Contactless Assessment of Earthquake-damaged Drainage Wells using Sentinel-1 SAR Backscatter Time-Series Analysis

Drainage wells stabilize landslide-prone slopes by lowering the groundwater level. Strong earthquakes can cause these structures undergo shear failure and lose their drainage function. After an earthquake, the functional state of drainage wells needs to be assessed promptly, yet direct field inspection in steep mountainous terrain is hazardous and time-consuming. Although SAR has been widely applied to monitor slope displacement, no established method diagnoses the functional state of drainage wells themselves from satellite data. Inspection based on SAR imagery offers a non-contact, wide-area, and rapid alternative. The above-ground part of a drainage well acts as a reflector that returns backscattered radiation from its location. When shear displaces this reflector from its original position, the backscattered radiation observed at that location decreases. Based on this principle, the time series of the VV-polarized backscatter coefficient (σ⁰) from Sentinel-1 C-band SAR is analyzed to detect shear failure of drainage wells. The proposed approach is applied to drainage wells affected by the 2024 Noto Peninsula Earthquake (Mj 7.6). Backscatter time series spanning the pre- and post-earthquake periods are extracted for one shear-damaged drainage well and seven structurally intact wells across the Noto Peninsula. The pre- and post-earthquake backscatter is compared statistically. The Crawford–Howell modified t-test yields t = -31.5 with 7 degrees of freedom and p = 8.4 × 10⁻⁹, and the Grubbs outlier test yields G = 2.66 with p = 7.6 × 10⁻⁸. Both tests indicate that the backscatter drop at the damaged well is significant relative to the intact population at p < 10⁻⁷. These results demonstrate that Sentinel-1 VV backscatter can detect shear failure of drainage wells from satellite data alone. Such inspection using SAR can identify damaged wells and prioritize them for detailed field inspection.