Soil Science Society of America Journal 63:809-815 (1999)
© 1999 Soil Science Society of America
DIVISION S-2-SOIL CHEMISTRY
Phosphorus Transformations and Their Relationships with Calcareous Soil Properties of Southern Western Australia
Abbas Samadia and
R.J. Gilkesa
a Soil Science and Plant Nutrition, The Univ. of Western Australia, Nedlands, WA 6009, Australia
soilsci{at}cyllene.uwa.edu.au
 |
ABSTRACT
|
|---|
The agronomic effectiveness of phosphate fertilizers is strongly affected by reactions of P with soil constituents. The transformation of P added to soil and the effect of soil properties on these transformations was investigated for 14 alkaline and calcareous soils from southern Western Australia. The decline of NaHCO3-extractable P (Olsen-P) with time followed a second order kinetic equation. The kinetic rate constant (k) increased with increasing oxalate-extractable Fe (Feo), citrate–dithionite–bicarbonate (CDB)-extractable Al and Fe (Ald and Fed), CaCO3-free clay content, cation-exchange capacity (CEC), and ratio of CDB-extractable Fe (Fed) to active CaCO3 equivalent (ACCE), and k decreased with increasing ACCE. A combination of these soil properties described 93% of the variation in rate constant, of which 78% of the variation was predicted by the Fed/ACCE ratio alone. A combination of clay related properties (Ald, clay, Feo, and CEC) described 62% of the variation in Ca2–P determined by specific extraction. Carbonate-related properties (ACCE and CCE) together described 71% variation in Ca8–P. Clay-related and carbonate-related properties jointly described 97% and 81% of the variation in Fe–P and Olsen-P respectively. Surface area (SA) and Ald together accounted for 43% of the variation in Al–P. Scanning electron microscopy analyses showed that added P was uniformly distributed in the soil matrix to the limit of the spatial resolution of the technique (
2 µm).
Abbreviations: ACCE, active calcium carbonate equivalent • BET, Brunauer, Emmett, and Teller • CDB, citrate–dithionite–bicarbonate • CCE, calcium carbonate equivalent • CEC, cation-exchange capacity • EDS, energy dispersive x-ray spectra • SA, surface area • XRD, x-ray diffraction
|
INTRODUCTION
|
|---|
MANY ATTEMPTS HAVE BEEN MADE to describe the sequence of reactions that lead to phosphate retention in calcareous soils (Freeman and Rowell, 1981; Castro and Torrent, 1995). Surface adsorption and precipitation are major P retention processes depressing the availability of applied P. For soils rich in CaCO3, the solubility of P may be controlled by solid phase dicalcium phosphate (Cole and Olsen, 1959) or by chemisorption of P on calcite, with the formation of a surface complex of calcium carbonate–P with a well defined chemical composition. For high P application rates, P availability to plants is negatively related to the amount of CaCO3 in soil but not to Fe oxide, CEC, or clay contents (Afif et al., 1993).
Phosphorus fertilization of some calcareous soils produces poorly soluble Ca–P compounds at the site of the fertilizer particle, and precipitated Ca–P forms in the vicinity of the particle (Larsen, 1967; Afif et al., 1993). Such P compounds may not necessarily dominate in calcareous soils containing other P-reactive constituents, such as Fe and Al oxides. For example, for soils varying in CaCO3 content from 8 to 244 g kg1, high energy adsorption capacities, based on the two-surface Langmuir equation, were closely associated with dithionite-soluble Fe (Holford and Mattingly, 1975). Phosphorus retention increases with the ratio of clay (or Fe oxides) to total (or active) CaCO3 equivalent (Castro and Torrent, 1995; Carreira and Lajtha, 1997).
The publications cited above have relied on so-called specific extractants to determine the forms of elements in soil, so that the above interpretations are somewhat speculative. Electron microscopy coupled with energy dispersive x-ray spectra (EDS) analysis is more specific and showed preferential associations of P with Fe, Si, and Mn in complex cutans and discrete, P-rich primary mineral grains (Qureshi et al., 1978). Norrish and Rosser (1983) observed grains of apatite in less weathered soils and grains of monazite, (Ce, La, Th), xenotime, YPO4, and plumbogummite in highly weathered soils, but these minerals accounted for only a small portion of the total soil P and most P was dispersed in the clay matrix. Similarly, Pierzynski et al. (1990a, 1990b) found that Al and Si were the dominant elements associated with P, while Ca, Fe, and Mn generally were present at low concentration in such associations.
It is evident that a wide range of associations of soil constituents with P may occur in soils and presumably affect the supply of P to plants. Little is known of the mechanism of P retention by calcareous soils in the south of Western Australia despite the common occurrence in the region of such soils and their requirement for heavy application of P fertilizers (McArthur, 1991).
|
Materials and methods
|
|---|
The 14 representative calcareous soil samples (0–10 cm) were collected from several sites in the low to medium rainfall Mediterranean climate areas in southwestern Australia and all are from the localities of Western Australia (WA) reference soil sites (McArthur, 1991). The calcium carbonate (calcite) in these soils is of pedogenic origin and consists of fine calcite, often of biological origin, and calcite concretions. The physical and chemical properties of air-dried <2-mm samples, the distribution of inorganic P fractions, and methods for their determination have been described elsewhere (Samadi and Gilkes, 1998). Table 1 shows values of properties that may be related to P dynamics. These are contents of carbonate-free clay, citrate–bicarbonate–dithionite (CBD) extractable Al and Fe (Ald and Fed), acid-oxalate extractable Al and Fe (Alo and Feo), total calcium carbonate equivalent (CCE), and the amount of active lime [the carbonate capable of reacting with neutral NH4–oxalate (Drouineau, 1942)], henceforth referred to as active calcium carbonate equivalent (ACCE). The organic C content ranged from 10 to 17 g kg1, and specific surface area ranged from 2.3 to 9.6 x 103 m2 kg1, as determined using the BET method (Carter et al., 1986). X-ray diffraction (XRD) analysis showed that the dominant clay mineral was kaolinite and that Al-substituted Fe oxides were present, which is consistent with the data of McArthur (1991). The major carbonate mineral was calcite with minor amounts of Mg substituting for Ca.
Incubation Study
Soil Incubation and Bicarbonate Extractable Phosphorus
Duplicate 5-g subsamples of each soil were treated with 300 mg P kg1 soil as a solution of KH2PO4, mixed thoroughly, and incubated up to 160 d at 25°C and field capacity. Controls without P addition were also included for each soil. Separate batches were taken after 0, 5, 20, 40, 80, and 160 d for extraction with 0.5 M NaHCO3 as a measure of plant available P (Olsen et al., 1954), and P was determined by the ascorbic acid method.
Fractionation
After 160 d of incubation, duplicate 1-g P-treated and untreated samples were analyzed using a sequential fractionation method developed for calcareous soils (Jiang and Gu, 1989), which has been described earlier (Samadi and Gilkes, 1998). The terms Ca2–P, Ca8–P, Al–P, Fe–P, O–P, and Ca10–P were used to simplify the more complicated terms 0.5 M NaHCO3-soluble P, 0.5 M NH4Ac-soluble P, 0.5 M NH4F-soluble P, 0.1 M NaOH–Na2CO3-soluble P, occluded-P, and 0.25 M H2SO4-soluble P, respectively.
Inorganic P fractionation data for the soils prior to treatment are presented in Table 2
along with total P and NaHCO3-extractable P (Olsen-P). The transformation of added phosphate into an inorganic fraction was expressed as a percentage of the added P, being calculated as 100(Pi – Pi0)/Pad; Pi and Pi0 are the amounts of inorganic P fractions in phosphate treated and untreated soil, and Pad is the amount of phosphate added.
Scanning Electron Microscopy
To study the distribution of P in the soil matrix, as well as associations of other elements with P, soil samples were impregnated with an epoxy resin and polished, and cylindrical blocks were prepared by normal petrological techniques, except that kerosene was used as a lubricant (to avoid hydration of colloidal material). A glass plate and 1200-mesh silicon paper were used for grinding. The uncovered section was polished successively with 6- and 1-mm diamond pastes and then coated with carbon in a vacuum coating unit. The samples were examined in a JEOL JSM-6400 instrument (JEOL Ltd., Tokyo) equipped with an x-ray energy dispersive system (EDS). Suitable operating conditions were found to be an accelerating voltage of 15 kv and an absorbed electron current of the order of 2 nA.
Statistical analyses were performed using the program StatView (Abacus Concepts, 1996).
|
Results
|
|---|
Interrelationships Among Soil Properties
For convenience, CEC, SA, clay, Fed, Ald, Feo, and Alo contents are referred to as clay-related properties, and CCE and ACCE are referred to as carbonate-related properties. As shown in Table 3
, there are positive relationships among some clay-related properties and negative relationships between carbonate-related properties and some clay-related properties.
Availability of the Applied Phosphorus
After incubation, the availability of added P (recovery by NaHCO3 extractant) differed widely among soils and decreased markedly with time of soil P contact (Table 4)
, as is commonly observed (Ryan et al., 1985a; Afif et al., 1993; Castro and Torrent, 1995). The trend of decreasing recovery of added P as a function of incubation time is quite uniform for each soil and is well described by the second order kinetic equation
, where P0 is the available P (mg kg1) at time zero, Pt is the available P at time t, k is the rate constant, and t is the reaction time in days. The rate constant (k) varied considerably among soils (Table 2). There were negative relationships between recovery of applied P as Olsen-P and the clay-related properties (Fed, Feo, Ald, CEC, and clay content), whereas there was a positive relationship with ACCE content (Table 5)
. These relationships support the hypothesis that recovery of P by bicarbonate is greater where P is likely to be present in calcium carbonate-related compounds as indicated by ACCE and CCE, and lower for P associated with clay-related compounds, as indicated by Feo, Fed, Ald, Alo, CEC, and clay content.
View this table:
[in this window]
[in a new window]
|
Table 5 Correlation coefficients (r) for linear relationships between soil properties and the recovery of available P after 1–160 d of incubation and rate constant (k)
|
|
There are positive relationships between k values and Feo, Fed, and clay content, and weak relationships with Ald and CEC (Table 5). Using a stepwise regression procedure, a combination of the Fed/ACCE ratio, Feo, Fed, Ald, and clay content describe 93% of the variation in rate constant, of which 78% of variation is explained by the Fed/ACCE ratio alone (Table 6)
. Figure 1 illustrates the effect of Fed/ACCE ratio on the rate constant.
Transformation of Added Phosphate
Range, mean, and standard deviation values of P extracted (mg kg1), and recovery (%) of applied phosphate as various inorganic P (Pi) fractions are presented in Table 7
. After 160 d of incubation, >50% of added P had been transformed into calcium phosphates (Ca2–P 40% and Ca8–P 16%), with mean recovery values for aluminum and iron phosphates (Al–P and Fe–P) of 23% and 13%, respectively. There were negligible recoveries of P in the citrate–dithionite–bicarbonate and sulfuric acid extracts [i.e., occluded-P (O–P) and apatite (Ca10 –P)].
View this table:
[in this window]
[in a new window]
|
Table 7 Range, mean, and standard deviation (SD) for 14 soils of amounts and percentages of added phosphate transformed into different P fractions (mg P kg-1 soil) after 160 d of incubation
|
|
Using a stepwise regression procedure, a combination the clay-related properties explain 61% of the variation in Ca2–P, of which 49% of the variation is explained by Ald alone (Table 8)
. Thus, we can hypothesize that the formation of the Ca2–P form of P is not primarily related to the abundance of CaCO3, but rather to the abundance of reactive forms of Al that compete with CaCO3 for adsorption of P. It is likely that Ald provides a measure of the amount of Al-substituted iron oxides that provide a major contribution to P sorption by soils (Carreira and Lajtha, 1997).
View this table:
[in this window]
[in a new window]
|
Table 8 Stepwise multiple regression equations between soil properties and percent recovery of added P in different inorganic P fractions
|
|
In a stepwise regression, the carbonate-related and clay-related properties (Ald and clay content) jointly accounted for 75% of the variation in Ca8–P, of which 65% of the variation was due to ACCE content alone (Table 8).
The transformation of added P into Al–P was related to the clay-related properties (Ald, Feo, clay, SA). These properties together predicated 79% of the variation in Al–P, while 34% of the variation was due to the Ald content alone (Table 8).
The percentage conversion of applied P into Fe–P was positively related to the Fed/ACCE ratio and to clay-related properties (Feo, Fed, Ald, SA, CEC, clay), but it was negatively related to the carbonate-related properties. A combination of clay-related properties and carbonate-related properties along with Fed/ACCE explained 96% of the variation in Fe–P, of which 58% of the variation in Fe–P was predicted by Fed/ACCE.
Interrelationships Among Phosphate Fractions and Their Relationships with Available Phosphorus
As can seen in Table 9
, there is a significant positive relationship between Ca2–P and Ca8–P, and there are negative relationships between Fe–P and Al–P. There is also a significant negative relationship between Ca8–P and Fe–P.
The recovery of available P (Olsen-P) was positively related to Ca2 –P and negatively related to Fe–P, but there were no relationships with other fractions (Table 9). Dicalcium–phosphorus and Fe–P jointly explained 53% of the variation in available P, with the partial contribution of Fe–P being 49% (Table 10) . The relationships between available P and Ca2–P is to be expected as both forms of P are determined by NaHCO3 extraction. Indeed, one might expect a closer relationship (than is indicated by r = 0.63), but the two extractions are determined under quite different conditions (Samadi and Gilkes, 1998).
View this table:
[in this window]
[in a new window]
|
Table 10 Stepwise multiple regression equations relating available P to percent recovery of added P as different forms after 160-d incubation
|
|
Scanning Electron Microscopy
Scanning electron microscopy and EDS data show that the soils under investigation consist mostly of Al and Si (kaolin), Fe (goethite, hematite), and Ca (calcite) compounds and that there is no evidence of local concentrations of P or spatial associations of other elements with P for both original soils (Fig. 2)
or soils pretreated with a P solution (Fig. 3 and 4)
. The Kon 5 soil contains little calcite (i.e., no high Ca matrix in Fig. 2 and 3), and the chemical data discussed above indicate that P retention is mostly by Fe and Al compounds. This interpretation is consistent with the apparently uniform distribution of P, as indicated by Fig. 3 and 4. The Kell 9 soil contains much calcite (high Ca regions in Fig. 4) in addition to Fe and Al compounds. For all soils x-ray element maps failed to show an association of P with soil matrix elements. Consequently, we may conclude that P is uniformly distributed in the soil matrix to the limit of spatial resolution of the technique (2 µm) in both the treated (300 mg P kg1) and untreated fertilized soils.
View larger version (90K):
[in this window]
[in a new window]
|
Fig. 2 SEM data for the polished surface of Kon 5 slightly calcareous soil. Secondary electron (SE) and backscattered electron (BS) images, element distribution maps for P, Al, Ca, Fe, and Si, and x-ray spectrum for a whole field of view
|
|
View larger version (92K):
[in this window]
[in a new window]
|
Fig. 3 SEM data for the polished surface of Kon 5 P-treated slightly calcareous soil. Secondary electron (SE) and backscattered electron (BS) images, element distribution maps for P, Al, Ca, Fe, and Si, and x-ray spectrum for whole field of view
|
|
View larger version (91K):
[in this window]
[in a new window]
|
Fig. 4 SEM data for the polished surface of Kell 9 P-treated highly calcareous soil. Secondary electron (SE) and backscattered electron (BS) images, element distribution maps for P, Al, Ca, Fe, and Si, and x-ray spectrum for whole field of view
|
|
|
Discussion
|
|---|
The rate of decrease in recovery of added P, as Olsen-P, was faster than in the field study of Hooker et al. (1980), but similar to that in the study of Ryan et al. (1985b). Such rate differences are related to the manner of fertilization and conditions of incubation (Barrow, 1974). The more rapid rate in the present study is due to thorough mixing, constant temperature, optimum moisture, and application of P in solution in contrast to P application as superphosphate pellets and variation within the environment affecting P dissolution and adsorption in the field (Hooker et al., 1980).
The high amounts of kaolinite and moderate amounts of Al-substituted iron oxides in the soils studied provide most of the P adsorbing surfaces, as shown by the high positive relationships between rate constant (k) and the clay-related properties (Feo, Fed, Alo, Ald, and clay percentage). In contrast, the negative relationship between ACCE and rate constant (k) shows that the calcite surface has a relatively low P-adsorption capacity and/or releases much adsorbed P to bicarbonate solution. Similar results have been reported by Castro and Torrent (1995) for calcareous soils from the Mediterranean region.
Although the ratio of acid-oxalate extractable Fe (Feo or amorphous Fe) to CDB-extractable Fe (Fed or crystalline Fe) in the soils is low, small quantities of amorphous iron oxides with a large surface area may have a great effect on initial P retention (Ryan et al., 1985b). This is apparent in the negative relationship between the recovery of P as Olsen-P and values of Feo for the 5-d incubation period (Table 5). For periods longer than 5 d and up to 160 d, the negative relationships between this measure of recovery and clay-related properties (Feo, Fed, Ald, and clay contents) and positive relationships with carbonate-related properties (ACCE) indicate that P availability is not a consequence of a single soil property, but that a combination of soil components determine retention of P. Under these circumstances, Fe and Al oxides and the edge surfaces of kaolin crystals provide adsorption sites of higher affinity than the surface of calcite (Pena and Torrent, 1990; Afif et al., 1993). While adsorption and subsequent P transformations might be expected to be related to the specific surface of solid phase CaCO3 (Holford and Mattingly, 1975), this effect, if it did occur, was masked by the dominant effect of Fe oxides (Ryan et al., 1985b).
The positive relationship between available P (as Olsen-P) and the Ca2–P (also NaHCO3 soluble P) indicates that this fraction may be a major contributor to available P. In contrast, the negative relationship between the available P and the Fe–P implies that iron–phosphate associations in these soils will be inferior sources of P for plant growth. Nevertheless, due to the strong retention of phosphate by iron oxides, it seems probable that iron oxides with adsorbed phosphate are a principle form of phosphate in many calcareous soils and must therefore be significant source of P for plants (Norrish and Rosser, 1983). However, the highly predictive multivariate relationships between available P and a combination of Pi fractions suggest that P availability is not controlled solely by this P fraction (Fe–P). This proposition should be evaluated by plant growth experiments and associated analyses of changes in P forms in the soil.
Scanning electron microscopy and EDS data show no evidence of local concentrations of P or of spatial associations of other elements with P in the soil matrix. This is due to (i) P being a relatively minor, dispersed component of the soil matrix and (ii) the low spatial resolution of the scanning electon microscope technique. We, therefore, concluded that much of the fertilizer P applied to these soils will quickly become adsorbed by Fe oxides and that much available P (Olsen-P) comes from Ca2–P.Abacus concepts. 1996; Qureshi Jenkins Davis Reds 1969
|
ACKNOWLEDGMENTS
|
|---|
The first author gratefully acknowledges the Minister of Culture and Higher Education of Islamic Republic of Iran for the receipt of a fellowship.
Received for publication February 27, 1998.
|
REFERENCES
|
|---|
- Abacus concepts StatView reference. Berkley, CA: Abacus Concepts, Inc., 1996.
- Afif E., Matar A., Torrent J. Availability of phosphate applied to calcareous soils of West Asia and North Africa. Soil Sci. Soc. Am. J. 1993;57:756-760.[Abstract/Free Full Text]
- Barrow N.J. Effect of previous addition of phosphate on phosphate adsorption by soils. Soil Sci. 1974;118:82-89.
- Carreira J.A., Lajtha K.L. Factors affecting phosphate sorption along a Mediterranean, dolomitic soil and vegetation chronosequence. Eur. J. Soil Sci. 1997;48:139-149.
- Carter, D.L., M.M. Mortland, and W.D. Kemper. 1986. Specific surface. p. 413–423. In A. Klute et al. (ed.) Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA, Madison, WI.
- Castro B., Torrent J. Phosphate availability in calcareous Vertisols and Inceptisols in relation to fertilizer type and soil properties. Fert. Res. 1995;40:109-119.
- Cole C.V., Olsen S.R. Phosphorus solubility in calcareous soils: II. effects of exchangeable phosphorus and soil texture on phosphorus solubility. Soil Sci. Soc. Am. Proc. 1959;23:119-121.
- Drouineau G. Dosage rapide du calcaire actif du sol; nouvelles donnees sur la separation et la nature des fractions calcaires. Ann. Agron 1942;12:441-450.
- Freeman J., Rowell D. The adsorption and precipitation of phosphate onto calcite. J. Soil Sci. 1981;32:75-78.
- Holford J.C.R., Mattingly G.E.G. The high and low-energy phosphate adsorbing surfaces in calcareous soils. J. Soil Sci. 1975;26:407-417.
- Hooker M.L., Peterson G.A., Sander D.H., Daigger L.A. Phosphate fractions in calcareous soils as altered by time and amounts of added phosphate. Soil Sci. Soc. Am. J. 1980;44:269-277.[Abstract/Free Full Text]
- Jiang B., Gu Y. A suggested fractionation scheme of inorganic phosphorus in calcareous soils. Fert. Res. 1989;20:159-165.
- Larsen S. Soil phosphorus. Adv. Agron. 1967;19:151-210.
- McArthur, W.M. 1991. Reference soils of South-Western Australia. S.S.S.A. Dep. of Agric. W.A.
- Norrish, K., and H. Rosser. 1983. Mineral phosphate. p. 335–361. In Soils: An Australian viewpoint. CSIRO–Academic Press, Melbourne.
- Olsen S.R., Cole C.V., Watanabe F.S., Dean L.A. Estimation of a variable phosphorus in soils by extraction with sodium bicarbonate. USDA Circ. 939. Washington, DC: U.S. Gov. Print. Office, 1954.
- Pena F., Torrent J. Predicting phosphate sorption in soil of Mediterranean regions. Fert. Res. 1990;23:173-179.
- Pierzynski G.M., Logan T.J., Bigham J.M., Traina S.J. Phosphorus chemistry and mineralogy in excessively fertilized soils: Description of phosphorus-rich particles. Soil Sci. Soc. Am. J. 1990;54:1583-1589 a.[Abstract/Free Full Text]
- Pierzynski G.M., Logan T.J., Bigham J.M., Traina S.J. Phosphorus chemistry and mineralogy in excessively fertilized soils: Quantitative analysis of phosphorus-rich particles. Soil Sci. Soc. Am. J. 1990;54:1576-1583 b.[Abstract/Free Full Text]
- Qureshi R.H., Jenkins D.A., Davis R.I., Reds J.A. Application of microprobe analysis to the study of phosphorus in soils. Nature (London) 1969;221:1142-1143.
- Ryan J., Hasan H.M., Bassiri M., Tabbara H.S. Availability and transformation of applied phosphorus in calcareous soils. Soil Sci. Soc. Am. J. 1985;51:1215-1220 a.
- Ryan J., Curtin D., Cheema M.A. Significance of iron oxides and calcium carbonate particle size in phosphorus sorption by calcareous soils. Soil Sci. Soc. Am. J. 1985;49:74-76 b.[Abstract/Free Full Text]
- Samadi A., Gilkes R.J. Forms of phosphorus in virgin and fertilized calcareous of Western Australia. Aust. J. Soil Res. 1998;36:585-601.
This article has been cited by other articles:
|
|
|
|
 
F. E. Rhoton and J. M. Bigham
Phosphate Adsorption by Ferrihydrite-Amended Soils
J. Environ. Qual.,
April 20, 2005;
34(3):
890 - 896.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. A. Ippolito, K. A. Barbarick, D. M. Heil, J. P. Chandler, and E. F. Redente
Phosphorus Retention Mechanisms of a Water Treatment Residual
J. Environ. Qual.,
September 1, 2003;
32(5):
1857 - 1864.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
2 Kell 9
; 
10 Bea 1
; 
15 Kon 5
