SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 11 April 2005
Published in Soil Sci Soc Am J 69:607-615 (2005)
DOI: 10.2136/sssaj2004.0156
© 2005 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Saavedra, C.
Right arrow Articles by Delgado, A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Saavedra, C.
Right arrow Articles by Delgado, A.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Saavedra, C.
Right arrow Articles by Delgado, A.
Related Collections
Right arrow Nutrient Cycling
Right arrow Phosphorus
Right arrow Soil Chemistry

Soil Chemistry

Phosphorus Fractions and Release Patterns in Typical Mediterranean Soils

Concepción Saavedra and Antonio Delgado*

Departamento de Ciencias Agroforestales, EUITA, Universidad de Sevilla, Ctra. Utrera Km 1, 41013 Sevilla, Spain

* Corresponding author (adelgado{at}us.es)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Phosphorus forms dictate the P release potential of soils in agronomic and environmental terms. The main purposes of this work were to study P forms by sequential chemical fractionation in soils typical of Mediterranean areas, establishing correlations between P fractions and soil properties, and identify the relationships of specific P fractions to P release potential as determined using P sinks (anion-exchange resins). To this end, three different fractionation methods were used. A comparison of the results obtained with the three sequential fractionation schemes provides useful information about P forms in representative Mediterranean soils, allowing the distinction of P fractions which include: (i) the more labile P forms (essentially adsorbed P), (ii) most of pedogenic Ca phosphates, (iii) most of low soluble pedogenic Ca phosphates, (iv) lithogenic Ca phosphates, (v) P occluded in poorly crystalline Fe oxides, and (vi) P occluded in crystalline Fe oxides. The ratio of the P fraction, which includes the more labile P forms (essentially adsorbed) to combined non-organic P fractions was negatively correlated with soil pH (r = –0.83, P < 0.001) and positively correlated with the portion of combined Fe fractions related to poorly crystalline oxides (r = 0.84, P < 0.001). The greatest of P desorbable from resins (Qmax, in the Johnson–Mehl equation) was related mainly to the combined P fractions including adsorbed P and precipitated Ca phosphates in the studied soils. The amount of P desorbed at 1 h (Q1h) accounted for a sizeable fraction of Qmax (between 29 and 59 %), the Q1h/Qmax ratio being positively correlated with the portion of combined non-organic P fractions related to adsorbed P (r = 0.66; P < 0.01), and negatively correlated with the portion of combined P fractions related to low soluble pedogenic Ca phosphates (r = –0.67; P < 0.01).

Abbreviations: Cit, citrate • CA, citrate ascorbate • CB, citrate bicarbonate • CBD, citrate bicarbonate dithionite • EDTA, ethylene dinitrilo tetracetic acid • mG, Golterman fractionation method as modified by Díaz-Espejo et al. (1999) • OAc, acetate • OP, organic P • OS, Olsen and Sommers (1982) sequential fractionation method • R, Ruíz et al. (1997) fractionation method • TP, total P • WP, water-extractable P


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
PHOSPHORUS FERTILIZER reactions determine the dominant potentially labile P forms in soils and hence the proportion of soil P that can be released to the soil solution or water reservoirs. This accounts for the agronomic or environmental significance of P in agricultural soils. In broad terms, P binding to soil particles results from surface adsorption or precipitation. The relative contribution of these two processes is related to soil properties and fertilizer management practices (Afif et al., 1993), precipitation of Ca phosphates is widely assumed to be the dominant reaction in calcareous soils at P concentrations above 10–4.5 M (Castro and Torrent, 1998; Tunesi et al., 1999).

Because of its low content in soil, the precipitated phase in soil resulting from P fertilizer addition has traditionally been determined with indirect methods based on phase diagrams (Fixen et al., 1983; Havlin and Westfall, 1984). Only 31P NMR spectroscopy has proved an effective direct method for studying P speciation in acid (Al and Ca related P, Lookman et al., 1996) and calcareous soils (Ca related P, Delgado et al., 2000; Delgado et al., 2002). Iron related P (Fe phosphates or P bound by Fe oxides), however, cannot be examined by this technique. Sequential chemical extraction methods have often been used to study the nature of P forms in soils and sediments. These methods are based on the selective extraction of operationally defined P fractions by using single extractants in a sequential manner. The discrimination of Ca related P (Ca phosphates) from Fe related P has been achieved in some cases (Olsen and Sommers, 1982; Golterman and Bouman, 1988). However, this information may be inadequate with a view to predicting the potential release of P from soils (e.g., no distinction is usually made between soluble and low soluble Ca phosphates or between adsorbed and occluded P in oxides). In calcareous soils, part of extracted P using NaOH (mostly adsorbed P, Olsen and Sommers, 1982) is readsorbed by calcite and then released in the following step (citrate bicarbonate), making it difficult to discriminate adsorbed P from P precipitated in soluble Ca phosphates (Williams et al., 1971). Although this shortcoming is shared by the sequential method of Ruíz et al. (1997), this method does distinguish low soluble pedogenic Ca phosphates from lithogenic fluorapatite, and P occluded in poorly crystalline Fe oxides from that occluded in crystalline Fe oxides.

In general, fractionation methods do not quantify the organic fraction by means of a specific extractant; rather, the fraction remaining after inorganic constituents have been removed is taken to be OP (Barbanti et al., 1994). To avoid underestimating OP as result of hydrolysis during removal of inorganic forms in sediments, Golterman (1976) proposed the use of chelating agents at the same pH as the sediment to characterize inorganic P forms (Fe bound P and Ca bound P).

The purposes of this work were to study: (i) P forms in typical Mediterranean soils using three different sequential fractionation methods, (ii) the differences in dominant P fractions between soil types and how these fractions are determined by soil properties, and (iii) the relationships between the P release potential, estimated by means of P sinks (resins), and P fractions in these soils.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soils
Seventeen soils from the Guadalquivir Valley, South Spain (Palma del Río, 37°43' N Lat., 5°13' W Long.), were selected in such a way as to include the most typical for Mediterranean areas as per the soil taxonomy (Soil Survey Staff, 1998). The soils were used for cultivation of typical Mediterranean crops (wheat [Triticum aestivum L.], sunflower [Helianthus annuus L.], beans [Phaseolus vulgaris L.], pea [Pisum sativum L.], and chickpea [Cicer arietinum L.]) under rainfed agriculture. No organic amendments (not even crop residues) had been applied to the soils over the last 30 yr. The typical P fertilizer rates for such a period ranged from 20 to 30 kg P ha–1 yr–1. Samples were collected from the Ap horizon (0–30 cm), air-dried and ground to pass through a 2-mm sieve before analysis.

General Soil Properties
Particle-size analyses were performed by using the pipette method following treatment with an HOAc–NaOAc buffer at pH 4.75 to remove carbonates (Gee and Bauder, 1986). Organic matter was determined by dichromate oxidation (Walkley and Black, 1934), and the cation exchange capacity (CEC) by using 1 M NH4OAc buffered at pH 7 (Sumner and Miller, 1996). Following extraction with 1 M NH4OAc (three consecutive washes), exchangeable Na, K, and Mg were determined, using flame photometry for K and Na, and atomic absorption spectroscopy for Mg. Owing to the presence of CaCO3, exchangeable Ca had to be estimated as the difference between CEC and the combination of all other exchangeable cations. The total CaCO3 equivalent (CCE) was determined from the weight loss on treatment with 6 M HCl. The pH and electrolytic conductivity were measured in water (1:2.5 and 1:5 soil/water ratio, respectively).

Phosphorus availability index was determined according to Olsen et al. (1954); the active carbon treatment was avoided and centrifugation (1000 x g) was used instead of filtering. Total P was determined using the molybdenum blue colorimetric method (Murphy and Riley, 1962) following nitric-perchloric digestion in a Kjeldahl block. Organic P was determined as the increase in 0.5 M H2SO4 extractable P after ashing at 550°C for 2 h (Kuo, 1996). Finally, water-extractable P (WP) was quantified by extraction of P at a soil/water ratio of 1:10 for 1 h.

Phosphorus Fractionation
Sequential soil P fractionation was performed following three different schemes. One was the classical scheme of Chang and Jackson (1957) as modified by Olsen and Sommers (1982)(OS fractionation method). The process involves four consecutive extractions with

  1. 0.1 M NaOH + 1 M NaCl (NaOH-P), which releases adsorbed P, P precipitated as Fe and Al phosphates, and P bound by Fe and Al organic complexes;
  2. 0.27 M Na citrate + 0.11 M NaHCO3 (CB-P), which extracts adsorbed P and highly soluble Ca phosphates, partly precipitated or adsorbed on calcite after the NaOH extraction in calcareous soils;
  3. 0.27 M Na citrate + 0.11 M NaHCO3 + 2% (w/v) Na dithionite (CBD), which dissolves Fe oxides releasing P adsorbed on high energy sites plus P occluded within the oxides; and
  4. 1 M HCl (HCl-P), which dissolves low soluble Ca-phosphates (mainly hydroxyapatite + fluorapatite).

The second scheme was that of Ruíz et al. (1997)(R fractionation method). It expands the previous method with some steps intended to discriminate P related to soluble Ca phosphates from that related to low soluble Ca phosphates, and P occluded within poorly crystalline from that occluded in crystalline Fe oxides. The process involves eight extractions using

  1. 0.1 M NaOH + 1 M NaCl (NaOH-P), which releases adsorbed P, P precipitated as Fe and Al phosphates, and P bound by Fe and Al organic complexes;
  2. 0.27 M Na citrate + 0.11 M NaHCO3 (CB-P), which extracts adsorbed P and highly soluble Ca-phosphates, partly precipitated or adsorbed on calcite after the NaOH extraction in calcareous soils;
  3. 0.25 M Na citrate at pH 6, and
  4. 0.2 M Na citrate pH 6 (Cit-P), which release pedogenic Ca phosphates not dissolved by CB;
  5. 0.2 M Na citrate + 0.05 M ascorbate at pH 6 (CA-P, "mild reductant soluble P"), which releases mostly P occluded in poorly crystalline Fe oxides;
  6. 0.27 M Na citrate + 0.11 M NaHCO3 + 2% Na dithionite (CBD-P, "reductant soluble P"), which releases P occluded in crystalline Fe oxides;
  7. 1 M NaOAc buffered at pH 4 (OAc-P), which releases residual pedogenic Ca phosphates previously not dissolved by citrate; and
  8. 1 M HCl, which dissolves most lithogenic apatite.

Occluded P includes P on sites that are not accessible to the outer solution (Torrent et al., 1990). Part of the "mild reductant" and "reductant" soluble P is not strictly occluded, and probably includes phosphate strongly adsorbed on Fe oxides and P related to organically complexed Fe (Ruíz et al., 1997). From an operational point of view, occluded P is considered the fraction released by the action of reductants after removing most of P adsorbed to Fe oxides and Fe-rich phosphate particles in previous extraction steps. The distinction of P occluded in poorly crystalline Fe oxides from that occluded in better crystalline oxides is achieved by using two different reductants sequentially: citrate ascorbate (CA), which dissolves poorly crystalline oxides more specifically than oxalate (Reyes and Torrent, 1997), and citrate bicarbonate dithionite (CBD), which dissolves Fe oxides (Mehra and Jackson, 1960).

The third fractionation scheme involves the use of chelating compounds to determine P related to Fe oxides and Ca-phosphates. It is a modification of the sediment P fractionation scheme of Golterman (1996) developed by Díaz-Espejo et al. (1999)(the mG fractionation method) and involves five fractionation steps:

  1. 0.05 M Ca-EDTA + 1% (w/v) Na dithionite (Ca-EDTA-P), for 1 h, to release Fe related P;
  2. 0.1 M Na2–EDTA (Na-EDTA-P), for 1.5 h, to release Ca bound P;
  3. 0.5 M H2SO4 (H2SO4–P), for 0.5 h;
  4. 2 M NaOH. for 2 h, to release OP alkali extractable; and
  5. 0.5 M H2SO4 + 8% (w/v) K2S2O8, to release residual OP.

The extraction time in the OS and the R fractionation schemes was 16 h except for the second extraction with Na-citrate (8 h), and the HCl extraction (1 h). Also, the CBD step in both schemes was performed at room temperature for 16 h, as recommended by Solís and Torrent (1989); this proved as effective as extracting at 363 K (Olsen and Sommers, 1982). The soil/extractant ratio was 1:40 in all fractionation methods. All extractions were performed in duplicate using polyethylene flasks at 298 K, except the extraction with 2 M NaOH in the modified Golterman scheme, which was performed at 363 K ("hot" NaOH extractable OP). After all extractions, the suspension was centrifuged at 1000 x g and the supernatant analyzed for molybdate reactive P (Murphy and Riley, 1962). The NaOH extracts in the first and second schemes were also analyzed for TP (nitric-perchloric digestion) to estimate the "cold" NaOH extractable OP (NaOH-OP). Total P in all the extracts obtained using the modified Golterman scheme was determined after sulfuric-persulphate digestion (Díaz-Espejo et al., 1999). Differences between molybdate reactive P and TP in the extracts can be mainly ascribed to OP (Golterman et al., 1998).

Iron was determined in the CB, Cit, CA, and CBD extracts of the Ruíz et al. (1997) scheme by using atomic absorption spectrometry. The combined Fe fractions can be assumed to represent Fe related to oxides. Citrate ascorbate extractable Fe (CA-Fe) was ascribed to poorly crystalline Fe oxides, and CBD extractable Fe (CBD-Fe) to crystalline Fe oxides (Reyes and Torrent, 1997; Ruíz et al., 1997).

Phosphorus Release Kinetics
The kinetics of P release was studied by using an anion-exchange resin saturated with Cl as P sink. Nylon bags holding 2.2 g of 20-50 mesh Dowex 1 x 4 resin (The DOW Chemical Co., Midland, MI) were placed in 60-mL polyethylene flasks containing 40 mL of 0.002 M CaCl2 and 1 g of soil. Flasks were placed in a reciprocating shaker oscillating at 1.2 Hz. At 0.25, 0.5, 1, 2, 5, 9 h, and at 1, 2, 4, 8, 16, 30, and 60 d, the resin bags were recovered and replaced with fresh Cl-resin. The resin was eluted twice with 0.5 M HCl to remove adsorbed P. Phosphorus in the eluate was determined using the above-described method. The experiment was performed in duplicate at 298 K.

Statistical Analyses
Phosphorus release data was fitted to time using the Johnson-Mehl equation (Goss, 1987):

where Q is the amount of P released at a time t, Qmax is the maximum amount of P that can be desorbed, and k and n are constants. Data were fitted using the Simplex procedure (CoHort software, Minneapolis, MN). Means were compared using Student's t test as implemented in Costat (CoHort software, Minneapolis, MN).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Soil Properties
Table 1 shows the classification of studied soils according to Soil Taxonomy (Soil Survey Staff, 1998) and their most salient properties. Alfisols (Haploxeralfs and Palexeralfs groups) were Aquic or Calcic. Although the surface horizon of these soils must originally have been free of carbonates, some samples contained significant amounts of these compounds as a result of erosion and mixing by effect of tillage (deep plowing). On average, Palexeralfs had the lowest clay content, cation exchange capacity, and carbonate content in the surface horizon of all soils. Inceptisols had a Vertic character (Vertic Calcixerepts). Aquic Haploxeralfs and Vertic Calcixerepts had the highest clay content. The sand content was negatively correlated with pH (r = –0.75, P < 0.001) and the carbonate content (r = –0.72, P < 0.01).


View this table:
[in this window]
[in a new window]
  		Table 1. General properties of the soils.{dagger}

 
Olsen P (Olsen et al., 1954), which constitutes a reasonable estimate of available P for plants in these types of soil (Delgado and Torrent, 1997), ranged from 18 to 63 mg kg–1, indicating that plant response to P fertilizer is not probable in these soils (Delgado and Torrent, 2000). Water-extractable P (WP) at a soil/water ratio of 1:10 can be considered an estimate of P release potential from soil to water (Torrent and Delgado, 2001). This index ranged from 1.13 to 15.1 mg kg–1. Olsen P and WP were highly correlated (r = 0.83, P < 0.001). Neither Olsen P nor WP was correlated with TP in soils.

Fractionation Results
The combined P fractions (non-organic) in the three schemes accounted for similar amounts of P (Fig. 1). The combined amounts of P extracted with mG (Ca-EDTA-P + Na-EDTA-P + H2SO4–P), however, were slightly greater: the combined P fractions obtained with OS and R accounted for 94 and 76%, on average, respectively, of the P fractions obtained with mG. One possible explanation for the smaller amounts of P obtained with R method may be increased hydrolysis of OP in the acid extractions of the OS and mG schemes (Barbanti et al., 1994). Hydrolysis of OP by HCl in R can be expected to be less marked as a result of the previous release of part of OP by citrate and citrate-ascorbate, which is not hydrolyzed, and therefore cannot contribute to molybdate reactive P through hydrolysis in the later acid extraction. In nine of the soils, this is supported by the fact that the combined Cit-P, AcO-P, and HCl-P fractions (mainly ascribed to low soluble Ca phosphates) in R were smaller than HCl-P in OS (Table 2).



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 1. (a) Relationship between the combined inorganic P fractions according to Golterman as modified by Díaz-Espejo et al. (1999) and the combined P fractions according to the other two fractionation methods used: Olsen and Sommers (1982), and Ruíz et al. (1997). (b) Relationship between the sum of extracted P in the first three steps in the fractionation method of Ruíz et al. (1997) and the sum of extracted P in the first two step in the modified Golterman fractionation method (Díaz-Espejo et al., 1999).

 

View this table:
[in this window]
[in a new window]
  		Table 2. Phosphorus fractions according to the three fractionation schemes: Ruíz et al. (R; 1997), Olsen and Sommers (OS; 1982), and Golterman as modified by Díaz-Espejo et al. (mG; 1999).

 
In broad terms, Ca-EDTA-P (mG) was smaller than CBD-P (mostly occluded P) in OS (Table 2). Because Ca-EDTA-P must include Fe related P (adsorbed, occluded, and Fe phosphates), one can assume that this extractant is not so effective in extracting Fe related P. This can be essentially ascribed to a lower concentration of Na dithionite added to Ca-EDTA and shorter extraction time relative to the OS and R schemes, which results in incomplete dissolution of Fe oxides. According to Barbanti et al. (1994), complete extraction of Fe related P using Ca-EDTA + dithionite probably requires repeated extractions.

The sum of the amounts extracted in the three first steps of R was similar to the sum of the first two steps of mG (Fig. 1) supporting the hypothesis of a reduced recovery of P occluded in Fe oxides by Ca-EDTA + dithionite. In these steps of both fractionation schemes it can be assumed that it is extracted most of adsorbed P, Fe and Al phosphates, soluble Ca phosphates, and low soluble pedogenic Ca phosphates. The amount of P extracted with Ca-EDTA + dithionite was highly correlated with the sum NaOH-P + CB-P in R (r = 0.82, P < 0.001), and accounted for 70%, on average, of this sum, which includes the more labile P fractions; also, the sum is usually highly correlated with availability indexes (Domínguez et al., 2001). In all studied soils, Ca-EDTA-P was the fraction exhibiting the strongest correlation with the availability index (Olsen P) and the amount of P released to water (WP, Fig. 2). The Ca-EDTA-P can therefore be assumed to include the more labile P forms (essentially adsorbed P, but no soluble precipitated Ca phosphates, which are not dissolved by this extractant). On the other hand, Cit-P in R scheme was highly correlated with Na-EDTA-P (r = 0.88, P < 0.001), and accounted for about 80% of the amount provided by this extraction. Probably, Na2–EDTA dissolves not only the low soluble Ca phosphates that are also dissolved by citrate, but also the soluble Ca phosphates dissolved by CB. Thus, the difference between Na-EDTA-P (mG scheme) and Cit-P (R scheme) can be supposed an estimate of the amount of P related to soluble Ca phosphates. The efficiency of Na2–EDTA in dissolving lithogenic Ca phosphates is supposed to be limited (Barbanti et al., 1994). Thus, at least a fraction of low soluble Ca phosphates (lithogenic apatite) is likely to be dissolved by H2SO4 in mG. Also, Fe related P not released in previous steps must contribute to extracted P in this last step. The nature of reactive P released by H2SO4 in mG must therefore be complex, and probably includes the result of the hydrolysis of some OP.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 2. Olsen P (Olsen et al., 1954) and water-extractable P (WP) as related to P extracted by Ca-EDTA + dithionite (Ca-EDTA-P, Golterman as modified by Díaz-Espejo et al., 1999).

 
Most extractable Fe (reducible by sequential extraction according to Ruíz et al., 1997) was related to crystalline Fe oxides (CBD-Fe); by contrast, the amount of Fe that can be ascribed to poorly crystalline Fe oxides (CA-Fe) accounted for only 4% on average of the combined Fe fractions (Table 3). Phosphorus extractable by CBD in the R scheme was highly correlated with the amount of CBD extractable Fe (Fig. 3); this suggests that occlusion of P in crystalline Fe oxides was related to the amount of this soil component formed through weathering. Citrate ascorbate-extractable P was not related to the amount of Fe extracted in this step and the P/Fe mole ratio in CA extracts was negatively correlated with the fraction of Fe extracted by CA (Fig. 3). This suggests that the amount of P occluded in poorly crystalline oxides was mainly related to the P concentration in solution rather than to the rate of oxide precipitation in soil. The P/Fe mole ratio of crystalline Fe oxides (Table 3) is much higher than in unfertilized horizons of similar soils of the area (Smeck et al., 1994), thus indicating some occlusion of P applied in fertilization. The high P/Fe mole ratio in CA extracts relative to CBD extracts (Table 3) indicates that poorly crystalline oxides contain more occluded P per unit mass of Fe than crystalline oxides and suggests that P fertilizer can be preferentially occluded in poorly crystalline Fe oxides (Ruíz et al., 1997; Domínguez et al., 2001).


View this table:
[in this window]
[in a new window]
  		Table 3. Iron fractions in soils according to Ruíz et al. (1997), and P/Fe mole ratio in Fe oxides.{dagger}

 


View larger version (16K):
[in this window]
[in a new window]
  		Fig. 3. (a) Citrate bicarbonate dithionite P (CBD-P) as related to the amount of Fe extracted in the same fractionation step (CBD-Fe) (Ruíz et al., 1997). (b) The P/Fe mole ratio in CA extracts as related to the ratio CA-Fe/combined Fe fractions.

 
Organic P determined as the increase in inorganic P on ashing (Kuo, 1996) was significantly smaller than the combination of OP fractions determined as per the mG scheme (Table 1 and 2; means significantly different, P < 0.05, n = 17). In broad terms, TP determined by the nitric-perchloric method (456 ± 181) was slightly smaller than the combined P fractions (organic and inorganic) provide by mG (634 ± 221), both estimates being highly correlated (r = 0.84, P < 0.001). This suggests that the combination of P fractions in mG can provide an accurate estimate for TP in soil, and also that the ashing procedure clearly underestimates the amount of OP. On average, the combined OP fractions as per mG accounted for 47% of the sum of both organic and non-organic P fractions in such a scheme, and for 60% of TP determined after nitric-perchloric digestion (Tables 1 and 2). This exposes the significance of the OP pool in typical Mediterranean soils, even though the organic matter content was lower than 20 g kg–1 in all the cases (Table 1). Most of this OP is probably related to non-labile forms such as inositol phosphates (phytates) or humic compounds (Condron et al., 1985; Golterman et al., 1998).

Organic P extracted by 0.1 M NaOH at room temperature (NaOH-OP; Table 2) accounted for 11% on average of OP in the soils based on the mG scheme. This fraction probably includes labile and moderately labile OP in soil (Daroub et al., 2000) and part of inositol phosphates (phytase hydrolyzable) since this is the dominant OP form in most agricultural soils (Condron et al., 1985). No significant relationships were observed between OP fractions and organic matter in soil.

Phosphorus Fractions as Related to Soils Properties
The highest Olsen P/TP ratio (0.12) and WP/TP ratio (0.02) were found in Aquic Palexeralfs. In these soils, the combination of NaOH-P and CB-P (R and OS schemes) amounted, on average, to one half of the combined P fractions (Table 2). Also these soils showed the highest ratio of Ca-EDTA-P, which seems to include the more labile P forms (mainly adsorbed P on Fe oxides), to combined non-organic P fractions (mG scheme) (0.40 ± 0.09, Table 2). This accounts for the fact that the Olsen P to TP ratio was the highest in Aquic Palexeralfs as these fractions must include the more labile P forms in soils (Delgado et al., 2000). In the other soils types, the significance of the P fractions related to pedogenic Ca phosphates (Cit-P in the R scheme and Na-EDTA-P in the mG scheme) (Table 2) accounts for substantial precipitation of Ca phosphates as a component of P reactions in soils.

The dominant P fractions are dictated by the soil properties affecting P dynamics. The ratio of Ca-EDTA-P to combined non-organic P fractions (mG scheme) was negatively correlated with soil pH (Fig. 4). The portion of the combined P fractions obtained with R that corresponded to low soluble pedogenic Ca phosphates (Cit-P) was, on the other hand, positively correlated with soil pH (Fig. 4). This suggests increased precipitation of Ca phosphates at increased pH in the soil. The more sandy soils exhibited, as it was stated above, the lowest CaCO3 contents and pH, suggesting P adsorption on Fe oxides being favored over precipitation of Ca phosphates in these soils.



View larger version (14K):
[in this window]
[in a new window]
  		Fig. 4. (a) Relationship between the ratio of P extracted by CA-EDTA + dithionite (Ca-EDTA-P) to the combined of P fractions (Díaz-Espejo et al., 1999) and soil pH. (b) Relationship between the ratio of P extracted by citrate (Cit-P) to the combined P fractions (Ruíz et al., 1997) and soil pH.

 
The ratio of the fraction related to adsorbed P (Ca-EDTA-P) to combined non-organic P fractions (mG scheme) was positively correlated with the portion of combined Fe fractions that corresponded to poorly crystalline Fe oxides (r = 0.84, P < 0.001). No significant correlation was found with the proportion of Fe in crystalline oxides. This indicates the significance of poorly crystalline Fe oxides contributing to P adsorption in studied soils. The ratio of CA-Fe to combined Fe fractions was negatively related to soil pH (r = 0.64, P < 0.01) and positively to sand content in the soil (r = 0.63, P < 0.01), contributing to explain the relevance of adsorbed P in soils with lower pH and higher sand content. In Aquic soils from South Spain, it has been observed that the proportion of Fe related to poorly crystalline Fe oxides, probably formed by rapid oxidation of Fe2+ after flooding, was higher than in non Aquic soils of the area, being these oxides the main surface contributing to P adsorption in these soils (Domínguez et al., 2001). In agreement with this finding, the Aquic Palexeralfs, which were the more sandy soils, exhibited the highest CA-Fe to combined Fe fractions ratio (0.063 ± 0.012).

Phosphorus Release Potential as Related to Phosphorus Fractions in Soils
Phosphorus sinks have been widely used to study P release potentials in soils. Sinks mimic P desorption from soil particles to a dilute electrolyte containing a low P concentration. The Qmax values estimated from the Johnson–Mehl equation by using resins as P sinks have been found to exceed isotopic exchangeable P, this suggests that sinks probably promote dissolution of metal phosphates (Delgado and Torrent, 2000). In fact, Qmax (Table 4) was related mainly to the combination of P fractions including adsorbed P and precipitated Ca phosphates in the studied soils (NaOH-P + CB-P + Cit-P, r = 0.67, P < 0.01; Ca-EDTA-P + Na-EDTA-P, r = 0.66, P < 0.01). Thus, even low soluble Ca phosphates will likely contribute to the amount of P desorbed in the long term to resins. Phosphorus desorbed in the short term (1 h, 1 d), was related mainly to P fractions including adsorbed P and soluble precipitated P; the amounts extracted at 1 h from resins was correlated with NaOH-P + CB-P (r = 0.54, P < 0.05) and Ca-EDTA-P (r = 0.51, P < 0.05); at 1 d, the amounts extracted were only correlated with NaOH-P + CB-P (r = 0.55, P < 0.05).


View this table:
[in this window]
[in a new window]
  		Table 4. Parameters of the Johnson–Mehl equation used to fit kinetic data (resin extraction) and amounts extracted at 1 h and 1 d.{dagger}

 
Desorbed P at 1 h (Q1h) accounted for a substantial fraction of Qmax (between 29 and 59%, Table 4). The Q1h to Qmax ratio was positively correlated with the portion of combined non-organic P fractions related to adsorbed P in mG scheme (Ca-EDTA-P) (Fig. 5). By contrast, the Q1h to Qmax ratio was negatively correlated with the fraction of combined P fractions in R scheme related to low soluble pedogenic Ca phosphates (Cit-P) (Fig. 5). Adsorbed P must therefore favor fast desorption from resins, in contrast to low soluble Ca phosphates, which are probably involved in slow P desorption from soil. This is consistent with previous findings of Delgado and Torrent (2000), who found an increased proportion of rapidly desorbed P in soils with an increased proportion of adsorbed P relative to soils with a higher proportion of precipitated P as Ca phosphates.



View larger version (15K):
[in this window]
[in a new window]
  		Fig. 5. (a) Ratio of P extracted by resins at 1 h (Q1h) to the maximum extractable P (Qmax in the Johnson–Mehl equation) as related to the ratio of P extracted by CA-EDTA + dithionite (Ca-EDTA-P) to the combined of P fractions (Díaz-Espejo et al., 1999) and to (b) the ratio of P extracted by citrate (Cit-P) to the combined P fractions (Ruíz et al., 1997).

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The three sequential fractionation schemes used allow the distinction of P fractions including: the more labile P forms (essentially adsorbed P, Ca-EDTA-P), non-lithogenic Ca phosphates (Na-EDTA-P), most of low soluble pedogenic Ca phosphates (Cit-P), lithogenic Ca phosphates (HCl-P in R scheme), P occluded in poorly crystalline Fe oxides (CA-P), and P occluded in crystalline Fe oxides (CBD-P in R scheme). The combined OP fractions in mG provided an accurate estimate of OP in soils; it accounted for approximately one half of TP in the studied soils.

The ratio of Ca-EDTA-P to combined non-organic P fractions (mG scheme) was negatively correlated with soil pH, and the ratio of Cit-P to combined P fractions (R scheme) positively correlated with soil pH, suggesting precipitation of Ca phosphates being favored over P adsorption at increased pH in soil.

The maximum amount of desorbable P using resins estimated from the Johnson–Mehl equation was related mainly to the combination of P fractions including adsorbed P and precipitated Ca phosphates in the studied soils (Ca-EDTA-P + Na-EDTA-P in the mG scheme, and NaOH-P + CB-P + Cit-P in the R scheme). Fractions related to adsorbed P (Ca-EDTA-P) determined the relative proportion of quickly released P using resins. By contrast, P fractions including low soluble Ca phosphates (Cit-P) were probably involved in the long-term P desorption from soil.


   ACKNOWLEDGMENTS
 
This work was funded by Spain's R & D Program (Plan Nacional I+D), Project AGF99-0574-CO1, and, supplementarily, by DAP (the Public Company for Agricultural and Fisheries Development of Andalusia), which also provided working facilities and access to soils. Mrs. Concepción Saavedra's work was also funded by the Spain's Ministry of Science and Technology through a research training grant.

Received for publication May 4, 2004.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
C. Saavedra, J. Velasco, P. Pajuelo, F. Perea, and A. Delgado
Effects of Tillage on Phosphorus Release Potential in a Spanish Vertisol
Soil Sci. Soc. Am. J., January 1, 2007; 71(1): 56 - 63.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Saavedra, C.
Right arrow Articles by Delgado, A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Saavedra, C.
Right arrow Articles by Delgado, A.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Saavedra, C.
Right arrow Articles by Delgado, A.
Related Collections
Right arrow Nutrient Cycling
Right arrow Phosphorus
Right arrow Soil Chemistry

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Journal of
Environmental Quality