Improving modeling of ecosystem gross primary productivity through re-optimizing temperature restrictions on...

The terrestrial ecosystem gross primary productivity (GPP) plays an important role in the global carbon cycle and ecosystem functions. However, the estimates of GPP still have large uncertainties due to insufficient understanding of the photosynthesis-temperature relationship and maximum light use efficiency (LUE_max). We used satellite-derived proxies of GPP to derive optimum, minimum, and maximum temperature for photosynthesis at the ecosystem scale, which was then used to construct a new temperature stress expression. This study improves the MODIS-based light use efficiency model through coupling the optimized LUE_max with the new proposed temperature stress expression. The new model (R² = 0.81, RMSE = 17.8 gC m^-2 (16 d)^-1) performed better than the MODIS GPP products (R² = 0.67, RMSE = 30.4 gC m^-2 (16 d)^-1), especially for evergreen broadleaf forests and croplands. The mean annual GPP over China is 5.7 ± 0.27 PgC, and the GPP significantly increased by 0.046 ± 0.006 PgC year^-1 during 2001-2018. This study provides a potential method for future projections of terrestrial ecosystem functioning.