A MODIS-based Photosynthetic Capacity Model to estimate gross primary production in Northern China and the Tibetan Plateau

• An solely MODIS-based GPP model (PCM) follows the logic of the LUE model.
• The PCM is a simplified form of the LUE model and is driven by EVI and LSWI.
• Predictive accuracy of the PCM is better than that of the MODIS GPP products.
• Meanwhile, it is comparable with that of the VPM and the potential PAR-based models.
• A linear function of model parameter offers a potential for regional GPP estimation.

Accurate quantification of the spatio-temporal variation of gross primary production (GPP) for terrestrial ecosystems is significant for ecosystem management and the study of the global carbon cycle. In this study, we propose a MODIS-based Photosynthetic Capacity Model (PCM) to estimate GPP in Northern China and the Tibetan Plateau. The PCM follows the logic of the light use efficiency model and is only driven by the Enhanced Vegetation Index (EVI) and the Land Surface Water Index (LSWI). Multi-year eddy CO2 flux data from five vegetation types in North China (temperate mixed forest, temperate steppe) and the Tibetan Plateau (alpine shrubland, alpine marsh and alpine meadow-steppe) were used for model parameterization and validation. In most cases, the seasonal and interannual variation in the simulated GPP agreed well with the observed GPP. Model comparisons showed that the predictive accuracy of the PCM was higher than that of the MODIS GPP products and was comparable with that of the Vegetation Photosynthesis Model (VPM) and the potential PAR-based GPP models. The model parameter (PCmax) of the PCM represents the maximum photosynthetic capacity, which showed a good linear relationship with the mean annual nighttime Land Surface Temperature (LSTan). With this linear function, the PCM-simulated GPP can explain approximately 93% of the variation in the flux-observed GPP across all five vegetation types. These analyses demonstrated the potential of the PCM as an alternative tool for regional GPP estimation.