Spatial And Temporal Variation Characteristics Of Soil Texture In A Pre-hill Areas In Beijing And Hebei And Analysis Of Its Genesis

Soil texture is a fundamental property characteristic of soils,and soils with similar texture often have similar fertility characteristics.Understanding the distribution and variability of soil texture at different spatial scales is essential for the evaluation of soil resources and the precise management of sustainability.Current research on the spatial distribution of soil texture is mostly focused on mountainous hilly areas,and relatively little research has been conducted on the gentler areas of pre-mountain alluvial fans and the different scales in such areas.In this thesis,we select the pre-mountain alluvial fan area in southwest Beijing and Hebei,and analyse the spatial variability of soil texture and its key environmental factors at medium(pre-mountain alluvial fan area)and large scales(eastern Taihang Mountains area in southwest Beijing and Hebei)to provide a high-precision data basis for soil resource evaluation and management,as well as a research basis for model mapping for large samples and large scale soil census mapping.Three methods,Ordinary Kriging(OK),Regression Kriging(RK)and Random Forest(RF),were selected at the mesoscale,and two methods,Extreme Gradient Boosting(XGB)and Random Forest(RF),were selected at the large scale,and the accuracy of the prediction model was evaluated by 50 validation sets at the mesoscale and 352 validation sets at the large scale,to analyse the spatial variability characteristics and prediction key environmental factors of the models.The main findings are as follows.(1)The random forest(RF)prediction model achieves optimal results at both mesoscale and large scale.The overall trend of soil sand content in the pre-mountain alluvial fan at the mesoscale is low in the southeast and high in the northwest,while the overall trend of powder and clay particles is high in the southeast and low in the northwest.(2)At the mesoscale of the pre-mountain alluvial fan,soil type(SOTY)and elevation(ELVA)contributed most to soil texture for sand and meal grains,while vegetation factors,surface temperature,wind action index(WEI)and flow accumulation(FLAC)contributed relatively less to soil texture.For clay grains,the accuracy of spatial autocorrelation was better than using environmental variables.This indicates that in the pre-mountain alluvial fan area where the topography is gentler,sand and powder grains are more sensitive to topographic variables;while clay grains have high spatial autocorrelation due to the fact that they are mostly in the southeastern region,where the topographic variation is small and the distribution pattern is fixed after being alluvialized by water flow.(3)Light loamy soils are most widely distributed in the southwestern part of Beijing and Hebei,probably due to the influence of the soil texture of the Loess Plateau;the distribution of sandy soils and heavy loamy soils is regional,with sandy soils mainly distributed near the watershed and surrounded by sandy loamy soils,heavy loamy soils mainly distributed in the south-eastern part near the lake,and medium loamy soils located in the transition area.The overall spatial distribution of soils is from west to east and from light to heavy in texture.Soil parent material(PAMA)makes the highest contribution to soil texture,followed by topographic and climatic factors.At the provincial scale,climate has shown the ability to improve the accuracy of soil texture predictions.Soil texture is lighter in wind and sandy alluvium;clay and clayey alluvium;sandy loam and light loam in yellow gneiss parent material;lighter soil texture in the upper part of river alluvium,and progressively heavier soil texture in the lower part of river alluvium to lake area.