No abstract

<p><span>Imaging seismic velocity of the Earth has been implemented widely for years. The majority of these studies are based on linear or non-linear methods that minimize the difference between seismic observations and predictions of these observations from simplified models of the Earth (tomographic models). Another family of methods, based on the work of Backus & Gilbert (1968), constrains Earth models by maximizing their resolution. A numerically tractable version of such linear local averaging methods, called SOLA, was recently been adapted to seismic tomography by Zaroli (2016). When correctly implemented, SOLA tends to reduce artifacts caused by uneven path coverage. It also provides information about model uncertainties and resolutions. </span></p><p><span>We are the first to have applied the SOLA Backus-Gilbert method to group velocity dispersion tomography of the Northwest Iranian plateau. We used Rayleigh wave dispersion curves obtained from vertical component seismograms of local and regional M ≥ 4.5</span> <span>earthquakes that occurred from 2010 to 2021. We also used cross-correlations of ambient seismic noise from January 2013 to the end of December 2015. We allowed the resolution to vary with location and adapted the target resolution based on the local path density. We included data uncertainties based on the location uncertainties of the earthquakes and on the energy in the dispersion curves at each period. We selected the trade-off parameter between model resolution and model uncertainties using a standard L-curve.</span></p><p><span>We present group velocity maps at periods between 10 and 50 seconds as well as maps of model resolution lengths and uncertainties. We also present maps that mask regions where the anomalies are within the uncertainties to highlight the strongly anomalous regions. Our short-period maps reveal the relatively lower velocities in eastern Anatolia and western parts of NW Iran can be explained by partially melt zones in the crust, in accordance with the study of keshin (2003) who proposed extensive melting in the crust because of the interaction of hot asthenosphere with the Eastern Anatolian Accretionary Complex. Also, higher velocity anomalies along the Sanandaj-Srijan metamorphic zone (SSZ), can be related to the sedimentary and metamorphic Paleozoic-Cretaceous rocks. The low velocities observed along the Zagros fault thrust belt are also well correlated with high and shallow seismicity in this zone (Maggi et al 2000) which implies the presence of an upper crust tectonically very active.</span></p><p><span>Our long-period maps reveal high-velocity anomalies beneath the Alborz and low-velocity zone in SSZ. The low-velocity anomalies are mainly due to a thin lithosphere or the absence of a lithospheric mantle, while high velocities can be related to the presence of a stable continental mantle lid or an oceanic-like lithosphere. </span></p><p><strong><span>Keywords: </span></strong><span>SOLA Backus-Gilbert, Group Velocity, Inverse theory, North-West of Iran, Tomography.</span></p>

No abstract