Biomass Distribution Characteristics of Picea Schrenkiana Var.Tianschanica By Integrating Ordinary Kriging and Machine Learning

Forest biomass modeling is the current mainstream forest biomass measurement method, and is also one of the hot issues in modern forestry scientific research and production practice. In this study, the biomass of Picea schrenkiana forest in the west of Tianshan Mountains had been taken as the research object. The method of drawing a box map had been used to quantitatively explore the impact of terrain, climate and soil on the biomass. The method of integrating ordinary kriging and machine learning had been used to predict the spatial distribution of Picea schrenkiana biomass. The results had shown that there is a close relationship between the changes of environmental variables, which jointly determines the spatial distribution of the biomass of Picea tianshanica. Picea tianshanica prefers to be located in the middle mountain shady slope area with an altitude of 1500-2700m, and the graybrown forest soil has good water conservation capacity and high nutrients. Moreover, there is a spatial clustering of the biomass of Picea. The high-value clustering of the biomass is concentrated in Tex, Gongliu and Xinyuan areas. The low value clusters of biomasses are concentrated on the south slope of the southern Tianshan Mountains, while the high/low value clusters of biomasses in the northern Tianshan Mountains are staggered and concentrated on a small scale.