Application of particle swarm optimization to 3D Euler deconvolution and 3D modeling of gravity data-a case study from Biga and can towns, NW Turkey

Abstract

The subsurface density distribution in the NW Turkey, between Biga and can towns, is interpreted using 3D Euler deconvolution and 3D modeling methods. Particle swarm optimization is used for the both methods to show its ability to solve a wide range of geophysical gravity problems with minor changes. During 3D modeling, the desired level of smoothness is forced using a smoothing filter, and a penalty is added to the cost function to prevent artifacts. Using these measures, the method is found to be able to recover models from gravity data successfully. 3D Euler solutions revealed possible branches of the known faults. 3D modeling suggests densities up to rho =2.9 g/cm(3) for the basement rocks and <= 2.6 g/cm(3) for the sedimentary cover. According to the recovered model, 3D Euler solutions for structural index value of 0 has a better fit to the sedimentary thicknesses. The study area is placed in the Intra-Pontide suture zone, formed by the collision of the Sakarya continent with the Istanbul Zone, and the recovered model indicates that the can-Biga Fault Zone may correspond to the Intra-Pontide suture in the region. The model also shows that the earthquakes in the region are mostly occurring near the boundaries of the low- and high-density areas. Even though the previous studies have shown this relation using conductivity models, the results indicate that the same relation can also be inferred from gravity data.