Water and Heavy Metal Transport in Roadside Soils

Pedosphere 15(6): 746-753, 2005ISSN 1002-0160/CN 32-1315/P回2005 SCIENCE PRESS, BEIJINGWater and Heavy Metal Transport in Roadside Soils*1B. KOCHER', G. WESSOLEK2 and H. STOFFREGEN21 Federal Highway Research Institute, Bruederstrafe 53, 51427 Bergisch Gladbach (Germany). E-mail: kocher@bast.de2 Institute of Ecology, Department of Soil Protection, Technical University of Berlin, Salzufer 12, 10587 Berlin (Germany)(Received February 22, 2005; revised October 6, 2005)ABSTRACTRoads with very high trafic loads in regions where soils are low in both pH and sorption capacity might be a sourceof percolation water loaded with heavy metals. Looking at some“worst case" scenarios, this study focused on the input oftraffic related pollutants and on Pb, Cd, Cu, Zn, Ni and Cr concentrations in the soil matrix and soil solution, respectively.The analysis also included pH and electrical conductivity and at some sites DOC. The investigations were carried out onsandy soils with more or less low pH values at four motorway sites in Germany. The average of daily trafi was about50000 up to 90 000 vehicles. Soil pore water was collected in two soil depths and at four distances from the road. The pHin general decreased with increasing distance from the roadside. The elevated pH near the roadside was presumably causedby deposition of dust and weathering residues of the road asphalt, as well as by infltration of salt that was used duringwinter time. At these road sites, increased heavy metal concentrations in the soil matrix as well as in the soil solutionwere found. However, the concentrations seldom exceeded reference values of the German Soil Protection Act. The soilsolution concentrations tended to increase from the road edge to 10 m distance, whereas the concentration in the soilmatrix decreased. Elevated DOC concentrations corresponded with elevated Cu concentrations but did not substantiallychange this tendency. High soil water percolation rates were found near the roads. Thus, even low metal concentrationsof percolation water could yield high metal loads in a narrow area beside the road.Key Wornds: heavy metals, percolation water, roadside soils, trafficINTRODUCTIONStreet runoff shows elevated concentrations of heavy metals and many organic contaminants ascompared to normal precipitation (Harrison and Johnston, 1985; Harrison and Wilson, 1985a, b; Krauthand Klein, 1982; Makepeace et al, 1995; Wigington et al, 1986). These substances in runof, as well asdry deposition and spray water, are transported to the adjacent soil. Many investigations have proventhat the trafic-related soil contamination depends on the road's traffic volume and distance betweenthe sampling points and the road. However, elevated soil metal concentrations are found near the road,decreasing to background values with increasing distance from the road (Harrison and Johnston, 1985;Lagerwerf and Specht, 1970; Wheeler and Rolfe, 1979; Wigington et al, 1986). .Quite a few measurements of groundwater concentrations near roads have been reported (Granato etal, 1995; Mikkelsen et al, 1996; Golwer, 1973, 1995). Most of them deal with loamy soils and did not takeinto consideration the mixing of percolation water and groundwater. Only a few direct measurementsof percolation water concentrations near roadsides have been completed (e.g., Reinirkens, 1996). Thesestudies considered soils with medium pH at roads with fairly low traffic loads. Prior to this study, it wasknown that road runoff had pH values between 6 and 9 (Barbosa and Hvitved Jacobsen, 1999; Legretand Pagotto, 1999; Lygren et al, 1984)，but the impact from this runoff on spatial distribution of pHin roadside soils had not yet been determined. Another point, which has also been poorly investigated,is the effect from road runoff infiltration into a narrov中国煤 化工Sperschneider (992)found downward hydraulic gradients in a road bank duno work on infiltrationdepth or amounts of percolation water near roads hMYHCNMHGookingatsome"worstcase”scenarios, which unfortunately are quite widespread in northern Germany, this study focuses onthe input of traffic related contaminants. Solute transport of heavy metals (Pb, Cd, Cu, Zn, Ni and*1 Project supported by the German Federal Highway Research Institute (No. 15.305/1998/GRB).SOIL WATER AND HEAVY METAL TRANSPORT747Cr) to the groundwater, as well as pH, electrical conductivity and at some sites DOC, were analyzed insandy soils with low pH values at four motorways in Germany.MATERIALS AND METHODSThe sites chosen for this study were situated in the federal states of Lower Saxony and Berlin (bothGermany). Only sites with no changes in the roadside for more than 10 years were selected. Table Igives an overview of the sites and their soil properties.TABLE IRoad and soil properties of the four study sitesSiteLandAverageGround SoilSoilDistance Samp- pH of soilAverage AverageNo.usedailywaterdominant typfrom the ling. pH of soil electricaltypetraficlevelfractiondepth CaCl2 H2O poreconductivityvolumebelowsolution of soil pore(ADT)surfacesolutionvehicles day-1 mm1APine70000>5SandHumic0.5 6.19 7.23 7.39440(Berkhof forestPodzol1.5 5.80 7.15 7.30449BAB 7)0.54.36 4.42 5.358451.5 4.504.328732AFallow 900001.0-1.3 SandPlinthic0.5 6.13 7.05 7.071455(Mellen- pastureGleysol1.06.60 7.04 7.181842dorf6.006.50 7.293721.0 6.00 6.50 6.373373AFallow 70000>4Stagnic0.5 6.05 7.68 7.64974Cambisol1.5.62 7.38 7.09970bostel100.5 4.91 5.61 6.86268BAB 7) .1.0 4.19 5.13 5.842284AMixed 90000> 10 SandSpodo-0.6.71 8.49 7.715200(AVUS，forestDystric1.96.998.79 7.23 ,3390BAB 115)Cambisol 100.54.47 5.16 5.403234.215.06 6.66423At each site, 32 suction cups (sintered Al2O3, medium pore diameter 0.45 pum) were installed at fourdistances from the road edge and in two depths (0.5 and 1.5 m), with four cups for every depth anddistance. Every four weeks within 1999 and 2000 the samples were collected in Duran glass bottles thathad been rinsed in HNO3 and dried. After sampling, a suction of 150 up to 300 hPa was applied to thesuction cups, and the resulting pore water samples were collected at the next sampling date. Electricalconductivity and pH were measured immediately after sampling and one drop of HNO3 suprapur per50 mL was added to preserve the samples. Graphite-tube atomic absorption spectrophotometry (AAS)was used to measure concentrations of Pb, Cd, Cu, Ni and Cr, while flame AAS was utilized for Zn.These were all directly measured from the samples without filtration. Some of the samples were fltered(0.45 μum) and DOC concentrations were measured by using a Shimadzu TOC-5050A.Soil samples were taken at each site from sampling pits at 1, 2.5, 5 and 10 m distances from theroad edge. For every soil profile, the whole area covered by one horizon was taken as one mixed sampleof 2- 2.5 kg soil material. After air drying, a representative sample of the fine earth (sieved > 2 mm)was taken for measuring the pH. Thus, 10 g of air-dried soil material was suspended in 25 mL CaCl2(0.01 mol L-1), another 10 g in deionized water. After four中国煤化工V were measured.In addition to that, TDR probes (n = 4) and tensiomemeasure the soilwater content and matrix potential. They were installed inYHCN MH Gso cm) and at 1.5m distance from the road edge. Tensiometer readings were used to calculate hydraulic gradients, andTDR measurements were used to calculate soil water balance.To estimate the long term average infiltration depth and water balance, climatic data from the748B. KOCHER et al.Northeast Region of Germany were used. The calculation was based on experiences and followingassumptions (Wessolek and Facklam, 1997): precipitation at 600 L m- 2 year' - 1，road surface of asphalt12 m wide, actual evaporation from the road about 100 L m -2 year -1 , potential evaporation at 620 Lm-2 year-1, infiltration into the road surface at 50 L m-2 year-1, grassland vegetation, soil consistingof medium sand, spray water at 20% of road drainage and runoff at 80% of road drainage. More detailsare described in Kocher and Wessolek, 2002.Spatial distribution of metal concentrations in the soil pore water was compared with the distributionof pH. The data and feld observations were used to distinguish three different zones of water infiltrationand heavy metal transport.RESULTS AND DISCUSSIONSoil water content and hydraulic gradients of two depths (0.37 and 0.65 m) for a position 1.5 m farfrom the road edge at Site 4A are shown in Fig. 1. The soil water content during the measuring periodwas relatively high, even during the summer months July and August. Symptoms of water shortage (dryplants, low matric potentials, etc.) were not observed. An explanation is the high input of runoff andspray water from the road at rainy days (Kocher and Wessolek, 2002). Although the precipitation inthis region was only 430 mm for 1999, water in the layer 0.5 to 0.8 m below surface percolated downwardfor 6 of the 7 measuring dates, and the hydraulic gradients in 0.65 m depth were above zero most of the .time. Only in a short period the water flow in this layer was directed upwards (Fig. 1).300「oWater content0.37 m△Gradient 0.37 m」250l OWater content 0.65 m X Gradient 065 m苣2200.4150t08-12100Direction of water flowin0.5 to 0.8 m depth:-16圣50↓↓↓-2-20Jul.1999 Aug.1999 Oct.1 999 Dec.1 999Jan.2000 Mar.2000Fig. 1 Soil water content, hydraulic gradient and water flow direction of two depths (0.37 and 0.65 m) at 1.5 m distancefrom the road edge for one of the four study sites (Site 4A).To get a better understanding of the soil water behavior depending on the distance from the roadedge, long-term mean annual infiltration depth was calculated (Fig.2). In the road bank, which isdirectly located at the road (0 -1 m), road runoff water from precipitation mainly infuences waterbalance. The infiltration depth of percolation water reached depths of up to 12 m within one year. Itdecreased to 1 2 m in the area, only infuenced by spray water, which was measured (not shown) to beabout 5 m wide. At > 5 m distance from the road edge, the annual infiltration of soil water was onlyinfuenced by rainfall and reached depths of about 1 m. More details of water balance prediction andmeasurements are described in Kocher and Wessolek, 2002.These data hinted to high amounts of percolation water in the road bank. These high rates of waterpercolation, resulting from road runoff with pH of 6 to 9, caused elevated soil pH levels in the first fewmeters beside the sealed road edge (Table I, Fig. 3).中国煤化工tes directly at the roadedge (measured in water) reached 7 or 8 and decreased(10 m) with increasingdistance from the road edge. Reasons for the scatteri:YHc N M H G diferent amount andfrequency of runoff infiltration that infuences soil moisture. Both were infuenced by distance, steep-ness of the road bank and ditch, vegetation, groundwater table, and annual as well as seasonal weatherconditions. However, on soils with neutral background pH (pH about 7) no references for elevated pHSOIL WATER AND HEAVY METAL TRANSPORT749Distance from the road edge (m)245A-runoff influenced area己0tB-area influenced by spray water2t14Fig. 2 Estimated mean annual infiltration depth of percolation water in the soil of one of the four study sites (Site 4A).Site 1ASite 2ASite 3ASite 4A宁昌宁守:卓导中↑古6宁宁占皇4t2412.551014.8112.551012.551(Fig. 3 pH of the soil solutions for different distances from the road edge at the four study sites.values near roadsides have been found.The metal concentrations in the soil matrices decrease with increasing distance from the roadside(Kocher and Wessolek, 2002). Comparing the spatial distribution of metal concentrations in the soilpore water with the distribution of pH, different patterns were found for the diferent substances andsites (Figs.4- 6). At Sites 1A and 4A the Cd concentration in the soil solution increased with increasingdistance from the roadside. This tendency could not be observed for Site 2A, and only for 5 m distanceat Site 3A (Fig. 4). However, if the Cd concentration of the soil solution was compared with their pHvalue, as done in Fig. 5 for all samples, it can be figured out that the occurrence of high Cd concentrationhad more relevance for low pH conditions. This trend between pH and the concentration of Cd wasobserved for the samples of Sites 1A, 3A and 4A (Fig. 5). The main reason for this was the range of pHcovered by the samples at the four sites: Sites 1A, 3A, and 4A had a range of 2.8 up to 3.3 pH steps,whereas the pH at Site 2A showed differences of only 1.4 pH steps (0.1 and 0.9 percentiles in Table II).At Site 2A, with a shallow ground water level of 1.0 -1.3 m below surface during the whole year (TableI)，pH changed only slightly with the distance from the road edge because the background pH at 10m distance was higher than at the other sites, around 6.5- -7.0 (Table I, Fig. 3); therefore, no relationbetween pH and soil solute concentrations of Cd occurred (Fig. 5).1000，出主，宁0.1中国煤化工0.010.001FYHCNMHG_2.5 54.8 102.5Fig. 4 Cd in the soil solutions for different distances from the road edge at the four study sites.750B. KOCHER et al.00 rSite 1ASite 2A、Site 3ASite 4AC0.0.010.0013 45678456784.56789Fig. 5 Cd versus pH in the soil solutions for the four study sites.1000Site1A 。Site 2A。Site 3A10030.13.456784.5678456789pHFig. 6 Cu versus pH in the soil solutions for the four study sites.Near the roads the pH- elevating effect of street runoff (Fig. 3) was strong enough to prevent criticalCd concentrations in the soil pore solution (Figs. 3 and 4), according to the Federal German Soil Pro-tection Act (1998, Table II). The same can be said about most of the other heavy metals investigated.Only Cu and Zn showed a relatively larger number of elevated concentrations (Table II).The general relational distributions of pH and Cd concentrations in soil solutions were not foundfor Cu. On the contrary, at three of the four sites a weak tendency of Cu concentrations to increasewith pH values were found (Fig. 6). A possible reason for the weak relation between pH and Cu wasthe high infuence of DOC on mobilization of Cu (Kordel et al, 1997). The DOC data (Fig. 7, TableII) supported this for Sites 2A and 3A, less for Site 1A and not at all for Site 4A. The source of theelevated DOC concentrations at Sites 2A, 4A and partly 3A may be mixed-in organic soil material ordumps of the roadside topsoil, which had to be removed at roads with very high trafic load every six toten years. In former times, this specific material was dumped in the ditch quite often. The soil profilesof Sites 2A, 3A and 4A up to 5 m distance from the road edge showed peat layers and fossil A horizons.They were also spotted with organic soil material up to more than 1 m depth, probably mixed in whileconstructing or repairing the road (Kocher and Wessolek, 2002).Generalizations and zonal classificationSeveral generalizations can be drawn from the data and field observations. Near broad roads, thelong infuence of percolating road runoff changed the formerly low soil pH to near or above neutral. Theamount of infiltration and the soil moisture regime varied strongly with the distance from the roadsideas well as the form of the bank and ditch, causing a great variety of pH and metal concentrations inthe soil solutions. Even when the soil substrate was quite homogeneous, pH did not provide a sufficientexplanation for mobile metal concentrations. The conceications of Cu mobility,but also no satisfactory reason. DOC concentrations中国煤化工w more disturbed sol .profles.TYHCNMHGAt the sites investigated, and at probably other ones with low background soil pH, the pH of thesoil solution reflected a situation that combined a formerly low pH of soil material and an elevated pHin the road runoff, which had been infiltrating for many years. Three zones were distinguishable: ZoneSOIL WATER AND HEAVY METAL TRANSPORT751TABLE IIMetal concentrations in the soil solutions at the four study sitesiteStatisticspFEC .DOCPb)dCuNiCrZμS cm-1 mg L-1μgL-1mgL-1lAMinimum3.8785.240.010.050.020.0020.1 percentiles4.222288.530.07 0.052.350.590.380.0080.5 percentiles6.1242220.80.53 0.47 6.003.991.090.064Mean5.9352125.11.6317.18.162.090.2000.9 percentiles7.4851.62.344.4025.021.15.300.640Maximum8.25346591.480.727.866510314.03.84028425932862302822A3.9314314.10.030.0015.9720815.90.134.361.000.650.0056.88 .35219.20.700.1111.73.681.770.0206.749847.31.60 0.2822.07.350.0477.4013141313.530.9149.419.16.920.1388.1058402021.83.1018065.166.60.3402071742042051671963A3.973.980.04 0.010.10.002 .4.651354.500.101.90.510.307024.80.577.32.551.310.0186.3669624.41.662.1512.36.271.940.1657.51135848.13.297.5226.120.34.250.418 .8.07404066.944.824.911445.014.52.010n19218421941571641934.010430.20.009537348.60.05 0.01 3.40.920.5.5 percentiles7.37221084.00.65 0.2410.73.171.420.0377.04311384.91.7026.02.730.1168.0069814.433.0268.46.720.3648.471909713741.68.44 .52.017.20.690 .190183197195159160186Limit according to the Federal German Soil Protection Act (1998) 25500.500100000 t，Site 1ASite 2A。oSite 3ASite 4A。0ty8°嚣5101001000100 1000100 1000 10 100 1000DOC (mgL")Fig. 7 Cu versus DOC in the soil solutions for the four study sites.A, nearly continually influenced by runoff (0 to about 1.5 m distance from the road edge), where pH insoil solutions were nearly neutral (7-7.5) and metal concentrations comparably low; Zone B, reachingfrom 1.5 to 5 m distance and infuenced by spray water and qprobably duringheavy rain, where pH in the soil solution was intermediate,中国煤化Iitration, changingoften; and Zone C, almost never infuenced by road runoff,MHCN M H Gcovering> 5 m tomore than 10 m distance, where the pH was low, corresponding to low background soil pH and metalconcentrations in the soil solution were relatively high. In Zones B and C remarkable changes of pHin soil solutions occurred, depending on occasional high infiltration rates of road runoff during heavy752B. KOCHER et al.rainfall and complete drying afterwards. However, this applied only to soils with low background pH.CONCLUSIONSSpecial characteristics of soils near roads included contaminant input occurring continuously byaerial deposition and intermittently by road runoff and spray water. High infltration and percolationrates, as well as high soil moisture, could be found in the first meter from the road edge most of thetime and occasionally up to a distance of several meters.The estimation of infiltration depth and hydraulic gradients showed that high percolation rates hadto be expected near the road edge, which, although the metal concentrations in soil solutions near theroadside were low, could lead to remarkable metal loads. One problem in calculating the traffic inducedcontribution was the dificulty of identifying background values in the percolation water because theconcentrations in general increased with increasing distance for the investigated distance up to 10 m fromthe road edge. Combining the percolation rate and the solute concentration to get an estimation of thetotal metal load provided more valuable hints than the sole consideration of soil solution concentrations.For this reason, more measurements of the percolation rate and solute concentration under definedroadside conditions should be undertaken.The“worst case" scenarios chosen for this study confirmed that in most cases the contribution oftraffic did not lead to dangerous metal inputs into the groundwater. The sensitive sites chosen herewere not as sensitive as presumed from the surrounding soils. Thus, only in very small groundwatercatchments metal loads might cause a problem.REFERENCESBarbosa, A. E. and Hvitved-Jacobsen, T. 1999. Highway runoff and potential for removal of heavy metals in an infiltrationpond in Portugal. Sci. Total Environ. 235: 151-159.Golwer, A. 1973. Interference of roads with groundwater. Zeitschrift der Deutsch. Geolog. Ges. (in German). 124: 435-446.Golwer, A. 1995. Traffic routes and their risk for ground water. Eclogae Geol. Helv. (in German). 88(2): 403- 419.Granato, G. E, Church, P. E. and Stone, V. J.1995. Mobilization of major and trace constituents of highway runoff ingroundwater potentially caused by deicing chemical migration. Transp. Res. Rec. 1483. pp. 92-103.Harrison, r. M. and Johnston, W. R. 1985. Deposition fuxes of lead, cadmium, copper and polynuclear aromatichydrocarbons (PAH) on the verges of a major highway. Sci. Total Environ. 46: 121-135.Harrison, R. M. and Wilson, s. J. 1985a. The chemical composition of highway drainage waters. I. Major ions and selectedtrace metals. Sci. Total Environ. 43: 63-77.Harrison, R. M. and Wilson, S. J. 1985b. The chemical composition of highway drainage waters. II. Chemical associationsof metals in the suspended sediments. Sci. Total Environ. 43: 79 -87.Kocher, B. and Wessolek, G. 2002. Transport of Traffic -Related Contaminants with Percolation Water. Final ReportFE 05.118/1997/GRB. Forschung StraBenbau und StraBenverkehrstechnik No. 864 (in German). Federal Ministry ofTrafic, Bonn, Germany. 99pp.Kordel, W., Dassenakis, M., Lintelmann, J. and Padberg, s. 1997. The importance of natural organic material forenvironmental processes in waters and soils. Pure and Applied Chemistry. 69(7): 1571-1 600.Krauth, K. and Klein, H. 1982. Investigations on the Quality of Surface Water from Motorways. Parts I and II. ForschungStraBenbau und StraBenverkehrstechnik No. 363 (in German). Federal Ministry of Traffic, Bonn, Germany. 73pp.Lagerwerf, J. V. and Specht, A. W.1970. Contamination of roadside soil and vegetation with cadmium, nickel, lead, andzinc. Environ. Sci. Technol. 4: 583-586.Legret, M. and Pagotto, C. 1999. Evaluation of pollutant loadings in the runoff waters from a major rural highway. Sci.Total Environ. 235: 143- -150.Lygren, E, Gjessing, E. and Berglind, L. 1984. Pollution transport from a highway. Sci. Total Environ. 33: 147-159.Makepeace, D. K., Smith, D. W. and Stanley, s. J. 1995. Uri中国煤化工mary of contaminant dataCritical Reviews in Environmental Science and Technology.Mikkelsen, P. S., Hafiger, M., Ochs, M., Tell, J. C, Jacobson,fHC N M H Grimental asessment of soiland groundwater contamination from two old infiltration systems for road run-off in Switzerland. Sci. Total Environ.189/190: 341 -347.Reinirkens, P. 1996. Analysis of emissions through trafic volume in roadside soils and their effects on seepage water. Sci.Total Environ. 189/190: 361-369.SOIL WATER AND HEAVY METAL TRANSPORTSpeerschneider, R. 1992. Dynamics of water and chloride in roadside soils: Impact of microstructure (texture) and surfacesealing (in German). Ph.D. Dissertation, Univ. of Hannover, Germany. 104pp.Wheeler, G. L. and Rolfe, G. L. 1979. The relationship between daily traffic volume and the distribution of lead in roadsidesoil and vegetation. Environmental Pollution. 18: 265- 274.Wessolek, G. and Facklam, M.1997. Site properties and water balance of sealed surfaces. Z. Pflanzenernihr. Bodenk.(in German). 160: 41- 46.Wigington, P. J., Cifford, W. R. and Grizard, T. J.1986. Accumulation of selected trace metals in soils of urban runoffswale drains. Water Resources Bulletin. 22: 73- -79.中国煤化工MYHCNMHG