| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Apex Environmental Inc., Lenexa, KS 66215
b Department of Plant and Soil Sciences, Oklahoma State University, Stillwater, OK 74078
c Department of Statistics, Oklahoma State University, Stillwater, OK 74078
d Department of Agronomy and Environmental Management, Louisiana State University, Baton Rouge, LA 70803
e STV Incorporated, 80 Ferry Boulevard, Stratford, CT 06615
* Corresponding author (hailin.zhang{at}okstate.edu)
Received for publication December 17, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: DAP, diammonium phosphate • DRP, dissolved reactive phosphorus • M3P, Mehlich-3 phosphorus • Psat, phosphorus saturation index calculated from Mehlich-3 phosphorus and phosphorus sorption maxima • PSIox, ammonium oxalate phosphorus saturation index • PSIWSP, water-soluble phosphorus saturation index • Smax, phosphorus adsorption maximum • STP, soil test phosphorus • TSS, total suspended solids • WSP, water-soluble phosphorus
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
A better understanding of P transport processes will provide useful information for the development of site-specific P management strategies. The P risk index is a tool used to assess various landforms and management practices for potential risk of P movement to water bodies (Lemunyon and Gilbert, 1993). The P risk index considers P source and transport factors (Lemunyon and Gilbert, 1993). Many states have developed or are currently in the process of developing state specific P risk indices (Sharpley et al., 2002). Generally, the factors contributing to P loss potential are weighted and the sum of the weighted factors is the derived relative risk index. However, the weights and relative P loss risk index have been derived primarily from professional judgment, as there is limited site-specific data (Lemunyon and Gilbert, 1993).
Many researchers have found that the P content of the surface soil directly influences the loss of P in runoff (Romkens and Nelson, 1974; Daniel et al., 1994). Other researchers have shown significant relationships between soil P and runoff P concentration (Wendt and Alberts, 1984; Sharpley, 1995; Pote et al., 1999; Cox and Hendricks, 2000).
As there is a relationship between runoff P and soil test P, there is a critical soil test P level. Soil test P levels above this critical point will negatively influence water quality. A common soil P–runoff P management strategy is to identify this critical level in soil and prevent further manure or commercial fertilizer P addition. However, research has indicated that a universal soil P critical limit will not adequately protect water quality because P loss is not constant across all soils. It has been found that the critical levels and relationships between runoff P and soil P are soil specific and dependent on soil and site characteristics (Daniel et al., 1994; Sharpley et al., 1996; Pote et al., 1999; Cox and Hendricks, 2000). Consequently, it is the goal of many state and federal water quality agencies to identify soil P critical levels for benchmark soils (Daniel et al., 1994).
In a review paper, Sharpley et al. (1996) compared the findings from eight studies and found that the relationship between runoff dissolved reactive phosphorus (DRP) and Bray 1 P varied markedly among soils tested, indicating that soil type influences the relationship between runoff P and soil P. Cox and Hendricks (2000) found that two soils with clay contents of 5 and 32% would require soil Mehlich-3 phosphorus (M3P) values of 253 and >700 mg dm–3, respectively, to produce runoff DRP concentrations of 1 mg L–1. Using ammonium oxalate–extractable Fe, Al, and P, Hooda et al. (2000) found that ammonium oxalate P saturation index (PSIox) was the most significant soil property for predicting water desorbable P from contrasting soils. Therefore, runoff P–soil P relationships used to assess the potential for P loss in runoff will probably have to be soil and site specific (Sharpley, 1995).
While many researchers have illustrated the relationship between runoff P and soil P among a wide range of soils, few have compared multiple soils within one study. Using 10 Oklahoma soils, Sharpley (1995) illustrated soil specific relationships between runoff P and M3P. Conversely, Sharpley found that a single regression equation could describe the relationship between DRP and soil P sorption maxima (Smax) for all 10 soils investigated. This study suggests that soil P saturation based on sorption maxima determines the potential for P loss in runoff (Sharpley, 1995).
Using three Ultisols, Pote et al. (1999) also found soil specific relationships between runoff DRP and soil water-soluble phosphorus (WSP). Pote also discovered that normalizing runoff DRP to runoff depth (cm) could correct the differences observed between the three evaluated soils. This finding suggests that the relationship between runoff P and soil P is dependent on soil hydrology.
Researchers have shown that the physical and chemical properties of soil may profoundly influence the relationship between runoff P and soil P. Consequently, it is important to differentiate which properties may have the greatest impact on the runoff P–soil P relationship. Such information is essential to the development of a practical and effective site-specific P risk index assessment and management tool. The objectives of this study were to (i) evaluate the impact of soil P on runoff P in three soil series representative of the most important agricultural soils in Oklahoma across three major land resource areas and (ii) determine the effect of soil physical and chemical properties on runoff P.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
The boxes were saturated using the rainfall simulator (75 mm h–1 until ponding was observed, approximately 10 min) and excess water drained naturally 24 h before rainfall simulation. Rainfall was applied to soil boxes on a 5% slope until 30 min of measurable runoff was collected. From each collection container, total runoff volume was recorded for each soil box and a representative runoff sample (500 mL) was obtained for analyses. In addition, an aliquot (500 mL) of well-mixed runoff from one of the replicates for all soils was collected every 5 min throughout the 30 min runoff duration to evaluate changes in runoff volume and P concentration over time. Soil samples were collected from each box before rainfall simulation, air-dried, sieved to pass a 2-mm screen, and analyzed for different forms of P, which were then correlated with runoff P.
Soil and Runoff Analyses
Mehlich 3–extractable phosphorus (M3P; Mehlich, 1984), water-soluble phosphorus (WSP; Self-Davis et al., 2000), P sorption maxima (Smax), ammonium oxalate–extractable phosphorus (Pox), Al (Alox), and Fe (Feox), texture (Gee and Bauder, 1986), soil organic matter (SOM; Ben-Dor and Banin, 1989), and pH (Thomas, 1996) were determined for the three soils series evaluated. Duplicate analyses were conducted on the study soils. Soil characterization results are summarized in Table 1.
Phosphorus adsorption isotherms were determined according to the method of Graetz and Nair (2000). One gram of soil sample was equilibrated with 25 mL 0.0, 0.5, 1.0, 5.0, 10.0, 15.0, and 20.0 mg P L–1 in 0.01 M CaCl2 solution in 50-mL centrifuge tubes. The tubes were shaken for 24 h on an end-to-end shaker at 150 oscillations per min (opm). The samples were then centrifuged for 10 min at 5211 x g and the supernatant decanted. The P in solution was then quantified colorimetrically using the ascorbic acid method (Kuo, 1996). The amount of P adsorbed was determined by the difference between the initial and final amounts of P in solution. Duplicate analyses were conducted on all study soils.
Phosphorus adsorption isotherms were determined with the linearized form of the Langmuir equation (Eq. [1]):
 | [1] |
The term Smax was calculated by regressing C/S versus C, where C is the equilibrium solution P concentration and S is adsorbed P. The reciprocal of the slope of the linear regression is Smax (Olsen and Watanabe, 1957; Syers et al., 1973).
Acidified ammonium oxalate–extractable P, Al, and Fe (Pox, Alox, Feox) were determined by shaking 1.5-g samples of soil with 30 mL of 0.5 M (COONH4)2·H2O at pH 3.0 in 50-mL centrifuge tubes (Schoumans, 2000) for 2 h, in the dark, on an end-to-end shaker at 150 opm and centrifuged for 10 min at 5211 x g. Supernatants were analyzed for P, Al, and Fe using inductively coupled plasma–atomic emission spectroscopy (ICP–AES). The PSIox was computed using the P, Al, and Fe contents (mmol kg–1) according to Eq. [2] (Schoumans, 2000):
 | [2] |
 | [3] |
Another P saturation index similar to Psat (PSIWSP) was computed using WSP according to Eq. [4]:
 | [4] |
Immediately after rainfall simulation, an aliquot of runoff water sample was filtered (0.45 µm) and analyzed colorimetrically (Murphy and Riley, 1962) to determine dissolved reactive phosphorus (DRP) (Pote and Daniel, 2000). Total phosphorus (TP) from each runoff sample was determined by digesting 25 mL of runoff at 175°C with 1 mL concentrated H2SO4 and 5 mL concentrated HNO3 until a total volume of 1 mL remained (Pote and Daniel, 2000). All digested samples were neutralized and analyzed for P colorimetrically (Murphy and Riley, 1962). Particulate phosphorus (PP) was calculated by subtracting DRP from TP. Runoff TP and DRP loads (mg) were computed by multiplying DRP and TP concentration (mg L–1) by total runoff volume (L).
Total suspended solids (TSS) were determined for all runoff water samples by vacuum filtering (0.45 µm) 50 mL of well-mixed runoff water sample and drying the vacuum filter cup and filter paper at 95°C. Runoff TSS load (mg) was calculated by multiplying TSS (mg L–1) by runoff volume (L) for all runoff water samples.
Statistical Analysis
Analysis of variance (ANOVA) was performed using PROC GLM to determine significant treatment effects within soils (SAS Institute, 2001). When significance was indicated, means were separated by a Fisher's Least Significant Difference procedure using a 0.05 significance level. For comparisons between soils, treatments were classified into four ordinal groups (treatment categories) since the actual treatment level varied from soil to soil. The simple effects of soil given treatment category were assessed with a SLICE option in an LSMEANS statement. Differences in soils given treatment category were further assessed with pairwise t test from a DIFF option (at a 0.05 significance level) if the simple effect from the slice was significant (p < 0.05)
Separate simple linear regressions of DRP on M3P, WSP, Psat, PSIox, and PSIWSP were performed controlling for soil in PROC GLM. Soil was added into the model as a class variable, and soil by independent variable interaction included to assess whether the slopes associated with the three soils differed. A SOLUTION option in the MODEL statement was used to obtain estimates of the slope to compare their values. Other relationships were evaluated using PROC CORR.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The Effect of Fertilizer Phosphorus Addition on Soil Test Phosphorus
Soil samples were collected from all treatments and analyzed periodically over a year to assess the effect of fertilizer P on soil test phosphorus (STP) over time. After fertilizer application, soil M3P increased sharply as the amount of fertilizer P increased (Table 2). The increased M3P in the Kirkland soil decreased slowly after the first 30 d with time until it reached steady state by 210 d for the 180 mg P kg–1 treatment (Fig. 1)
. However, it appears that the 580 and 1080 mg P kg–1 treatments were not at steady state at 210 d and M3P was still gradually declining after 210 d. Similar results were observed for the other two soil series. This suggests some P exists either in solution or as undissolved P fertilizer shortly after P fertilizer application. This added P is easily extractable and susceptible to dissolution and transport during a runoff event resulting in greater P runoff risk. With time, in general, added fertilizer P is gradually precipitated as Fe and Al phosphates in low pH soils or as Ca phosphates in higher pH soils. Ultimately, apatite compounds may form over great periods of time rendering soil P relatively insoluble (Pierzynski et al., 1994; Havlin et al., 1999). As runoff P is directly related to soil test P, the length of time between P application and first runoff event is very important (Sharpley et al., 1994; Daniel et al., 1994). This suggests that as soil test P was highest shortly after P fertilizer application, runoff P would also be highest immediately after fertilizer application.
|
0.97, p < 0.001) existed between M3P and fertilizer P added for all three soils (Fig. 2A) . In addition, a highly significant relationship (r2 = 0.92, p < 0.001) existed between M3P and fertilizer P added for the soils combined (Fig. 2B). Similarly, DeLaune et al. (2004) found a strong positive linear relationship existed between M3P and P added as triple super phosphate (range of 0.0–1344 mg P kg–1) for a Captina silt loam soil. Slopes for the relationships in our study ranged from 0.69 to 0.80 and are similar to the slope reported by DeLaune et al. (2004). The slope reported by DeLaune et al. (2004) for this relationship was 0.26 mg M3P kg–1 per kg P ha–1 added, which is equivalent to a slope of 0.59 when the P added is expressed on a mg P kg–1 basis.
|
Based on the changes of M3P averaged over all three soils one year after P fertilizer addition, it was determined that soil M3P would be raised 1 mg kg–1 by the addition of 1.35 mg kg–1 fertilizer P (soils combined) or a fertilizer P to M3P ratio of 1.35:1. Specifically, the Kirkland and Dennis series had the same results requiring a ratio of 1.25 while the Richfield series required 1.45:1. From a long-term soil fertility field study, Johnson et al. (1998) found that the addition of as much as 6 mg kg–1 fertilizer P is required to raise soil M3P 1 mg kg–1. It is difficult to compare the results of this study with the field data since the field study lasted for more than 25 yr and field P loss beyond crop removal was not included. Our study collected data in a controlled environment over a period of one year and removed many environmental factors that would influence P dynamics such as microorganisms, plants, organic matter reactions, etc. As such, our estimation that soil M3P would be raised 1 mg kg–1 by the addition of 1.35 mg kg–1 fertilizer is probably an underestimation of the amount of fertilizer that would be required in a natural field system where these factors are present. Therefore, the estimation made by Johnson et al. (1998) is probably a more accurate representation of the amount of fertilizer needed in a field situation to raise soil M3P by 1 mg kg–1. Producers and researchers would be able to predict STP if the relationship between STP and P addition can be established, but short-term indoor controlled variable experiments may result in different relationships from those derived from long-term field data and care should be exercised in extrapolating these results to field situations.
Runoff Sample Characteristics
Runoff volumes did not vary (p > 0.05) within treatments for each individual soil series (Table 3). Similarly, runoff volumes were similar (p > 0.05) between soils for most treatments (Table 4). Total suspended solids were statistically equivalent (p > 0.05) within treatments for each individual soil series (Table 3) as well as between soils (Table 4). Conversely, DRP varied within treatments (p < 0.05) for each individual soil series and increased with increasing P addition (Table 3). The DRP also varied (p < 0.05) between soils for most treatments with the exception of the control plots, which showed no difference between soils (Table 4). Similar to DRP, particulate P and total P increased with increasing P addition within treatments for each individual soil series (Table 3). No differences were observed (p > 0.05) between soils for levels of particulate P at Treatments 1 and 2 (Table 4). At higher treatments (i.e., Treatments 3 and 4), particulate P was greater for the Kirkland soil (p < 0.05) as compared to the Richfield and Dennis soils, which were statistically equivalent (p > 0.05) (Table 4). Similar patterns were observed between soils for total P.
|
|
|
Runoff DRP ranged from 0.11 to 3.8 mg L–1 for all treatments (Table 3) and was highly correlated (p < 0.001) with M3P for the Richfield (r2 = 0.95), Kirkland (r2 = 0.92), and Dennis (r2 = 0.95) soil series (Fig. 3A) . Our results for individual soil series are similar to those of other researchers who have reported highly significant relationships between runoff DRP and M3P for individual soil series (Sharpley, 1995; Pote et al., 1996, 1999; Torbert et al., 2002; DeLaune et al., 2004; Kleinman et al., 2004). Slopes for the regressions between runoff DRP and soil P concentrations are referred to as extraction coefficients and are typically used as model inputs for P transport models and P site assessment indices (Sharpley et al., 2002; Kleinman et al., 2004). Extraction coefficients for our study ranged from 0.002 to 0.005. The extraction coefficients for our study are similar to those reported by Sharpley (1995) who reported extraction coefficients ranged from 0.0016 to 0.0072 for unvegetated packed boxes of 10 Oklahoma soils. Our results are consistent with those of Fang et al. (2002) whose extraction coefficient was 0.006 for a study involving unvegetated packed boxes of 10 alkaline soils.
|
Pote et al. (1996) reported a highly significant relationship between DRP and WSP for a Captina silt loam soil in Arkansas that received P inputs as swine and poultry manure. Similarly in our study, DRP was highly related to WSP (p < 0.001) for the individual soil series (Fig. 3B). Correlation coefficients for the individual soil series were: Richfield (r2 = 0.96), Kirkland (r2 = 0.88), and Dennis (r2 = 0.93) (Fig. 3B). Our results are similar to those of McDowell and Sharpley (2001) and Fang et al. (2002) who found highly significant relationships between DRP and WSP. Statistical analysis revealed that each soil series expressed a significantly different relationship (p < 0.05) between DRP and WSP. These findings are consistent with those found by Pote et al. (1999) among three Ultisols and other researchers who have studied the relationship between DRP and WSP (Sharpley, 1995; Cox and Hendricks, 2000; Torbert et al., 2002).
Relationships between Normalized Runoff Phosphorus and Soil Phosphorus
In an attempt to develop one regression equation to describe the relationship between runoff DRP and soil P for a combination of soils, several researchers have normalized DRP and then examined the relationships between DRP and soil P. Pote et al. (1999) normalized runoff DRP concentrations for three Ultisols to correct for observed DRP–soil P relationship differences among soils. In their study, Pote et al. (1999) found that by normalizing runoff DRP concentration one regression equation could be used to adequately describe the relationships between runoff DRP and water-soluble P. Kleinman et al. (2004) normalized DRP by dividing it by rainfall, runoff depth, area, and various combinations of these factors. Their study found that normalizing DRP did not significantly improve the relationships between DRP and M3P. Similarly to Kleinman et al. (2004), normalization of DRP by rainfall, runoff depth, area, and various combinations of these factors did not significantly improve the relationship between DRP and M3P in our study (Table 6). Unlike the study by Pote et al. (1999), normalization of DRP did not improve the relationship between DRP and water-soluble P for our study (Table 6). As the Pote et al. (1999) study was performed on three Ultisols in a field situation, the hydrological properties of the soils evaluated were probably not impacted to the extent of the soils in this study. The soils in this study were removed from the field and sieved. Both of these actions may have influenced the physical properties (e.g., bulk density, pore volume, etc.) of the soils. In addition, this study worked with a much wider range of M3P and WSP and the differences between the regression lines seem to be amplified at high levels of M3P and WSP. Lastly, all soil surfaces in this study had no surface cover while the sites evaluated by Pote et al. (1999) were under well-established tall fescue (Festuca arundinacea Schreb).
|
|
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Highly significant relationships existed between M3P and fertilizer P added as well as between WSP and fertilizer P added. It was found the addition of about 1.35 mg fertilizer P for each kg soil raised M3P one unit (mg kg–1) one year after P addition. As this was an indoor experiment, many environmental variables were controlled thusly providing an accurate assessment of the actual amount of fertilizer P required to raise soil M3P by one unit. This information may also prove useful for future controlled environment experiments involving the addition of fertilizer P to achieve a predetermined soil M3P level, and to estimate time required to reach a critical M3P level for farmers. However, short-term indoor controlled variable experiments may result in different relationships from those derived from long-term field studies and care should be exercised in extrapolating these results to field situations.
Runoff volumes did not vary within treatment for individual soil series nor among soils. Similarly, there was no difference between total suspended solids within treatments for individual soils and no difference between soils for the different treatments. However, DRP increased with increasing P addition within treatments for the soils. The DRP also varied between soils and DRP for Kirkland and Richfield was greater than DRP for the Dennis soil.
Highly significant relationships were found between DRP and M3P and between DRP and WSP for the individual soils. The difference between the slopes of the regressions for both the DRP–M3P relationship and DRP–WSP relationship for different soils indicates the relationships are probably soil specific. Normalization of data did not significantly improve the relationships between DRP and M3P or between DRP and WSP.
Highly significant relationships were found between DRP and Psat and between DRP and PSIWSP for individual soil series. Similarly, a highly significant relationship was observed between DRP and PSIox for individual soil series. Statistical analyses indicated significant differences in the slopes (p < 0.05) for the DRP–Psat relationship and the DRP–PSIWSP relationship. The slopes of the regressions for the Kirkland and Richfield soil series were statistically equivalent (p > 0.05) while the slope of the regression for the Dennis was statistically different (p < 0.05) than the other two soils for the DRP–PSIox relationship indicating that PSIox may be a more appropriate saturation index. Highly significant relationships existed between DRP load and M3P for the individual soil series. Similar relationships were found between DRP load and WSP for the individual soil series. Summarizing, statistical analyses revealed significant differences (p < 0.05) for the slopes of the regressions for the DRP–M3P, DRP–WSP, DRP–PSIox, DRP–PSIWSP, and DRP–Psat relationships indicating the relationships will probably have to be soil specific to be useful in management decisions.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  J. L. Schroder, H. Zhang, D. Zhou, N. Basta, W. R. Raun, M. E. Payton, and A. Zazulak The Effect of Long-Term Annual Application of Biosolids on Soil Properties, Phosphorus, and Metals Soil Sci. Soc. Am. J., January 11, 2008; 72(1): 73 - 82. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. K. Udeigwe, J. J. Wang, and H. Zhang Predicting Runoff of Suspended Solids and Particulate Phosphorus for Selected Louisiana Soils Using Simple Soil Tests J. Environ. Qual., July 17, 2007; 36(5): 1310 - 1317. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. J. Chardon, G. H. Aalderink, and C. van der Salm Phosphorus Leaching from Cow Manure Patches on Soil Columns J. Environ. Qual., January 9, 2007; 36(1): 17 - 22. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Zhang, J. L. Schroder, R. L. Davis, J. J. Wang, M. E. Payton, W. E. Thomason, Y. Tang, and W. R. Raun Phosphorus Loss in Runoff from Long-term Continuous Wheat Fertility Trials Soil Sci. Soc. Am. J., December 2, 2005; 70(1): 163 - 171. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
