Effectiveness of Extractants for Bioavailable Phosphorus in Tropical Soils Amended with Sewage Sludge

Urban wastes such as sewage sludge can be an economically viable alternative source for providing macroand micronutrients to plants in tropical conditions. Sewage sludge is normally rich in phosphorus (P), which is present in soils mainly in organic forms, so that it is very important to establish methods for estimating its availability to plants. This study aimed to test three extractants that simulate P-uptake by maize (Zea mays) cropped in plots after 13 consecutive years of fertilization with sewage sludge, in a cycle of fertilized sugarcane (Saccharum L.) amended with sewage sludge and organic compost. Soil samples were collected at depths of 0– 10, 10–20, and 20–40 cm inMarch 2010 from the two experimental areas. Soil P was extracted via ion exchange resin,Mehlich-I, and 0.025MH 2 SO 4 and determined via colorimetry.Maize and sugarcane diagnostic leaves were collected in the experiments, subjected to nitric-perchloric digestion, and the leaf-P content was determined via colorimetry. No significant correlations were found between phosphorus extracted from soils and phosphorus concentrations in diagnostic leaves. Resin extracted larger amounts of P in the short-term experiment, while acidic extractants yielded larger amounts in the long-term experiment.


Introduction
Applying urban wastes such as sewage sludge to agricultural soils is an economically viable alternative for their disposal, and amending soils with sewage sludge benefits crops and poses few environmental risks [1,2].Concerns remain, however, regarding the possible presence of trace elements, toxic organic compounds, or pathogenic microorganisms in sewage sludge [3].
Sewage sludge can increase soil concentrations of certain nutrients, such as phosphorus (P) [4].Current Brazilian legislation [5] mandates that rates of sewage sludge be calculated based on crop nitrogen (N) requirements and does not take into consideration P needs or the fate of the P applied with wastes [6].Because the P : N ratio of waste is typically higher than that required by plants [7], soils amended with sludge usually accumulate P depending on climate and soil properties that command soil P adsorption and desorption.Ippolito et al. [8] showed that fertilization with sewage sludge based on crop nitrogen requirement can add to the soil more P that maize plants can remove.
Studies of long-term soil amendment with sewage sludge of agricultural soils are crucial because in addition to P accumulation soil amendment with sludge can potentially cause incremental and unpredictable effects such as increased levels of organic matter and toxic element accumulation [1].Ideally, a P soil extractant should operate regardless of varying mineralogical, biological, and chemical attributes of soils.Many authors have used parameters of plant growth such as the P content and plant production to test the extent to which planned responses are correlated with P extracted from soils [9].In many cases, however, P extractants commonly used in laboratories are incapable of extracting from the soil quantities of the element correlated with those accumulated in plants.One possible reason for this variation is the fact that many extractants remove inorganic P while underestimating the influence of organic P in plant nutrition.In tropical soils, the organic material has high affinity for Fe and Al oxide surfaces [10], causing P fixation.
Available P contents may be underestimated if the extractant is not able to oxidize the organic fraction resulting from the quick decomposition of the fresh organic material or if organic matter are slightly modified, that are associated with the oxides surface [11].In soils with low P content in which organic P is an important source of nutrient to plant nutrition, the selection of extractants that are able to estimate the available P may not be an easy task [12].As of now, there is no known efficient extractant for organic P present in residues which is capable of estimating the labile organic P fraction that can be absorbed by the plants [13].
The aim of this study was to compare the effectiveness of three extractants widely used for the extraction of phosphorus in tropical soils considering the quantification of (i) available P in soils amended with sewage sludge for 13 consecutive years and the maize P-uptake and (ii) available P in soils amended with a single application of sewage sludge and compost sewage sludge and sugarcane P-uptake.

Materials and Methods
The maize experiment was part of an ongoing study established in the 1997-1998 growing season at Jaboticabal, in the state of São Paulo (SP), Brazil, on a Typic Eutrorthox, with the aim of evaluating the effects of 13 consecutive years of amendments with sewage sludge.The original treatments were a control (no sludge and no fertilizer added), 2.5 t ha −1 , 5 t ha −1 , and 10 t ha −1 of sewage sludge (dry basis).The 5 t ha −1 rate was established to supply the amount of nitrogen (N) required by maize crops, with the assumption that 1/3 of the N in the waste would become available to plants during the first year.Starting in the second year, mineral fertilizer was added each year to supply plant needs in NPK according to soil chemical analysis and the indications of van Raij et al. [14].Starting in the fourth growing season, the 2.5 t ha −1 rate was changed to 20 t ha −1 to induce trace element phytotoxicity in plants.During the 13 years before the soil sampling, the cumulative amounts of sewage sludge added to the soils totaled 65, 130, and 207.5 t ha −1 for the 5, 10, and 20 t ha −1 treatments, respectively.The experiment consists of 60-m 2 plots in randomized complete blocks with four treatments (varying rates of sewage sludge) and five replicates.In 2009, the annual application of sewage sludge took place in December.The sludge had a pH H2O of 5.8 and contained 81.3% water and the following contents (dry matter): 246.7 g kg −1 C, 20.3 g kg −1 P, 24.8 g kg −1 N, 2.4 g kg −1 K, 1.0 g kg −1 Na, 15.9 g kg −1 Ca, and 4.2 g kg −1 Mg.Heavy metal concentrations in the sludge (dry matter) were 5.1 mg kg −1 Cd, 19.6 mg kg −1 Co, 531.5 mg kg −1 Cr, 669.0 mg kg −1 Cu, 34,526.6 mg kg −1 Fe, 320.2 mg kg −1 Mn, 290.7 mg kg −1 Ni, 106.6 mg kg −1 Pb, and 1,398.5 mg kg −1 Zn.The plots with sewage sludge were fertilized with mineral fertilizers in order to receive the same amounts of NPK as the control.Sewage sludge was spread over the entire area and incorporated to 0-10 cm, while the mineral fertilizers were applied in the furrow.
The second field experiment was part of a study underway at Piracicaba (SP), Brazil, on a Typic Hapludalfs in a sugarcane plantation that had been recently harvested (third harvest) at the time of the sampling.Soil acidity was corrected with 2.5 t ha −1 of dolomitic lime (15% water content) and sewage sludge and organic compost were applied on a single occasion on top of the sugarcane straw left behind after harvest.The sewage sludge contained (dry matter) 340 g kg −1 C, 22 g kg −1 N, 11 g kg −1 P, 5 g kg −1 K, 14 g kg −1 Ca, 3 g kg −1 Mg, and 9 g kg −1 S. Metal concentrations were 95 mg kg −1 Cu, 625 mg kg −1 Mn, and 715 mg kg −1 Zn.The composted sewage sludge had (dry matter) 210 g kg −1 C, 18 g kg −1 N, 11 g kg −1 P, 3 g kg −1 K, 20 g kg −1 Ca, 3 g kg −1 Mg, and 4.5 g kg −1 S, as well as 70 mg kg −1 Cu, 475 mg kg −1 Mn, and 527 mg kg −1 Zn.The treatments were complemented with 120 kg ha −1 of K 2 O, since the sanitary sludge had low levels of potassium.The experiment consisted of subdivided plots with three replicates.The treatments (sludge or compost) were (i) no waste applied; (ii) 50% of the rate recommended by Brazil's National Environmental Council (CONAMA) for amendment with sewage sludge applied; (iii) 100% of the recommended rate applied; and (iv) 200% of the recommended rate applied.The CONAMA regulations are based on the assumption that 20% and 10% of organic N in anaerobically digested sewage sludge and organic compost, respectively, will be mineralized.To supply 100 kg ha −1 N for the sugarcane crop, 67 t ha −1 of sludge (wet basis) and 300 t ha −1 of compost (wet basis) were used; the water content of these organic sources was 73% and 60%, respectively.The experiment consisted of 24 plots.Each plot contained five rows of sugarcane measuring 7 m long and spaced 1.4 m between rows.The three central rows and 5 m to either side of them represent the primary sampling area of the plots.
Diagnostic leaves of maize (the leaf opposite to and below the ear) and sugarcane (plus three leaf) cultivated in the 2009-2010 growing season were collected from the two experiments to quantify P concentrations.Plant material was harvested, cleaned with deionized water, placed in paper bags, and dried at 60 ∘ C in a forced-circulation oven until reaching constant mass.Samples were ground in Wiley mills and subsequently subjected to nitric-perchloric digestion (0.5 g of plant material + 5 mL HNO 3 + 1 mL HClO 4 ) adapted from Johnson and Ulrich [15].P concentrations of the extracts were determined by colorimetry (ammonium vanadate-molybdate method).
Soil samples were collected in March 2010 from the two experimental sites: (i) the soil amended with sewage sludge (cultivated with maize) and (ii) the soil amended with sewage sludge composted with straw pruning tree and sludge (cultivated with sugarcane).Twenty subsamples were collected at each of the three soil layers (0-10, 10-20, and 20-40 cm) and mixed to make a composite sample for each depth.In this study, we examined the surface layer (0-10 cm) to observe how higher levels of organic matter influenced the contents of available P to plants.Samples taken for soil fertility analyses are typically collected at a depth of 0-20 cm because fertilization is commonly performed based on tables that consider soil sampling at this depth.Soil samples were air-dried and sifted through a 2 mm (10 mesh) screen.Phosphorus was extracted with ion exchange resin [16], Mehlich-I (0.05 M HCl + 0.0125 M H 2 SO 4 ) [17], and 0.025 M H 2 SO 4 [18].The latter is the extractant most commonly used to predict levels of P available to sugarcane crops in Brazilian soils.Phosphorus content of the extracts was determined by colorimetry.
The results were compared via analysis of variance and correlation tests.For all analyses SAS software was used [19] and significance levels of 1% and 5%.

Phosphorus Concentrations in Maize and Sugarcane
Diagnostic Leaves.The P concentration in maize and sugarcane diagnostic leaves did not differ between different rates of sewage sludge and/or composted sewage sludge (Table 1).The P concentrations in diagnostic leaves were in the adequate range for maize (2-4 g kg −1 ) and sugarcane nutrition (1.5-3.0 g kg −1 ) [14].Sewage sludge application for a long time did not add amounts of the element in the plant above the values recommended for the culture.The element may have been either strongly adsorbed by soil colloids or may not have accumulated at the part of the plant analyzed.Phosphorus bioavailability in soil is directly dependent on its lability in the residues [20,21], and this can explain the absence of increment in the element concentration in the diagnostic leaves.
No significant correlations were found between the P extracted from soils by the extractants and that in the diagnostic leaves of maize and sugarcane suggesting that this sampled part of the plant is not adjusted for the soils we have studied, which have been amended continuously with a waste that is rich in organic matter and phosphorus.Some authors suggest other leaves for the nutritional diagnostic of maize and sugarcane plants [22].The excess of extracted P, where found to occur, may have accumulated in another part of the plant.In Brazil, one special leaf has been used to predict the nutritional state for various crops, to facilitate the recommendation of additional fertilization during the plant growing cycle.However, sometimes there is no correlation between the nutrient absorbed by the plant and its content in the diagnosis leaf, depending on the crop considered.For maize, this is relatively well defined, whereas for sugarcane there is some controversy.The use of diagnostic leaves for nutritional diagnosis of cultures is often used in agricultural soils not receiving organic compounds over a long period, which may explain the absence of correlation of nutrients in soil and plant.
In highly weathered soils that are rich in Fe and Al oxides, P is strongly retained by colloids and the extractants we used may not be capable of removing the available quantity.This complicates fertilization management and may explain the weak correlations observed between P concentrations extracted by common extractants and those absorbed by plants [23].
Sewage sludge possesses a vast gamut of physical and chemical properties.Schroder et al. [24] comment that the average level of total P in the samples of sewage sludge between 1993 and 2005 was 38.6 g kg −1 , ranging between 29.8 and 59.5 g kg −1 .Maguire et al. [25] concluded that forecasting P availability in sewage sludge can be a useful tool for predicting the extractable P in soils treated with this type of residue.They investigated the relationship between processes applied in the treatment of sewage sludge and the bioavailability of P and verified that the extraction pattern is different for soils depending on whether they are amended or not with residue (and amended or not with Fe or Al) and that the need for such extraction tends to reduce in line with the incubation period of soil with the residue.
Kamprath and Watson [26] compiled various correlations between the amount of P extracted by Bray-I solution and the amounts absorbed by plants and found a range of values from −0.10 to 0.94.Their study is a good example of the difficulty of finding a P extractant that is effective across a broad range of plants and/or soils.Similarly, Sharpley et al. [27] compiled data on correlation coefficients and found values ranging from 0.41 to 0.93 between P absorption by plants and P extracted by anionic exchange membrane, Olsen, Bray-I, strips impregnated with Fe oxide, acetic acid, and dilute sulfuric acid.This lack of correlation between the P extracted from the soil and the P absorbed by the plant occurs in the majority of cases, where the extractants had been developed for a system designed for traditional crops, without significant presence of organic material, and this could impose difficulties for the extraction of organic P such that at the stage where the element is fractioned this P content is included in the fraction extracted by HCl.Huang et al.
[28] managed, at least in part, to correct and eliminate the organic P in the HCl fraction, but this value was included in the residual fraction.This demonstrates the complexity of cultivation systems with a significant presence of organic material as far as it relates to a prediction of bioavailable P for the plants, once great variation becomes apparent in the attributes of the soils and residuals used as alternative sources of nutrients.
No extractant provided a reliable representation of the quantity of available P in a wide-range of soils.In addition, the fact that P forms in soils change quickly following the addition of organic waste material containing the element (e.g., when P is added as manure or sludge) raises many questions about the validity of extraction methods [29,30].

Soil Phosphorus Content in the Long-Term Experiment.
At depths of 0-10 and 10-20 cm levels of P-Mehlich-I were higher than those extracted by H 2 SO 4 , which in turn were higher than those extracted by resin.At a depth of 20-40 cm the highest P contents were obtained with P-resin, followed by P-Mehlich-I and P-H 2 SO 4 ( < 0.01) (Table 2).Soils formed in tropical conditions are corrected with limestone to the 0-20 cm layer, having the highest probability of P being bound to calcium forms.The Mehlich-1 extracts the Ca-P fraction and to a lesser extent the fractions of P-Al and Fe-P.These findings could explain the high content extracted by Mehlich-1 in the surface soil and the low extraction capacity in the subsurface layer, in which the concentrations of Fe and Al are higher.The same behavior was observed in tropical soils by Viégas et al. [31] observing high correlation among the P levels extracted by Mehlich 1, with the action of two acids (HCl and H 2 SO 4 ) and the increase in the soil pH, which decreased when the correlation with resin extractant was considered possibly due to the consumption of acidity (Mehlich 1) and exchange anions SO 4 2− .Mehlich-I solution is a mixture of strong acids with a pH of approximately 1.2.As an extractant based on strong acids, it is expected to mostly extract P bound to Ca, and, at lower levels, P bound to Fe and Al.Due to the nature of the acid-Mehlich I, it is expected that in addition to the resin extracted fraction, extracted from the fraction bound to calcium, amounts of nonlabile P can also be extracted, which can cause an overestimating of the levels of P available in soils for agricultural crops.In clayey soils where kaolinite predominates, extractant exhaustion has been observed, leading to an underestimating of P levels and the readsorption of phosphate to the soil colloids after the adsorption of chloride and sulfate ions by sites that were not initially occupied by phosphate [32].
The effect of the rates of sewage sludge was detected by all three extractants studied (Table 3).Even after 13 years of amendment with sludge, these soils showed increasing levels of P with increasing amounts of waste (accumulative effect).Sludge may diminish adsorption of P to soils because organic ions may compete with phosphate for adsorption sites [33].
Paz-Ferreiro et al. [34] carried out an experiment to assess the effect of cattle manure (30, 60, and 90 t ha −1 , dry basis) on nutrient accumulation in the surface layer of an acidic soil with high levels of OM and P and found that levels of P extracted by anion exchange resin were lower than those extracted by Mehlich-III.The P levels extracted by the resin and Mehlich-III methods (determined by colorimetry or ICP) increased as the rates of cattle manure increased.Kidd et al. [35] carried out a greenhouse experiment to assess P availability to maize and two species of wild plants in soils treated with sewage sludge for more than ten years and found that amendment with sludge was associated with increased levels of Olsen P. The P absorption was higher in plants grown on amended soils (possibly luxury consumption by these plants), but the difference was lower for maize plants, suggesting that the excess P added to the soils cultivated with maize posed a higher risk for losses through leaching.
Between 1988 and 2000 Mantovi et al. [36] studied the application of liquid sewage sludge, dehydrated sludge, and composted sewage sludge (5 and 10 t ha −1 yr −1 ) to a silty loam soil in Italy cultivated in rotating crops of wheat, beets, and maize and observed a sharp increase in P contents extracted by Olsen.This finding could be explained not only by the large rates of sludge applied, but also by the increasing availability of the nutrient with time in the treated soils [37].Mantovi et al. [36] reported that wheat grain from amended soils showed higher P concentrations than those of plants treated with mineral fertilizers.Citak and Sonmez [38] also noted a greater concentration of nutrients in the edible part of cabbage (Brassica oleracea) when grown in soil amended with organic residues (bovine and chicken manure and blood meal) when compared to vegetables composted with mineral fertilizers.
Despite the beneficial effects for plants, excess of mobile P represents an environmental risk [39] that can lead to the eutrophication of lakes and streams [40] in sandy soils.McDowell et al. [13] quantified available P to Lolium and Pinus using 12 extractants (five saline and seven resins) for seven soil types and reported that the mineralization of organic P was best characterized by increases in the correlation coefficients between extractable P and P absorbed by plants when organic and inorganic P were considered together.McDowell et al. [13] argued that the easily extractable forms of organic P in soils contribute, in the short term, to the absorption of P by plants and that those forms of P should be considered in common tests to predict the availability of P in soils.
We observed positive linear correlations between the methods we tested, with the highest correlation coefficients between Mehlich-I and H 2 SO 4 , but we did not detect any significant correlation between P levels extracted from soils and P concentration in diagnostic leaves by plants, probably because the extractants we used did not take into account the organic P added to the soils (Table 4).

Phosphorus Content in the
Short-Term Experiment.In the experiment in which soils received a single application of sewage sludge and composted sewage sludge, resin extracted the greatest amounts of P from samples collected at depths of 0-10 and 20-40 cm, followed by Mehlich-I and H 2 SO 4 , which occurred due to less contact time of the residue with the soil, and the predominance of more soluble P, which is poorly retained by the soil and easily extracted by the resin.At a depth of 10-20 cm, P extracted by resin and Mehlich-I were similar and greater than those extracted by H 2 SO 4 (Table 2).Variation in the rates of sludge and composted sewage sludge had no effect on the extracted phosphorus contents.This was mostly due not only to the short duration of the experiment (i.e., only three months between waste application and soil sampling), but also to the fact that waste was not incorporated into the soil (Table 5).After the initial increases, the pool of P extracted by laboratory tests declined gradually with time (due to the slow precipitation of mineral P).Lucero et al. [41], Reddy et al. [42], and Griffin et al. [43] reported differences in the amounts of P extracted from soils when the element was applied as inorganic fertilizers (KH 2 PO 4 , NH 4 HPO 4 ) and manure.The effects of these differences on the availability of P have been attributed to variation in the speciation of the element, pH, the solubility of the applied sources, the quality and quantity of organic matter, and the presence of Al, Fe, and Ca [44,45].
In two years Chiba et al. [46] applied rates of sewage sludge (with or without incorporation and at amounts based on N contents) to an Ultisol planted with sugarcane and showed that P contents were not altered by amendment.This result probably reflects that sewage sludge in that experiment remained on the surface of the soil, largely out of contact with the soil biota capable of breaking it down.Both thermal drying and exposure to sun tend to slow the degradation rate of sewage sludge, due to a reduction in water content and a consequent increase in the stability of organic compounds.This may also explain why different rates of amendment had no effect on levels of biologically available P in this short-term experiment, since the waste was placed on the soil surface in a single application.Any change in the phase or decomposition of organic material attributable to heat or dryness can lead to the formation of P compounds with metals and or minerals, although more thorough studies may be necessary to elucidate the mechanisms behind these alterations [47].In a study of the sequential extraction of P in sewage sludge and poultry litter, He et al. [47] concluded that the solubility of P contained in the sewage sludge was controlled mainly by the metals Al, Mn, and Zn, with Al exerting the greatest influence due to its greater concentration, which should be a relevant consideration when working with soils that are highly weathered in moist tropical zones.
We observed positive linear correlations with a high correlation coefficient between extractants, but not between extractant and P levels in diagnostic leaves (Table 6).The strongest correlations were observed between acidic extractants, which provided the same effect on the release of the organic P on the soil surface.The effects of changes caused by the application of sewage sludge vary not only with the duration of application, but also with variation in speciation of the element, in pH, in the solubility of the applied source, in the quantity and quality of organic matter, in the presence of Al, Fe, and Ca, and other factors.

Conclusion
(i) No significant correlations were observed between P extracted from soils by the three extractants and P concentration in the diagnostic leaves of maize and sugarcane plants grown in those soils, suggesting that (1) the plant parts sampled were not appropriate for these evaluating the availability of P in tropical soils that are amended with sewage sludge or composted sewage sludge or (2) the extractants used were developed for conventional crops, and they do not appear to be effective in systems where there is an elevated level of organic material.This underlines the importance to us of the supply of nutrient to plants, especially in soils where the level of the element is low.
(ii) Resin extracted greater amounts of P from soils in the short-term experiment (P in more labile forms), while the acidic extractants removed greater amounts of the nutrient in the long-term experiment (P in less bioavailable fractions).
(iii) An alternative for improving the extractants with the aim of rendering them useable in tillage systems where there is a presence of organic material is to develop potentially available mechanisms for extracting organic P or quantify the P in the residuals as well.

Table 1 :
Phosphorus concentrations in diagnostic leaves of maize plants grown in a Typic Eutrorthox amended annually with sewage sludge for 13 consecutive years and in diagnostic leaves of sugarcane plants grown in a Typic Hapludalf amended once with sewage sludge and sludge compost ( = 4).

Table 2 :
Mean extracted phosphorus by Mehlich-I, H 2 SO 4 , and resin in the two experiments described in the text: one with amendment of sewage sludge and the other with the amendment of sewage sludge and composted sewage sludge ( = 4).

Table 3 :
Mean extracted phosphorus by Mehlich-I, H 2 SO 4 , and resin from samples collected at three depths from soils treated with varying rates of sewage sludge, in the long-term experiment ( = 4).

Table 4 :
Correlation matrix for phosphorus extracted from soils at depths of 0-10, 10-20, and 20-40 cm and phosphorus levels in diagnostic leaves of maize plants grown in the long-term experiment ( = 4).

Table 5 :
Mean of phosphorus concentrations extracted by Mehlich-I, H 2 SO 4 , and resin from samples collected at three depths from soils treated with varying rates of sewage sludge and composted sewage sludge, in the short-term experiment ( = 4).Values with the same letter within a column in the same depth are not different (Tukey,  < 0.05).

Table 6 :
Correlation matrix for phosphorus extracted from soils at depths of 0-10, 10-20, and 20-40 cm and phosphorus levels in diagnostic leaves of sugarcane plants grown in those soils, for the short-term experiment ( = 4).