沙地整治下榆林土地利用及土壤有机质时空分异特征

毛乌素沙地是典型的生态脆弱区，近年来针对其在榆林境内的沙地整治利用取得显著成效，也对土壤环境产生了深刻影响。为了探究沙地不同整治利用方式对土壤有机质的影响，该研究选取榆林市显性沙地，利用多光谱遥感影像及相关光谱指数，结合沙地土地利用变化特征，通过XGBoost机器学习方法，反演1990-2020年土壤有机质含量；分析不同土地类型下土壤有机质含量变化，通过半变异函数揭示了其空间变异性，厘清人为因素和自然环境的影响程度。结果表明，30 a间榆林5 460 km2沙地中超过半数得到整治和利用，沙地-草地是最主要的地类转变方式，建设用地面积增长最迅速；沙区土壤有机质含量上升，但整体呈现先增加后降低的趋势，有机质均值由0.34%增长至0.79%，近10年降低至0.51%；榆林沙区土壤有机质具有较强的空间自相关性。起初，人为利用对其有积极作用，但随着沙地的利用强度增大，对土壤有机质产生负向作用，进而致使其含量下降，面临土地退化危机。建议加强退化林草的修复改良，放缓建设用地开发力度，研究以期为沙地整治提供理论和实践借鉴意义，保护榆林沙地土壤环境安全。

Spatial-temporal characteristics of land use and soil organic matter in Yulin under sandy land remediation

Abstract: The Mu Us Sandy Land is a typical ecologically fragile area, and its remediation and utilization in Yulin has achieved remarkable results in recent years, which also has a profound impact on the soil environment. As an important indicator ofsoil fertility and productivity, it is important to monitor the dynamic changes of soil organic matter from large scale space and long time series to find out the trend of soil organic matter under different natural conditions and anthropogenic influence, and help decision makers to understand the stability and security of soil ecosystem in time. The purpose of this study is to determine the characteristics of land use changes in the conspicuous sandy land in Yulin City in the past three decades, to investigate the changes in soil organic matter content under different land types transformed from sandy land, and to clarify the magnitude of the effects of different remediation and utilization methods on soil organic matter in sandy land. In this study, the dominant sandy land in Yulin city was selected, and the land use transformation characteristics of the dominant sandy land in Yulin city were analyzed by calculating its land use dynamic attitude; using the information of each waveband of multispectral remote sensing images and related spectral indices, combining the factors related to natural conditions and land use change characteristics of sandy land, by comparing the fitting accuracy of three machine learning methods, namely decision tree classification, random forest classification and XGBoost, finally The XGBoost method was selected to invert the soil organic matter content from 1990 to 2020; analyze the change of soil organic matter content and spatial distribution characteristics under different land types, reveal its spatial variability by semi-variance function, calculate the average content change of soil organic matter under different land use types transformed from sandy land in the study area, and clarify the influence of anthropogenic factors and natural environment on desert soil organic matter. The results show that more than half of the sandy land in Yulin City was remediated and utilized in the first three decades, the sandy land transformation was the fastest in the first decade, sandy-grassland was the most important land transformation method, and the construction land area grew the fastest, with the growth rate exceeding 70% at one time; the multispectral remote sensing using XGBoost machine learning method The inversion can better estimate the soil organic matter content, and the inversion error is within 13% by comparing with the relevant studies on soil organic matter content measurement in Yulin City in the past. The average value of soil organic matter of land use types represented by arable land and water area reached nearly 0.8%, and after 2010, the soil organic matter of all land use types decreased significantly, and the average value of soil organic matter decreased to 0.51% in 2020. Soil organic matter in the Yulin sand region has a strong spatial autoregulation and is mainly influenced by natural environmental factors such as temperature, precipitation and topography. Initially, anthropogenic use had a positive impact on it, but as the intensity of sand use increased, it had a negative impact on soil organic matter, which in turn led to a decline in its content and a crisis of land degradation.This study recommends strengthening the restoration and improvement of degraded forest and grass, slowing down the development efforts, and reducing human activities, in order to provide theoretical and practical implications for sandy land remediation, protect the soil environmental safety of Yulin sandy land, and realize the harmonious coexistence between human and nature.