Research On Source Landscape Apportionment Of Soil Heavy Metals At County Scale

Soil environment health is the important foundation for maintaining ecosystem function and ensuring the safety of agricultural production.With the advancement of urbanization and rapid development of the industrial economy,soil pollution has increased in prominence,especially from heavy metals,which severely endangers the quality of soil environment.Exploring the pollution sources and spatial distribution characteristics of regional soil heavy metals can facilitate better understanding of the causes of pollution formation,and presents great significance for realizing the control of pollution sources and management of soil heavy metal pollution.Therefore,this study takes one mining and metallurgy city in Hubei Province to carry out research on source landscape apportionment of soil Co,Cu,Mn,Pb,Zn,and Cd.On the basis of soil sampling data and environmental variables,combined with multiple spatial analysis models and machine learning methods,we identified the pollution sources of soil heavy metals at the landscape level,explored the spatial distribution characteristics and influencing factors of heavy metals,and quantified the contribution of pollution sources.This study aims to provide the technical and data support for the prevention and control of regional soil heavy metal pollution.The main results and conclusions are as follows:(1)The receptor model and spatial correlation analysis method were combined to identify the pollution sources of soil heavy metals.Based on the soil sample data and environmental variables related to pollution sources,the source apportionment results with Principal Component Analysis and Absolute Principal Component Score receptor model and Random Forest algorithm showed that soil Co,Cu,Mn,Pb,Zn and Cd mainly originate from two industrial and mining pollution sources with the explained variance of 40.72% and 33.48%,respectively.Then,the global bivariate Moran’s I and geographical detector method were used to identify the pollution source landscapes,by analyzing the spatial correlation between landscapes and soil heavy metal accumulations.Ultimately,this study identified four industrial and mining source landscapes that significantly correlate with the accumulation of soil Co,Cu,Mn,Pb,Zn,and Cd accumulations.(2)The influencing factors of soil heavy metal pollutions were explored on the basis of “source” and “sink” theory,and the spatial distribution of heavy metal contents were estimated by geo-statistical and machine learning methods.The results showed that industrial and mining activities were the dominant factors influencing the spatial distribution of contents of studied heavy metals.The dominant “source” factor affecting the spatial distribution of soil heavy metals is the distance from the source landscapes with an average contribution of 17.74%.In addition,the cation exchange and soil organic carbon content are the main “sink” factors affecting the spatial distribution of soil heavy metals,with average contributions of 7.32% and 5.9%,respectively.In general,the high values of soil heavy metals are highly similar to the distribution of industrial and mining lands.In terms of prediction accuracy,the Random Forest model and Extreme?Gradient?Boosting model have higher prediction accuracies than the Ordinary Kriging method.The highest prediction accuracies for soil heavy metals are Co(r=0.74),Cu(r=0.81),Mn(r=0.66),Pb(r=0.75),Zn(r=0.70),and Cd(r=0.70).For the spatial distribution total content of soil heavy metals,the high contents are mainly distributed as follows: Co and Cu in the central region;Mn in the midwestern and southeastern region;Pb,Zn,and Cd in the south-eastern and northern regions of the study area.For the spatial distributions of the different pollution source contributions,the high contributions of the first source are mainly distributed in the southeastern part,and that of the second source are mainly distributed in the central area.Overall,the distribution patterns verify the results of pollution source types identified by the Random Forest algorithm.Based on the spatial distribution pattern and influencing factors of contributions from other sources,the other sources of Co,Cu and Mn are inferred to include industrial and mining sources,while those of Pb,Zn and Cd may be the natural sources.(3)Based on the identified source landscapes and quantified soil heavy metal accumulations by using receptor model and spatial correlation analysis,combining the environmental factors of the “source”,“pathway”,and “receptor” links of soil heavy metal accumulation process,a source landscape apportionment model for regional soil heavy metals was developed by using neural network model,and was used to quantify the contributions of different source landscapes to different soil heavy metal accumulations.The results showed that soil Co(53.87%)and Cu(56.04%)accumulations in the study area were ascribed to the atmospheric deposition from source landscape 3,soil Mn accumulations(57.19%)were ascribed to the atmospheric deposition contribution from source landscapes 3 and 4,and soil Pb(68.35%),Zn(44.35%),and Cd(79.79%)accumulations were mainly contributed by the atmospheric deposition from source landscapes 3 and 5.The influence ranges of the source landscapes are as follows: source landscape 3 to soil Co accumulations was approximately 6 km;source landscapes 3 and 4 to Cu and Mn accumulations were largest with 4–6 km;and source landscapes 3 and 5 were the largest for Pb,Zn,and Cd with range of 4–6 km.The correlation coefficients between the actual and predicted values for the studied heavy metals were 0.6–0.73,indicating high accuracy.This study combined receptor model and spatial correlation analysis to identify pollution sources from the source type level to landscape level;revealed the spatial distribution and influencing factors of soil heavy metals based on “source” and “sink”theory;quantified the contribution of pollution source landscapes based on the source landscape apportionment model of soil heavy metals developed by combining environmental factors in multiple links of soil heavy metal accumulation process.The results of this study would provide an important basis for the prevention and control of regional soil heavy metal pollution.