The Science of the Total Environment 235 Ž1999. 151]159
Highway runoff and potential for removal of heavy metals in an infiltration pond in Portugal A.E. BarbosaU , T. Hvitved-Jacobsen En¨ ironmental Engineering Laboratory, Aalborg Uni¨ ersity, Sohngaardsholms¨ ej 57, DK-9000 Aalborg, Denmark
Abstract Highway runoff from IP 4, a mountain road in the north-east of Portugal, has been monitored using a system integrating a raingauge, a flowmeter and an automatic water sampler. Average daily traffic ŽADT. is 6000 and the study catchment has 5970 m2 of total area and 2500 m2 of road pavement. A single stormwater outlet discharges into an infiltration pond with overflow to a creek. Sampling was carried out before the runoff water entered the pond. Among the parameters analysed in the runoff water, the heavy metals cadmium ŽCd., chromium ŽCr., copper ŽCu., lead ŽPb. and zinc ŽZn. were emphasised because of their toxicity. Concentrations of Cd and Cr were usually lower than the detection limit Ž1 mgrl.. Copper levels found were between 1 and 54 mgrl; lead from 1 to 200 mgrl and zinc from 50 to 1460 mgrl. A first flush effect was observed, meaning that the first 50% of the runoff volume for each event typically transported between 61 and 69% of the total suspended solids, Zn, Cu and Pb loads. Runoff water is totally infiltrated into the pond and heavy metals are being sorbed to the soil. Soils used in infiltration ponds should have specific characteristics in order to perform effectively and ensure groundwater protection. Not only well-known soil texture and composition characteristics are relevant: the soil sorption capacity } the extension and reversibility of the processes } is of main importance in this kind of highway runoff treatment. Experiments concerning the sorption of Zn, Cu and Pb to soils, followed by desorption at pH values of 2, 4 and 6 were conducted in the laboratory. These experiments were performed with the soil existing at the highway IP 4 infiltration pond and with two other common types of Portuguese soils. The three types of soil showed different behaviours, which must be related to their characteristics; the soil pH seemed to play a significant role in controlling the Zn, Cu and Pb sorption processes. As expected, a lowering of the pH value increased the desorption rate. The infiltration pond soil is the one with the lowest sorption capacity, however, it showed a relatively high sorption strength which means that it is considered reliable, concerning highway runoff treatment and groundwater protection. Q 1999 Elsevier Science B.V. All rights reserved. Keywords: Highway runoff; Total suspended solids; Heavy metals; Copper; Lead; Zinc; Detention pond; Soils; Infiltration; Sorption; Portugal
U
Corresponding author. Tel.: q45-96358493; fax: q45-98142555. E-mail address:
[email protected] ŽA.E. Barbosa.
0048-9697r99r$ - see front matter Q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 Ž 9 9 . 0 0 2 0 8 - 9
152
A.E. Barbosa, T. H¨ it¨ ed-Jacobsen r The Science of the Total En¨ ironment 235 (1999) 151]159
1. Introduction Highway runoff water may be responsible for serious environmental impacts, especially in the long term. Highway runoff monitoring studies have evidenced already the types of pollutants present and proved how much pollutant loads and concentrations depend on road characteristics and rainfall pattern ŽHarrison et al., 1985; Hamilton et al., 1987; Revitt et al., 1987; Stotz, 1987; Dannecker et al., 1990; Hvitved-Jacobsen and Yousef, 1991; Hewitt and Rashed, 1992; Bardin et al., 1996.. If there is a risk that a road will cause environmental problems, a solution for highway runoff treatment must be found. Although different solutions have different layouts and design, the processes behind them are more or less the same: detention of runoff for a period of time which is sufficient to allow sedimentation of pollutants or infiltration and sorption to soils ŽHvitved-Jacobsen et al., 1994; Federal Highway Administration, 1996; Mikkelsen et al., 1996.. Biological and chemical processes may also play an important role ŽYousef et al., 1985; Breen et al., 1994; Mungur et al., 1995; Federal Highway Administration, 1996.. The present work concerns a section of the highway IP 4 in the north-east of Portugal and a detentionrinfiltration pond receiving runoff water from this section. The IP 4 road crosses a mountain area where groundwater is intensively used. In Portugal the land transportation network is based in roads and situations where water resources can be affected by highway pollutant runoff are common. No procedures exist to solve these problems and there is no national practice established on highway runoff monitoring. A main objective of the present work is to contribute to this practice. This present work deals with a monitoring programme designed to provide data on IP 4 runoff water } pollutant concentrations and loads } and laboratory research, concerning soil sorption processes. Analysis focused on heavy metals because of their toxicity. Although five metals, characteristic of the highway environment, were considered } Cd, Cu, Cr, Pb and Zn } results showed that Cd and Cr were below detection
limit Ž1 mgrl. and these were therefore excluded from further studies. Laboratory experiments performed to assess the potential and the effectiveness of treatment at the infiltration pond consisted in determining the soil sorption capacity and testing the reversibility of the process under changing pH conditions which is known to influence metal sorption to soils, especially under acid rain conditions. For this purpose, sorption to soils of a known amount of a metal was followed by desorption at three different pH values: 2, 4 and 6. Results for Zn, Cu and Pb are reported. As the sorption characteristics of a soil are dependent on its origin and composition, two other common Portuguese soils were also included in the experiments allowing a more general objective and extended discussion of the results.
2. The study site in north-east Portugal The study site is located in a mountain, semirural area with no other local sources of atmospheric pollution apart from the road itself. The closest town, Vila Real with 14 000 inhabitants, is located 14 km to the west. The road section is approximately 200 m long, with a longitudinal slope from 1.7 to 3.3%. Lanes are bounded by paved hardshoulders, and the study catchment has 5970 m2 of total area and 2500 m2 of road pavement. Average daily traffic ŽADT. volume is 6000. The catchment has a single stormwater outlet, discharging into an infiltration pond. The pond consists of two separate compartments, the first one with approximately 200 m3 and the second with a capacity of 400 m3. Although the first pond was designed as a pre-sedimentation compartment, the high infiltration rate of the soil and an over-sizing of the system make it a rather infrequent event that runoff water passes to the second pond, where the last phase of the treatment should occur. The system was designed as a wet pond but it is performing as an infiltration pond.
A.E. Barbosa, T. H¨ it¨ ed-Jacobsen r The Science of the Total En¨ ironment 235 (1999) 151]159
3. Material and methods 3.1. Monitoring system The field monitoring system consists of a raingauge, a flowmeter and an automatic water sampler placed at the pond’s inlet. Sampling was done before the runoff water entered the pond and was performed at different time intervals during the event. At an early stage of the project, manual sampling of the highway runoff was done, including sampling at the road. A few rainfall samples were also collected, using a simple procedure: a plastic bottle and a funnel were placed close to the road. Samples were transported to the laboratory and prepared for analysis within approximately 2]3 h. On two occasions sampling occurred in the evening and samples were transported to the laboratory the following morning. Total heavy metals were analysed after extraction in aqueous solution at pH- 2, followed by filtration through a 0.45-mm filter. Analyses were performed using either flame or graphite furnace atomic absorption spectrophotometer ŽAAS., depending on the concentration levels. Analytical procedures were adopted according to APHA, AWWA, WEF Ž1995.. Titrisol standards were used to prepare metal standard solutions. The calibration curves were prepared with a blank and more than three different dilutions of the standard. A blank and a standard were analysed after each set of approximately 10 samples. Replicate measurements on subsamples were used to control the results. 3.2. Soil sorption and desorption experiments Three different Portuguese soils were used for these experiments, named after the places where the samples were collected. Vila Real soil was taken in the study site area and it is similar to the one existing at the pond. The other two soils were from Costa da Caparica and Estremoz in central Portugal. These three soil samples represent typical Portuguese soils. Soil samples were collected within the top 25 cm rejecting the most superficial
153
2]3-cm layer. Sampling places were not too close to roads, neither under the influence of other sources of atmospheric pollution and have not been exposed to agriculture. The soil characteristics were determined at the Portuguese National, Institute for Agriculture Research ŽInstituto Nacional de Investigac¸˜ ao Agraria ´ . following their standard procedures } the soil pH was measured in an aqueous suspension of 1 part of soil to 2.5 parts of water. In order to improve the precision of the sorption experiments, soil samples were sieved through a nylon mesh to a fraction - 0.5 mm ŽChristensen, 1984. and air dried before performing the experiments. A soil suspension ratio of 1 g of dry matter to 100 g of de-ionised water was used in all the experiments and preliminary tests were performed to establish the procedures to be adopted. Sorption was performed during 24 h in a 10y3 M NaNO3 solution with 1.8 mgrl of Zn, Cu or Pb. Sorption was done using 5 g of soil. The soil suspension was shaken at times intervals of 1, 2, 3 and 5 h, to allow establishment of an equilibrium. At these times and at 24 h, the pH was measured and a sample was taken, filtered through a 0.45mm filter, acidified to pH- 2 and kept at 48C before analysis. Immediately after completion of each sorption experiment, soil and solution were separated and the soil was air dried before starting the desorption experiment. Batch desorption experiments were performed with 3 g of metal enriched soil under permanent stirring for 24 h in a 10y3 M NaNO3 solution. pH was set to the desired value Ž2, 4 or 6. at the beginning of the experiment. A pH measuring system and alarm settings permitted automatic pH control with nitric acid ŽHNO3 . or sodium hydroxide ŽNaOH. solutions within a difference of "0.03 of the desired value, in the cases of the Vila Real and Estremoz soils. In the case of the Costa da Caparica soil, the interval was "0.60 for pH values of 4 and 6 because otherwise the soil would require addition of a too big volume of the pH correction solution. Nevertheless, when samples were taken it was checked that the pH was 4 or 6 " 0.03. Samples were filtered and handled the way described before. The volume of acid or
154
A.E. Barbosa, T. H¨ it¨ ed-Jacobsen r The Science of the Total En¨ ironment 235 (1999) 151]159
base used in each batch reactor was taken into account in the calculations. A desorption at pH 2 was performed with the original soil samples to determine the existing metal levels and ensure that the work was done with clean soils.
4. Results and discussion 4.1. IP 4 runoff quality The concentrations of pollutants in IP 4 runoff water are summarised in Table 1. These results are from a total of 10 events, four of them with manual sampling. The average event mean concentrations ŽEMC. for TSS, Zn, Pb and Cu are presented in Table 2. Only some of the events provided all the information needed to calculate EMC values. The wide range of values observed is expected to be the result of variations in conditions such as traffic, interevent dry period, intensity and duration of rainfall and wind. The level of concentrations is within the range observed in similar studies ŽHarrison et al., 1985; Hamilton et al., 1987; Revitt et al., 1987; Stotz, 1987; Dannecker et al., 1990; Hvitved-Jacobsen and Yousef, 1991; Hewitt and Rashed, 1992; Bardin et al., 1996; Federal Highway Administration, 1996.. Nevertheless, compared with levels of Pb and Cu, Zn seems higher than expected. A possible explanation could be the existence of another source of Zn, independent of the traffic, which is the metallic guard-rails found on mountain roads. Two facts
support this hypothesis: Zn levels were higher at the beginning of the monitoring work, shortly after the road had been opened, and decreased 1 year later; additionally it was observed that dissolved Zn showed a much higher variation than dissolved Pb or Cu. The results also showed that the pollutants concentrations typically are higher at the beginning of an event and decrease afterwards. When plotting the pollutant load transported versus the percentage of runoff volume, the typical curve of the first flush effect was seen: the percentage of pollutant load transported was always higher than the corresponding percentage of runoff volume. This is illustrated in Fig. 1 for Cu and Pb and Table 3 presents some figures relating the percentage of TSS, Pb, Zn and Cu loads and the corresponding percentage of runoff volume, which transported these pollutants. The first half of the volume of the runoff from each event is transporting between 61 and 69% of TSS, Pb, Zn and Cu loads. According to the figures in Table 3, Pb and Zn exhibit a stronger first flush effect when compared to TSS and Cu. TSS and Zn parameters were below detection limit in 50% or more of the samples which represent some loss of information. Curves obtained for Pb and Cu must be closer to what may be expected in reality, because more data were available. Nevertheless, the existence of a first flush effect for all pollutants analysed is evident. This is a relevant information for detention pond volume design. Three precipitation samples were collected at the beginning of the storm events to provide
Table 1 Pollutant concentration levels in IP 4 highway runoff water Parameter
No. samples
Range of values
Average
S.D.
pH Conductivity ŽmSrcm. T. hardness Žmg CaCO3 rl. Total suspended solids Žmgrl. Pb Žmgrl. Zn Žmgrl. Cu Žmgrl.
127 127 123 73 123 123 123
5.9]7.2 8.8]183.8 1.5]67.2 - 8]147 - 1]199.5 - 50]1462 - 1]54.3
6.4 43.0 8.2 19.3 10.8 172 10.7
0.3 26.7 9.0 6.9 23.6 272 9.5
A.E. Barbosa, T. H¨ it¨ ed-Jacobsen r The Science of the Total En¨ ironment 235 (1999) 151]159 Table 2 Average of event mean concentration Parameter
No. events
Average EMC
S.D.
TSS Žmgrl. Pb Žmgrl. Zn Žmgrl. Cu Žmgrl.
3 5 3 5
8 12.5 308 24.1
14 5.2 178 9.2
information on the atmospheric pollution contribution to IP 4 runoff loads. Results for heavy metals concentrations are depicted in Table 4. Results from rainfall analysis indicate clearly the presence of metals in the atmosphere close to the road, probably including resuspended particles from the road. Concentrations may depend on several variables such as wind, rainfall characteristics and volume sampled. 4.2. Soil sorption and desorption experiments The characteristics of the three Portuguese soils are described in Table 5. Both the Vila Real and the Costa da Caparica soil are very sandy. The Costa da Caparica soil includes carbonates, which is the reason why the pH level in the desorption experiments could not be kept constant as easily as for the other two soils. The Estremoz soil distinguishes in having the highest content of silt, clay, organic matter and cation exchange capacity ŽCEC.. These parameters, together with soil pH, are the ones usually correlated with soil sorption capacity for metal cations ŽKorte et al., 1975; Soldatini et al., 1976; Farrah and Pickering, 1977; Garcıa-Miragaya and Page, 1978; Christensen, ´
155
Table 3 Relationship between percentage of pollutant versus percentage of runoff volume. Results from a best fit line for five or three events % Runoff
40 50 70 80
% Pollutant transported TSS
Pb
Zn
Cu
50 61 81 91
58 69 90 99
54 65 87 97
51 61 81 90
1984, 1989; Billett et al., 1991; Krosshavn et al., 1993; Zhu and Alva, 1993; Zhang et al., 1997.. There is usually a positive correlation between the soil potential capacity for metal cation sorption and the soil pH, organic matter Žo.m.., silt, clay and CEC. The effect of pH seems to be significant: for increased soil pH values there is an increase in the negative charge of the soil, less competition with protons for sorption sites and formation of hydroxyl complexes in solution. However, exceptions to this effect have been noticed ŽKorte et al., 1975; Garcıa-Miragaya and ´ Page, 1978; Billett et al., 1991.. There is proof of different behaviours for distinct metal cations ŽBillett et al., 1991; El-Hassanin et al., 1993; Krosshavn et al., 1993; Zhu and Alva, 1993.. Some authors concluded that Zn and Cu have similar sorption behaviours ŽBillett et al., 1991; Zhang et al., 1997.. There is much evidence of a compared solubility as follows: Zn ) Cu ) Pb, presenting Pb a very strong binding capacity ŽKorte et al., 1975; El-Hassanin et al., 1993; Krosshavn et al., 1993.. There are references in the literature to variations in metal sorption to
Fig. 1. Percentage of Pb and Cu load transported versus percentage of runoff volume. Values based on five events.
A.E. Barbosa, T. H¨ it¨ ed-Jacobsen r The Science of the Total En¨ ironment 235 (1999) 151]159
156
Table 4 Cencentrations of heavy metals in precipitation sample Event
Sample vol. Žml.
Pb Žmgrl.
Zn Žmgrl.
Cu Žmgrl.
1 2 3
80 20 120
3.0 26.2 21.9
- 50 144 54
5.7 32.9 25.3
soils due to the presence of calcium ŽCa2q . or carbonates ŽKorte et al., 1975; Riffaldi et al., 1976; El-Hassanin et al., 1993; Zhu and Alva, 1993.. The desorption experiment at pH 2 with the original clean samples showed levels of Zn, Cu and Pb within the range observed in normal arable soils ŽTjell and Hovmand, 1981; Billett et al., 1991.. The highest level of Pb Ž8 mgrg. was observed in the Estremoz soil. The Costa da Caparica soil had the highest content of Zn Ž31 mgrg. and Cu Ž4 mgrg.. Table 6 shows pH values measured during the sorption experiments. Usually the pH increased from the beginning of the experiment to the end, when the equilibrium was reached. As can be seen, the three soils kept their relatively different pH-values of Table 5. Generally, the Pb sorption showed lower pH values than the one with Zn, and the Cu sorption presented intermediate values. The results from the sorption and desorption experiments are shown in Fig. 2. The Costa da Caparica and the Estremoz soils have } compared with the Vila Real soil } a higher and almost similar sorption capacity. They sorbed almost the total amount of Zn, Cu and Pb available in solution Ž0.18 mgrg. whereas the Vila Real
soil sorbed an average of 0.06 mg Znrg and 0. 12 mg rg of Cu and Pb. The desorption results showed a clear inverse correlation between the desorbed metal and the pH. Zn desorption from the Estremoz soil, at pH 2 resulted in a value slightly higher than the sorbed Zn, which should be explained by experimental errors. The amount desorbed at pH 6 was very low and always - 8 mgrg in all cases. It is also possible to see that the soils presented a similar behaviour for Cu and Pb. The mass of metal retained by the soils at pH 2 ranged from 0.07 to 0 mgrg soil, the lowest values respecting to Zn. The value 0.07 mgrg was obtained for Pb retention by the Estremoz soil. The retention efficiencies of metals vary from 40 to 0%; the three lowest results concerning Zn. The average values of metal retention at pH 2 for the three soils were almost identical for Pb and Cu Ž0.43 mg Curg soil and 0.40 mg Pbrg soil. as it was the efficiency of retention Ž29% Cu and 25% Pb., whereas Zn presented much lower values: 0.002 mgrg soil and 4%. It should be noticed that the Vila Real soil although having the lowest retention capacity it exhibited the highest retention efficiency at pH 2 for Zn Ž10%. and Cu Ž33%.. The Estremoz soil is the most efficient one in retaining Pb Ž40%.. As a general conclusion of the sorption and desorption experiments, the Vila Real soil is reliable for use in infiltration ponds. The Estremoz soil compared to the Costa da Caparica soil has significantly higher values for clay, silt, o.m. and CEC which should therefore correspond to a higher number of sorption sites. The similar results obtained in the sorption experiments for these two soils could thus be ex-
Table 5 Characteristics of the three Portuguese soils Žfraction F 0.5 mm. Soil
pH ŽH2 O.
Organic matter Ž%.
Coarse sand Ž%.
Sand Ž%.
Silt Ž%.
Clay Ž%.
CaCO3 Ž%.
CEC ŽmEq.r100 g.
Vila Real C. Caparica Estrernoz
4.20 7.60 5.40
1.40 0.75 3.90
27.5 65.0 8.6
50.1 31.4 38.9
16.7 2.0 34.4
5.7 1.6 18.1
n.d.a 3.26 n.d.
10.44 8.43 14.97
a
n.d., not detectable.
A.E. Barbosa, T. H¨ it¨ ed-Jacobsen r The Science of the Total En¨ ironment 235 (1999) 151]159 Table 6 pH values measured during the sorption experiments
Vila Real
C. Caparica
Estremoz
pH variation
pH values at the end
Zn 4.75]5.40 Cu 4.48]4.93 Pb 3.66]4.19 Zn 6.60]7.41 Cu 5.90]7.75 Pb 4.42]6.91 Zn 6.01]6.40 Cu 5.61]6.87 Pb 4.71]5.49
4.77]5.00 4.50]4.68 4.05]4.19 7.15]7.41 7.58]7.75 6.80]6.91 6.18]6.40 6.20]6.87 5.29]5.49
plained by the soil pH values: the Costa da Caparica soil ŽpH 7.60. having a higher potential for sorption than the Estremoz soil ŽpH 5.40.. The Vila Real soil has intermediate values for clay and silt, o.m. and CEC compared with the other two soils. On the other hand it is the soil with the lowest pH Ž4.20. and it showed a high metal retention strength, at the pH values used in the desorption experiments. These results seem to indicate that the soil pH plays an important role in controlling metal up-
157
take by the three Portuguese soils. The Estremoz soil efficiency in adsorbing Pb could be attributed to the soil silt and clay content. Other factors can also affectrcontrol the process such as the existence of carbonates, the type of clay and metal background levels.
5. Conclusions The level of total suspended solids, Zn, Pb and Cu found in IP 4 runoff water are within the range of concentrations found in similar studies. A relatively high level of Zn could be explained by the existence of metallic guard-rails. The results also showed the existence of a first flush effect with the first half of the runoff volume for each event transporting typically between 61 and 69% of the total suspended solids, Zn, Cu and Pb. This information is useful for detention pond volume design. Well-known soil characteristics in terms of texture and composition are considered important for selection of soils used for infiltration ponds.
Fig. 2. Sorption of Zn, Cu and Pb Žmaximum of 0.18 mg Merg soil. by the three Portuguese soils and desorption at different pH values.
158
A.E. Barbosa, T. H¨ it¨ ed-Jacobsen r The Science of the Total En¨ ironment 235 (1999) 151]159
However, these characteristics do not provide sufficient information concerning sorption and accumulation of heavy metals. A procedure for determination of sorption and desorption of heavy metals to soil particles is evaluated. Soils which exhibit a high sorption capacity and a high resistance to desorption at low pH are considered ideal for pollutant removal in infiltration ponds. Three different and common types of Portuguese soils have been included in this study. These soils show different characteristics in terms of their ability to remove and accumulate Zn, Cu and Pb. The soil type which exists at the IP 4 infiltration pond } Vila Real } presents the lowest sorption capacity, nevertheless when compared with the Costa da Caparica and the Estremoz soils it showed a relatively high metal retention strength at low pH values. The work done introduces information which is useful in designing infiltration andror detention ponds for highway runoff water treatment, particularly under conditions found in southern Europe. Further studies concerning the physical and water transport characteristics of the soils will be needed.
Acknowledgements The Portuguese National Road Administration ŽJunta Autonoma de Estradas, JAE. supported ´ part of the monitoring work. Highway runoff analysis were performed at Chemistry Section of the University of Tras-os-Montes and Alto Douro ´ ŽUTAD., Portugal. The Ph.D scholarship was granted by the PRAXIS XXI Programme. References APHA, AWWA, WEF. Standard methods for the examination of water and wastewater. Maryland: American Public Health Association, 1995. Bardin JP, Cres FN, Touzo A, Verjat JL. Pollution measurements in a stormwater settling and infiltration basin. Hanover, Germany: The Seventh International Conference on Urban Storm Drainage, 1996;III:1611]1616. Billett MF, Fitzpatrick EA, Cresser MS. Long-term changes in the Cu, Pb and Zn content of forest soil organic horizons
from north-east Scotland. Water Air Soil Pollut 1991;59:179]191. Breen PF, Mag V, Seymour BS. The combination of flood-retarding basin and wetland to manage the impact of urban runoff. Water Sci Tech 1994;29Ž4.:103]109. Christensen TH. Cadmium soil sorption at low concentrations: I. Effect of time, cadmium load, pH and calcium. Water Air Soil Pollut 1984;21:105]114. Christensen TH. Cadmium soil sorption at low concentrations, vol. VIII. Correlation with soil parameters. Water Air Soil Pollut 1989;44:71]82. Dannecker W, Au M, Stechmann H. Substance load in rainwater runoff from different streets in Hamburg. Sci Total Environ 1990;93:385]392. El-Hassanin AS, Labid TM, Dobal AT. Potential Pb, Cd, Zn and B contamination of sandy soils after different irrigation periods with sewage effluent. Water Air Soil Pollut 1993;66:239]249. Farrah H, Pickering WF. Influence of clay-solute interactions on aqueous heavy metal ion levels. Water Air Soil Pollut 1977;8:189]197. Federal Highway Administration. Evaluation and management of highway runoff water quality. Publication No. FHWA-PD-96-032. Washington: US Department of Transportation, 1996:457. Garcıa-Miragaya J, Page AL. Sorption of trace quantities of ´ cadmium by soils with different chemical and mineralogical composition. Water Air Soil Pollut 1978;9:289]299. Hamilton RS, Revitt DM, Warren RS, Duggan MJ. Metal mass balance studies within a small highway dominated catchment. Sci Total Environ 1987;59:365]368. Harrison RM, Johnston WR, Ralph JC, Wilson SJ. The budget of lead, copper and cadmium for a major highway. Sci Total Environ 1985;46:137]145. Hewitt CN, Rashed MB. Removal rates of selected pollutants in the runoff waters from a major rural highway. Water Res 1992;26Ž3.:311]319. Hvitved-Jacobsen T, Yousef YA. Highway runoff quality, environmental impacts and control. In: Hamilton RS, Harrison RM, editors. Highway pollution. Elsevier, 1991:165]208. Hvitved-Jacobsen T, Johansen NB, Yousef YA. Treatment systems for urban and highway run-off in Denmark. Sci Total Environ 1994;146r147:499]506. Korte NE, Skopp JS, Niebla EE, Fuller WH. A baseline study on trace metal elution from diverse soil types. Water Air Soil Pollut 1975;5:149]156. Krosshavn M, Steinnes E, Varskog P. Binding of Cd, Cu, Pb and Zn in soil organic matter with different vegetational background. Water Air Soil Pollut 1993;71:185]193. Mikkelsen PS, Hafliger M, Ochs M, Jacobsen P, Tjell JC, ¨ Boller M. Pollution of soil and groundwater from infiltration of highly contaminated stormwater } A case study. Hannover, Germany: The Seventh International Conference on Urban Storm Drainage, 1996;II:707]712. Mungur AS, Shutes RBE, Revitt DM, House MA. An assessment of metal removal from highway runoff by a natural wetland. Water Sci Tech 1995;32Ž3.:169]175.
A.E. Barbosa, T. H¨ it¨ ed-Jacobsen r The Science of the Total En¨ ironment 235 (1999) 151]159 Revitt DM, Hamilton RS, Warren RS. The transport of heavy metals within a small urban catchment. Sci Total Environ 1987;93:329]337. Riffaldi R, Levi-Minzi R, Soldatini GF. Pb adsorption by soils, vol. II. Specific adsorption. Water Air Soil Pollut 1976;6:119]128. Soldatini GF, Riffaldi R, Levi-Minzi R. Pb adsorption by soils, vol. I. Adsorption as measured by the Langmuir and Freundlich isotherms. Water Air Soil Pollut 1976;6: 111]118. Stotz G. Investigations of the properties of the surface water run-off from federal highways in the FRG. Sci Total Environ 1987;59:329]337. Tjell JC, Hovmand MF. Metalkoncentrationer i Dansk
159
Dyrkede Jorder. In: Polyteknisk Forlag, editor. Slammets Jordbrugsanvendelse. Lyngby: III Delrapporter, 1981: 95]100. Yousef YA, Wanielista MP, Harper HH. Fate of pollutants in retentionrdetention ponds. In: Wanielista MP, Yousef YA, editors. Stormwater management ‘an update’. University of Central Florida, Environmental Systems Engineering Institute, Publication No. 85]1, 1985:259]275. Zhang M, Alva AK, Calvert DV. Chemical association of Cu, Zn, Mn and Pb in selected sandy citrus soils. Soil Sci 1997;162Ž3.:181]188. Zhu B, Alva AK. Differential adsorption of trace metals by soils as influenced by exchangeable cations and ionic strength. Soil Sci 1993;155Ž1.:61]66.
