Abstract
The uncertainty in the multiscale validation of coarse-pixel satellite albedo product significantly reduces the reliability of validation results but has not yet been quantitatively assessed. To close this knowledge gap, this study first analyzed the main uncertainty factors in the procedure of multiscale validation, and then proposed the quantification method for measuring the uncertainty caused by different factors. Finally, the propagation law of these uncertainty factors into the final uncertainty of validation results was established. Our analysis demonstrates that the uncertainty of multiscale validation mainly comes from the errors in high-resolution albedo maps, the errors generated in the upscaling process, and the errors caused by the geolocation mismatch between reference albedo and satellite albedo. These uncertainty factors are uncorrelated with each other and linearly propagated into the final uncertainty of validation results. The influence of these uncertainty factors partly depends on their magnitudes and partly depends on their weighting coefficients. The systematic errors in high-resolution albedo maps propagate into the errors of pixel scale ground “truth” linearly with slope coefficients equal to or slightly smaller than 1, depending on the std. of noise. And then the pixel scale ground “truth” propagates into the final validation results with the weighting coefficient close to −1. But the non-systematic errors are more likely to be balanced out through the upscaling process in the case of a small subpixel size. The errors generated in the upscaling procedure are generally within the range of −3%-3%. Moreover, their negative weighting coefficients are very small, with a magnitude of less than 0.01. Consequently, the contribution of the errors generated in the upscaling process is negligibly small. By contrast, the errors caused by geolocation mismatch are significant, with the values ranging from −10% to 25%, which exceed the 5% accuracy requirement of satellite albedo products proposed by GCOS. Furthermore, its contribution to the overall uncertainty of validation results is significant, with large weighting coefficients close to 1. The magnitudes of these uncertainty factors and their weighting coefficients show dependence on subpixel sizes. Generally, a fine subpixel size is preferred in multiscale validation because it not only reduces the influence of non-systematic errors of high-resolution albedo maps but also enables more stable and reliable upscaling results. This study deepens the consciousness and understanding of the uncertainty embedded in multiscale validation results, which takes an important step toward more reliable validation over heterogeneous surfaces.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.