Joint Transmission and Reflection Traveltime Tomography in Imaging Permafrost and Thermokarst lakes at Northwest Territories of Canada

Abstract
We propose an application of joint tomographic inversion using both transmitted and reflected arrivals to map permafrost velocity structures at Mackenzie Delta, Northwest Territories of Canada. Our tomography algorithm is based on a grid based solver of Eikonal equation, Huygens’ Principle and the adjoint method. The grid based solver assigning a traveltime to each grid avoids the shadow zone problem in classical ray tracing and is well adapted for parallelization. The adjoint method provides the gradient of a given objective functional without explicitly estimating Fréchet derivative matrix. Combined with Huygens’ Principle, the tomographic inversion can simultaneously use first and later arrivals to optimize the final model. We first demonstrate the performance of the joint tomography algorithm on a two-dimensional (2D) synthetic model with velocity increasing with depth and decreasing afterwards mimicking a typical permafrost environment. The field 2D seismic survey, covering over 20 kilometers with up to 4 kilometer offset, was acquired from Mackenzie delta, Northwest Territories of Canada, where the shallow subsurface is characterized by high velocity permafrost and low velocity thermokarst lakes. Our results show the potential of the joint tomography in characterizing multi-scale heterogeneous velocity structures in the presence of permafrost.

Characterization of Arctic gas hydrate using AVO techniques – A gear shifting project

A gear shifting project means that Full Waveform Inversion project including the modeling of wave propagation using Spectral/hp Element method will be put aside to accommodate immediate institutional interest other than long term personal interest. This is not bad at all, since AVO is an intermediate step toward full quantification of elastic parameters of subsurface geology. Observe that a bunch of AVO related terminology exist in literature: AVO analysis, AVO inversioin, AVO waveform inversion, full pre-stack inversion, simultaneous pre-stack inversion, pre-stack waveform inversion (PWI), etc.; not to be confused with Full Waveform inversion (FWI), the ultimate goal of seismic inversion. AVO related inversion is an approximation in which the forward modeling engine is not full wave equation but simplified versions (such as Zoeppritz’s equation, reflectivity method, Born&Rytov approximatino, De Wolf approximation, etc.) to speed up calculation, enhance reliability and reduce non-linearity of  inversion. AVO related techniques will be clarified in a separated post.

Conventional seismic reflection method assumed that seismic signals are band-limited normal incidence reflection coefficient series with appropriate traveltime and amplitude variations due to the propagation effect. Thus seismic gathers were usually subject to stacking and further processing to satisfy the above assumption. As early as 1982, Ostrander demonstrated that gas sand reflection coefficients vary in an anomalous fashion with increasing offset and showed that this anomalous behavior can be used as a hydrocarbon indicator. Since then, amplitude variation with offset analysis, termed AVO, became a popular technique in hydrocarbon industry as a lithology indicator and reservoir characterization tool.

Considering that different gas distributions yield different AVO behaviors due to the velocity dispersion in partially gas-saturated sediments, geophysicists in early 1990s started to apply AVO technique to analyze Bottom Simulation Reflector (BSR) generated by free gas underneath gas hydrate bearing sediment in marine environment and estimated gas hydrate and/or gas saturations. However, AVO techniques have not yet been widely applied to characterize Arctic gas hydrate accumulation due to the lack of observable BSRs and coherent reflections. At Geological Survey of Canada, I investigated four lines of 2D high resolution seismic data which in contrast to conventional on-shore seismic data, possessed strong and coherent reflections associated with gas hydrate occurrence. I intend to evaluate various AVO attributes as possible gas hydrate indicator. Petrophysical properties of gas hydrate bearing sediments can be further extract from the prestack seismic data, if data permitting.