摘要
通过对贵州区域不同类型的自然土壤,按淋溶层(A)、淀积层(B)、母质层(C)进行剖面采样及As含量分析,并采用R语言随机森林回归模型、ArcGIS地统计学半变异函数分析及克里金插值法对数据进行统计分析,探讨贵州区域不同土层土壤As含量空间异质性及影响因素。结果表明,贵州区域土壤As含量存在明显的空间异质性,在水平分布上,土壤As的空间块金系数呈A层(0.61)>B层(0.22)>C层(0.13),A层属中等强度空间自相关性,B、C层空间自相关性强烈,特别是C层;土壤As含量以低、中As区(≤30 mg/kg)为主,高As区(>30 mg/kg)仅分布于黔西南、黔中、黔东南和铜仁市的局部地区。在垂直分布上,土壤As含量的变异系数呈C层(0.89)>B层(0.87)>A层(0.75),表明土壤剖面中As含量的空间异质性随土层深度加深而增强。随机森林回归模型排序结果表明,岩性、地层年代、土壤类型是影响不同土层As含量的3个主要因素(P<0.01),三者之间存在密切的地球化学关系,其中岩性为首要影响因素,碳酸盐岩区域土壤As含量高于碎屑岩区域土壤,随着区域岩性组合中碳酸盐岩占比的增加,土壤As含量出现明显的累积特征。不同土壤类型的剖面层次(A、B、C)中As平均含量的大小顺序为石灰土>黄壤>黄棕壤>棕壤>红壤>紫色土,其中石灰土C层的As含量显著高于A层(P<0.05),其余土壤类型的剖面层次之间As含量未出现显著性的差别(P>0.05)。
By conducting profile sampling of different types of natural soils in Guizhou region according to the eluvial horizon(A),illuvial horizon(B),and parent material horizon(C),and analyzing arsenic(As)content,this study applies the R-language Random Forest regression model,ArcGIS geostatistical semivariogram analysis,and Kriging interpolation method for statistical data analysis to explore the spatial heterogeneity of soil As content in different soil horizons of Guizhou region and its influencing factors.The results showed significant spatial heterogeneity in soil As content in Guizhou.In horizontal distribution,the spatial nugget coefficients of soil As followed the order A horizon(0.61)>B horizon(0.22)>C horizon(0.13).The A horizon exhibited moderate spatial autocorrelation,while the B and C horizons showed strong spatial autocorrelation,particularly the C horizon.Most soil As content fell within low-to-medium As zones(≤30 mg/kg),with high As zones only distributed in local areas of Southwest Guizhou,Central Guizhou,Southeast Guizhou,and Tongren City.In vertical distribution,the coefficients of variation(CV)of soil As content were ordered as C horizon(0.89)>B horizon(0.87)>A horizon(0.75),indicating that the spatial heterogeneity of As content in soil profiles increased with deeper soil layers.The Random Forest regression model ranking showed that lithology,chronological age,and soil type were the three main factors affecting As content in different soil layers(P<0.01),with close geochemical relationships among them.Lithology was the primary influencing factor:Soil As content in carbonate rock areas was higher than that in clastic rock areas,and As content showed obvious accumulation with the increasing proportion of carbonate rocks in regional lithological combinations.The order of average As content across profile horizons(A,B,C)in different soil types was calcareous soil>yellow soil>yellow-brown soil>brown soil>red soil>purple soil.Specifically,As content in the C horizon of calcareous soil was significantly higher than that in the A horizon(P<0.05),while no significant differences in As content were observed among profile horizons of other soil types(P>0.05).
作者
胡琳
刘方
朱健
陈祖拥
刘元生
HU Lin;LIU Fang;ZHU Jian;CHEN Zuyong;LIU Yuansheng(College of Resources and Environmental Engineering,Guizhou University,Guiyang,Guizhou 550025,China;Key Laboratory of Karst Geological Resources and Environment,Ministry of Education,Guiyang,Guizhou 550025,China;Field Scientific Observation and Research Station of Guizhou Karst Environmental Ecosystem,Ministry of Education,Guiyang,Guizhou 550025,China)
出处
《贵州师范大学学报(自然科学版)》
北大核心
2026年第1期24-33,共10页
Journal of Guizhou Normal University:Natural Sciences
基金
国家重点专项部省联动计划课题(2022YFD1901505)。
关键词
AS
空间分布
贵州
岩性
随机森林回归模型
As
spatial distribution
Guizhou
lithology
Random Forest regression model
