| Abstract Submission No. | ABS-4.1-0155 |
| Title of Abstract | PyGIMLi-Based Inversion of Resistivity Data for Groundwater Assessment: A Comparative Analysis with Res2DInv |
| Authors | Ronak Dahiya, Tanvi Arora* |
| Organisation | CSIR-National Geophysical Research Institute |
| Address | V.P.O Garhi Sisana Sonipat, Haryana, India Pincode: 131408 E-mail: mail2ronakdahiya@gmail.com |
| Country | India |
| Presentation | Poster |
| Abstract | ERT is a common geophysical tool used to map variations in subsurface resistivity and to evaluate groundwater potential in complex geological settings. The reliability of interpretations depends on the inversion method. In this study, ERT data were processed and inverted using the Python-based open-source PyGIMLi framework to assess groundwater conditions in a semi-arid region experiencing long-term water scarcity. Four 2D profiles were acquired in the Wenner¿Schlumberger configuration to balance depth penetration and lateral resolution. Data preprocessing included data quality control through variance checks, outlier removal, and median smoothing to stabilize the inversion. In PyGIMLi, the inversion was carried out using a Gauss¿Newton least-squares approach with regularization term controlled by λ, balancing smoothness (L2 norm) and robustness (L1 norm). The inversion was run in multiple iterations, with adjustments to damping and mesh, until we reached a stable solution. We observed that all four profiles show models with RMS and χ² errors below 5%, which suggests the data quality and fit were quite good. On the other hand, Res2DInv employs a grid-based finite-difference mesh and a more constrained inversion routine, optimized for computational efficiency and user accessibility. From our comparison, PyGIMLi gave us more control over inversion settings and error handling. This made the results look more reliable for hydrogeological interpretation. Its open-source design can also be easily combined with other geophysical and hydrological datasets for joint inversion. Res2DInv is fast and easy to use, but it offers limited flexibility for detailed error evaluation and model customization. This study highlights the importance of the chosen inversion method in deriving robust resistivity models for groundwater assessment. A comparative analysis of the inversion routines shows that the PyGIMLi approach achieved better resolution and lower error. |
| Keywords | Electrical Resistivity Tomography (ERT), PyGIMLi, Gauss¿Newton least-squares approach, regularization term, L1 & L2 norm |