نوع مقاله : مقاله پژوهشی
نویسندگان
گروه فیزیک زمین، مؤسسه ژئوفیزیک، دانشگاه تهران، تهران، ایران.
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Spectral induced polarization (SIP) is a useful tool in geophysical exploration for understanding the capacitive properties of materials beneath the surface. Unlike conventional methods, SIP analysis can be done in both the time domain and frequency domain. In the time domain, it measures the decay of electrical potential after transmitting a direct current pulse, while in the frequency domain, it measures the phase shift of an alternating current. The Cole-Cole model (CCM) is widely used to analyze SIP data, aiding in the comprehension of subsurface properties in different geological settings. Initially introduced by Cole and Cole (1941) and subsequently expanded upon by Pelton et al (1978), this model provides a description of the complex resistivity of materials. While initially developed for mineralized rock, CCM has been successfully adapted to characterize sedimentary formations lacking electronically conducting components. In such cases, the polarization arises from interactions between pore fluids and electrically charged mineral surfaces, forming an electric double layer. It is well established that frequency-dependent induced polarization measurements offer additional spectral information beyond a single measure of induced polarization amplitude, even though the universal mechanism is not fully understood. This spectral information, derived from the shape of the frequency response, can be linked to petrophysical and geochemical properties of the Earth’s subsurface, such as soil texture, water saturation, hydraulic conductivity, pH, and the dissemination of metallic minerals, through empirical relationships. In addition to advances in the fundamental understanding of induced polarization phenomena, the SIP method has seen significant progress and development across various research areas in recent years, including forward modeling, inversion, and equipment. However, the success of the SIP method is strongly dependent on providing a reliable and precise inversion algorithm aimed at retrieving the CCM parameters. Inverse problem theory refers to a mathematical framework that addresses the extraction of information about a parameterized physical system using observational data, theoretical relationships between model parameters and data (i.e., forward problem), and prior knowledge. To ensure accurate interpretation of the estimated models, it is crucial to understand the correlation between the parameters in the subsurface models. This research is significant because it explores the dependency and correlation of the CCM parameters using a Bayesian approach in a 2.5D inversion framework specifically designed for SIP data. The motivation for studying correlation analysis between model parameters arises from the challenges that high parameter correlation can pose to Markov chain Monte Carlo (McMC) sampling algorithms in probabilistic models. In other words, the objective is to enhance the understanding of subsurface properties and provide a more reliable interpretation of the estimated models by thoroughly analyzing parameter interdependencies. A novel 2.5D inversion code specifically developed for SIP data is introduced, leveraging Python-based libraries and advanced statistical methods. Through synthetic modeling and McMC sampling, the robustness of this approach across various subsurface scenarios is evaluated, including a homogeneous earth model, a two-layer medium, and a model featuring two anomalies within a homogeneous background. Our method enables the extraction of CCM parameters that reflect electrical properties, offering deeper insights into complex geological formations. Visualizations of McMC chains and corner plots effectively reveal the interdependencies among CCM parameters, illustrating the convergence and reliability of the parameter estimates. Validation against synthetic models highlights the precision and effectiveness of the proposed methodology. Overall, this study demonstrates the potential of Bayesian inversion to improve the interpretation of geophysical data and offers valuable insights into the correlation between CCM parameters across different geological environments.
کلیدواژهها [English]
)