1

Evaluation of amendment addition and tree planting as measures to remediate 1contaminated soils: the Guadiamar case study (SW Spain)2

3

Madejón, Pa*; Domínguez, MT

a,b; Gil-Martínez, M

a; Navarro-Fernández, CM

a; Montiel-4

Rozas, MMa,c; Madejón, E

a; Murillo, JM

a; Cabrera, F

a; Marañón, T

a 5

6

aIRNAS, CSIC, Avenida Reina Mercedes 10, 41012 Sevilla, Spain 7

b Department of Crystallography, Mineralogy and Agricultural Chemistry, Edaphology 8Unit, University of Sevilla, Prof. García González St., 41012, Sevilla, Spain9

cCEBAS, CSIC, Campus Universitario de Espinardo, CP 30100 PO Box 164, Murcia, 10Spain 11

* Corresponding author: pmadejon@irnase.csic.es 12

13

ABSTRACT 14

Soil pollution is one of the main environmental concerns at a global scale, and different 15

measures have been proposed for its prevention and remediation. In this paper we 16

evaluate two measures – amendment addition and tree planting – applied to remediate 17

soils contaminated by trace elements (TE). The Guadiamar study site (SW Spain) is a 18

well-known example of a large-scale cleaning and remediation program intended to 19

rehabilitate about 3000 ha of soils affected by a mine-spill in 1998. Firstly, we present 20

the results of a long-term experiment in which two types of amendments – sugar beet 21

lime (SL) and biosolid compost (BC) – were added to the spill-affected soil. Three 22

different treatments were established: SL, BC, and NA (non-amended control). The 23

experiment has been running since 2002 and new soil samples were taken in 2016. In 24

general, the amendments increased soil pH and total organic carbon, and reduced TE 25

availability. The available TE concentrations (CaCl2 extraction) decreased drastically 26

with time in all cases, but the evolution differed among treatments. The treatment 27

Postprint of: Catena (166) 34-43 (2018)

2

effectiveness was evaluated by measuring plant biomass, TE concentrations in plant 28

shoots, and soil-plant transfer. Secondly, we present results detailing the effects of 29

seven tree species (three - Populus alba, Celtis australis, and Fraxinus angustifolia -30

were deciduous and four -Quercus ilex, Olea europaea, Ceratonia siliqua, and Pinus 31

pinea - evergreen) on remediated soils. In 2014 (about 15 years after planting) we32

measured TE concentrations in leaves and roots and related them with TE 33

concentrations (total and available) in soil. There were significant differences among the 34

studied tree species in the uptake and accumulation of TE in leaves and roots. Soil pH 35

differed among tree species and with respect to adjacent non-planted sites. There was a 36

negative exponential relationship between soil pH and availability of Cd, Cu, and Zn. 37

The bioconcentration factor (BCF, root:soil) and the translocation factor (TF, leaf:root) 38

were calculated to evaluate the potential of each tree species to stabilize TE in soil. 39

Finally, we propose the phytoremediation of TE-contaminated soils as a three-stage 40

process, including addition of soil amendments, planting of trees, and monitoring of TE 41

dynamics.42

Keywords: 43

Phytoremediation, , Trace elements, Aznalcóllar mine spill, Bioconcentration factor 44

45

1. Introduction 46

Soil pollution is one of the main environmental problems at a global scale. It 47

affects 2.5 million sites in the European Union alone, representing a management cost48

equivalent to 6 million € per year (Panagos et al., 2013; EEA, 2014). The losses49

associated with soil contamination are due not only to remediation expenses, but also to 50

its detrimental effects on soil health and fertility, which limit the suitability of 51

contaminated soils for certain land uses, such as agriculture. Given the environmental 52

3

and economic dimensions of this land degradation process, soil contamination has been 53

included as a priority in the EU environmental agenda (Virto et al., 2015), so that the 54

diagnosis and prevention of soil pollution was a priority task in the implementation of55

the European Soil Thematic Strategy (European Commission, 2012).56

Phytoremediation is an efficient, low-cost, and eco-friendly rehabilitation 57

strategy, which uses plants and their associated microorganisms to remove, detoxify, 58

and retain both organic and inorganic contaminants, especially those present in soil59

(Bolan et al., 2011). Moreover, phytoremediation promotes soil stabilization and carbon 60

sequestration, and might be used to generate added value from the degraded land 61

through biomass/biofuel production, if the selected plant species are able to produce 62

high biomass (Ali et al., 2013; Domínguez et al., 2017; Wang et al., 2017).63

Among the phytoremediation technologies, the phytostabilization approach 64

seems to be the most feasible option for the management of large areas contaminated by 65

trace elements (TE) (Mendez and Maier, 2008). This approach combines the use of soil 66

amendments and plants to immobilize pollutants in the soil, thus reducing the risks of 67

their transfer through the food web. The implementation of a phytostabilization program 68

can be considered a three-stage process whose final success will depend on: first, the 69

selection of soil amendments, second, the choice of the most suitable plant species for 70

afforestation, and third, the maintenance of a proper monitoring scheme to detect 71

changes in the mobility of TE in the system (Domínguez et al., 2008; Bolan et al., 72

2011). 73

As the first step in the phytostabilization process, soil amendments should be 74

applied to improve the physical, chemical, and biological properties of the degraded 75

soils, and to facilitate plant establishment. Inorganic amendments (mainly calcium 76

carbonate amendments) improve certain physical and chemical properties; for example, 77

4

by increasing pH and maintaining soluble metals at low levels (Janoš et al., 2010; 78

Pourrut et al., 2011). Organic amendments also improve physical and chemical 79

properties, as well as increasing nutritional status, water infiltration rates, water-holding 80

capacity, and soil pH, thereby indirectly affecting the adsorption and complexation of 81

metals in contaminated soils (Tordoff et al., 2000). In addition, the use of organic 82

amendments implies the enrichment of nutrients and carbon for microbial communities 83

that could assist in the establishment of plant cover. 84

In the next step, phytostabilization requires the introduction of native plants, 85

preferably with an extensive root system, which must not translocate high amounts of86

TE into their aerial tissues. These plants will help to stabilize metals in soils through 87

different mechanisms such as the precipitation of metals in less available forms, their 88

complexation with organic products, their sorption onto root surfaces, or accumulation89

inside root tissues (Mendez and Maier, 2008). Moreover, a successful amendment 90

application and plant establishment will lead to a diverse and/or more active microbial 91

community, which will improve soil quality and promote carbon sequestration in the 92

soil (Shrestha and Lal, 2006; Bastida et al., 2015; Montiel-Rozas et al., 2016a, b). 93

Finally, once the remediation measures have been applied, it is necessary to 94

carry out periodic monitoring and evaluation of the contaminant dynamics within the 95

system and of the risk of the TE being introduced into the food web; if necessary, new 96

remediation measures should be taken (Bolan et al., 2011). 97

Within the European RECARE project, the effectiveness of different measures 98

with regard to the remediation of degraded soils, affected by different threats, was99

evaluated in a series of land remediation experiments across Europe (see other articles 100

in this special issue). In this article we present the results obtained in two long-term 101

studies that have evaluated the effectiveness and limitations of two consecutive and102

5

complementary measures – amendment addition and tree planting – to remediate TE-103

contaminated soils. These measures have been evaluated in the Guadiamar Green 104

Corridor (SW Spain), an example of large-scale phytoremediation. 105

Specific objectives were: 1) to test the effects of two types of amendments on 106

the TE stabilization and soil quality improvement in this contaminated site, 14 years 107

after their first application; 2) to evaluate the potential of different tree species for TE 108

phytostabilization in these contaminated soils; and 3) based on these findings, to discuss 109

the effectiveness of applying both measures on a large scale (in the Guadiamar Green 110

Corridor) and the importance of monitoring TE dynamics in the system. For both 111

remediation measures, we studied three common aspects: i) changes in soil pH and TE 112

availability, ii) effects on soil organic carbon, and iii) transfer of TE between the soil 113

and plant (roots and leaves) systems. 114

115

2. Materials and methods 116

2. 1. Study area 117

The study sites and experimental plots were located inside the area affected by 118

the Aznalcóllar mine-spill, in SW Spain (Figure S1). In that accident (in April 1998) a 119

waste dam collapsed, releasing ca. 6 ×106m

3of slurry and contaminating about 4,000 120

ha with sludge rich in TE (Grimalt et al., 1999; Cabrera et al., 1999; Madejón et al.,121

2018). After the mine spill, the Andalusian Government expropriated the affected 122

private lands, ceasing their arable and livestock uses, to minimize the health risk. Then, 123

the affected soils were cleaned up and reclaimed, representing a large-scale case of 124

phytoremediation (Domínguez et al., 2008). Currently, the management of these lands is 125

focused on the conservation of biodiversity and the establishment of an ecological 126

6

corridor connecting the Doñana National Park and the Sierra Morena Mountains, named 127

“Protected Landscape of the Guadiamar Green Corridor” and occupying about 2,700 ha. 128

The climate is typically Mediterranean, with mild, rainy winters (about 500 mm 129

mean annual rainfall) and hot, dry summers. The mean annual daily temperature is 17130

ºC, with a maximum of 33.5 ºC in July and a minimum of 5.2 ºC in January131

(Domínguez et al., 2008).132

Geologically, the Guadiamar Green Corridor (GGC) is located on the south-133

eastern edge of the Iberian Pyrite Belt (IPB), which constitutes the largest and most 134

important volcanogenic massive sulfide province in Western Europe. This mine belt 135

extends 200 km, from south-western Portugal to the West of Seville, in Spain. The136

Aznalcóllar mine complex is situated on the northern edge of the Guadalquivir Tertiary 137

basin, where transgressive Miocene sediments cover Paleozoic materials (more 138

information in Gallart et al., 1999; López-Pamo et al., 1999).139

The soils in the study area correspond to the Mediterranean edaphic zone, being 140

very heterogeneous due to a highly variable lithology and geomorphology. The alluvial 141

soils are calcareous and non-calcareous Typic and Aquic Xerofluvents, with sandy and 142

sandy-loam textures. The soils of the lower terraces are Typic and Aquic Haploxeralfs 143

and Aquic Xerofluvents, while in the higher terraces Typic Rhodoxeralf soils associated 144

with sandy soils and Pseudogley are also present. Between the terrace and alluvial soils 145

there are eroded soils classified as Calcixerollic Xerochrepts (soil taxonomy according146

to Soil Survey Staff, 1994, see Cabrera et al., 1999; Nagel et al., 2003). 147

148

2.2. Experimental design 149

7

2.2.1. Measure 1 - Amendment addition 150

Two types of amendments - one organic and another inorganic - were tested. 151

The organic amendment was biosolid compost (BC) and the inorganic amendment was 152

sugar beet lime (SL), a residual material from sugar beet processing with 70-80% (dry 153

basis) CaCO3. The most relevant characteristics of these amendments are shown in 154

Madejón et al. (2009).155

The experimental site was located at El Vicario (37º26′16.73′′ N, 6º13′3.97′′ W), 156

inside the GGC. The complete design consisted of eight plots of 7 m × 8 m each, with a 157

space 1 m long and 2 m wide between plots (Figure S2). The experiment was set up in 158

October 2002. Three treatments were established: SL (30 t ha−1

yr−1), BC (30 t ha

−1159

yr−1), and NA (non-amended control). The application rates were comparable to those 160

applied to the whole of the spill-affected area, the GGC (Antón-Pacheco et al., 2001). 161

The amendments were applied in two consecutive years (October 2002 and 2003) and 162

incorporated into the top 15 cm of the soil using a rotary tiller (RL328, Honda). The 163

non-amended subplots were tilled in an identical manner. The experiment had a 164

completely randomized block design with three replicates per treatment.165

Soil samples were taken at 0-15 cm depth from each plot. For each plot three 166

soil cores were taken (using a spiral auger of 2.5 cm diameter) to make a composite 167

sample representative of each plot. The soils were sampled in autumn 2002 (before 168

amendment addition), 2007, and 2011 (see previous results in Xiong et al., 2015). A 169

new sampling was performed in 2016 to explore the long-term trends.170

Plant surveys and samplings in the experimental plot were carried out in the 171

spring of 2007 and 2011 (Pérez de Mora et al., 2011; Xiong et al., 2015). The three 172

8

most dominant plant species were Poa annua, Lamarckia aurea, and Cynodon dactylon. 173

Shoots were harvested for chemical analysis (see methods below).174

175

2.2.2. Measure 2 - Tree planting 176

After the contaminated soil had been cleaned-up and remediated, the whole 177

GGC was planted with native shrub and tree species (Domínguez et al., 2008, 2010). In 178

fact, we evaluated the combination of amendments and tree planting. We tested the 179

effects of planting different tree species (considered as “treatments”) on the availability 180

of TE in the soil, and compared the results with soils from “open”, non-planted sites. 181

This study was conducted in a site of about 14 ha inside the GGC (37° 23' 10′′182

N, 6° 13' 40′′ W), where we randomly selected five replicates of each of the seven tree 183

species afforested in the plot (35 trees in total) (Figure S3). Three species were 184

deciduous (Populus alba, Celtis australis, and Fraxinus angustifolia) and four were 185

evergreen (Quercus ilex, Olea europaea, Ceratonia siliqua, and Pinus pinea) (hereafter 186

we use the genus name for simplicity). In autumn 2014, we sampled fully expanded 187

leaves at three points in the outer canopy; these were mixed to make one composite 188

sample per tree. We collected the forest floor (litter) biomass inside three 25x25 cm 189

quadrats around each tree and mixed it to make one composite sample per tree. We 190

excavated roots at several points around the tree trunk and made one composite sample 191

per tree. We sampled the soil at two depths (0-10 cm and 10-30 cm), using a spiral 192

auger of 2.5 cm diameter, at three points under the tree canopy and mixed the samples193

to make one composite sample per tree and depth. The samples of forest floor, roots,194

and soil were taken in late spring 2014. 195

196

9

2.4. Laboratory analyses 197

The soil samples were initially air-dried, crushed, and sieved (<2 mm). Prior to 198

the determination of pseudototal TE concentrations, the air-dried soil samples were 199

ground to <60 μm. Soil pH was determined, using a CRISON micro pH 2002 probe and 200

meter, in a 1/2.5 sample/1 M KCl suspension. The total organic carbon (TOC) in soil 201

was determined following dichromate oxidation of samples and titration with ferrous 202

ammonium sulfate (Walkley and Black, 1934).203

Soil CaCl2-extractable TE concentrations were determined in 0.01 M CaCl2 at a 204

1:10 soil/solution ratio, after shaking for 3h (Houba et al., 2000). For analysis of the 205

pseudototal TE concentrations in the soil and amendments, samples (<60 μm) were 206

digested with aqua regia (1:3 conc. HNO3:HCl) in a microwave oven (Microwave 207

Laboratory Station Mileston ETHOS 900, Milestone s.r.l., Sorisole, Italy). 208

Plant material was washed for 15 s with 0.1 M HCl, followed by a 10-s washing 209

with distilled water, and then oven-dried at 70 ºC. Dried plant material was ground and 210

passed through a 500-μm stainless steel sieve. Plant TE were extracted by wet oxidation 211

with concentrated HNO3 in a microwave digester.212

The TE concentrations in the plant and soil digests and soil CaCl2 extracts were 213

measured by inductively coupled plasma optical emission spectrometry (ICP-OES) with 214

a Varian ICP 720-ES (simultaneous ICP-OES with axially viewed plasma).215

The quality of the analyses was assessed using reference samples. For soil we 216

used sample ERM-CC141 (loam soil) and obtained recoveries from 95% to 101%. For 217

plant analysis we used the reference INCT-ONTL-5 (tobacco leaves) and obtained 218

recovery rates between 90 and 103%.219

10

220

2.5 Data analysis 221

The mean and standard error (SE) were determined for all variables and 222

treatments. Previously, the data were tested for normality (Shapiro-Wilk test) and 223

homoscedasticity (Levene test) and log-transformed if necessary. The effects of the 224

treatments (amendment type or tree species) on soil parameters (TE concentrations, pH,225

and TOC) were analyzed by one-way analysis of variance (ANOVA). Post hoc analyses 226

were based on Tukey’s test (or Dunnett’s T3 test if the variances were unequal) and the 227

LSD test for soil pH under trees. The significance level used was 0.05. The bivariate 228

relationships between pH and TE availability were fitted to linear or non-linear 229

functions. All the analyses were carried out with SPSS 20.0 software for Windows 230

(SPSS Inc., USA).231

To evaluate the potential for phytoremediation it is important to know the 232

relationships between the concentrations of TE in soil and those in plants. There are 233

many indexes proposed in the literature, with different names and ways to calculate 234

them (see review in Buscaroli, 2017). We used here the “bioconcentration factor” 235

(BCF), calculated as the ratio between the TE concentrations in plant shoots and the 236

concentrations (pseudototal) in soil, for the amendment experiment. In the case of the 237

tree plantations we distinguished two subindexes: BCFroot, obtained by dividing the TE 238

concentrations in roots by the pseudototal TE concentrations in soil and BCFleaf, 239

obtained by dividing the leaf concentrations by the soil concentrations. In addition, we 240

calculated the “translocation factor” (TF) as the ratio between the TE concentrations in 241

the tree leaves and those in the roots.242

243

11

3. Results and discussion 244

3.1. Changes in pH and trace element availability 245

The main target in the stabilization of TE contaminated soils is to increase pH 246

and thereby reduce the availability of cationic TE (Adriano, 2001). Next, we present the 247

results and discuss the effects of two remediation measures - amendments and tree 248

planting -on soil pH and TE availability.249

250

3.1.1. Effectiveness of amendment addition 251

The soil of the Vicario experimental site at the initial stage was very acid (pH 252

around 3.5, see pH values of NA plots, Figure 1); therefore, an increase in soil pH was a 253

key target for its remediation. The addition of the amendments increased pH 254

significantly and their effects lasted with time. Sugar beet lime (SL) raised soil pH to 255

close to neutral, and those values were maintained for at least 12 years without any 256

further amendment application (Figure 1). In fact, there were no significant differences 257

among the values obtained in 2007, 2011, and 2016 for this treatment. The organic 258

amendment (BC) also increased soil pH and maintained similar values over time (no 259

significant differences in the values among the different samplings were observed), but 260

it was less effective than SL for soil alkalization (Figure 1). In summary, the pH 261

increment from 2002 to 2016 due to SL was 111% and for BC it was 43%.262

The pseudototal TE concentrations at the beginning of the study (in 2002) were: 263

As 202, Cd 4.4, Cu 119, Pb 471, and Zn 381 (mean values in mg kg-1). These values 264

were much higher than the background values in the area (Table S1; see also 265

pseudototal values in 2016 for comparison). Therefore, due to their strong acidity and 266

12

high concentrations of total TE, the availability of TE and the potential toxicity risk of 267

these soils were expected to be high.268

The effectiveness of the amendments was thus evaluated by the soil pH 269

increment (as mentioned above) and consequent decrease in TE availability. For all 270

treatments, the available concentrations of As and Pb (CaCl2-extractable) were below 271

the detection limits (0.01 mg L−1, equivalent to 0.1 mg kg

−1), reflecting the low 272

solubility of these elements. The available concentrations (CaCl2-extractable) of Cd, Cu,273

and Zn in the amended soils showed a strong, significant decrease compared to those in 274

the unamended plots (NA, Figure 2). In accordance with the results obtained for soil 275

pH, the reduction of TE availability (from 2002 to 2016) was stronger for the SL (a 276

reduction of 99% in the Cd, Cu, and Zn concentrations) than for the BC treatment (a 277

reduction of 73% for Cd, 98% for Cu, and 85% for Zn) (Figure 2). The soils treated 278

with SL became more alkaline, so the cationic TE would have precipitated, forming 279

complexes and secondary minerals (Mench et al., 1994). In the case of the organic 280

amendment (BC), the soils also became more alkaline, although to a lesser extent;281

therefore, TE availability was intermediate (Figure 2). The liming effect of organic 282

amendments can be related to proton consumption during decarboxylation of organic 283

acid anions as plant residues decompose, with proton consumption by functional groups 284

associated with organic material, or to specific adsorption of organic molecules by 285

ligand exchange with the release of OH− ions (Madejón et al., 2010). Moreover, organic 286

matter forms insoluble and stable complexes with TE and also absorbs TE by the 287

formation of organic colloids, both processes contributing to the reduction of TE 288

availability in soil; this may partially explain the effects of the organic amendments on 289

TE solubility.290

13

The effectiveness of soil amendments, in the short-term, has been widely 291

demonstrated (Pérez de Mora et al., 2006; Ciadamidaro et al., 2016). However, less 292

information is available concerning the stability and longevity of these remediation 293

treatments, especially when organic amendments are applied. These new data (14 years 294

after the first addition) support the positive and long-term effect of BC on soil 295

remediation. Although the effectiveness of the amendments in the reduction of toxicity 296

and contamination is undoubted, it is also interesting to note the moderate decrease in 297

the available TE concentrations with time in the NA plots, a sign of natural remediation 298

in these soils (Figure 2).299

300

3.1.2. Effectiveness of tree planting 301

Trees were planted in 1999-2000, after sludge removal and the addition of 302

amendments. Fourteen years after tree plantation, in 2014, the mean topsoil pH value at 303

the afforested experimental plot was 4.18, ranging from 2.63 to 6.05. A significant 304

difference was found among studied plant species in soil pH underneath the trees305

(ANOVA F=2.61 and p=0.03). Soil acidity was strongest (pH was lowest) under Pinus306

and Quercus species. By contrast, under Ceratonia, Fraxinus, and Populus species soil 307

pH was higher. The “open”, non-planted sites (with herbaceous cover and no trees), as 308

well as Celtis and Olea species, showed intermediate soil pH values (Figure 3).309

Relatively high pseudototal TE values in the superficial soil (0-10 cm, Table S1) 310

indicate the persistence of the residual contamination. Overall values (grouping all 311

samples) for As ranged between 65 and 317 with a mean of 145, mean Cd was 0.80 312

with a maximum of 1.30, mean Cu was 191 with a maximum of 253, and for Zn the 313

mean was 243 with a maximum of 364 (values expressed in mg kg-1). There was no 314

14

significant difference among the studied tree species or between them and the open 315

sites, indicating a general TE contamination.316

The As and Pb availability in the soil underneath the trees was below the 317

detection limits (see detection levels in section 3.1.1.). The higher soil pH values under 318

some tree species gave an exponential reduction of Cd, Cu, and Zn availability (Figure 319

4). Therefore, the trees of some species (like Ceratonia, Fraxinus, and Populus) were 320

effective at increasing pH and reducing TE availability in the soil underneath them. In 321

contrast, other tree species (Pinus and Quercus) had lower soil pH in comparison with 322

the open sites and therefore were less effective at reducing TE availability. Lastly,323

Celtis and Olea species showed a neutral effect on soil pH and TE availability. 324

Trees are ecosystem engineers, modifying soil physico-chemical properties, 325

nutrient cycling, and microbial communities (Aponte et al., 2013). In the soil 326

underneath the trees the net result of proton (H+) release or consumption will determine 327

the increase or decrease in soil acidification. That change in soil pH is associated with 328

tree species traits like root respiration, exudation of organic acids, the quantity and 329

chemistry of litterfall, and fine roots (Russell et al., 2017). Because a rise in soil pH 330

enhances the binding of TE and hence reduces their bioavailability (Domínguez et al., 331

2009; Bolan et al., 2011), those tree species inducing soil alkalization will be more 332

effective in the phytostabilization of cationic TE-contaminated land.333

334

3.2. Soil organic carbon 335

Soil organic matter contributed to stabilization of TE, decreasing both their 336

release from the soil into the environment and their bioavailability (Bolan et al., 2011). 337

However, soil carbon loss is one of the main threats to soil quality at a global scale 338

15

(Sanderman et al., 2017). Therefore, in addition to soil contamination, we are concerned 339

with the impact of the two remediation measures on soil carbon loss, as a target threat.340

We evaluated the effectiveness of amendments addition and tree plantations with 341

respect to increasing soil organic carbon, and hence remediating soils. Both remediation 342

measures – amendments (mainly BC) and tree plantations – strongly increased soil 343

carbon, which has associated additional benefits like the mitigation of climate change 344

(C sequestration) and the improvement of soil communities (more active and diverse)345

(see below). 346

347

3.2.1. Effectiveness of amendments addition 348

A deficient organic carbon content is generally considered a major factor 349

inducing desertification in Mediterranean soils; 2% being the threshold for the pre-350

desertification stage (COM, 2002). The initial values of the TOC content in the 351

experimental site were around 1% (<10 g kg-1 in 2002, Figure 5), indicating a high level 352

of soil degradation (partly due to the cleaning and removal of the top soil). The 353

application of both soil amendments increased the TOC significantly, compared to the 354

non-amended plots (Figure 5). In the first period (2002 to 2007), the soil TOC increased 355

rapidly, especially with the BC treatment. In the longer-term (2011), the BC effects on 356

the soil TOC decreased slightly (by 16% from the 2007 value), but from 2011 to 2016 357

the value was maintained (Figure 5). Although the increase of the soil organic carbon358

content was rapid and strong for the BC (as expected), a clear effect was also found for 359

the inorganic SL. In these soils (SL-treated), there was a gradual increase in soil TOC, 360

from 12.6 to 16.4 g kg-1 in nine years. This increase (although less marked than with361

BC) may not be directly related to the chemical effect but may be a consequence of the 362

16

increasing vegetation cover caused by this amendment; thus, the accumulated plant 363

biomass in SL plots was much higher than in NA plots (Madejón et al., 2009).364

This increase in the soil organic matter was accompanied by the improvement of 365

soil fertility and the enhancement of the vegetation cover, as indicated before. A well-366

developed vegetation cover may reduce wind and water erosion and also the deposition 367

of contaminated dust, therefore decreasing the potential intake of TE by biota (Bargagli, 368

1998). In the amended plots, a sort of “assisted natural remediation” developed, the369

greater vegetation cover indicating the improvement in soil quality over time. On the 370

other hand, the sparse vegetation established in the NA plots reflected a “natural 371

remediation” of the soil.372

In general, amendment additions affect soil microbial communities (Marschner 373

et al., 2003). Their improvement of the soil chemical conditions entails a reduction of 374

the environmental filtering by high TE concentrations. Thus, in soils with elevated TE 375

concentrations, microbial communities will be composed of stress-tolerant organisms376

(Op de Beeck et al., 2015). In the amended plots, the reduction of the contamination 377

affected the community structure of the arbuscular mycorrhizal fungal community,378

increasing the phylogenetic diversity - which reached its maximum value with the BC379

treatment (Montiel-Rozas et al., 2016b). In addition, the application of organic 380

amendments favoured carbon sequestration in a degraded soil and improved its physical 381

and chemical conditions (Montiel-Rozas et al., 2016a).382

383

3.2.2. Effectiveness of tree planting 384

There was a trend for the studied soils to be richer in total organic carbon under 385

the different tree species (with exception of Pinus) than in the open, non-planted sites 386

17

(Figure 6). However, this difference was not statistically significant (ANOVA F=1.51, 387

p=0.20), due to high interspecific variation. The average TOC at the soil surface for all 388

samples (15.7 g kg-1) represented an increment of 55% compared to the deeper soil (10-389

30 cm); this TOC enrichment being greater under Ceratonia trees (mean of 83%) than 390

in open sites (mean of 56%) (Figure 6). 391

Despite the strong differences in the litter accumulation and the quality (C:N) of 392

the organic matter among the tree species (Table S2), their effect on TOC in the topsoil 393

was weak (Figure 6). Probably, the time elapsed (16 years) was not enough to detect a 394

significant and strong tree signal in the soil. In a review of 70 studies, Li et al. (2012) 395

found that the tree age was a major factor explaining the dynamics of soil C 396

(significantly increased after 30 years) and N (significant after 50 years). 397

Trees have a strong effect on nutrient cycling, inducing feedback processes and 398

affecting soil biodiversity (Aponte et al., 2013). Soil carbon was a major factor 399

determining the composition of fungal communities associated with Quercus trees 400

planted in the GGC (López-García et al., 2018).401

402

3.3. Transfer of trace elements from soil to plant 403

The “plant factor” is crucial in determining the success of phytostabilization 404

technology (Bolan et al., 2011). Plant traits influence biomass production, the405

development of root systems, the transfer of TE from soil to plant tissues (with the 406

contrasting strategies of accumulators and excluders), and their translocation from roots407

to shoots. Hence, the soil-plant relationships were explored for both remediation 408

measures -amendment addition and tree plantations. 409

18

410

3.3.1 Effects of amendments 411

In general, herbaceous plants growing in amended soils accumulated lower TE 412

concentrations than those in NA soils (except in some cases for BC; data for 2011 and 413

2015 in Table S3). In amended soils, the TE concentrations were in the normal range for 414

plants (Chaney, 1989), excepting As in Lamarckia and Poa species. The extreme As 415

concentrations for Lamarckia in NA plots reached phytotoxic levels (Table S3; 416

according to Chaney, 1989;, Pérez de Mora et al., 2011; Xiong et al., 2015). 417

The BCF values calculated for the TE in the three grass species were relatively 418

low (Table 1). In general, the addition of amendments reduced these values, compared419

to the control (NA), supporting the effectiveness of these amendments in the 420

phytostabilization process; for example, the BCF of Cd in Poa decreased by about 26% 421

for SL and 48% for BC compared with the NA values. There were significant 422

differences among elements in the transfer from soil to plant. Thus, Cd and Zn are more 423

mobile elements and, therefore, showed higher BCF values; by contrast, the transfer of 424

As and Pb, less mobile elements, was very low (Table 1).425

Amendment addition to the contaminated soils reduced TE transfer into plants 426

(due to the reduction of TE availability in soils, Figure 2); consequently, the plants had 427

limited accumulation of TE in their aerial tissues (Table S3) and therefore represented a 428

low risk for the food web. In addition, successive extractions of TE by plants each 429

growing season (if they are harvested or grazed) diminish their availability with time430

(McGrath et al., 2002).431

432

19

3.3.2. Effects of tree planting 433

In general, the BCF values differed among tree species, organs, and elements 434

(Table 1). The values were higher for roots (overall mean of 0.58) than for leaves 435

(0.24). Among the studied TE, Cd showed the highest BCF values (BCFroot>1) for the 436

roots of all species. Among the tree species, Populus accumulated the highest437

concentrations of Cd and Zn in leaves (BCFleaf >1).438

The translocation factors (TF) also differed among tree species and elements 439

(Figure S4 and Table S4). The TF values were relatively low (below 0.5) for Pb, Cu,440

and As, for all species, while they were above one for Cd and Zn in some species 441

(maximum TF of 5.8 for Cd and 9.1 for Zn, both in Populus trees). 442

A general criterion in the selection of the plants best suited for the 443

phytostabilization of TE-contaminated soils is to choose tree species that keep the TF as 444

low as possible, having TF values <1 (Mendez and Maier, 2008; Bolan et al., 2011). In 445

this study area, most of the species fulfilled that criterion and were suited to 446

phytostabilization, with the exception of Populus species - which accumulated high 447

concentrations of Cd (TF 2.16) and Zn in leaves. For example, the tree species with the 448

lowest TF for Cd were Fraxinus (0.02), Olea (0.05), and Pinus (0.05).449

A second criterion is that the TE taken up from soils by the plants should be 450

retained preferentially in the roots. All the tree species had relatively high transfer of Cd 451

from soil to roots (BCFroots>1), and also of Cu and Zn (although BCFroots<1); however, 452

the BCFroots values were low for As and Pb. Among the tree species, Pinus had the 453

highest BCFroots for Cd, and Celtis the highest for Cu and Zn (Table 1).454

The planting of trees can be a very effective measure for the remediation of TE-455

contaminated soils (Pulford and Dickinson, 2005). However, not all tree species have 456

20

the same capacity to stabilize TE in the soil. Several functional traits should be 457

considered in the selection of tree species: they must be tolerant of high TE 458

concentrations in soil, produce a high amount of biomass, have a large and deep root 459

system, extract TE from soil and retain them in the roots and wood, and translocate a 460

very low fraction to the leaves and fruits, to minimize the risk of toxicity in the food 461

web (Mendez and Maier, 2008; Bolan et al., 2011; Madejón et al., 2017).462

463

3.4. Implementation of a phytoremediation process 464

The phytoremediation scheme in the Guadiamar study site comprised several 465

stages, which can be extrapolated to other cases of TE contamination. As an initial and 466

crucial measure, the contamination source must be stopped. In the Guadiamar case, 467

mine activities ceased with important social and economic impacts. The broken dam 468

was filled and sealed; the sludge had been cleaned up and the most-contaminated top 469

layer of the soil removed (being dumped into the open pit of the mine). All these 470

operations had a huge economical cost; about 58 million euros (Arenas et al., 2005).471

Within the phytoremediation process, the first stage was to immobilize in the 472

soil, as much as possible, the potentially toxic elements (As, Pb, Cd and Tl) by the 473

addition of amendments. The second stage was planting trees and shrubs for the long-474

term stabilization of TE and soil enrichment with organic matter. The third stage 475

involved the monitoring of TE dynamics in the system and the proposal of new 476

remediation measures if needed (Figure 7). 477

The use of soil amendments brought several benefits during the remediation 478

process. Inorganic amendments rich in carbonates (such as sugar beet lime) were most 479

effective at increasing pH and inducing the immobilization of cationic TE (Pérez de 480

21

Mora et al., 2011). On the other hand, the amendments rich in organic matter (like 481

biosolid compost), besides increasing the soil pH and decreasing the availability of TE, 482

compensated the loss of organic matter that could lead to desertification under semiarid 483

climate. A combination of both types of amendments was applied at large-scale in the 484

study site (Madejón et al., 2018). Amendment addition also enhanced the soil 485

community structure, by reducing the toxicity of TE and increasing the C sequestration 486

(Pérez de Mora et al., 2011; Montiel-Rozas et al., 2016a, b). In addition, most of these 487

amendments are generated from industrial or urban residues, with low costs, although488

these costs could be increased by the transportation from their generation point and by 489

application at the contaminated site; in any case, their use has a social benefit because it 490

contributes to the circular economy (European Commission, 2015). 491

A disadvantage associated with lime-based amendments is the gradual 492

reduction of its effect over the time due to dissolution and leaching, an effect which can 493

be accelerated by acid deposition (Ruttens et al., 2010). In the case of organic 494

amendments, the medium- and long-term evolution of the low-solubility compounds 495

formed between organic matter and TE is of utmost importance. The solubility of these 496

compounds could decrease with time, favoring TE stabilization; however, the opposite 497

process can also take place, leading to an increase in TE solubility over time as the 498

added organic matter is mineralized thus acting as a ‘chemical time bomb’ (Madejón et 499

al., 2010). Moreover, it is necessary to verify the adequacy and the effect of different 500

types of amendments, appropriate application doses, and the persistence of amendment 501

effects over time (Pérez de Mora et al., 2007; Madejón et al., 2010) and at each study 502

case. The duration of the amendments also depend on land use and climatological 503

conditions (Mench et al., 2003). For this reason, the unknown temporary character of 504

22

the amendment makes necessary the monitoring of TE in soil and wild vegetation to 505

ensure their stabilization and evaluate the need for further applications.506

Tree saplings (1-2 years old) were planted with plastic protectors on cleaned and 507

amended soils, on favourable season (autumn-winter). During the first years there was 508

some mortality of saplings, mainly due to high irradiance and drought stress during the 509

warm and dry Mediterranean summer (Domínguez et al., 2010). In these adverse 510

conditions, shrubs (natural or previously planted) can be used as nurse to facilitate the 511

stablishment of tree saplings (Domínguez et al., 2015). The restoration and plantation of 512

more than 40 woody species, at a density of 700-900 plants/ha, across the 2707 ha 513

costed about 22 million euros. This expense was added to the cost of purchasing the 514

spill-affected land (about 66 million euros) to avoid contamination of the food (Arenas 515

et al., 2005). In other cases, when this high amount of public money cannot be afforded, 516

farmers may grow alternative crops (not to be consumed) like fiber (cotton), biofuel, 517

pharmaceutical or ornamental in the contaminated and remediated land (Antoniadis et 518

al., 2017). Indeed, the use of plants growing at the GGC for energy production through 519

combustion and gasification of biomass has been recently evaluated (Dominguez et al., 520

2017), showing that this alternative use of vegetation for biofuel production could pose 521

an added value for land managers.522

The choice of the tree species is crucial for the phytostabilization success. For 523

some tree species (Fraxinus, Ceratonia and Populus) soil pH increased underneath the 524

trees, and thus reduced the soil availability of Cd, Cu and Zn. In contrast, other species 525

(Pinus and Quercus) tended to acidify the soil and thus increased the mobility of TE in 526

the ecosystem; therefore, they are less suitable for TE phytostabilization (Marañón et 527

al., 2015). 528

23

All the studied tree species had high uptake and retention of Cd in their roots, 529

and all (but Populus) had low translocation from roots to leaves, thus suiting the 530

criterion for phytostabilization. However, the particular case of Populus and its 531

accumulation of TE in aboveground organs (leaves, fruits, seeds) may represent a 532

toxicity risk for the food web; therefore such TE accumulator species are less suitable 533

for phytostabilization and should be monitored carefully (van Nevel et al., 2011). The 534

evaluation of the tree effects on ecosystem properties (as part of the third stage, see 535

below) may suggest to change the species composition with time, following an adaptive 536

management (Chazdon, 2008).537

Planted trees are also effective against the loss of soil organic matter and 538

contribute strongly to the sequestration of carbon in the soil, thus mitigating climate 539

change (Pan et al., 2011). Most of the species studied here tended to increase the soil 540

organic carbon underneath the trees, compared with adjacent open, non-afforested sites. 541

However, the effectiveness (tree signal) was weak because soil carbon sequestration is a 542

slow process, with significant effects not occurring until more than 30 years after 543

planting (Li et al., 2012).544

Tree plantations contribute to coping with other soil threats such as biodiversity 545

loss. The patches of soil under planted trees are “islands of fertility” that increase soil 546

biodiversity (Manning et al., 2006). Soil carbon was a key factor determining the 547

composition of ectomycorrhizal fungal communities (López-García et al., 2018). Those 548

fungal associations can contribute to TE sequestration in the rhizosphere, and to 549

alleviate the TE stress in the trees, thus representing a microbial-assisted 550

phytoremediation (Sarwar et al., 2017). 551

24

The proposed third stage - monitoring and evaluation - is crucial to achieve the 552

successful phytoremediation of TE-contaminated soils in the long-term (Bolan et al., 553

2011). The existence of residual contamination in Guadiamar river banks (e.g., 554

important levels of As and Pb), despite the remediation operations, impose that those 555

sites should be identified in detail, mapped, and subjected to a new clean-up and 556

addition of amendments (Domínguez et al., 2016). River banks are especially sensitive 557

sites; they need to be protected with vegetation to prevent water erosion, reduce the 558

flood risk and improve the water quality (Madejón et al., 2017). Since TE remain in the 559

affected area, although with reduced mobility, the regular and long-term monitoring of 560

the status of these potentially toxic elements in the ecosystem (water, soil, plants and 561

animals) is obligatory (Bolan et al., 2011; Figure 7).562

For the application of this phytoremediation process to other contaminated areas 563

some challenges and limitations need to be considered (see reviews by Bolan et al., 564

2011; Mahar et al., 2016; Antoniadis et al., 2017). In general, phytostabilization is a 565

simple and cost-effective approach, although in the Guadiamar case the expropriation of 566

the contaminated lands to change their use made it expensive. When the contamination 567

level is very high, the growth and survival of plants can be impaired; then new 568

amendment addition should be applied to further reduce TE availability, and more 569

tolerant species should be planted. The need for repeated amendment additions and 570

plants of specific tolerant species (not necessarily available at commercial nurseries) 571

might pose a higher cost for the phytostabilization of heavily contaminated lands. As 572

phytostabilization is a very slow process and the TE status in the soil-plant system is 573

dynamic, this management approach requires a long-term monitoring. Some tree species 574

can have long-term adverse effects, such as reducing soil pH and mobilizing TE. Other 575

tree species can accumulate TE in leaves and fruits posing a risk for the food web. Tree 576

25

responses in multielement contamination sites may be specific, requiring research on 577

their tolerance to contamination, and on their effects on ecosystem properties. The 578

adaptive management of the contaminated site is a learning process based on the 579

continuous evaluation of the results and the modification of the applied measures 580

according to monitoring outcomes.581

582

4. Conclusions 583

Soil contamination by TE is a major environmental and health concern. A three-584

stage phytoremediation approach is proposed based on the experience acquired during 585

the restoration of the Guadiamar Green Corridor (Spain). First stage is the addition of 586

soil amendments. Inorganic amendments rich in lime are very effective at increasing 587

soil pH and reducing TE availability (mobility); while organic amendments are 588

particularly effective at compensating the loss of soil organic matter. Second stage is 589

tree planting, to stabilize TE in soil, mainly through their retention in the roots. 590

Different tree species differ in their suitability for soil remediation; those increasing soil 591

pH, accumulating TE in the rhizosphere, and having low TE translocation from roots to 592

shoots should be selected. Third stage is the long-term monitoring of TE availability in 593

soils and their concentrations in organisms (flora and fauna). The continuous evaluation 594

of the results will modify the application of the remediation measures in an adaptive 595

management cycle. 596

597

Acknowledgements 598

26

We have received support from the European Union Seventh Framework 599

Programme (FP7/2007-2013) under grant agreement n° 603498 (RECARE) and from 600

Spanish Ministry of Economy and Competitiveness under RESTECO (CGL2014-601

52858-R) and BIORESMED (AGL2014-55717-R) projects. We thank the staff of 602

Guadiamar Green Corridor for facilities and support to carry out the different studies, to 603

the Group SoilPlant of the IRNAS, CSIC, especially J. M. Alegre for their technical 604

assistance, and the IRNAS´Analysis Service for chemical analyses of soil and plant 605

samples. MTD thanks University of Sevilla for her postdoctoral fellowship (V Plan 606

Propio).607

608

References 609

Adriano, D.C., 2001. Trace Elements in Terrestrial Environments: Biogeochemistry, 610

Bioavailability and Risks of Metals, second ed., Springer, New York, 2001611

Ali, H., Khan, E., Sajad, M.A., 2013. Phytoremediation of heavy metals-concepts and 612

applications. Chemosphere 91, 869-881.613

Antón-Pacheco, C., Arranz, J.C., Barettino, D., Carrero, G., Jiménez, M., Gómez, J.A., 614

Gumiel, J.C., López-Pamo, E., Martín Rubí, J.A., Martínez Pledel, B., De Miguel, 615

E., Moreno, J., Ortiz, G., Rejas, J.G., Silgado, A., Vázquez, E., 2001. Actuaciones 616

para el reconocimiento y retirada de los lodos depositados sobre el terreno, y su 617

restauración edáfica y morfológica. Boletín Geológico y Minero, 112 (Número 618

Especial), 93-121 (in Spanish). 619

Antoniadis, V., Levizou, E., Shaheen, S. M., Ok, Y. S., Sebastian, A., Baum, C., Prasad, 620

M.N.V., Wenzel, W.W., Rinklebe, J. 2017. Trace elements in the soil-plant 621

27

interface: Phytoavailability, translocation, and phytoremediation–A review. Earth-622

Science Reviews, 171, 621-645.623

Aponte, C., García, L.V., Marañón, T., 2013. Tree species effects on nutrient cycling 624

and soil biota: A feedback mechanism favouring species coexistence. For. Ecol. 625

Manage. 309, 36-46.626

Arenas, J.M., Carrascal, F., Gil, A., Montes, C. 2005. Breve historia de la construcción 627

del Corredor Verde del Guadiamar. In: C. Montes, F. Carrascal (eds.), La 628

restauración ecológica del Río Guadiamar y el proyecto del Corredor Verde, pp. 629

29-64. Consejería de Medio Ambiente, Junta de Andalucía. 630

Bargagli, R. 1998. Trace Elements in Terrestrial Plants. An Ecophysiological Approach 631

to Monitoring and Biorecovery, Springer-Verlag, Berlin.632

Bastida, F., Selevsek, N., Torres, I. F, Hernández, T., García, C., 2015. Soil restoration 633

with organic amendments: linking cellular functionality and ecosystem processes. 634

Sci. Rep. 5, 15550635

Bolan NS, Park JH, Robinson B, Naidu R, Huh KY., 2011. Phytostabilization. A green 636

approach to contaminant containment. Adv. Agron. 112, 145-204.637

Buscaroli, A. 2017. An overview of indexes to evaluate terrestrial plants for 638

phytoremediation purposes. Ecol. Indic. 82, 367-380. 639

Cabrera, F., Clemente, L., Díaz-Barrientos, E., López, R., Murillo, J.M., 1999. Heavy 640

metal pollution of soils affected by the Guadiamar toxic flood. Sci. Total Environ. 641

242, 117–129.642

Chaney, R.L. 1989. Toxic element accumulation in soils and crops: protecting soil 643

fertility and agricultural food-chains, in: Bar-Yosef, B., Barrow, N.J., Goldshmid, 644

J. (Eds), Inorganic Contaminants in the Vadose Zone. Springer-Verlag, Berlin, pp. 645

140–158646

28

Chazdon, R. L. 2008. Beyond deforestation: restoring forests and ecosystem services on 647

degraded lands. Science, 320, 1458-1460.648

Ciadamidaro, L., Madejón, P., Camacho, F., Fernández Boy, E., Madejón, E., 2016. 649

Organic compost to improve contaminated soil quality and plant fertility. Organic 650

compost to improve contaminated soil quality and plant fertility. Soil Sci. 181, 651

487-493.652

COM, 2002. Communication from the Commission to the Council, the European 653

Parliament, the Economic and Social Committee and the Committee of the 654

Regions. Towards a thematic strategy for soil protection. Brussels 16 (4), 1–35.655

Domínguez, M.T., Marañón, T., Murillo, J.M., Schulin, R., Robinson, B.H., 2008. 656

Trace element accumulation in woody plants of the Guadiamar Valley, SW Spain: 657

A large-scale phytomanagement case study, Environ. Pollut. 152, 50-59.658

Domínguez, M.T., Madrid, F., Marañón, T., Murillo, J.M., 2009. Cadmium availability 659

in soils and retention in oak roots: potential for phytostabilisation. Chemosphere 660

76, 480-486. 661

Domínguez, M.T., Madejón, P., Marañón, T., Murillo, J.M., 2010. Afforestation of a 662

trace-element polluted area in SW Spain: woody plant performance and trace 663

element accumulation. Eur. J. For. Res. 129, 47-59.664

Domínguez, M. T., Pérez-Ramos, I. M., Murillo, J. M., & Marañón, T. 2015. 665

Facilitating the afforestation of Mediterranean polluted soils by nurse shrubs. J. 666

Environ. Manage. 161, 276-286.667

Domínguez, M.T., Alegre, J.M., Madejón, P., Madejón, E., Burgos, P., Cabrera, F., 668

Marañón, T., Murillo, J.M., 2016. River banks and channels as hotspots of soil 669

pollution after large-scale remediation of a river basin. Geoderma 261, 133-140.670

29

Domínguez, M.T., Montiel-Rozas, M.M., Madejón, P., Diaz, M.J., Madejón, E. 2017. 671

The potential of native species as bioenergy crops on trace-element contaminated 672

Mediterranean lands. Sci. Total Environ. 590-591, 29-39. 673

European Commission, 2012. Report from the Commission to the European Parliament, 674

the Council, the Economic and Social Committee and Committee of the Regions. 675

Published February 2012. Available at, http://eur-lex.europa.eu/legal-676

content/EN/TXT/ PDF/?uri=CELEX:52012DC0046&from=EN (accessed on 30 677

October 2017).678

European Commission, 2015. Closing the loop -An EU action plan for the Circular 679

Economy. European Commission, Brussels.680

European Environment Agency (EEA), 2014. Progress in Management of Contaminated 681

Sites (LSI 003)-Assessment. Published May 2014. Available at, http://www.eea. 682

europa.eu/data-andmaps/indicators/progress-in-management-of-683

contaminatedsites-3/assessment (accessed on 30 October 2017).684

Gallart, F., Benito, G., Martín-Vide, J.P., Benito, A., Prió, J.M., Regüés, D., 1999. 685

Fluvial geomorphology and hydrology in the dispersal and fate of pyrite mud 686

particles released by the Aznalcóllar mine tailings spill. Sci. Total Environ. 242, 687

13–26.688

Grimalt, J.O., Ferrer, M., Macpherson, E., 1999. The mine tailing accident in 689

Aznalcóllar. Sci. Total Environ. 242, 3-11.690

Houba, V.J.G., Temminghoff, E.J.M., Gaikhorst, G.A, Van Vark, W., 2000. Soil 691

analysis procedures using 0.01 M calcium chloride as extraction reagent. 692

Commun. Soil Sci. Plant Anal. 31, 1299–1396693

30

Janoš, P., Vávrová, J., Herzogová, L., Pilařová, V., 2010. Effects of inorganic and 694

organic amendments on the mobility (leachability) of heavy metals in 695

contaminated soil: a sequential extraction study. Geoderma 159, 335-341.696

Li, D., Niu, S. and Luo, Y., 2012. Global patterns of the dynamics of soil carbon and 697

nitrogen stocks following afforestation: a meta-analysis, New Phytol. 195, 172-698

181.699

López-García, A., Gil-Martínez, M., Navarro-Fernández, C.M., Kjøller, R., Azcón-700

Aguilar, C., Domínguez, M.T., Marañón, T. 2018. Functional diversity of 701

ectomycorrhizal fungal communities is reduced by trace element contamination. 702

Soil Biol Biochem (submitted). 703

López-Pamo, E., Barettino, D., Antón-Pacheco, C., Ortiz, G., Arránz, J.C., Gumiel, J.C., 704

Martínez-Pledel, B., Aparicio M., Montouto, O., 1999. The extent of the 705

Aznalcóllar pyritic sludge spill and its effects on soils. Sci. Total Environ. 242, 706

57-88707

Madejón, E., Madejón, P., Burgos, P., Pérez de Mora, A., Cabrera, F., 2009. Trace 708

elements, pH and organic matter evolution in contaminated soils under assisted 709

natural remediation: A 4-year field study. J. Hazard. Mat. 162, 931-938.710

Madejón, P., Pérez-de-Mora, A., Burgos, P., Cabrera, F., Lepp, N.W., Madejón, E., 711

2010. Do amended, polluted soils require re-treatment for sustainable risk 712

reduction? -Evidence from field experiments. Geoderma 159, 174-181713

Madejón, P., Marañón, T., Navarro-Fernández, C.M., Domínguez, M.T., Alegre, J.M., 714

Robinson, B., Murillo, J.M. 2017. Potential of Eucalyptus camaldulensis for 715

phytostabilization and biomonitoring of trace element contaminated soils. PLoS 716

ONE, 12 (6), art. no. e0180240 717

31

Madejón, P., Domínguez, M.T., Madejón, E., Cabrera, F., Marañón, T., Murillo ,J.M. 718

2018. Soil-plant relationships and contamination by trace elements: A review of 719

twenty years of experimentation and monitoring after the Aznalcóllar (SW Spain) 720

mine accident. Sci. Total Environ. 625, 50-63.721

Mahar, A., Wang, P., Ali, A., Kumar, M., Hussain, A., Wang, Q., Li, R., Zhang, Z. 722

2016. Challenges and opportunities in the phytoremediation of heavy metals 723

contaminated soils: a review. Ecotoxicol. Environ. Saf. 126, 111-121.724

Manning, A. D., Fischer, J., Lindenmayer, D.B., 2006. Scattered trees are keystone 725

structures – implications for conservation. Biol. Conserv. 132, 311–321.726

Marañón, T., Navarro-Fernández, C.M., Domínguez, M.T., Madejón, P., Murillo, J.M., 727

2015. How the soil chemical composition is affected by seven tree species planted 728

at a contaminated and remediated site. Web Ecol. 15, 45-48.729

Marschner, P., Kandeler, E., Marschner, B., 2003. Structure and function of the soil 730

microbial community in a long-term fertilizer experiment. Soil Biol. Biochem. 35, 731

453–461732

McGrath, S. P., Zhao, J. and Lombi, E. 2002. Phytoremediation of metals, metalloids, 733

and radionuclides. Adv. Agron. 75, 1–56.734

Mench, M., Vangronsveld, J., Didier, V., Clijsters, H., 1994. Evaluation of metal 735

mobility, plant availability and immobilization by chemical agents in a limed-silty 736

soil. Environ. Pollut. 86, 279–286.737

Mench, M., Bussiere, S., Boisson, J., Castaing, E., Vangronsveld, J., Ruttens, A., De 738

Koe, T., Bleeker, P., Assuncao, A., Manceau, A. 2003. Progress in remediation 739

and revegetation of the barren Jales gold mine spoil after in situ treatments. Plant 740

Soil, 249, 187-202741

32

Mendez, M.O., Maier, R.M., 2008. Phytoremediation of mine tailings in temperate and 742

arid environments. Rev. Environ. Sci. BioTechnol. 7, 47-59.743

Montiel-Rozas, M.M., Panettieri, M., Madejón, P., Madejón, E., 2016a. Carbon 744

sequestration in restored soils by applying organic amendments. Land Degrad. 745

Dev. 27: 620–629746

Montiel-Rozas, M.M., López-García, A., Kjøller, R., Madejón, E., Rosendahl, S., 747

2016b. Organic amendments increase phylogenetic diversity of arbuscular 748

mycorrhizal fungi in acid soil contaminated by trace elements. Mycorrhiza 26, 749

575-585.750

Nagel, I., Lang, F., Kaupenjohann, M., Pfeffer, K.-H., Cabrera, F., Clemente, L., 2003. 751

Guadiamar toxic flood: factors that govern heavy metal distribution in soils. 752

Water Air Soil Pollut. 143, 211-224.753

Op De Beeck, M., Ruytinx, J., Smits, M.M., Vangronsveld, J., Colpaert, J.V., Rineau, 754

F., 2015. Belowground fungal communities in pioneer Scots pine stands growing 755

on heavy metal polluted and non-polluted soils. Soil Biol. Biochem. 86, 58–66.756

Pan, Y., Birdsey, R.A., Fang, J., Hayes, D., 2011. A large and persistent carbon sink in 757

the world´s forests. Science 333: 988-993.758

Panagos, P., Liedekerke, M.V., Yigini, Y.,Montanarella, L., 2013. Contaminated sites in 759

Europe: review of the current situation based on data collected through a 760

European network. Int. J. Environ. Res. Public Health, 158764.761

Pérez-de-Mora, A., Madejón, E., Burgos, P., Cabrera, F., 2006. Trace element 762

availability and plant growth in a mine-spill contaminated soil under assisted 763

natural remediation I. Soils. Sci. Total Environ. 363, 28-37.764

33

Pérez-de-Mora, A., Burgos, P., Cabrera, F., Madejón, E., 2007. "In situ" amendments 765

and revegetation reduce trace element leaching in a contaminated soil. Water Air 766

Soil Pollut. 185, 209-222.767

Pérez-de-Mora, A., Madejón, P., Burgos, P., Cabrera, F., Lepp, N.W., Madejón, E., 768

2011. Phytostabilization of semiarid soils residually contaminated with trace 769

elements using by-products: Sustainability and risks. Environ. Pollut. 159, 3018-770

3027771

Pourrut, B., Lopareva-Pohu, A., Pruvot, C., Garçon, G., Verdin, A., Waterlot, C., Bidar, 772

G., Shirali, P., Douay, F., 2011. Assessment of fly ash-aided phytostabilisation of 773

highly contaminated soils after an 8-year field trial: Part 2. Influence on plants. 774

Sci. Total Environ. 409, 4504-4510.775

Pulford, I. D., Dickinson, N. M., 2005. Phytoremediation Technologies Using Trees, in: 776

Prasad, M.N.V., Sajwan, K.S., Naidu, R. (Eds.), Trace elements in the 777

environment: biogeochemistry, biotechnology, and bioremediation. Boca Raton, 778

CRC Press, pp. 383-404.779

Russell, A. E., Hall, S. J., Raich, J. W. 2017. Tropical tree species traits drive soil cation 780

dynamics via effects on pH: A proposed conceptual framework. Ecol. Monogr (in 781

press). DOI: 10.1002/ecm.1274782

Ruttens, A., Adriaensen, K., Meers, E., De Vocht, A., Geebelen, W., Carleer, R., 783

Mench, M., Vangronsveld, J., 2010. Long-term sustainability of metal 784

immobilization by soil amendments: cyclonic ashes versus lime addition. Environ. 785

Pollut. 158, 1428-1434.786

Sanderman, J., Hengl, T., Fiske, G. J., 2017. Soil carbon debt of 12,000 years of human 787

land use. Proc. Natl. Acad. Sci. U.S.A. 114, 9575-9580.788

34

Sarwar, N., Imran, M., Shaheen, M. R., Ishaque, W., Kamran, M. A., Matloob, A., 789

Rehim, A., Hussain, S. 2017. Phytoremediation strategies for soils contaminated 790

with heavy metals: Modifications and future perspectives. Chemosphere, 171, 791

710-721.792

Shrestha R, Lal R. 2006. Ecosystem carbon budgeting and soil carbon sequestration in 793

reclaimed mine soil. Environ. Int. 32: 781–796. 794

Tordoff, G., Baker, A., Willis, A., 2000. Current approaches to the revegetation and 795

reclamation of metalliferous mine wastes. Chemosphere 41, 219-228.796

Van Nevel, L., J. Mertens, J. Staelens, A. de Schrijver, F.M.G. Tack, S de Neve, E. 797

Meers, K. Verheyen, 2011. Elevated Cd and Zn uptake by aspen limits the 798

phytostabilization potential compared to five other tree species. Ecol. Eng. 37, 799

1072-1080.800

Virto, I., Imaz, M.J., Fernández-Ugalde, O., Gartzia-Bengoetxea, N., Enrique, A., 801

Bescansa, P., 2015. Soil degradation and soil quality in Western Europe: current 802

situation and future perspectives. Sustainability 7, 313–365.803

Walkley, A., Black, I. A.1934. An examination of the Degtjareff method for 804

determining soil organic matter, and a proposed modification of the chromic acid 805

titration method. Soil Sci. 37, 29–38.806

Wang, L., Ji, B., Hu, Y., Liu, R., Sun, W., 2017. A review on in situ phytoremediation 807

of mine tailings. Chemosphere 184, 594-600.808

Xiong, J., Madejón, P., Madejón, E., Cabrera, F., 2015. Assisted natural remediation of 809

a trace element-contaminated acid soil: An eight-year field study. Pedosphere 25, 810

250-262.811

Table 1. Bioconcentration factors calculated as [Element]plant/[Element]soil ratios based

on pseudototal values. For the amendment experiment the whole shoot was analysed

and the response to the two amendment types are shown. In the tree plantation, the BCF

indexes for the seven tree species are calculated separately for leaves and roots. Values

>1 are marked with bold letter.

Plant species As Cd Cu Pb Zn

Amendments Type

Lamarckia* NA 0.01 0.23 0.10 0.04 0.48

SL <0.01 0.14 0.06 0.01 0.31

BC 0.01 0.20 0.08 0.03 0.40

Poa* NA 0.02 0.27 0.07 0.01 0.42

SL 0.01 0.20 0.05 0.01 0.29

BC <0.01 0.14 0.04 <0.01 0.23

Cynodon NA <0.01 0.08 0.03 <0.01 0.22SL <0.01 0.02 0.03 <0.01 0.08BC <0.01 0.10 0.03 <0.01 0.26

Tree plantations Organ

Celtis Leaf <0.01 0.27 0.06 0.01 0.28Root 0.15 1.12 0.72 0.33 0.81

Ceratonia Leaf <0.01 0.29 0.02 <0.01 0.14Root 0.04 1.35 0.49 0.07 0.45

Fraxinus Leaf 0.01 0.05 0.05 <0.01 0.08Root 0.07 2.61 0.56 0.15 0.75

Olea Leaf <0.01 0.06 0.04 <0.01 0.18Root 0.08 1.19 0.59 0.14 0.60

Pinus Leaf <0.01 0.18 0.02 <0.01 0.29Root 0.05 3.76 0.49 0.09 0.65

Populus Leaf 0.01 2.96 0.05 <0.01 1.84Root 0.06 1.77 0.45 0.10 0.50

Quercus Leaf <0.01 0.19 0.04 <0.01 0.25Root 0.01 1.06 0.24 0.02 0.32

* Data from Pérez de Mora et al. (2011)

Figure Captions

Fig. 1. Evolution over time of soil pH after application of amendments (SL sugar beet

lime, and BC biosolid compost) in comparison with unamended soils (NA; mean and

SE bar). For each year significant differences between treatments are indicated with

different letters (p<0.05). For each treatment there were no significant differences with

time (2007-2011-2016).

Fig. 2. Evolution of available (CaCl2- extracted) trace element concentrations (mean and

SE bar) over time in soils after application of amendments (SL sugar beet lime, and BC

biosolid compost) in comparison with unamended soils (NA). For each year significant

differences between treatments are indicated with different letters (p<0.05). For each

treatment there were no significant differences with time (2007-2011-2016).

Fig. 3. Mean and SE values of soil pH (0-10 cm depth) underneath seven types of tree

species and comparison with adjacent open sites (indicated with a dotted line; data of

2014 samples). When mean values are significantly different (according to LSD test)

they are indicated by different letters.

Fig. 4. Relationships between soil pH and availability of Cd, Cu and Zn in soil (0-10

cm) under different tree species, and in adjacent open sites (data of 2014 samples).

Fig. 5. Evolution over time of total organic carbon (TOC, mean +SE) after application

of amendments (SL sugar beet lime, and BC biosolid compost) in comparison with

unamended soils (NA). For each year significant differences between treatments are

indicated with different letters (p<0.05). For each treatment there were no significant

differences with time (2007-2011-2016).

Fig. 6. Soil total organic carbon (TOC) under seven tree species, at two soil depths (0-

10 cm and 10-30cm), and comparison with adjacent open sites (indicated with a dotted

line; data of 2014 samples).

Fig. 7. The implementation of a soil phytoremediation process, after putting an end to

the contamination source, should include three stages: amendment addition, tree

plantation, and monitoring of trace elements bioavailability. In case of residual

contamination and toxicity risk, new remediation operations should be applied.

Figure 1

Year

pH

va

lue

s

0

1

2

3

4

5

6

7

8

9

NASLBC

2002 2007 2011 2016

a

c

b

b

a

c

b

a

b

Figure 2

Cd

co

nc

en

tra

tio

n (

mg

kg

-1)

0.0

0.1

0.2

0.3

0.4

2002 2007 2011 2016

Cu

co

nc

en

tra

tio

n (

mg

kg

-1)

0.0

0.2

0.4

0.6

3.0

6.0

9.0

12.0

15.0

18.0

2002 2007 2011 2016

Zn

co

nc

en

tra

tio

n (

mg

kg

-1)

0.0

0.2

0.4

0.6

0.8

1.0

20.0

40.0

60.0

80.0

100.0

120.0

2002 2007 2011 2016

c

a

b

a

a

a

b

a

ab

c

a

b

a

a

a

b

a a

c

a

b

b

a

a

b

a

ab

Year

Figure 3

Cel

tis

Cer

atonia

Fraxi

nusOle

a

Pinus

Populus

Quer

cus

Open

pH

valu

es

0

1

2

3

4

5

6

7

ab

ab

ab

aa a

b

b

Figure 4

Soil pH

2 3 4 5 6 7

Cd

co

ncen

trati

on

(m

g k

g-1

)

0.0

0.1

0.2

0.3

0.4

0.5

Soil pH

2 3 4 5 6 7

Cu

co

ncen

trati

on

(m

g k

g-1

)

0

5

10

15

20

25

30

Soil pH

2 3 4 5 6 7

Zn

co

ncen

trati

on

(m

g k

g-1

)

0

20

40

60

80

100

120

Figure 5

Year

TO

C (

g k

g-1

)

0

3

6

9

12

15

18

21

NASLBC

2002 2007 2011 2016

a

ab

b

a

b

b

a

bb

Figure 6

Cel

tis

Cer

atonia

Fraxi

nusOle

a

Pinus

Populus

Quer

cus

Open

TO

C (

g k

g-1

)

0

5

10

15

20

25

0-10 cm10-30 cm

Figure 7