Study on simulation and spatial differentiation of vegetation GPP in the Loess Plateau from the perspective of terrain and geomorphology
[Objective]The study aims to reveal the spatiotemporal patterns of gross primary productivity(GPP)of vegetation under the influence of terrain undulation,to further delve into the interaction mechanisms between topography and vegetation GPP,and to provide new perspectives for the simulation of vegetation carbon flux and the study of its heterogeneity.[Methods]Machine learning models were employed to construct a vegetation GPP simulation model based on macro-scale terrain factors.Spectral models were utilized to extract the spatial spectra of GPP in six typical geomorphic sample areas,and qualitative and quantitative analysis methods were applied to investigate their spatial heterogeneity.[Results]The XGBoost model in this study showed the XGBoost model in this study showed the improved accuracy in vegetation GPP simulation.The R2 of the model with macroscopic topographic features increased by 11.26％over the conventional feature set and by 0.94％over the micro-topographic feature set.Correspondingly,the RMSE was reduced by 21.27％and 2.27％,respectively.From 2003 to 2023,Loess Plateau vegetation GPP rose by 19.12％,with higher values in the southeast and lower values in the northwest.GPP varied notably across six typical regions,generally peaking after an initial decline with increasing terrain ruggedness.[Conclusion]Topographic factors play a crucial role in simulating vegetation GPP,with macroscopic topographic factors more effectively revealing the impact of terrain undulation on GPP than microscopic ones.
万方数据
维普
