SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

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 (10)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Dobermann, A.
Right arrow Articles by Thevs, N.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Dobermann, A.
Right arrow Articles by Thevs, N.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Dobermann, A.
Right arrow Articles by Thevs, N.
Related Collections
Right arrow Soil Fertility and Productivity
Right arrow Tropical Soils
Right arrow Nutrient Management
Right arrow Soil Analysis
Soil Science Society of America Journal 66:652-660 (2002)
© 2002 Soil Science Society of America

DIVISION S-8—NUTRIENT MANAGEMENT & SOIL & PLANT ANALYSIS

Phosphorus Fertilizer Effects on Soil Phosphorus Pools in Acid Upland Soils

Achim Dobermanna, Thomas George*,b and Niels Thevsc

a Dept. of Agronomy and Horticulture, Univ. of Nebraska, P. O. Box 830915, Lincoln, NE 68583-0915
b Univ. of Hawaii–IRRI Collaboration, Soil, and Water Sciences Division, International Rice Research Institute (IRRI) DAPO Box 7777, Metro Manila, Philippines
c Flurstr. 240, 22549 Hamburg, Germany

* Corresponding author (t.george{at}cgiar.org)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Opinions vary on fertilization strategies in part because of uncertainties in methods assessing P supply across sites. We quantified the fate and extractability of fertilizer P after two to four crops with four to five P levels applied to upland rice (Oryza sativa L.)–soybean [Glycine max (L.) Merr.] rotations in three experiments in Asia. Soil P pools were measured by Mehlich-1 extractant, a modified Hedley fractionation and by mixed-bed resin capsules after 1 and 14 d (resin adsorption quantity, RAQ-P1 and RAQ-P14). Without P addition, 84% of the total P was in the NaOH-Po and residual-P fractions across sites. Phosphorus fertilization increased Mehlich-1 P, resin-P, NaOH-Pi, H2SO4-P, RAQ-P1, and RAQ-P14 across sites, whereas NaOH-Po and residual-P were unchanged. The sum of resin-P and NaOH-Pi increased from 10% to between 20 and 30% of the total soil P. Mehlich-1 P and resin P increased similarly across sites and fitted quadratic models: the increase in Mehlich-1 P (mg kg-1 per kg P ha-1) ranged from 0.050 at low P rates to >0.125 at >400 kg P ha-1. The increases per unit P of RAQ-P, NaOH-Pi, and H2SO4-P varied among sites. Oxalate-extractable Fe accounted for most of the variation in NaOH-Pi and RAQ-P. Changes in soil P pools in tropical upland Oxisols and Ultisols following P addition are likely better reflected by NaOH-Pi and RAQ-P than Mehlich-1 P and resin P. Improvements in soil P tests are needed to better discriminate the changes in P pools from fertilization across soils.

Abbreviations: BC, buffer coefficient • IRRI, International Rice Research Institute • LTPE, Long-term P Experiments • PDSS, P decision support system • RAQ, resin adsorption quantity • SD, standard deviation


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
PHOSPHORUS FERTILIZATION is a key component of increasing soil productivity in upland cropping systems in Southeast Asia. However, opinions differ on strategies to best manage P in the highly weathered acid soils that typify these systems. Earlier studies suggested that very large initial additions of P were needed to maximize crop yields and that the excess P applied had a large long-term residual value (Kamprath, 1967). Yost et al. (1981) reported a residual efficiency of 40 to 60% with large applications of P in a strongly P-fixing Brazilian Oxisol. The benefits from large applications are attributed not only to correcting P deficiency but also to secondary effects such as increases in pH, CEC, and organic matter content (Sanchez and Uehara, 1980). However, without extremely large additions, benefits decline rapidly. Therefore, frequent small applications of P have been proposed as more economical in the long term (Cassman et al., 1993; Linquist et al., 1996). Adsorption of additional P decreases as the quantity of P already adsorbed increases (Barrow, 1990), so that repeated additions have cumulative benefits. However, there have been few P response studies across multiple sites, making it difficult to separate the more general from the more location-specific factors affecting the fate of P. The reasons for different findings between studies have been summarized by Linquist et al. (1996). Varying associations among soil P pools and plant P uptake have been documented (Hedley et al., 1994; Schmidt et al., 1996). Added P is usually rapidly adsorbed on the surfaces of Fe and Al oxides which is followed by very slow immobilization in other forms and within soil particles (Hedley et al., 1994; Linquist et al., 1997b). In acid P-fixing soils, alkali-soluble inorganic-P fractions represent a sink for P during excess application, but also a P source under conditions of deficient supply (Hedley et al., 1994; Schmidt et al., 1996). Effects of soil particle size (Tiessen et al., 1983) and aggregation (Linquist et al., 1997b; Wang et al., 2001) on plant availability of soil P have also been reported.

Identifying appropriate P management strategies requires a good understanding of the fate of applied P. Also needed are reliable methods for estimating plant available of added P in different soil pools. A P decision support system (PDSS) for managing P in acid tropical upland soils has been developed, but predictions of fertilizer P requirements using this system are associated with errors caused by uncertainties about critical soil test levels and the buffer coefficient (BC) that the PDSS uses (Chen et al., 1997). Though rapid soil P extractions such as Mehlich-1 (Nelson et al., 1953) or Bray-1 (Bray and Kurtz, 1945) can be useful when calibrated against crop performance at particular sites, their value in predicting the availability of recently applied and residual P across sites is doubtful (Linquist et al., 1997b). Critical soil test levels vary among sites, from year to year at the same site, or may even increase over time (Linquist et al., 1996). Gradual changes over time are expected because of slow reactions of fertilizer P with the soil, which include both precipitation in highly insoluble forms and slow diffusion to adsorption sites within soil (Barrow, 1983). The rates and extent of these reactions vary greatly between soils, particularly with texture, clay mineralogy, pH, aggregate size, and stability (Sanchez and Uehara, 1980; Cox, 1994; Linquist et al., 1997b). A BC for applied P, defined as the increase in extractable soil P per unit applied fertilizer P, can be estimated by adding P to a soil and measuring the increase in extractable P fractions with appropriate reagents under laboratory conditions or in the field. However, measurements on crushed soil samples may be of limited value for predicting P fertilizer requirements in the field because BCs vary greatly with soil aggregation, with soils with a high proportion of small aggregates tending to fix more P (Linquist et al., 1997b; Wang et al., 2001).

Processes such as plant-induced P solubilization (Hedley et al., 1994; Kirk, 1999), increased P availability after wetting of a dry soil (Yadvinder-Singh et al., 2000), or mineralization of organic P fractions (Linquist et al., 1997a) are of importance for P availability to plants in the field. Previous studies in lowland rice soils suggested that a dynamic soil test reflecting the soil P diffusion characteristics such as the resin capsule method might provide more realistic measurements of P availability to plants at different soil moisture levels (Dobermann et al., 1994; Skogley and Dobermann, 1996; Dobermann et al., 1996; Pampolino and Hatano, 2000). It is of interest then to examine whether such a resin capsule method would be appropriate to measure P diffusion dynamics in upland soils.

In this paper, we report initial results from three sites of International Rice Research Institute's (IRRI's) Long-term P Experiments (LTPE) on effects of increasing soil P levels from applied P on changes in soil P fractions in acid, P-fixing upland soils in tropical Asia. Our specific objectives were to (i) quantify the distribution of added P in different soil P pools, (ii) quantify the relationship between changes in soil P pools and amount of P added, and (iii) assess whether a method based on P diffusion to an ion exchange resin provides additional information about plant available P in P-fixing soils.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Site Characteristics
Field experiments were established at Matalom (Philippines) and Sitiung (Indonesia) in 1994 and at Siniloan (Philippines) in 1995. The experiments were conducted to understand long-term P dynamics and the effects of P on other components of soil fertility in highly weathered soils in the humid tropics (George et al., 2000).

Table 1 summarizes climate and soil conditions at the LTPE sites. At Matalom, February to April are usually dry months and a short dry spell occurs during August or September. The highest rainfall occurs in September or October. Rainfall is evenly distributed during the year in Siniloan and Sitiung. The Matalom site was a grass fallow infrequently grown to traditional upland rice with no prior inputs of fertilizer. The Siniloan site was a low-lying field abandoned from cultivation for ~5 yr, prior to which it was grown to 1 yr of pineapple (Ananas somosus L. Merr.), 2 yr of dry-season vegetables, and 4 yr of a lowland rice–maize (zea mays L.) rotation with some input of animal manure. The Siniloan field occasionally floods for brief periods during the wet season because of shallow groundwater level (70 cm below surface). The field at the Sitiung site was an uncultivated fallow following a lime (1.5 Mg ha-1) experiment conducted 5 yr ago. At Matalom, clay content increases from about 150 g kg-1 in the Ap horizon to >600 g kg-1 below a 30-cm depth, but drainage is still adequate. Siniloan soil is high in clay content, which increases with depth. Soils at Sitiung are deep well-structured, and well-drained clay. At all three sites, initial extractable Mehlich-1 P (0.05 M HCl + 0.0125 M H2SO4, 1:10) was below 3 mg P kg-1. Exchangeable acidity and Al were very high at Siniloan (Table 1). At Sitiung, the soil was limed with 3000 kg CaCO3 ha-1 at the beginning of the experiment, causing the pH (KCl) to increase from 3.7 in 1994 to 5.0 in 1996 and the exchangeable Al content to decrease from 2.0 to 0.03 cmolc kg-1. Soil pH after liming at Sitiung was more variable (standard deviation [SD ] 0.3 pH units) than at Siniloan and Matalom (SD 0.1 pH units). Mehlich-1 P BC (change in Mehlich-1 extractable P (mg L-1 soil) per unit of applied P (mg L-1 soil)) determined from a 10-d soil incubation indicated that soils were highly P-fixing with a BC of 0.07 at Siniloan, 0.11 at Sitiung, and 0.16 at Matalom (George et al., 2000).


View this table:
[in this window]
[in a new window]
  		Table 1. Site and soil characteristics of the three long-term P experiments. Values shown are means ± standard deviations of initial 0- to 15-cm topsoil samples.

 
Experimental Design and Field Management
At each site, a field experiment was established with four or five soil P levels and two crop sequences—either upland rice–nodulated soybean or upland rice–nonnodulated soybean—with treatments arranged in a split plot in a randomized complete block design (Table 2). There were four to six replications and plot sizes were ~100 m2. The changes in soil P discussed in this paper refer to the initial experimental period of 1994 through 1995 to 1996 and include four crops at Matalom (two upland rice, two soybean), three crops at Sitiung (two upland rice, one soybean), and two crops at Siniloan (one upland rice, one soybean). Phosphorus was applied as single (Matalom, Siniloan) or triple superphosphate (Sitiung) incorporated to 15 cm of soil before planting. The soil P levels were set at below and above a Mehlich-1 P level of 6 mg kg-1 resulting in 0, 6, 12, 24, and 48 mg P kg-1 soil. Phosphorus was subsequently applied in amounts to achieve comparable soil Mehlich-1 P levels across sites (Table 2). Required P was calculated using BC from a short-term lab P-isotherm procedure. Phosphorus additions ranged from 0 to 354 kg P ha-1 at Matalom, 0 to 166 kg P ha-1 at Siniloan, and 0 to 412 kg P ha-1 at Sitiung. Other nutrients (N, K, Zn, and Mg) were also applied (George et al., 2000).


View this table:
[in this window]
[in a new window]
  		Table 2. Application rates and total inputs of P in three long-term P experiments.

 
Sampling and Analysis of Soil P
In May 1996, soil samples from 0- to 0.15-m depth were collected at each site, air-dried, ground, and passed through a 2-mm sieve. In each plot, five soil cores were combined into one composite sample. All samples were analyzed for Mehlich-1 P (Nelson et al., 1953) by shaking 3 g of soil in 30 mL of 0.05 M HCl + 0.0125 M H2SO4 extractant for 5 min. Other soil properties (Table 1) were measured by procedures described by van Reeuwijk (1992). A modified Hedley fractionation method suitable for highly weathered Ultisols and Oxisols was used to determine sequentially five different P fractions on 1.0-g samples (Hedley et al., 1994).

Resin-P.
Inorganic P that is freely plant-available (Saggar et al., 1990) and determined after 16-h shaking soil end-over-end at 25 °C in 25 mL of deionized water containing a 120 by 23 mm strip each of anion (HCO-3 form) and cation (H+ form) exchange resin membrane (about 0.5 meq of exchange capacity each). The membranes were removed with a plastic tweezer and adherent soil particles washed into the sample tube with 9 mL of deionized water. Phosphorus was recovered from the anion resin membranes by shaking them in 20 mL of 0.5 M HCl for 1 h.

NaOH-Pi.
Inorganic P associated with Fe and Al probably through chemisorption to Fe and Al components. One mL of 3.5 M NaOH was added to the suspensions from Step 1, resulting in 35 ml of 0.1 M NaOH. The samples were shaken for 16 h and centrifuged for 5 min at 9056 x g. In a 5-ml subsample of the clear supernatant, the organic matter was precipitated with three drops of concentrated HCl to obtain the NaOH-Pi fraction. From another 10-ml subsample of the supernatant, the total NaOH-extractable P was determined by digesting and refluxing the subsample with 8 ml of concentrated H2SO4 for 2 h at 250 °C, repeatedly adding 0.5 mL of 30% (v/v) H2O2 and heating to 250 °C until the solution was clear.

NaOH-Po.
Labile organic P. NaOH-Po was calculated as NaOH-Pi subtracted from the total NaOH-extractable P.

H2SO4-P.
P associated with negatively charged oxide surfaces through exchangeable cations and some of the occluded P. Thirty-five milliliters of 0.5 M H2SO4 were added to the soil residue from Step 2. The samples were shaken for 16 h and centrifuged for 5 min at 9056 x g.

Residual-P.
The remainder of the occluded P and the more recalcitrant organic forms. The soil residue from Step 4 was digested with concentrated H2SO4 and H2O2 as described in Step 2 and the digests were diluted to 50 mL with deionized water.

Phosphorus absorption by mixed-bed resin capsules was measured in vitro using a laboratory incubation technique (Dobermann et al., 1994). Spherical resin capsules (PST-1, UNIBEST, Inc., Bozeman, MT) with a total surface area of 11.4 cm2 and 0.12 cmolc of cation (H+-form) and 0.10 cmolc of anion (OH--form) exchange capacity (Amberlite IRN-150, Rohm and Haas Co., Philadelphia, PA) were used (Yang et al., 1991). Roughly 400 g of each soil sample was made into a saturated soil paste with deionized water. To ensure uniform water saturation, all samples were equilibrated at 30 °C for 1 d. Then the soil paste was distributed into four 120-mL high-density polyethylene bottles and in each a resin capsule was inserted in the soil so as to be covered by at least 1 cm of soil on all sides. The bottles were incubated at 30 °C for 14 d with regular replacement of water lost by evaporation. After 1 d and 14 d, capsules were retrieved from the soil in duplicates, washed carefully in a stream of deionized water, and placed in 60-mL high-density polyethylene bottles. Phosphorus absorbed by the capsules was recovered by shaking in three sequential batches of 20 mL of 2 M HCl each. Phosphorus concentrations in all soil or resin extracts and digests were determined colorimetrically (Murphy and Riley, 1962). Results of the resin analysis, averaged from two capsules per time and P treatment, were expressed as RAQ (µmol P cm-2 capsule surface area).

Statistical Analysis
Using PROC GLM of SAS (SAS Institute Inc., 1988), analysis of variance was conducted to assess differences among the P treatments. Previous studies with more time steps indicated that nutrient absorption kinetics by a resin capsule fitted a fractional power-function equation (Dobermann et al., 1994).


[1]
where RAQ-Pt is the quantity (µmol cm-2) of P absorbed by the resin at time t (d), and RAQ-P1 is the quantity of P absorbed after 1 d of incubation. Using the second measurement of RAQ after 14 d (RAQ-P14), b was calculated as:

[2]

The time (in days) to adsorb twice the amount of RAQ-P1 equals 21/b., hence, the smaller the value of b, the slower the increase. A large b value, however, does not necessarily mean a large cumulative P release, because RAQ-P1 may be very small.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Changes in Soil Phosphorus Pools because of Phosphorus Fertilization
Total soil P content in unfertilized plots was largest at Siniloan, which had large amounts of P in the NaOH-Po pool associated with a high soil organic matter content (Table 3). Without P addition, ~84% of the total P was found in the NaOH-Po and residual-P pools at all sites (Table 3). Sodium hydroxide-Po was positively correlated with soil organic C content (r = 0.93, P < 0.001), dithionite-extractable Al (r = 0.92, P < 0.001), and clay content (r = 0.78, P < 0.001). Only 1.5 to 2.4 mg P kg-1 (>0.5% of total P) was found in the most labile fraction, resin-P. Resin-P in unfertilized soils was not significantly correlated with soil texture, organic matter content, pH, or dithionite-extractable Fe and Al (data not shown).


View this table:
[in this window]
[in a new window]
  		Table 3. Phosphorus fractions and total P (sum of fractions) in different P treatments at three sites. Values shown are means ± standard deviations (SD) measured in topsoil (0–15 cm) samples collected in 1996.

 
Excepting a few instances, increasing rates of P significantly increased Mehlich-1 P, resin-P, NaOH-Pi, H2SO4-P, RAQ-P1, and RAQ-P14 at all sites, whereas NaOH-Po and residual-P remained unchanged (Tables 3 and 4). The sum of resin-P and NaOH-Pi increased from about 10% (before P application) to 20 to 30% of total P in the treatments with the largest P application at each site (Table 3). Resin-P and NaOH-Pi (r = 0.69, P < 0.001) and NaOH-Pi and H2SO4-P (r = 0.84, P < 0.001) were positively correlated (Table 5), indicating that these fractions mainly measured P in equilibrium with the soil solution. NaOH-Po was most strongly correlated with residual-P (r = 0.69, P < 0.001). The high correlation between resin-P and Mehlich-1 P (r = 0.86, P < 0.001) suggests that both measure a similar pool of labile P. Our results agree with distributions of P fractions in similar soils in other studies (Beck and Sanchez, 1994; Cajuste et al., 1994; Hedley et al., 1994; Linquist et al., 1997a). Added P had little impact on organic-P fractions but greatly influenced soluble inorganic-P fractions (Hedley et al., 1994; Linquist et al., 1997b).


View this table:
[in this window]
[in a new window]
  		Table 4. Mehlich-1 extractable P, resin adsorption quantify after 1-d (RAQ-P1) and 14-d incubation (RAQ-P14), and the b coefficient describing P adsorption on the resin capsule in different P treatments at three sites. Values shown are means ± standard deviations (SD) measured in topsoil (0–15 cm) samples collected in June 1996.

 

View this table:
[in this window]
[in a new window]
  		Table 5. Linear correlations (r) between different soil P tests (all P treatments, N = 66).

 
Phosphorus Absorption by Resin Capsules
The amount of P absorbed after 1 d (RAQ-P1) was an indicator of readily extractable soil P, either initially in solution or in rapid equilibrium with the soil solution. The sensitivity of RAQ-P1 to P addition (Table 4) and the positive correlations of RAQ-P1 with Mehlich-1 P, resin-P, NaOH-Pi, and H2SO4-P (Table 5) support its usefulness as a measure of readily extractable P. However, the relationship between resin-P and RAQ-P1 differed among sites (Fig. 1) . On the limed soil at Sitiung, resin-P per unit RAQ-P1 was about three times larger than on the very acidic soils at Matalom and Siniloan. Presumably, the pH increase at Sitiung induced by liming (from pH [KCl] 3.7–5.0) was associated with an increase in the most labile P fraction, resin-P. Similar results were found for the relationship between resin-P and RAQ-P14, whereas there was no such difference among sites when RAQ-P was correlated with Mehlich-1 P (data not shown).



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 1. Relationships between P adsorption on resin membranes (resin-P of the Hedley procedure) and P absorption by resin capsules during a 1-d incubation (RAQ-P1).

 
Our study seem to indicate that the difference between RAQ-P14 and RAQ-P1 ({Delta}RAQ-P) may be interpreted as an index of P solubilized from more recalcitrant soil P fractions. The {Delta}RAQ-P increased with P addition and was highly correlated with NaOH-Pi (r = 0.74, P < 0.001) and H2SO4-P (r = 0.81, P < 0.001), possibly because of solubilization of surface-bound and some occluded P by acid generated in the vicinity of the H+/OH- presaturated mixed-bed resin capsule and diffusing away from it (Skogley and Dobermann, 1996). Moreover, {Delta}RAQ-P was weakly positively correlated with NaOH-Po and residual-P (r = 0.35–0.36, P < 0.01), whereas correlations of rapid soil tests such as Mehlich-1 P, resin-P, or RAQ-P1 with these two mostly organic P fractions were not significant (Table 5). We hypothesize that {Delta}RAQ-P represents a reservoir of soil P that becomes available through processes such as acidification, temporary reducing conditions because of wetness, and some mineralization of organic matter. However, whether a resin capsule properly mimics plant root functions requires further study because it acts as a strong sink for P and a strong source of acidity, which may lead to unrealistic pH changes in its vicinity.

In the absence of P fertilization, the RAQ-P14 values measured in the three acid upland soils were very low (0.01–0.02 µmol P cm-2, Table 4). Phosphorus application in the highest P treatments did increase RAQ-P14 7- to 18-fold, but the absolute values still remained in the low range of 0.08 to 0.18 µmol P cm-2. This is in comparison to a median RAQ-P14 of 0.23 µmol P cm-2 for about 400 widely differing lowland rice soil samples measured since 1992 (A. Dobermann, unpublished data, 1998). Previous research suggested that the total amount of P absorbed after 14 d (RAQ-P14) is an indicator of the total P-supplying capacity of the soil that includes initially available P and subsequently solubilizable P (Dobermann et al., 1994; Dobermann et al., 1996). Thus, the low RAQ-P14 values confirm the low P-supplying capacity of the acid upland soils in our study.

The coefficient b in Eq. [1] is a measure of the soil's capacity to maintain a P flux to a strong sink that simultaneously releases acid. In principle, b might be affected by the soil's P status and by soil attributes affecting the transport of acid away from the resin, its P-solubilizing effect, and transport of solubilized P back to the resin. The b coefficients (Eq. [1]) did not differ significantly between P treatments within a site (Table 4). The average b values were 0.35 at Sitiung, 0.40 at Matalom and 0.46 at Siniloan. The fact that the b values were stable within sites but differed between sites suggests b to be a measure of inherent soil attributes independent of soil P status; however, it should be noted that b values at Matalom were somewhat variable among P treatments especially at the highest P rate. Figure 2 shows the kinetics of P absorption by resin capsules at Sitiung. Although b did not vary among P treatments, P addition increased the initial (RAQ-P1) and cumulative P absorption over time. After an initial absorption period of 1 to 2 d, the change in the daily absorption rate ({delta}2RAQ/{delta}t2) became almost zero in all treatments, suggesting that steady-state P absorption conditions had been reached. This is expected because, as the zone of acidification and P solubilization spreads away from the resin, the P concentration gradient near the resin, and hence the flux into the resin tend to become constant. Further, when site b value was the highest, soil pH was the least: b 0.35 at pH 5.8 in Sitiung, b = 0.40 at pH 4.8 in Matalom and b = 0.46 at pH 4.5 in Siniloan. This suggests b to be inversely related to soil pH or positively related to the content of active, oxalate-extractable Fe or both.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 2. Kinetics of P adsorption on resin capsules in different P treatments at Sitiung. The (a) kinetics, (b) first- and (c) second-derivative functions, were estimated assuming that adsorption of P followed a power function of the type RAQ-Pt = RAQ-P1 tb. In the table, within a column, values of RAQ-P1 or b with the same letter are not significantly different at P < 0.05.

 
We used water-saturated soil sieved to 2-mm size in our RAQ measurements. While we avoided complete destruction of aggregates as in the case of P extractions done in suspensions, it is expected that some aggregate destruction had occurred. The Matalom soil was less aggregated with 30% of the total soil weight consisting of aggregates of <0.25-mm size and with larger aggregates contributing proportionately less to the total soil weight. In contrast, the soils from Siniloan and Sitiung were more aggregated with 35% of the soil weight made up of aggregates >2.0 mm (T. George, unpublished data, 1996). Therefore, the RAQ measurements are likely to be influenced by the sample soil aggregation status by influencing rates of solute transport to and from the resin. Consequently, the P release kinetics measured are probably not identical to those occurring under conditions of natural moisture and aggregation. But, given that RAQ-P1 was a sensitive index of P release from P fertilizer, and b was relatively stable within a site, the resin capsule could potentially provide additional information on soil P-supplying capacity that differs from other simple soil tests. It is possible to use the resin capsules in situ in undisturbed soil to avoid aggregate destruction (Dobermann et al., 1997) but a problem is then to define the mass or volume of soil that has been assayed. Further correlation and calibration research is required before resin capsule methods can be used routinely for upland soils.

Estimating Changes in Soil Test Phosphorus From Net Phosphorus Input
The increases in Mehlich-1 P and resin P (mg P kg-1 soil) with net input of P (kg P ha-1) fitted quadratic models across sites with high r2 values (Fig. 3a, b) . It should be noted that our P rates increased these extractable-soil P pools well beyond what is considered to be required by upland rice or soybean used in the experiments. For example, the Mehlich-1 P model indicated that a net input of about 50 kg P ha-1 was required to increase Mehlich-1 P from about 2 mg P kg-1 to the tentative critical level of 5 mg P kg-1 soil for upland rice. The increase in Mehlich-1 P with increasing net P input increased from about 0.050 mg kg-1 per kg P ha-1 at low P rates to more than 0.125 mg kg-1 per kg P ha-1 at rates above 400 kg P ha-1. The variability associated with the quadratic models at the highest P rates is large but these empirical models support the notion that at higher P additions the proportion of P added is more extractable.



View larger version (34K):
[in this window]
[in a new window]
  		Fig. 3. Increase in soil P test levels as a function of the total net amount of P (fertilizer P minus crop P removal) added prior to the collection of soil samples in 1996. The solid lines and equations show fitted models for estimating the amount of net P input to raise soil P test levels regardless of whether P was applied once or in several doses. All models were fitted using treatment means (symbols).

 
Given the similarity of the relationship between Mehlich-1 P or resin-P with net P input across sites, a possible next step would be to consider whether such relationships could be the basis for making fertilizer P recommendations for Ultisols and Oxisols in decision support systems such as the PDSS. However, the question remains whether the good fit across the three sites is partly a result of grinding of soil samples prior to P extraction which minimized the differences between soils in reactive surfaces attributable to soil aggregation, pH, and active Fe (Wang et al., 2001). Soil properties, method and amount of P applied, time from P application, and sample processing are all expected to influence the relationship between soil test P and P applied. As a consequence, many different models have been presented in other studies, including linear (McCollum, 1991), piecewise linear (Zhang and Mackenzie, 1997), quadratic (Linquist et al., 1996), and other functions. Further, site differences in P extractability were revealed when NaOH-Pi and RAQ-P were considered (Table 3 and 4), both are P pools expected to be more sensitive to P addition. The NaOH-Pi pool contains most of the P applied and the diffusion-sensitive, RAQ-P, probably mostly draws P out of this same pool. But, both NaOH-Pi and RAQ-P are P pools not commonly measured in routine soil testing.

The site differences in RAQ-P per unit net P input followed the order Siniloan > Matalom > Sitiung. The order of sites was the same for the oxalate-extractable Fe (Table 1) but soil pH followed the reverse order of Sitiung > Matalom > Siniloan (Table 1). Phosphorus input also caused steeper increases in NaOH-Pi and H2SO4-P (data not shown) at Siniloan, the highest oxalate-Fe, lowest soil-pH site, than at Matalom and Sitiung. While, the reasons for these differences are not easily discernible, it is presumed that more fixation of applied P as NaOH-extractable P on the surface of Fe oxides occurred in Siniloan soil with high active Fe than at the other sites. Thus, it is likely that NaOH-Pi and RAQ-P more accurately reflected the differences in P extractability between sites and possibly soil P available for plant uptake. Hedley et al. (1994) presents some evidence for this. They found that most of the P taken up by upland rice was P solubilized from the NaOH-Pi pool by root-induced changes.

Based on the association found between oxalate-Fe and NaOH-Pi and RAQ-P pools, we attempted estimating these P pools using regressions that included oxalate-Fe. The increase in NaOH-Pi or RAQ-P per unit net P input (net-P, kg ha-1) were described by the regression equations:

[3]

[4]

[5]
where ox-Fe is the oxalate-extractable Fe (g kg-1). Equations [3] through [5] indicate that net P input and oxalate-Fe together could explain most of the variation in NaOH-Pi and RAQ-P pools.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
In field experiments on three Ultisols and Oxisols, the application of fertilizer P mainly increased inorganic P associated with Al and Fe oxide surfaces. While P application increased soil test levels of Mehlich-1 P and resin-P similarly at all three sites, the P fractionation and release studies indicated site differences in the more P-addition sensitive extractable P pools such as NaOH-Pi and RAQ-P. Phosphorus release kinetics were related to other soil properties such as active Fe content, P solubilization from the NaOH-Pi pool, and probably aggregation. Oxalate-extractable Fe together with net P input explained most of the variation in NaOH-Pi and RAQ-P. Therefore, further studies should be made to improve soil tests for assessing P supply capacity in acid, P-fixing soils including the use of oxalate-extractable P (Guo et al., 1999). Dynamic soil tests such as a resin capsule potentially provide additional information about P availability that is not easily obtainable with conventional soil extractions. Their future potential lies mainly in studying the dynamics of P and other elements under conditions of natural soil aggregation and moisture, including moisture by nutrient interactions and their effects on plant nutrition (Pampolino and Hatano, 2000). Standardized methodologies need to be developed for this, including improved approaches for interpretation of dynamic measurements.


   ACKNOWLEDGMENTS
 
We acknowledge the assistance provided by A. Almendras, Visayas State College of Agriculture (ViSCA), Leyte, Philippines, W. Khatib, Sukarami Assessment Institute for Agricultural Technology (SAIAT), Padang, Indonesia, and R. Magbanua, IRRI, Philippines, in the collection of soil samples from the Long-term P Experiment sites at Matalom, Sitiung, and Siniloan, respectively. We thank Prof. H. Wiechmann (University of Hamburg) for supporting N. Thevs during his M.Sc. thesis program. We are grateful to Drs. Roland Buresh and Guy Kirk (IRRI) for providing comments on an earlier draft of this paper. Joint contribution of the International Rice Research Institute (IRRI) and the University of Hawaii/Soil Management CRSP/NifTAL.

Received for publication April 2, 2001.


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




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
C. R. Bond, R. O. Maguire, and J. L. Havlin
Change in Soluble Phosphorus in Soils following Fertilization is Dependent on Initial Mehlich-3 Phosphorus
J. Environ. Qual., August 9, 2006; 35(5): 1818 - 1824.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
T. V. Karpinets, D. J. Greenwood, and J. T. Ammons
Predictive Mechanistic Model of Soil Phosphorus Dynamics with Readily Available Inputs
Soil Sci. Soc. Am. J., March 1, 2004; 68(2): 644 - 653.
[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 (10)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Dobermann, A.
Right arrow Articles by Thevs, N.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Dobermann, A.
Right arrow Articles by Thevs, N.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Dobermann, A.
Right arrow Articles by Thevs, N.
Related Collections
Right arrow Soil Fertility and Productivity
Right arrow Tropical Soils
Right arrow Nutrient Management
Right arrow Soil Analysis

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