Yanze Yang, Xiankun Ge, Jie Chen, Mengjie Xu, Lei Yang
In remote sensing, multi-sensor data fusion enhances environmental monitoring by integrating complementary observations. A critical step in this integration is spatial resampling to a common scale. Although often regarded as a routine preprocessing operation, resampling can become a significant source of radiometric uncertainty, systematically altering scene radiance during scale transformation, especially in heterogeneous aquatic environments. In this study, we evaluate resampling-induced radiometric uncertainty and assess the physical advantages of flux-conserving resampling in multi-scale aquatic remote sensing. Using the radiometrically stable Landsat 8 OLI sensor as a reference platform, this study develops a radiometric stability-based framework to evaluate multi-scale resampling methods. Radiometric consistency in the visible bands was first evaluated using a Rayleigh scattering calibration, allowing a systematic comparison of four resampling methods across multiple spatial scales. Normalized water-leaving radiance was then retrieved using the Satellite Signal in the Solar Spectrum (6S) radiative transfer model and validated against in situ AERONET-OC measurements. Our results indicate that radiometric consistency decreases with increasing scale, while flux-conserving resampling maintains higher stability and preserves the spatiotemporal characteristics of water radiance. These findings highlight the importance of flux-conserving resampling for multi-scale radiometric fidelity and establish the proposed framework as a reference for reliable multi-source data fusion and quantitative inversion in aquatic remote sensing and beyond.