SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 8 June 2007
Published in Soil Sci Soc Am J 71:1128-1136 (2007)
DOI: 10.2136/sssaj2006.0222
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Brand-Klibanski, S.
Right arrow Articles by Shenker, M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Brand-Klibanski, S.
Right arrow Articles by Shenker, M.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Brand-Klibanski, S.
Right arrow Articles by Shenker, M.
Related Collections
Right arrow Soil Organic Matter
Right arrow Volcanic Soils

SOIL CHEMISTRY

Overestimation of Phosphorus Adsorption Capacity in Reduced Soils: An Artifact of Typical Batch Adsorption Experiments

S. Brand-Klibanskia, M. I. Litaorb and M. Shenkera,*

a Dep. of Soil and Water Sciences, Faculty of Agricultural, Food and Environmental Quality Sciences, The Hebrew Univ. of Jerusalem, Rehovot 76100, Israel
b Dep. of Environmental Sciences, Tel-Hai Academic College, Upper Galilee 12210, Israel

* Corresponding author (Shenker{at}agri.huji.ac.il).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Although soil reduction often results in P release to soil solutions, many researchers have observed an increased maximum P adsorption (Smax) following soil reduction. We hypothesized that this result is an experimental artifact caused by exposure of the reduced soils to aerobic conditions and by the use of high P additions, which may result in precipitation. Four semiarid altered wetland soils were incubated under reduced conditions, followed by reoxidation, and their P-adsorption characteristics were measured under atmospheric and N2–atmosphere conditions. During the reductive incubation, soluble P and Fe concentrations increased. In one of the soils, P and Fe were monitored after reoxidation and both were found to decrease. The reduction–reoxidation cycle has led to increased Smax values. Under an N2 atmosphere, the equilibrium P concentrations at zero adsorption (EPC0) of all soils were higher than those determined under atmospheric conditions, whereas no significant changes were observed in Smax values. Oversaturation of the equilibrating solutions with respect to P minerals suggested P precipitation and overestimation of Smax at high added P concentrations under both aerobic and N2–atmosphere conditions. We conclude that aerobic batch experiments of reduced soils are affected by P adsorption to newly in-tube-formed ferric oxides. In accordance, we stress the importance of the EPC0 rather than the Smax as an informative measure of P adsorption, and the need for using low-P experiments and maintaining anaerobic conditions in evaluating P adsorption of reduced soils.

Abbreviations: CBD, citrate–bicarbonate–dithionite • CDP, cultivated degraded peat soil • EC, electrical conductivity • EPC0, equilibrium P concentration at zero adsorption • FDP, fallow degraded peat soil • Fed and Feox, Fe extracted by citrate–bicarbonate–dithionite and oxalate, respectively • NDNCP, nondegraded noncalcareous peat soil • OM, organic matter • Smax, maximum P adsorption


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Adsorption of P by Fe oxides and hydroxides (hereafter referred to as Fe oxides) and other highly reactive soil constituents may reduce P concentrations in the soil solution and minimize its movement through the soil matrix to the root surface and deeper layers, or from soils to waterways. Hence, a characterization of P adsorption in soils is of great importance for agronomists, as well as for the preservation of water quality, especially in areas susceptible to P leaching. Phosphorus adsorption in soils is influenced by physical and chemical factors, among which alternating reduction–oxidation conditions in soils and the associated transformations of Fe oxides have been documented as highly significant (Torrent, 1997; Rhue and Harris, 1999). Flooding a soil, which results in its reduction, followed by drainage and subsequent soil oxidation, generally results in a marked increase in P adsorption. This effect was attributed to decreased crystallinity of Fe oxides, especially after repeated cycles of drying and rewetting (Willett and Higgins, 1978; Holford and Patrick, 1981; Sah and Mikkelsen, 1986). There is uncertainty, however, concerning the effect of rewetting on P-adsorption properties of formerly drained soils. Phosphorus release from soils on rewetting and the subsequent development of reducing conditions are well-documented phenomena (Moore and Reddy, 1994; Reddy et al., 1998; Shenker et al., 2005). There are many other studies, however, that report an increase in soil P-sorption capacity following soil reduction (Khalid et al., 1977; Holford and Patrick, 1981; Villapando and Graetz, 2001). Various mechanisms have been suggested to explain the latter phenomenon, among them P adsorption to Al oxides (Villapando and Graetz, 2001), Ca carbonates (Moore and Reddy, 1994; Pant and Reddy, 2001), and clay minerals (Singh and Tabatabai, 1977), or P binding by Al and Fe organic complexes (Richardson and Vaithiyanathan, 1995).

Some studies have attributed the enhanced P retention under reduced conditions to P adsorption to newly precipitated amorphous ferrous oxides (Khalid et al., 1977; Willett and Higgins, 1978; Holford and Patrick, 1981; Jugsujinda et al., 1995). These studies suggested that amorphous ferrous oxides have a greater surface area and a larger number of adsorption sites than the original well-crystallized ferric oxides, and thus have a higher P-adsorption capacity. None of these studies, however, included any direct evidence confirming the higher P-adsorption capacity of ferrous oxides compared with ferric oxides. Moreover, many of those studies reported a concurrent increase in soluble P and Fe concentrations due to dissolution of Fe oxides and P release (Jugsujinda et al., 1995; Holford and Patrick, 1981; Khalid et al., 1977), a widespread phenomenon observed also by others (Nair et al., 1999; Reddy et al., 1998; Shenker et al., 2005). To elucidate this apparent contradiction, we hypothesized that the increase in P adsorption following soil reduction is an artifact of the conditions under which the adsorption experiments are conducted. Phosphorus may be adsorbing to ferric oxides that are formed during the adsorption experiment. Furthermore, P-adsorption measurements might be skewed by the high concentrations of P used in adsorption experiments, leading to P precipitation, which results in an erroneous overestimation of P adsorption. This estimation is based on unrealistically high P concentrations and it does not reflect the true maximum P adsorption (Smax) of the soils. Distinguishing between adsorption and precipitation is important because of their different kinetics. Whereas adsorption occurs within minutes to hours, precipitation is a slower process, especially under low P concentrations. The common P concentrations encountered in soil pore water are rather low (0.0003–0.3 mg L–1, Beek and van Riemsdijk, 1982) compared with those usually used in adsorption experiments, up to 100 mg L–1 (Graetz and Nair, 2000) or even 1200 mg L–1 (Zhou and Li, 2001). These high P concentrations may result in rapid chemical precipitation. Several studies have shown that P is also retained by surface precipitation on selected minerals such as goethite (Li and Stanforth, 2000; Ler and Stanforth, 2003), calcite (Freeman and Rowell, 1981), and gibbsite (van Riemsdijk and Lyklema, 1980), and that this precipitation may take place below saturation of any defined P mineral.

Veith and Sposito (1977) showed that precipitation may result in a statistically significant conformity to a Langmuir adsorption isotherm. Therefore, conformity of reaction data to adsorption isotherms such as Langmuir or Freundlich does not constitute proof of an adsorption reaction (Barrow, 1978). Because adsorption and precipitation are different mechanisms of P retention, the parameters in the adsorption isotherm cannot be interpreted in terms of adsorption reactions without additional independent evidence that clearly demonstrates that only adsorption is involved. The term P sorption capacity includes both adsorption and precipitation, but using this term may mask the difference between two distinct processes that may have different effects on the environment.

Our study focused on soils of an altered wetland in the Hula Valley, Israel (Fig. 1), that was drained in the 1950s, intensively cultivated for four decades, and partially rewetted and flooded in 1994 with the construction of a small (1000-ha) lake (Lake Agmon) in the midst of the altered wetland (Litaor et al., 2003). The creation of the new lake elevated the groundwater table in the center of the Hula Valley by at least 60 cm (Tsipris and Meron, 1998) and occasionally even to the surface, creating strongly reduced conditions (Shenker et al., 2005). Increased P concentrations were recorded in the water of the newly formed Lake Agmon during the first years after its construction (up to >400 µg L–1 total P and >100 µg L–1 dissolved P, Gophen et al., 1998). Increased soluble P concentrations in soil solutions upon rewetting of soils around the lake (up to 1500 µg L–1) was found to result from the reductive dissolution of ferric oxides, which is a major P-adsorbing phase (Shenker et al., 2005), and may explain the increased annual P loads from the valley to the Jordan River that were reported by Rom (1999). The implications of increased P loading from the Hula Valley is of major environmental concern because the valley (about 175 km2) is currently drained by a system of artificial canals, which empty into the Jordan River. The Jordan River carries as much as 75% of the annual freshwater input to Lake Kinneret, which, in turn, provides about 35% of the country's drinking water. The environmental consequences are especially serious since algal growth and productivity in Lake Kinneret are mainly controlled by P (Serruya and Berman, 1975).


Figure 1
View larger version (53K):
[in this window]
[in a new window]

  		Fig. 1. The study site in the Hula Valley, showing Lake Agmon, the drainage canals, and the sampling locations of cultivated degraded peat (CDP), fallow degraded peat (FDP), nondegraded, noncalcareous peat (NDNCP), and marl soils. Coordinates according to the Israel grid are given in kilometers.

 
Because the preferentially horizontal belowground flow through wide cracks in altered wetland soils is important (Litaor et al., 2006), it is crucial to separate the rapid process of P adsorption from the slower mechanisms of P precipitation in reduced subsoils. We suspect that, under the conventional experimental conditions under which P adsorption is evaluated, the distinction between these two processes might be questionable. Accordingly, the objective of this research was to evaluate the influence of the experimental conditions under which P adsorption is studied on the P adsorption parameters of reduced soils.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Study Soils
The peat soils of the Hula Valley are predominantly Histosols (approximately 1860 ha), which have been classified into Medifibrists, Medihemists, Medisaprists, and "Conflagrated Histosol"; each of these is subclassified according to its lime content: without lime, with minimal lime, and with lime (Israel Ministry of Agriculture, 1986). There are also numerous organo-mineral soils that have developed from marl-lacustrine deposits or a mixture of decomposed swampy plant material with alluvium. Some have redoximorphic features characteristics of long-term rewetting (Litaor et al., 2003). The soils for this study were sampled within 200 to 1700 m from Lake Agmon (Fig. 1) and represent the major soil types as well as current and previous land-use practices: (i) a nondegraded noncalcareous peat soil that has been permanently below the water table (NDNCP); (ii) a shallow, slightly calcareous, intensively cultivated, degraded peat soil overlying organo-mineral material developed along the transitional area between the former swamps and Hula Lake (CDP); (iii) a fallow degraded peat soil similar to the cultivated peat soil but which has not been cultivated for the last 7 yr (FDP), and (iv) a marl calcareous soil representing the depositional environs of the former Hula Lake (marl). The NDNCP peat was taken from a depth of 200 cm in a deep peat profile, well below the upper layer of the degraded peat, which is about 1 m thick. The other three soils were sampled from a depth of 5 to 25 cm below the surface. Before incubation, the soils were thoroughly hand mixed and large pieces of organic debris and other coarse materials were removed.

Soil Incubation in a Biogeochemical Microcosm
To study the effect of alternating redox conditions on P behavior in the rewetted area of the Hula Valley, the soils were packed in a controlled microcosm under reduced conditions followed by reoxidation. Two microcosms were constructed for each soil. The experiment began by flooding the soils until a low and steady redox potential (Eh ≤ –200 mV) was established for 8 to 16 wk. Then one microcosm for each soil was exposed to air and drained gradually until the establishment of a high and steady redox potential (Eh ~ 400 mV). After the incubations, the soils were freeze-dried to maintain the redox and mineralogical conditions for further analyses and for adsorption experiments. The soils at the end of each incubation step are referred to as reduced and reoxidized soils, respectively.

The microcosms (Fig. 2) were constructed from a 1.6-L sealed plastic vessel, 15 cm in diameter and 19 cm in height, covered with aluminum foil to prevent light and avoid algal and weed growth during the incubation. The soils were packed in the microcosms at a bulk density similar to that found in the field (1.01–1.09 kg dry weight m–3) and saturated gradually with deionized water from a porous cup. The water table was set at about 1 cm above the soil surface. Natural soil conditions were simulated by mixing the soils with ground wheat (Triticum aestivum L.) straw at a ratio of 1 g straw kg–1 dry soil, as suggested by Patrick and Henderson (1981) and applied by Shenker et al. (2005). The straw served as a labile microbial substrate and an electron donor.


Figure 2
View larger version (15K):
[in this window]
[in a new window]

  		Fig. 2. Microcosm setup: (a) darkened 19-cm high, 15-cm-diameter plastic vessel; (b) redox electrode; (c) pH electrode; (d) porous cup; (e) tube and tap, connected to syringe; (f) CO2 effluent tube; (g) electrical-resistance adaptors, connected to a data logger; (h) data logger; (i) sealing cover.

 
The pH was monitored throughout the incubation period with a pH electrode for semisolids (pH combined electrode, Model U-05998–20, Cole-Parmer, Vernon Hills, IL) and the redox potential by a wide-Pt-band redox electrode (ORP combined electrodes, Cole-Parmer). All electrodes were connected through electrical-resistance adaptors to a data logger (Multilog, Fourier Systems, Rosh HaAyin, Israel). The pH and redox potentials were recorded every 10 min. A similar setup of continuous pH and redox measurements in soil was previously shown to be reliable for long periods (Eshel and Banin, 2002; Shenker et al., 2005). At the end of the incubation period, electrodes were rinsed and checked in buffer solutions. The recorded values for each electrode were corrected for differences from the initial calibration as described by Shenker et al. (2005).

Soil solutions were sampled periodically during the incubation period using water-sampling tubes, each connected inside the column to a porous cup and at the outer end to a 10-mL sampling syringe. We used ultra-high-molecular-weight polyethylene porous cups with pore sizes of 7 to 40 µm (Porex Corp., Fairborn, GA), which had been tested and found not to adsorb P (Shenker et al., 2005). During the reductive incubation, soil-solution samples (~45 mL) were collected periodically followed by injection of the same volume of deionized water through the porous cup sampler to prevent clogging of the cup. The dead volume of the sampler cup and tube was discarded on sampling. The CO2 emitted from the soils during the incubation was vented out through a water column to minimize O2 inflow (Fig. 2). Reoxidation was established by opening the effluent tube [Fig. 2(f)] and by sampling the soil solution without returning the sampled volume. For the FDP, soil solution samples were removed from the reoxidized microcosms by applying low vacuum (~10–30 cm H2O suction), thus allowing the collection of more samples than from the other microcosms.

Laboratory Analysis
Microcosm water samples were filtered through 0.45-µm pore filters immediately after sampling, followed by a measurement of electrical conductivity (EC) and preservation of the sample by acidification (pH ~2) for further analyses. Filtered but not acidified subsamples were analyzed within 1 min after sampling at 562 nm for Fe(II) concentration using the ferrozine complex method (Stookey, 1970) and for dissolved inorganic P using the ammonium molybdate colorimetric method (Murphy and Riley, 1962). Total dissolved Fe, Ca, Mn, and Al were determined by inductively coupled plasma–atomic emission spectroscopy (ICP–AES; Spectro, Kleve, Germany) in the filtered and acidified samples.

Unless otherwise stated, methods of analysis were as given by Sparks et al. (1996): soil pH was measured in 1:10 (w/v) soil/water extracts with a glass electrode; organic matter (OM) content was determined by weight loss under dry combustion at 400°C; CaCO3 content by volumetric CO2 emission in 1.3 M HCl; total P content was determined after wet combustion with HClO4 and HNO3; free oxide contents of Fe, Mn, and Al were determined by citrate–bicarbonate–dithionite (CBD) extraction while the amorphous and poorly crystallized hydroxide contents of Fe, Mn, and Al were determined by oxalate extraction. The CBD and oxalate extracts were analyzed for P as a measure of the P associated with free and amorphous oxides, respectively. Gypsum content was estimated by the water-dissolution technique (Loeppert and Suarez, 1996) following a modification described in Shenker et al. (2005). All soil tests were performed in triplicate.

Adsorption Experiments
The P-adsorption characteristics of the studied soils were evaluated with the Langmuir and Tempkin isotherms using batch experiments as described by Graetz and Nair (2000) with slight modifications. The adsorption experiments were conducted with 1-g freeze-dried soil samples in duplicate at 25°C, using a soil/solution ratio of 1:25 (w/v). Initial P concentrations (C0) in the added solutions ranged from 0 to 40 mg L–1 (0–30 mg L–1 in the NDNCP peat because of its lower adsorption capacity) and included the following concentrations: 0, 0.5, 1, 2, 4, 6, 8, 10, 15, 20, 25, 30, and 40 mg L–1, all prepared from analytical-grade KH2PO4 dissolved in 0.03 M KCl. The suspensions were shaken for 24 h in 50–mL polyethylene tubes on a reciprocal shaker, followed by centrifugation at 1500 x g for 20 min. The filtered supernatant (0.45 µm) was analyzed for P using the Murphy and Riley (1962) method. Experiments similar to these aerobic adsorption experiments were conducted also in a glove box under N2 atmosphere to study P adsorption by the reduced soils. For this set of experiments, the partial pressure of O2 ranged between 0.02 and 0.2 kPa (~0.1–1% of atmospheric partial pressure of O2) and O2 was removed from all solutions by N2 bubbling for 2 h before each experiment. These experiments are referred to as anaerobic adsorption experiments.

The Smax (mg kg–1) and the binding energy constant (k, L mg–1) were estimated by fitting the experimental values of S (P adsorbed after 24-h equilibration, mg kg–1) and Ct (P concentration in the solution after 24-h equilibration, mg L–1) to the Langmuir model (Eq. [1]) after correcting for the initial adsorbed P (S0, mg kg–1). The value of S0 was evaluated by linear extrapolation, using a least-squares fitting procedure for the experimental data of C0 < 4 mg L–1, according to Reddy et al. (1998).

Formula 1[1]
where total adsorbed P (St) is the sum of S and S0.

The equilibrium P concentration at zero adsorption (EPC0) dictates whether P would be adsorbed or desorbed as solution moves through the soil (Reddy et al., 1998; Nair et al., 1999); its value was estimated by fitting the experimental values of S and Ct from the low C0 range (C0 < 10 mg L–1) to the Tempkin model (Eq. [2]), which explicitly estimates the EPC0 (Litaor et al., 2005):

Formula 2[2]

The parameters Smax, k, EPC0, and their standard errors were estimated by nonlinear, least-squares fitting of the data points to the above models by using the SigmaPlot curve-fitting routine (SPSS, 2002).

Geochemical Modeling of the Equilibrating Adsorption Solution
The pH, EC, and total dissolved elements were measured as described above in the equilibrium solutions of the adsorption experiment with C0 values of 0, 0.5, 10, 20, and 40 mg P L–1. Geochemical modeling was conducted using Visual MINTEQ, Version 2.30 (Gustafsson, 2004) to evaluate the likelihood of P precipitation during the adsorption experiment. The modeling for the oxidized soils was based on the equilibrium solutions from the aerobic experiments, while for the reduced soils the equilibrium solutions from the N2–atmosphere experiments were considered. The ionic strength was estimated from the EC according to Griffin and Jurinak (1973) and the saturation indices with respect to various P minerals were calculated according to the thermodynamic data of Lindsay (1979) and the redox state of the soils [i.e., Fe and Mn were considered Fe(II) and Mn(II) in the reduced soils and Fe(III) and Mn(IV) in the oxidized soils].


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
General Soil Characteristics
We used four soils, differing largely in their characteristics according to their parent material and land-use history (Table 1), to test our hypothesis. The lowest OM content (5%) was found in the marl, the highest (66%) in the NDNCP, and intermediate OM contents were found in the degraded peat soils, CDP and FDP. The CaCO3 content ranged from zero in the NDNCP to 498 g kg–1 in the marl. The soil pH ranged from slightly acidic in the NDNCP to slightly basic in the marl. The CDP and FDP soils had high gypsum contents, while the NDNCP and marl contained only traces of gypsum. The degraded peat soils contained the largest amount of free and amorphous Fe oxides. In all soils, the content of free (0.05–1 g kg–1) and amorphous (0.03–0.25 g kg–1) Mn oxides was small relative to that of the Fe oxides. The contents of free (0.51–4.24 g kg–1) and amorphous (0.65–6.33 g kg–1) Al oxides were also rather low. Although Mn and Al oxides can also be an adsorbing phase for P (Krairapanond et al., 1993; Jugsujinda et al., 1995), their lower content compared with Fe oxides makes them a secondary candidate for P adsorption. Total P in the soils ranged from 0.98 to 1.22 g kg–1 in the surface soils (CDP, FDP, and marl), while a smaller value was found in the NDNCP. The amount of P extracted by the CBD procedure (Pd) ranged from 23 to 88% of total P, and that extracted by the oxalate procedure (Pox) ranged from 31 to 64% of total P, suggesting that adsorbed P is a major part of the P reserve in these soils. In the marl, the Pox was higher than the Pd, even though CBD extracted more Fe than the oxalate. The higher Pox in this calcareous soil may result from P transformation from Ca-P to adsorbed P during the CaCO3 removal step of the oxalate extraction procedure (Sparks et al., 1996), or it may result from coprecipitation of the extracted P with Ca during the CBD extraction. Such precipitation is not expected in the oxalate extraction because Ca2+ is effectively removed by the oxalate. Thus, we assume that the adsorbed P in this soil might be overestimated by the oxalate extraction, or it might be underestimated by the CBD extraction. This effect was found also in some of the extractions of the degraded peat soils (CDP and FDP, Table 1) that are characterized by lower CaCO3 content, but not in the NDNCP, which had the lowest CaCO3 content.


View this table:
[in this window]
[in a new window]

  		Table 1. Selected soil properties before incubation and following flooding and drainage of the microcosms.

 
Soil Incubation under Reduced and Reoxidized Conditions
The total concentrations of Fe and P in the solutions from the four soils are depicted in Fig. 3, along with the –log of the e– activity in molar units (pe) + pH profiles. Within 1 to 3 d from the start of the experiment, a steep decrease in pe + pH values was observed (0.4–3.7 units, from initial values of 13.1–15.9). As incubations proceeded, the pe + pH values decreased further, to about 3, indicating highly reduced conditions, and Fe and P concentrations increased. Draining and initiation of the reoxidation process (marked in Fig. 3 by the arrows on the x axes) led to a significant increase in the pe + pH values (9.7 to 19.0). In the FDP soil, P and Fe concentrations in the soil solutions were monitored after reoxidation and both were found to decrease. The four soils differed in the rate of redox-potential change, the rates of P and Fe release into soil solution, and the extent to which their concentrations changed during the incubations. The differences among the soils are explained by the amount of extractable Fe and P (Table 1). During flooding of the FDP, large amounts of Fe dissolved into the soil solution, resulting in 18.9 mg L–1 Fe in the solution, in accordance with the high Feox and Fed contents in this soil. The Fe-reductive dissolution was followed by considerable (0.93 mg L–1) release of desorbed P, as shown by Shenker et al. (2005) in similar soils. Despite the high content of Fe oxides in the CDP and the higher P content of this soil (Table 1), the lower dissolution of Fe during the reductive incubation (1.23 mg L–1) led to lower release of P (0.21 mg L–1). A similar pattern was observed in the NDNCP, where low Fe-oxide content resulted in low release of P and Fe (0.65 and 0.48 mg L–1, respectively). The flooding of the marl resulted in high concentrations of P (0.71 mg L–1), in accordance with the high P content of this soil.


Figure 3
View larger version (29K):
[in this window]
[in a new window]

  		Fig. 3. Timeline of –log of the e– activity in molar units (pe) + pH (solid line) and Fe and P concentrations during soil incubation under reduced and reoxidized conditions in the cultivated degraded peat (CDP), fallow degraded peat (FDP), nondegraded, noncalcareous peat (NDNCP), and marl soils. Time of draining and initiation of the reoxidative incubation is marked by an arrow on each x axis. To fit into the graph, the Fe concentration values were reduced by a factor of 2 for the CDP, FDP, and NDNCP soils, and by a factor of 20 for the marl soil, as shown on the y axes.

 
The similar trends observed in the different soils support our hypothesis that the observed processes characterize common phenomena occurring in the tested soils despite their different characteristics. The concurrent increase of P and Fe concentrations during the reductive incubation suggest that reductive dissolution of ferric oxides is the main mechanism of P release from the oxidized soils. Similarly, the concurrent decrease of P and Fe concentration during the reoxidative period suggests that the main P-retention mechanism is precipitation of amorphous ferric oxides followed by P adsorption on the newly formed oxides or Fe–P coprecipitation.

Reduction and reoxidation during the incubation had only a minor effect on soil pH and contents of OM, CaCO3, and gypsum (Table 1). Despite the pronounced reductive Fe dissolution during the reductive incubation (Fig. 3), Feox and Fed concentrations did not indicate any consistent change.

Phosphorus Adsorption
The Langmuir and Tempkin models fitted well the P adsorption by the four soils (Table 2), which differed considerably in their adsorption characteristics. The Fe-oxide-rich CDP and FDP soils had higher adsorption capacity, as indicated by their Smax and k values, than the Fe-oxide-poor NDNCP and marl soils. Low EPC0 values were associated with the high-P-adsorbing CDP and FDP soils. The higher EPC0 values of the NDNCP and marl soils, coupled with their lower Smax may result in higher P mobility.


View this table:
[in this window]
[in a new window]

  		Table 2. Phosphorus adsorption parameters determined under aerobic and anaerobic conditions.

 
Despite the different P-adsorption characteristics, all soils exhibited similar trends following their reduction and subsequent reoxidation. In all cases, the Smax values increased by the reduction–oxidation cycle, suggesting that repeated cycles of flooding and draining, which are common in drained wetlands, may enhance P adsorption. These trends are well explained by the Fe reduction–oxidation processes that lead to Fe-oxide dissolution on reduction and precipitation on oxidation (in accordance with Fig. 3). Other researchers (e.g., Sah and Mikkelsen, 1986) have attributed this phenomenon to the transformation of crystalline Fe oxides to amorphous Fe oxides with a higher surface area, capable of larger P adsorption. Holford and Patrick (1981) found this transformation to be enhanced when the pH and Eh of the soils are low enough during soil flooding to maintain high Fe concentrations. In contrast, no clear trend was found in the EPC0 values of the soils. While in the NDNCP and marl, EPC0 values decreased following soil reoxidation, in the degraded peat soils (FDP and CDP), which had rather low EPC0 values, this trend was not found, probably because of the difficulty in measuring the extremely low P concentrations at the end of the equilibrium period (Ct) in the lower P additions.

In the three surface soils (CDP, FDP, and marl), Smax values obtained under aerobic conditions apparently increased following soil reduction compared with the preincubated soils (Table 2). In the formerly reduced NDNCP soil, this trend was not significant. The other adsorption parameters were less conclusive. The apparent increase in the soils' P-adsorption capacity on reduction is similar to the observations of many researchers (e.g., Willett and Higgins, 1978; Holford and Patrick, 1981; Villapando and Graetz, 2001). Nevertheless, it contradicts the finding of increased P concentrations in all tested soils during the reduction period (Fig. 3). Hence, we hypothesized that the apparent increase in the soils' P-adsorption capacity on reduction reflects an artifact resulting from the rapid oxidation of ferrous Fe and the precipitation of ferric oxides on exposure of the reduced soils to air. These in-tube-created oxides form new adsorption sites, resulting in an overestimation of the P-adsorption capacity of the reduced soils.

To verify this potential error, the adsorption experiments with the reduced soils (including the preincubated NDNCP, which was reduced in its initial state) were also conducted under a N2 atmosphere. The results are summarized in Table 2. In accordance with our hypothesis, the EPC0 values of all the reduced soils were significantly (F = 4, P < 0.05) higher under anaerobic conditions (Table 2). For the NDNCP in its initial state, the increase was not significant, probably indicating that some oxidation took place in the time between sampling in the field and transfer to the lab. All of the EPC0 values determined under anaerobic conditions were lower than the soluble P concentrations measured in the microcosms during the reductive incubation. This difference reflects the higher partial pressure of O2 in the N2 chamber (20.3–203 Pa) than in the reduced microcosms. Indeed, the theoretical partial pressure of O2 at a pe + pH value of 4, similar to the values measured in the reduced microcosms, corresponds to 6.12 MPa (Lindsay, 1979). This extreme value was not established in the N2 chamber. Hence, traces of O2 could have oxidized the ferrous Fe and led to the precipitation of amorphous ferric oxides followed by an increase in P adsorption. Only in the marl soil, increased EPC0 of the reduced soil, compared with the preincubated soil, was not observed when determined under N2 atmosphere. This might be explained by the highest Fe concentration in its soil solution during the reductive incubation (Fig. 3). We assume that because of its higher ferrous Fe concentration, this soil was more susceptible to oxidation during the adsorption experiment, even under the low O2 partial pressure that prevailed in the N2 chamber. Unlike the trends in the EPC0 values, no significant differences were observed between the Smax values determined under anaerobic vs. aerobic conditions; the k values exhibited contrasting trends, which in most cases were insignificant (Table 2).

Higher EPC0 values in soils subject to reductive incubation vs. oxidative incubation have been reported for sediments (Pant and Reddy, 2001), wetland soils (Reddy et al., 1998), and spodic-horizon soils (Nair et al., 1999). Generally, the reductive incubations also exhibited smaller Smax and k values. In a few cases, however, especially in soils with high P loading, Smax values have been found to increase under reductive conditions (Reddy et al., 1998; Nair et al., 1999; Villapando and Graetz, 2001). In line with these observations, it has been found that at high concentrations of added P (100 mg L–1), more P is sorbed under reduced conditions than under oxidized conditions (Patrick and Khalid, 1974; Khalid et al., 1977). Similar findings were reported for two anaerobic peat soils from the Hula Valley: under anaerobic conditions, P sorption was higher at the high end of the P addition (C0 = 70 mg L–1) and lower at the low end of the P addition (C0 < 8 mg L–1) compared with P sorption under aerobic conditions (Litaor et al., 2005). Further, it was found in that study that much of the sorbed P above C0 of 8 mg L–1 had resided in the Ca-P phase, as verified by P fractionation in the solid phase following the sorption experiment.

Geochemical Modeling
On the basis of the different tendencies exhibited by the Smax and EPC0 values, we hypothesized that the apparent higher P adsorption in the higher range resulted from precipitation. We tested this possibility by calculating the saturation index of selected P minerals in the equilibrium solutions of the adsorption experiments (Fig. 4). In all soils, redox states, and levels of added P, except at the lowest P concentration in the incubated marl soils, the solutions were oversaturated with respect to hydroxyapatite [Ca5(PO4)3OH]. In all the reduced soils, including the preincubated NDNCP, at all levels of added P, the solutions were oversaturated also with respect to MnHPO4. Overall, the geochemical analyses indicate the potential importance of Ca-P solid phases in governing P concentrations in the equilibrating adsorption solutions under both reduced and reoxidized conditions, while under reduced conditions MnHPO4 is of importance as well. The source of the Ca might be calcite in the marl soil and gypsum in the peat soils. Despite the low content of Mn oxides in the studied soils, the Mn released due to the reductive dissolution of Mn oxides may have precipitated with P as MnHPO4 during the adsorption experiments of the reduced soils.


Figure 4
View larger version (37K):
[in this window]
[in a new window]

  		Fig. 4. Saturation indices (SI) of P minerals in the equilibrated adsorption solutions of preincubated, reduced, and reoxidized(A) cultivated degraded peat (CDP), (B) fallow degraded peat (FDP), (C) nondegraded noncalcareous peat (NDNCP), and (D) marl soils. The SI values for the P minerals were calculated according to the thermodynamic stability constants of Lindsay (1979): brushite, CaHPO4·2H2O; octacalcium phosphate, Ca4H(PO4)3·2.5H2O; beta-calcium phosphate, ß-Ca3(PO4)2; hydroxyapatite, Ca5(PO4)3OH; MnHPO4; strengite, FePO4·2H2O; and vivianite, Fe3(PO4)2·8H2O. Data for the oxidized soils were taken from adsorption experiments conducted under aerobic conditions and data for the reduced soils were taken from experiments conducted under N2 atmosphere.

 
Oversaturation of the equilibrated solutions with respect to P minerals indicates possible precipitation of P from the solutions during the adsorption experiment. Similar findings have been reported by Hundal (1988), Olila and Reddy (1997), and House and Denison (2000). Although geochemical modeling cannot prove actual precipitation, and despite the short time of the experiment on a pedological time scale, native crystalline seeds of P minerals may induce precipitation during P-adsorption experiments, as has been suggested by Hundal (1988) and Sposito (1984). Phosphorus precipitation subsequent to or during adsorption reactions has also been demonstrated by direct mineralogical analysis of pure systems (van Riemsdijk and Lyklema, 1980; Freeman and Rowell, 1981). Potential P precipitation during adsorption experiments was also evident from an analysis of P fractions following adsorption experiments with Histosols from the Hula Valley (Litaor et al., 2005). In that study, a consistent and substantial increase in the Ca-P fraction of the soil P was observed in the tested soils after they had been subjected to adsorption experiments. Thus, Ca-P precipitation simultaneous with P adsorption was concluded. Overall, the cited findings along with our results suggest that, under the experimental conditions, the process of adsorption at the higher P concentrations cannot be distinguished from the process of precipitation, and accordingly, Smax might be overestimated by the use of high P additions. Nanzyo et al. (2004) demonstrated that P adsorption by five soils treated with dithionite increased when the samples were exposed to air and further increased with the addition of ferrous Fe. They found precipitation in the equilibrium solutions and identified the precipitate as amorphous Fe phosphate. Based on those results, Nanzyo et al. (2004) inferred that the increase in P adsorption by reduced soils is due to Fe oxidation with O2 from the air and its precipitation with P, thus leading to an artificial increase in P adsorption. Enhanced P adsorption following soil reduction, as reported by Holford and Patrick (1981), Khalid et al. (1977), and Patrick and Khalid (1974), was determined from experiments with P additions of up to 100 mg L–1. We speculate that by using such high P concentrations, P precipitation may lead to overestimations of the adsorbed P and to erroneous conclusions of increased P adsorption following soil reduction. The geochemical analyses (Fig. 4) suggest that MnHPO4 precipitation may become a dominant P-retention mechanism in the reduced soils. In accordance, an apparent increase of Smax was found in all the reduced soils compared with the preincubated soils regardless to the aerobic or anaerobic conditions used in the adsorption experiments. Although in all soils, the solutions were also oversaturated with respect to hydroxyapatite, the slow precipitation rate of this mineral, especially in the presence of organic substances (Inskeep and Silvertooth, 1998), excludes its precipitation as an important P-retention mechanism during the short equilibration time (24 h).

The saturation indices (Fig. 4) suggest that, at low P concentrations, the possibility of P retention by precipitation is minimal, thus this range better represents P adsorption and better estimates the EPC0. Figure 4 clearly shows, however, that even at zero added P, some oversaturation occurs; therefore the chosen range (0–10 mg L–1) of the initial added P concentrations in our adsorption experiment is a compromise used to minimize the possibility of P retention by precipitation. This range is also a better reflection of the P concentrations encountered in the field.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The results of this study clearly demonstrate the considerable effect of alternating redox conditions on soil P-adsorption properties. Drainage of previously flooded soil and the development of oxidative conditions are known to result in a simultaneous decrease in P and Fe concentrations in the soil solutions. These trends, coupled with the decrease in EPC0 values and the increase in Smax following the drainage, suggest that reduction–oxidation cycles caused by repeated cycles of flooding and drying enhance soil P adsorption, apparently caused by decreased ferric oxide crystallinity. The decrease in the redox potential during soil flooding coupled with the simultaneous increase in the concentrations of P and Fe in the soil solutions indicate that soil reduction leads to P release due to the reductive dissolution of ferric oxides. Based on our results, we conclude that the increased P adsorption by reduced soils that has been reported in many studies may reflect an artifact due to oxidation during the adsorption experiments. Increased P adsorption in reduced subsoil layers might be expected to protect percolated water from being loaded with P, but according to our findings, this is not the case. Phosphorus retention may take place in the reduced subsoil layers, but it occurs mainly via the slower and less effective process of precipitation, depending on the available counterelements (e.g., Ca, Mn, Fe) in the soil solution. The difference between these two retention processes might be of importance in cases of preferential flow, which allows short-time residence of the solution in the soil.

We postulate that EPC0, rather than Smax, is a meaningful parameter that accurately describes P adsorption in the environment. While Smax is actually never encountered in soils and sediments, the EPC0 value is more closely related to the actual P concentrations in pore water and is believed to well represent equilibrium with the solid phase. To estimate the EPC0 of a soil, low P concentrations should be used in a batch experimental setup, and for reduced soils, the experimental conditions should exclude O2.


   ACKNOWLEDGMENTS
 
This research was supported in part by a grant (GLOWA–Jordan River) from the Israeli Ministry of Science and Technology and the German Bundesministerium fuer Bildung und Forschung (BMBF).


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication June 11, 2006.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
C. C. Hoffmann, C. Kjaergaard, J. Uusi-Kamppa, H. C. B. Hansen, and B. Kronvang
Phosphorus Retention in Riparian Buffers: Review of Their Efficiency
J. Environ. Qual., August 24, 2009; 38(5): 1942 - 1955.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Brand-Klibanski, S.
Right arrow Articles by Shenker, M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Brand-Klibanski, S.
Right arrow Articles by Shenker, M.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Brand-Klibanski, S.
Right arrow Articles by Shenker, M.
Related Collections
Right arrow Soil Organic Matter
Right arrow Volcanic Soils

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome