Forms oF InorganIc PhosPhorus In soIl under dIFFerent long term soIl tIllage systems and wInter croPs

the cultivation of crops with different capacity of P uptake and use under long-term soil tillage systems can affect the distribution of P cycling and inorganic forms in the soil, as a result of higher or lower use efficiency of P applied in fertilizers. the purpose of this study was to evaluate the effect of long-term cultivation of different winter species under tillage systems on the distribution of inorganic P forms in the soil. In 1986, the experiment was initiated with six winter crops (blue lupin, hairy vetch, oat, oilseed radish, wheat and fallow) on a rhodic hapludox in southwestern Paraná, under no-tillage (nt) and conventional tillage (ct). the application of phosphate fertilizer in nt rows increased inorganic P in the labile and moderately labile forms, and soil disturbance in ct redistributed the applied P in the deeper layers, increasing the moderately labile P concentration in the subsurface layers. Black oat and blue lupin were the most efficient P-recyclers and under nt, they increased the labile P content in the soil surface layers.


IntroductIon
In natural ecosystems on young, little weathered soils, P availability is controlled by the dissolution of primary mineral phosphate (tiessen et al., 1984).however, in more weathered soils, the decomposition and mineralization of forest litter and soil organic matter come to be the primary sources of P, since mineral P is retained mostly in less available forms (Vincent et al., 2010).In the conversion of natural ecosystems to agricultural areas, the native plants, well-adapted to low levels of available P are substituted by crops with high nutrient requirements, with lower capacity and strategies for P uptake and use.therefore, in view of the crop demand these environments are generally P-deficient, which limits yields and requires phosphate supply by fertilization.
In production systems with regular fertilizer application, the total P content in the soil increases.the form in which the applied phosphate accumulates in the soil depends on the applied and exported amount of P, and on the soil type and management.When the soil is cultivated for a long time under no-tillage (nt), P is accumulated in the soil surface (redel et al., 2007;Zamuner et al., 2008), since the fertilizer is applied in the planting row, with minimum soil disturbance and distributed in a lower soil volume than in conventional tillage (ct) (rheinheimer & anghinoni, 2001).With this accumulation of P under nt, the sites of higher P affinity in the soil surface layer are gradually saturated, decreasing the binding energy of P and increasing its availability (rheinheimer et al., 2003).
due to the differences in the chemical nature and binding energy of P in the soil, its desorption capabilities can differ.routine methods usually estimate P availability by a single extraction of available P with chemical extractants, such as mehlich-1 or anion exchange resin (aer) (cQFsrs/ sc, 2004).however, to evaluate the potential availability, a single extraction can be insufficient.In this sense, gatiboni et al. (2002) reported that the first extraction with aer from a clay soil estimated only 37 % of potentially available P, and that seven successive extractions were necessary to exhaust it completely.rheinheimer et al. (2000) found that after 10 successive extractions with aer, the inorganic P extracted by naoh (0.1 and 0.5 mol l -1 ) was the main buffer of the potentially available P, approximately 80 %. therefore, data on P accumulation or depletion in the long term may be obtained with stronger extractants, such as 0.1 mol l -1 naoh (Kuo et al., 2005;gatiboni et al., 2007;takeda et al., 2009).the P uptake capacity differs among crops.these differences are the result of the varied acquisition strategies, as expressed in differences in the root morphology (root density, length, diameter, hair), in varying degrees among the species, besides differences in the mycorrhiza associations, and in the ph of the rhizosphere and exudation of lowmolecular-weight phosphatases and organic acids (lajtha & harrison, 1995;Bayon et al., 2006;Wang r. Bras. ci. solo, 36:271-281, 2012et al., 2008;li et al., 2010;rose et al., 2010).these differences may affect the processes of dissolution/ precipitation, adsorption/desorption, diffusive transport, soil/root contact and P mineralization (horst et al., 2001. however, for rheinheimer & anghinoni (2001), the effect of growing different annual crop species on the dynamics of P inorganic fractions was low, restricted to the accumulation of more labile P forms in grass treatments without n application, due to the lower P export in low grain yields.
the purpose of this study was to evaluate (i) the effect of different soil tillage systems and (ii) the effect of different winter species on the distribution of soil inorganic P forms in a very clayey soil in long term cultivation.

materIal and methods
a long-term experiment was established in 1986 at the experimental station of Pato Branco of the agronomic Institute of Paraná (IaPar), in southwestern Paraná state, Brazil (52 o 41' W, 26 o 07' s, 700 m asl).the soil of the experimental site is a very acid oxisol (rhodic hapludox), with high clay content (mineralogical composition in table 1).according to Köppen, the climate was classified as cfb -subhumid tropical.the average monthly temperature and rainfall at the experimental station in the period 1979-2009 and in 2009, respectively, are listed in table 2. the experimental area was covered by subtropical forest until 1976, when it was cleared to grow maize (Zea mays l.) and beans (Phaseolus vulgaris l.) for 10 years in a conventional system, without terracing, prior to the experiment.From the winter of 1986, experimental treatments combining different soil tillage and winter cover crops were initiated in the area, in a split-plot design in three blocks.the winter cover crops consisted of: blue lupin (Lupinus angustifolius l.), hairy vetch (Vicia villosa roth), black oat (Avena strigosa schreb.),radish (Raphanus sativus l.), winter wheat (Triticum aestivum l.), and fallow.tillage treatments were conventional tillage (ct), with one disk plow and two disk harrowings, and continuous no-tillage (nt).these winter cover crops were sown in the main plots (in the years 1986-1990, 1992-1994, 1999-2001, 2005 and 2008).In some winter growing seasons (1991, 1995, 1996, 1998, 2006, and 2009), black oat was planted in the entire experimental area; in 1993 the soil was left fallow in the winter; and black oat + radish were sown in all main plots (in 1997, 2002, 2003, 2004, and 2007).the cover crops  were cut at flowering with a rolling stalk chopper (lupin, hairy vetch, oats and radish).herbicide was applied in the fallow treatment to kill weeds.In some cases, cutting was completed with herbicide application.Wheat grain was harvested until 1995 (7 crops).after wheat harvest, straw was left on the soil surface as mulch or incorporated before planting the summer crop.summer crops of maize and soybean were planted after the winter species.
summer crops were fertilized each year; the total amount of fertilizer applied in the 23 years was 1510 kg ha -1 P 2 o 5 , 805 kg ha -1 K 2 o, and 501 kg ha -1 n. generally, the P, K and one-third of the n fertilizer were applied at planting, and two-thirds of n 45 days after planting.during this period, lime was applied in six years (1.0; 2.0; 3.0; 1.5; 2.0; and 2.0 mg ha -1 of dolomitic lime in all plots, in 1989,  1992, 1995, 1999, 2001 and 2006, respectively), with a total amount of 11.5 mg ha -1 .soil was sampled in october 2009, at the beginning of black oat flowering.two trenches were opened per plot (0.5 x 0.5 m) and soil collected from five layers: 0-0.05; 0.05-0.10;0.10-0.20;0.20-0.30;and 0.30-0.40m. the subsamples of the two trenches of each plot were mixed to form a composite sample.soil was also sampled in three trenches from the adjacent undisturbed forest, for comparison.the soil samples were dried at ±60 °c, sieved through a 2 mm screen and stored.undisturbed soil samples of each layer from one trench per plot were collected with metal cylinders (0.03 x 0.0612 m), to determine the soil bulk density.soil inorganic P forms were estimated by fractionation, as described by hedley et al. (1982) and modified by rheinheimer (2000).the P forms were sequentially extracted by the following extractants: anion exchange resin (aer) (ar 103 QdP 434 Ionics Inc.), nahco 3 0.5 mol l -1 at ph 8.5, 0.1 mol l -1 naoh, 1.0 mol l -1 hcl, and 0.5 mol l -1 naoh.the residual soil was divided into two parts.In one, the total P was estimated by digestion with h 2 so 4 and h 2 o 2 in the presence of saturated mgcl 2 (Brookes & Powlson, 1982) and in the other, organic P was estimated by the ignition method (olsen & sommers, 1982).thus, the residual inorganic P was obtained by the difference between the total P and the residual organic P. the inorganic P extracted by aer and nahco 3 was considered to be labile; the inorganic P extracted by naoh was considered moderately labile and the P extracted by hcl and residual inorganic P forms were considered nonlabile.
the P desorption capability was evaluated (soil layers 0-0.05, 0.05-0.10 and 0.10-0.20 m) by successive extractions with aer, by filling 0.200 g of each soil sample into 15 ml Falcon tubes, in triplicate, with 10 ml of distilled water and one sheet of bicarbonate-saturated aer. the suspension was stirred for 16 h in an end-overend shaker at ±25 o c. the resin sheets were then removed and washed with a distilled water flush and then eluted in 10 ml of 0.5 mol l -1 hcl. the operation was repeated until the amount of extracted P stabilized near zero.With the values of cumulative desorbed P, the equation of first-order kinetics proposed by mcKean & Warren (1996) was modified as follows: where β is the maximum amount desorbed or potentially available P, α is the amount of P desorbed in the first extraction, k is the desorption rate constant, t is the extraction time.
the inorganic P was extracted with distilled water by stirring 1.000 g of soil in 10 ml of distilled water for 2 h in a horizontal shaker.total P was estimated by digestion with h 2 so 4 and h 2 o 2 in the presence of saturated mgcl 2 (olsen & sommers, 1982).the inorganic P in the acid extracts was determined according to the method of murphy & riley (1962) and in the alkali extracts as proposed by dick & tabatabai (1977).the soil P balance was assessed considering the amount of P applied, the amount of P exported and the content of potentially available P for the following crop combinations: a) black oat/lupin under nt; b) vetch/oilseed radish/fallow under nt; c) wheat under nt; and d) average of the crops under ct. the amount applied was calculated based on the amount applied in each summer crop and the number of crops between 1986 and 2009, and the amount of P exported was calculated according to the average yield of each crop during the experimental period and the average P grain content.the amount of potentially available P in the 0-0.20 m soil layer was calculated based on the amount of potentially available P (β) obtained by the mcKean & Warren (1996) equation, soil density and thickness of each layer.the relationship between P applied and potentially available P plus P exported was considered as recovered P. the P losses by erosion and the amount of P applied and exported before the beginning of the experiment (1976 to 1986) were not considered in the balance.a statistical model of three-factor randomized blocks with three replications was used.data were subjected to analysis of variance, in a split-split plot design (two soil managements x six cropping sequences x five soil layers).the tillage systems r.Bras.ci.solo, 36:271-281, 2012 were treated as main plots, the winter crops as split plots and soil layers as split-split-plots.When the treatment effects were significant at 5 % probability of error by the F test, the differences between means of tillage systems, winter crops and soil layers were compared by tukey´s test (p < 0.05).the statistical analysis were performed with assistat software, version 7.6 Beta 2011 (silva & azevedo, 2006).
results and dIscussIon effect of tillage system under nt, the content of inorganic P recovered by all extractors in the 0-0.05 m soil layer was higher than in the soil under ct, except for the residual fraction (table 3).In the ct, the content  (1) Forest (control) data were not included in the statistical analyses. (2)Values in brackets are relative to the total soil P. ct: conventional tillage; nt: no-tillage; * p < 0.05; ** p < 0.01; means followed by same letter in the column, comparing soil management for each depth, and means followed by the same capital letter in the same row, comparing the depth within each soil management, were not significantly different at p < 0.05 by tukey's test.
tales tiecher et al. of P extracted by 0.1 mol l -1 naoh from the 0.20-0.30m layer and the content of total P in the layers 0.10-0.20 and 0.20-0.30m, were higher than in nt.under nt, the labile P extracted with aer from the 0-0.5 m layer was 4.4 times higher than under ct, and due to the very low vertical mobility of P, the P content in the soil layers of 0.10-0.20 and 0.20-0.30m was only half of the content found in the soil under ct.
total P content was higher in the soil under crops than under native vegetation, in all layers evaluated (table 3). the proportion of inorganic P to the amount of total P in the cultivated soil was lower than in the forest (control) soil (49 and 57 % on average, respectively).
the annual application of phosphate fertilizer in the rows of the system without soil disturbance (nt) resulted in higher contents of total P and all inorganic P forms, except the recalcitrant non-labile fraction in the soil surface layer than in ct.It is also believed that soil and P losses by erosion are lower in the nt and the P taken up by plants from deeper layers is accumulated on the surface after the decomposition of their residues (rheinheimer & anghinoni, 2001), which may have contributed to the results.the amount of P extracted successively by aer decreased exponentially in all cases.only in the 0-0.05 m soil layer the amount of potentially available P (β) was higher in nt than in ct (table 4).
the P exchange sites in the soil are finite and have different affinities for phosphate ions.In nt, soluble phosphate applied in the planting rows leads to a region of saturated reactive sites, where much of the phosphate applied is adsorbed with low energy and becomes easily available.In this microregion, there is an increase in ph, deprotonation of functional groups in the surface, creation of negative and reduction of positive charges (Yu, 1997), making them less reactive.In addition, total organic carbon content in the soil surface layer under nt is higher and may decrease the P adsorption sites through competition with low-molecular-weight organic acids (Jones, 1998;redel et al., 2007).the set of these factors results in lower energy retention of phosphate when the fertilized soil is not tilled and explains the higher content of P extracted by water, aer and bicarbonate (labile P forms) in the 0-0.05 m layer in nt than in ct.accordingly, in the desorption of P from the soil with aer, rheinheimer et al. (2003)  to the extent that it runs out, the removed amount decreases, since the binding energy of the phosphate with colloids increases, leading to an exponentially decaying curve according to the number of successive extractions, as observed in this study.although the P availability for plants is enhanced under nt, the accumulation of phosphate on the soil surface poses a great risk of eutrophication of water bodies due to the P-rich colloidal particles carried away by runoff (Panuelas et al., 2009).
By the fertilizer application to the soil under nt, to always the same layer and by the frequent liming of the soil surface in this experiment, the P concentrations in the region under influence of the diffusion of soil-fertilizer reaction products increased to high enough levels for the neoformation of secondary calcium phosphates of low solubility (rheinheimer & anghinoni, 2001), explaining the higher content of P extracted by 1.0 mol l -1 hcl, from nt than from plowed soil.soil plowing prior to each crop in ct distributed the P applied more evenly in a greater soil volume, resulting in higher total P content in the layers 0.10-0.20 and 0.20-0.30m than in nt.thus, in ct the higher affinity sites for P in the subsurface are saturated, although much less than in the soil surface layer under nt.however, this was sufficient to increase the level of moderately labile P extracted by 0.1 mol l -1 naoh from the 0.20-0.30m layer.this demonstrates that P adsorbed in the soil subsurface under ct has a lower binding energy than under nt.nevertheless, the higher entry of P in subsurface ct was not enough to change the content of the P forms of lower lability, in the forms extracted by 0.5 mol l -1 naoh and the residual fraction.
the amount of P exported in maize and soybean grains was 252 kg ha -1 in nt and 232 kg ha -1 in ct, despite the non-significant differences in yields (calegari, 2006).the P theoretically recovered (potentially available + exported) was higher in the nt (63 %) than in ct (48 %) (table 5).
the comparison of the theoretical balance of P between the tillage systems indicates that in soils fertilized for several years under nt, the P that is not exported by crops remains in the soil in forms that are potentially more readily available than in the soil under ct, as observed by rheinheimer et al. (2000).this occurs because aside from the P distribution in a larger soil volume in the ct, P applied (2) exported (3) available (4)  recovered (5)   kg ha  (1) Yeld data obtained from calegari ( 2006). (2)amount of P applied in fertilizer from 1986 to 2009 (30 kg ha -1 of P/ maize or soybean crop). (3)amount of P exported in grains, considering a content of 2.2, 3.9 and 2.8 g kg -1 P of maize, soybean and wheat grain, respectively (cQFsrs/sc, 2004). (4)available P in the 0-0.20 m layer = maximum desorption of P with anion exchange resin (aer), considering soil mass as a function of soil bulk density and the layer volume. (5)recovered P = exported P + P desorbed with aer. ( 6 average of all winter crops.
r. Bras.ci.solo, 36:271-281, 2012 more soil and P can be lost by erosion and, as in the present study, less P exported by grains in the ct.
the increase in total P content together with the decreasing proportion of inorganic P in crop compared to forest soil indicated that P applied via fertilizer was not accumulated in inorganic forms.

effect of winter crops
under blue lupin and oat under nt in the winter the 0-0.05 m layer contained more P extracted by distilled water, aer, nahco 3 and hcl, than under the other winter crops.In the layer 0.05-0.10m, higher levels of P were extracted by hcl and by aer only under blue lupin (tables 6 and 7). the recovery of P extracted by 0.1 mol l -1 naoh was also higher under blue lupin, independent of the soil management and sampled layer (table 8), compared to the other winter crops.however, the cultivation of different plant species did not alter the content of lower lability P forms (0.5 mol l -1 naoh and residual fraction) or the total content of P in soil.
In the nt, the amount of potentially available P was higher in the soil layers down to 0.10 m under blue lupin and oat, and lower in the soil layer of 0-0.05 m under wheat, compared to the other winter crops (table 4). the desorption rate constant was not affected by the winter crops.
the higher content of labile P forms in soil surface layers under nt with black oat and blue lupin cannot be attributed to the lower export of P by grain yield of summer crops, as reported by rheinheimer & anghinoni (2001).In this study, the amount/grain yield of maize (8 crops) and soybeans (9 crops) grains produced in the soil under nt was not affected by the winter crops (calegari, 2006).nevertheless, oat produced large amounts of dry matter from 1986 to 2005; over the course of the 11 years in which all winter species were grown, black oat accumulated 15 mg ha -1 of residue more than the other cultivated species (calegari, 2006).these residues probably cycled a larger amount of P, resulting in the higher content of labile P forms in soil surface layers under nt.lupin is known for excreting large amounts of organic acids such as citrate, malate and fumarate, even at relatively high P availability (Bayon et al., 2006), which promotes the mobilization of P by ligand exchange and/or by dissolving and occupying adsorption sites.consequently, lupin can access inorganic P forms that are inaccessible to other plants.thus, lupin may also cycle large amounts of P, which may have increased the concentration of labile P forms in the soil surface layers under nt.however, in the ct, plowing distributed the organic residues in the deeper layers, with the consequent reactions of P with the soil after mineralization, which is why the content of labile P forms under black oat and lupin was not higher.
In a comparison of the amount of potentially available P in the 0-0.05 m layer of the winter crops, three groups were formed: black oat and lupin, with the highest amount of potentially available soil P, followed by oilseed radish, vetch and fallow, and wheat, which in turn, had the lowest content of potentially available P among the winter crops.however, the amount of P recovered theoretically (potentially available + exported) from soil under wheat was very similar to the amount recovered from the soil with oilseed radish, vetch and fallow (61 and 60 % respectively), probably due to the P exported by seven wheat harvests, between 1986 and 1995.conclusIons 1. Phosphate fertilization in the row in nt increased inorganic P in the labile and moderately labile forms, and soil disturbance in the ct redistributed P applied to deeper layers, promoting an increase in the moderately labile P concentration in soil subsurface layers.
2. Black oat and blue lupin were most efficient in P cycling, and when grown under nt increased the content of labile P in soil surface layers.   (Forest (control) data were not included in the statistical analyses; 2 Values in brackets are relative to the total soil P. ns : not significant; ** p < 0.01; means followed by the same letter in the column comparing winter crops were not significantly different at p < 0.05 by tukey's test.

table 5 .
Phosphorus balance after 23 years of different soil management and winter crops of a

table 4 . desorption maximum with anion exchange resin (β) and desorption rate constant (k) after 23 years of different soil management and winter crops of a rhodic hapludox
observed that the first extraction removes phosphate adsorbed with less energy and (1)Forest (control) data were not included in the statistical analyses.ct:conventional tillage; nt: no-tillage; ** p < 0.01; means followed by same letter in the column, comparing soil management for each depth, means followed by same capital letter in the column comparing soil depth for each soil management, and means followed by the same capital letter in the same row, comparing the winter crop within each soil management in each depth were not significantly different at p < 0.05 by tukey's test.r.Bras.ci.solo, 36:271-281, 2012

table 6 . Phosphorus extracted by distilled water and by anion exchange resin (aer) after 23 years of different soil management and winter crops of a rhodic hapludox
: not significant; ** p < 0.01; means followed by same letter in the column, comparing soil management for each depth, and means followed by the same capital letter in the same row, comparing the winter crop within each soil management in each depth were not significantly different at p < 0.05 by tukey's test; 1 Forest (control) data were not included in the statistical analyses. ns