Phosphorus Availability for Irrigated Rice Cultivated Under No-Tillage and Different Phosphate Sources

The objective of this work was to evaluate the response of irrigated rice to phosphate fertilization using triple superphosphate (TSP) and Arad phosphate rock (APR), and the phosphate residual effects of upland crops in no-tillage system on the following rice crop. Also, it aimed to evaluate the efficiency of Mehlich-1 and anion exchange resin as soil P extractors. Two experiments were conducted in Albaqualf soil under irrigated conditions in the southern region of Rio Grande do Sul State. The experiments were designed as random blocks with four replications and the treatments were displayed as a 2 × 2 factorial (TSP and APR, with and without annual P application as TSP). No yield responses to phosphate were observed. Phosphate fertilization performed on upland crops (maize and soybeans) presented a residual effect on the rice crop even after four years of consecutive cropping under no-tillage. The extractors Mehlich-1 and AER were equally efficient in the evaluation of P availability for the rice crop. P content values obtained by both methods did show a significant correlation with accumulated plant P. The APR presents a similar performance as the TSP in regard to phosphate nutrition in irrigated rice when rotated with upland crops under no-tillage system.


Introduction
Soluble phosphate fertilizers are readily dissolved in the soil releasing high P concentrations, leading to higher P uptake and fast growth of plants.However, these phosphate sources loose their efficiency over time since P adsorption occurs converting it to non-labile forms.Natural reactive phosphates slowly get into solution and normally are more efficient to be used at medium term.
Combinations of soluble and rock phosphate can consist in a good strategy for the supply of P to the irrigated rice cultivated as a single crop or in crop rotations (Ferreira et al., 2007).Some reports have shown a positive effect on rice grain yield after four years of phosphate rock (PR) fertilizer use, indicating a residual effect of P for the following crop (Gomes et al., 2005).
The one-year residual effect is higher after two successive applications in comparison to only one application, after two years or more for both situations (Nakamura et al., 2016).It is possible that the residual effect remains for longer, but it was not possible to detect in this experiment.use ferric oxide as electron receptors in the respiration process, reducing them to ferrous oxides.The latter are more soluble, promoting P desorption (Sousa et al., 2009).In such an environment, P uptake by plants can occur.
Many P extraction methods have produced good results regarding the phosphate fertilizer used for the upland crop (Chang, 1978).However, in flooded soils, P levels increase significantly due to soil reduction, but varying according to soil type.Since most methods include the analysis of dry soil samples, they miss possible changes in P availability after flooding, not reflecting a true condition of P availability to rice in flooded soils.The Mehlich-1 extractor (H 2 SO 4 0.025N + HCl 0.05N) is the most commonly used method in Brazil.This method is based on maintaining H + activity in the solution.This activity is sufficient for the partial dissolution of calcium phosphate and Fe and Al oxides (Lindsay, 1979).Although presenting only a fair predictive value for the P availability in State soils cultivated with rice, no other method has provided significantly better results in order to justify its replacement (Silva et al., 2008).
The method of P extraction based on ion exchange resins was initially proposed by Amer et al. (1955).In Brazil, only the State of São Paulo uses this method in routine labs.Ion exchange resins are synthetic products that present a tridimensional hydrocarbon chain net containing functionally charged groups (-NR 3 + OH -) that simulate a strong alkali compound which dissociates in any pH value (Silva & Raij, 1999).In the P extraction process, the element is transferred to the resin, causing its adsorption (Raij et al., 1986).
An advantage of this method as opposed to Mehlich-1 is that the extraction is persistent and it uses water in a similar manner to P uptake by plant roots and does not use any energetic chemical reagents which could dissolve non-labile phosphates.According to the Committee of Soil Chemistry and Fertility-RS/SC (2004), this method is indicated for the diagnoses of P availability in soils that were fertilized with phosphate rocks over two years.The interpretation is though performed independent from Clay or soil flooding levels.
The objectives of this work were to evaluate the response of irrigated rice to phosphate fertilizer using triple superphosphate (TSP) and Arad phosphate rock (APR).Also, to evaluate the residual effect of P applied on upland crops under no-tillage system to the following rice crop and the efficiencies of Mehlich-1 and anion exchange resin methods as soil P extractors.

Material and Methods
In order to achieve the proposed objectives, two experiments were performed in the southern region of Rio Grande do Sul State.One experiment was conducted for a period of 5 years in a Typic Albaqualf soil (Soil Survey Staff, 2010) in the Field belonging to the Lowland Station of Embrapa Temperate Climate.The experiments were designed as random blocks with four replications in a 2 × 2 factorial (Triple Superphosphate-TSP and Arad Phosphate Rock-APR, with and without annual application of P as TSP).Also, a control treatment was included without any phosphate fertilizer application.One experiment was conducted with soybean and ryegrass on the summer and winter, respectively.The other experiment had maize in the summer and ryegrass + white clover in the winter.The area covered by each plot was 20 m 2 .
The upland crops were cultivated in the same area for four years under no-tillage cropping.Winter crops were desiccated with a total action herbicide and the plant residues remained on the soil surface.During the summer of the fifth year, all plots were sown with rice cultivar Atalanta under no-tillage on the desiccated residue of upland crops.P doses were established as a function of an initial soil analysis performed before setting the experiments: water pH: 5.5 (soil/water ratio of 1:1); O.M.: 2.2% (Walkley-Black wet oxidation method); K: 39.9 mg dm -3 , Na: 48 mg dm -3 and P: 1.5 mg dm -3 (Mehlich-1); Al: 0.6 cmol c dm -3 , Ca: 2.2 cmol c dm -3 and Mg: 1.3 cmol c dm -3 (1 mol L -1 KCl); and clay content: 20%.Phosphate fertilizers were in the order of 110 kg ha -1 for maize and 120 kg ha -1 for soybean.K and N fertilizers were also applied for maize (100 kg ha -1 KCl and 130 kg ha -1 N as urea, respectively) and K for soybean (90 kg ha -1 KCl).
Before the rice crop, at the fifth year of experiments, soil samples were collected at two depths (0-10 cm and 0-20 cm).Soil samples were dried and sieved (2 mm sieve).In the soil, P content was assessed by Mehlich-1 and AER on 5.0 × 2.0 cm layers.Plant tissues were analyzed for P content (g kg -1 ) according to the method described by Tedesco et al. (1995).A total of 50 flag leaves were collected per plot at full flowering stage.At the end of the crop cycle, a 0.50 m linear sample of plants was collected per plot and dried at 65 °C for 72 hours.After, P and shoot (culms, leaves and grains) dry matter contents were measured.Grain harvest for yield assessment were performed after physiological maturation in an area of 8.40 m 2 per plot.
Statistical analyses were performed by partitioning the variation among treatments (ANOVA) considering a 2 x 2 factorial, the additional treatment, as well as the design (random blocks with four replications).Factors consisted as follows: Factor A-P 2 O 5 source (A1: TSP and A2: APR) and factor B-phosphate fertilizer application (B1: without P 2 O 5 reapplication and B2: with P 2 O 5 reapplication).The additional treatment consisted of a control without P application.
In the general conditions used, the model can be expressed by the following equation: Where, Yij: response observed on the plot that received the ith treatment on the jth block; m: experiment mean value; ti: effect of the ith treatment; bj: effect of jth block; EEij: experimental error that received the ith treatment at the jth block.
For the partitioning of treatment effects that form the factorial structure, the procedures described by Zimmermann (2004) were used: use of orthogonal contrasts in order to isolate the factorial effects and partitioning of factorial into principal effects and their interactions.According to Zimmermann (2004), when the experimental factors show only two levels, there is no need to perform a test complementary to F for the comparison of means.For correlation analyses, the Winstat software was used (Machado, 2001).

Results and Discussion
The Phosphate fertilization increased P contents in the soil as detected by both extraction methods.These contents were different from the control without fertilizer application (Table 1).P contents extracted by Mehlich-1 were inferior to the critical levels of 6.0 mg dm -3 in the maize desiccated residue area for flooded soils (Committee of Soil Chemistry and Fertility-RS/SC, 2004).In the soybean desiccated residue area, P contents were higher than 6.0 mg kg -1 at both depths (0-10 cm and 0-20 cm) when there was reapplication of phosphate fertilizer.
Table 1.Soil extracted P contents using Mehlich-1 and anion Exchange resin (AER) from two depths of a Typic Albaqualf after harvesting of maize and soybean with and without phosphate application (w/r and wo/r) Phosphate Fertilizers Depths 0-10 cm 0-20 cm 0-10 cm 0-20 cm wo/r w/r wo/r w/r wo/r w/r wo/r w/r  Note.* Control differs from remaining treatments.Means followed by distinct capital letters on the column and low capital letters on the row differ statistically with the F test (< 5%).Control: without P 2 O 5 application, TSP: recommended dose of P 2 O 5 as TSP with and without annual reapplication; APR: recommended dose of P 2 O 5 as PR from Arad with and without annual reapplication with TSP.
There were no differences between P sources regardless the depth and area analyzed, but the annual reapplication of the recommended dose lead to increases in soil P contents (Table 1).The extracted P contents obtained by AER were higher than the critical level only in those treatments with annual fertilizer reapplication for both areas analyzed.The critical level suggested by the Committee of Soil Chemistry and Fertility-RS/SC ( 2004) is 20 mg dm -3 (Table 1).As opposed to the results from the Mehlich-1 extractor, there were differences in P contents extracted by the AER between the two sources, where the APR lead to higher P contents in the maize desiccated ER) 0 cm have since to the sis of soil P availability in soils fertilized with phosphate rock after two years.On the other hand, since Mehlich-1 is an acid extraction method, it could dissolve non-labile phosphates.However, due to the simplicity and speed of this method, Mehlich-1 is the most used in Brazil, and is based on maintaining enough H + activity in the solution for the partial dissolution of calcium phosphates and Fe and Al oxides (Lindsay, 1979).The P extraction method using the AER is based on desorption or dissolution of P from the soil solid phase for reestablishing the equilibrium with the P dissolved in the soil suspension (Silva & Raij, 1996).The reading of P contents extracted by the resin is performed independently from Clay content or soil flooding (Committee of Soil Chemistry and Fertility-RS/SC, 2004).
According to Skogley and Dobermann (1996) in the AER method, the P extracted is linked to the colloid surface, until the electrochemical equilibrium between the soil-solution-resin is obtained.The Mehlich-1 method can cause the solution of soil stable P compounds in the soil that do not contribute to P in solution, especially calcium phosphates (Raij et al., 1986;Oliveira et al., 2015) Nevertheless, the present study did not present higher P contents when the Mehlich-1 was compared to the AER method (Table 1).
A report from Cardoso (2007) disagrees with Silva and Raij (1996), which state that one of the advantages of the AER method is to avoid the overestimation of P availability as much as the acid extractors in soils fertilized with phosphate rocks.This author observed lower correlation coefficients between the P accumulated in the rice plants and the P extracted by the resin in a soil fertilized with phosphate rock (r = 0.47) when compared to the triple superphosphate (r = 0.67).
The phosphate fertilizer, independently from reapplication and P sources increased P content in the rice plants (Table 2).The higher P contents in flag leaves and rice shoots were observed when there was an annual reapplication of phosphate fertilizer.Differences between phosphate sources in the soybean desiccated residue and were only observed on shoot P contents, always showing lower values for APR when compared to TSP.  -------------------------------------P content in shoot (g kg -1 ) - -------------------------------- Note.* Control differs from remaining treatments.Means followed by distinct capital letters on the column and low capital letters on the row differ statistically by the F test (< 5%).Control: without P 2 O 5 application, TSP: recommended dose of P 2 O 5 as TSP with and without annual reapplication; APR: recommended dose of P 2 O 5 as PR from Arad with and without annual reapplication with TSP.
Rice flag leaf P contents were kept below sufficient levels determined by the Committee of Soil Chemistry and Fertility-RS/SC (2004) for irrigated rice, which ranges from 2.4 to 4.8 g kg -1 .However, the critical level of P in the flag leaf for modern rice cultivars according to Dobermann and Fairhurst (2000), and SOSBAI ( 2016) is 1.8 g kg -1 and 1.7 g kg -1 , respectively.If this criterion is used, the treatments with annual phosphate reapplication would be supplying the rice plants, while in other treatments (control and fertilizer applied only in the first year) plants would be P deficient.If one considers that the fertilizer amounts used in the treatments with fertilizer reapplication were sufficient to nourish the rice plants, the levels suggested by Dobermann and Fairhurst (2000), and SOSBAI (2016) seem to be more adequate than those suggested by the Committee of Soil Chemistry and Fertility-RS/SC (2004).-----------Maize   --------------------4.03Ab4.05Ab ------------------- ------------------------------Accu 9.94Aa 9.54Aa -------------- ) and after so values ranging lants (Table 3 n higher plant esiccated resid h annual TSP r plants from no ilizer (w/r and wo/r  -----------------------------------8.92Aa6.24Ba ------------------  In the maize desiccated residue area (Figure 3), the shoot accumulated P was better correlated with P contents extracted with AER (r = 0.77 and r = 0.78) than with Mehlich-1 (r = 0.59 and r = 0.68) method.However, in the soybean stubble crop area (Figure 4), an opposite result was observed when compared to the maize area, where the rice plant P accumulation correlated better with P values obtained from the Mehlich-1 extractor (r = 0.75 and r = 0.76) rather than with AER (r = 0.51 and r = 0.53).The higher correlation observed between P accumulated and the P content obtained from AER in the maize area could have been caused by a concentration of data points in two regions of the graph, related to low and high P concentrations extracted by the resin.The occurrence of these two regions is related to treatments with and without annual P reapplication.Those are different from P values in Mehlich-1 extractions where there is a better distribution of data points along the straight line.
Other irrigated rice studies have shown that the P availability predictions by Mehlich-1 and AER are equivalent.
Correlation coefficients between P values detected and plant P levels of 0.45 and 0.70 have been observed for Mehlich-1 and AER, respectively (Silva et al., 2008).On another report evaluating three lowland soils (Typic Albaqualf, Vertic Albaqualf and Typic Endoaqualf), where different TSP and Daoui phosphate rock doses were applied, correlation levels of 0.58 and 0.61 were observed for Mehlich-1 and AER, respectively (Cardoso, 2007).
Even though the Mehlich-1 method is considered only fair for the prediction of P availability from Rio Grande do Sul State soils, no other method has shown a convincing superior performance in order to justify its replacement (Silva et al., 2008).
Grain yield (Table 4) was not affected by any of the applied treatments, suggesting that P amounts available in the soil were enough to achieve similar yields.Although P is usually a limiting factor for plants cultivated in highly drained soils, it seems not to be a limiting factor for irrigated rice, since its availability is higher under reducing conditions (Vahl & Sousa, 2004).Increases in grain yields obtained from phosphate fertilizer experiments carried out on Albaqualf soil rarely reached over 20% (Vahl, 2004).
Upland crops cultivated before rice did respond to the phosphate fertilizer, regardless of the source used.For maize and soybean, the grain yields were 100% and 50% higher than the control (no fertilizer), respectively (data not shown).These results indicate that upland crops respond better to phosphate fertilizer than irrigated rice.
According to Lang et al. (2016), the gradual corrective phosphate fertilization in the row resulted in higher yield and grain quality of upland rice when compared to its absence.Also, in the upland rice crop, the use of triple superphosphate promoted increases in the number of tillers and panicles.
The lack of response of rice to the fertilizer harms the comparison of phosphate sources used.However, in general P contents in the soil and in the plants were equivalent between sources.Therefore, it is possible to infer that the use of phosphate rocks on upland crops preceding rice can be a viable alternative.Other studies have shown tha (Gomes et The P cont 5).Thus, a rice plants deficiency and Fairhu nutrition o Figure 5 Such indic with P con are related conditions

Rice grain level of de
The phosp the irrigate The Mehli the irrigat accumulat The Arad irrigated ri

Acknowle
The FAPE work.