Time-lapse monitoring of localized changes within heterogeneous media with scattered waves

Abstract

Time-lapse monitoring of geological and mechanical media has been the focus of various studies over the past four decades because of the information that the inferred changes within the medium provides insight into the dynamic characteristics of the medium. Time-lapse changes within a medium can be used to characterize the temporal evolution of the medium, evaluate the forces driving the changes within the medium and make predictions on the future state of the monitored medium. The detectability of the changes within a material depends on the characteristics of the change to be imaged, the sensitivity of the monitoring data to the change, and the time-lapse monitoring parameters such as the monitoring source-receiver array and the spectral content of the monitoring waves. Various time-lapse monitoring tools have been used to monitor changes within media ranging from the earth's surface to tumors within the human body. These monitoring tools include the use of 4D active surveys were an imprint of the change within the medium is extracted from the time-lapse surveys and the use of interferometric techniques that use singly or multiply scattered waves. My major goal in this study is to image and localize changes present within a scattering medium using time-lapse multiply scattered waves generated within the monitored medium. The changes to be imaged are generally localized in space. This work is an extension of coda wave interferometry. Coda wave interferometry focuses on the identification and extraction of average velocity change occurring within a scattering medium. Due to the non-linear char- acteristics of multiply scattered waves and limited information of the origin of the multiply scattered waves, coda wave interferometry resolves the average velocity change within the scattering medium with no or limited indication of the location of the change. In this study, I demonstrate that time-lapse changes can be imaged and localized within scattering media using travel-time changes or decorrelation estimated from the time-lapse coda waves. The imaging algorithm is defined to invert for the location and magnitude of changes within both statistically homogeneous and statistically heterogeneous scattering media. The imaging of the localized change requires an appropriate computation of the sensitivity of the scattered waves to the monitored change. I develop a novel approach to compute the sensitivity kernel needed to image localized changes present within a scattering medium. I compute the sensitivity kernel, using an a-priori scattering model that has similar statistical properties as the actual medium, by computing the intensity of the scattered waves generated at both the source and the receiver locations. This approach of the kernel computation allows one to compute the sensitivity kernel for any heterogeneous scattering medium with a prescribed boundary condition. Generating the kernel with the a priori model prevents one from invoking a homogeneity assumption of the scattering model. I apply the imaging algorithm on both numerical and laboratory experiments. The numerical experiment provides an opportunity to evaluate the resolution of the monitored change for each coda lapse time. In the laboratory experiment, I accurately resolve the change induced within two concrete blocks due to localize stress and heat changes. I also monitored velocity changes present within the subsurface beneath the Eastern section of the Basin and Range Province, Western US. Time-lapse monitoring with coda waves, generated with repeating active sources, over a period of four months suggests of the presence of a maximum velocity change of approximately 0.2% with the Eastern section of the Basin and Range Province. This observed velocity change are likely induced by the deformation within the Basin and Range Province.