Are spatial distributions of major elements in soil influenced by human landscapes?

Huan Yu?Zhengwei He?Zeming Shi?Bo Kong

Abstract The present study attempted to evaluate the influence of human activity on major elements(Na2O,MgO,Al2O3,SiO2,K2O,CaO,Fe2O3),and to find a method to explore correlations between major elements and human disturbances,according to geospatial theories and methods.The study results indicate that landscapes influence major elements in diverse ways:Al2O3is closely related to road and mine landscapes;strong relationships exist between MgO,Fe2O3,CaO,and SiO2and roads;Na2O,SiO2,and Fe2O3are unrelated to city landscapes;and Na2O is unrelated to road and mine landscapes.

Keywords Major elements·Spatial distribution·Geographical background·Human landscape·Geographic information system·Remote sensing

1 Introduction

The major elements in virgin soil are generally dependent on the lithology of the parent material and the pedological and geochemical processes of soil formation(Mitchell 1960;Hardy and Cornu 2006).Major element concentrations in soils are influenced by natural factors,such as features of the soil parent material,the processes of weathering and biocycling,and wet and dry atmospheric deposition(Cortizas et al.2003).Various solutions and chemical indices have been established and applied to the quantitative evaluation of chemical weathering intensity,most of which are based on major element analyses(Qiu et al.2014).Many studies are based on the assumption that major elements in soil are mainly controlled by natural processes(Taylor and McLennan 1985;Huang and Gong 2001;Zhang 2011;Palma et al.2013).

However,farming,traveling,mining,industrial production,and human settlements have a critical influence on the geochemical,physical,and biochemical properties of soil,especially on soil elements(Kelepertsis et al.2001;Caravaca et al.2002;Takamatsu et al.2010;Alexakis and Gamvroula 2014;Ye et al.2014).In the past 20 years,major element distribution in soil subjected to various human disturbances has garnered considerable attention.Li and Thornton(2001)investigated the influence of mining and smelting activities on some major elements(Mn,Fe,Al,Ca,and P)in soils.Popovic et al.(2001)studied the leaching behavior of major elements through coal ash transportation in a power plant.Lucho-Constantino et al.(2005)estimated the distribution and accumulation of major elements in agricultural soils that had been irrigated with raw waste waters for about 20 years.A comprehensive chemical characterization of 27 fertilizers of different types used in Spain was conducted by Otero et al.(2005)to identify and characterize sources of contamination based on major,minor,and trace element analysis.Reimann et al.(2012)studied the total concentrations of the major elements(Na2O,MgO,Al2O3,SiO2,K2O,CaO,TiO2,MnO,Fe2O3,and P2O5)in grazing land and agricultural soils,and derived some rules around the influence of human activity on those elements.

Most past research supports the legitimacy of using quantitative geochemical methods according to mathematical statistics to evaluate the influence of human activity on soil elements(Cuadrado and Perillo 1997;Villaescusa Celaya et al.2000;Cevik et al.2009).Moreover,many studies have demonstrated that element distribution characteristics in soil can be effectively analyzed by geospatial methods(Eze et al.2010;Bai et al.2011;Lin et al.2011;Bastami et al.2012;Nanos and Martín 2012).The objectives of this paper are to(a)evaluate the influence of human activity on spatial distribution characteristics of major elements,and(b)develop a method for exploring the spatial relationships between major elements and influencing factors based on geospatial theories and methods.

2 Materials and methods

2.1 Location of study area

The study region is28°55′–30°27′N and 105°20′–106°22′E,in Chongqing Municipality(Fig.1).We chose this region based on convenience of transportation and for its representative economic status,landscape,ecosystem,and presence of conflict between people and land.

2.2 Data

Mine,road,and building landscapes were delineated using remote sensing image interpretation,with buffering regions averaging 2000 m.Land form,stratigraphy,and soil data were obtained through digitizing a thematic map.

In total,2314 soil samples were gathered at the study area in 2010.Parameters were tested using Geochemical Survey Specifications,conducted by the Chinese Geological Survey.

Fig.1 Location of study area in Chongqing,China

2.3 Methods

Na2O,MgO,Al2O3,SiO2,K2O,CaO,and Fe2O3spatial distribution were obtained by interpolating soil point samples through spatial analysis(Simanton and Osborn 1980)using ArcGIS software.

Geochemical anomalies were based on regional geochemical background values and Na2O,MgO,Al2O3,SiO2,K2O,CaO,and Fe2O3spatial distribution data,in terms of the Geochemical Survey Specifications;one example is shown in Fig.2.Through the comprehensive analysis of element spatial distributions and regional geochemical background values,geochemical anomalies were identified according to the Specifications of the Multi-purpose Regional Geochemical Survey,executed by the Chinese Geological Survey.When sampling data had a normal distribution,the ranges of regional background values were identified by arithmetic mean(X)with 2 standard deviations(S):X±2S.When sampling data had a lognormal distribution,the ranges of regional background values were identified by geometric mean(Xg):Xg × Sg±2.Values that went beyond the change range of backgrounds were considered to be geochemical anomalies.The anomalies were applied to explore these element correlations with geographical factors and human landscapes,to explore their influence on major element spatial distributions.

Fig.2 Distribution of Fe2O3in context of mines,roads,building lands,and rivers

The human disturbance factors were then analyzed by a distance decay function and regression methods.Distance decay describes the effect of distance on cultural or spatial interactions,with the effect decreasing as distance increases.The spatial distributions of the elements were derived and the spatial relationships between the elements in soils and human landscapes obtained.To illustrate correlations between element anomalies and landscapes of human disturbances scientifically,the Pearson method was used to calculate product moment correlation coefficients between the element anomaly area ratio and the distance to landscapes of human disturbance(Pearson 1895).Correlation coefficients have a value between+1 and-1,where 1 is total positive linear correlation,-1 is total negative linear correlation,and 0 is no linear correlation.

3 Results and discussion

3.1 Natural background analysis

The ratios of Na2O,MgO,Al2O3,SiO2,K2O,CaO,and Fe2O3anomalies in the different soil types,landform types,and geological times were calculated by spatial analysis(Table 1).

Most of the Al2O3,CaO,K2O,MgO,SiO2,and Fe2O3anomalies were detected in the landforms ofUplifting Folded Low MountainsandEroded or Denuded Hills,which cover more than 97%of this region;more than 74%of the anomalies were detected inYellow SoilandPaddy Soil;and more than 76%were in theLate Triassic–Early Jurassic.Thus,Uplifting Folded Low Mountains,Eroded or Denuded Hills,Yellow Soil,Paddy Soil,andLate Triassic–Early Jurassicwere considered natural backgrounds in further analysis of human disturbance factors.As more than 99%of Na2O anomalies were in theEroded or Denuded Hills,PaddyorPurple Soils,andMiddle Jurassic,these were considered natural background in further analysis of human disturbances for Na2O.

At the same time,theUplifting Folded Low Mountains,Yellow Soil,andLate Triassic–Early Jurassiconly occupy 8.76%,7.69%,and 11.83%of the entire study area,respectively.This indicates that Al2O3,CaO,K2O,MgO,SiO2,and Fe2O3might have been affected by certain natural or human factors.Whether or not the anomalies were caused by human interference requires further analysis.

3.2 Human disturbance analysis

Anomaly distribution data of Al2O3,CaO,K2O,MgO,Na2O,SiO2,and Fe2O3and buffer region spatial data were overlapped to calculate the ratios of anomalies falling in each buffer region;an example is shown in Fig.3.Correlation coefficients between the anomaly area ratio distributions and the distance to human disturbance landscapes were calculated based on the Pearson method(Table 2).

3.2.1 City landscape

The Al2O3,CaO,K2O,and MgO anomaly area distributions of cities continually fluctuated with distance for all landform,soil,and geological formation types,and the regularities were vague,indicating an infirm relationship.However,Al2O3returned a high coefficient and a lowPvalue forEroded or Denuded HillsandPaddy Soil;CaO returned a high coefficient and a lowPvalue for all four of these natural background factors;K2O returend a high coefficient and a lowPvalue forEroded or Denuded Hills,Paddy Soil,andLate Triassic–Early Jurassic;MgO returned a high coefficient and a lowPvalue forEroded or Denuded HillsandLate Triassic–Early Jurassic.Collectively these results preclude confirmation of correlations between the city landscapes and Al2O3,CaO,K2O,and MgO.

Similarly,the distributions of Na2O,SiO2,and Fe2O3anomalies also continually fluctuated with distance from landform,soil,and geological formation types,and with vague regularities.However,correlation coefficients of Na2O were all below 0.33 andPvalues all above 0.52,indicating a weak relationship and suggesting that Na2O was not affected by city landscapes.Low correlation coefficients and highPvalues of SiO2and Fe2O3also reflect a weak relationship and demonstrate that city landscape does not affect SiO2and Fe2O3.Previous work demonstrated that trace elements presented significantly higher concentrations in urban soils than in control soils,with the highest concentrations correlating with land use type;major elements did not show a similar phenomenon(Khalil et al.2013).

3.2.2 Road landscape

Al2O3,MgO,Fe2O3,CaO,and SiO2anomalies constantly decreased with distance from roads across landform,soil,and geological formation types,which is consistent with the rule that effects of disturbance decrease further from roads.Moreover,both Al2O3and Fe2O3showed a high coefficient and a lowPvalue for all the landform,soil,and geological formation types(Table 2);the correlation coefficients of MgO were all above 0.96 and theirPvalues below 0.05 for all the natural backgrounds,indicating a strong relationship and demonstrating that Al2O3,MgO,and Fe2O3were affected by roads.Similarly,the correlation coefficients of CaO and SiO2were all above 0.93 forall the natural backgrounds.The close relationship between Al,Ca,and Mg with roads is supported by previous work(Rybak 2015);these elements originate mainly from windblown road dust(Szczepaniak and Biziuk 2003)or are emitted by traffic(Zechmeister et al.2006).Furthermore,it has been proven that Al and Fe often originate from the wear of metallic vehicle parts and from road dust resuspension in urban areas(Vukovi?et al.2013).In addition,Ca and Fe are considered the most mobile elements and affected by a variety of natural and human factors(Gregorauskiene and Kadunas 2006).

Table 1 The proportions of the anomaly area falling in different geographical backgrounds

Fig.3 Distributions of Fe2O3 anomalies at different distances to roads under main geographical backgrounds

The distribution of K2O anomalies continually fluctuated with distance inYellow Soil,indicating an infirm relationship.However,the correlation coefficients of K2O were all above 0.89 and theirPvalues below 0.05 for all the landform,soil,and geological formation types.Thus,correlations between the road landscape and K2O were unable to be clearly confirmed.

Na2O anomalies also continually fluctuated with increasing distance for all the landform,soil,and geological formation types and the regularities were vague,indicating another infirm relationship.Low correlationcoefficients or highPvalues of Na2O were observed,indicating an infirm relationship and demonstrating that Na2O was not affected by the road landscape.

Table 2 Pearson correlation coefficients between distance to human disturbance landscapes and area ratio distributions of the anomalies

3.2.3 Mine landscape

Al2O3anomaly distributions constantly decreased with distance from mines across landform,soil,and geological formation types,which is consistent with the rule that disturbance decreases further from mines.In addition,correlation coefficients of Al2O3were all above 0.94 and theirPvalues below 0.01 for all the landform,soil,and geological formation types(Table 2),indicating a firm relationship and demonstrating that Al2O3was influenced by mine landscapes.A strong correlation between Al and mine landscapes or mining activity has been established by several previous studies(Li and Thornton 2001;Santos et al.2015;Valente et al.2016).

The distributions of CaO anomalies continually fluctuated with distance inEroded or Denuded HillsandYellow Soil;K2O anomalies continually fluctuated with distance in Paddy Soil;MgO anomalies continually fluctuated with distance inEroded or Denuded Hills,Fe2O3anomalies continually fluctuated with distance inYellow SoilandLate Triassic–Early Jurassic,all indicating infirm relationships.The correlation coefficients of CaO,K2O,and Fe2O3were all above 0.89 and theirPvalues below 0.05 for all the landform,soil,and geological formation types.Thecorrelation coefficients of MgO were all above 0.94 and theirPvalues below 0.01 for all the landform,soil,and geological formation types.Duly,correlations between the mine landscapes and CaO,K2O,Fe2O3,and MgO were unable to be clearly confirmed.

The distribution of Na2O anomalies continually fluctuated with distance across landform,soil,and geological formation types,and the regularities were vague,indicating another infirm relationship.The correlation coefficients of Na2O were all below 0.23 and thePvalues all above 0.62,demonstrating that Na2O was not disturbed by the mine landscape.

The spatial distribution of SiO2anomalies continually fluctuated with distance inUplifting Folded Low Mountains,Eroded or Denuded Hills,Yellow Soil,andPaddy Soil—an infirm relationship.SiO2returned a high coefficient and a lowPvalue inUplifting Folded Low Mountains,Yellow SoilandLate Triassic–Early Jurassicbut a highPvalue inEroded or Denuded HillsandPaddy Soil.Thus,correlations between the mine landscape and SiO2were unable to be clearly confirmed.

A table showing the relationships of the different landscapes and major element distributions in soil is shown in Table 3.

4 Conclusions

Element anomalies are primarily produced by geographical influences or disturbances involving human activities.On the premise of the natural background analysis,the correlations between Na2O,MgO,Al2O3,SiO2,K2O,CaO,and Fe2O3anomalies and human disturbance landscapes,i.e.,cities,roads,and mines,were explored to show that major elements are influenced by different landscapes in diverse ways.Al2O3had a strong correlation with road and mine landscapes;MgO,Fe2O3,CaO,and SiO2had a strong correlation with road landscapes that affected these elements significantly;Na2O,SiO2,and Fe2O3had a weak relationship with city landscapes;Na2O had a weak relationship with road and mine landscapes.

This study proves that a response mechanism exploration of major element migration and human disturbance landscape using geospatial theories and methods is practical.However,correlations between Al2O3,CaO,K2O,and MgO and city landscapes;correlations between K2O and road landscapes;and correlations between CaO,K2O,MgO,SiO2,and Fe2O3and mine landscapes could not be determined through the present methods,and will require further work.

AcknowledgementsThis study was supported by the Youth Science Foundation(Grant Nos.41101174 and 41301094),the Lead Strategic Project of the Chinese Academy of Sciences (GrantNo.XDB03030507),the Hundred Young Talents Program of the Institute of Mountain Hazards and Environment(Grant No.SDSQB-2015-02)and the Open Fund for Key Laboratory of Geoscience Spatial Information Technology of Ministry of Land and Resources(Grant No.KLGSIT2016-01).We felt grateful to Southeast Sichuan Geological Team for offering us the experimental data.