Nutr Cycl Agroecosyst (2007) 77:283–292
DOI 10.1007/s10705-006-9066-2
ORIGINAL PAPER
The Olsen P method as an agronomic and environmental
test for predicting phosphate release from acid soils
Maria do Carmo Horta Æ José Torrent
Received: 8 June 2006 / Accepted: 17 October 2006 / Published online: 13 January 2007
Springer Science+Business Media B.V. 2006
Abstract The role of soil phosphorus (P) in the
eutrophication of fresh water systems is well
established. It is crucial therefore to assess the
potential loss of P from soil in the various
scenarios where soil can come into contact with
water. To date, such assessment has often been
based on soil P tests that are used for agronomic
purposes (e.g. fertilizer recommendations). The
purpose of this work was to examine the usefulness of one such test (viz. the Olsen test, which is
based on extraction with bicarbonate) for predicting not only the amount of soil P available to
plants, but also that which can be desorbed to
water in a group of 32 Portuguese soils, of which
29 were acid and 3 calcareous. To this end, we (i)
assessed the total amount of phytoavailable P in
soil by successively pot-cropping Chinese cabbage, buckwheat and rye; and (ii) measured the
amount of phosphate-P desorbed to a dilute
electrolyte mimicking fresh water over periods
of up to 218 days at soil:solution ratios of 1:100,
1:1000 and 1:10000. Total phytoavailable P and
M. do C. Horta
Escola Superior Agrária, Quinta Sra. Mércules
6000 Castelo Branco, Portugal
J. Torrent (&)
Departamento de Ciencias y Recursos Agrı́colas y
Forestales, Universidad de Córdoba, Edificio C4,
Campus de Rabanales, 14071 Cordoba, Spain
e-mail: [email protected]
Olsen P were found to bear a quadratic relationship, with Olsen’s extractant underestimating the
content in phytoavailable P of soils with high
Olsen P contents relatively to soils with low
contents. The ‘‘change point’’ at which phytoavailable P began to increase rapidly per unit
change in Olsen P was 53 mg Olsen P kg–1 soil.
For the acid soils, a significant quadratic relationship was found between the amount of P desorbed to water and Olsen P at the three
soil:solution ratios studied. However, these relationships became less significant when only the
soils with an Olsen P value of less than 50 mg kg–1
were considered. For the acid soils, the change
point at which P input to water began to increase
rapidly per unit change in Olsen P was 20, 61 and
57 mg kg–1 at the 1:100, 1:1000 and 1:10000 ratio,
respectively. At comparable Olsen P values, the
calcareous soils released more phosphate to water
than the acid soils. On the basis of our results, we
suggest the following environmental threshold
values for Olsen P in acid soils: 20 mg kg–1 for P
desorption scenarios where the soil:solution ratio
is high (e.g. drainage water) and 50 mg kg–1 for
desorption scenarios where the soil:solution ratio
is low (e.g., runoff, water in reservoirs). Both
values are higher than the agronomic threshold
above which plants are well supplied with P.
Keywords Acid soils Eutrophication Olsen P Phosphate Phosphorus Soil P test
123
284
Introduction
The application of fertilizer phosphorus (P) to
soil in amounts exceeding plant uptake can lead
to accumulation and subsequent loss of P to water
(Newman 1997; Sibbesen and Runge-Metzger
1995). Freshwater eutrophication caused by phosphate P can thus be partly ascribed to non-point
source pollution from agricultural land (Sharpley
et al. 1994; Sharpley 1995; Sharpley and Tunney
2000).
The amount of phosphate P desorbed from soil
to water depends, among other factors, on the
total amount of desorbable P (P0), the soil:solution ratio (S) and the time of contact (t), and can
be approximated by the equation P desorbed = KP0S–atb (Sharpley et al. 1981; Sharpley
and Ahuja 1983; Torrent and Delgado 2001),
where K, a and b are constants typical of each
soil. Thus, the amount of P desorbed to water
ultimately depends on the ‘‘desorption scenario’’
(the term ‘‘desorption’’ is used here to designate
all processes by which P is released from the solid
phase to the solution). Small S values and contact
times on the minute to hour scale are generally
involved in the case of runoff; large S values and
times on the day/week scale in the case of fast/
slowly percolating drainage water; and small S
values and times of several months are when P is
released from a soil sediment lying at the bottom
of a large standing water body (e.g. a reservoir).
Various approaches have been used to estimate the environmental risk posed by soil P
irrespective of the particular P desorption scenario. For acid sandy soils, Van der Zee et al.
(1987), and van der Zee and van Riemsdijk (1988)
proposed the ‘‘degree of phosphorus saturation’’
(DPS), which is given by DPS = Pox/0.5(Feox +
Alox), the subscript ‘‘ox’’ denoting oxalateextractable, and Pox, Feox and Alox being
expressed in mmol kg–1. The Feox + Alox combination is a measure of the content in major
phosphate-sorbing soil components of acid soils
(viz. poorly crystalline Fe and Al oxides, and
organic Fe and Al complexes) and Pox represents
P associated with these sorbents. A DPS value of
0.25 is generally taken to be the critical value
above which the risk of P loss in drainage
increases significantly. Various agronomic soil P
123
Nutr Cycl Agroecosyst (2007) 77:283–292
tests (viz. Mehlich-1 and –3, Bray–Kurtz, Olsen)
have also been used to calculate DPS (see
Maguire et al. 2005). Heckrath et al. (1995) studied the quantity/intensity relationship between
the value of Olsen P (Olsen et al. 1954)—the
quantity factor—and the concentration of P in
drainage water—the intensity factor—in one soil
and established an Olsen P ‘‘change point’’ above
which P concentration in drainage water increased rapidly. The change point is the intersection of the two regression lines best fitting the
quantity/intensity relationship. The change point
or, in general, any empirically established
‘‘threshold’’ above which losses of P to water
are considered to be unacceptably high, is a
function not only of the particular P test used, but
also of soil type and desorption scenario (Maguire
et al. 2005).
Almost twenty soil P tests are currently in use
for agronomic purposes (Tunney et al. 1998) of
which five including the Olsen test have been
employed to establish ‘‘environmental’’ threshold
values (Sharpley et al. 2002). The Olsen test,
which is based on extraction with 0.5 M sodium
bicarbonate, is an effective choice for agronomic
purposes in regions where soils are neutral or
calcareous. However, relatively less is known
about its usefulness for acid soils, whether for
agronomic or environmental purposes. The purpose of this work was to evaluate the ability of the
Olsen method to predict the P release capacity of
Portuguese acid soils. To this end, laboratory and
plant growth chamber experiments were conducted with a view to mimicking various scenarios
where desorption of P from soil to water bears
agronomic or environmental significance.
Materials and methods
Soil sampling and analysis
We collected 32 samples from representative soil
units of Portugal. Samples were obtained from A
or upper B horizons of soils developed on widely
different parent materials that were in use for
cropping or grazing. The soils belonged to the
following reference groups as per the WRB
(FAO 1998): Fluvisols (3), Regosols (3), Lithosols
Nutr Cycl Agroecosyst (2007) 77:283–292
(1), Vertisols (3), Luvisols (8), and Cambisols
(14). Three calcareous soils (two Vertisols and
one Luvisol) were included in the study for the
purpose of comparison with the acid soils. Samples were air dried, ground to pass through a
2-mm sieve, and analyzed (in duplicate) in the
laboratory for particle size distribution (pipette
method), organic matter (dichromate oxidation),
pH (potentiometric measurements in a 1:2.5
soil:water suspension), cation exchange properties (Soil Survey Staff 1992), calcium carbonate
equivalent (van Wesemael 1955), citrate/bicarbonate/dithionite-extractable Fe (Fed; Mehra and
Jackson 1960), and acid oxalate-extractable Al,
Fe and P (Alox, Feox, Pox; Schwertmann 1964).
Total P (Pt), organic P (Po), and inorganic P (Pi)
were determined according to Olsen and Sommers (1982). The method of Fox and Kamprath
(1970) was used to construct the P sorption (i.e.
quantity/intensity) curves except that 0.002 M
CaCl2 was used as the supporting electrolyte.
Sorption data were fitted to the Freundlich
equation, X = Acb, where X is phosphate-P
sorbed (mg P kg–1 soil), c is the solution phosphate concentration (mg P L–1), and A and b are
adjustable parameters. The equilibrium P concentration (EPC) was measured in the 1:1 soil:water extract obtained by shaking the soil
suspension 30 min per day over a period of
6 days. The degree of P saturation (DPS) was
calculated as indicated in the Introduction section. In all experiments, orthophosphate P in
solution was determined by the molybdate blue
method of Murphy and Riley (1962).
Pot cropping experiment
A pot cropping experiment in a growth chamber
was done in order to study how the Olsen P of the
soil changed as the result of P uptake (i.e. P
desorption) by plants. For each soil, six 200 ml
pots were filled with a substrate consisting of 150 g
acid-washed quartz sand and the following
amounts of soil: 0.5Q, 0.75Q, Q, 1.5Q, 2Q, and
3Q, Q being the amount of soil containing 3 mg of
total plant-available P, which corresponded to
1.5 g of plant dry matter containing 2 mg P kg–1;
total available P in acid soils is assumed to be five
times Olsen P on average (Delgado and Torrent
285
1997). For soils with low (<15 mg kg–1) Olsen P
values, the number of pots was reduced to 3–5 so
that the amount of soil per pot did not exceed
60 g. The pots were placed in a growth chamber
provided with a photoperiod of 16 h with day/
night temperatures of ~26/18C and a photosynthetically active radiation of ~100 lmol m–2 s–1,
and cropped successively over 30-day periods with
Chinese cabbage (Brassica pekinensis cv. Kasumi), rye (Secale cereale L.). A third cropping with
buckwheat (Fagopyrum esculentum Moench.) was
carried out when Olsen P was >10 mg kg–1. These
species were selected on the grounds of their high
P absorption capacity and adaptability to poor soil
conditions (Dechasa et al. 2003; Nanzyo et al.
2002). The plants (from germinated seeds) were
allowed to grow for 30 days before being removed
from the substrate, cleaned from soil and sand
particles, dried at 65C, and weighed. During
cropping, the pots were watered to field capacity
daily and supplied with half-strength Hoagland
solution on a regular basis to an overall amount of
45 mg of N for the crop. The concentration of P in
dry matter was determined following digestion
with nitric/perchloric acid (Zasosky and Burau
1977). After each cropping, and for each pot: (i),
the soil/sand mixture was dried and analysed for
Olsen P, the result being corrected for the
proportion of soil in the mixture, and (ii) the
cumulative P uptake (i.e. the amount of P taken
up by plants since the beginning of the experiment) was recorded. After completing the experiment, the 6–15 data points from each soil were
fitted to a logarithmic curve of the form: Cumulative P uptake = a + bln(Olsen P), where a and b
are parameters typical for each soil.
Experiments of desorption of P to a dilute
electrolyte
Long-term P desorption experiments were carried
out in order to mimic various scenarios of soil P
desorption to water. Thus, 1:100, 1:1000 and
1:10000 soil:solution ratios and respective volumes of 0.1, 1, and 1 L were used to mimic
conditions typical of drainage water, runoff water
and large bodies of standing water over a soil
sediment (e.g. lakes and reservoirs), respectively.
Before the experiment the soil was finely ground
123
286
(<1 mm) and thoroughly homogenised. The supporting electrolyte was 0.002 M CaCl2, the ionic
strength of which is similar in to that of many
continental fresh waters. The soil suspensions
were placed in polyethylene bottles and shaken at
2 Hz for 30 min the first 2 days and then once a
week. Portions of each suspension (2, 4, and 6 ml
for the 1:100, 1:1000, and 1:10000 soil:solution
ratio, respectively) were taken at day 14, 29, 59,
89, 119, 149, 179 and 218, and centrifuged at
~1.04 · 105 m s–2 for 15 min before determining
the concentration of molybdate reactive orthophosphate P in the supernatant.
Statististical analyses
Statistical analyses (viz. linear regression and
stepwise multiple regression) were done using
STATISTIX 7 (Analytical Software 2000). A
split-line model was used to describe some relationships. In this model, the curve is divided into
two linear sections so that the unaccounted
variance of the dependent variable is minimized;
the intersection of these two lines defines then the
so-called ‘change point’ (Heckrath et al. 1995).
For nonlinear regression models, curve fitting was
based on the Simplex procedure (CoHort Software 1995). In this work, the symbols *, **, and
*** denote significance at the 0.05, 0.01, and 0.001
probability (P) levels, respectively.
Results and discusssion
Soil properties
Table 1 summarizes the properties of the 32 soils.
As can be seen, texture, organic carbon content,
exchange properties, and pH varied over wide
ranges. On average, the soils were light-textured,
moderately acidic, and relatively poor in organic
carbon. The wide range spanned by the total P
concentration was a result of differences in parent
material and fertilizer practices. The differences
in fertilizer P application were clearly reflected in
the Olsen P values, which ranged from 4 to
116 mg kg–1 (i.e. from a very low to a very high P
fertility class). The soils also differed considerably
in P sorption and buffering capacity as a result of
123
Nutr Cycl Agroecosyst (2007) 77:283–292
differences in the concentrations of Fed, Feox, and
Alox. Such concentrations provide a measure of
those of the most active P-sorbing soil components (basically, oxides and organic complexes of
Fe and Al, and edges of silicate clays). The values
shown in Table 1 changed little when the three
calcareous soils were excluded, except that the
maximum pH (water) was 7.1 and the maximum
clay content 360 g kg– 1.
Plant P uptake in relation to Olsen P
At the end of the pot experiment, Olsen P had
decreased by 73% on average (Table 2); however,
the decrease was only ~30% in P-poor soils. The
cumulative P uptake by plants as a function of
Olsen P generally fitted the logarithmic model
described before, as illustrated in Figure 1 for soil
CH-209. We took phytoavailable P to be the
cumulative amount of P taken up by plants when
Olsen P had reached a value of 6 mg kg–1, as
estimated from the fitted logarithmic curve. The
rationale for this choice was that previous P
uptake experiments had shown plant growth and
P uptake to be substantially reduced, but not
necessarily stopped near this value (Matar et al.
1992; Delgado 1996). The content in the sodefined phytoavailable P ranged widely (Table 2),
as expected from the wide range of Olsen P
values exhibited by the soils.
The P buffering capacity under plant uptake
conditions was calculated as the absolute value of
the derivative of the fitted logarithmic equation.
The P buffering capacity at Olsen P = 6 mg kg–1
(BC6, Table 2) differed widely among soils, with
an average value of 6 mg mg–1 (i.e., plants took up
6 mg P kg–1 soil for a decrease of 1 mg kg–1 in
Olsen P). If the four soils with Olsen P >50 mg
P kg–1 and the calcareous soils were excluded, the
average was 3.6 mg mg–1. The average value for
the three calcareous soils was 2.6 mg mg–1. These
average values are similar to the values obtained
by Delgado and Torrent (1997) for moderately
acidic and calcareous European soils, respectively.
Phytoavailable and Olsen P were found to be
strongly correlated (R = 0.93, P < 0.001). The
relationship between the two variables (excluding
the calcareous soils) can be described by a
quadratic equation (Fig. 2a); however, the
Nutr Cycl Agroecosyst (2007) 77:283–292
287
Table 1 Selected properties of the soils
Property
Unit
Minimum
Maximum
Mean ± SD
Median
Clay
Sand
Organic C
pHwater
CCEa
ECEC
Exchangeable Al
Base saturation
Feda
Feoxa
Aloxa
Poxa
Olsen P
Total P
Inorganic P
Organic P
Aa
BC1a
DPS
EPCa
g kg–1
g kg–1
g kg–1
2.5
207
1.7
5.0
0
0.7
<0.01
21
0.2
1.3
2.5
0.5
2.5
91
26
5
25
5
3
0.02
565
997
30.5
8.0
165
74
1.89
100
31.3
82.7
282.6
35.5
116.1
1728
1312
384
580
157
93
4.4
133 ± 117
686 ± 185
10.3 ± 6.2
5.9 ± 0.8
84 ± 72b
11.9 ± 14.5
0.26 ± 0.50
61 ± 23
7.9 ± 9.4
25.5 ± 19.6
36.2 ± 54.7
6.17 ± 7.25
29.4 ± 28.8
475 ± 364
251 ± 289
129 ± 87
146 ± 108
40 ± 36
26 ± 21
0.39 ± 0.49
91
741
8.5
5.7
61b
6.7
0.44
55
3.0
17.9
20.2
4.20
20.2
365
146
111
109
29
19
0.19
g kg–1
cmolc kg–1
cmolc kg–1
%
g kg–1
mmol kg–1
mmol kg–1
mmol kg–1
mg kg–1
mg kg–1
mg kg–1
mg kg–1
mg kg–1
L kg–1
%
mg L–1
a
Abbreviations: CCE, calcium carbonate equivalent; ECEC, effective cation exchange capacity; subscript d, citrate/
bicarbonate/dithionite-extractable; subscript ox, acid oxalate-extractable; A and BC1, phosphate-P sorbed and phosphate
buffering capacity, respectively, at an equilibrium concentration of 1 mg P L– 1 (calculated from the fitted Freundlich
curve); DPS, degree of P saturation; EPC, equilibrium P concentration
b
Three soils
Table 2 Selected phosphorus desorption properties
Property
Unit
Phytoavailable Pa
DOlsen Pb
BC6b
P1:100b
P1:1000b
P1:10000b
mg
%
mg
mg
mg
mg
kg–
1
mg–1
kg– 1
kg– 1
kg– 1
Minimum
Maximum
Mean ± SD
Median
0.1
29
0.1
1.2
3.6
4.5
746.1
93
39.7
85.6
283.3
464.5
85.2
73
6.1
12.5
44.5
61.9
17.8
76
2.8
5.7
33.9
23.5
±
±
±
±
±
±
163.2
13
8.3
16.6
50.5
96.7
a
Cumulative amount of P taken up by plants when Olsen P reached a value of 6 mg kg–1, as estimated from the fitted
Cumulative P uptake/Olsen P curve
b
D Olsen P, decrease in change in Olsen P as a result of cropping; BC6, plant-uptake buffer capacity at Olsen P = 6 mg kg– 1;
P1:100, P1:1000, P1:10000, amount of P released to 0.002 M CaCl2 at a 1:100, 1:1000, and 1:10000 soil:solution ratio, respectively
significance of such an equation is weaker if the
soils with Olsen P >50 mg kg– 1 are also excluded (Figure 2b). The change point at which
phytoavailable P began to increase rapidly per
unit change in Olsen P was 53 mg kg–1 soil
(Fig. 2c). The data points for the calcareous soils
(Olsen P values = 10.4–11.5 mg kg–1) lay close to
the regression line for the acid soils (Fig. 2c). This
suggests that the relation between phytoavailable
P and Olsen P might be relatively similar for acid
and calcareous soils, at least for soils with Olsen P
values in the 10–12 mg kg–1 range.
As noted by Menon et al. (1988), Lin et al.
(1991), Kleinman et al. (2001) and Bah et al.
(2003) among others, Olsen’s method performs
well in alkaline and moderately weathered soils,
the latter being the case with many Portuguese
acid soils. In summary, bicarbonate seems capable of extracting an amount of P from these soils
that relates to the amount of phosphate released
123
288
Nutr Cycl Agroecosyst (2007) 77:283–292
a
b
Fig. 1 Relationship between cumulative plant P uptake
and Olsen P for soil CH-209
to the solution and taken up by plants in a period
appropriate for plant growth. However, one
cannot expect bicarbonate to be a good proxy
for the different capacity of plants to take up the
different phytoavailable P forms present in acid
soils (mainly phosphate adsorbed on Al– and
Fe–OH mineral surfaces and Al- and Fe-rich
metal phosphates). Because the proportions of
the various P forms in the studied soils are likely
to range widely, the goodness of the relationship
between phytoavailable and Olsen P is limited, as
clearly shown by Fig. 2.
One interesting conclusion from Fig. 2 is that
Olsen’s extractant underestimates the concentration of phytoavailable P in soils with high Olsen P
values relative to soils poor in this P form. This, as
noted previously, was reflected in the high BC6
values for the soils rich relative to those poor in
Olsen P. Similar results were reported by Delgado (1996) for a group of overfertilized European soils exhibiting a wide pH range. This
underestimation can be partly due to the low
soil:solution ratio (1:20) of the extractant, which
can lead to phosphate supersaturation in the
extraction medium —mainly in P-rich soils— and
readsorption of phosphate (Barrow and Shaw
1976). Mineralization of organic P, which occurs
to a great extent in some manure-fertilized soils,
might also supply some phosphate to the solution.
This might in turn contribute to the underestimation because soluble organic P goes undetected in
the usual analyses for molybdate-reactive P in
Olsen’s extractant.
123
c
Fig. 2 Relationship between phytoavailable P and Olsen
P for (a) the acid soils [a quadratic fit is shown]; (b) the
acid soils with Olsen P <50 mg kg–1 [a quadratic fit is
shown]; and (c) the acid soils [a ‘‘change-point’’ fit is
shown]. Filled triangles: acid soils; open inverse triangles:
calcareous soils (not used in the curve or split-line fitting).
The significance level is indicated in brackets after the
corresponding equation
Relationship between the amount of P
released to water and Olsen P
The amount of P desorbed from the soil to
0.002 M CaCl2 (i.e. ‘‘simulated fresh water’’) was
found to be a function of the soil:solution ratio
Nutr Cycl Agroecosyst (2007) 77:283–292
Fig. 3 Time course of P desorption to 0.002 M CaCl2 at a
soil:solution ratio of 1:100, 1:1000, and 1:10000 for soils
CH-204, CH-210 and CH-222
and time of contact. The time course of P
desorption (and the corresponding concentration
of P in solution) differed widely among soils, as
illustrated by Fig. 3, and generally exhibited wide
oscillations. Microbial activity—no biocide was
added to the suspensions—precipitation of new
phosphate phases, and readsorption of desorbed
289
phosphate are the likely sources of such oscillations. The amount of P desorbed generally
increased with decreasing soil:solution ratio, as
previously found by Yli-Halla et al. (2002), for
example. This was not the case with some soils
with Olsen P < 8 mg kg–1 or with a high P
buffering capacity and low P saturation, where
the concentration of P in solution was very low
and hence (i) strongly influenced by microbial
activity and (ii) subject to significant analytical
errors. The concentration of P in solution was of
the same order of magnitude after a few days of
contact as it was at the end of the experiment. In
order to compare soils, we used the average P
concentration in solution between days 149 and
218 to calculate the amount of P desorbed in the
long term.
The relationships between long-term desorbed
P and Olsen P in the acid soils at the different
soil:solution ratios are shown in Fig. 4. The
change-points at which the amount of desorbed
P began to increase rapidly per unit change in
Olsen P were 20, 61 and 57 mg Olsen P kg–1 soil
for the 1:100, 1:1000 and 1:10000 soil:solution
ratio, respectively. The 20 mg kg–1 value for the
1:100 soil:solution ratio, which somewhat mimics
a scenario of water percolating through the soil, is
similar to that found by Zhang et al. (2005) as
critical for the release of P from urban soils to
drainage water; however, this value should be
taken with the caveat that the regression line for
the higher Olsen P values in Fig. 4a is not
significant, and the difference in slope between
the two lines is significant only at P = 0.15. Other
reported Olsen P change point values for drainage water are 60 mg kg–1 (Heckrath et al. 1995),
36 mg kg–1 (McDowell and Sharpley 2001), and
10–20 mg kg–1 (Hesketh and Brookes 2000;
McDowell et al. 2001). This apparent discrepancy
is obviously due to differences in soil:solution
ratio, time of soil:water contact, temperature, and
soil type and depth among the different P
desorption scenarios were the change point was
determined. The solution P concentration at the
change point was 0.051 mg L–1, which would
correspond to the loss of only about 50 g P ha–1
per 100 mm of drainage water. Therefore,
20 mg kg–1 can be adopted as a conservative
threshold value for environmental purposes. It
123
290
Nutr Cycl Agroecosyst (2007) 77:283–292
a)
b)
c)
Fig. 4 Relationship between P desorbed at a soil solution
ratio of (a) 1:100, (b) 1:1000, and (c) 1:10000 with Olsen P
for the acid soils (the change point is shown). Filled
triangles: acid soils; open inverse triangles: calcareous soils
(not used in the split-line fitting). The significance level is
indicated in brackets after the corresponding equation
should be noted that this value is acceptable from
an agronomic standpoint because typical threshold values for soils well supplied with P are
15–20 mg kg–1.
Although the significance of regression lines
for the higher Olsen P values was low both for the
123
1:1000 and the 1:10000 soil:solution ratios
(P < 0.25 and <0.10, respectively), the slopes of
the regression lines used to calculate the change
point were significantly different (P < 0.01 and
<0.001, respectively). On the basis of these results
and the values of the change points, a conservative value for 50 mg kg–1 can be proposed as a
‘safe’ change point for the corresponding desorption scenarios.
Based on Figure 4, the change point is consistent with a P concentration in solution of
~0.05 mg L–1 for the 1:1000 ratio and
~0.06 mg L–1for the 1:10000 ratio. It should be
noted that 0.05 mg L–1 is considered to be the
critical level in overflow water if eutrophication of
water bodies by P is to be prevented (Golterman
and Oude 1991).
Quadratic curves of the form P desorbed =
a + b · (Olsen P)2, where a and b are positive,
were found to fit the relationships for the acid
soils depicted in Fig. 4. Such curves were found to
explain 65, 73 and 92% of the variance in the
amount of P desorbed at the 1:100; 1:1000 and
1:10000 ratios, respectively. However, if the soils
with Olsen P > 50 mg kg–1 were excluded, then
the percent variance accounted for decreased to
56, 29 and 35%, respectively. It must also be
noted that the data point for the calcareous soils
in Fig. 4 lay markedly above the regression lines
for the acid soils for the 1:1000 and the 1:10000
soil:solution ratios. This can be tentatively
ascribed to the dissolution of Ca phosphates
—which are a major form of accumulation of P
in these soils—in highly diluted suspensions.
The poor performance of Olsen P in predicting
desorption of P to water at small Olsen P
values—particularly at low soil:solution ratios—
makes it advisable to consider other soil properties related to P dynamics in order to improve
predictions. Thus, we found the following regression equations to account for 60, 73 and 55%,
respectively, of the variance in the amount of P
desorbed from the acid soils with Olsen
P < 50 mg kg–1:
P desorbed at 1 : 100 ¼ 4:8 þ 0:0096
ðOlsen PÞ2 0:097 FeOX
ð1Þ
Nutr Cycl Agroecosyst (2007) 77:283–292
P desorbed at 1 : 1000 ¼ 11 þ 0:0175
ðOlsen PÞ2 þ 0:070 Fed 0:074 Alox
P desorbed at 1 : 10000 ¼ 1 þ 0:025
ðOlsen PÞ2 þ 0:070 Fed
291
ð2Þ
ð3Þ
where desorbed P and Olsen P are expressed in
mg kg–1, and Fed, Alox and Feox in mmol kg–1.
Equations (1–3) suggest that P desorption from
soil to water is influenced by the nature and
extension of the P-sorbing surfaces, —mainly
those with surface Al– and Fe–OH groups—as
measured by the amounts of Al and Fe dissolved
in acid oxalate or citrate/bicarbonate/dithionite.
The different signs of the regression coefficients
of Fed, Alox and Feox suggest that, depending on
the soil:solution ratio, the sorbing surfaces either
release the sorbed phosphate or act as sinks of the
phosphate released from other sources (e.g.
dissolving metal phosphates).
The Olsen P method as an agronomic
and environmental soil P test
The transfer of P from soil to water is controlled,
among other factors, by (i) the total amount of
desorbable P; (ii) the partition of P in the soil
solid phase between adsorbed and precipitated
forms; (iii) the kinetics of phosphate desorption
from various adsorption sites; and (iv) the dissolution kinetics of metal phosphates—mainly those
rich in Al, Fe, and Ca. Olsen’s procedure, which is
based on a short extraction with bicarbonate,
seems capable of mimicking desorption of the
different forms of soil P to water in some
scenarios, but not in others. In our study, Olsen
P proved a reasonably good predictor of the
amount of P released to the soil solution as a
response to depletion through uptake by plants. It
was also an acceptable predictor for P inputs to
water at long times and low soil:solution ratios.
Our experiments showed that Olsen P is more
useful as an agronomic P test than as an environmental one in acid soils. However, they allowed
us to tentatively establish some environmentally
useful threshold values, namely: ~20 mg kg–1 for
P desorption scenarios where the soil:solution
ratio is high (e.g. drainage) and ~50 mg kg–1 for
scenarios involving dilute soil suspensions (e.g.
runoff and water in reservoirs). It should be noted
that both values are higher than the ‘critical’ (i.e.
threshold) levels above which most cultivated
plants are generally considered to be well supplied with P (Sharpley et al. 2002). In summary,
there seems to be some margin for meeting plant
needs with no significant risk of eutrophication by
phosphate.
Acknowledgements The senior author acknowledges
award of a grant from PRODEP III (2/2001) supported
by the Portuguese government and the European Union.
References
Analytical Software (2000) Statistix 7 User’s Manual.
Analytical Software, Tallahassee, FL
Bah AR, Zaharah AR, Hussin MHA, Halimi MS (2003)
Phosphorus status of amended soil as assessed by
conventional and isotopic methods. Commun Soil Sci
Plant Anal 34:2659–2681
Barrow NJ, Shaw TC (1976) Sodium bicarbonate as an
extractant for soil phosphate, I. Separation of the
factors affecting the amount of phosphate displaced
from soil from those affecting secondary adsorption.
Geoderma 16:91–107
CoHort Software (1995) Coplot manual. CoHort Software,
Minneapolis, MN
Dechasa N, Schenk MK, Claassen N, Ateingrobe B (2003)
Phosphorus efficiency of cabbage (Brassica oleraceae
L. var. capitata), carrot (Daucus carota L) and potato
(Solanum tuberosum L). Plant Soil 250:215–224
Delgado A (1996) Liberación de fosfato en suelos sobrefertilizados de la Union Europea. PhD Thesis, Universidad de Córdoba, Córdoba, Spain (in Spanish)
Delgado A, Torrent J (1997) Phosphate-rich soils in the
European Union: estimating total plant-available
phosphorus. Eur J Agron 6:205–214
Fao, Isric, Isss (1998) World Reference Base for Soil
Resources. Fao, Rome
Fox RL, Kamprath EJ (1970) Phosphate sorption isotherms for evaluating the phosphate requirements of
soils. Soil Sci Soc Am Proc 34:903–906
Golterman HL, Oude NT (1991) Eutrophication of lakes,
rivers and coastal seas. In: Hutzinger O (eds), The
Handbook of Environmental Chemistry, Vol. 5, Part
A. Springer-Verlag, Berlin, pp 80–124
Heckrath G, Brookes PC, Poulton PR, Goulding KWT
(1995) Phosphorus leaching from soils containing
different phosphorus concentrations in the Broadbalk
experiment. J Environ Qual 24:904–910
Hesketh N, Brookes PC (2000) Development of an
indicator for risk of phosphorus leaching. J Environ
Qual 29:105–110
Kleinman PJA, Sharpley AN, Garley K, Jarrel WM, Kuo
S, Menon RG, Myers R, Reddy KR, Skogley EO
123
292
(2001) Interlaboratory comparison of soil phosphorus
extracted by various soil test methods. Commun Soil
Sci Plant Anal 32:2325–2345
Lin TH, Ho SB, Houng KH (1991) The use of iron oxideimpregnated filter paper for the extraction of available phosphorus from Taiwan soils. Plant Soil
133:219–226
Maguire RO, Chardon WJ, Simard RR (2005) Assessing
potential environmental impacts of soil phosphorus by
soil testing. In: Sims JT, Sharpley AN (eds), Phosphorus: Agriculture and the Environment. ASA
CSSA SSSA, Madison WI, pp 145–180
Matar A, Torrent J, Ryan J (1992) Soil and fertilizer
phosphorus and crop responses in the dryland Mediterranean zone. Adv Soil Sci 18:82–146
McDowell R, Sharpley AN (2001) Aproximating phosphorus release from soils to surface runoff and
subsurface drainage. J Environ Qual 30:508–520
McDowell R, Sharpley AN, Brookes P, Poulton P (2001)
Relationship between soil test phosphorus and phosphorus release to solution. Soil Sci 166:137–149
Mehra OP, Jackson ML (1960) Iron oxide removal from
soils and clays by a dithionite-citrate system buffered
with sodium bicarbonate. Clays Clay Miner 7:317–327
Menon RG, Hammond LL, Sissing HA (1988) Determination of plant-available phosphorus by the iron
hydroxide-impregnated filter paper (Pi) soil test. Soil
Sci Soc Am J 53:110–115
Murphy J, Riley J (1962) A modified single solution
method for the determination of phosphate in natural
waters. Anal Chim Acta 27:31–36
Nanzyo M, Shibata Y, Nobuko W (2002) Complete
contact of Brassica roots with phosphorus fertilizer
in a phosphorus-deficient soil. Soil Sci Plant Nutr
48:847–853
Newman EI (1997) Phosphorus balance of contrasting
farming systems, past and present Can food production be sustainable? J Appl Ecol 34:1334–1347
Olsen S, Cole C, Watanabe F, Dean L (1954) Estimation
of available phosphorus in soils by extraction with
sodium bicarbonate. USDA Circular Nr 939, US Gov.
Print. Office, Washington, D.C.
Olsen SR, Sommers LE (1982) Phosphorus. In: Page AL,
et al (eds), Methods of Soil Analysis, Part 2, 2nd edn,
Agron Monogr 9. ASA and ASSA, Madison WI, pp
403–430
Schwertmann U (1964) Differenzierung der Eisenoxide
des Bodens durch Extraktion mit AmmoniumoxalatLösung. Z Pflanzenernähr Düng Bodenkd 105:194–
202
Sibbesen E, Runge-Metzger A (1995) Phosphorus balance
in European agriculture—Status and policy options.
In: Tiessen H (eds), SCOPE 54: Phosphorus in the
123
Nutr Cycl Agroecosyst (2007) 77:283–292
Global Environment. Transfers, Cycles and Management. John Wiley & Sons, New York, pp 43–58
Sharpley AN (1995) Dependence of runoff phosphorus on
extractable soil phosphorus. J Environ Qual 24:920–
926
Sharpley AN, Ahuja LR (1983) A diffusion interpretation
of soil phosphorus desorption. Soil Sci 135:322–326
Sharpley AN, Tunney H (2000) Phosphorus research
strategies to meet agricultural and environmental
challenges of the 21st century. J Environ Qual
29:176–181
Sharpley AN, Ahuja LR, Yamamoto M, Menzel RG
(1981) The kinetics of phosphorus desorption from
soil. Soil Sci Soc Am J 47:462–467
Sharpley AN, Kleinman PJA, McDowell RW, Gitau M,
Bryant RB (2002) Modelling phosphorus transport in
agricultural watersheds: Processes and possibilities. J
Soil Water Conserv 57:425–439
Sharpley AN, Chapra SC, Wedepohl R, Sims JT, Daniel
TC, Reddy KR (1994) Managing agricultural phosphorus for protection of surface waters: Issues and
options. J Environ Qual 23:437–451
Soil Survey Staff (1992) Soil survey laboratory methods
manual. Soil Survey Investigations Report 42.
USDA–SCS, National Soil Survey Center, Lincoln,
NE
Torrent J, Delgado A (2001) Using phosphorus concentration in the soil solution to predict phosphorus
desorption to water. J Environ Qual 30:1829–1835
Tunney H, Breeuwsma A, Withers PJA, Ehlert PAI (1998)
Phosphorus fertilizer strategies: present and future.
In: Tunney H, et al (eds), Phosphorus Loss from Soil
to Water, Chp 8. CAB International, Wallingford UK,
pp 177–203
van der Zee SEATM, Fokking LGJ, van Riemsdijk WH
(1987) A new technique for assessment of reversibly
adsorbed phosphate. Soil Sci Soc Am J 51:599–604
van der Zee SEATM, van Riemsdijk WH (1988) Model
for long-term phosphate reaction kinetics in soil. J
Environ Qual 17:35–41
van Wesemael JCh (1955) De bepaling van het calciumcarbonaat-gehalte van gronden. Chem Weekblad
51:35–36
Yli-Halla M, Hartikainen H, Vaatainen P (2002) Depletion of soil phosphorus as assessed by several indices
of phosphorus supplying power. Eur J Soil Sci 53:431–
438
Zasosky RJ, Burau RG (1977) A rapid nitric–perchloric
acid digestion method for multielement tissue analysis. Commun Soil Sci Plant Anal 8:425–436
Zhang GL, Burghardt W, Yang JL (2005) Chemical
criteria to assess risk of phosphorus leaching from
urban soils. Pedosphere 15:72–77