Abstract: This study investigated Phosphorus (P) sorption characteristics of 60 soil samples formed over 6 different parent materials in south-eastern Nigeria relating that to some soil properties. Soils were analyzed for pH, clay, Bulk Density (BD), Organic Carbon (OC), Mehlich-3 extractable calcium (CaM3) ammonium oxalate and Mehlich-3 extractable P, aluminium (Al) and iron (Fe). The oxalate and Mehlich-3 extractable elements were used in estimating Phosphorus Saturation Index (PSI). There were significant relationships between P saturation indices and soil attributes in the study site at p&le0.01 and p&le0.05. However, these correlations were least in shale-derived soils due to high content of CaM3. A P-sorption capacity model resulted from a multiple regression analysis using soil attributes as independent variables. The model showed high predictive ability (r2 = 0.77; RMSE = 3.70) with little underestimation (bias = -0.09).
INTRODUCTION
Phosphorus (P) is a critical element in natural and agricultural ecosystems throughout the world. It is a key component of cellular compounds and vital to both plant and animal life. It is next to nitrogen considering its importance for profitable yields (Brady and Weil, 1999). Although P is considered to be relatively immobile in the soil system (Johnson et al., 1997), there are mechanisms through which it can leave the soil through plant uptake, loss in surface runoff, erosion of sediment and leaching through the soil (2005). Application of P-fertilizers in excess of crop requirement leads to its build-up in soils and increases the chances of P-loss from the soil system (Sharpley et al., 1999) and its entry into the runoff is a primary source of non-point soil pollution (Daniel et al., 1994) and eutrophication of water bodies (Sharpley et al., 1999).
The degree to which P becomes available in soils is governed by one or a combination of chemical and mineralogical properties, such as clay type content, iron and aluminium oxides, organic carbon, pH and calcium carbonate content (Burt et al., 2002). Significant correlations have been reported between soil P adsorption and soil pH (Brennan et al., 1994; Dodor and Oya, 2000; Khare et al., 2004), organic carbon (Dodor and Oya, 2000), aluminium and iron oxide content (Sanyal et al., 1993; Borling et al., 2001). The presence of organic matter reduces the phosphate sorption capacity of highly weathered tropical soils such as oxisols and ultisols possibly due to direct result of competition for sorption sites between phosphate and organic ligands (Gillman et al., 1989). But these tropical soils have suffered drastic reductions in soil organic matter due to soil erosion and inappropriate land uses (Mbagwu, 1992; Lal, 1995; Styczen, 1992) thereby making P-predictions inconsistent in most tropical soils (Oti, 2002).
Most soil properties used in predicting soil P are interrelated and this makes it difficult to identify the principal components contributing to P-availability in soils. Several studies used path analysis (Cramer and Wehner, 2000; Zheng et al., 2002; Garcia del Moral et al., 2003) and multiple regressions to evaluate the effect of different combinations of soil properties on P adsorption (Brennan et al., 1994; Dodor and Oya, 2000; Borling et al., 2001).
The major objective of this study was to ascertain sorption characteristics in highly weathered, soils of south-eastern Nigeria and relating components of the prediction to P-sorption capacity.
MATERIALS AND METHODS
Site
The lowland states of south-eastern Nigeria, namely Abia and Imo States
lying between latitudes 4°401 and 8°151N and longitudes
6°401 and 8°151 E (Federal Department of Agricultural
Land resources, 1985) were chosen for the study. The parent materials underlying
soils of the study area are Alluvium, Coastal Plain Sands, Shale, Lower Coal
Measures, Upper Coal Measures and Falsebedded Sandstones (Ofomata, 1987). The
site belongs to the humid tropics and its climate and vegetation pattern have
been reported by Igwe and Stahr (2004). Pedogenic characterization of soils
of the area has been done using 30 pedons and soil classes identified include
Typic Udipsamments (Dystric Fluvisols), Arenic Paleudults/Typic Paleudults (Dystric
Nitisols), Arenic Rhodic Paleudults (Dystric Acrisols), Dystrustepts/Typic Paleustults
(Dystric Cambisols), Vertic Paleudults (Vertic Nitisols) and Typic Paleustults
(Plinthic Nitisols) (Onweremadu, 2006). These classifications were done using
Soil Taxonomy (Soil Survey Staff, 2003) and correlated with FAO/UNESCO legend,
(FAO, 1998) implying that these are highly weathered soils. Farming is a major
socio-economic activity in the area.
Field Sampling
Sixty soil samples were collected representing the diversity of soils of
the study area under 4 years of fallow. Ten soil samples were collected over
each of the 6 main parent materials of the study site. The soils were sampled
from plow layer (0-20 cm) air-dried and sieved to pass a 2 mm mesh sieve.
Laboratory Determinations
Soil pH was measured in 1: 2.5 suspension of soil in 0.1 M KCl (Hendershot
et al., 1993). Total C (g kg-1) was measured by combustion
with an Elementar CNS analyzer (Elementar America, Inc. Mt. Laurel, N.J) according
to the methods described by Nelson and Sommers (1996). Percent clay content
was determined by Bouyoucos hydrometer method (Gee and Or, 2002). Bulk density
was estimated by core method of Grossman and Reinsch (2002).
Phosphorous adsorption isotherm was estimated according to the method of Graetz and Nair (2000). A gram of soil sample was equilibrated with 25 mL of varying concentrations of P in 0.01 CaCl2 solution in 50 mL centrifuge tubes. The following concentrations of the solutions were used: 0.0, 0.5, 1.0, 10.00, 15.0 and 20.0 mg P l-1.
The centrifuge tubes were shaken for 1 day on an end-to-end shaker at 150 oscillations per minute. The samples were centrifuged for 10 min at 5211 x g and the supernatant was decanted. Quantification of P in solution was done colourimetrically according to the procedures of the ascorbic acid method (Kuo, 1996). Phosphorous adsorbed was measured as the difference between initial and final amounts of P in solution. All these analyses were duplicated on all studied soils.
Extractable P, Aluminium (Al) and iron (Fe) were determined by shaking duplicate 1.5 g samples of soils with 30 mL of 0.5 M acidified ammonium oxalate (COONH4)2. H2O at pH 3.0 in 50 mL centrifuge tubes (Schoumans, 2000). Later samples were shaken for 120 min in the dark on an end-to-end shaker at 150 oscillations per minute and centrifuged for 10 min at 5211 x g. Resulting supernatants were analyzed for P, Al and Fe using a TJA-9000 inductively coupled plasma-atomic emission spectroscopy (ICP-AES).
Again extractable P, Al and Fe were estimated by shaking duplicate 2 g samples of soil and 20 mL Mehlich-3 solution (M3) in 50 mL centrifuge tubes for 10 min on an end-to-end shaker at 150 oscillations per minute. Then, samples were centrifuged at 5211 x g and supernatants analyzed by ICP-AES.
Computations
Phosphorus Saturation Index (PSI) was calculated based on ammonium oxalate
extraction (Schoumans, 2000). It is given by the formula:
Where PSIox = P saturation index (oxalate)
Pox, Alox and Feox = oxalate extractable P,
Al and Fe, respectively
Phosphorous saturation index was also computed using Mehlich-3 extractions of
P, Al and Fe (Kleinman and Sharpley, 2002). This is given as follows:
Where PSIM3 = P saturation index (Mehlich-3)
PM3, AlM3 and FeM3 = Mehlich-3 extractable
P, Al and Fe, respectively
Data Analyses
Soil data were subjected to correlation and multiple regression analyses
using SAS computer package (SAS, 2000). Root Mean Square Error (RMSE) was used
to evaluate average prediction uncertainly of the model while bias of the model
was used to assess overestimation or underestimation (Moldrup et al.,
2004).
RESULTS AND DISCUSSION
Soil Properties
Clay content ranged from 60-400 g kg-1 indicating variability
in the genesis of soils. Sandier parent materials such as Upper Coal Measures,
Alluvium and Coastal Plain Sands gave rise to less clayey soils. Soils were
moderately to strongly acidic (4.2-5.8) which could be attributed to the combined
influence of parent materials and climate on pedogenesis. Shale-derived soils
in the study area have intercalations of limestones which are inherently alkaline.
However, heavy rainfall and the attendant leaching may have washed away basic cations from the pedon thereby making soils acidic. Ranged after bulk density from 1.31 to 1.44 M gm-3 in response to differential distribution of organic carbon among soil groups (Table 1).
Phosphorous Sorption Characteristics
Phosphorous extracted by acidified ammonium oxalate (Pox) ranged
from 51.0 to 151.0 mg kg-1 soil with a median value of 120 mg kg-1
while Alox and Feox varied from 191.0 to 951.0 and 221.0
to 2201.0 mg kg-1 soils, respectively. Phosphorus saturation index
using ammonium oxalate extractant varied from 4.0 to 31.0%. Mehlch-3 extraction
showed variation in distribution and as follows: PM3 (13-70 mg kg-1),
AlM3 (191-661 mg kg-1), FeM3 (28-341 mg kg-1)
and PSlM3 (1.9-240%). These distributions were considered irrespective
of soil group (Table 2).
| Table 1: | Selected properties of soils studied (mean values of soil groups) |
PM = Parent Material, BD = Bulk Density, OC = Organic Carbon, a = soil group represents all soils collected over the same parent material |
|
| Table 2: | Phosphorus sorption characteristics of soils studied (mean values of soil groups) |
Phosphorus saturation (Psat) in the study site ranged from 7 to 75%. High Psat values indicate that soils derived from Lower Coal Measures and Falsebedded Sandstones are greatly filled with P when compared with other soil groups. Phosphorus saturation is a measure of potential desorbability of soil P (Beauchemim and Simard, 1999). Phosphorus desorption correlates highly with P saturation (Sibbesen and Sharpley, 1997) stating that P description increases with higher degrees of P saturation. The environmental implication of this in the study site is that more P is available for runoff and consequent non-point pollution and eutrophication in soils formed over Lower Coal Measures, Falsebedded Sandstones and Upper Coal Measures.
Table 3 shows the degree of relationship between Psat and some soil properties, indicating a varying influence of soil properties on Degree of P Saturation (DPS). In these tropical soils, organic carbon and clay exhibited profound relationships with Psat having r = -0.90 and r = +0.82, respectively. Similar findings were made by Dodor and Oya (2000). The presence of organic matter reduces the phosphate sorption capacity of oxisols and ultisols (Gillman et al., 1989). This could be the direct result of competition for sorption sites between phosphates and organic ligands. Organic matter (organic carbon) reduces positive charge on variable charge surfaces by lowering pHo (pH value at which the net charge due to variable charge components is zero) thus decreasing attraction of phosphates to the soil surface. The influence of organic matter on cation exchange capacity is not as pronounced in ultisols as in oxisols, as it has little effect on pHo.
There was significant negative correlation between Ca and Psat (r = -0.46), showing that competition exists between them and other polyvalent cations and anions. Giesler et al. (2005) reported that competition between phosphate and other anions may affect P sorption capacity thereby confirming the works of Shang et al. (1990) and Celi et al. (2003) that organophosphorus compounds, such as inositol phosphates. Again, ionic strength influences nature of competition. Previous studies (Beck et al., 1999; Arbestain et al., 2002) have shown that phosphate can displace Dissolved Organic Carbon (DOC) from sorption sites.
| Table 3: | Correlation coefficients between P saturation (Psat) with selected soil properties (n = 60) |
vs = Versus, ** Significant at p<0.01, * Significant at p<0.05 |
|
| Table 4: | Correlation coefficients between phosphorus saturation indices (n = 60) |
vs = Versus, ** Significant at p&le0.01, * Significant at p&le0.05 |
|
Decrease in ionic strength decreases DOC concentration (Kalbitz et al., 2000) which is explained by a reduction in divalent cations, such as Ca2+ thereby increasing phosphate sorption. Highly weathered soils of the humid tropics have modest amounts of Ca, Mg and K hence characterized by high P sorption capacity.
Relationships Among Psat Indices
Table 4 shows the degree of correlation between Psat indices
in the study site. Phosphorus saturation using acidified ammonium oxalate was
highly correlated (r = 0.89; &ge 0.01) with PSlM3 for all 60 soil
samples from the study site. These results are consistent with the reports of
Maguire and Sims (2002). The same trend was observed between Psat and PSlox
(r = 0.82; p&le0.01). Again, P sorption characteristics had significant
positive correlations in the study area. However, correlation coefficients between
P saturation indices varied among soil groups (Table 5). Least
correlation values were recorded on soils derived from Shale and Low Coal Measures.
These soil groups contained higher values of Ca extracted with M3, having 525.3
and 387.2 mg kg-1, respectively, suggesting that soils contained
greater concentrations of carbonates which may have interfered with Mehlich-3
and oxalate extractions of Al and Fe. Soils derived from Alluvium, Coastal Plain
Sands and Upper Coal Measures are highly weathered based on low silt-clay ratios
reported by Onweremadu (2006).
These highly weathered soils are generally subjected to harsh climate to produce clay fractions consisting of low activity minerals such as kaolinite, Al and Fe oxides (Gillman et al., 1989). These Al and Fe form surface sites where organophosphates sorb (Cali et al., 2003; Giesler et al., 2005).
Modelling
Step-wise multiple regression analysis using soil pH, OC, Alox,
Feox, clay and B.D. were used to attempt modelling of P sorption
capacity in these tropical soils (Table 6). With fairly low
coefficient of alienation (1-R2 = 0.23), high prediction certainty
was achieved (RMSE = 3.70) with little underestimation (bias = -0.09). Because
the soils of the study site have an abundance of Fe and Al (Igwe et al.,
2005) which affect solubility of phosphorus (Giesler et al., 2005) this
model becomes relevant in estimating P availability to plants and micro-organisms
(Giesler, 2002) and ground water discharge rate (Pellerin, 2002) and possible
transfer of P from terrestrial to aquatic ecosystem (Lyons et al., 1998).
| Table 5: | Correlation coefficients between phosphorus saturation indices among soil groups (n = 10) |
| Table 6: | Phosphorus sorption capacity model attributes |
CONCLUSION
A good knowledge of P sorption capacity of highly weathered tropical soils is helpful in agronomyand environment. Soil properties such as pH OC, clay, Al and Fe oxides, Ca and BD showed significant relationships with P sorption characteristics. However, there relationships differed among soil groups in the study site with soils formed over Shale and Lower Coal Measures having least relationship. The measured soil attributes had a good relationship with P sorption characteristics hence of great importance in modelling and monitoring agricultural and environmental activities as they affect soils.
ACKNOWLEDGMENT
The researcher is grateful to Professor F.O.R. Akamigbo and technical staff of the Department of Soil Science, University of Nigeria Nsukka for their assistance in this study.