基于地理加权回归模型的土壤有机碳密度影响因子分析

 为深入探讨土壤碳库的空间变异特征、准确评价区域碳库的影响机制,选择内蒙古自治区赤峰市敖汉旗黄花甸子流域为研究对象,基于研究区的实地采样数据,结合遥感与地理信息系统技术,采用地理加权回归模型对土壤有机碳密度及其影响因子进行拟合,探求不同环境因子对土壤有机碳密度影响的空间变异特性。结果表明,研究区土壤有机碳密度在1.91~16.63 kg/m²范围内变化,其平均密度为7.42 kg/m²;各影响因子对土壤有机碳密度的影响程度由高到低依次为海拔 >坡度 >归一化植被指数(NDVI)>与道路或村庄最短距离(DIST);各因子对土壤有机碳的影响随空间位置的变化存在明显的差异,其中海拔、坡度与土壤有机碳密度呈现负相关,而NDVI、DIST 与土壤有机碳密度呈现正相关。

Analysis on influence factors of soil organic carbon density using a geographically weighted regression model

This research was conducted in Huanghuadianzi watershed in Aohan Chifeng, Inner Mongolia. The influence factors of soil organic carbon density were mainly divided into human factors and natural factors; altitude, slop, normalized differential vegetation index (NDVI) and the shortest distance from path or the village (DIST) were selected as the influence factors. Based on field data samples of the study area, both remote sensing and geographic information system were applied. A geographically weighted regression model was used to study the spatial variations of soil organic carbon density and the different environmental factors. The results showed that the soil organic carbon density changed in the study area from 1.91 to 16.63 kg/m², with an average density 7.42 kg/m². The influence degrees of soil organic carbon density in different driving factors ranked as altitude >slop >NDVI >DIST. The influence of each factor on the soil organic carbon changed with spatial difference. Altitude and slope respectively showed a positive and negative correlation with soil organic carbon density. In general soil organic carbon density decreased with the increasing of altitude and slope in most of the study area and the correlation coefficients were -0.436 and -0.223, while positive effect were only in a few areas. On the other hand, the NDVI and DIST showed a positive correlation with soil organic carbon density, with the correlation coefficients of NDVI being from 1.37 to 1.45 and DIST being from 0.15 to 0.47. In order to analyze the spatial variation of each influence factor, a map of the regression coefficient distribution of the environmental factors and soil organic carbon density in the study area was provided, which provided a scientific basis for the efficient utilization of soil and the development of precision agriculture according to the local conditions.