An Insight into The Internal Consistency of MODIS Global Leaf Area Index Products

Abstract

The evaluation and validation of Climate Data Records (CDRs) derived from remote sensing play crucial roles in their generation and applications. However, many existing evaluation schemes rely on simplistic, spatiotemporally invariant metrics to assess the products’ overall quality, which leads to the long-term neglect of intra-product inconsistencies stemming from observation conditions, algorithmic differences, and sensor degradation. Leaf Area Index (LAI) is a crucial variable for land surface and climate modeling, and the intra-product inconsistency will increase the uncertainty in related studies. In order to improve the evaluation scheme of LAI products and ensure their reliability, we propose a new perspective for evaluating global LAI time series. In this study, we utilize the MODIS C6.1 LAI product as an example to infer its internal consistency through cross-comparisons among different sensors and spatiotemporal correlations between two adjacent years of the product. We found that compared to the main algorithm, the backup algorithm of the MODIS LAI product tends to underestimate the retrieval results. This inconsistency is particularly pronounced in tropical regions but relatively minor in most other areas. Additionally, these inconsistencies can lead to unusual fluctuations in the LAI time series, impacting the magnitude and direction of short-term vegetation monitoring. However, the influence on long-term trend analyses are negligible. Therefore, special attention should be given to the intra-product consistency in certain studies. In conclusion, the evaluation perspective proposed in this study is of great significance for improving the LAI evaluation scheme and ensuring the use and improvement of remote sensing products.