A MODIS-based ecosystem respiration model and its application in optimizing vegetation photosynthesis and respiration model: A case study of two terrestrial ecosystems in Northern China

Turbulent flux data observed in surface layer during growing seasons at Xilinhaote National Climatic Observatory and Jinzhou Agroecosystem Observatory and remote sensing data were analyzed to acquire main environmental factors and biological factors which drive the ecosystem respiration (R_eco). Then the key driven factors of R_eco were selected to optimize a semi-empirical ecosystem respiration model. Based on the new ecosystem respiration model, respiration part of Vegetation Photosynthesis and Respiration Model (VPRM) was optimized and its simulation effect of net ecosystem exchange (NEE) was validated in a semi-arid grassland ecosystem and a maize cropland ecosystem.

Compared to the linear temperature model, the nocturnal R_eco simulated by the new ecosystem respiration model agreed remarkably better with the observed R_eco (at Xilinhaote site, R² increased from 0.08 to 0.61 in 2010-2012; at Jinzhou site, R² increased from 0.13 to 0.55 in 2010). And the new ecosystem respiration model showed similar performance in predicting nocturnal R_eco (at Xilinhaote site, R² increased from 0.32 to 0.57 in 2013; at Jinzhou site, R² increased from 0.33 to 0.61 in 2011).

This study also indicates that optimization of the respiration part of VPRM can improve the simulation effect of NEE during nighttime of the growing seasons in a semi-arid grassland ecosystem and a maize cropland ecosystem, R² between the modeled NEE and the observed NEE increased from 0.30 to 0.57 in the semi-arid grassland ecosystem and increased from 0.03 to 0.48 in the maize cropland ecosystem. However, in the whole time of the growing seasons, little difference was found between the modelled NEE by the original VPRM model and that by our modified VPRM model, probably for the reason that daytime NEE is mainly dominated by vegetation photosynthesis.