Abstract: Terrain effects cause significant deviations to the surface reflectance of remote sensing data, especially in the scenario of large solar zenith angle and significant surface heterogeneity. The traditional terrain correction methods frequently fall short in fully compensating for the scattering effects induced by uneven terrain. To address the inadequacy of terrain correction methods in instances of intense surface scattering, the SS-Minnaert model, a terrain correction approach that accounts for solar zenith angle and surface heterogeneity, is developed in response to a detailed analysis of terrain-induced errors in vegetated regions. The model's accuracy is validated using reflectance simulations from the three-dimensional discrete anisotropic radiative transfer (DART) model. The results indicate that the SS-Minnaert model reduces RMSE by 0.036 to 0.092 in the green, red, and near-infrared bands for solar zenith angles greater than 40°, and by 0.004 to 0.031 for angles less than 40°. Further quantitative analysis demonstrates the SS-Minnaert model's superior performance in terrain correction over mainstream models at solar zenith angles beyond 40°. For solar zenith angles below 40°, the SS-Minnaert model exhibits comparable topographic correction proficiency to established VECA and Teillet models in the near-infrared band. In the red and green bands, the performance is even better. In summary, the SS-Minnaert model markedly enhances the Minnaert model by integrating the dual influences of solar zenith angle and surface heterogeneity, demonstrating a terrain correction effect superior across multiple bands, particularly at high solar zenith angles where the improvement is notably significant. 

Key words: terrain correction, three-dimensional radiation transfer model, forest canopy, minnaert model, scattering effect