Mapping global leaf inclination angle (LIA) based on field measurement data

Abstract. Leaf inclination angle (LIA), the angle between the leaf surface normal and zenith directions, is a vital trait in radiative transfer, rainfall interception, evapotranspiration, photosynthesis, and hydrological processes. Due to the difficulty of obtaining large-scale field measurement data, LIA is typically assumed to follow the spherical leaf distribution or simply considered to be constant for different plant types. However, the appropriateness of these simplifications and the global LIA distribution are still unknown. This study compiled global LIA measurements and generated the first global 500 m mean LIA (MLA) product by gap-filling the LIA measurement data using a random forest regressor. Different generation strategies were employed for noncrops and crops. The MLA product was evaluated by validating the nadir leaf projection function (G(0)) derived from the MLA product with high-resolution reference data. The global MLA is 41.47°±9.55°, and the value increases with latitude. The MLAs for different vegetation types follow the order of cereal crops (54.65°) > broadleaf crops (52.35°) > deciduous needleleaf forest (50.05°) > shrubland (49.23°) > evergreen needleleaf forest (47.13°) ≈ grassland (47.12°) > deciduous broadleaf forest (41.23°) > evergreen broadleaf forest (34.40°). Cross-validation shows that the predicted MLA presents a medium consistency (r=0.75, RMSE = 7.15°) with the validation samples for noncrops, whereas crops show relatively lower correspondence (r=0.48 and 0.60 for broadleaf crops and cereal crops, respectively) because of the limited LIA measurements and strong seasonality. The global mean G(0) is 0.68±0.11. The global G(0) distribution is out of phase with that of the MLA and agrees moderately with the reference data (r=0.62, RMSE = 0.15). This study shows that the common spherical and constant LIA assumptions may underestimate the interception of most vegetation types. The MLA and G(0) products derived in this study could enhance our knowledge of global LIA and should greatly facilitate remote sensing retrieval and land surface modeling studies. The global MLA and G(0) products can be accessed at https://doi.org/10.5281/zenodo.12739662 (Li and Fang, 2025).