Improving light use efficiency models via the introduction of both the diffuse fraction and radiation scalar.

Surface solar radiation, both its components and intensity, is pivotal in determining vegetation light use efficiency (LUE) and is essential for accurately estimating gross primary production (GPP) in ecosystems. This study introduces two key parameters: the diffuse photosynthetic photon flux density fraction (f) to account for the diffuse fertilization effect (DFE) and the radiation scalar to reflect the impact of radiation intensity on leaf-level LUE. Leveraging these parameters, we developed two novel LUE models: the Big-leaf Diffuse-fraction Radiation-scalar LUE (BDR-LUE) model, adapted from traditional big-leaf LUE models, and the Two-leaf Diffuse-fraction Radiation-scalar LUE (TDR-LUE) model, based on conventional two-leaf LUE frameworks. These models were calibrated and validated using data from 32 FLUXNET sites representing six vegetation types with available diffuse PPFD measurements. The results reveal the following key findings: (1) Both new models, particularly the TDR-LUE model, deliver superior performance compared to conventional big-leaf and two-leaf LUE models; (2) The BDR-LUE model reduces the root mean square error (RMSE) by at least 12.75 % compared to the conventional Eddy Covariance-LUE (EC-LUE) model; (3) The TDR-LUE model achieves an RMSE reduction of at least 13.54 % compared to the established Two-Leaf LUE (TL-LUE) model; (4) Both models effectively capture annual and monthly variations in vegetation GPP; (5) While the BDR-LUE model outperforms the TDR-LUE model for croplands, it shows lower accuracy for other vegetation types. These findings underscore the importance of integrating fdPPFD and the radiation scalar into LUE models. The proposed models demonstrate substantial potential for enhancing GPP estimates in terrestrial ecosystems.