HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Yadvinder-Singh, Articles by Khind, C.S. Search for Related Content PubMed Articles by Yadvinder-Singh, Articles by Khind, C.S. GeoRef GeoRef Citation Agricola Articles by Yadvinder-Singh, Articles by Khind, C.S.

DIVISION S-4-SOIL FERTILITY & PLANT NUTRITION

Optimal Phosphorus Management Strategies for Wheat–Rice Cropping on a Loamy Sand

^a Dep. of Soils, Punjab Agricultural Univ., Ludhiana 141 004, India
^b Soil and Water Sciences Division, IRRI, P.O. Box 3127, MCPO, 1271 Makati City, Philippines (Present address: Univ. of Nebraska, IANR, Dep. of Agronomy, P.O. Box 830915, Lincoln, NE 68583-0915)
^c Texas A&M University, Agricultural Exp. Stn., Route 3, Box 219, Lubbock, TX 79401 USA

adobermann2{at}unl.edu

	ABSTRACT

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Knowledge about optimal P rates for wheat (Triticum aestivum L.)–rice (Oryza sativa L.) cropping is insufficient because of nutrient availability differences between aerobic and anaerobic soil. We assessed P management strategies in a wheat–rice rotation on a Typic Ustochrept at Ludhiana, India. Seven P fertilizer treatments applied to wheat and rice, respectively, (P0-0, P0-26, P13-13, P26-0, P26-13, P39-0, and P26-26; treatment abbreviations used include P applied to wheat followed by P applied to rice, both in kg ha^-1) were compared from 1990 to 1997. Grain yield and seasonal P accumulation by wheat were highest for higher P rates and remained stable in treatments with P applied to wheat. Phosphorus application to rice increased P accumulation by rice, but did not consistently increase rice yields because flooding decreased soil P sorption and increased P diffusion resulting in higher P supply to rice relative to wheat. Indigenous soil P supply measured in wheat was 5.8 to 8.0 kg P ha^-1, as compared with 14.9 to 18.1 kg P ha^-1 in rice. Phosphorus adsorbed by ion-exchange resin capsules placed in situ was five times greater under rice than under wheat. Applying only 26 kg P ha^-1 to wheat and no P to rice was not economical and led to a negative P balance and a decline in soil P. Applying 32 kg P ha^-1 to wheat and 15 kg P ha^-1 to rice was optimal for achieving short-term economic and long-term agronomic goals when both grain and straw were removed from the field. These findings require further validation at other sites, at higher rice yield levels, and for different straw management.

Abbreviations: AE_P, agronomic efficiency of applied P • ANOVA, analysis of variance • FP_r, amount P fertilizer applied to rice • FP_w, amount P fertilizer applied to wheat • IE_P, internal efficiency of P • IPS, indigenous P supply • P_P, price of P fertilizer • P_r, price of rice grain • P_w, price of wheat grain • RAQ, resin adsorption quantity • RE_P, apparent first crop recovery efficiency of applied P

	INTRODUCTION

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
IN IRRIGATED AREAS of the Indo-Gangetic plain, rice followed by wheat is one of the most common crop rotations. Rice–wheat systems evolved during the Green Revolution and currently occupy 9.1 million ha of North India (Woodhead et al., 1994). Wheat straw is usually completely removed from the soil, whereas rice straw is generally burned in the field. Long-term changes in yields, productivity, nutrient efficiency, and soil fertility in this rice–wheat rotation are not fully understood.

We address the issue of how to manage P in a rice–wheat rotation in order to maintain yields and profits at a high level, while sustaining or improving soil P status. Key questions are (i) How much P fertilizer should be applied in one rotation cycle? and (ii) How much P should be applied to the wheat and rice phases of this rotation? How much P to apply on wheat and rice depends on the crop P demand, potentially available soil P resources, and the chemical processes causing differences in soil P supply under aerobic and anaerobic conditions. In most lowland rice soils, P availability initially increases upon flooding (Ponnamperuma, 1972; Willett, 1986). On clay soils of temperate regions, drainage following soil submergence induced P deficiency in upland crops following rice (Willett et al., 1978; Brandon and Mikkelsen, 1979; Willett, 1979), mainly due to changes among inorganic P fractions (Sah and Mickelsen, 1986). The significance of similar processes for soil P supply in rice–wheat rotations with more frequent flood–drain cycles on coarse-textured soils with low organic matter content is little understood. Soil P depletion due to a negative P input–output balance may occur when both grain and straw are removed from the fields, but its impact on yields of wheat and rice may differ.

The main objective of our study was to assess strategies for P management in a wheat–rice rotation by combining economic profitability with agronomic sustainability and improved understanding of the differences in soil P supply under wheat and rice crops. Specific objectives were (i) to quantify changes in yield, plant P accumulation, P input–output balance, and available P during 7 yr of wheat–rice rotation; (ii) to assess differences in soil P supply under wheat and rice; and (iii) to evaluate the profitability and sustainability of wheat–rice cropping under different P management strategies. Our analysis is restricted to a single-site experiment, but may serve as an example for similar studies at other sites to arrive at broader conclusions on P management for wheat–rice rotations.

	Materials and methods

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Site Characteristics and Experimental Design
A long-term experiment studying direct, cumulative, and residual effects of applying P and K fertilizer to a wheat–rice rotation was established in 1990 at the Punjab Agricultural University farm, Ludhiana (30°56' N, 75°52' E) in Northwestern India. The soil at the experimental site is a well-drained Fatehpur loamy sand (Typic Ustochrept) with 660 g kg^-1 sand, 190 g kg^-1 silt, and 150 g kg^-1 clay in the surface horizon. Initial soil properties of composite samples taken from 0- to 15-cm soil depth were 3.2 g kg^-1 organic C (Walkley, 1947), pH (H₂O) 7.9 (1:2 soil/water suspension), 5.2 g kg^-1 free CaCO₃, 0.23 dS m^-1 electrical conductivity (1:2 soil/water suspension), 0.6 g kg^-1 total N, 4 mg P kg^-1 (Olsen et al., 1954), and 40 mg K kg^-1 (1 M NH₄OAc-extractable K, pH 7). On a volume basis, average Olsen P values and standard deviations of three composite samples were 10.0 ± 0.6 kg P ha^-1.

Wheat is grown during the cooler winter-spring period (November–April), whereas rice is grown during the summer monsoon (June–October, Fig. 1) . The long-term experiment consists of 17 treatments with different P and K rates applied to wheat and rice in a randomized complete block design with three replications (19.8 m² plot size). In this paper, we present data of the first 14 successive crops, seven wheat and seven rice crops harvested from 1991 to 1997. We restrict our analysis to seven P treatments (Table 1) .

View larger version (29K): [in this window] [in a new window]   Fig. 1 Average monthly minimum and maximum temperatures, monthly rainfall and cumulative monthly pan evaporation at Ludhiana, Punjab, India

After wheat harvest, the field was fallow up to early June, when flooding and land preparation for the rice crop started. In mid June, 5- to 6-wk-old rice seedlings (cv. PR106 initially, cv. PR111 since 1995) were transplanted in the plots at a spacing of 20 by 15 cm. One-third of total N (150 kg N ha^-1) was applied at transplanting, and the remainder top-dressed 3 and 6 wk after transplanting. All P and K fertilizer was incorporated into the soil before the last puddling. Zinc (11.5 kg Zn ha^-1 as ZnSO₄ · 7H₂O ha^-1) was applied to rice in alternate years. During the rice season, more than 400 mm of water were received from rainfall (Fig. 1). Supplemental irrigation was provided using both well and canal water. Plots were kept flooded for 3 wk after transplanting; thereafter, rice was irrigated at 2-d intervals. Although soil did not remain flooded for more than 8 to 10 h after irrigation, anaerobic conditions prevailed for more than 75% of the rice growth period.

Rice and wheat varieties were changed in 1995, but all varieties grown were modern semidwarf types with similar yield potential and harvest index. Hand-weeding was done in both crops. Although pest control followed standard practices, yield losses occurred in several years, particularly in the 1994 to 1996 rice crops. At harvest, plants were cut close to the ground surface. Rice and wheat straw was removed from the plots after harvest.

Grain yield, straw yield, and P accumulation in grain and straw were determined from samples collected in a 13.5-m² harvest area in wheat and 15-m² harvest area in rice. Grain and straw subsamples were oven-dried at 70°C and finely ground. All plant samples were digested in a mixture of HNO₃, HClO₄, and H₂SO₄ (9:3:1) and the P content was determined colorimetrically (Jackson, 1962). Grain yields were adjusted to a standard moisture content of 0.14 kg H₂O kg^-1 grain.

Soil Phosphorus Measurements
Soil samples were collected from all experimental plots (0–15 cm soil depth) shortly after rice harvest in 1992, 1994, 1995, 1996, and 1997. Air-dried soil samples of replicate plots were extracted with 0.5 M NaHCO₃ for measuring Olsen P (Olsen et al., 1954). Results are expressed in kilograms P per hectare after adjustment for bulk density.

Additional soil P measurements were obtained from the 1996-1997 wheat and 1997 rice crops. Olsen P was measured twice, after the wheat harvest (April 1997) and after the rice harvest (October 1997). Phosphorus absorption by mixed-bed resin capsules was measured in situ in wheat and rice (Dobermann et al., 1997). Spherical resin capsules (PST-1, UNIBEST, Inc., Bozeman, MT) with a total surface area of 11.4 cm² and 0.12 cmol_c of cation (H⁺-form) and 0.10 cmol_c of anion (OH^--form) exchange capacity (Amberlite IRN-150, Rohm and Haas Co., Philadelphia, PA) were used (Yang et al., 1991). After the first irrigation of wheat, on 2 Dec. 1996 (24 d after sowing), three resin capsules were placed between wheat rows at a depth of 5 cm in each replicate plot of the P0-0, P0-26, P26-0, P26-13, and P39-0 treatments. Capsules were removed individually at 1, 7, and 14 d after placement. During the 1997 rice season, in situ measurements of resin P were conducted under flooded conditions in the replicate plots of the P0-0, P0-26, P13-13, P26-0, and P26-13 treatments. Three resin capsules were placed in the middle of four rice hills 34 d after transplanting and retrieved 1, 7, or 14 d later. Phosphorus absorbed by the capsules was recovered by sequential shaking in three batches of 20 mL of 2 M HCl each and determined colorimetrically (Murphy and Riley, 1962). Results of the resin analysis are expressed as resin adsorption quantity (RAQ, µmol P cm^-2 capsule surface area) and represent the average of three replicate capsules per placement time and P treatment.

Calculations and Statistical Analysis
Agronomic efficiency (AE_P, kg grain kg^-1 P applied) and first crop recovery efficiency (RE_P, kg plant P kg^-1 P applied) were estimated for each crop grown by the difference method:

(1)

(2)

where Y_P is the measured grain yield in a treatment with P applied, Y₀ is the measured grain yield without P application (P0-0), FP is the fertilizer P input, and UP is the total P accumulation in grain and straw for the two treatments, all expressed as kilograms per hectare.

Assuming that P input from irrigation and rainfall was small and equivalent to P leaching losses (Dobermann et al., 1998), and taking into account that plants were cut at surface level with all straw removed, we estimated the partial P input–output balance (kg P ha^-1) for each crop as

(3)

To test the significance of treatment differences and possible changes with time, we compared analysis of variance (ANOVA) done for single cropping seasons with a repeated measures design in which year was included as a factor. Although grain yields, particularly those of rice, varied among years, relative differences among the P treatments did not differ significantly. Therefore, results shown are restricted to a single-factor ANOVA performed separately for each rice and wheat crop (Fig. 2) as well as an analysis of data pooled for the whole experimental period (Table 2) . For the latter, measured (yield, P accumulation) or calculated (P balance, profit) values of each crop during the period of 1991 to 1997 were summed for each replicate plot. Analysis of variance (with P regime as factor) was then performed on these cumulative values. This analysis is interpreted as the average response to the different P management strategies during the 7-yr experimental period.

View larger version (38K): [in this window] [in a new window]   Fig. 2 Trends of grain yield and total P accumulation of wheat and rice in P treatments of the wheat–rice long-term experiment at Ludhiana during 1991 to 1997. Means of seven P treatments are shown. The legend shows the amount of P applied to wheat (first number) and rice (second number) in kg P ha-1. The low rice yield in 1995 was due to severe pest damage. Bars show the LSD0.05

View this table: [in this window] [in a new window]   Table 2 Cumulative amount of grain harvested, cumulative P removal with grain and straw, the estimated total P input–output balance of wheat and rice, and economics of different P management strategies in the field experiment at Ludhiana during 1991 to 1997.

(4)

(5)

where is the total profit for one wheat–rice cycle (U.S. $ ha^-1 yr^-1), P_w and P_r are the values of grain (average farm gate prices of wheat and rice, U.S. $ kg^-1), Y is the grain yield (kg ha^-1 crop^-1), Y₀ is the grain yield without P application (kg ha^-1 crop^-1), P_P is the average market price of P fertilizer (U.S. $ kg^-1), FP is the amount of fertilizer P applied (kg ha^-1 crop^-1), and the subscripts "w" and "r" denote wheat and rice in the annual crop rotation. We used prices of U.S. $0.55 kg^-1 P, U.S. $0.133 kg^-1 wheat, U.S. $0.120 kg^-1 rice for all economic calculations and did not include fixed costs of fertilizer application or interest rates for loans to buy fertilizer (Colwell, 1994).

Using the average yields measured during the 7-yr period we modeled the average yield response of wheat and rice to P applied to either crop as:

(6)

(7)

where Y is the grain yield of wheat or rice (kg ha^-1), Y₀ is the grain yield without P fertilizer applied (kg ha^-1), FP_w is the amount of P applied to a wheat crop (kg ha^-1), FP_r is the amount of P applied to rice (kg ha^-1), and a, b, c, d, and e are the regression coefficients fitted to wheat (w) or rice (r). This quadratic model treats P from fresh fertilizer and a residual fraction of P applied to the previous crop similar to modeling the interactions of two different nutrients (Colwell, 1994). We assumed that fertilizer P applied to wheat will react differently than that applied to rice because of the differences in P availability and crop needs in the anaerobic phases.

We estimated optimal rates of P use depending on prices of P fertilizer and grain. For response functions obeying the law of diminishing returns, the economically optimal P rate is achieved at (Colwell, 1994). The optimal combination of FP_wr is the point on the three-dimensional response function where the slope of the function equals the ratio of the P price to the average price of rice and wheat. Differentiating Eq. [6] and [7] the optimal P rate of wheat is achieved at

(8)

and that of rice at

(9)

By substitution of FP_w or FP_r, Eq. [8] and [9] can be combined and solved to determine the optimal FP_w or FP_r for any combination of prices of P fertilizer (P_P), wheat (P_w), and rice (P_r):

(10)

(11)

	Results

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Grain Yield, Phosphorus Uptake, and Phosphorus Use Efficiency of Wheat
Grain yield and P accumulation of wheat changed little with time and increased with the amount of P fertilizer applied to wheat (Fig. 2). Phosphorus accumulation and wheat yield in treatments receiving 26 or 39 kg P ha^-1 to wheat (P26-0, P26-13, P26-26, P39-0) were significantly greater than in the P0-0, P0-26, and P13-13 treatments (Fig. 2 and Table 2). Averaged across the whole period, increasing P applied to wheat from 26 to 39 kg P ha^-1 did not significantly increase grain yield or P uptake. However, since 1995, grain yield and P accumulation by wheat were always greater in the treatments receiving a total P input of 39 to 52 kg P ha^-1 yr^-1 (P39-0, P26-13, and P26-26) than in P26-0 (Fig. 2). The only exception to this trend was treatment P0-26, in which the wheat yield significantly increased from 2.4 Mg ha^-1 in 1991 to 4.5 Mg ha^-1 in 1997 (295 kg ha^-1 yr^-1, P < 0.001). This trend was consistent with an increase in Olsen P from 10 kg P ha^-1 in 1991 to 14 kg P ha^-1 in 1997 (Fig. 3) .

View larger version (16K): [in this window] [in a new window]   Fig. 3 Annual P input–output balance and available P (Olsen P) measured after rice harvest in seven P treatments of the wheat–rice long-term experiment at Ludhiana during 1991 to 1997. The legend shows the amount of P applied to wheat (first number) and rice (second number) in kg P ha-1. Bars show the LSD0.05. The initial Olsen P content measured in 1990 on three composite samples is shown as mean ± standard deviation

Grain Yield, Phosphorus Uptake, and Phosphorus Use Efficiency of Rice
The yield response to P applied to rice was smaller than in wheat and not consistent across years. Yearly variations are attributed to yield losses caused by pests in 3 of 7 yr. Significant treatment differences in rice P accumulation occurred in most years (Fig. 2), but the internal efficiency of P (IE_P, kg grain per kg P accumulated) remained low, indicating that constraints other than P limited rice yields. Rice yields across all P treatments ranged from 5.2 to 6.9 Mg ha^-1 from 1991 to 1993, but were low from 1994 to 1996 (Fig. 2). In 1994, rice yield was low in spite of adequate P uptake because a severe incidence of sheath blight (Rhizoctonia solani Kühn) resulted in poor grain filling. In 1995 and 1996, rice yields and P uptake were low due to severe stem borer damage (Scirpophaga incertulas Walker and Chilo suppressalis Walker). During the first 4 yr of the experiment, rice P accumulation changed little and increased with increasing applied P (Fig. 2). In treatments receiving the highest rice P rate (P0-26, P26-26), accumulation of P by rice was 4 to 5 kg ha^-1 greater than in the treatments with P applied only to wheat (P26-0).

Differences in cumulative rice grain yield were not statistically significant across all treatments, although significant differences in the P accumulation occurred (Table 2). Treatments where 26 kg P ha^-1 was applied to rice (P0-26, P26-26) had significantly higher rice yields and rice P accumulation than the P0-0, P26-0, and P39-0 treatments in 2 and 4 yr, respectively (Fig. 2). Orthogonal comparison of these two treatment groups showed that average rice yield was significantly larger when 26 kg P ha^-1 was applied to rice than when no P was applied . Pooled across years, the average AE_P was 17 kg kg^-1 where 13 kg P ha^-1 was applied to rice and 15 kg kg^-1 where 26 kg P ha^-1 was applied. These values are about seven- to eightfold lower than in wheat. The average RE_P for all treatments with P applied to rice was 0.16 kg kg^-1 (range 0.01–0.34 kg kg^-1), and there were no significant treatment differences.

Flooding the soil released sufficient P for rice at the yield levels achieved in our experiment, especially when P was applied to wheat. From 1991 to 1997, the average IE_P for each P treatment ranged from 262 kg grain kg^-1 P in P26-26 treatment to 304 kg grain kg^-1 P in the P0-0 treatment. Average IE_P for all treatments by year ranged from 235 kg grain kg^-1 P in 1994 to 316 kg grain kg^-1 P in 1993. This is comparable with optimal values reported for rice of 385 kg kg^-1 when growth and yield are not limited or reduced by the supply of water and other nutrients or incidence of pests (Witt et al., 1999). Possible reasons for the low IE_P in our experiment include (i) insufficient congruence between N supply and crop N demand, (ii) insufficient K supply, and (iii) insufficient control of pests. Rice yields, P accumulation, and internal P efficiency increased in 1997, so it remains unclear whether the decline observed during 1994 to 1996 represents a significant trend due to some gradual change in soil fertility.

Phosphorus Input–Output Balance and Changes in Available Soil Phosphorus
The cumulative P balance after 7 yr ranged from -166 kg P ha^-1 in the P0-0 treatment to +111 kg P ha^-1 in the P26-26 treatment (Table 2). The P balance was positive in treatments in which 39 kg P ha^-1 yr^-1 was applied (Table 2, Fig. 3). The cumulative P balance for treatments receiving 26 kg P ha^-1 yr^-1 was -33 to -40 kg P ha^-1. Changes in the Olsen P content values occurred most frequently during the first 2 to 4 yr (Fig. 3). Olsen P values increased in treatments in which (i) the total P input was 39 kg P ha^-1 yr^-1 and (ii) at least one-third of the total applied P was to rice (Fig. 3). The largest increase was from 10 to 17.5 kg P ha^-1 in the P26-26 treatment, whereas Olsen P values declined by 20 and 10% in the P0-0 and P26-0 treatments, respectively (Fig. 3).

Phosphorus fertilizer applied to rice had a lower first crop recovery efficiency (RE_P) than that applied to wheat, but remained largely in forms that were available for the succeeding wheat crop. This can be seen in the continuous increase of wheat yields and a corresponding increase in Olsen P values in the P0-26 treatment, although the cumulative P input–output balance for this treatment was -33 kg P ha^-1 (Fig. 2 and 3, Table 2). In contrast, P applied to wheat appeared to be fixed in more recalcitrant pools and was not reflected in the Olsen P values measured after rice. In the P39-0 treatment, we observed no change in Olsen P values measured after rice, although the cumulative P balance was +42 kg P ha^-1 (Fig. 3).

Comparing Soil Phosphorus Supply in Wheat and Rice
Plant P accumulation in the P0-0 treatment was used as a measure of the indigenous P supply (IPS, kg P ha^-1 crop^-1). The IPS includes soil P supply from different soil layers and unknown amounts from rain and irrigation water (Dobermann and White, 1999). This definition assumes that other nutrients or water were not limiting P uptake; that is, differences in IPS between wheat and rice were purely the result of a different soil P supply and differences in root P acquisition. The IPS for wheat ranged from 5.8 to 8.0 kg P ha^-1 and that for rice from 14.9 to 18.1 kg P ha^-1 between 1991 and 1997 (P0-0 in Fig. 2).

Additional P measurements made during the 1996-1997 wheat and 1997 rice crops confirmed these differences. The RAQ P values measured after 14 d were three to five times higher in rice than in wheat, whereas no such differences were found in Olsen P values measured after rice or wheat (Table 3) . In wheat, RAQ values at 1 and 7 d were not influenced by the fertilizer P treatments. After 14 d, RAQ P was greater in the P26-13 treatment than in P0-0 and P0-26 treatments. Differences in RAQ P between treatments receiving 26 or 39 kg P ha^-1 applied to wheat were not significant (data not shown). In rice, RAQ P values at 1 d after placement did not differ significantly among treatments. However, after 14 d, RAQ P values were significantly lower in P0-0 treatment when compared with treatments receiving P (Table 3). The trend for RAQ P after 14 d to be greater in P0-26 and P26-13 than in P26-0 was not statistically significant.

View this table: [in this window] [in a new window]   Table 3 Resin adsorption quantity of P (RAQ P) and Olsen P in P treatments of the 1996–1997 wheat–rice season.

Four fertilizer P management strategies (P26-0, P26-13, P39-0, and P26-26) were statistically similar in terms of the cumulative profit (Table 2) and the temporal variability of wheat yields and profit (Fig. 4) , but superior to P0-0, P0-26, and P13-13. However, in many years (five for wheat, two for rice), grain yields and P accumulation of wheat and rice and the annual profit were highest in P26-26 (Fig. 2 and 4), and this treatment was the only one with (Table 2). In all other P treatments, P rates were not at an economic optimum because d/dFP_WR was either >0 (P26-0, P26-13) or <0 (P39-0; Table 2).

View larger version (31K): [in this window] [in a new window]   Fig. 4 Variability of grain yield of (a) wheat and (b) rice, and of (c) the profit and (d) increase in profit compared to the P0-0 treatment. Values shown are measured values of three replicate plots per P treatment for a period of 7 yr (n = 21 data points per treatment)

(12)

(13)

Within the range of P levels tested in our experiment, Eq. [12] and [13] indicate maximum possible yield responses to P of 2.72 Mg ha^-1 in wheat, but only 0.43 Mg ha^-1 in rice. Using Eq. [10] and [11] resulted in optimal P fertilizer rates of and at price levels of U.S. $0.55 kg P^-1, U.S. $0.12 kg rice^-1, and U.S. $0.133 kg wheat^-1. At this level of FP_w and FP_r, predicted values were 5.3 Mg wheat ha^-1, 5.2 Mg rice ha^-1, and an annual profit of U.S. $1305 ha^-1.

The ratio of the highest/lowest annual world market price was 1.45 for triple superphosphate, 1.31 for rice, and 1.40 for wheat during the past decade (D. Dawe, 1999, personal communication). Therefore, we simulated the influence of price changes (P_P, P_w, and P_r) on optimal rates of FP_w and FP_r for a range of ± 20% around the average prices used above (Table 5) . Price changes greatly affected the predicted annual profit, but had less influence on the optimal fertilizer rates. For all scenarios studied, the optimal FP_w ranged from 30.9 to 32.3 kg ha^-1 and the optimal FP_r ranged from 13.0 to 16.9 kg ha^-1 (Table 5, Scenarios 1–4).

View this table: [in this window] [in a new window]   Table 5 Influence of prices of P fertilizer (Pp), wheat grain (Pw), and rice grain (Pr) on optimal rates of fertilizer P applied to wheat (FPw) and rice (FPr), predicted grain yield of wheat (Yw) and rice (Yt), and the profit in the wheat–rice experiment at Ludhiana.

Therefore, as an example, we assumed that with better management it is possible to increase rice yields by 15% to levels achieved during the first years of the experiment (6.1 Mg ha^-1 in treatments with sufficient P supply; Fig. 2). Assuming no change in the yield without P applied, such an increase in yield would change the response curve of rice to:

(14)

In this case, optimal P fertilizer rates were at current average price levels. Predicted yields of 5.4 Mg wheat ha^-1 and 6.1 Mg rice ha^-1 would greatly increase the annual profit to U.S. $1447 ha^-1 yr^-1.

	Discussion

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Understanding Soil Phosphorus Dynamics in Wheat–Rice Rotation
The IPS and RAQ P in rice were much greater than in wheat, and differences in RAQ between wheat and rice increased with time (Fig. 2 and Table 3). Reduction of Fe (III) phosphate compounds and increased solubility of Ca-P compounds after flooding due to pH decrease (Willett, 1986; Kirk et al., 1990; Sanyal and De Datta, 1991), increased P sorption during the wheat phase (Brandon and Mikkelsen, 1979; Willett, 1979; Sah and Mikkelsen, 1986), root-induced solubilization of acid-soluble P by rice (Saleque and Kirk, 1995), and differences in soil P diffusion (Turner and Gilliam, 1976) are possible explanations for the higher IPS measured in rice than in wheat. Many of these processes are affected by higher temperatures during the rice phase than in wheat (Gill and Meelu, 1983). Considering, the slightly alkaline pH (7.9) and low active Fe content (1.8 g kg^-1 oxalate-extractable Fe) of our soil, most P under wheat was probably in the Ca-P form. The increase in solution P after flooding mainly resulted from a pH decrease increasing the solubility of Ca-P.

Increased P sorption following drainage is likely to be a cause of low IPS in wheat, but was smaller in our soil than on fine-textured soils where this was studied in the past. A key process is the formation of poorly crystalline amorphous Fe hydroxides (ferrihydrites) with high P sorption capacity, which leads to an increase in the sorption or occlusion of native and fertilizer P (Willett and Higgins, 1978; Willett, 1979; Sah and Mikkelsen, 1986; Willett et al., 1988). Intensity and reversibility of this process depend on factors determining the dynamics of soil redox potential in flood–drain cycles, particularly temperature, soil texture, organic matter content, and the content of reducible Fe (Sah and Mikkelsen, 1989; Sah et al., 1989). Our soil was flooded for 110 d yr^-1, but soil reduction was less strong because of relatively high percolation rates, low soil organic matter content (3.2 g kg^-1 organic C), loamy sand texture, slightly alkaline pH, and low active Fe content. The period from draining the field to sowing of wheat was 40 to 50 d and usually dry and hot. Therefore, it is likely that little accumulation of amorphous Fe took place and reoxidation of the soil occurred quickly, frequently, and for long periods. Treatment P0-26 provides evidence that increased P sorption was not strong in our soil. In this treatment, applying P to rice but not to wheat resulted in a continuous increase in wheat yields.

The RAQ data can be used to assess the significance of acid-induced P solubilization and increased P diffusion rates for the greater P supply under rice. The RAQ values after 1 d are a measure of short-term P supply mainly related to the P concentration in soil solution and easily desorbable P forms, whereas RAQ after 14 d measures the cumulative P supply for a longer period, including solubilization of acid-soluble P and differences in P diffusion rates to a strong sink such as a resin or a root (Dobermann et al., 1996). In a flooded soil in which NH⁺₄ is the dominant form of N, mixed-bed resin capsules will adsorb more cations than anions and will consequently release more H⁺ than OH^- into the soil. The resultant acidification will cause desorption of P from the soil and increase P adsorption by the resin capsule (Skogley and Dobermann, 1996), similar to the P solubilization induced by rice roots (Kirk and Saleque, 1995). On the loamy sand at Ludhiana with pH 7.9, acidification in the rhizosphere or around a resin capsule is likely to be only moderate. Acid generated will quickly diffuse away from the root (Kirk and Saleque, 1995) or capsule surface so that acid-induced P solubilization can probably not explain the large differences in RAQ and IPS between wheat and rice. Therefore, increased P diffusion due to greater P solution concentrations and greater mobility of phosphate ions appears to be a major mechanism causing greater RAQ and IPS values in rice than in wheat.

Sustainable Phosphorus Management Strategies in Wheat–Rice Rotation
The original recommendation for wheat–rice rotation in North India was to apply 26 kg P to wheat and 13 kg P to rice, based on results of field trials conducted during 1972 to 1977 (Saggar et al., 1985). Later on, the P recommendation for wheat–rice rotation was revised as to apply 26 kg P to wheat and no P to rice (Gill and Meelu, 1983). However, in our study, the P26-0 treatment resulted in lower P accumulation and profit, a negative P balance, decline in available soil P, and lower AE_P and RE_P of wheat, although, on average, rice yields were not significantly smaller than in treatments such as P26-13 or P26-26 (Table 2). In a comparable experiment conducted from 1983 to 1987 at the same site, Olsen P values declined from 15 to 13 kg P ha^-1 in the P26-0 treatment (Kolar and Grewal, 1989). A sustainable P management strategy must ensure (i) high and stable overall food production, (ii) high annual profit, and (iii) sufficient P supply for potential yield increases. Key agronomic considerations derived from our experiment are:

1. Treatments with a total P input of 26 kg P ha^-1 yr^-1 were inferior with regard to yields, P uptake, overall profit, the P input–output balance, and maintenance or improvement of soil fertility (Tables 2 and 3; Fig. 2, 3, and 4). Only treatments with a total P input of 39 to 52 kg P ha^-1 yr^-1 had a positive P balance and led to maintenance or an increase in soil P.
   
2. The initial Olsen P content of our soils was low so that sustainable management should aim at increasing it to a level of . Discounting the inferior treatment P0-26 (low wheat yield), this range of available P was only achieved in P26-13 and P26-26 (Fig. 3).
   
3. The total P input should be close to high yields occasionally achieved to ensure sufficient supply in favorable years and after other constraints to growth are removed. Attainable yields of 6 Mg ha^-1 wheat and 7 Mg ha^-1 rice require a total P accumulation of 42 kg P ha^-1 yr^-1 (17 kg P ha^-1 by wheat and 25 kg P ha^-1 by rice).
   
4. Soil P supply to wheat is lower than in rice. Any increase in the solution P concentration by applying more P fertilizer will benefit wheat. On soils testing <11 kg P ha^-1 Olsen P, a significant response of wheat to P up to 39 kg P ha^-1 has been observed at several locations in Northwest India (Azad et al., 1993). Band application of P below the seed and sufficient irrigation should be standard practice to increase soil P supply to wheat plants.
   
5. Strategies with no P applied to rice were less sustainable in agronomic and economic terms, even though the response of rice to fresh fertilizer P was inconsistent because factors other than P limited rice yields. Current levels of IPS were sufficient to achieve rice yields of 5 to 5.5 Mg ha^-1 without P application. However, raising the rice yield beyond that level will require applying P to rice.
   

Summarizing, the agronomic indicators in our experiment suggest that (i) the total P input should be in the range of 40 to 50 kg P ha^-1 yr^-1, (ii) at least 26 kg P ha^-1 must be applied to wheat for achieving wheat yields of >5 Mg ha^-1, and (iii) not more than 15 to 25 kg P ha^-1 should be applied to rice. Economically, an optimal P regime across varying prices was P32-15 (32 kg P ha^-1 applied to wheat and 15 kg P ha^-1 applied to rice; Table 5). This economic optimum matched all the agronomic criteria for a sustainable P management strategy.

	Conclusions

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Yields and P uptake of wheat changed little with time and were positively related to the annual P input. Phosphorus application to rice increased plant P uptake, but did not increase rice yields in most years because flooding increased soil P supply sufficiently for present yield levels. Rice yields were low in 3 yr, mainly due to increasing pest incidence and development of other nutrient deficiencies such as K. Achieving a positive P input–output balance required applications of 39 kg P ha^-1 yr^-1, which led to an increased available soil P where P was applied to rice and wheat. The IPS under wheat was 6 to 8 kg P ha^-1, compared with 15 to 18 kg P ha^-1 under rice. In situ-placed resin capsules provided a sensitive index of soil P supply because the capsules accounted for differences in the IPS between the aerobic (wheat) and the anaerobic phases (rice) in this cropping system.

We have demonstrated a framework for assessing P management strategies. A strategy of P32-15 was found to be optimal and sustainable in agronomic and economical terms for wheat–rice cropping on a loamy sand with low initial soil P supply. However, this new recommendation needs to be tested at other sites against the present recommendation of P26-0 and revised after several years to account for the change in IPS that may have occurred due to a positive or negative P input–output balance. More research is needed to confirm that P applied to rice has a beneficial residual effect on wheat on coarse-textured soils.StatSoft. 1998

	ACKNOWLEDGMENTS

 
We thank Dr. David Dawe (IRRI) for commenting on the economic analysis, Drs. Guy Kirk and Chirstian Witt (IRRI) for reviewing an earlier draft of this paper, and three anonymous reviewers and the Associate Editor for helpful comments.

Received for publication June 4, 1999.

	REFERENCES

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 

• Azad, A.S., Bijay-Singh, and Yadvinder-Singh. 1993. Response of wheat to graded doses of N, P and K in soils testing low, medium and high with respect to P and K in Gurdaspur district of Punjab. J. Potassium Res. 9:266–270.
• Brandon D.M., Mikkelsen D.S. Phosphorus transformations in alternately flooded California soils: I. Cause of plant phosphorus deficiency in rice rotation crops and correction methods. Soil Sci. Soc. Am. J. 1979;43:989-994.[Abstract/Free Full Text]
• Colwell J.D. Estimating fertilizer requirements. A quantitative approach. Wallingford, UK: CAB Int, 1994.
• Dobermann A., Cassman K.G., Mamaril C.P., Sheehy J.E. Management of phosphorus, potassium and sulfur in intensive, irrigated lowland rice. Field Crops Res. 1998;56:113-138.
• Dobermann A., Cassman K.G., Sta P.C., Cruz M.A.A., Adviento, Pampolino M.F. Fertilizer inputs, nutrient balance, and soil nutrient-supplying power in intensive, irrigated rice systems. III. Phosphorus. Nutr. Cycl. Agroecosyst. 1996;46:111-125.
• Dobermann A., Pampolino M.F., Adviento M.A.A. Resin capsules for on-site assessment of soil nutrient supply in lowland rice fields. Soil Sci. Soc. Am. J. 1997;61:1202-1213.[Abstract/Free Full Text]
• Dobermann A., White P.F. Strategies for nutrient management in irrigated and rainfed lowland rice systems. Nutr. Cycl. Agroecosyst. 1999;53:1-18.
• Gill H.S., Meelu O.P. Studies on the utilization of P and causes for its differential response to rice–wheat rotation. Plant Soil 1983;74:211-222.
• Jackson M.L. Soil chemical analysis. Englewood Cliffs, NJ: Prentice-Hall, 1962.
• Kirk G.J.D., Saleque M.A. Solubilization of phosphate by rice plants growing in reduced soil: Prediction of the amount solubilized and the resultant increase in uptake. Eur. J. Soil Sci. 1995;46:247-255.
• Kirk, G.J.D., T.R. Yu, and F.A. Choudhury. 1990. Phosphorus chemistry in relation to water regime. p. 211–223. In Phosphorus requirements for sustainable agriculture in Asia and Oceania. IRRI, Manila, Philippines.
• Kolar J.S., Grewal H.S. Phosphorus management of a rice–wheat cropping system. Fert. Res. 1989;20:27-32.
• Murphy J., Riley J.P. A modified single solution method for determination of phosphate in natural waters. Anal. Chim. Acta 1962;27:31-36.
• Olsen S.R., Cole C.V., Watanabe F.S., Dean L.A. Estimation of available phosphorus in soils by extraction with sodium bicarbonate. Washington, DC: U.S. Gov. Print Office, 1954 USDA Circ. 939..
• Ponnamperuma F.N. The chemistry of submerged soils. Adv. Agron. 1972;24:29-96.
• Saggar S., Meelu O.P., Dev G. Effect of P levels applied in different phases in rice–wheat rotation. Indian J. Agron. 1985;30:206.
• Sah R.N., Mikkelsen D.S. Transformation of inorganic phosphorus during the flooding and draining cycles of soil. Soil Sci. Soc. Am. J. 1986;50:62-67.[Abstract/Free Full Text]
• Sah R.N., Mikkelsen D.S. Phosphorus behavior in flooded–drained soils. I. Effects on phosphorus sorption. Soil Sci. Soc. Am. J. 1989;53:1718-1722.[Abstract/Free Full Text]
• Sah R.N., Mikkelsen D.S., Hafez A.A. Phosphorus behavior in flooded–drained soils. II. Iron transformation and phosphorus sorption. Soil Sci. Soc. Am. J. 1989;53:1723-1729.[Abstract/Free Full Text]
• Saleque M.A., Kirk G.J.D. Root-induced solubilization of phosphate in the rhizosphere of lowland rice. New Phytol 1995;129:325-336.
• Sanyal S.K., De Datta S.K. Chemistry of phosphorus transformations in soil. Adv. Soil Sci. 1991;16:1-120.
• Skogley E.O., Dobermann A. Synthetic ion-exchange resins—Soil and environmental studies. J. Environ. Qual. 1996;25:13-24.
• StatSoft. 1998. Statistica 5.1. Statsoft, Tulsa, OK.
• Turner F.T., Gilliam J.W. Increased P diffusion as an explanation of increased P availability in flooded rice soils. Plant Soil 1976;45:365-377.
• Walkley A. A critical examination of a rapid method for determining organic carbon in soils: Effect of variations in digestion conditions and of inorganic soil constituents. Soil Sci. 1947;63:251-263.
• Willett I.R. The effects of flooding for rice culture on soil chemical properties and subsequent maize growth. Plant Soil 1979;52:373-383.
• Willett, I.R. 1986. Phosphorus dynamics in relation to redox processes in flooded soils. p. 748–755. In Trans. XIII. Congr. of the Int. Soc. of Soil Sci. ISSS, Hamburg, Germany.
• Willett I.R., Chartres C.J., Nguyen T.T. Migration of phosphate into aggregated particles of ferrihydrite. J. Soil Sci. 1988;39:275-282.
• Willett I.R., Higgins M.L. Phosphate sorption by reduced and reoxidized rice soils. Aust. J. Soil Res. 1978;16:319-326.
• Willett I.R., Muirhead W.A., Higgins M.L. The effects of rice growing on soil phosphorus immobilization. Aust. J. Exp. Agric. Anim. Husb. 1978;18:270-275.
• Witt C., Dobermann A., Abdulrachman S., Gines H.C., Wang G.H., Nagarajan R., Satawathananont S., Son T.T., Tan P.S., Tiem L.V., Simbahan G.C., Olk D.C. Internal nutrient efficiencies of irrigated lowland rice in tropical and subtropical Asia. Field Crops Res. 1999;63:113-138.
• Woodhead, T., R. Huke, E. Huke, and L. Balababa. 1994. Rice–wheat atlas of India. IRRI, CIMMYT, Los Baños, Philippines.
• Yang J.E., Skogley E.O., Georgitis S.J., Schaff B.E., Ferguson A.H. Phytoavailability soil test: Development and verification of theory. Soil Sci. Soc. Am. J. 1991;55:1358-1365.[Abstract/Free Full Text]

This article has been cited by other articles:

				R.K. Gupta, Yadvinder-Singh, J.K. Ladha, Bijay-Singh, J. Singh, G. Singh, and H. Pathak Yield and Phosphorus Transformations in a Rice Wheat System with Crop Residue and Phosphorus Management Soil Sci. Soc. Am. J., August 9, 2007; 71(5): 1500 - 1507. [Abstract] [Full Text] [PDF]

				A. Dobermann, T. George, and N. Thevs Phosphorus Fertilizer Effects on Soil Phosphorus Pools in Acid Upland Soils Soil Sci. Soc. Am. J., March 1, 2002; 66(2): 652 - 660. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Yadvinder-Singh, Articles by Khind, C.S. Search for Related Content PubMed Articles by Yadvinder-Singh, Articles by Khind, C.S. GeoRef GeoRef Citation Agricola Articles by Yadvinder-Singh, Articles by Khind, C.S.