Enhancing Uncertainty Assessment of LAI Remote Sensing Products Through Multiple Collocation Methods and Error Separation Techniques

Information relating to errors in Leaf Area Index (LAI) remote sensing products is crucial for their applications, yet such information is often difficult to obtain due to the limited availability of field measurements for comprehensive validation. This study is the first to evaluate the applicability of the Multiple Collocation (MC) method for LAI product validation and further analyzes error correlation strategies for products at the same scale and the impact of rescaling procedure. Using the MC approach, we assessed the uncertainties of GLASS, MCD15A2H, VNP15A2H, and CLMS LAI products across different land cover types in the Heihe River Basin. The results show that: (1) The MC method with error separation, reduces underestimation of product uncertainty and significantly improves fairness and objectivity in practical applications. (2) The introduction of scaled RMSE as an evaluation metric enabled a clear differentiation between the performance of static error handling and the quantification of dynamic errors. (3) The evaluation framework effectively captures the dynamic errors of products, though the rescaling process requires further optimization. The improved error characterization and validation framework presented here offers a more precise and broadly applicable intercomparison method for LAI product validation, thereby guiding the production of more reliable LAI products.