful to understand the mechanics of earthquakes. However, some earthquake source parameters such as fault strength evolution (τ f ) and frictional dissipated power (Q̇ = σN μss u̇ in MW/m2, where σN is the normal stress, μss is the friction coefficient at steady state and u̇ is the slip rate) cannot be derived by the inversion of seismic waves. Nevertheless, τ f andQ̇ are extremely important to improve knowledge about earthquakes, since they control the moment release rate, the temperature increase in the slip zone and therefore the activation of coseismic fault dynamic weakening mechanisms. The study of frictional melts, preserved in the geological record as pseudotachylytes and exhumed by orogenic activity, allows the observation of entire fault segments from few to hundreds meters that can be studied to retrieve τ f (Sibson, 1975; Di Toro et al., 2006). In a previous study (Nielsen et al., 2010), the authors proposed a theoretical model that linkedQ̇ to the pseudotachylyte-host rock boundary micro-roughness. Frictional dissipated power, indeed, is proportional to shear heating: increasingQ̇ increases the temperature gradient perpendicular to the fault during coseismic slip. Since rocks are usually composed by different minerals with different melting temperatures, a high thermal gradient (highQ̇) will cause the minerals to melt uniformly near the sliding surface (i.e., independently of their melting points), resulting in a relatively smooth pseudotachylyte-wall rock boundary. On the other hand, a gentle temperature gradient (lowQ̇) with widely-spaced isotherms will mainly melt those minerals with low melting points, generating a rough boundary. A first attempt at estimating frictional dissipated power using pseudotachylyte microroughness was performed in the M.Sci. thesis of Castagna (2012), where the author proved that artificial pseudotachylytes produced with SHIVA (a powerful rotary shear apparatus designed to reproduce in the laboratory deformation conditions typical of natural seismic fault slip; Di Toro et al., 2010) become smoother with increasingQ̇. However, this relationship (pseudotachylyte-wall rock smoothing with increasingQ̇) was not well-established in natural pseudotachylytes and it was not possible to estimate frictional dissipated power in natural faults. Moreover, micro-roughness on natural pseudotachylytes was found to be considerably higher than the one found in experimental pseudotachylytes, making a direct comparison between natural and experimental faults very challenging. In this thesis, I try to refine the relationship between micro-roughness and frictional dissipated power in natural pseudotachylytes to get an estimate ofQ̇. In order to do so, I collect five new samples from the Gole Larghe fault zone (GLFZ, Adamello batholith, Southern Alps) and perform a detailed microstructural analysis of 22 carefully selected samples from the large collection of pseudotachylytes from the GLFZ produced in the years by the research group of Prof. Di Toro. One of the main goals was also to identify the physical processes acting on the pseudotachylyte-host rock boundary during seismic rupture propagation and slip. I also revise and improve a MATLAB®script used to quantitatively describe the micro-roughness (expressed as the characteristic asperities height, ω0, and the average asperities radius, λave). The microstructural analysis, together with the refinement of the micro-roughness measurement methodology, lead to a robust relationship between the micro-roughness and frictional dissipated power in natural faults, though a precise estimate ofQ̇ using the experimental data for the calibration is still not possible. In fact, micro-roughness of natural pseudotachylytes is also related to (1) the wall-rock damage induced by the propagation of the seismic rupture, which is not reproduced in the laboratory experiments simulating seismic slip and (2) the initial roughness of the natural faults, which are rougher than the experimental ones (the latter are well-polished for experimental issues). However, for natural near-ideal faults which, as discussed in this thesis, satisfy the assumptions for the theoretical model proposed by Nielsen et al. (2010), I estimated for the GLFZ ancient earthquakes frictional dissipated power values of 156±90 MW m2 . This estimate, which is the first field-based frictional dissipated power estimate for an earthquake ever produced, is slightly larger than theQ̇ predicted (10-100 MW m2) from simplified earthquake models (e.g., Sibson, 1980).

Estimate of earthquake source parameters from an exhumed ancient seismogenic fault (Gole Larghe Fault Zone, Italy)

Lazari, Francesco
2021/2022

Abstract

Seismological inversion methods are extremely useful to understand the mechanics of earthquakes. However, some earthquake source parameters such as fault strength evolution (τ f ) and frictional dissipated power (Q̇ = σN μss u̇ in MW/m2, where σN is the normal stress, μss is the friction coefficient at steady state and u̇ is the slip rate) cannot be derived by the inversion of seismic waves. Nevertheless, τ f andQ̇ are extremely important to improve knowledge about earthquakes, since they control the moment release rate, the temperature increase in the slip zone and therefore the activation of coseismic fault dynamic weakening mechanisms. The study of frictional melts, preserved in the geological record as pseudotachylytes and exhumed by orogenic activity, allows the observation of entire fault segments from few to hundreds meters that can be studied to retrieve τ f (Sibson, 1975; Di Toro et al., 2006). In a previous study (Nielsen et al., 2010), the authors proposed a theoretical model that linkedQ̇ to the pseudotachylyte-host rock boundary micro-roughness. Frictional dissipated power, indeed, is proportional to shear heating: increasingQ̇ increases the temperature gradient perpendicular to the fault during coseismic slip. Since rocks are usually composed by different minerals with different melting temperatures, a high thermal gradient (highQ̇) will cause the minerals to melt uniformly near the sliding surface (i.e., independently of their melting points), resulting in a relatively smooth pseudotachylyte-wall rock boundary. On the other hand, a gentle temperature gradient (lowQ̇) with widely-spaced isotherms will mainly melt those minerals with low melting points, generating a rough boundary. A first attempt at estimating frictional dissipated power using pseudotachylyte microroughness was performed in the M.Sci. thesis of Castagna (2012), where the author proved that artificial pseudotachylytes produced with SHIVA (a powerful rotary shear apparatus designed to reproduce in the laboratory deformation conditions typical of natural seismic fault slip; Di Toro et al., 2010) become smoother with increasingQ̇. However, this relationship (pseudotachylyte-wall rock smoothing with increasingQ̇) was not well-established in natural pseudotachylytes and it was not possible to estimate frictional dissipated power in natural faults. Moreover, micro-roughness on natural pseudotachylytes was found to be considerably higher than the one found in experimental pseudotachylytes, making a direct comparison between natural and experimental faults very challenging. In this thesis, I try to refine the relationship between micro-roughness and frictional dissipated power in natural pseudotachylytes to get an estimate ofQ̇. In order to do so, I collect five new samples from the Gole Larghe fault zone (GLFZ, Adamello batholith, Southern Alps) and perform a detailed microstructural analysis of 22 carefully selected samples from the large collection of pseudotachylytes from the GLFZ produced in the years by the research group of Prof. Di Toro. One of the main goals was also to identify the physical processes acting on the pseudotachylyte-host rock boundary during seismic rupture propagation and slip. I also revise and improve a MATLAB®script used to quantitatively describe the micro-roughness (expressed as the characteristic asperities height, ω0, and the average asperities radius, λave). The microstructural analysis, together with the refinement of the micro-roughness measurement methodology, lead to a robust relationship between the micro-roughness and frictional dissipated power in natural faults, though a precise estimate ofQ̇ using the experimental data for the calibration is still not possible. In fact, micro-roughness of natural pseudotachylytes is also related to (1) the wall-rock damage induced by the propagation of the seismic rupture, which is not reproduced in the laboratory experiments simulating seismic slip and (2) the initial roughness of the natural faults, which are rougher than the experimental ones (the latter are well-polished for experimental issues). However, for natural near-ideal faults which, as discussed in this thesis, satisfy the assumptions for the theoretical model proposed by Nielsen et al. (2010), I estimated for the GLFZ ancient earthquakes frictional dissipated power values of 156±90 MW m2 . This estimate, which is the first field-based frictional dissipated power estimate for an earthquake ever produced, is slightly larger than theQ̇ predicted (10-100 MW m2) from simplified earthquake models (e.g., Sibson, 1980).
2021-07-20
126
File in questo prodotto:
File Dimensione Formato  
Lazari_Francesco.pdf

accesso aperto

Dimensione 12.59 MB
Formato Adobe PDF
12.59 MB Adobe PDF Visualizza/Apri

The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12608/21363