Abstract Determining long‐term soil erosion and deposition rates and understanding landform evolution are important for managing both natural and human‐modified landscapes, including postmining rehabilitation sites. Although various landform evolution models (LEMs) have been developed to simulate erosion processes and landscape change, relatively few studies have directly compared modelled outputs with field‐based erosion estimates. This study evaluates and compares two LEMs, (i) SIBERIA, which is widely used in the Australian mining sector, and (ii) SSSPAM, a newer coupled soilscape–LEM, together with the well‐established soil erosion model (the revised universal soil loss equation). A formerly grazed hillslope in the Upper Hunter region of New South Wales was used as the case study, and a high‐resolution light detection and ranging (LiDAR)‐derived digital elevation model was used as the landscape input. Field‐based erosion rates were quantified using sediment yield data from a catchment dam and 137 Cs isotopic analysis. Sediment trap measurements indicated erosion rates ranging from 0.43 to 0.61 t/ha/year, while 137 Cs results showed erosion and deposition rates of up to 1.5 and 1.1 t/ha/year, respectively. SIBERIA predicted erosion rates of 1.07 t/ha/year under dense vegetation cover and 3.74 t/ha/year under moderate cover, while SSSPAM estimated 0.35 and 2.43 t/ha/year for the same conditions. The RUSLE model predicted an average erosion rate of 2.23 t/ha/year, with values ranging from 0.58 to 4.65 t/ha/year. Overall, the modelled erosion estimates were broadly consistent with the field observations, demonstrating the ability of both SIBERIA and SSSPAM to reproduce realistic hillslope erosion rates. These findings support the use of LEMs as valuable tools for guiding sustainable land management and rehabilitation practices in both natural and constructed landscapes.
Senanayake et al. (Mon,) studied this question.