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


     

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 Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (14)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Litaor, M. I.
Right arrow Articles by Shenker, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Litaor, M. I.
Right arrow Articles by Shenker, M.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Litaor, M. I.
Right arrow Articles by Shenker, M.
Related Collections
Right arrow Spatial Distribution
Right arrow Wetlands and Aquatic Processes
Right arrow Preferential Flow
Right arrow Phosphorus
Journal of Environmental Quality 32:335-343 (2003)
© 2003 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America

TECHNICAL REPORTS
Waste Management

Spatial Analysis of Phosphorus Sorption Capacity in a Semiarid Altered Wetland

M. I. Litaor*,a, O. Reichmannb, M. Belzera, K. Auerswaldc, A. Nishrid and M. Shenkerb

a Dep. of Biotechnology and Environmental Sciences, Tel-Hai Academic College, Upper Galilee 12210, Israel
b Dep. of Soil and Water Sciences, Faculty of Agricultural, Food and Environmental Quality Sciences, The Hebrew University of Jerusalem, Rehovot 76100, Israel
c Dep. of Grassland Science, Technical Univ. Muenchen, Am Hochanger 1, D-85350 Freising-Weihenstephan, Germany
d Kinneret Limnological Laboratory, POB 345 Tiberias, Israel

* Corresponding author (litaori{at}telhai.ac.il)

Received for publication December 20, 2001.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The observed increase in phosphorus (P) loading into the Jordan River could increase eutrophication processes in Lake Kinneret, the only freshwater lake in Israel, which provides 25% of the country's drinking water. The P may originate from the peat soils of the highly altered Hula Valley's semiarid wetland ecosystem through which the Jordan River runs. The objectives of this research were to ascertain the sorption capacity of these soils and to identify areas with high potential for P release from soils to ground water. We extracted 80 soil samples collected across the valley with ammonium oxalate and determined the ratio of extractable P to Fe and Al, from which we derived the degree of phosphorus saturation (DPS). A relatively low DPS (<15%) was observed in Histosols compared with the high DPS (>30%) observed in many of the hydromorphic organo–mineral soils. We used a sequential Gaussian simulation technique to assess the spatial pattern of the DPS and found that the Histosols have a low probability (<10%) of exceeding the widely used environmental DPS threshold of 25%. The areas characterized by mineral soils, such as hydromorphic Vertisols and various marl redoximorphic soils, have a high probability (>60%) of exceeding the threshold value. The ability to predict the concentrations of dissolved P in ground water based on DPS values was somewhat impaired because of the preferential flow characteristics in this altered wetland.

Abbreviations: ccdf, conditional cumulative distribution function • DPS, degree of P saturation


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
FRESHWATER WETLANDS are nutrient sinks that efficiently process and store N and P, thus reducing the potential of eutrophication in downstream lakes (Richardson, 1985). The draining of Lake Hula and the elimination of its surrounding wetlands during the mid 1950s, to increase the arable land in northern Israel, removed this crucial nutrient sink. Consequently, an increased volume of suspended material and nitrate loading were observed from the Hula basin into the Jordan River, which runs through the basin, and Lake Kinneret, the only freshwater lake in Israel, which provides as much as 25% of the whole country's drinking water. The biogeochemical cycles of C and N in the peat soils of the Hula Valley and the development of agrotechnical methods to reduce nitrate seepage from these soils have received much attention in past years (Avnimelch et al., 1978; Brenner et al., 1978). However, only scant attention has been paid to P leaching mechanisms. Phosphorus loading into the Jordan River has nearly doubled in the last 30 years (Rom, 1999) and P has been shown to be the principal nutrient regulating algal growth in downstream Lake Kinneret (Serruya et al., 1969; Serruya and Berman, 1976).

The behavior of P in the peat soils of the Hula Valley is complex, partly because most of the 54 soil series identified in the valley have not yet reached a steady state (Israel Ministry of Agriculture, 1986). In the past 40 yr, since the drainage of the Hula Valley, the soil pH has increased steadily from 5 to 7, while the organic matter content has decreased from 50 to 70% before the drainage to 25 to 30% in 2000 (Litaor et al., 2003a). The steady change in pH and organic matter content has undoubtedly affected the physico–chemical behavior of P and its availability to plants. Indeed, many of the crops in the Hula Valley exhibit strong P deficiency (Gephen et al., 1985), resulting in the common practice of high P application rates by farmers. The soil surveyors of the Israel Ministry of Agriculture (1986) hypothesized that P released during the oxidation of organic matter is coprecipitated by Fe, Mn, and Al, which are also released into the soil solution during the oxidation of organic matter, or adsorbed by the newly formed oxides and hydroxides of these metals. These sesquioxides are insoluble under aerobic conditions, and the level of available P has therefore been considerably depressed. The recent (1994) reflooding of the least fertile area of the former wetland created a small eutrophic lake, locally known as Lake Agmon, and elevated ground water in a vast area in the central section of the Hula Valley (Tsipris and Meron, 1998). The elevated ground water increased the reductive dissolution of iron oxides and hydroxides, followed by P release to the ground water (Litaor et al., 2003b). This P release may have partially contributed to the increase in P loading into the Jordan River in recent years observed by Rom (1999).

Sequential extraction experiments showed that P in the soils of the study area is associated mainly with iron oxides and Ca-P solid phases (Litaor et al., 2003b). These results suggest that P solubility is a product of sorption–desorption processes with iron oxides and precipitation–dissolution of Ca-P phases. Since the average pH of the soil system has steadily increased from 5 to 7 over the years, P concentration in soil solutions should be decreasing if chemical equilibria of Ca-P phases control its solubility (Lindsay, 1979). However, the observed simultaneous increase in ferrous and P concentrations in ground water following the decrease in redox potential (<-200 mV) in the reflooded area suggests that the increase in P concentration is related to desorption of P from iron oxyhydroxides rather than dissolution of Ca-P solids.

The construction of isotherm plots is a widely used technique to evaluate the sorption maximum and point of equilibrium P concentration (EPC0), at which P sorption equals P desorption under aerobic or anaerobic conditions (Wolf et al., 1985). Based on the observation that as the amount of P adsorbed by a soil increases, the EPC0 of the soil, and consequently the risk of P losses to shallow ground water and surface water also increase, many authors have addressed the ratio of actual adsorbed P to the P sorption maximum as a crucial soil characteristic that regulates P losses and retention (Breeuwsma and Silva, 1992; Lookman et al., 1995; Young and Ross, 2001; Zhou and Li, 2001). The generation of full P adsorption isotherm plots to determine the terms for calculating this soil characteristic is labor-intensive and time-consuming. It can therefore only be used to characterize representative soils, which is insufficient to provide the required information regarding P availability and potential mobility from the soils to ground water on a regional scale. This information is vital for decision- and policy-making by agricultural and natural resources management. To evaluate the spatial heterogeneity of P adsorption processes across a landscape, a quicker, more suitable technique is required; hence many simplified methods have been suggested to estimate both the actual adsorbed P and the P sorption maximum of soils. For example, Breeuwsma and Silva (1992) developed a test referred to as the degree of phosphorus saturation (DPS), which relates the soil P sorption maximum to an extractable soil P concentration with the following relationships:

[1]

The extractable soil P is estimated with 0.2 M ammonium oxalate buffered at pH 3.0, whereas the P sorption maximum is computed from an adsorption isotherm. To avoid the need for the laborious assessment of the P sorption maximum parameter from the adsorption isotherm, Breeuwsma and Silva (1992) estimated the P sorption maximum with the oxalate-extractable Al (Alox) and Fe (Feox) values and expressed the DPS as:

[2]

Other estimations for actual adsorbed P in various soils include: 0.1 M NaOH–extractable P (Sallade and Sims, 1997); 1.25 M NH4–acetate + 0.03 M NH4F–extractable P (Young and Ross, 2001); and Olsen P (Zhou and Li, 2001). The P sorption maximum has also been estimated by other methods. Most of these are based on a single-point sorption determination under various conditions of initial P concentration, ratios of soil to solution, and time of equilibration (Sallade and Sims, 1997; Young and Ross, 2001; Zhou and Li, 2001).

The concept of using DPS based on oxalate extraction of P, Al, and Fe was used by Lookman et al. (1995), who showed that the relative size of the quickly desorbing P pool (i.e., extractable P) increases as the DPS increases. They also suggested that nearly all Pox is potentially desorbable and that the DPS is directly relevant to the estimation of leaching losses from P-laden agricultural soils. More recently, the DPS technique has been used to estimate the potential for P release from soils to surface runoff and subsurface discharge (Pote et al., 1996; Sharpley et al., 1996; Leinweber et al., 1997; Nair et al., 1998).

The objective of this study was to quantify the spatial P sorption behavior using the DPS approach on a regional scale. Since P release from the soils of the Hula Valley has been found to be closely related to the presence and chemistry of iron oxyhydroxides (Litaor et al., 2003b), the oxalate-based DPS method was followed in this study. A quantification of the spatial structure would facilitate a generalization of the sorption capacity from a few isotherms taken at several selected sites to the required scale. The quantification of the thematic and probability maps with stochastic simulation techniques should provide an important tool for better P management in altered wetlands.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Study Site
The Hula Valley is the northernmost segment of the Jordan–Arava Rift Valley, approximately 70 m above sea level. It is approximately 175 km2 in area and it is currently drained by a system of artificial canals, which empty into the Jordan River at the southern end of the valley (Fig. 1) . Following the draining of the swamp, a surface layer of peat with an average thickness of 4 to 6 m was exposed to the atmosphere. Subsequently, organic matter oxidation began, inducing internal conflagration followed by major land subsidence (2–4 m) and partial elimination of the original drainage pattern, destruction of soil structure, increased topsoil erosion by wind, and a continuous decline in soil quality and fertility. To reverse some of the negative consequences of the drainage of Lake Hula and the surrounding wetlands, a small 100-ha lake was engineered in 1994, covering the least agriculturally productive peat soils in the Hula Valley (Fig. 1).



View larger version (56K):
[in this window]
[in a new window]
  		Fig. 1. Sampling locations in the Hula Valley, Israel.

 
The soils of the Hula Valley are predominantly Histosols (approximately 1860 ha) that were classified into four major great groups: (i) Medifibrists, (ii) Medihemists, (iii) Medisaprists, and (iv) Conflagrated Histosols. The group "Conflagrated Histosols" was created to recognize the importance of fires in the accelerated oxidation of these drained soils. There are also numerous organo–mineral soils that have developed from marl-lacustrine deposits and some have redoximorphic features characteristic of long-term flooding (Litaor et al., 2003a).

Field Sampling
We collected 80 soil samples that best represent the important soil great groups observed in the Hula Valley during the winter of 2000. Most of the soil samples were located near observation wells with known coordinates (Fig. 1). Samples located away from the observation wells were georeferenced with a Geographical Positioning System (eTrex GPS; Garmin International, Taiwan) (accurate up to 50 m). The soil samples were collected with an auger from a depth of below 50 cm, to avoid the plowed layer, and beneath the rhizosphere, where biological activity may enhance P solubility. All samples were air-dried and ground to pass through a 2.0-mm sieve and subsequently ground to pass through a 0.5-mm sieve.

Laboratory Analyses
The soil pH was determined in 1:10 soil and water suspensions with a glass electrode and the organic matter content was determined by the dry combustion method (Nelson and Sommers, 1982). The extractable soil P was determined with a 0.5-g soil sample using extraction solution of 30 mL, 0.5 M NaHCO3 (pH 8.5). The extractable P was determined colorimetrically with a Spectronic (Rochester, NY) Model 20 Genesis spectrophotometer following Murphy and Riley's (1962) method. Sesquioxides (Fe, Al) were extracted with the ammonium oxalate procedure described by Ross and Wang (1993). In short, we put 1 g of dry soil in extraction bottles with a 50-mL solution of ammonium oxalate (0.2 M) and oxalic acid (0.2 M) at pH 3. The bottles were wrapped with aluminum foil and placed on the shaker for 2 h, then centrifuged for 20 min at 1800 rpm. Each sample was filtered through a 0.45-µm-pore-diameter cellulose acetate membrane, and 0.5 mL concentrated HNO3 and a few drops of H2SO4 were added, to reduce interference by organic matter. Samples were kept at 4°C until elements (i.e., Fe, Al, P) were determined by inductively coupled plasma–atomic emission spectroscopy (ICP–AES; Spectro, Kleve, Germany). All elemental analyses were done in triplicate.

Selected samples were extracted sequentially with a 30-mL solution of 1.0 M ammonium acetate at pH 5.5 for several hours to remove Ca solid phases from the sample and then with ammonium oxalate treatment as described above. The concentrations of Fe, Al, P, and Ca in the two extracts were analyzed by ICP–AES.

The precision and sample homogeneity (i.e., a measure of the amount of error in the data attributable to the sampling technique) of the DPS were quantified by calculating the relative percent difference (RPD). The RPD is the quotient of the difference between field duplicates and the average of those results, expressed as a percentage. Duplicate soil samples were collected approximately every 10 samples. The RPD for DPS varied between 0.9 and 24%.

Sorption Experiments
The sorption characteristics of P in the Histosols were evaluated with the Langmuir isotherm model. Soil sorption experiments were performed in triplicates at 25°C, with a soil to solution ratio of 2 g to 25 mL of 0.03 M KCl (Richardson and Vaithiyanathan, 1995). The concentrations of P additions ranged from 0 to 70 (0, 0.5, 1, 2, 4, 6, 8, 10, 20, 40, and 70) mg L-1, similar to the sorption experiment steps described by Nair et al. (1984). The suspensions were equilibrated for 24 h in polyethylene tubes on a reciprocal shaker, followed by centrifugation at 1500 x g for 20 min. An aliquot of filtered (0.45 µm) supernatant was analyzed for P with the ammonium-molybdate colorimetric method (Murphy and Riley, 1962).

The sorption of P to the soils was calculated with the Langmuir isotherm model:

[3]
where Smax is the maximum amount of solute sorbed (mg kg-1), k is the constant related to binding energy (L mg-1), S0 is the initial adsorbed P present, Ct is the concentration of P in the solution after a 24-h equilibration (mg L-1), and ST is the total amount of sorbed P (mg kg-1). The terms S0, Smax, and k were estimated with the SigmaPlot curve-fitting routine (SPSS, 2000).

Ground Water
Ground water was collected seasonally from 1999 to 2001 at 60 observation wells located across the Hula Valley. The wells were installed to collect ground water from 2 to 5 m below the surface. The water samples were immediately filtered (0.45 µm), kept in an iced cooler, and transferred to the laboratory at the end of each sampling day. The dissolved P concentrations were determined colorimetrically by standard methods. The ground water level, electrical conductivity, temperature, and pH were measured in the field.

Spatial Analysis
The spatial analysis of the DPS was conducted with the stochastic simulation methodology described by Deutsch and Journel (1992). Simulation differs from the more commonly used kriging techniques. While kriging is well suited where soil properties change gradually or where it is necessary to know only the average of rather large blocks, it fails where the soil property exhibits many small-scale variations, especially where the degree of fluctuation varies in space. The krige standard deviation cannot reproduce a realistic picture of this fluctuation because it is based on a general semivariogram and therefore depends only on the local layout of data points and not on the local variation of data. Where significant fluctuations beyond the resolution of a map occur, it is not possible and unnecessary to reproduce a detailed picture of a variable in space but the spatial likelihood of exceeding a certain ecological or regulatory threshold has to be known.

The most straightforward algorithm for generating realizations of the DPS is a sequential Gaussian simulation (SGS). The DPS was simulated sequentially following its conditional cumulative distribution function (ccdf), with a simple kriging system of equations. The conditioning data consisted of all original data and all previously simulated values found within a neighborhood of the location being simulated. Prior to the simulation, the dataset was transformed with a normal score transform routine. The normal score transform is a function that transforms the nonnormal measured data (i.e., DPS) into a standard Gaussian ccdf (called Y variable) with zero mean and unit variance. The spatial analysis of the Y variable was conducted with the traditional variogram equation. We used the trial-and-error approach to fit a model to the experimental variogram, because it requires prior understanding of the spatial arrangement of the data and wise selection of the variogram parameters, such as the range and nugget effect. The range is the distance at which the variogram reaches a plateau and the nugget effect is the separation distance at the origin of the variogram that cannot be further separated due to short-scale variability.

There are several steps involved in the execution of the sequential Gaussian simulation after data transformation and construction of the variogram. First, a random path passing through each node of a 280- x 450-m selected grid was defined. At each node, we specified a number of neighboring conditioning data, including both original data and previously simulated grid node values. The final specifications were set to 2 for a minimum and 10 for a maximum number of original data and 50 for previously simulated data. This specification represents an optimum number of nodes, based on the variogram range and the spatial arrangement of the original data.

The next step included the use of a simple kriging system of equations coupled with the normal score variogram model to compute the mean and variance of the ccdf of the DPS at location u. A simulated DPS value was generated from this ccdf and then added to the data set. The next node was visited and looped until all nodes were simulated. This sequence of steps produced the first realization. To build multiple realizations, the previous sequence was repeated 100 times with a different random path for each realization. The simulated normal values of the 100 sequential Gaussian simulations were backtransformed for the original variable. The post-processing analysis produced two output files. The first was an E-type estimate (mean of the ccdf), which is a point-by-point average of the multiple realizations. The E-type estimate was computed with the following equation:

[4]
where z*E(u) is the E-type estimate at location u, z'k is the kth class mean, and zk and zk+1 are class bounds that define the interval of variability of the DPS. The notation F[u;zk|(n)] represents a function of both the threshold value zk and the available information |(n) (see Deutsch and Journel, 1992 for details). Next, the ccdf for any probability of exceeding a threshold value of interest was computed:

[5]
where the threshold value of P sorption capacity was derived from the literature (e.g., 25% suggested by Breeuwsma and Silva, 1992) or adsorption isotherms of selected typical sites. Because the E-type estimates are average values, 100 simulations were considered a reasonable number of realizations.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
General Characteristics
Sesquioxide concentrations and organic matter contents are depicted according to depositional environs (Table 1). The Histosols that were formed in a swamp exhibited the highest amount of Fe and Al oxyhydroxides compared with the organo–mineral soils that were developed on marl and hydromorphic Vertisols. The Histosols contained a considerable amount of organic matter compared with the organo–mineral soils and the hydromorphic Vertisols. The Histosols exhibited the lowest DPS (Table 1), followed by the organo–mineral soils and the hydromorphic Vertisols.


View this table:
[in this window]
[in a new window]
  		Table 1. Summary statistics of the concentrations of organic matter (OM), Fe, Al, and degree of phosphorus saturation (DPS) in 80 soils of the study area. Values given are means ± standard deviations.

 
The oxalate-extractable P, P sorption maximum, and DPS values of 12 soils representing the Histosols and organo–mineral marl soils are summarized in Table 2. In general, the sorption maxima of the Histosols (Soils 2–9) were significantly higher than those of the organo–mineral soils (Soils 10, 11). The differences in P sorption maxima resulted from the higher sesquioxide and organic matter contents in the Histosols compared with the organo–mineral soils (Table 1). There was a strong correlation (r2 = 0.88, P < 0.001) between the DPS as estimated from the oxalate-extractable P and Fe + Al contents (Eq. [2]) and the DPS values computed from the ratio of oxalate-extractable P to P sorption maximum (Eq. [1]).


View this table:
[in this window]
[in a new window]
  		Table 2. Oxalate-extractable P, P sorption maximum, and degree of phosphorus saturation (DPS) values obtained from 12 selected soil sites representing the Histosols (2–9) and organo–mineral marl soils (10, 11).

 
The results of the sequential extraction with ammonium acetate followed by ammonium oxalate to remove Fe, Al, P, and Ca from the soil samples are summarized in Table 3. These results are also compared with samples that were extracted only with ammonium oxalate, without the ammonium acetate pretreatment. The pretreatment with ammonium acetate extracted large amounts of Ca from the samples with only small amounts of Fe, Al, and P. On the other hand, the ammonium oxalate extraction removed similar amounts of Fe, Al, P, and Ca from the untreated and pretreated samples. The concentrations of Fe, Al, and P extracted by the ammonium oxalate were significantly higher than the P removed by the ammonium acetate. In most samples, the P extracted by the ammonium acetate method was less than 5% compared with the P concentration extracted by the oxalate method. The results suggest that ammonium oxalate is phase-specific, that is, it successfully removes sesquioxides and other elements, such as P that have adsorbed onto these solid phases. The contribution of the Ca-P minerals to the overall P concentration in the ammonium oxalate extract was small. These findings further suggest that the estimates of DPS with the oxalate method are appropriate for the soils of the study area and that further spatial analysis with oxalate-extractable DPS is warranted.


View this table:
[in this window]
[in a new window]
  		Table 3. Concentrations of Fe, Al, P, and Ca extracted from Histosols (1–9) and organo–mineral soils (10 and 11) with ammonium oxalate with and without ammonium acetate pretreatment.

 
Spatial Modeling
The spatial structure of the sesquioxide content and the normal scores of the DPS are depicted in Fig. 2 . The experimental variograms of Feox and Alox exhibit a small relative-nugget effect (nugget/sill < 20%) and a long range of spatial dependency. The Caox variogram exhibited a complete nugget effect (not shown), which suggests that no spatial correlation exists for Ca in the soils extracted by ammonium oxalate. A spherical model best fit the experimental variograms (Fig. 2a,b) derived from the ccdf of the sesquioxides, and a linear model best described the DPS transformed into its normal score values (Fig. 2c). The similarity between the experimental variograms of the sesquioxides and the DPS suggests that the soil processes forming the sesquioxides and governing the DPS are operational across similar distances. The lack of spatial continuity of Caox suggests that the contribution of Ca-P solid phases to the DPS under the extraction conditions is insignificant.



View larger version (16K):
[in this window]
[in a new window]
  		Fig. 2. General relative omnidirectional variograms of oxalate-extractable Fe (a) and Al (b) and a traditional variogram of the normal score values of the degree of phosphorus saturation (DPS) (c). In the construction of these variograms, we used 70 georeferenced locations across the study area with a unit separation that varied between 1200 and 800 m, with a lag tolerance of 600 to 400 m, and a bandwidth of 500 m.

 
Thematic Maps
The E-type estimates of the DPS showed no clear trend among the Histosols (Fig. 3) , all remaining below 20%. Two areas exhibited exceptionally high DPS. The first area was characterized by the hydromorphic Vertisols, which are located in the northeastern section of the Hula Valley, whereas the second consisted mainly of the marl gley soils found in the southeast corner of the valley near the canals' confluence. To further test the accuracy of this contour map, we compared the E-type estimates with the 12 reference soils (Table 3), which were deliberately excluded from the original ccdf from which the map was constructed. All the soils exhibited DPS values significantly smaller than 20%, thus lending further credence to the map illustrated in Fig. 3.



View larger version (49K):
[in this window]
[in a new window]
  		Fig. 3. E-type estimates of the degree of phosphorus saturation (DPS) averaged from 100 realizations derived from the DPS data.

 
The high DPS values observed in the hydromorphic Vertisols strongly attest to the high potential for P leaching from this area to ground water and drainage waterways. An independent verification of this finding and our spatial interpretation was recently provided by a nutrient and water balance exercise for Lake Agmon (MIGAL, 2001). Lake Agmon receives its water from a reconstructed streambed of the Jordan River, a major drainage "Z" canal and via a pumping system from the eastern Hula drainage area, all of which include hydromorphic Vertisols. The highest concentrations of total P and dissolved P were observed in the eastern Hula drainage system (mean total P = 246 ± 170 µg L-1 and mean dissolved P = 190 ± 245 µg L-1). The other contributing sources, such as the Jordan River (mean total P = 200 ± 128 µg L-1 and mean dissolved P = 117 ± 98 µg L-1) and the "Z" canal (mean total P = 134 ± 78 µg L-1 and mean dissolved P = 68 ± 69 µg L-1), exhibited significantly lower concentrations, probably because of the low DPS of the soils in their drainage areas.

Breeuwsma and Silva (1992) concluded for Dutch soils that a DPS of 25% or higher is a sign of deterioration of soil quality. This value has become a widely accepted threshold in soil and water studies that assess the potential for eutrophication. With this DPS standard we built a probability of exceedance map (Fig. 4) to evaluate the areas most susceptible to P leaching. The probability of exceeding a threshold of 25% in the deep, shallow, and eastern Histosols was low. On the other hand, the hydromorphic Vertisols and the marl gley areas were assigned 60% probability of exceeding this threshold and were thus classified as areas susceptible to P leaching (Fig. 4). This information should be included in any future agro- or ecomanagement strategies aimed at minimizing P losses from soils to waterways.



View larger version (66K):
[in this window]
[in a new window]
  		Fig. 4. A contour map of the probability that a given degree of phosphorus saturation (DPS) value exceeds the widely used threshold of 25%.

 
Ground Water
Several studies have shown that the DPS in Histosols can be used as an indicator of the level of dissolved P in ground water. For example, Breeuwsma and Silva (1992) showed that soils with a P sorption capacity of 25% or more maintain a P concentration of 0.1 mg L-1 in ground water. Summary plots based on the median, quartile, and extreme values of dissolved P in the ground water below the soils of the study area are depicted in Fig. 5a according to soil types. The box in Fig. 5 represents the interquartiles range, which contains 50% of the dissolved P values. The whiskers are lines that extend from the box to the highest and lowest values, excluding outliers and extremes. The line across the box represents the median. The outliers are defined as dissolved P values between 1.5 and 3 box lengths and the extreme values as 3 box lengths from the interquartile range. The dissolved P concentrations in the ground water clearly showed a nonnormal distribution, with numerous outliers and extreme values (Fig. 5a). Most of these values were recorded in wells located in Histosols and marl soils where the DPS values were low. The main source for the outliers and extreme values of dissolved P in the ground water is a sampling campaign conducted immediately after the Histosols and marl soils had been fertilized with P and irrigated during the spring of 1999. Since the Histosols in this location are characterized by high P sorption capacity and a low DPS, this rapid enrichment of P in the ground water can only be explained by a preferential flow mechanism that transported a large amount of applied P through cracks and fissures. We excavated 15 deep pits (approximately 5 m) adjacent to most of the observation wells in which high dissolved P values had been measured and observed large cracks, a few centimeters wide, at each location. These cracks and fissures are ubiquitous in most of the Histosols overlying marl and marl soils as a consequence of the drainage and subsequent agromanagement. They seem to be a permanent hydrological feature that exerts significant control over recharge rates in the study area (unpublished data).



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 5. Summary statistics of dissolved phosphorus (DP) concentrations in ground water.

 
If the ground water samples collected immediately after P fertilization are removed from the analysis (Fig. 5b), the average concentrations of dissolved P measured in wells located in Histosols and marl soils are 190 ± 220 and 200 ± 182 µg L-1, respectively. These levels are the lowest in the study area and are in concurrence with the low DPS values of these soils. The average concentrations of dissolved P in ground water collected in wells located in gley marl and hydromorphic Vertisols are 882 ± 775 and 364 ± 362 µg L-1, respectively. These results are in reasonable agreement with the high DPS values of these soils. The overall ground water results suggest that the prediction of dissolved P from DPS in altered wetlands could be difficult if the history of P application is not known and the solute transport by preferential flow is ignored.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Results of the spatial analysis of the DPS in an altered semiarid wetland ecosystem support several conclusions that have useful environmental implications. All the Histosols and marl soils exhibited low DPS, whereas the redoximorphic organo–mineral soils had much higher DPS and are thus potentially more susceptible to P loss to ground water. The magnitude and spatial pattern of the sesquioxides, organic matter, and DPS provide important information that would explain the sorption behavior of P in the peat soils. Spatial analysis with the sequential Gaussian simulation technique suggested that all the Histosols have a low probability (<10%) of exceeding the commonly used threshold of 0.25. The areas characterized by organo–mineral soils, such as hydromorphic Vertisols and marl gley soils, have a high probability (>60%) of exceeding the threshold value of 0.25. The ability to predict the dissolved P concentrations in ground water based on the DPS values is difficult because of the preferential flow characteristics in this altered wetland, but even in cases of significant preferential flow, the residence time of the water within the field may last for many hours and the soil DPS may be of great importance, as was shown by the extreme values of dissolved P in the organo–mineral soils with the higher DPS.


   ACKNOWLEDGMENTS
 
This research was supported in part by the EU project PROWATER, EVK1-CT1999-00036 and the Jewish Agency for Higher Education Program in Eastern Galilee.


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




This article has been cited by other articles:


Home page
 Soil Sci.Home page
S. Grunwald and K. R. Reddy
Spatial Behavior of Phosphorus and Nitrogen in a Subtropical Wetland
Soil Sci. Soc. Am. J., June 18, 2008; 72(4): 1174 - 1183.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
M. I. Litaor, G. Eshel, O. Reichmann, and M. Shenker
Hydrological Control of Phosphorus Mobility in Altered Wetland Soils
Soil Sci. Soc. Am. J., September 20, 2006; 70(6): 1975 - 1982.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
M. I. Litaor, O. Reichmann, A. Haim, K. Auerswald, and M. Shenker
Sorption Characteristics of Phosphorus in Peat Soils of a Semiarid Altered Wetland
Soil Sci. Soc. Am. J., August 25, 2005; 69(5): 1658 - 1665.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
M. I. Litaor, O. Reichmann, K. Auerswald, A. Haim, and M. Shenker
The Geochemistry of Phosphorus in Peat Soils of a Semiarid Altered Wetland
Soil Sci. Soc. Am. J., November 1, 2004; 68(6): 2078 - 2085.
[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 Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (14)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Litaor, M. I.
Right arrow Articles by Shenker, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Litaor, M. I.
Right arrow Articles by Shenker, M.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Litaor, M. I.
Right arrow Articles by Shenker, M.
Related Collections
Right arrow Spatial Distribution
Right arrow Wetlands and Aquatic Processes
Right arrow Preferential Flow
Right arrow Phosphorus

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