Comparison of Solar-Induced Chlorophyll Fluorescence and Light Use Efficiency Models for Gross Primary Productivity Estimation on Three Mid-latitude Grassland Sites in North America

Solar-induced chlorophyll fluorescence (SIF) is a novel technique for gross primary productivity (GPP) estimation, which has great potential to track seasonal and inter-annual changes in GPP in different ecosystems. Currently, the key problem is the scale mismatch between traditional coarse-resolution SIF measurements and the GPP derived from eddy covariance (EC) flux towers. Using SIF data acquired by the Orbiting Carbon Observatory-2 (OCO-2) of three grassland sites in Kansas, Minnesota, and Arizona in USA, the relationship between the SIF and GPP was analyzed in detail on the basis of the different SIF bands, timescales, observation modes, and viewing zenith angles (VZAs). According to the results, the OCO-2 SIF and GPP had a strong positive correlation. Furthermore, SIF 757 was more highly correlated with GPP than SIF 771 was, and the correlation at the daily scale was higher than the correlation at the instantaneous (overpass) scale. The different VZAs had a strong impact on the correlation between SIF and GPP, especially when the zenith angle was greater than 40°. Compared to the vegetation index, SIF can more efficiently track the absorbed light during vegetation photosynthesis and represent the impact of environmental stress on the grassland GPP. Finally, the tower GPP (GPP_EC) was used to validate the estimated GPP produced by the SIF 757 , vegetation index model (VI), and vegetation photosynthesis model (VPM). The results showed that the estimated GPP produced by SIF 757 had the highest accuracy, with a determination coefficient (R 2 ) of 0.845 and a root-mean-squared error (RMSE) of 1.58 g C m −2 day −1 , which illustrates that the data product of the OCO-2 SIF has great potential for estimating the grassland ecosystem GPP.