Evaluating the spatiotemporal patterns of GPP and tree growth for their response to CO₂ fertilization effects in mid-latitude forests of China

The investigation of the spatiotemporal variation trend of the atmospheric CO₂ fertilization effect (β) has emerged as a prominent topic of interest on a global scale in recent times. Nevertheless, the spatiotemporal patterns of β remain unclear. Herein, we selected the mid-latitude forests of China as the designated study region. Accordingly, remote sensing Gross Primary Productivity (GPP) products were used along with model-based GPP simulation results and tree-ring data in this study. This was combined with the random forest algorithm and a moving window approach to assess the spatiotemporal patterns of vegetation productivity and tree growth responses to atmospheric CO₂ variations between 1982 and 2015. Our findings suggest that from 1982 to 2015, the estimated β derived from the two remote sensing GPP products demonstrated a declining trend. In particular, the EC-LUE GPP exhibited a decrease rate of −0.46%.100 ppm⁻¹yr⁻¹, while the NIRv GPP showed a decrease rate of −0.04%.100ppm⁻¹yr⁻¹. Similarly, the findings from the estimation based on models also indicated a decline in β, with an average decrease rate of −0.08%.100 ppm⁻¹yr⁻¹ across a total of 18 models. Based on the analysis of tree rings from 16 sites, it was observed that the radial growth response of vegetation to atmospheric CO₂ exhibited a decline with an average decrease rate of −0.81%.100 ppm⁻¹yr⁻¹. We speculated that the observed trend in <i>β</i> is primarily driven by LAI and forest age.