Over the past decades, benefiting from the development of computing capacity and the free access to Landsat and Sentinel imagery, several fine-resolution global land cover (GLC) products (with a resolution of 10 m or 30 m) have been developed (GlobeLand30, FROM-GLC30, GLC_FCS30, FROM-GLC10, European Space Agency (ESA) WorldCover and ESRI Land Cover). However, there is still a lack of consistency analysis or comprehensive accuracy assessment using a common validation dataset for these GLC products. In this study, a novel stratified random sampling GLC validation dataset (SRS_Val) containing 79,112 validation samples was developed using a visual interpretation method, significantly increasing the number of samples of heterogeneous regions and rare land-cover types. Then, we quantitatively assessed the accuracy of these six GLC products using the developed SRS_Val dataset at global and regional scales. The results reveal that ESA WorldCover achieved the highest overall accuracy (of 70.54% ± 9%) among the global 10 m land cover products, followed by FROM-GLC10 (68.95% ± 8%) and ESRI Land Cover (58.90% ± 7%) and that GLC_FCS30 had the best overall accuracy (of 72.55% ± 9%) among the global 30 m land cover datasets, followed by GlobeLand30 (69.96% ± 9%) and FROM-GLC30 (66.30% ± 8%). The mapping accuracy of the GLC products decreased significantly with the increased heterogeneity of landscapes, and all GLC products had poor mapping accuracies in countries with heterogeneous landscapes, such as some countries in Central and Southern Africa. Finally, we investigated the consistency of six GLC products from the perspective of area distributions and spatial patterns. It was found that the area consistencies between the five GLC products (except ESRI Land Cover) were greater than 85% and that the six GLC products showed large discrepancies in area consistency for grassland, shrubland, wetlands and bare land. In terms of spatial patterns, the totally inconsistent pixel proportions of the 10 m and 30 m GLC products were 23.58% and 14.12%, respectively, and these inconsistent pixels were mainly distributed in transition zones, complex terrains regions, heterogeneous landscapes, or mixed land-cover types. Therefore, the SRS_Val dataset well supports the quantitative evaluation of fine-resolution GLC products, and the assessment results provide users with quantitative metrics to select GLC products suitable for their needs.

Previous studies on global carbon emissions from forest loss have been marked by great discrepancies due to uncertainties regarding the lost area and the densities of different carbon pools. In this study, we employed a new global 30 m land cover dynamic dataset (GLC_FCS30D) to improve the assessment of forest loss areas; then, we combined multi-sourced carbon stock products to enhance the information on carbon density. Afterwards, we estimated the global carbon emissions from forest loss over the period of 1985–2020 based on the method recommended by the Intergovernmental Panel on Climate Change Guidelines (IPCC). The results indicate that global forest loss continued to accelerate over the past 35 years, totaling about 582.17 Mha and leading to total committed carbon emissions of 35.22 ± 9.38 PgC. Tropical zones dominated global carbon emissions (~2/3) due to their higher carbon density and greater forest loss. Furthermore, global emissions more than doubled in the period of 2015–2020 (1.77 ± 0.44 PgC/yr) compared to those in 1985–2000 (0.69 ± 0.21 PgC/yr). Notably, the forest loss at high altitudes (i.e., above 1000 m) more than tripled in mountainous regions, resulting in more pronounced carbon emissions in these areas. Therefore, the accelerating trend of global carbon emissions from forest loss indicates that great challenges still remain for achieving the COP 26 Declaration to halt forest loss by 2030.

Post-processing optimization refers to the refinement of land cover products by applying specific rules or algorithms to minimize erroneous changes in land cover types caused by classification uncertainty or interannual phenological variations. Global land cover (GLC) mapping has gained significant attention over the past decade, but current GLC time-series products suffer from considerable inconsistencies in mapping results between different epochs, leading to severe erroneous changes. Here, we aimed to design a novel post-processing approach by combining multi-source data to optimize the GLC_FCS30D product, which represents a groundbreaking improvement in GLC dynamic mapping at a resolution of 30 m. First, spatiotemporal filtering with a window size of 3 × 3 × 3 was applied to reduce the “salt-and-pepper” effect. Second, a temporal consistency optimization algorithm based on LandTrendr was used to identify land cover changes across the entire time series and eliminate excessively frequent erroneous changes. Third, certain land cover transitions between easily misclassified types were optimized using logical rules and multi-source data. Specifically, the illogical wetland-related transitions (wetland–water and wetland–forest) were corrected using a simple replacement rule. To address the noticeable erroneous changes in arid and semi-arid regions, the erroneous land cover transitions involving bare areas, sparse vegetation, grassland, and shrubland were corrected by combining NDVI and precipitation data. Finally, the performance of our post-processing optimization approach was evaluated and quantified. The proposed approach successfully reduced the cumulative change area from 7537.00 million hectares (Mha) in the GLC_FCS30D product without optimization to 1981.00 Mha in the GLC_FCS30D product with optimization, eliminating 5556.00 Mha of erroneous changes across 26 epochs. Furthermore, the overall accuracy of the mapping was also improved from 73.04% to 74.24% for the Land Cover Classification System (LCCS) level-1 validation system. Erroneous changes in GLC_FCS30D were considerably mitigated with the post-processing optimization method, providing more reliable insights into GLC changes from 1985 to 2022 at a 30 m resolution.