AHI LAI (2024-2025) & AHI LAI QC (2015-2025)

We selected high-quality MODIS LAI data as the benchmark dataset. Biome-specific Random Forest models were trained to retrieve LAI from Himawari-8/9 observations. Subsequently, a modified singular spectrum analysis (mSSA) algorithm incorporating spatiotemporal weighting was applied to fill the data gaps. Finally, a spatial downscaling method was implemented using MODIS spatial texture features. Through this framework, we produced the AHI LAI product. This product features a 500 m spatial resolution and a 4-day temporal resolution, covering the Asia-Pacific region from 2015 to 2025. Validation against field LAI measurements and cross-comparison with existing products indicate that the AHI LAI product aligns well with polar-orbiting satellite products in terms of spatial distribution patterns and seasonal phenological variations. In cloud-prone regions and complex terrains, the AHI LAI product exhibits smoother temporal profiles and lower data dispersion compared to the MODIS LAI product. Furthermore, evaluating the product performance during drought events indicates that the AHI LAI product captures timely vegetation responses to moisture stress.