Monitoring the Interface and Thermal Properties of Double‐Layered Soils: Numerical Inversion‐Based Heat‐Pulse Approaches

ABSTRACT The heat pulse (HP) technique is widely employed for characterizing soil physical properties and processes. However, traditional analytical models, which assume homogeneous media, face challenges in interpreting HP signals from soils with heterogeneous thermal properties. To overcome this limitation, we propose two novel numerical inversion approaches (optimization‐based and an iterative local updating ensemble smoother [ILUES] data assimilation method) to estimate the interface location from heater probe ( L ) and resolve distinct thermal properties in double‐layered soils. Validation across 12 experimental scenarios demonstrated that both methods significantly outperform the Deol et al. method, reducing the root mean square error (RMSE) in L estimation by 13.92% and 59.05%, respectively. Additionally, the ILUES method provided more accurate estimates of thermal conductivity and heat capacity than the optimization method, with mean RMSEs below 0.27 W m −1 K −1 and 0.13 MJ m −3 °C −1 for the dry layer and 0.39 W m −1 K −1 and 0.069 MJ m −3 °C −1 for the wet layer. The ILUES method further enabled quantification of parameter uncertainty, revealing narrow confidence intervals for all five estimated parameters, with the lowest uncertainty associated with L . This study advances HP applications in heterogeneous soils by enabling spatially resolved thermal property estimation, with direct implications for evaporation studies and precision agriculture.