Content, Distribution, and Solubility of Cadmium in Arable and Forest Soils Martin K. Andersen,* Anne Refsgaard, Karsten Raulund-Rasmussen, Bjarne W. Strobel, and Hans C. B. Hansen ABSTRACT exchange capacity (CEC), clay content, organic matter, Afforestation of former farmland decreases soil pH and thus in- and other metal ions present (Christensen, 1989; Wil- creases the solubility of Cd in the soil, which may cause Cd leaching kens and Loch, 1997; McBride et al., 1997; Ma et al., to streams and groundwater. The Cd concentration in soil and soil so- 1997; Sza ´ ´kova et al., 1999). Cadmium release from soil in- lution were determined in 11 pairs of Danish arable and forest soil creases substantially when pH drops below 4.5 (Berg- profiles representing three different texture classes (sand, loamy sand, kvist et al., 1989; McBride et al., 1997). This is especially and sandy loam). The soil pH did not change or decrease with depth true for sandy soils with low CEC, low acid neutral- through the arable profiles, but did increase with depth in the forest ization capacity, and low ability of the subsoil to sorb profiles. Significantly higher Cd contents were found in the upper 30 cm of the arable soil compared with that of the forest soil. The total Cd ions. Such soils may be vulnerable to Cd leaching soil Cd concentrations correlated with the effective cation-exchange after afforestation. capacity (ECEC), clay content, and organic matter content, but not the Arable land receives Cd from the application of ferti- soil pH. The soil solution pH was unchanged or decreasing downwards lizers, lime, and to some extent from sewage sludge (Bak through the arable profiles, but increasing with depth in the forest et al., 1997). Atmospheric deposition is also a substan- profiles. The soil solution concentration of Cd was significantly higher tial source of Cd to the soil. Atmospheric deposition of in the forest soils than in the arable soils. The Cd concentration in Cd is higher in forest areas compared with arable land the soil solution decreased as pH increased. Both total soil and soil (Wilcke et al., 1999). A considerable reduction of the solution Cd concentrations were higher in the sandy loam soils com- atmospheric Cd deposition in Denmark has taken place pared with the loamy sand and sand soils. It is concluded that afforesta- during the last decade (Hovmand and Kemp, 2000), tion may lead to higher soil solution concentrations of Cd as decreasing pH and ECEC diminish Cd retention and reduces Cd concentrations but years of high atmospheric Cd deposition may have in the forest topsoils. increased the total Cd content in the soil. Whether the Cd added from the various sources results in a net loss or accumulation of Cd from the soil profile will depend on the soil parameters important to Cd solubility, for L arge areas of intensively cultivated agricultural land are afforested in northern Europe to reduce both the agricultural production and the release of nutri- example, pH and the amount of sorption sites available for Cd. Studies of the relation between Cd concentrations in ents to the surrounding aquatic environment. The areas the soil solid and solution phases of natural (unpolluted) afforested in Denmark are primarily low productivity soil systems are scarce. To determine the effects of af- agricultural soils and areas that directly impact ground- forestation on Cd concentrations, solubility and distri- water reservoirs, the latter for protection of ground- bution in soil, comparisons between the two land-uses water from pesticides and other unwanted compounds were made with soils of different texture. The total soil used in agriculture. Unfortunately, soil acidification fol- and solution Cd concentrations of 11 pairs of Danish ar- lowing afforestation may cause mobilization of heavy able and afforested soil profiles were determined as metals in the soil, especially the toxic Cd that might be well as other soil properties, that is, pH, organic matter leached to streams and groundwater (Egli et al., 1999; content, and exchangeable base cations. The soil proper- Jug et al., 1999). In the afforested areas, lime is no longer ties important to Cd retention and solubility in these applied and in addition an increased production of or- soils are discussed. ganic acids and atmospheric acid deposition causes a de- cline in pH, which increases Cd solubility. Although in- creased Cd solubility as a consequence of complexation MATERIALS AND METHODS with dissolved organic matter in the afforested soils, Sampling seems negligible (Strobel et al., 2001). Soil samples were collected from November 1998 to January Cadmium can be bound in soil by simple electrostatic 2000 from 11 sites representing three different textural classes: forces or intimately associated with metal oxides, car- sand, loamy sand, and sandy loam (Table 1). The soils vary with bonates, and organic matter. It is also found that the Cd respect to acidity, clay content, and CEC. Each site is com- solubility increases as pH decreases (Christensen, 1989; prised of adjacent arable and forest profiles developed on the Chlopecka et al., 1996). A number of investigations have same parent material. The forest plots vary in tree species and shown Cd solubility to be dependent also on the cation- stand age but they were all planted on former farmland. All profiles were excavated to a depth of at least 125 cm and clas- sified according to Soil Taxonomy (Soil Survey Staff, 1997). M.K. Andersen, B.W. Strobel, and H.C.B. Hansen, Chemistry Dep., Soil was sampled from each horizon in all the profiles (a The Royal Veterinary and Agricultural Univ., Thorvaldsensvej 40, DK-1871 Frederiksberg C; A. Refsgaard and K. Raulund-Rasmussen, total of four to eight horizons per profile). At seven of the Danish Forest and Landscape Research Institute, Hoersholm 11 sites soil solutions were isolated from soil samples taken Kongevej 11, DK-2970 Hoersholm. Received 20 June 2001.*Corre- sponding author (email@example.com). Abbreviations: ECEC, effective cation-exchange capacity; ICP-OES; inductively coupled plasma optical emission spectra; ISO, Interna- Published in Soil Sci. Soc. Am. J. 66:1829–1835 (2002). tional Organization for Standardization. 1829 1830 SOIL SCI. SOC. AM. J., VOL. 66, NOVEMBER–DECEMBER 2002 Table 1. Site description and classification according to soil taxonomy (1997). Site Land use UTM coordinates Classification Texture class Sediment Nørland Arable 32VNH114144 Typic Durorthod Sand Fluvial sand Forest Oxyaquic Haplorthod Sand Fluvial sand Klosterhede Arable 32VMH681589 Typic Haplorthod Sand Fluvial sand Forest Typic Haplorthod Loamy sand Fluvial sand Baldersbæk Arable 32UMG947668 Psammentic Haplumbrept Sand Fluvial sand Forest Typic Durorthod Sand Fluvial sand Kompedal Arable 32VNH187294 Psammentic Haplumbrept Sand Fluvial sand Forest Typic Durorthod Sand Fluvial sand Løvbakke Arable 32UNG997668 Oxyaquic Haplumbrept Loamy sand Till Forest Typic Haplorthod Loamy sand Till Tisted Nørskov Arable 32VHN622944 Entic Haplumbrept Loamy sand Till Forest Entic Haplumbrept Loamy sand Till Løvenholm Arable 32VHN944581 Oxyaquic Glossudalf Loamy sand Till Forest Entic Haplumbrept Loamy sand Till Strødam Arable 33UUC30153 Psammentic Haplumbrept Loamy sand Till Forest Typic Udipsamment Loamy sand Till Jels Arable 32UNG141372 Typic Paleudult Sandy loam Till Forest Typic Paleudult Loamy sand Till Stenholtsvang Arable 33UUC348045 Oxyaquic Hapludalf Sandy loam Till Forest Oxyaquic Hapludalf Sandy loam Till Christianssæde Arable 32UPF522726 Typic Hapludalf Sandy loam Till Forest Mollic Hapludalf Sandy loam Till from each horizon in the profiles and from the remaining NH4NO3 and exchangeable acidity was determined by end- four sites (Tisted Nørskov, Løvenholm, Christianssæde, and point titration of the NH4NO3 extract to pH 6.00 (Stuanes et Stenholtsvang) soil solutions were isolated from soil samples al., 1984). The ECEC of the soils was calculated as the sum collected using a soil auger (Eijkelkamp Agrisearch Equip- of charges of the base cations and the exchangeable acidity. ment, Giesbeek, NL) and only from three depths (0–5, 20–25, The soil pH was determined in 0.01 M CaCl2 with a soil/ and 70–90 cm). The soil samples were stored in polyethylene solution ratio of 1:2.5 using a Metrohm (Herisau, Switzerland) bags at 5 C. The natural moist soil was placed in a polyethylene glass electrode (6.0202.110) connected to a Metrohm 691 pH cup with holes in the bottom and this cup was placed on top meter. Soil bulk density was determined by drying soil from of a polytetrafluorethene (Teflon) cup. By centrifugation for each horizon sampled in metal rings of a known volume. 20 min. at 3200 g (3500 rpm) (Allegra 6R, Beckman Coulter, Palo Alto, CA), the soil solution was isolated from the natural Soil Solution Analyses moist soil samples into the Teflon cup (Davies and Davies, 1963). The soil solutions were filtered through a 0.45- m cellu- The Cd concentrations in the soil solutions were determined lose membrane filter (ME 25, Schleicher & Schnell, Dassel, by graphite furnace atomic absorption spectrometry (GFAAS) Germany), transferred into polyethylene containers, and stored [Perkin Elmer (Wellesley, MA) 5100, Zeeman (Wellesley, at 4 C until analysis. The centrifuged soil were transferred MA) 5100]. For conservation of the samples, the soil solution into paper bags, dried at 55 C, sieved through a 2-mm screen for Cd analysis were acidified to 1.0% (wt./wt.) HNO3 (supra and stored in polyethylene containers. All laboratory equip- pure). The detection limit was 0.05 g L 1 and maximum al- ment used for experiments were soaked in 6.5% HNO3 (ana- lowed RSD value [RSD (standard deviation)/(mean value) lytical grade) for 1 h and rinsed thoroughly in MilliQ-water. 100] between the measurements was 5%. Every sixth sample run was a control of known Cd concentration. The concentra- Soil Solid Analyses tion of total organic C (TOC) in the soil solutions was deter- mined using Shimadzu (Kyoto, Japan) 500 total C analyzer, The Cd concentrations in the soil samples were determined and soil solution pH was determined immediately after the cen- by aqua regia extraction [International Organization for Stan- trifugation, using the same equipment as described for soil dardization (ISO), 1995]. The soils were ground in an agate pH. The concentrations of Al, Ca, K, Mg, and Na were deter- mortar and 2.50 g of soil were transferred into borosilicate mined by ICP-OES (Optima 3000 XL axial view, AS 90, Perkin tubes and 15.0 mL 30% (wt./wt.) HCl (supra pure) was added Elmer, Norwalk, CT). followed by 5.0 mL concentrated HNO3 (supra pure), then the suspensions were left covered in a heating block for 2 d at room temperature. The temperature was then slowly in- Calculations and Statistics creased to 90 C and the soils were digested for 2 h. After Average profiles of soil pH and soil Cd concentrations as cooling the solutions were filtered [Whatman 42 (Kent, UK) well as soil solution pH and soil solution Cd concentration filter, prewashed in 12.5 mM EDTA and rinsed in MilliQ were generated for each texture class (Fig. 1 and 2). Because water] and diluted to 50 mL using MilliQ water. The concen- the horizons in the different profiles were not sampled at the trations of Cd in the digestates were determined using induc- same depths a data point was calculated for each centimeter, tively coupled plasma optical emission spectra (ICP-OES; as the mean value of all the observations at this depth in all the Optima 3000 XL axial view, AS 90, Perkin Elmer, Norwalk, profiles in the respective texture class. Each of these calculated CT) with a detection limit of 1 g Cd L 1 equal to 20 g Cd points were then averaged across the five overlying and five kg 1 soil. The determination of aqua-regia extractable Cd was underlying points to smooth the curve. The average solution carried out in duplicates. Determination of soil texture was pH and Cd concentration profiles were generated for every done using the hydrometer method (Gee and Bauder, 1986). centimeter using the same procedure. Total contents of C and N in the soil were determined using The volume based Cd concentration, hereafter referred to a CN-analyzer (LECO CNS-2000, St. Joseph, MI). Exchange- as Cd content in the soil was calculated for four depths at able base cations were determined by extraction with 1.0 M each profile: 0 to 30, 30 to 60, 60 to 90, and 90 to 120 cm, by ANDERSEN ET AL.: CONTENT, DISTRIBUTION, AND SOLUBILITY OF CADMIUM 1831 Fig. 1. Vertical distribution of Cd contents ( g g 1) and pH of (a, b) arable and (c, d) forest soils in three texture classes. multiplying the mass-based soil Cd concentration found in wash plains in the western part of Denmark and the till each horizon with the soil bulk density and the relative part soil types are dominated by moraine soils in the eastern that the horizon comprises of the selected soil layer. The aver- part of the country. Norway spruce [Picea abies (L.) age soil solution concentrations were calculated for the same H. Karst.] is the dominating forest vegetation in Western four depths. Differences between the textural classes were analyzed depthwise by a general linear model (SAS Institute, Denmark whereas deciduous forest of beech (Fagus 1999). A similar analysis was made within each combination of sylvatica L.) and oak (Quercus robur L.) is the most land use and texture class to analyze the differences between common in Eastern Denmark. The soil profiles were the depths. In both analyses the Cd content in the soil and the grouped in three texture classes based on the texture soil solution concentration were log-transformed to normalize in the top 0 to 30 cm of the soil. Nørlund, Klosterhede, and homogenize the variances. Back-transformed mean values Baldersbæk, and Kompedal were characterized as sand of the total soil content and soil solution concentration of Cd soils and the clay content in the horizons in the profiles are reported. To test the difference between the forest and arable soils ranged between 2 and 9%. Løvbakke, Tisted Nørskov, the arable/forest ratios of the total Cd contents and the soil Løvenholm, and Strødam were in the loamy sand group solution concentrations in each soil profile pair at the four with clay contents between 5 and 17% and the sandy depths were calculated. These figures were then tested depth- loam soils comprised of Jels, Stenholtsvang, and Chris- wise within each texture class to vary significantly from one tianssæde with clay contents between 8 and 28% as an indicator of different content. The test was carried out (Tables 1 and 2). The content of soil organic C in the as a t-test (MEAN-procedure) (SAS Institute, 1999). top horizons of sand soils were in the range from 13 to Linear correlations between the soil Cd concentrations and 62 g kg 1, in the loamy sand soils between 6 and 36 g the ECEC, organic C, and the clay content were carried out by GLM (SAS Institute, 1999). Cadmium concentrations be- kg 1 and in the sandy loam soils from 12 to 26 g kg 1 low the detection limit of 20 g kg 1 in soil solids or 0.05 g (Table 2). There was not found any significant differ- L 1 in soil solutions were included in the statistical calculations ence between the organic matter content in forest and with half the detection limit. arable soil. RESULTS Soil Solids Sites The Cd concentrations in the O-horizons were be- The soils in the present study represent typical Danish tween 115 and 321 g kg 1, equivalent to a Cd content soils. The sandy soils are predominately found on out- of 0.8 and 2.3 mg m 2. The contribution of the O-hori- 1832 SOIL SCI. SOC. AM. J., VOL. 66, NOVEMBER–DECEMBER 2002 Fig. 2. Vertical distribution of soil solution concentration of Cd ( g L 1) and pH in solution of (a, b) arable and (c, d) forest soils in three texture classes. zons to the total Cd content of the soil profiles was ble 2). In arable soils the Cd concentration in the topsoil negligible; therefore data from the O-horizons are ex- were up to five times higher than in the lower horizons cluded from all calculations based on soil depth. and a significantly higher Cd content was found in the The Cd concentration in the top horizon of the min- topsoil for the sand and loamy sand soils (Table 3). The eral soils was in the range from 24 to 293 g kg 1 (Ta- Cd content did not vary significantly between depths in Table 2. Selected mean results from each horizon type (A, B, or C) within each land-use and texture class. The standard deviations are given in parenthesis. Textural class, Number of ECEC Cd (AR)† Ca‡ 1 1 Land-use Horizon horizons n pH soil C g kg Clay % molc g gg 1 molc g 1 Density Mg m 3 Sand, arable A 6 5.28 (0.55) 23.9 (15.6) 3.9 (1.3) 516 (354) 0.090 (0.071) 473 (327) 1.28 (0.28) B 14 4.82 (0.33) 4.6 (7.6) 3.0 (1.3) 80 (122) 0.021 (0.014) 64 (109) 1.50 (0.08, n 13) C 4 4.88 (0.15) 0.5 (0.3) 1.8 (0.5) 18 (3) 0.012 (0.013) 9 (3) 1.50 (0.09) Sand, forest A 6 3.05 (0.17) 26.2 (23.4) 2.6 (1.1) 51 (44) 0.064 (0.055) 13 (14) 1.23 (0.19) B 13 3.96 (0.59) 19.6 (22.1) 4.6 (2.4) 39 (40) 0.069 (0.063) 4 (5) 1.26 (0.28) C 5 4.56 (0.07) 1.2 (0.6) 2.3 (0.9) 7 (1) 0.024 (0.011) 1 (0) 1.49 (0.12) Loamy sand, arable A 5 5.53 (0.62) 21.6 (5.6) 7.4 (2.1) 519 (529) 0.223 (0.119) 468 (486) 1.32 (0.08) B 10 5.36 (0.66) 6.0 (7.2) 9.4 (4.0) 177 (200) 0.039 (0.015) 153 (186) 1.44 (0.19) C 6 4.86 (0.58) 0.6 (0.4) 6.8 (4.5) 40 (77) 0.019 (0.013) 35 (70) 1.59 (0.15, n 4) Loamy sand, forest A 4 3.62 (0.23) 19.3 (12.5) 6.1 (1.2) 74 (117) 0.051 (0.019) 42 (70) 1.16 (0.31) B 10 4.36 (0.48) 6.4 (4.5) 7.0 (3.0) 33 (23) 0.052 (0.023) 8 (7) 1.38 (0.16) C 7 4.63 (0.37) 1.2 (1.0) 7.6 (5.2) 44 (48) 0.040 (0.031) 17 (23) 1.57 (0.09, n 5) Sandy loam, arable A 3 6.17 (0.47) 16.2 (5.1) 13.2 (6.1) 373 (480) 0.195 (0.051) 327 (422) 1.54 ( , n 2) B 7 6.21 (0.47) 2.9 (1.9) 14.6 (5.1) 266 (367) 0.085 (0.060) 222 (303) 1.61 (0.13, n 4) C 3 6.07 (1.54) 1.6 (1.0) 17.5 (9.2) 336 (429) 0.121 (0.137) 265 (323) 1.80 ( , n 2) Sandy loam, forest A 4 4.46 (0.67) 15.5 (8.6) 14.1 (4.3) 85 (30) 0.127 (0.090) 52 (27) 1.24 (0.37) B 7 5.23 (0.99) 3.1 (2.6) 21.6 (5.4) 118 (36) 0.094 (0.063) 71 (50) 1.63 (0.09) C 3 5.57 (1.99) 1.8 (1.2) 16.8 (6.6) 138 (64) 0.134 (0.128) 67 (48) 1.77 ( , n 2) † Half the detection limit is used when concentration is below the detection limit. ‡ Exchangeable Ca2 . ANDERSEN ET AL.: CONTENT, DISTRIBUTION, AND SOLUBILITY OF CADMIUM 1833 Table 3. Cadmium content (back-transformations of the logarithm means) in soils and mean values of Cd concentration in soil solution at four depths of the soil profiles investigated. Different letters indicate significant difference (p 0.05) of Cd content and Cd con- centration between texture classes within each soil depth and land-use. Different numbers indicate significant difference (p 0.05) of Cd content and Cd concentration between depths within each texture class and land-use. Arable Forest Arable/Forest Cd in soil Cd in soil solution Cd in soil Cd in soil solution Depth Texture class mg m 2 gL 1 mg m 2 gL 1 Solid Solution a 0–30 cm: Sand 141 1 0.14 92 0.59 1.55 0.23* ab Loamy sand 206 1 0.20 1 68 0.90 12 3.17* 0.21* b Sandy loam 330 0.06 150 1.19 1.47 0.11** 30–60 cm: Sand 32 a 2 0.09 a 35 0.34 0.70 0.21 Loamy sand 67 ab 2 0.08 a 2 43 1.72 1 1.94 0.05* Sandy loam 164 b 0.02 b 167 2.69 0.92 0.01† 60–90 cm: Sand 24 a 2 0.15 31 a 0.47 0.63 0.13 b Loamy sand 66 2 0.08 2 44 ab 0.43 2 1.14 0.17* b Sandy loam 136 0.08 167 b 3.46 0.82 0.03** 90–120 cm: Sand 29 a 2 0.14 36 a 0.31 0.92 0.39 Loamy sand 61 ab 2 0.06 2 47 ab 0.312 1.03 0.21 b Sandy loam 134 0.30 174 b 3.46 0.89 0.08† * and ** indicate p 0.05 and p 0.01 respectively, for the t-test of the hypotheses that Arable/Forest-ratio 1. † No statistical difference due to missing values. the arable sandy loam soils (Table 3, Fig. 1a). The forest (Fig. 2c). In the 0- to 30-cm layer the Cd concentration in soils, on the other hand, showed a uniform Cd distribu- the forest soil solutions were higher than in the arable tion with depth, even though the sand soils showed a soils and the same was found for the loamy sand and slightly elevated level of the Cd content in the topsoil sandy loam soils in the 30- to 60- and 60- to 90-cm layers (Table 3, Fig. 1c). (Table 3). All the layers of the arable sandy loam soils and the The pH in the soil solutions of arable sandy loam 60- to 90- and 90- to 120-cm layers of the forest sandy soils were about 7 in the upper layers and about 5.5 in loam soils had significantly higher Cd content than the the lower layers. In the soil solution of sand and loamy loamy sand and sand soils (Table 3, Fig. 1a,c). Among sand soils, the pH was below 5 (Fig. 2b). There was a the arable soils, the sand soils had a lower Cd content tendency of the pH values to decrease with depth in in the topsoil, than the loamy sand and sandy loam soils. arable soils and increase with depth in forest soils This was not the case in the forest soils (Table 3). (Fig. 2b, d). The variations of the pH between texture The average Cd content in the 0- to 30-cm layer of classes in the forest soil solutions were small (Fig. 2d). the loamy sand soil was significantly higher in the arable soils than in the forest soils. The same trends were also found in the 0- to 30-cm layer in the two other texture DISCUSSION classes, but they were not significant (Table 3). No clear Soil Solids trends were found for the deeper layers. In the arable soils, pH was almost constant throughout The concentrations of Cd in the soils studied were the profile except for a gradually decrease with depth in within the ordinary range for natural soils in Denmark, the sandy loam soils (Fig. 1b). The upper 20 cm of the which were found to be 0.07 mg kg 1 for forested sand forest soils was more acid compared with the subsoil soils and 0.22 mg kg 1 for arable loam soil (Bak et al., (Fig. 1d). 1997). Higher Cd contents were found in the topsoil than in the lower layers of the arable soils, whereas no difference between topsoil and subsoil was found in the Soil Solutions forested profiles. This suggests that agriculture tends to The Cd concentrations in soil solutions were in the enhance the Cd retention in the topsoil, or that the range of 0.05 to 4.0 g L 1 (Fig. 2). In the arable arable soils receive a higher Cd input in the topsoil than loamy sand soils, a significant higher concentration was the forest soils. A significantly higher Cd content was found in the 0- to 30-cm layer than in the subsoil and found in arable soils than in forest soils within the upper in the loamy sand forest soils significant higher concen- 30 cm of the soil (p 0.046), but not within the upper trations were found in the 0- to 30- and 30- to 60-cm 120 cm (p 0.612) when all the texture classes were layers than in the deeper layers (Table 3). The Cd con- included in the calculation. An average of 45% of the centration did not differ between the four depths in the total Cd in the arable soil profile from 0 to 120 cm was sand soils and the sandy loam soils. In the loamy sand located in the top 30 cm, whereas only an average of soils the Cd concentration was higher in the upper soil 30% was found in the upper 30 cm in forest soils. Higher layers than in the lower soil layers (Table 3). contents of Cd in arable soils than in forest soils have The Cd concentrations in the soil solutions of ar- been reported previously (Bak et al., 1997; Romkens ¨ able soils were low ( 0.2 g L 1 ) in all texture classes and Salomons, 1998). The higher Cd contents found in (Fig. 2a), whereas the solution concentrations of Cd in the top layers of the arable soils compared with the the forest soils were higher (between 0.2 and 3.3 g L 1) forest soils are likely to originate from fertilizer used in 1834 SOIL SCI. SOC. AM. J., VOL. 66, NOVEMBER–DECEMBER 2002 Table 4. Regression models of the logarithm to the total Cd con- centration ( g g 1 ) in soil and the soil parameters effective cation-exchange capacity (ECEC, cmol kg 1, clay, %, and organic C, g kg 1). Model R2 p Arable and forest soils log Cd 1.206 0.008 ECEC 0.011 C 0.033 Clay 0.001 ECEC Clay 0.54 0.0001 Arable soils log Cd 1.256 0.010 ECEC 0.61 0.0001 Forest soils log Cd 1.319 0.25 Clay 0.012 C 0.41 0.0001 intensive agriculture (Holmgren et al., 1993) or because Fig. 3. The concentration of Cd ( g L 1) versus pH in the soil solu- of more intensive leaching from the forest profile tions from arable and forest soils. Log Cd 0.81 0.30pH, (Bergkvist et al., 1989). R2 0.39. The forested sand soils also had a significantly higher content of Cd in topsoil than in subsoil, whereas this concentration in soil solutions were found by Bergkvist was not the case in the loamy sand and sandy loam soils (1987) in Swedish soils and by Gooddy et al. (1995) in (Table 3). The higher Cd content in the upper layers of investigations of an acid sandy forest soil in southern the forest sand soils was consistent with a high content England. Solution concentrations above 1.0 g L 1 were of soil organic matter. The organic-rich Bh-horizons found in the forest soils from Jels, Løvbakke, and Strø- were also enriched with Cd, suggesting that Cd is re- dam that are all sandy loam or loamy sand soils. The tained by organic matter as observed by Wilkens and concentrations in these soils rose with depth as illus- Loch (1997). For the forested sand soils, the higher trated in Fig. 2c. In the sand soils, the general pattern content of organic matter seemed to retain the Cd de- of the Cd concentration was a decrease with soil depth spite the low pH. This corresponds well with the models (Fig. 2c). One single horizon in Baldersbæk had a high presented in Table 4, where the soil C content is in- concentration of Cd (4.0 g L 1). cluded in the modeling of the Cd concentration in the Variations in Cd concentrations among texture forest soils. classes were negligible in arable soils but pronounced Significantly higher Cd contents were found in the in forest soils (Fig. 2a, c). The Cd concentrations in the sandy loam soils than in the sand and loamy sand soils. subsurface horizons of the forest soils were higher in This was most pronounced in the forest soils (Table 3). the sandy loam soils compared with the sand soils, but no Higher clay contents cause a higher buffer capacity to- significant difference between the texture classes were wards acidification, and thus forest soils on sandy loam found, presumably because of the few observations (Ta- will exhibit a slower pH decrease, than more coarse- ble 3). Thus, the concentration of mobile Cd in soil so- textured soils. The higher total Cd content in the sandy lution seems to increase with the content of clay in soil. loam soil than in the sand and loamy sand soil can arise The low Cd concentrations in soil solutions in all the from a higher capacity to sorb the Cd added to the soil arable profiles are attributed to the relatively high pH and a higher Cd content in the parent material. This (Fig. 2). The soil texture as well as the pH in soil solution corresponds well with the significant correlations found seemed to have an influence on the Cd concentration. between the Cd concentration and clay and ECEC (Ta- Lower pH values limit the accumulation of Cd in soil, ble 4). The high correlation between the soil Cd concen- and a significant correlation (p 0.0001, R 2 0.39) tration and the ECEC in the arable soils and the clay was found between the log Cd concentration and pH content in the forest soils suggest that a high density of in soil solution (Fig. 3). The pH was found not to have sorption sites is important to Cd sorption (Table 4). In any influence on the total Cd content in the soil, but a a sequential extraction study of Cd in Danish forest and significant influence on the Cd in solution, and thus on arable soils, Andersen et al. (2002) found that Cd was the solubility of Cd in soil. The general understanding bound only in easily mobile fractions of the soil, which has been that Cd solubility decreases at high pH and eliminates the possibility that the high correlations be- with high clay content that both favors a high sorption tween Cd and clay content is caused by Cd being oc- ´ ´ capacity (Sanchez-Martın and Sa ´nchez-Camazano, 1993; cluded in the clay minerals. Other authors have also Boekhold et al., 1993; Egli et al., 1999; Sterckeman et found that Cd concentration in soil correlates with al., 2000). The Cd solubility in this study decreased with ´ ´ ECEC and clay content (Sanchez-Martın and Sanchez- ´ increasing pH, but in contrast to the above findings an Camazano, 1993; Springob and Bottcher, 1998; Stercke- ¨ increasing Cd concentration in the soil solution was man et al., 2000). The soil pH was not found to be sig- found as the clay percentage increased. This may be nificant when included in the statistical analyses, which explained by the higher Cd content found in the loamy leads to the conclusion that the Cd concentration is less sand and sandy loam soils. dependent on the soil pH (Egli et al., 1999). The present study reports actual field concentrations of Cd in unpolluted soils and soil solutions from various Soil Solution soil types and horizons. The variable origin of Cd in soil The majority of the measured Cd concentrations in makes the correlation with soil parameters more com- soil solutions were below 1.0 g L 1. Similar levels of Cd ´ plex (Sanchez-Camazano et al., 1998; Sauve et al., 2000). ANDERSEN ET AL.: CONTENT, DISTRIBUTION, AND SOLUBILITY OF CADMIUM 1835 The significance of other soil variables than pH on the Holmgren, G.G.S., M.W. Meyer, R.L. Chaney, and R.B. Daniels. 1993. Cadmium, lead, zinc, copper, and nickel in agricultural soils Cd solubility has been reported with ECEC and organic of the United States of America. J. Environ. Qual. 22:335–348. matter content being the most important ones (Buchter Hovmand, M., and K. Kemp. 1998. Rep. no. 313. p. 1–28. In Tungmetal ´ et al., 1989; Sauve et al., 2000). nedfald i Danmark. (In Danish.) Danmarks Miljøundersøgelser, Cadmium mobility is increased by afforestation. De- Copenhagen. creasing soil solution pH increase Cd in solution, but International Organization for Standardization. 1995. Soil quality— Extraction of trace elements soluble in aqua regia. ISO 11466. ISO, high ECEC and content of clay and organic matter in Geneva, Switzerland. the soil can counteract this by binding Cd in the soil. Jug, A., F. Makeschin, K.E. Rehfuss, and C. Hofmann-Schielle. 1999. Leaching of soil solution with the concentration levels Short-rotation plantations of balsam poplars, aspen and willows of Cd found in this investigation of soil with background on former arable land in the Federal Republic of Germany. III. levels of Cd, does not pose a risk to the surrounding en- Soil ecological effects. For. Ecol. Manage. 121:85–99. Ma, L.Q., F. Tan, and W.G. Harris. 1997. Concentrations and distribu- vironment. tions of eleven metals in Florida soils. J. Environ. Qual. 26:769–775. ´ McBride, M.B., S. Sauve, and W. Hendershot. 1997. Solubility control ACKNOWLEDGMENTS of Cu, Zn, Cd and Pb in contaminated soils. Eur. J. Soil Sci. 48: 337–346. This research was partially supported by a grant from the Romkens, P.F.A.M., and W. Salomons. 1998. Cd, Cu, and Zn solubility ¨ European Commission (FAIR3-CT96-1983) and a grant from in arable and forest soils: Consequences of land use changes for the Danish Forest and Nature Agency. metal mobility and risk assessment. Soil Sci. 163:859–871. Sanchez-Camazano, M., M.J. Sanchez-Martin, and L.F. Lorenzo. 1998. REFERENCES Significance of soil properties for content and distribution of cad- mium and lead in natural calcareous soils. Sci. Total Environ. 218: Andersen, M.K., K. Raulund-Rasmussen, B.W. Strobel, and H.C.B. 217–226. Hansen. 2002. Heavy metal distribution and fractionation in pairs SAS Institute. 1999. SAS System. Version 8. SAS, Cary, NC. of arable and afforested soils. Eur. J. Soil Sci. 53(3):491–502. ´ Sauve, S., W. Hendershot, and H.E. Allen. 2000. Solid-solution par- Bak, J., J. Jensen, M.M. Larsen, and J. Scott-Fordsmand. 1997. A titioning of metals in contaminated soils: Dependence on pH to- heavy metal monitoring-programme in Denmark. Sci. Total Envi- tal metal burden, and organic matter. Environ. Sci. Technol. 34: ron. 207:179–186. 1125–1131. Bergkvist, B. 1987. Soil solution chemistry and metal budgets of spruce ´ ´ ´ Sanchez-Martın, M.J., and M. Sanchez-Camazano. 1993. Adsorption forest ecosystems in S. Sweden. Water Air Soil Pollut. 33:131–154. and mobility of cadmium in natural, uncultivated soils. J. Environ. Bergkvist, B., L. Folkeson, and D. Berggren. 1989. Fluxes of Cu, Zn, Qual. 22:737–742. Pb, Cd, Cr, and Ni in temperate forest ecosystems. Water Air Soil Soil Survey Staff.1997. Keys to soil taxonomy. Pocahontas Press, Pollut. 47:217–286. Blacksburg, VA. Boekhold, A.E., E.J.M. Temminghoff, and S.E.A.T.M. Van der Zee. Springob, G., and J. Bottcher. 1998. Parameterization and regionaliza- ¨ 1993. Influence of electrolyte composition and pH on cadmium tion of Cd sorption characteristics of sandy soils. I. Freundlich type sorption by an acid sandy soil. J. Soil Sci. 44:85–96. parameters. Z. Pflanzenernaehr. Bodenkd. 161:681–687. Buchter, B., B. Davidoff, M.C. Amacher, C. Hinz, I.K. Iskandar, and Sterckeman, T., F. Douay, N. Proix, and H. Fourrier. 2000. Vertical H.M. Selim. 1989. Correlation of Freundlich Kd and n retention distribution of Cd, Pb and Zn in soils near smelters in the North parameters with soils and elements. Soil Sci. 148:370–379. of France. Environ. Pollut. 107:377–389. Chlopecka, A., J.R. Bacon, M.J. Wilson, and J. Kay. 1996. Forms of Strobel, B.W., H.C.B. Hansen, O.K. Borggaard, M.K. Andersen, and cadmium, lead, and zinc in contaminated soils from southwest K. Raulund-Rasmussen. 2001. Cadmium and copper release kinet- Poland. J. Environ. Qual. 25:69–79. ics in relation to afforestation of cultivated soil. Geochim. Cosmo- Christensen, T.H. 1989. Cadmium soil sorption at low concentrations: chim Acta 65:1233–1242. VIII. Correlation with soil parameters. Water Air Soil Pollut. 44: Stuanes, A.O., G. Ogner, and M. Opem. 1984. Ammonium nitrate as 71–82. extractant for soil exchangeable cations, exchangeable acidity and Davies, B.E., and R.I. Davies. 1963. A simple centrifugation method aluminum. Commun. Soil Sci. Plant Anal. 15:773–778. for obtaining small samples of soil solution. Nature 198:216–217. ´ ´ ´ ´ ´ Szakova, J., P. Tlustos, J. Balık, D. Pavlıkova, and V. Vanek. 1999. Egli, M., P. Fitze, and M. Oswald. 1999. Changes in heavy metal The sequential analytical procedure as a tool for evaluation af As, contents in an acidic forest soil affected by depletion of soil organic Cd and Zn mobility in soil. Fresenius J. Anal. Chem. 363:594–595. matter within the time span 1969–93. Environ. Pollut. 105:367–379. Wilcke, W., C. Guscherker, J. Kobza, and W. Zech. 1999. Heavy metal Gee, G.W., and J.W. Bauder. 1986. Particle-size analysis. p. 383–411. concentrations, partitioning, and storage in Slovak forest and ar- In A. Klute (ed.) Methods of soil analysis. Part 1. 2nd ed. ASA able soils along a deposition gradient. J. Plant Nutr. Soil Sci. 162: and SSSA, Madison, WI. 223–229. Gooddy, D.C., P. Shand, D.G. Kinniburgh, and W.H. Van Riemsdijk. Wilkens, B.J., and J.P.G. Loch. 1997. Accumulation of cadmium and 1995. Field-based partition coefficients for trace elements in soil zinc from diffuse immission on acid sandy soils, as a function of solutions. Eur. J. Soil Sci. 46:265–285. soil composition. Water Air Soil Pollut. 96:1–16.