| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
USDA-ARS, Northwest Irrigation and Soils Research Laboratory, 3793 N. 3600 E., Kimberly, ID 83341
* Corresponding author (bturner{at}ifas.ufl.edu).
Received for publication September 17, 2003.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: MRP, molybdate-reactive phosphorus • NMWL, nominal molecular-weight limit
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Conventionally, it is considered desirable to know the concentration of free phosphate in runoff, because this is the form most readily available to algae in waterbodies (Reynolds, 1984). This is estimated by crude fractionation based on membrane filtration (typically 0.45- or 0.2-µm pore size) and reaction with molybdate (Murphy and Riley, 1962), but rarely provides an accurate value of the true dissolved phosphate concentration in environmental samples, because a continuum of colloidal particles pass through the membrane filter. Phosphorus associated with such colloids can contribute to the MRP fraction following solubilization in the acidic medium of the reaction solution. Some acid-labile organic phosphates can also be hydrolyzed, but their contribution is negligible in the analysis of soil and soil solution (Bowman, 1989; Hens and Merckx, 2001).
Colloids are operationally defined as particles between 1 µm and 1 nm in size (Kretzschmar et al., 1999). This is somewhat arbitrary, but useful for defining particles with specific properties relating to contaminant transport in the environment. In particular, colloids have a large specific surface area, are mobile in the soil, and remain in solution for long periods of time, although sorption, flocculation, and deposition of colloids can also occur (Kretzschmar et al., 1999). The presence of colloids in runoff therefore has marked implications for understanding the P transfer process, because they facilitate the transport of contaminants such as P from the soil. They also have considerable implications for understanding the effect of P transfer on waterbodies. For example, Hudson et al. (2000) recently demonstrated that dissolved phosphate concentrations in large Canadian lakes were overestimated by orders of magnitude by molybdate colorimetry. This suggests the importance of colloidal P species in maintaining P availability to organisms and highlights the need for more detailed information on P speciation in runoff from terrestrial environments.
Colloidal P compounds can comprise a substantial component of the filterable MRP in soil solution (Shand et al., 2000; Hens and Merckx, 2002), soil water extracts (Sinaj et al., 1998), agricultural runoff, and river water (Haygarth et al., 1997). There is relatively little information on the nature of colloidal P compounds, but they can include primary and secondary P minerals, P occluded within mineral or organic compounds, P sorbed to soil particles or organic molecules, including clay–humic–metal complexes, and live bacterial cells (biocolloids) or cell fragments (Kretzschmar et al., 1999).
Most research into colloidal P in soil waters has focused on acidic and neutral soils with relatively high organic matter concentrations, in which colloidal P typically occurs as humic–metal–phosphate complexes (Gerke, 1992; Dolfing et al., 1999; Hens and Merckx, 2001). However, no information exists on colloidal P in runoff from irrigated soils of the semiarid western United States, which typically contain considerable amounts of CaCO3 and relatively small concentrations of organic C (<10 g kg–1). Such soils are agronomically important, because irrigated agriculture produces nearly 40% of the total U.S. crop value from only 15% of the total cropped land (Bajwa et al., 1992). Much of this land is irrigated by sprinkler systems; for example, in the Pacific Northwest region about two-thirds of the irrigated land (1.7 million ha) is irrigated by sprinklers (USDA National Agricultural Statistics Service, 1999). The aim of this study was to quantify the importance of colloidal MRP in P transfer from calcareous arable soils of the western United States. To achieve this we used micro- and ultrafiltration to measure an operationally defined colloidal MRP fraction in surface runoff generated by simulated sprinkler irrigation, then investigated the influence of soil properties by determining colloidal MRP in water extracts of a range of western U.S. soils.
|
METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Irrigation continued until 30 min of continuous runoff was collected. Total runoff was collected from each subplot in 12-L drums at 5-min intervals from the onset of runoff (determined as continuous flow of water from the collection points). Runoff samples for colloidal MRP analysis were collected in 60-mL plastic bottles and filtered immediately in the field through glass microfiber filters (GF/B; Whatman, Maidstone, UK) to remove suspended sediment. These filters have an approximate pore size of 1 µm, but preliminary tests showed that larger particles routinely passed through the filters. The filtered samples were returned to the laboratory on ice for further filtration and analysis (see below). In addition, parallel samples were filtered in the field through 0.45-µm cellulose-nitrate membranes (Millipore, Billerica, MA) for comparison with 1-µm filtered samples.
Runoff samples from the Portneuf soil were ultrafiltered and analyzed within 6 h of collection, while samples from the other soils were ultrafiltered within 6 h, stored at 4°C, and analyzed within 7 d. We considered this longer storage time acceptable given the high ionic strength in runoff from these calcareous soils, which minimizes changes associated with the determination of filterable MRP (Jarvie et al., 2002). However, the possibility that extended storage influenced colloids (e.g., aggregation, flocculation, sorption–desorption of P) cannot be ruled out. Samples for colloidal MRP determination were collected differently for the three soils. For the Portneuf soil, colloidal P was determined on samples from one subplot only, but all six 5-min samples were analyzed and reported values are flow-weighted means. For the Warden and Greenleaf soils, colloidal P was determined on a single sample from each subplot after 30 min of runoff, and reported values are means of these two samples.
Six portions of soil were taken from the top 2 to 3 cm of each subplot before each rainfall simulation experiment and combined to produce a single bulked sample. These subplot samples were analyzed separately and the results averaged to produce a single value for each main plot. All soils were air-dried at 30°C for 7 d and stored at ambient laboratory temperature for 3 mo before analysis by four soil-test P procedures. Olsen P, an agronomic test for plant-available P, was determined by extraction with 0.5 M NaHCO3 (adjusted to pH 8.5) for 30 min in a 1:20 soil to solution ratio (Olsen et al., 1954) and filtration through Whatman no. 42 filter papers before analysis. Water-extractable P was determined in two soil to solution ratios (1:10 and 1:200) using deionized water. Samples were shaken horizontally (160 min–1) for 1 h and centrifuged at 3000 x g for 15 min. Calcium chloride–extractable P was determined by extraction with 0.01 M CaCl2 for 1 h in a 1:10 soil to solution ratio. Both water and CaCl2 extracts were filtered through 0.45-µm cellulose-acetate membranes (Nalgene, Rochester, NY) before analysis. In all procedures, extractions were conducted at 20°C and MRP was determined by the method of Murphy and Riley (1962). Soil pH was determined in a 1:2 soil to deionized water mixture, and moisture content was determined by drying a 100-g sample at 105°C for 24 h. Organic C concentrations were determined by dichromate digestion (Nelson and Sommers, 1982) and CaCO3 concentrations by acid titration (Allison and Moodie, 1965).
Water Extracts of Western U.S. Soils
Twelve soils, selected to provide a range of properties typical of irrigated arable soils, were sampled to 30 cm from locations around the western United States during 2000 (Table 2). All soils were under irrigated arable cropping except the native Portneuf, which was under native (nonirrigated) sagebrush (Artemisia tridentata Nutt.). The sites were semiarid, with mean annual temperatures from 5.0°C at Fargo, ND, to 13.8°C at Amarillo, TX, and mean annual precipitation from 209 mm at Othello, WA, to 547 mm at Pullman, WA. The soils contained organic C concentrations between 4.5 and 18.2 g C kg–1 dry soil, CaCO3 concentrations between <1 and 480 g kg–1 dry soil, and pH values between 5.16 and 8.11. The soils were air-dried (30°C), sieved (<2 mm), and stored at ambient laboratory temperature before analysis. Water- and CaCl2–extractable MRP were determined by mixing 5 g of soil and 1 L of solution (1:200 soil to solution ratio) in rotating drums (20 rpm) for 4 h. Samples were filtered through 0.2-µm cellulose-acetate membranes (Nalgene) and analyzed for MRP (see below).
|
Colloidal Molybdate-Reactive Phosphorus Analysis
After prefiltration, all samples (runoff and water extracts) were microfiltered through 1.0-µm Cyclopore Track Etched Membranes (Whatman) to define an upper size limit for colloidal MRP. These polycarbonate surface-capture membranes have precise cylindrical pores that provide a distinct particle size cut-off. Separations are therefore accurate and reproducible compared with those using standard membranes (e.g., cellulose-nitrate). All membranes were 47 mm in diameter, and sample loading (1–2 mL cm–2) was minimized to reduce clogging. As well as the 1.0-µm membranes, water extracts were also filtered through three other pore-size Cyclopore membranes to fractionate large colloidal particles (Table 3).
|
Runoff samples (<1 µm) were ultrafiltered directly through a 3000-NMWL membrane to provide an operationally defined colloidal fraction (i.e., 1 µm–1 nm). Water extracts were micro- and ultrafiltered through a cascade of membranes to provide detailed information on colloidal size fractions (Table 3). Filtration was performed in parallel (i.e., aliquots of each water extract were filtered independently through a single membrane). All filtrates were analyzed for MRP, total P, and major cations. Total P and soluble cations (Al, Ca, Fe, K, Mg, Si) were determined by inductively coupled plasma–optical emission spectrometry (ICP–OES), with limits of detection (determined as the mean + 1.96 standard deviations of 23 blanks) being (µM): Al = 0.08, Ca = 0.25, Fe = 0.02, K = 0.26, Mg = 0.08, P = 0.06, and Si = 0.07. Molybdate-reactive P (limit of detection = 0.12 µM) was determined at 880 nm using a 1-cm cell and a 12-min reaction time (Murphy and Riley, 1962). Molybdate-unreactive P, which includes organic P, inorganic polyphosphates, and acid-resistant colloidal P, was calculated as the difference between total P and MRP, but concentrations were negligible in almost all samples and are not reported.
It is possible that Si interference may have influenced MRP determination in this study, but we consider this unlikely. In particular, interference is negligible in solutions with relatively low Si to P ratios similar to those analyzed here, and is also minimized when a short reaction time and a relatively strong H2SO4 concentration are used (as here) in the colorimetric procedure (Ciavatta et al., 1990).
Acidification of Water Extracts
Water extracts of three soils (Portneuf topsoil, Portneuf subsoil, Greenleaf) were further investigated by acidification before ultrafiltration. Filtered samples (<1 µm) were acidified with HCl to various pH values to pH 2. After 10 min the solutions were neutralized with NaOH (to avoid damage to ultrafiltration membranes) and colloidal MRP was determined by ultrafiltration through a 1-nm (3000-NMWL) membrane.
Statistical Analysis
Flow-weighted mean filterable MRP and sediment concentrations were calculated using concentration and flow volume data from 5-min samples. Statistical differences in regression models describing the relationship between the concentrations of soil-test P and colloidal MRP in runoff were determined using procedures described by Neter and Wasserman (1974) to indicate the significance of differences in gradients and intercepts. Data from dry and wet runs were pooled before analysis.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
Concentrations of colloidal Ca and Mg in water extracts were considerably greater in the more calcareous soils. For example, concentrations of colloidal Ca ranged from 1.18 µM in extracts of the acidic Palouse silt loam, to 95.63 µM in extracts of the highly calcareous conventionally managed Portneuf subsoil (Table 5). Colloidal Ca concentrations were proportionally smallest for the native Portneuf soil (7%) and greatest for the conventionally fertilized Portneuf subsoil (41%). Water extracts of the latter soil also contained the greatest concentration of colloidal Mg (18.24 µM). Proportions of colloidal Mg were smallest for the manured Portneuf subsoil (29%) and greatest for the Brinegar loam (53%). Most colloidal Ca occurred in the 1.0- to 0.4-µm and 3- to 0.3-nm fractions, whereas colloidal Mg was more evenly distributed across the whole colloidal size range (Fig. 3).
Colloidal MRP concentrations in water extracts were not significantly correlated with any of the colloidal elements, but were strongly positively correlated with soil-test P concentrations (Fig. 4). Indeed, similar relationships existed between soil-test P and colloidal MRP for both agronomic (Olsen P) and environmental (water and CaCl2) soil-test P procedures. Proportions of colloidal MRP were greater in alkaline soils containing large concentrations of CaCO3 and smaller concentrations of organic C (Fig. 4).
|
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The importance of colloidal MRP in runoff from calcareous soils was unexpected, because colloidal MRP is conventionally perceived as being most important in neutral or acidic soils with relatively high organic matter concentrations, in which humic–metal–phosphate complexes predominate (e.g., Gerke, 1992; Hens and Merckx, 2001). For example, water extracts of three acidic soils rich in Al and Fe contained much greater proportions of colloidal MRP than an extract of a calcareous Xerochrept from Albania (Sinaj et al., 1998). In the current study, it is possible that the use of reverse-osmosis water in irrigation simulation experiments facilitated colloid detachment, because soil dispersion is favored by low-ionic-strength water (Kretzschmar et al., 1999). This suggests that colloidal MRP might be quantitatively less important in runoff generated by the relatively high-ionic-strength irrigation water used in the western United States, the main source of runoff from farmland in this semiarid region.
Colloidal MRP was associated with two size fractions. The larger fraction (1.0–0.2 µm) was mainly associated with Al and Fe, so probably represented phosphate sorbed to clay-sized particles. This fraction was almost completely removed by filtration through a 0.2-µm membrane, and probably constituted the majority of the colloids in runoff from the Warden soil. A second smaller size fraction (3–0.3 nm) was associated with colloidal Ca and Mg, suggesting the presence of Ca- and Mg-phosphate minerals. These smaller particles probably constituted the majority of the colloids in runoff from the more calcareous Greenleaf and Portneuf soils. Based on the mineralogy of these calcareous soils, it is probable that colloidal compounds included ß-tricalcium phosphate, octocalcium phosphate, and possibly some hydroxyapatite (Robbins et al., 1999). It is unlikely that Mg-phosphates would contribute, because they are relatively soluble and only likely to occur ephemerally following phosphate fertilization (Lindsay et al., 1962). However, it is possible that Mg is associated with some of the Ca-phosphate minerals, which may explain the strong correlations between colloidal concentrations of MRP and Mg. Furthermore, it is interesting to note that Ca- and Mg-phosphate complexes can exist in solution between pH 7 and 9 if sufficient Ca and Mg ions are present (Lindsay, 1979). Molecular weights of such complexes are <1000, suggesting that at least some of the MRP that passed through the finest pore-size filter membrane (1000 NMWL, 0.3 nm) was not free phosphate. Confirmation of the contribution of Ca- and Mg-phosphates to the colloidal MRP fraction requires detailed investigation of colloidal particles by analytical techniques such as scanning electron microscopy.
The presence of large concentrations of colloidal P in surface runoff places an important limitation on our understanding of P transfer from western U.S. soils. For example, large concentrations of colloidal MRP suggest a greater mobility of MRP in the environment, because phosphate is readily removed from solution by sorption to soil and sediments, whereas colloids are much more mobile and can escape relatively easily to surface water systems (Kretzschmar et al., 1999). Such mobility can facilitate P leaching to depth in soils that would otherwise retain phosphate in the profile (Turner and Haygarth, 2000). In addition, evaluation of mechanisms involved in phosphate desorption is difficult when a large proportion of the filterable MRP is associated with colloidal particles and, therefore, not technically solubilized (Hamon and McLaughlin, 2002).
The presence of large concentrations of colloidal P in runoff also has important implications for the impact of P transfer on waterbodies. Free phosphate is considered to be readily bioavailable to aquatic microorganisms and, therefore, to pose the greatest threat to water quality (Reynolds, 1984; Haygarth and Jarvis, 1999). However, if much of the MRP in runoff from calcareous soils actually consists of mineral phosphates that become solubilized only at pH values unlikely to be encountered in the environment, then at least a fraction of the MRP cannot be considered readily bioavailable. This suggests that simple assumptions concerning the potential bioavailability of MRP in runoff from calcareous soils may be misleading, but does not preclude the possibility that aquatic microorganisms possess mechanisms to access colloidal P (e.g., by inducing local changes in pH), nor that phosphate may be released from colloids following changes in the carrying solution (e.g., sediment to solution ratio). Indeed, it is likely that the free phosphate concentration in solution fluctuates continuously in response to a range of physical, chemical, and biological processes during transport (Baldwin et al., 2002). It may be possible to estimate the potential bioavailability of the colloidal P fraction by adaptation of an anion-exchange resin procedure described recently for the particulate fraction (Uusitalo et al., 2003).
It seems difficult to overcome problems associated with the presence of colloidal P species in runoff, at least for routine analysis. Ion chromatography provides a much closer approximation of free phosphate in solution, but is relatively impractical for routine analysis of large numbers of samples, and phosphate sorption to colloidal particles can still interfere in the procedure (Sinaj et al., 1998). The use of 0.2- rather than 0.45-µm membrane filters will eliminate interference from fine clays (as often used by limnologists and recommended by Sinaj et al., 1998), but not the presence of the finer colloidal fraction. Ultrafiltration provides a more accurate measure of free phosphate, but lengthens analytical time to impractical levels for routine analysis. However, ultrafiltration or ion chromatography of a small number of representative samples will at least indicate the likely importance of colloidal MRP in the particular environment under investigation, and is recommended for future studies of MRP in runoff from agricultural land.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Current address: Soil and Water Science Department, University of Florida, 106 Newell Hall, P.O. Box 110510, Gainesville, FL 32611. 
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  A. N. Sharpley, P. J. A. Kleinman, A. L. Heathwaite, W. J. Gburek, G. J. Folmar, and J. P. Schmidt Phosphorus Loss from an Agricultural Watershed as a Function of Storm Size J. Environ. Qual., March 1, 2008; 37(2): 362 - 368. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. G. Warren, C. J. Penn, J. M. McGrath, and K. Sistani The Impact of Alum Addition on Organic P Transformations in Poultry Litter and Litter-Amended Soil J. Environ. Qual., March 1, 2008; 37(2): 469 - 476. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Kleinman, D. Sullivan, A. Wolf, R. Brandt, Z. Dou, H. Elliott, J. Kovar, A. Leytem, R. Maguire, P. Moore, et al. Selection of a Water-Extractable Phosphorus Test for Manures and Biosolids as an Indicator of Runoff Loss Potential J. Environ. Qual., July 17, 2007; 36(5): 1357 - 1367. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Schelde, L. W. de Jonge, C. Kjaergaard, M. Laegdsmand, and G. H. Rubaek Effects of Manure Application and Plowing on Transport of Colloids and Phosphorus to Tile Drains Vadose Zone J., March 8, 2006; 5(1): 445 - 458. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. F. Koopmans, W. J. Chardon, and C. van der Salm Disturbance of Water-Extractable Phosphorus Determination by Colloidal Particles in a Heavy Clay Soil from the Netherlands J. Environ. Qual., July 5, 2005; 34(4): 1446 - 1450. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
