Estimation Of Global Terrestrial Gross Primary Productivity Based On Solar-induced Chlorophyll Fluorescence

Gross Primary Productivity(GPP)is a key parameter for the research of global carbon cycle and global change.The remote sensing-based method is the mainstream approach to estimates GPP of terrestrial ecosystems.Solar-induced chlorophyll fluorescence(SIF)is an optical signal directly related to vegetation photosynthesis,which is emitted by the photosystems after absorbed solar energy.With the development of remote sensing technology and SIF extraction algorithms,multiple satellite platforms have successfully acquired global SIF products,opening up new ways to monitor vegetation growth status and estimate terrestrial ecosystem GPP.However,the relationship between SIF and GPP at the canopy level is also affected by many factors such as canopy structure,plant species and external environmental stress,resulting in the canopy GPP-SIF relationship and the change of this relationship on time scale is still unclear,and the most widely used GPP-SIF empirical linear estimation model lacks universality.Based on the above background,we first analyzed the fluorescence emission mechanism at different spatial scales,and analyzed the theoretical basis of the linking between chlorophyll fluorescence and photosynthesis based on PAM empirical research.On the basis of the GPP-SIF empirical linear estimation model,we introduced some factors affecting the photosynthetic capacity and canopy SIF emission,to construct a theoretical model of GPP estimation based on near-infrared fluorescence.Combining GOME-2 SIF products,FLUXNET2015 dataset and MODIS products,the theoretical model is constructed on different vegetation types and the estimation accuracy verification analysis is carried out.The main conclusions of this paper are as follows:(1)A large number of empirical studies,such as active fluorescence measurements based on PAM,have shown that ?P+?NPQ and ?F are approximately constant under normal vegetation growth.The photochemical reflection vegetation index PRI based on remote sensing has great potential for estimating LUE,but at the canopy scale,it is also affected by factors such as canopy structure and environmental stress.Although the near-infrared chlorophyll fluorescence is rarely affected by chlorophyll reabsorption in leaves and canopy,its scattering characteristics are still affected by factors such as canopy leaf area index and leaf dip angle distribution.Based on the above theoretical analysis,vegetation index that representing canopy structure characteristics and PRI are introduced into the GPP-SIF empirical linear estimation model,and a nonlinear theoretical model between GPP and near-infrared fluorescence is established.(2)Based on the measured GPPEC,GOME-2 SIF and MODIS vegetation index data,our theoretical model was constructed by using different vegetation indices and the accuracy analysis was carried out.The results show that our model is basically similar to the empirical linear estimation model,with the highest estimation accuracy on deciduous broad-leaved forests and poor estimation accuracy in coniferous forests and evergreen broad-leaved forests.However,whether it is on a single site or on multiple sites of the same planting type,the estimation accuracy of our model on all vegetation types is greatly improved compared with the empirical linear estimation model.On the other hand,our model can better reflect the seasonal variation characteristics of different vegetation types GPP represented by each site.(3)Combined with the global monthly SIF product and the monthly vegetation index product with 0.5° resolution,the 2010 global monthly GPP dataset was estimated by our model and compared with MODIS-GPP.The GPP estimation results of our model are consistent with the existing research results in the annual total GPP and spatial distribution,and can accurately reflect the global GPP seasonal variation characteristics.This proves to some extent the validity of our model.