Abstract: Accurately grasping the moisture content of shallow soil and its variation process holds significant importance in multiple fields, including geotechnical engineering and environmental engineering geology. Utilizing natural temperature information to estimate soil moisture represents a novel approach suitable for long-distance and large-scale monitoring. To address the issues of low accuracy in soil moisture estimation resulting from the limited resolution of soil temperature measurement and reconstruction in conventional methods, this study proposes a new approach based on distributed fiber-optic temperature sensing (FO-DTS) technology. By integrating an explicit finite difference algorithm and the Markov chain Monte Carlo (MCMC) method, a novel method for shallow soil temperature reconstruction and moisture inversion is proposed and validated through in-situ pit tests. The results indicate that: (1) The high spatial and temporal resolution temperatures obtained through FO-DTS allow the explicit finite difference algorithm to effectively reconstruct temperature distributions at various soil depths, with a temperature reconstruction residual error of approximately 0.2 ℃. (2) The MCMC inversion optimization algorithm accurately estimates the soil thermal diffusivity, leading to an estimation error of only 7% in soil moisture. (3) The estimated shallow soil moisture effectively reflects water migration changes induced by weather variations. This new method achieves high-precision soil moisture estimation and demonstrates broad applicability.

Key words: fiber optic distributed temperature sensing, explicit finite difference, Markov chain Monte Carlo, moisture