Effect of cumulative fertilizer dressings on the phosphorus status of mineral soils I Changes in inorganic phosphorus fractions

Surface soil samples were collected from 16 P fertilization trials before onset of the experiments and after seven years of cultivation. The changes in the inorganic P fractions were investigated in plots amended annually with 0, 30 or 60 kg of P ha~'. In the clay soils, cultivation without P fertilization depleted the NH4F-extractable and NaOH-extractable P reserves by 22—69 kg ha-1 ; in the coarser soils, the respective depletion was B—l4o8 —140 kg ha-1 . H 2 S04 -soluble P decreased in seven soils by 16—34 kg ha -1 . In the plots amended totally with 210 or 420 kg of P ha-1 , on the other hand, these P fractions increased by 24—174 and 46—368 kg ha-1 , respectively. The higher the P dressing was, the more the added P tended to accumulate in the fluoride-soluble form as compared to the alkali-soluble form. Index words: P accumulation, residual P, inorganic P fractions


Introduction
Effective P sorption by A 1 and Fe com- pounds is typical of Finnish mineral soils (Kaila 1963, 1964, 1965, Hartikainen 1979).Efficient retention restricts losses by leaching, but also diminishes the availability of P to plants.Laturi (1977) calculated that the an- nual uptake of added P by yields averages 30 %, but much lower estimations have been published recently (Saarela and Sippola   1987).Thus, during recent decades, a succes- sive build-up of residual P has taken place in Finnish cultivated soils amended with large quantities of fertilizer P.
The long-term effect of fertilization inten- sity on different inorganic P reserves has been studied only sparsely under field condi- tions in which management practices, partic- ularly ploughing, cause dilution of residual P in the uppermost layers.Store dressing with rock phosphate has been found to result in a quite permanent P enrichment in acid-soluble form (see Hänninen and Kaila 1960, Kaila  1969), whereas superphosphate has been demonstrated to accumulate in NH 4 F-soluble and NaOH-soluble fractions (Kaila 1961).There are, however, no data available on the changes that occur in various P fractions in plots cultivated for several years without P fertilization.The present study reports the P accumulation found in a seven-year field ex- periment with increasing superphosphate treatments and the depletion in P reserves in plots that received no P fertilizers.

Material and methods
The soil samples were collected from 16 P fertilization trials carried out by the Institute of Agricultural Chemistry and Physics, Agricultural Research Centre, at several re- search stations in Finland.The trials involved five treatments where the P quantity added annually varied from 0 to 60 kg ha~'.The experimental crops were mainly cereals, but pea and grasses were also cultivated.
The fields were composed of four blocks, each of which contained all the treatments.Five of the fields represented clay soils (>3O % clay fraction <2 pm), 11 being coarser mineral soils.Characteristics of the soils are given in Table 1.The particle-size composition of mineral material was deter- mined according to the pipette method of Elonen (1971), and the organic C content was measured by a modified Alten wet com- bustion method (Graham 1948).Aluminium (Al 0) and iron (Feo) were extracted with an acid (pH 3.3) 0.05 M NH 4 -oxalate solution at a soil to solution ratio of 1:20.
Before the start of the trial, a volume of soil was gathered from each experimental field, air-dried and used as a control sample.The other samples were taken at the end of the trial from plots fertilized for seven years with 0, 30 or 60 kg of P ha~' annually.Each of the air-dried and 2 mm sieved subsamples was analyzed separately for pH in a 1:2.5 (w/v) 0.01 M CaCl 2 suspension and for inorganic P fractions by a modified Chang and Jackson procedure (Hartikainen 1979).The P con- centration in the extracts was determined by the molybdenum blue-stannous chloride meth- od of Kaila (1955).

Results and discussion
The inorganic P fractions in soils at the be- ginning of the trials and the changes that oc- curred during the seven-year period as a result of different fertilization backgrounds are shown in Table 2.The results, given in kg ha~' per plough layer of o -2 o 0 -20 cm, were calculated by assuming the bulk density of soil to be 1 kg dm~3.These are only rough esti- mates, because the sampling sites in the fields after the trials cannot exactly coincide with those before the experiments.Furthermore, owing to the great variation between therepli- cates, the changes in the P fractions, though great, were not always statistically significant.In the following, the uncertain changes were not included in the calculations.
Al-P and Fe-P in Table 2 refer to the NH 4 F-soluble and NaOH-soluble fractions assumed to represent P mainly bound by hydrated A 1 and Fe oxides, respectively.Sev- en years of cultivation without P addition depleted these reserves by 2-15 °7o in the clay soils and by 2-17 % in the coarser soils.In absolute quantities, the reduction in the clay soils ranged from 22 to 69 kg ha~', in the Table 2. P fractions in experimental soils (kg ha -1 ) and the changes in them during seven years of cul- tivation.
Changes in plots that received totally P kg ha -1 Soil P frac-Con- lion trol 0 210 420  The H 2 S0 4 -soluble fraction, denoted as Ca-P in Table 2, decreased significantly in seven soils, the depletion being 16-34 kg ha -1 .
The slight increase in these reserves found in soil 9 may be attributable to inaccuracies in sampling.
It is noteworthy that in soil 8, the acid- soluble fraction was 2-lo fold higher than in the other soils, and remained unaffected by the fertilization regimens.Heavy dressings with rock phosphate may result in marked increases in H 2 S0 4 -P (cf.Kaila 1969), but ac- cording to the information available at the Maaninka research station, no store dressing has been performed during past decades (K.Rinne oral communication, 19 Öct., 1988).
On the other hand, the experimental field is located near the region where apatite is mined.Therefore it is likely that the high acid-soluble fraction was derived from native apatite reserves of the soil material.In acid Finnish soils, the H 2 S0 4 -soluble fraction is consid- ered to represent mainly primary apatitic P (Kaila 1964), even though in soils treated with P fertilizers the acid extractant may also dissolve secondary Ca phosphates (Kaila  1961).In studies where P is added e.g. as a K compound, this fraction has, however, been found to be quite inactive (Kaila 1964, Har-  tikainen 1979).
The P fertilization increased the NH4Fsoluble and NaOH-soluble reserves in all but four soils, for which no significant changes in these fractions were observed.The recovery of fertilizer P amounted to 11-83 % and 11-88 % of the total P additions of 210 and 420 kg ha respectively.The accumulation in the A 1 and Fe bound forms was highest in the very acid soil 11 (pH in CaCl 2 4.2) rich in oxalate-soluble Fe.In the untreated control samples, the ratio of NH 4F-P to NaOH-P correlated moderately with the ratio of Al c to Fe H , the value of r being o.7s*** (n=ls).When soil 12, which was very high in Al 0 , was included, the r value rose to o.96***.Nevertheless, the paired t-statistics revealed that in the soils amended with the lower P quantity, the NH 4 F-P/NaOH-P ratio was significantly higher than in the control sam- ples, but lower than in the soils treated with heavy P dressings.This finding indicates that the higher the P dressing was, the more intensively the added P tended to accumulate in the fluoride-soluble fraction.
The acid-soluble P seemed to increase only in soils 3, 9 and 12.This does not mean, how- ever, that no accumulation of fertilizer P as Ca compounds took place in other soils.Kai- la (1961) demonstrated that in soils recently dressed with superphosphate, dicalcium phosphate, a reaction product of monocalcium phosphate, may be attacked by the NH 4 F ex- tractant in the earlier phase of the fractiona- tion procedure.
In the plots amended annually with 60 kg of P per ha -1 , the P addition markedly exceeded the P withdrawal by yields, estimated on the basis of nutrient uptake data of Jaak- kola et ai.(1982).In many soils the recovery of residual P, however, remained quite small.This may be attributable to the uneven distribution of added P in soils, owing to the place- ment of fertilizers in bands.The uneven dis- tribution, in turn, was reflected in the large variation between replicates, which decreased the statistical significance of the observed changes.