A modified two-leaf light use efficiency model for improving the simulation of GPP using a radiation scalar

Two-leaf light use efficiency (TL-LUE) models are efficient methods to simulate regional and global gross primary productivity (GPP). A TL-LUE model has previously been shown to outperform the big-leaf MOD17 model through separate consideration of the contributions of sunlit and shaded leaves. However, the impacts of radiation intensity on LUE are inadequately considered in the TL-LUE model, and the maximum LUEs of sunlit and shaded leaves are assigned as different constants, which often induce large uncertainties. Therefore, a TL-LUE model modified with a radiation scalar (RTL-LUE) is developed in this paper. The same maximum LUE is used for both sunlit and shaded leaves, and the difference in LUE between sunlit and shaded leaf groups is determined by the same radiation scalar. The RTL-LUE model was calibrated and validated at global 169 FLUXNET eddy covariance (EC) sites. Results indicate that although GPP simulations from the TL-LUE model match well with the EC GPP, the RTL-LUE model can further improve the simulation, for half-hour, 8-day, and yearly time scales. The TL-LUE model tends to overestimate GPP under conditions of high incoming photosynthetically active radiation (PAR), because the radiation-independent LUE values for both sunlit and shaded leaves are only suitable for low-medium (e.g., average) incoming PAR conditions. The errors in the RTL-LUE model show lower sensitivity to PAR, and its GPP simulations can better track the diurnal and seasonal variations of EC GPP by alleviating the overestimation at noon and growing seasons associated with the TL-LUE model. This study demonstrates the necessity of considering a radiation scalar in GPP simulation in LUE models even if the first-order effect of radiation is already considered through differentiating sunlit and shaded leaves. The simple RTL-LUE developed in this study would be a useful alternative to complex process-based models for global carbon cycle research.