Isotopically exchangeable Al in coastal lowland acid sulfate soils

Science of the Total Environment 542 (2016) 129–135

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Isotopically exchangeable Al in coastal lowland acid sulfate soils Yliane A.M. Yvanes-Giuliani a,b, D. Fink b, J. Rose c, T. David Waite a, Richard N. Collins a,⁎ a b c

UNSW Water Research Centre, School of Civil and Environmental Engineering, UNSW Australia, Sydney, NSW 2052, Australia Centre Européen de Recherche et d'Enseignement des Géosciences de l'Environnement, Aix-Marseille Université, Aix en Provence, France Institute for Environmental Research, Australian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC, NSW 2232, Australia

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Isotopically exchangeable Al was compared to 1 M KCl or 0.2 M CuCl2 extractable Al. • 1 M KCl always underestimated isotopically exchangeable Al concentrations. • 0.2 M CuCl2 mobilised non-isotopically exchangeable Al • 1 M KCl values require correction of ~ 1.7 to reflect exchangeable Al concentrations.

a r t i c l e

i n f o

Article history: Received 21 July 2015 Received in revised form 8 October 2015 Accepted 9 October 2015 Available online 28 October 2015 Editor: D. Barcelo Keywords: Exchangeable aluminium Isotope exchange Extractable aluminium

a b s t r a c t Periodic discharges of high concentrations of aluminium (Al) causing fish kills and other adverse effects occur worldwide in waterways affected by coastal lowland acid sulfate soils (CLASS). The exchangeability — a metal's ability to readily transfer between the soil solid- and solution-phases — of Al in these soils is therefore of particular importance as it has implications for metal transport, plant availability and toxicity to living organisms. In the present study, the concentrations of isotopically exchangeable Al (E values) were measured in 27 CLASS and compared with common salt extractions (i.e. KCl and CuCl2) used to estimate exchangeable soil pools of Al. E values of Al were high in the soils, ranging from 357 to 3040 mg·kg−1. Exchangeable concentrations estimated using 1 M KCl were consistently lower than measured E values, although a reasonable correlation was obtained between the two values (E = 1.68 × AlKCl, r2 = 0.66, n = 25). The addition of a 0.2 M CuCl2 extraction step improved the 1:1 agreement between extractable and isotopically exchangeable Al concentrations, but lead to significant mobilisation of non-isotopically exchangeable Al in surficial ‘organic-rich’ CLASS having E values b 1000 mg·kg−1. It was concluded that currently used (i.e. 1 M KCl) methodology severely underestimates exchangeable Al and total actual acidity values in CLASS and should be corrected by a factor similar to the one determined here. © 2015 Elsevier B.V. All rights reserved.

⁎ Corresponding author. E-mail address: [email protected] (R.N. Collins).

http://dx.doi.org/10.1016/j.scitotenv.2015.10.051 0048-9697/© 2015 Elsevier B.V. All rights reserved.

130

Y.A.M. Yvanes-Giuliani et al. / Science of the Total Environment 542 (2016) 129–135

1. Introduction

2.2. Sample analyses

High concentrations of labile and reactive aluminium (Al) in coastal lowland acid sulfate soils (CLASS) are responsible for the periodic discharge of large quantities of Al (Macdonald et al., 2007; Yvanes-Giuliani et al., 2014) into adjacent aquatic systems resulting in adverse toxic effects to living organisms. Some of the more labile forms of Al in CLASS, such as the water soluble and exchangeable fractions, are of particular interest as they are the most easily mobilised into drains following rainfall events (Yvanes-Giuliani et al., 2014). Measuring ‘water soluble’ Al is relatively straightforward, however, quantifying the exchangeable fraction of Al is more challenging. While there is no consensus on how to measure the most available fraction of contaminants in soils, various chemical extractions have been used to try and isolate the ‘exchangeable’ pool of aluminium. The standard extraction solution for determining exchangeable acid-generating cations, principally Al, in CLASS is the use of 1 M KCl (Ahern et al., 2004). Chemical extractions have also been regularly used to chemically parameterise CLASS (Claff et al., 2011; Johnston et al., 2010; Yvanes-Giuliani et al., 2014); however, data on the lability of Al in CLASS is quasi non-existent (Johnston et al., 2010; Yvanes-Giuliani et al., 2014). Although chemical extractions have some limitations, they are the most convenient and therefore the most widely used tool to estimate the potentially bioavailable and phytoavailable fraction of metals in soils. Isotope exchange can be used to determine the concentration of (isotopically) exchangeable elements in soils and, because it does not disturb the geochemical equilibrium, can be considered a reference method to quantify exchangeable concentrations of a range of elements in CLASS and other soil types (Hamon et al., 2002). Possibly due to limited supply, financial constraints and the requirement of specialist counting techniques (e.g. Accelerator Mass Spectrometry (AMS)), isotopic tracer studies of Al are rare (Kleja et al., 2005; Kotzé et al., 1984) and comparable isotope exchange data on Al in CLASS is currently lacking. Exchangeable concentrations estimated from KCl extraction (the most commonly used method for CLASS) have never been compared to the concentrations of Al that are isotopically exchangeable at prevailing geochemical conditions. A significant lack of understanding regarding the nature of the Al brought into solution from chemical extractions therefore exists. Furthermore, it is unclear if the concentrations of exchangeable Al determined from use of chemical extractants have any resemblance to actual exchangeable concentrations in the soils. As such, the main objective of this study was to compare the use of conventional extraction methods for measurement of exchangeable Al with the use of 26 Al isotope exchange.

General chemical analysis methods, sample preparation procedures and methods for determination of pH, cation exchange capacity (CEC, 1 M KCl), soil total organic carbon content (TOC, LECO CHN TruSpec analyser) and ‘pseudo-total’ Al concentrations (Altot, aqua ‘regia’) were identical to those described by Yvanes-Giuliani et al. (2014). A summary of selected soil chemical characteristics of the soils are presented in Appendix A, Table A1.

2. Materials and methods 2.1. Study area and sample collection Twenty-seven samples were collected on sugar cane farms across three catchments in north eastern New South Wales, Australia (Appendix A, Fig. A1): Black's drain and Ledday's creek on the Tweed valley floodplain (samples in the B and L series) and Christie's creek, a coastal catchment ~20 km south of the Tweed valley (samples in the C series). A CLASS profile in these farms is typically composed of an organic rich surface soil horizon (0 to ~30 cm), an oxidised and acidified (actual) acid sulfate soil horizon (~ 30 to ~ 60 cm), a transition zone (60 to ~ 120 cm) and a potential acid sulfate soil horizon (− 120 cm and below) (Collins et al., 2010; Jones et al., 2011; Kinsela et al., 2011; Kinsela and Melville, 2004; Macdonald et al., 2007). Multiple distinct samples were collected from surface (14) and actual acid sulfate soil (13) layers at 15 different sites in order to provide a range of soil variability. In some cases, both soil layers were collected from a site whereas at other times only a surface or actual acid sulfate soil layer was collected.

2.3. Theoretical background to isotope exchange Isotope exchange techniques can be used to determine the labile fraction (i.e. the solid phase exchangeable fraction in dynamic equilibrium with the solution phase over a defined period of time) of elements in soils. A tracer isotope is added to a soil suspension and allowed to equilibrate with the soil solution for a defined amount of time during which the isotope redistributes itself between the solid and solution phases in an identical fashion to the corresponding ‘native’ exchangeable element. The isotopically exchangeable concentration (E) is most often defined as (Hamon et al., 2002):
 v −1 E mg
kg ¼ Kd  cs þ cs  m

ð1Þ

where Kd is the partitioning coefficient of the isotope tracer in L·kg−1 soil, cs is the concentration of the stable element of interest in mg·L−1, and v/m is the solution-to-solid ratio of the suspension (L·kg−1) in which isotope exchange is conducted. Kd is usually calculated as:
 R−r v −1 Kd L
kg  ¼ r m

ð2Þ

where R is the quantity of the isotope tracer added to the sample and r is the quantity of tracer remaining in solution after equilibration. In the present study, however, because we measure a ratio — 26 Al:27Al — rather than directly measuring the concentration of 26Al remaining in solution, Eq. (2) may be modified as follows: 


−1

Kd L
kg

26



26

¼

   . . AMS 26 Al 27 Al − Al 27 Al v
   AMS 26 m Al 27 Al

ð3Þ



where 27 Al is the ratio calculated based on the quantity of tracer added Al 26 AlAMS

is the ratio measured by AMS after equilibration. In Eq. (3), R (i.e. 26Al⁎) and r (i.e. 26AlAMS) are just expressed relative to the concentration of carrier (27Al) added for AMS measurement. to the sample and 27 Al

2.4. 26Al spiking, equilibration and extractions Isotope exchange experiments were conducted on all 27 soil samples with 12 soils analysed in duplicate in order to provide an indication of sample variability. For the isotope exchange experiments, approximately 5 g (dry weight equivalent) of field moist soil was suspended in N 18.2 MΩ·cm−1 water (to obtain a soil:solution of 1:5) and allowed to pre-equilibrate on an orbital shaker for 24 h. The 26Al used for these experiments was purchased from Los Alamos National Laboratory, NM, USA. An appropriate amount of 26Al from a diluted spike solution (5.74 × 10−12– 1.78 × 10−11 mol 26Al and 2.02 × 10−10–6.26 × 10−10 mol of carrier 27 Al, Appendix A, Table A2) was added to each soil suspension and allowed to equilibrate for a further 24 h, the typical duration used when CLASS are extracted with 1 M KCl (Ahern et al., 2004). After centrifugation (15 min at 3000 rpm) and filtration (0.22 μm PVDF durapore membrane

Y.A.M. Yvanes-Giuliani et al. / Science of the Total Environment 542 (2016) 129–135

filters, Millipore, Merk Pty, Ltd.), aliquots were acidified and analysed by ICP–OES (Agilent Varian vista pro 710) in order to determine stable 27Al concentrations. Subsamples were also used for 26Al:27Al ratio measurement via Accelerator Mass Spectrometry (AMS) as described further below. The remaining supernatant was discarded and the soils resuspended in approximately 25 mL of 1 M KCl. After 24 h of equilibration, the supernatant phase was separated from the solid phase via centrifugation, filtered and acidified in a manner similar to that used for the water extracts. The remaining supernatant was again discarded and the soils were mixed with approximately 25 mL of 0.2 M CuCl2 and shaken for 24 h before being centrifuged, filtered and acidified. Elemental analyses on the extraction solutions were performed by ICP–OES. To facilitate the comparison of both approaches (i.e. isotopically determined E values and exchangeable concentrations estimated with chemical extractants), the same contact time between the soil and isotope and between the soil and chemical extractant was used (i.e. 24 h). A schematic showing sample spiking, extraction and preparation procedures is provided in Appendix A, Fig. A2. 2.5. 26Al sample preparation and AMS analyses A suitable amount (2–10 mL, Appendix A, Table A2) of high purity Al (99.9995%, 1000 mg·L− 1 Al standard, Choice Analytical, NSW, Australia) was added to the aliquots preserved for 26Al AMS analyses. The addition of an appropriate amount of carrier was crucial both to ensure the recovery of enough Al oxide for AMS analyses and to minimise the possibility of 26Al contamination of the Australian Nuclear Science and Technology Organisation's Cosmogenics Laboratory and accelerator. The steps required to produce Al2O3 for AMS analyses involves a series of precipitation, dissolution and purification steps in order to separate and purify Al from the soil solution phase (Child et al., 2000). After recovery of the aluminium oxide from calcination and after mixing with an appropriate amount of Nb (99.95%, 325 mesh in a mass ratio of Al2O3:Nb of 1:3), the samples were pressed into copper AMS targets. Accelerator Mass Spectrometry analyses were performed at the Australian National Tandem Accelerator for Applied Research (ANTARES) (Fink et al., 2004; Fink and Smith, 2007). The 26Al:27Al ratios of the chemical procedural blanks were between 7 · 10− 15 and 1 · 10−12. While the latter ratio is high for typical AMS measurement, it was orders of magnitude lower than most of the samples (Appendix A, Table A3) and only for 3 samples was the procedural blank N 5% of the 26Al:27Al ratio measured in the sample (0.03–9.71%, median of 0.68%). Background correction was performed using the procedural blanks ratios. The results were also normalised relative to standards KN-A-01-4-1 or KN-A-01-4-2 prepared by Nishiizumi (2004) that are routinely used at ANTARES (a comparison of the nominal ratio as well as measurements performed at ANTARES is presented in Appendix A, Table A4). In-house standards were also prepared from the 26Al spike solution and high purity 27Al standard with the expected and measured ratios provided in Appendix A, Table A5. The % error in the measured 26 Al:27Al ratios in the samples was 3.61% on average with errors generally below b 5% except for samples that had ratios in the low 10−11 (6 samples, maximum error of 6.67%). The measured ratios and errors are shown in Appendix A, Table A3. 27

3. Results and discussion In this study, surface soils are referred to as ‘organic-rich’ samples as they contain higher concentrations of total organic carbon (TOC) than the deeper actual acid sulfate soil samples. These soils, however, are not actually “organic soils” as the accepted threshold for such soils is usually 20% TOC (d.w. soil equivalent) (Soil Survey Staff, 2014). The maximum recorded for these soils was 14.8% with an average TOC value of 9% (Yvanes-Giuliani et al., 2014). Regardless, the deeper actual acid sulfate soil horizons contained, on average, less than half the TOC of the surficial soils, hence the use of ‘organic-rich’ to describe these soils. 3.1. % sorbed Al The percentage of Al that is isotopically exchangeable and associated with the soil solid phase can be calculated as follows (Degryse et al., 2009): %Alsorbed ¼

Kd Kd þ

v  100: m

ð4Þ

With the exception of two samples, over 98% of the 26Al, and hence the isotopically exchangeable ‘native’ 27Al, was associated with the solid-phase over the natural pH range of the samples (Fig. 1) despite Al solution phase concentrations reaching values as high as 1.18 mM (Appendix A, Table A2). 3.2. Partitioning coefficients — Kd The Kd values calculated here represent the partitioning of 26Al between the isotopically exchangeable solid phase and the solution phase as they are derived from addition of the isotope tracer to soil samples. These values are often referred to as Klab d in order to differentiate them from Kd values that refer to the total metal concentration in the solid phase (i.e. Mtot), namely Ktot d (Degryse et al., 2009), i.e.:

Kd

tot


 −1 Mtot mg
kg
: ¼ ½MH2 O mg
L−1

ð5Þ

The Kd of 26Al varied by more than 2 orders of magnitude in the soils — 62 to 8904 L·kg−1 — across all sampling sites (Fig. 2 with individual results reported in Appendix A, Table A6). In the organic-rich surface soils, the Kd values were much higher (average of 4803 ± 2232 L·kg− 1) than in the actual acid sulfate soil layers (average of 810 ± 557 L·kg−1) which could either be due to the high sorption affinity of soil organic matter (SOM) for metals and Al in particular and/or to

2.6. Statistical analyses One- and two-parameter (multiple) linear regressions were determined with the data analysis tool of Microsoft Excel 2010 to examine correlations between isotopically exchangeable Al E and Kd values with various soil properties. The statistical significances (i.e. probability (p)) of the correlations were determined with critical values for the Pearson product–moment correlation coefficient for a two-tailed test with degrees of freedom (df) = n − 2. The data were log transformed to meet the normality assumption when appropriate for the multiple regression tests.

131

Fig. 1. Percentage of 26Al sorbed to the soil solid phase relative to pH.

132

Y.A.M. Yvanes-Giuliani et al. / Science of the Total Environment 542 (2016) 129–135

the higher pH values of these soils (Sauvé et al., 2003; Sparks, 2003). While lower Kd values may be related to a higher solubility of Al, the observation of high Kd values in the organic-rich surface soils does not necessarily mean that Al solubility is low in these soils, only that there are very large concentrations of exchangeable Al adsorbed on the solid phase compared to the Al concentrations in solution. Indeed the concentrations of Al in solution after isotopic equilibration were relatively high in all samples (0.14–31.8 mg·L−1, average of 2.81 mg·L−1). A weaker retention of Al by the solid phase was observed in the actual acid sulfate soil samples (pH 3.0–4.0) where the distribution of Al between the solid and liquid phases was highly pH dependent (Fig. 2) and a sorption maxima of 26Al occurred at pH 4–4.5. When compared to Fig. 1, it is clear that examination of the Kd rather than the proportion of Al sorbed allows for more sensitive analysis of the trends in the partitioning of isotopically exchangeable Al in the soils as a function of pH. In our previous study (Yvanes-Giuliani et al., 2014), it was concluded that a larger proportion of the exchangeable pool of Al in ‘organic’ surficial CLASS is present in the solution phase at pH values N 4.8 as a result of complexation with dissolved organic carbon (DOC). Indeed, it can be noted in Fig. 2 that Kd values trended towards lower values after the occurrence of a sorption maximum between pH 4.3–4.7. As discussed above, the partitioning of Al between the solid and solution phases was highly correlated with pH. Correlations with other soil parameters (CEC, TOC and ‘pseudo-total’ Al content) were also investigated via simple and multiple empirical regression analyses. The regression equations and statistical indicators are presented in Table 1 and plots of the statistically significant regressions can be found in Appendix A, Fig. A3. Using a stepwise multiple linear regression, it was found that pH accounted for the majority of the variation in the data. While the regression could be improved by the addition of a second component, CEC, its contribution was relatively weak as can be seen by the small coefficient associated with this parameter in Table 1. Although SOM is known to be an important soil sorbent for a range of metals, including Al (Degryse et al., 2009; Donisa et al., 2003; Sparks, 2003), the dependence of Kd on total organic carbon (TOC) content was insignificant (p N 0.05). 3.3. Isotopically exchangeable concentrations — E values The concentration of isotopically exchangeable Al measured in both the organic-rich and acidic soils ranged from 357 mg·kg− 1 to 3040 mg·kg−1 (mean: 1645 ± 709 mg·kg−1) (Fig. 3, see Appendix A, Table A6 for individual results). Due to the limited number of relevant studies, these results cannot be easily compared to literature values.

Although Kd and E values spanning several orders of magnitude have been reported for metals in soils exhibiting a range of characteristics (Degryse et al., 2009; Sauvé et al., 1999), the current E values are similar, although higher, than those reported previously for Al in Spodosols in central and southern Sweden (Kleja et al., 2005). Based on the pH and TOC of the Spodosol soils (mainly B horizon) examined by Kleja et al. (2005), it would be expected that they would present attributes similar to the AASS investigated here, with the exception that AASS also contain high sulfate concentrations. High concentrations of isotopically exchangeable Al concentrations are present in the soils and represent a small but significant fraction of the ‘pseudo-total’ pool of Al (see Appendix A, Table A6 for a comparison of the E values to Altot). Indeed, in spite of the clear dominance of the inert inorganic pool of Al in the studied soils, the E values are considerable (4.81 ± 2.27% of Altot). It must be emphasised that while this may not seem to be a significant fraction of the total pool of soil Al, these soils are clay based and therefore contain very large quantities of aluminosilicates. Soils in this area are mostly made of sandy clay loam or light to medium clays which have a clay content of up to 60% (Jenkins and Morand, 2004). The soils at Black's drain and Christie's creek are light to medium clay soils (35–45% clay) while the soils at Ledday's creek generally have a slightly more sandy texture with a lower (20–45%) clay content (ASRIS, 2011; Lin et al., 1998). That up to 9.5% of the Al could be present as a labile form that is readily dissolved from the aluminosilicates initially present in the parent material is significant (Appendix A, Table A6). In contrast to the partitioning coefficients calculated for the soils, the variability of E values between soils was quite low (less than one order of magnitude difference). It was observed, however, that the E values were slightly lower on average in the organic-rich soils than in the acidic soils (i.e. 1401 ± 707 mg·kg−1 and 1926 ± 622 mg·kg−1 respectively). Empirical simple and multiple linear regression analyses were also used to evaluate correlations between the E values and selected soil chemical characteristics, namely pH, CEC, TOC and ‘pseudo-total’ Al content. The results of these regression analyses are reported in Table 2 and plots of the significant correlations are presented in Appendix A, Figs. A4 and A5. Similar to the Al partitioning coefficients (Kd), the regression of E values with TOC content was insignificant (p N 0.05). Reasonably strong correlations were only obtained with pH and CEC in the organic-rich soils (Table 2 and Appendix A, Fig. A4), however, it is clear that the E value correlated best with pH (Table 2 and Appendix A, Fig. A4). However, linear regressions of the E values with pH varied between soil types: the linear relationship explained more of the variability amongst the organic-rich samples (65%) than amongst the acidic samples (only 13%). The slopes of both regression equations are also very different, clearly highlighting the difference in behaviour between the acidic and organic soils. Although statistically significant, the regression of the E values with the CEC of the soils explained less than 30% of the variation and deviated quite significantly from a 1:1 relationship (Table 2 and Appendix A, Fig. A5). 3.4. Comparison of E values and chemical extractants for estimation of exchangeable Al Simple salt extractions for estimating exchangeable metal concentrations in soils can be appealing due to their ease of use compared to Table 1 Empirical regression equations of the log-transformed partitioning coefficient (Kd) of Al with selected soil parameters. The regression equations are for all soil samples (both organic and acidic soils, n = 27) in decreasing order of significance.

26

Fig. 2. Partitioning coefficients (Kd) of Al in the ‘organic’ and ‘acidic’ soil samples expressed as a function of pH. The Kd values have been log transformed. The quadratic equation used to fit the data was log(Kd) = − 15.66 + 8.56 × pH – 0.95 × pH2, r2 = 0.90, p b 0.01.

Regression

SEE

r2

p

logKd = 1.11 × pH – 1.73 × logCEC – 0.81 logKd = 0.90 × pH – 0.44 logKd = 0.02 × Altot + 2.36 logKd = 0.06 × CEC + 2.30

0.30 0.32 0.52 0.53

0.73 0.68 0.16 0.13

1.9E−7 1.2E−7 0.04 0.06

Y.A.M. Yvanes-Giuliani et al. / Science of the Total Environment 542 (2016) 129–135

Fig. 3. Isotopically exchangeable concentrations of Al in the soil samples as a function of pH. Linear regressions of E values against pH are shown for both sets of soil horizons. The results for each individual soil are also presented in Appendix A, Table A6.

isotope exchange experiments. However, a suitable extractant can be hard to find. Identification of a single extractant or even a combination of extractants that can specifically isolate the true ‘labile’ fraction of a soil is often impossible (Degryse et al., 2009; Sauvé et al., 1999). Indeed, many ‘mineral’ fractions may be responsible for the lability of aluminium. The optimum electrolyte would solubilise all forms of labile Al without bringing non-labile Al into solution. As mentioned earlier, KCl is widely used to estimate exchangeable Al (Garcia-Rodeja et al., 2004; Lin et al., 1998; Walna et al., 2005) and is also a reference method for CLASS parameterisation (Ahern et al., 2004). Potassium chloride, however, does not extract all of the labile Al pool in many soil types. In particular, KCl has been found to be unsuitable for the estimation of ‘exchangeable’ Al concentrations in soils with high organic content and in variable charge soils (i.e. rich in iron oxy(hydroxides)) (GarciaRodeja et al., 2004). A range of other extractants has also been used for estimating ‘labile’ pools of Al. In this study, CuCl2 was employed following findings that the concentrations extracted by KCl could be up to 9-fold lower than those extracted with CuCl2 in CLASS, especially those with relatively high organic contents (Yvanes-Giuliani et al., 2014). One of the main objectives of this study was to evaluate how these two extractants performed compared to a ‘reference’ method such as isotope exchange. As can be seen in Fig. 4, the concentrations of Al extracted with KCl always underestimated isotopically exchangeable concentrations of Al and, on average, were 58.3 ± 15.3% of the corresponding E values (individual results for the soil samples are provided in Appendix A, Table A6). Recoveries of isotopically exchangeable Al by KCl was exceedingly poor in organic-rich soils having E values N2600 mg·kg−1 (27–30%, Appendix A, Table A6). Exchangeable Al concentrations estimated using KCl were up to almost four times lower than that measured using isotope exchange with the difference most pronounced in organic-rich soils. By following with a 0.2 M CuCl2 extraction, significantly more Al was extracted, particularly from the organic-rich soils. As such, it would be expected that the quantity of Al extracted by 0.2 M CuCl2 would be similar regardless of whether this extraction solution was used in isolation

133

or after extraction with 1 M KCl. Extractable Al concentrations were twice as high, on average, from the organic-rich soils compared to the acidic soils (1100 mg·kg−1 vs 560 mg·kg−1). The concentrations extracted by a combination of KCl and CuCl2 were, for the most part, in better 1:1 agreement with the E values (Fig. 4) but the results were very different for both soil types. In the acidic samples, AlKCl + CuCl2 was in very good agreement with, and provided a good estimation of, isotopically exchangeable Al concentrations (Appendix A, Table A6). In the organic soils, however, CuCl2 extracted up to ~430% of the isotopically exchangeable concentration of Al in organic soils having E values b1000 mg·kg−1 (Appendix A, Table A6) indicating that CuCl2 was extracting non-isotopically exchangeable Al from these samples. This is clearly shown in Fig. 5 where it can be observed that KCl-extracted Al is largely uninfluenced by the size of the isotopically exchangeable pool of Al, but CuCl2-extracted Al is highly correlated. This result was unexpected based on literature that has reported Al extracted by CuCl2 is typically associated with weaker organic complexes as opposed to relatively inert complexes (Garcia-Rodeja et al., 2004; Kleja et al., 2005; Walna et al., 2005). While the possibility remains that some of the Al mobilised with this extraction step could be from inorganic mineral dissolution, previous studies have found that CuCl2 usually extracts the more labile Al-SOM fraction and does not significantly dissolve other mineral forms of Al (i.e. non-exchangeable interlayer Al or Al minerals) that are solubilised by extractants such as Na4P2O7 (Garcia-Rodeja et al., 2004; Kleja et al., 2005; Walna et al., 2005). If Al extracted by 0.2 M CuCl2 is considered to be representative of Al complexed with solid-phase SOM, then these results suggest that CuCl2 can also extract Al-SOM complexes of a more inert nature (i.e. Al that is non-isotopically exchangeable within 24 h) and, in this case, appears to be related to organic-rich CLASS having relatively low isotopically exchangeable Al concentrations. While Al extracted by CuCl2 may not be fully isotopically exchangeable, organic complexes and colloids are considered potentially highly reactive (Walna et al., 2005) and are still of interest when evaluating the pool of Al likely to become a source of mobile Al. In fact, organic matter colloids are highly mobile and can, in some instances, enhance the transport of Al and other metals over larger distances (Aucour et al., 2003; Lead and Wilkinson, 2006; Tavakkoli et al., 2013). From this perspective, the concentrations extracted with CuCl2 are arguably still relevant. 3.5. Practical implications The results presented in this study clearly highlight the potential bias introduced by using either KCl or CuCl2 to estimate exchangeable Al concentrations in soils. Titratable actual acidity (TAA) determinations of the actual acidity (AA) using KCl are routinely undertaken on acid

Table 2 Empirical regression equations of E values for all the samples, the organic-rich surface soils and the actual acid sulfate soil samples with selected soil chemical characteristics. Regression All samples E = 97.3 × CEC + 3230 Organic-rich E = −1900 × pH + 9810 E = −127 × CEC + 3650 Acidic E = 41.0 × Altot + 478 E = 857 × pH −1160

SEE

r2

n

p-Value

641

0.24

27

0.01

451 531

0.65 0.52

14 14

4.9E−4 3.8E−3

524 606

0.35 0.13

13 13

0.03 0.23

Fig. 4. Comparison of the isotopically exchangeable concentrations of Al (E value) with the concentrations of Al extracted with 1 M KCl and 1 M KCl followed by 0.2 M CuCl2. The 1:1 line is shown in black. 95% confidence limits of the regressions are shown as the dashed lines. Samples noted with an * were not included in the regression analyses as described in Section 3.4. Individual results for the soil samples are provided in Appendix A, Table A6.

134

Y.A.M. Yvanes-Giuliani et al. / Science of the Total Environment 542 (2016) 129–135

Fig. 5. Comparison of the % isotopically exchangeable Al (E value) extracted with 1 M KCl or 1 M KCl followed by 0.2 M CuCl2 as a function of the E value. Individual results are provided in Appendix A, Table A6.

sulfate soils based on the assumption that KCl will extract exchangeable cations responsible for the generation of acidity that will subsequently require neutralisation (Ahern et al., 2004; Kinsela and Melville, 2004; Oates and Kamprath, 1983; Vithana et al., 2013). These determinations are used to estimate the amount of lime required to limit the persistence and transport of acidity from acid sulfate soils. Although the present results demonstrate that the current methodology is not suited for this purpose, it is recognised that lime calculations typically include a 1.5 safety factor (Dear et al., 2014) which may go some way to overcoming this limitation. Extractions with CuCl2 have previously been found to be a better tool to more appropriately determine liming requirements (Oates and Kamprath, 1983). The current results would seem to corroborate these findings, at least for the acidic soils, if the efficiency of liming practices is directly linked to the exchangeable acidity of Al in these soils. The isotopically exchangeable concentrations and those estimated using relatively simple salt-extractions remain the best estimates of active Al in these soils on short timescales. In CLASS, other larger pools of Al are expected to become a source of mobile Al over years, decades or even centuries (i.e. reactive secondary Fe and Al minerals and aluminosilicates) under particular conditions (Yvanes-Giuliani et al., 2014). Previous research has also shown that the size of the exchangeable pool of Al is time-dependant and that, after long equilibration periods (32 days), the dissolution of Al-bearing minerals and interlayer Al replenished the exchangeable pool (Kleja et al., 2005). The authors raised a valid point when suggesting that perhaps short equilibration times used were not relevant on soil formation/transformation time scales. However, when assessing the potential for Al release into drainage systems and estuaries, exchangeable concentrations obtained after 24 h (as used in this study) appear to be appropriate as transport of contaminants would be occurring quickly. In particular, CLASS in northeastern NSW are frequently subject to flash floods during which surface and subsurface runoff discharges into drains and high flows remove excess water promptly from fields with or without the use of pumps (Macdonald et al., 2007; Wilson et al., 1999). In such circumstances, shorter extraction times (less than 2 days) are more relevant as they are more realistic of field situations, yet leave enough time for the solutions to be in dynamic equilibrium with the solid phase. Irrespective of the methodology used, the exchangeable pool in these soils is clearly important quantitatively. This pool represents Al that is likely to come into solution during episodic rainfall - storms

and floods are common and relatively frequent in the floodplains of north-eastern NSW (Kinsela and Melville, 2004; Wilson et al., 1999). This pool of Al therefore represents a large source of available and mobile Al, which has implications for plant uptake and, even more importantly, for Al transport to surrounding surface waters and ecosystems. Lability is often used to assess the bioavailability and possible toxic effects of metals as uptake by organisms and plants is strongly dependent on the strength of metal bonds to soil constituents (Hamon et al., 2002; Scheifler et al., 2002; Wilson, 2011). In the case of Al, the current results would suggest that there is a high potential for the release of latent acidity through the release of exchangeable Al into surrounding aquatic systems, but also for toxic effects to aquatic organisms, and notably fish as they are particularly sensitive to Al (Wilson, 2011). Significant concentrations of labile Al are present in both soil horizons investigated in this study and, although they are higher at depth, transport of Al through surface runoff and lateral subsurface leaching could be expected in the surface soils when they become saturated (as a result of rainfall and flooding). Most of the samples in this study were collected b 5 m away from agricultural drainage lines; therefore transport of Al to neighbouring aquatic systems could be significant. Vertical capillary transport, however, has been found to be significant (Minh et al., 1998; Nath et al., 2013; Rosicky et al., 2004) and has led to the accumulation of large amounts of amorphous/poorly crystalline aluminium minerals and other Al-containing minerals in the surficial horizons (Yvanes-Giuliani et al., 2014) and could therefore exacerbate off-site transport through surface runoff. 4. Conclusions The E values obtained in this study of 27 CLASS samples were high and ranged from 357 to 3040 mg·kg−1. The exchangeable pool of Al was dominated by that associated with the soil–solid phase. This was, however, not a result of low Al solubility (i.e. Al was 0.14– 31.8 mg·L−1 in solution) but rather a result of the sheer amount of exchangeable Al associated with the solid phase (N98% for almost all samples). The partitioning coefficient (Kd) and E values of Al were both pH dependent, although pH was more significantly correlated to Kd as would be expected. While isotopically exchangeable concentrations remain the most relevant measure of reactive metals in soils, routine determination of these concentrations is impracticable and, in the case of Al, time and

Y.A.M. Yvanes-Giuliani et al. / Science of the Total Environment 542 (2016) 129–135

resource intensive. For this reason, isotopically exchangeable concentrations of Al were compared with the concentrations extracted using KCl and CuCl2 solutions. While both extraction solutions have drawbacks in their application to estimating exchangeable Al in organicrich CLASS, continued use of 1 M KCl with a correction factor of at least ~ 1.7 or ~ 1.1 for that extracted with a sequential 1 M KCl and 0.2 M CuCl2 extraction would seem appropriate for most actual acid sulfate soils. Acknowledgements This research was supported by the Australian Research Council's Linkage Projects funding scheme (LP110100480) with industry partners the Tweed Shire Council, NSW Canegrowers' Association and NSW Milling Cooperative. Yliane Yvanes-Giuliani acknowledges the support of an Australian Institute of Nuclear Science and Engineering (AINSE) postgraduate research award (ALNSTU10105). Dr Richard Collins is the recipient of an Australian Research Council Future Fellowship (FT110100067). The cooperation of landowners Robert Quirk, Ross Hardy and Nicola Stainlay during sample collection is also acknowledged. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2015.10.051. References Ahern, C.R., McElnea, A.E., Sullivan, L.A., 2004. Acid sulfate soils laboratory methods guidelines. Queensland Acid Sulfate Soils Manual 2004. Queensland Department of Natural Resources, Mines and Energy, Indooroopilly, QLD, Australia. ASRIS (2011), Australian Soil Resource Information System. At http://www.asris.csiro.au/ index_ie.html. Aucour, A.M., Tao, F.X., Moreira-Turcq, P., Seyler, P., Sheppard, S., Benedetti, M.F., 2003. The Amazon River: behaviour of metals (Fe, Al, Mn) and dissolved organic matter in the initial mixing at the Rio Negro/Solimões confluence. Chem. Geol. 197, 271–285. Child, D., Elliott, G., Mifsud, C., Smith, A.M., Fink, D., 2000. Sample processing for earth science studies at ANTARES. Nuclear Instruments and Methods in Physics Research, Section B: Beam Interactions with Materials and Atoms 172, 856–860. Claff, S.R., Burton, E.D., Sullivan, L.A., Bush, R.T., 2011. Metal partitioning dynamics during the oxidation and acidification of sulfidic soil. Chem. Geol. 286, 146–157. Collins, R.N., Jones, A.M., Waite, T.D., 2010. Schwertmannite stability in acidified coastal environments. Geochim. Cosmochim. Acta 74, 482–496. Dear, S.E., Ahern, C.R., O'Brien, L.E., Dobos, S.K., McElnea, A.E., Moore, N.G., Watling, K.M., 2014. Queensland Acid Sulfate Soil Technical Manual: Soil Management Guidelines. Department of Science, Information Technology, Innovation and the Arts, Queensland Government, Australia, Brisbane. Degryse, F., Smoulders, E., Parker, D.R., 2009. Partitioning of metals (Cd, Co, Cu, Ni, Pb, Zn) in soils: concepts, methodologies, prediction and applications — a review. Eur. J. Soil Sci. 60, 590–612. Donisa, C., Mocanu, R., Steinnes, E., 2003. Distribution of some major and minor elements between fulvic and humic acid fractions in natural soils. Geoderma 111, 75–84. Fink, D., Hotchkis, M., Hua, Q., Jacobsen, G., Smith, A.M., Zoppi, U., Child, D., Mifsud, C., van der Gaast, H., Williams, A., et al., 2004. The ANTARES AMS facility at ANSTO. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 223–224, 109–115. Fink, D., Smith, A., 2007. An inter-comparison of 10Be and 26Al AMS reference standards and the 10Be half-life. Nuclear Instruments and Methods in Physics Research, Section B: Beam Interactions with Materials and Atoms 259, 600–609. Garcia-Rodeja, E., Novoa, J.C., Pontevedra, X., Martinez-Cortizas, A., Buurman, P., 2004. Aluminium fractionation of European volcanic soils by selective dissolution techniques. Catena 56, 155–183. Hamon, R.E., Bertrand, I., McLaughlin, M.J., 2002. Use and abuse of isotopic exchange data in soil chemistry. Soil Research 40, 1371–1381.

135

Jenkins, B., Morand, D., 2004. An overview of acid-sodic soils in two regions of New South Wales, Australia. In: Singh, B. (Ed.), Supersoil 2004: 3rd Australian New Zealand Soils Conference; University of Sydney, Australia, 5–9 December 2004. Johnston, S.G., Burton, E.D., Bush, R.T., Keene, A.F., Sullivan, L.A., Smith, D., McElnea, A.E., Ahern, C.R., Powell, B., 2010. Abundance and fractionation of Al, Fe and trace metals following tidal inundation of a tropical acid sulfate soil. Appl. Geochem. 25, 323–335. Jones, A.M., Collins, R.N., Waite, T.D., 2011. Mineral species control of aluminum solubility in sulfate-rich acidic waters. Geochim. Cosmochim. Acta 75, 965–977. Kinsela, A.S., Melville, M.D., 2004. Mechanisms of acid sulfate soil oxidation and leaching under sugarcane cropping. Soil Research 42, 569–578. Kinsela, A.S., Collins, R.N., Waite, T.D., 2011. Speciation and transport of arsenic in an acid sulfate soil-dominated catchment, eastern Australia. Chemosphere 82, 879–887. Kleja, D.B., Standring, W., Oughton, D.H., Gustafsson, J.P., Fifield, K., Fraser, A.R., 2005. Assessment of isotopically exchangeable Al in soil materials using 26Al tracer. Geochim. Cosmochim. Acta 69, 5263–5277. Kotzé, W.A.G., Joubert, M., de Villiers, J.F., van der Westhuizen, M., van der Bank, D., 1984. Measurement of exchangeable aluminium in soils and clay minerals by isotopic exchange. South African Journal of Plant and Soil 1, 57–60. Lead, J.R., Wilkinson, K.J., 2006. Aquatic colloids and nanoparticles: current knowledge and future trends. Environ. Chem. 3, 159–171. Lin, C., Melville, M.D., Islam, M.M., Wilson, B.P., Yang, X., van Oploo, P., 1998. Chemical controls on acid discharges from acid sulfate soils under sugarcane cropping in an eastern Australian estuarine floodplain. Environ. Pollut. 103, 269–276. Macdonald, B.C.T., White, I., Åström, M., Keene, A.F., Melville, M.D., ReynoldS, J.K., 2007. Discharge of weathering products from acid sulfate soils after a rainfall event, Tweed River, eastern Australia. Appl. Geochem. 22, 2695–2705. Minh, L.Q., Tuong, T.P., van Mensvoort, M.E.F., Bouma, J., 1998. Soil and water table management effects on aluminum dynamics in an acid sulphate soil in Vietnam. Agric. Ecosyst. Environ. 68, 255–262. Nath, B., Lillicrap, A.M., Ellis, L.C., Boland, D.D., Oldham, C.E., 2013. Hydrological and chemical connectivity dynamics in a groundwater-dependent ecosystem impacted by acid sulfate soils. Water Resour. Res. 49, 441–457. Nishiizumi, K., 2004. Preparation of 26Al AMS standards. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 223–224, 388–392. Oates, K.M., Kamprath, E.J., 1983. Soil acidity and liming: I. Effect of the extracting solution cation and pH on the removal of aluminum from acid soils. Soil Sci. Soc. Am. J. 47, 686–689. Rosicky, M.A., Sullivan, L.A., Slavich, P.G., Hughes, M., 2004. Factors contributing to the acid sulfate soil scalding process in the coastal floodplains of New South Wales, Australia. Soil Research 42, 587–594. Sauvé, S., Norvell, W.A., McBride, M., Hendershot, W., 1999. Speciation and complexation of cadmium in extracted soil solutions. Environ. Sci. Technol. 34, 291–296. Sauvé, S., Manna, S., Turmel, M.-C., Roy, A.G., Courchesne, F., 2003. Solid-solution partitioning of Cd, Cu, Ni, Pb, and Zn in the organic horizons of a forest soil. Environ. Sci. Technol. 37, 5191–5196. Scheifler, R., Schwartz, C., Echevarria, G., de Vaufleury, A., Badot, P.M., Morel, J.L., 2002. “Nonavailable” soil cadmium is bioavailable to snails: evidence from isotopic dilution experiments. Environ. Sci. Technol. 37, 81–86. Soil Survey Staff, 2014. Keys to soil taxonomy. USDA-Natural Resources Conservation Service, 12th ed. (Washington, DC). Sparks, D.L., 2003. 3 — chemistry of soil organic matter. In: Sparks, D.L. (Ed.), Environmental Soil Chemistry, second edition Burlington Academic Press, pp. 75–113. Tavakkoli, E., Donner, E., Juhasz, A., Naidu, R., Lombi, E., 2013. A radio-isotopic dilution technique for functional characterisation of the associations between inorganic contaminants and water-dispersible naturally occurring soil colloids. Environ. Chem. 10, 341–348. Vithana, C.L., Sullivan, L.A., Bush, R.T., Burton, E.D., 2013. Acidity fractions in acid sulfate soils and sediments: contributions of schwertmannite and jarosite. Soil Research 51, 203–214. Walna, B., Spychalski, W., Siepak, J., 2005. Assessment of potentially reactive pools of aluminium in poor forest soils using two methods of fractionation analysis. J. Inorg. Biochem. 99, 1807–1816. Wilson, B.P., White, I., Melville, M.D., 1999. Floodplain hydrology, acid discharge and change in water quality associated with a drained acid sulfate soil. Mar. Freshw. Res. 50, 149–157. Wilson, R.W., 2011. Aluminum. Fish Physiology: Homeostasis and Toxicology of Nonessential Metals. Academic Press, San Diego, CA, USA, pp. 67–123. Yvanes-Giuliani, Y.A.M., Waite, T.D., Collins, R.N., 2014. Exchangeable and secondary mineral reactive pools of aluminium in coastal lowland acid sulfate soils. Sci. Total Environ. 485–486, 232–240.