High-Resolution Global land surface Albedo Component Datasets (HR-GACD) for Bare Soil, Non-photosynthetic Vegetation, and Photosynthetic Vegetation

Land surface albedo (LSA) plays a fundamental role in the terrestrial energy-water-carbon nexus, with its variability driven by contributions from bare soil (SL), non-photosynthetic vegetation (NPV), and photosynthetic vegetation (PV). However, current global land surface models (LSMs) primarily rely on static albedo maps or simplified look-up tables, which inadequately capture the spatiotemporal dynamics of LSA and its components. While satellite-derived LSA products provide real-time LSA information, they do not explicitly provide separate individual surface albedo components, and existing unmixing models struggle to isolate NPV impacts from bare soil backgrounds.

To bridge these gaps, we develops a generalized parameterization scheme and a global albedo component product across three broadbands (visible, NIR and shortwave) and seven MODIS spectral bands. By integrating a kernel-based BRDF model with an analytical approach, we explicitly account for the effects of soil moisture, solar zenith angle (SZA), and terrain on SL, NPV, and PV albedo. Furthermore, a pixel-wise optimization model refines the parameterization, enabling the generation of global albedo component maps at 500 m and 0.05° resolutions.

Validation results demonstrate the robustness of the proposed method, with RMSEs for bare soil albedo ranging between 0.009 and 0.013. The derived global component maps effectively capture both the spatial heterogeneity and the temporal variability of SL, NPV, and PV albedo. After separating the impact of NPV, subsequent analysis highlights the climatological influence of the updated SL albedo, which reduces biases in total net radiation by 8.5 W m-2 (12%) over arid and semi-arid regions, leading to improvements in soil evaporation simulations based on the Two-Source Energy Balance (TSEB) model. Moreover, the non-negligible impact of SZA on vegetation albedo underscores the importance of accounting for illumination geometry effects in high-resolution, hourly ecohydrological modeling. This study advances the understanding of component albedo dynamics, and the publicly available albedo component maps and parameterization framework will facilitate research on water budgets, wildfire behavior, carbon dynamics, and related environmental processes, offering broad implications for improving climate and hydrological simulations at high-resolution scales.

Due to Zenodo's storage limitations, we have uploaded only the BRDF and LSA component maps for representative transitional zones, ranging from arid and semi-arid regions to humid regions, including CONUS, the Sahel, and Australia. The "Download all" button is not working properly due to the large data volume. For batch downloading, full global datasets are available at http://glass.umd.edu/veg_soil_albedo (http://, not https://). The corresponding manuscript is currently under review in Remote Sensing of Environment. Codes for calculating MODIS LSA from BRDF parameters are included, along with the corresponding gridding code for converting MODIS sinusoidal projection to latitude/longitude. For additional details, please contact Dr. Aolin Jia at [e-mail].