Lawrence Berkeley National Laboratory

Surface waves are advantageous for mapping seismic structures of permafrost, in which irregular velocity gradients are common and thus the effectiveness of refraction methods are limited. Nevertheless, the complex velocity structures that are common in permafrost environments often yield unusual dispersion spectra, in which higher-order and leaky modes are dominant. Such unusual dispersion spectra were prevalent in the multichannel surface-wave data acquired fromour permafrost study site at Barrow, Alaska. Owing to the difficulties in picking and identifying dispersion curves from these dispersion spectra, conventional surface-wave inversion methods become problematic to apply. To overcome these difficulties, we adopted a full-wavefield method to invert for velocitymodels that can best fit the dispersion spectra instead of the dispersion curves. The inferred velocity models were consistent with collocated electric resistivity results and with subsequent confirmation cores, which indicated the reliability of the recovered seismic structures. The results revealed embedded low-velocity zones underlying the ice-rich permafrost at our study site - an unexpected feature considering the low ground temperatures of -10°C to -8°C. The low velocities in these zones (~70%-80% lower than the overlying ice-rich permafrost) were most likely caused by saline pore-waters that prevent the ground from freezing, and the resultant velocity structures are vivid examples of complex subsurface properties in permafrost terrain. We determined that full-wavefield inversion of surface waves, although carrying higher computational costs than conventional methods, can be an effective tool for delineating the seismic structures of permafrost.