SSSAJ Grow Your Career with SSSA
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)
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 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 ISI Web of Science (62)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sims, J. T.
Right arrow Articles by Pautler, M. C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Sims, J. T.
Right arrow Articles by Pautler, M. C.
Agricola
Right arrow Articles by Sims, J. T.
Right arrow Articles by Pautler, M. C.
Related Collections
Right arrow Water Pollution
Right arrow Phosphorus
Right arrow Soil Analysis
Soil Science Society of America Journal 66:2016-2032 (2002)
© 2002 Soil Science Society of America

DIVISION S-8—NUTRIENT MANAGEMENT & SOIL & PLANT ANALYSIS

Evaluation of Mehlich 3 as an Agri-Environmental Soil Phosphorus Test for the Mid-Atlantic United States of America

J. T. Sims*, R. O. Maguire, A. B. Leytem, K. L. Gartley and M. C. Pautler

Dep. Plant and Soil Sciences, Univ. Delaware, Newark, DE 19717-1303. Paper No. 1708 in the journal series of the Delaware Agric. Expt. Stn

* Corresponding author (jtsims{at}udel.edu)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 NOTES
 CONCLUSIONS
 REFERENCES
 
Laws and guidelines limiting P applications to cropland based on soil P exist in the Mid-Atlantic USA because of water quality concerns. We evaluated Mehlich 3 (M3) as an environmental soil P test using 465 soils typical to the Mid-Atlantic region and found M3-P accurately predicted water soluble P (WSP), desorbable P (Fe oxide strip P [FeO-P]), and total sorbed P (oxalate P). The M3-P saturation ratio (M3 [P/(Al+Fe)]) was linearly related to the well-established oxalate P saturation method (DPSox) and a M3 [P/(Al+Fe)] range of 0.10 to 0.15 corresponded to reported environmental limits for DPSox (25–40%). Rainfall simulation and column leaching studies showed M3 [P/(Al+Fe)] predicted runoff and leachate P concentrations better than M3-P. We suggest consideration of the following approach now used in Delaware for agri-environmental interpretation of M3-P and M3 [P/(Al+Fe)]: (i) Below optimum (crop response likely; M3-P <= 50 mg kg-1; M3 [P/(Al+Fe)] < 0.06); (ii) Optimum (economic response to P unlikely, recommendations for P rarely made; M3-P = 51–100 mg kg-1; M3 [P/(Al+Fe)] = 0.06–0.11); (iii) Above Optimum (soil P will not limit crop yields, no P recommended; M3-P > 100 mg kg-1; M3 [P/(Al+Fe)] > 0.11); (iv) Environmental (implement improved P management to reduce potential for nonpoint P pollution—in Delaware M3-P > 150 mg kg-1; M3 [P/(Al+Fe)] > 0.15 is now used). (v) Natural Resource Conservation (no P applied even if the potential water quality impact is low to conserve P, a finite natural resource).

Abbreviations: Alox, acid ammonium oxalate extractable Al • DRP, dissolved reactive P • FeO-P, Fe oxide strip P • Feox, acid ammonium oxalate extractable Fe • DPSox, degree of P saturation by the oxalate method • ICP-AES, inductively coupled plasma atomic emission spectroscopy • M3, Mehlich 3 • OM, organic matter • Pox, acid ammonium oxalate extractable P • PSI, P site index • TMDLs, total maximum daily loads • UDSTP, University of Delaware Soil Testing Program • WSP, water soluble P • ***, significant at the 0.001 probability level


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 NOTES
 CONCLUSIONS
 REFERENCES
 
MINIMIZING NONPOINT SOURCE POLLUTION of surface waters and shallow ground waters by P originating in agricultural soils is a national environmental priority in the USA and many other countries today (Parry, 1998; Sharpley et al., 2000; Withers et al., 2000). While the role of P in surface water eutrophication has been recognized for decades, efforts in the USA to identify agricultural settings where improved P management practices are needed to protect water quality have intensified in the past 5 yr. A number of factors have contributed to this increased emphasis on P management. For example, federal and state water quality monitoring and assessment studies completed in the past decade have consistently identified agriculture as a major source of P to many inland and coastal waters; many of these waters are also considered impaired because of excessive nutrient loading (Boesch et al., 2001; USGS, 1999). In the 1998 National Water Quality Inventory Report, the U.S. Environmental Protection Agency (USEPA) reported that agriculture was the leading cause of water quality impairment of the rivers and lakes in the USA and that nutrients were the leading stressor in lakes, ponds, and reservoirs (USEPA, 2000). In some states, in response to reports such as these, and to highly publicized pollution incidents such as fishkills and algal blooms (Burkholder et al., 1997), lawsuits have been initiated by environmental action groups against USEPA for failing to protect the waters of the USA as required by the Clean Water Act. These lawsuits have typically resulted in settlements that require state environmental agencies to establish total maximum daily loads (TMDLs) for nutrients and other pollutants in surface waters above which the waters are no longer considered fishable or swimmable. Once TMDLs are established, watershed-scale pollution control strategies must be implemented to reduce pollutant loading, from point and nonpoint sources, below the TMDL.

The situation in the Mid-Atlantic USA reflects much of what has occurred nationally. Water quality monitoring reports show surface waters impaired by nutrients, fishkills and algal blooms plague surface waters regularly, and TMDLs are either established or are in litigation (Shedlock et al., 1999; Sims and Coale, 2002). Four states in the region (Delaware, Maryland, Pennsylvania, and Virginia) have passed nutrient management laws that restrict agricultural land application of nutrients in some way (Beegle et al., 1997; Simpson, 1998; Sims, 2000a). A key component of all of these state actions has been the requirement of some form of P-based management by agriculture, and particularly by animal-based agricultural operations that land-apply manure. For example, the Delaware Nutrient Management Act of 1999 allows a maximum P application rate to high P soils equivalent to the "three year crop removal rate" (Sims, 1999); thus, high P soils in Delaware will usually only receive P applications once every 3 yr. The Virginia Poultry Waste Management Bill of 1999 targets the state's poultry industry and mandates that, depending upon soil test P level, manure P application rates shall not exceed the greater of crop nutrient needs or crop nutrient removal. Regulations recently promulgated under the Maryland Water Quality Improvement Act of 1998 require that a P Site Index (PSI) assessment (a field-specific evaluation of the effects of transport, source, and management factors on the potential for P loss to waters; Lemunyon and Gilbert, 1993) be conducted for any soil with a soil test (Mehlich 1) P concentration more than three times the agronomic critical value of 25 mg kg-1 (i.e., if Mehlich 1 P is > 75 mg P kg-1). In addition to state actions, national changes have occurred. For example, the USDA Natural Resources Conservation Service (NRCS) has revised its Code 590 nutrient management standard to provide three options to identify the need for P-based management: (i) agronomic soil test P; (ii) a threshold soil P value; and (iii) the PSI. As any of these measures of the potential for P loss from soil to water increases, the amount of manure or wastewater P that can be applied decreases. The USEPA has proposed new regulations for Concentrated Animal Feeding Operations (CAFOs) that mandate P-based management for the land application of manures and wastewaters (USEPA, 2001).

It seems apparent that these changes in state and federal laws and guidelines will require more intensive P management by agriculture and that fields requiring greater management will be identified by some form of soil P testing, either alone or in combination with other site and management factors. Ideally, standard agronomic soil tests used to identify crop P needs for optimum yields could also be used in an environmental assessment of the P loss potential of a field. If not, soil testing laboratories would need to evaluate and verify the validity of environmental soil P tests (e.g., degree of soil P saturation, WSP, easily desorbable P). They would also need to conduct separate analyses of the same soil sample for agronomic and environmental purposes, which would be time-consuming and expensive. While there is justifiable concern that agronomic soil P tests were not developed for environmental purposes, recent studies show that the agronomic P tests are often well-correlated with soil P measured by environmental P tests as well as with measurements of soluble P and P in runoff and leaching waters (Chardon and Faassen, 1999; Hesketh and Brookes, 2000; Hooda et al., 2000; McDowell and Sharpley, 2001a; Pote et al., 1996; Sims et al., 2000).

The Mehlich 3 extractant is widely used in the Mid-Atlantic USA and in other states as an agronomic soil test for P and other plant nutrients. It has also recently been proposed as an "agri-environmental soil test" for eastern Canada, where it was shown to accurately assess soil P saturation and predict potato (Solanum tuberosum L.) yield responses to P (Bolinder et al., 1998; Khiari et al., 2000). Our objective in this study was to evaluate, and propose interpretive guidelines for, the use of Mehlich 3 as an agri-environmental soil P test for the Mid-Atlantic USA.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 NOTES
 CONCLUSIONS
 REFERENCES
 
Soil Sample Collection and Data Set Organization
The data set used in this study was obtained by combining selected results from five studies we conducted during 1998 through 2001. Brief descriptions of each study follow:

Study 1.

This study evaluated the relationship between soil test P, soil P saturation, and soluble P using 120 agricultural soil samples (topsoils, recommended sample depth is 0–20 cm) obtained from the University of Delaware Soil Testing Program's archives (Pautler and Sims, 2000).

Study 2.

This study characterized soil P concentrations and P saturation in the profiles of agricultural soils located on 20 farms in southern Delaware (K.L. Gartley and J.T. Sims, University of Delaware, unpublished data, 1998). At four farms, we also obtained samples from forested areas immediately adjacent to the agricultural fields. Soil samples were collected at each site at depths of 0 to 20, 20 to 40, 40 to 60, and 60 to 90 cm.

Study 3.

This study was conducted to compare the effects of long-term biosolids applications on soil P (Maguire et al., 2000). We collected soil samples from biosolids-amended soils and adjacent setback areas (areas with similar soils and crop management practices, but where biosolids were not applied) at 17 sites in Delaware, Maryland, and Virginia. Soil samples were collected at each site at depths of 0 to 5, 5 to 20, 20 to 40, and 40 to 60 cm.

Study 4.

This study was conducted to determine the relationship between soil P and runoff P using a rainfall simulation runoff box method (A.B. Leytem and J.T. Sims, University of Delaware, unpublished data, 2001). Bulk soil samples were collected from the 0- to 20-cm depth of five soil series in Delaware that ranged in soil test P concentrations and also varied in soil texture and organic matter (OM) content. The soil series used included: Butlertown silt loam (fine-silty, mixed, semiactive, mesic Typic Fragiudults), Evesboro loamy sand (mesic, coated Typic Quartzipsamments), Matapeake silt loam (fine-silty, mixed, semiactive, mesic Typic Hapludults), Pocomoke sandy loam (coarse-loamy, siliceous, thermic Typic Umbraquults), and Sassafras sandy loam (fine-loamy, siliceous, semiactive, mesic Typic Hapludults). After collection all soils were air-dried, sieved to pass a 7-mm screen, and poured into wooden runoff boxes approximately 100 by 20 by 5 cm in size, leveled, and presaturated 24 h before being placed under a rainfall simulator. Three replications were used for each soil. Soil samples were taken from the runoff boxes to the depth of 5 cm immediately before rainfall simulation for analyses, as described below. The rainfall simulator consisted of a single "Tee Jet" HH-SS-50WSQ nozzle (Spraying Systems Co., Wheaton, Il) attached to a 3 by 3 by 3 m metal frame, and calibrated to achieve an intensity of 7.5 cm h-1 at 90% uniformity. Prepared soil runoff boxes were placed randomly under the rainfall simulator on steel racks adjusted to a 5% slope. Rainfall events were 30 min long and all runoff was collected directly into 9-L plastic containers.

Study 5.

This study evaluated the relationship between a range of environmental soil P tests and P losses in leaching using undisturbed soil columns (15-cm i.d., 20-cm depth) collected from the same five soil series used in the rainfall simulation study described above (Maguire and Sims, 2002). Columns were collected in the same fields where the bulk soil samples for the runoff study had been obtained. At the time of column collection, soil samples were also obtained directly outside each column to a depth of 20 cm and composited for analysis, as described below. After collection, end caps with a drainage outlet were placed on the lower end of the undisturbed columns and the columns were placed in racks in the greenhouse. The columns were prewet with deionized water and left to drain to field capacity for 2 d. Each column was then leached with the equivalent of 5-mm rainfall using deionized water. Leachate samples were analyzed as described below.

Analyses of Soils, Runoff Waters, and Column Leachates
All soil samples were air-dried, sieved to pass a 2-mm stainless steel screen, and analyzed for (i) WSP (1:10 [w/w] soil/deionized water, 1-h reaction time, filtration with a 0.45-µM millipore membrane; Self-Davis et al., 2000); (ii) FeO-P (1:40 [w/w] soil/0.01 M CaCl2 + Fe-oxide coated filter paper strip, 16-h reaction time, followed by dissolving Fe and P from the filter paper strip for 1 h in 1 M H2SO4; Chardon et al., 1996); (iii) M3-P, Al, and Fe (M3; 1:10 [w/w] soil/0.2 M CH3COOH + 0.25 M NH4NO3 + 0.015 M NH4F + 0.13 M HNO3 + 0.001 M EDTA, 5 min reaction time and filtration through Whatman #42 filter paper; Mehlich, 1984); and (iv) oxalate extractable P, Al, and Fe (Pox, Alox, Feox; 1:40 soil/0.2M acid ammonium oxalate [pH 3], 2-h reaction time in the dark; McKeague and Day, 1966). All soil extracts were analyzed for P, Al, and Fe by inductively coupled plasma atomic emission spectroscopy (ICP-AES). We also determined pH (1:1 soil/deionized water) and OM by loss on ignition for all soils using standard methods of the University of Delaware Soil Testing Program (Sims and Heckendorn, 1991).

Runoff water samples from Study 4 and column leachate samples from Study 5 were filtered through 0.45-µM millipore filter paper immediately after collection and analyzed for dissolved reactive P (DRP) by colorimetry (Murphy and Riley, 1962).

Soil Phosphorus Saturation Calculations
The degree of P saturation (DPSox) by the acid ammonium oxalate extraction method was calculated for all soils as follows (values in millimoles per kilogram; Schoumans, 2000):

[1]
where, for noncalcareous soils, {alpha} is an empirical parameter that can be used to relate total soil P sorption capacity to (Alox + Feox) (van der Zee et al., 1987). Values for {alpha} typically range from 0.4 to 0.6; based on previous work with Delaware soils and other studies in the literature with similar soils, we used an {alpha} value of 0.5 (Beauchemin and Simard, 1999; Lookman et al., 1995; Pautler and Sims, 2000; Schoumans, 2000). For sites in Study 2 where we collected subsoil samples, we also determined DPSox for the soil profile to a depth of 90 cm by first calculating the total mass of Pox, Alox, and Feox in each soil depth, per hectare:

[2]
and then summing across the soil depths to obtain the total mass of Pox, Alox, and Feox in the soil profile, per hectare:

[3]

From these values we calculated [DPSox]Profile as:

[4]

We estimated the bulk density for each soil layer from information on soil texture obtained for each soil series from the Sussex Country, Delaware soil survey manual (USDA-SCS, 1974). Values for bulk density ranged from 1300 (sandy clay loam) to 1600 kg m-3 (sand) as a function of the texture of the soil layer.

We also calculated two M3-P saturation ratios for all soils (all values in millimole per kilogram):

[5]

[6]

Note that we did not use an {alpha} value to determine the M3-P saturation ratios because we have no experimental evidence to support a quantitative value for this parameter. Therefore we report empirical saturation ratios for M3-P, not percentages of total P sorption capacity as with the oxalate method (DPSox). This approach is similar to that used by Khiari et al., (2000) to calculate a M3-P saturation index for acid, coarse-textured Canadian soils.

Statistical Analyses
All correlation and regression analyses were conducted by standard procedures of SAS Version 8.0 (SAS Institute, 1998). For runoff and column leaching studies, a split-line model, as described by McDowell and Sharpley (2001a) was used to determine change points for the relationships between M3 soil P saturation and DRP in runoff or leachate. The split-line model describes two linear relationships for the data, one below and one above the change point, and uses nonlinear regression techniques in SAS 8.0 to estimate the change point and the slopes and intercepts of these two lines.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
NOTES
 CONCLUSIONS
 REFERENCES
 
Soil Characteristics
The data set for this study consisted of 465 soils (335 topsoils and 130 subsoils) primarily obtained from crop land used to produce corn (Zea mays L.), small grain, and soybean (Glycine Max L.) (16 soils were from topsoils and subsoils in forests directly adjacent to agricultural fields). Major soil orders represented were Ultisols, Entisols, and Inceptisols; surface textures were sandy loams, loamy sands, and silt loams. In general, as is typical for many soils in the Mid-Atlantic region of the USA, the 465 soils were moderately acidic (median pH = 5.7), low in OM (median OM = 16 g kg-1), and high in soil test P (median M3-P = 115 mg kg-1; agronomic optimum M3-P = ~30–50 mg kg-1; SERA-IEG-6, 2001; Sims and Gartley, 1996; SPAC, 1999) (Table 1). Agricultural topsoils had higher median pH, OM, and M3-P values (5.9, 17 g kg-1, 158 mg kg-1) than subsoils (5.5, 10 g kg-1, 36 mg kg-1). In comparison, median values for ~18000 soils tested for pH, OM, and M3-P1 in 1993 through 1999 by the University of Delaware Soil Testing Program (UDSTP) were 6.0, 14 g kg-1, and 128 mg kg-1.


View this table:
[in this window]
[in a new window]
  		Table 1. Selected properties of the 465 soils used in the study.

 
We also characterized several forms of P, Al, and Fe and found that, on average, M3 extracted 53% of total sorbed P (Pox) and 84% of Alox, but only 19% of Feox; about 4 and 17% of M3-P was extracted as WSP or easily desorbable P (FeO-P) (Table 1). The DPSox averaged 39.5%, in the range reported to be of environmental concern (25–40%, using an {alpha} value of 0.5; Breeuswma et al., 1995; De Smet et al., 1996). Average M3 [P/(Al+Fe)] and M3 (P/Al) saturation ratios for the soils in our study were 0.15 and 0.16. In comparison, Khiari et al. (2000) characterized Quebec soils with M3 (P/Al) ratios greater than 0.15 as being in the high environmental risk group.

Examination of the results for the soils collected in Study 2 provides more detailed insight into the changes in soil P that have occurred in an agricultural setting dominated by geographically concentrated animal production, such as the Delmarva Peninsula. These soil samples were collected on poultry farms in southern Delaware where manure and fertilizer P have been applied for years for the production of grain crops. Because of the lack of alternatives to land application, and the unfavorable N/P ratio in the manure, P has been overapplied relative to crop needs, resulting in a buildup of soil P to values well above those needed for crop production (Table 2). This is a common scenario in the poultry producing region of Delmarva and other areas of concentrated animal production in the Mid-Atlantic region of the USA. More information on the general nature of agriculture, manure management, and water quality on the Delmarva peninsula is found elsewhere (Cabrera and Sims, 2000; Mozaffari and Sims, 1994; Sims et al., 2000).


View this table:
[in this window]
[in a new window]
  		Table 2. Melich 3 P concentrations in topsoils and estimated P saturation in soil profiles of selected soils from Delaware's coastal plain. Data are from Study 2.

 
As illustrated in Fig. 1a and 1b , which compare profile M3-P and DPSox in four agricultural fields with values in forest soils directly adjacent to these fields, this long-term overapplication of P has resulted in the movement of P downward in the soil profile and P saturation of the soil profiles to values above those associated with shallow ground water contamination (e.g., >25%; Table 2). For example, topsoil M3-P and (DPSox)Profile ranged from 14 to 39 mg P kg-1 and from 2 to 10%, respectively, in forests, compared with from 108 to 692 mg kg-1 and 12 to 103% in the adjacent crop land. We also found that topsoil M3-P values were significantly correlated with (DPSox)Profile and with soil WSP concentrations at the 60- to 90-cm depth (Fig. 1c). Topsoil M3-P values >~300 mg P kg-1 were consistently found to have (DPSox)Profile values >25% (r = 0.73***) and if M3-P was >200 mg P kg-1, soil WSP concentrations (60–90 cm) were consistently >1.0 mg kg-1 (equivalent to 0.10 mg WSP L-1, based on the 1:10 soil/solution ratio used to determine WSP). Our results support guidelines in the Netherlands where a critical (DPSox)Profile value of 25% (using an {alpha} value of 0.5) was recommended to prevent ground water P concentrations from exceeding an environmentally unacceptable value of 0.10 mg P L-1 (Breeuswma et al., 1995; van der Zee et al, 1990). A significant linear relationship was also found between soil WSP (60–90 cm) and (DPSox)Profile (y = 0.07x - 0.45; r2 = 0.77***). Based on this relationship a (DPSox)Profile value of 25% would result in a WSP concentration (60–90 cm) of 1.3 mg kg-1 (0.13 mg L-1). To further put this in perspective, UDSTP soil test summaries for1993 through 1999 showed that 17% of the soils tested in Delaware and 21% of those in Sussex County (location of the state's poultry industry and the soils in Fig. 1) had M3-P values >300 mg kg-1 and that 32% (statewide) and 37% (Sussex County) had M3-P values >200 mg kg-1.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 1. Comparison of the forms and distribution of soil P in the profiles of cropped and forested soils from four poultry farms in southern Delaware. (a) Changes in Mehlich 3 P with depth in the soil profile; (b) Changes in degree of P saturation by the oxalate method (DPSox) in the soil profiles; (c) Relationships between Mehlich 3 P in topsoils and DPSProfile and water soluble P (WSP) at the 60- to 90-cm soil depth.

 
Relationships Between Mehlich 3 Phosphorus and Water Soluble Phosphorus, Desorbable Phosphorus, and the Degree of Phosphorus Saturation
If an agronomic soil P test is also to be used for environmental purposes it should, at a minimum, be well correlated with the forms of soil P most susceptible to losses in runoff and leaching and with soil P saturation. For our data, when all 465 soils were considered, M3-P was found to be linearly related to WSP (r2 = 0.68***), FeO-P (r2 = 0.84***), and DPSox (r2 = 0.72***) (Fig. 2) . Separating the data for WSP, FeO-P, and DPSox into topsoils (r2 = 0.53***, 0.86***, 0.72***) and subsoils (r2 = 0.86***, 0.88***, and 0.69***) did not appreciably improve the predictive value of these equations. Based on the regression equations developed from the overall data set (Fig. 2), values of WSP, FeO-P, and DPSox associated with the agronomic critical value for M3-P used by the University of Delaware (50 mg kg-1) were 1.9, 13.6 mg kg-1, and 21.2%. We placed reference lines at these values on the graphs in Fig. 2 and found that the M3-P soil test accurately segregated these three measures of the potential environmental impact of soil P (soluble P, desorbable P, and P saturation) into two categories, those within and above the range where profitable crop response to applied P is likely. Our results showed that only 4, 3, and 2% of the soils had WSP, FeO-P, and DPSox values in the upper left quadrants (low M3-P and high WSP, FeO-P, DPSox) while only 8, 8, and 7% of the soils had WSP, FeO-P, and DPSox values in the lower right quadrants (high M3-P and low WSP, FeO-P, DPSox) of these graphs.



View larger version (32K):
[in this window]
[in a new window]
  		Fig. 2. Relationships between Mehlich 3 P and (a) water soluble P (WSP), (b)Fe oxide strip P (FeO-P), and (c) degree of P saturation via the oxalate method (DPSox) for all 465 soils used in the study. Dashed vertical lines represent the agronomic critical value for Melich 3-P (50 mg kg-1) and dashed horizontal lines represent the corresponding values for WSP, FeO-P, and DPSox, as derived from the linear regression equations in each graph.

 
Comparison of Mehlich 3 and Acid Ammonium Oxalate as Soil Tests for Phosphorus, Aluminum, Iron, and Soil Phosphorus Saturation
Knowledge of the degree of saturation of soils with P, in addition to the amount of plant available, soluble, or desorbable P, is important to long-range nutrient management planning. Not surprisingly, then, soil tests to characterize P saturation are receiving widespread interest (Beauchemin and Simard, 1999; Hooda et al., 2000; Kleinman et al., 1999; Sims et al., 2000). One of our goals in this study was to determine if the M3 soil test could be used as a rapid, inexpensive alternative to the well-established DPSox method for measuring soil P saturation. We had already shown, in earlier research with Delaware topsoils and several Dutch soils, that DPSox was well correlated with soluble P (r = 0.84***), FeO-P (r = 0.84***), and Mehlich 1 soil test P (r = 0.82***) and thus appeared suitable as an environmental soil P test for acidic, coarse-textured, low OM soils of the Mid-Atlantic USA (Pautler and Sims, 2000). For the 465 soils used in the present study, we again found that DPSox was significantly correlated with WSP and FeO-P (r = 0.86*** and 0.84***, data not shown). These results provide further evidence that DPSox can be a reliable indicator, for the Mid-Atlantic USA, of soils with greater potential to release soluble P to runoff or leaching waters and of soils that will be highly enriched in bioavailable P should erosion transport particulate P to surface waters. Unfortunately, as noted earlier, the logistical difficulties (e.g., longer shaking time, different reagents, separate extraction required for agronomic and environmental assessment of the same soil sample) associated with using DPSox as an indicator of soil P saturation have, at least to date, precluded its adoption by soil testing laboratories in the region; thus our interest in using a routine agronomic soil test to rapidly characterize P saturation.

With respect to the suitability of M3 as a P saturation test, as has been advocated by Bolinder et al. (1998) and Khiari et al. (2000), our results showed that, for all 465 soils, M3-P and Al were well correlated with Pox and Alox (r = 0.71*** and 0.85***) but that M3-Fe was only moderately correlated with Feox (r = 0.44***). Linear regression equations between M3 and oxalate P, Al, and Fe were statistically significant and accounted for 51, 71, and 19% of the variability in the relationships between the two extracting methods when all soils were considered (Table 3). Splitting the data set into topsoils and subsoils did not appreciably improve the predictive accuracy of regression equations between M3 and oxalate P, Al, and Fe (Table 3). In fact, the relationship between M3-P and Pox was noticeably worse in subsoils (r2 = 0.28***) than in topsoils (r2 = 0.52***).


View this table:
[in this window]
[in a new window]
  		Table 3. Comparison of Mehlich 3 and acid ammonium oxalate as extractants for P, Al, and Fe in soils of the Mid-Atlantic United States.

 
Observation of the data in Fig. 3c clearly indicated that some soils had very high Feox values and disproportionately low M3-Fe values; these same soils also had much lower M3-P/Pox ratios than others in the data set (Fig. 3a,c). A review of these soil samples showed that most were from Study 3 and had been obtained at sites where biosolids treated with Fe salts had been applied (a common practice in this region) and thus had markedly elevated Feox values relative to the remaining soils. Excluding 21 soils with Feox concentrations >40 mmol kg-1 improved the predictive accuracy of the regression equations between M3 and oxalate extractable P and Fe (Table 3).



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 3. Relationships between Mehlich 3 and acid ammonium oxalate extractable (a) P, (b) Al, and (c) Fe, for all 465 soils used in the study. Correlation coefficients in each graph are based on the entire data set.

 
Despite some of these problems we observed in the relative extractability of P, Al, and Fe, by M3 and oxalate, we found a remarkably good linear relationship (r2 = 0.92***) between the two soil P saturation ratios (M3 [(P/(Al+Fe)] and DPSox; Fig. 4a) when all soils were considered, including those with Feox > 40 mmol kg-1. The fact that the same soils had disproportionately low values for M3-P and Fe and that the high soil Feox concentrations did not affect the relative extractability of M3 and oxalate extractable Al explains this. Based on the regression equation in Fig. 4a, a M3 (P/(Al+Fe) saturation ratio range of 0.10 to 0.15 corresponded to DPSox of 25 to 40%, values used in other countries to identify soils sufficiently saturated with P to be of environmental concern (Chardon and van Faassen, 1999; De Smet et al., 1996). We also observed, as did Khiari et al. (2000), that there was an excellent relationship between the [M3 (P/Al)] ratio and the M3 [(P/(Al+Fe)] ratio (r2 = 0.99***; Fig 4b). Khiari et al. (2000) found that no significant improvement in the relationship between soluble P and the M3 (P/Al) ratio was obtained when M3-Fe was included in the regression analysis. They attributed this to the fact that, on a molar basis, M3-Fe represented only about 12% of M3-(Al+Fe) and thus had little impact on the statistical relationship between soluble P and soil P saturation. For our soils, M3-Fe represented ~7% of M3 (Al+Fe), on a molar basis. However, we feel that it is premature to only consider M3-Al when calculating P saturation for soils in the Mid-Atlantic region, where amorphous Fe is well-known to play an important role in the retention and release of soil P to plants and to runoff and leaching waters. We argue that obtaining more information on the Fe status of soils in areas where P management is important to water quality and crop production is worth the minor costs of simultaneously measuring this element, along with P and Al, in an ICP-AES extract.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 4. Relationships between (a) Mehlich 3 [P/(Al+Fe)] and degree of P saturation via the oxalate method (DPSox) and (b) Mehlich 3 (P/Al) and Mehlich 3 [P/(Al+Fe)] for all 465 soils used in the study.

 
Relationships Between Soil Phosphorus Saturation and Phosphorus Concentrations in Runoff and Leaching
The statistically significant relationships we observed between DPSox and the forms of soil P most susceptible to losses in runoff and by leaching (WSP, FeO-P), in combination with a growing body of research (Anderson and Xia, 2001; Beauchemin and Simard, 1999; Behrendt and Boekhold, 1993; Breeuwsma and Silva, 1992; Chardon and van Faassen, 1999; DeSmet et al., 1996; Hooda, 2000; Khiari et al., 2000; Lookman et al., 1995; McDowell and Sharpley, 2001b; Sims et al., 2000) support the value of DPSox as an environmental assessment tool. Further, the strong statistically significant linear relationships between the M3 [(P/(Al+Fe)] and [M3 (P/Al)] ratios and DPSox (Fig. 4) indicate that the many soil testing laboratories now using the M3 soil test for agronomic purposes could also develop environmental interpretations related to P saturation and the potential for nonpoint source pollution, as discussed in more detail below.

To further evaluate the concept that soils with high degrees of P saturation are of greater risk to water quality, we compared the M3 [(P/(Al+Fe)] ratio with (i) DRP in runoff from soils ranging in soil test P from low to excessive (based on agronomic criteria for crop response) using a rainfall simulation box study method (Study 4), and (ii) DRP in leachate from undisturbed columns, using some of the results of Maguire and Sims (2002) who evaluated the relationship between a number of soil P tests and P leaching (Study 5). The soils used in these two studies were collected from the same fields on the same farms and thus were similar in soil properties, soil test P, and P saturation. They were moderately acidic, low in OM, and were intentionally selected to provide a wide range in soil test P and soil P saturation (Table 4).


View this table:
[in this window]
[in a new window]
  		Table 4. Properties of the soils used in the rainfall simulation runoff box study and the soils used in the undisturbed column leaching study. Data are from Studies 4 and 5.{ddagger}

 
In both the runoff box and column leaching studies we observed distinct change points for the M3 [(P/(Al+Fe)] ratio, values above which DRP in runoff or column leachates increased rapidly (Fig. 5) . The runoff and column leaching data could be fit with statistically significant split-line models (r2 = 0.73*** and 0.87***). These split-line models quantified the change point and described linear relationships with different slopes for the P saturation–DRP relationships below and above the change point. We also found the split-line models using M3 [(P/(Al+Fe)] to be better predictors of DRP in runoff and column leachate than M3-P alone (Fig. 6) .



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 5. Relationships between Mehlich 3 [P/(Al+Fe)] and (a) dissolved reactive P (DRP) in runoff from soils in a rainfall simulation runoff box study (data are all reps of all 17 soils used in the study); and (b) dissolve reactive P (DRP) in leachate from undisturbed soil columns (reproduced from Maguire and Sims, 2002). Note that r2 values apply to the entire split-line model, not individual segments.

 


View larger version (17K):
[in this window]
[in a new window]
  		Fig. 6. Relationships between Mehlich 3 P and (a) dissolved reactive P (DRP) in runoff from soils in a rainfall simulation runoff box study (data are all reps of all 17 soils used in the study); and (b) DRP in leachate from undisturbed soil columns. Note that r2 values apply to the entire split-line model, not individual segments.

 
The M3 [(P/(Al+Fe)] change points for runoff and column DRP were determined to be 0.14 and 0.21. When expressed as M3 (P/Al) the change points were 0.16 and 0.23, similar to the value of 0.15 suggested by Khiari et al. (2000) as an environmental threshold for M3 (P/Al). For M3-P, the change point for column DRP was 242 mg kg-1. For runoff DRP and M3-P the data could not be fit with a split-line model and thus we could not determine a change point; however a significant correlation did exist between runoff DRP and M3-P and DRP values were appreciably higher when M3-P was >~150 mg kg-1 (Fig. 6). Higher values for M3 [(P/(Al+Fe)] change points for leaching compared with runoff likely reflect the greater potential for soluble P to be retained as it percolates through the column, compared with runoff from the surface 5 cm of soil. That is, lower P saturation change point values will be associated with soluble P losses in runoff compared with leaching. We recognize that results from runoff box and column studies must be interpreted with caution. However, our data and other recent research that has identified change points for DRP in runoff, leachate, and tile drain waters versus soil test P and for soil P saturation (Brookes et al., 1997; Heckrath et al., 1995; Hesketh and Brookes, 2000; McDowell and Sharpley, 2001a; McDowell and Sharpley 2001b; Pote et al., 1996) clearly provide support for the concept that as soils become increasingly saturated, a point is reached above which soluble P losses in surface runoff and leachate will become of environmental concern for ground and surface water pollution.

Agri-Environmental Interpretation of the Mehlich 3 Soil Test
There is increasing pressure in the Mid-Atlantic USA, and in many other states and countries, to develop P-based management practices for agriculture that would reduce the potential for nonpoint source P pollution of surface and shallow ground waters. Given this, there is clearly a need to develop a consistent approach to guide the environmental interpretation of soil P tests, such as the M3, as has already been done by those that rely upon soil tests to make agronomic recommendations. What should be the guiding principles for such an interpretation?

First, we should recognize that agronomic soil test criteria, those that identify when crop responses to applications of fertilizers, manures, biosolids, and other P sources are likely to be profitable, are well-established and accepted as accurate by farmers and their advisors. Most soil testing laboratories have developed soil test P critical values (defined as the "..soil test value that produces the best separation between responsive soils from those where the nutrient no longer limits yield"; Beegle et al., 1998) through local research and use them to identify when profitable crop responses to P inputs are unlikely (Beegle and Oravec, 1990; Mallarino, 1997; Sen-Tran et al., 1990). It is also well recognized that there are variations in the critical values and agronomic crop response ranges used for soil test P. These variations are caused by a number of factors, including the crop to be grown, soil type, other soil and nutrient management practices (e.g., tillage, irrigation, subsoil properties, and crop rotation), analytical procedures (e.g., use of colorimetric or ICP methods to determine P concentrations in soil test extracts, where ICP measurements typically result in higher P values; Eliason et al., 2001; Mallarino and Sawyer, 1999), and by differing philosophies of soil test interpretation (e.g., sufficiency level vs. "build-up and maintenance"; Sims; 2000b). Consider the following critical values reported for M3-P by some soil testing programs in the USA (all values in mg kg-1; where values were reported in pounds per acre we converted them to mg kg-1 by dividing by two): Arkansas (50); Delaware (50); Iowa (11–21, varies depending upon crop to be grown and subsoil P level); Kentucky (16); North Carolina (53); New Jersey (36); Oklahoma (32); Pennsylvania (30); Soil and Plant Analysis Council (50) (Heckman, 1998; PSU Agron. Guide, 2000; Sawyer et al., 1999; Sims and Gartley, 1996; SERA-IEG-6, 2001; SPAC, 1999). In general, farmers and their advisors recognize the validity of the agronomic criteria that are appropriate for their region and rarely recommend that fertilizer P be purchased and applied when soil test P exceeds the critical value. From an environmental perspective, following these guidelines should produce the minimum environmental impact from soil P that is possible if we are to attain economically optimum crop yields, assuming that soil conservation practices to reduce soil P losses by erosion are implemented effectively.

Second, the heart of the environmental issue with respect to overfertilization with P and the buildup of soil test P to values above those needed for economically optimum crop yields, is widely recognized to be our inability, to date, to develop alternatives to land application of agricultural and municipal by-product sources of P (manures, biosolids, wastewaters, composts; Lanyon, 2000; Parry, 1998; Sims, 1993; Sims and Coale, 2002; Tunney et al., 1997). While it would be unusual to find situations where farmers purchase and apply fertilizer P to soils where economic crop responses to this input are highly unlikely, they often must apply manure to these soils because no other options are available to them. Until recently, these by-products were mainly applied to meet crop N requirements, which inevitably led to overapplication of P and increases in all forms of soil P (soil test, water soluble, desorbable, soil P saturation; Beegle et al., 2000; Cabrera and Sims, 2000; Klausner, 1994; Sharpley et al., 1994).

Finally, as a result of intensive research efforts for the past decade, we now have enough knowledge about the relationships between soil P tests and potential water quality impacts to develop guidelines that reduce the environmental impact of agronomically unnecessary P additions and high P soils. Beyond this, environmental interpretations for soil P are now being requested by state and federal advisory agencies or mandated by regulatory programs, such as the state nutrient management laws in Delaware, Maryland, and Virginia, the NRCS Code 590 Nutrient Management Standard, and the CAFO regulations recently proposed by USEPA.

Given these considerations, the results of our study and other research, we suggest adoption of the approach recommended by Beegle et al. (1998) for agri-environmental interpretation of the M3-P soil test and M3 [P/(Al+Fe)] (note: M3 [P/(Al+Fe)] values that correspond to suggested M3-P values below were determined using data for all 465 soils as: M3 [P/(Al+Fe)] = (0.0009 x M3-P) + 0.018, r2 = 0.83***). Interpretations are presented for both soil P tests despite the fact that few soil testing laboratories use the M3 [P/(Al+Fe)] test today. However, this may change in the future because of the ease of determining P saturation by this method and if other research supports our conclusion that M3 [P/(Al+Fe)] is a better predictor of P losses in runoff and leaching than M3-P alone (Fig. 5 and 6). We illustrate the approach of Beegle et al. (1998) using the results of our study and the current soil testing criteria of the University of Delaware. We recognize that different numerical criteria for M3-P and M3 [P/(Al+Fe)] may be appropriate for other geographic areas and thus present this as a generalized set of guidelines that should be modified according to regional conditions.

Below Optimum.

As defined by Beegle et al. (1998) this is the soil test category where "The nutrient is considered deficient and will probably limit crop yield. There is a high to moderate probability of an economic yield response to the added nutrient. Recommendations are based on crop response and should build the soil into the optimum range over time". In Delaware, soils with M3P <= 50 mg kg-1 (M3 [P/(Al+Fe)] < 0.06) are considered below optimum. For our data, median values for DPSox, M3 [P/(Al+Fe)], FeO-P, and WSP in the below optimum category were 15%, 0.03, 6 mg kg-1, and 1.0 mg kg-1, respectively (Fig. 7a, 7d, 8a, and 8d) . The frequency distribution for these parameters (Fig. 7 and 8) shows that 89% of these soils had DPSox values less than the proposed environmental upper limit of 25% (Chardon and van Faassen, 1999) and that all samples were below the proposed environmentally excessive M3-P saturation ratio of 0.15 (Khiari et al., 2000). Additions of P to soils in the below optimum category should result in profitable increases in crop yields, hence inputs of fertilizer P or N-based application of by-product P sources (e.g., manures, biosolids) are recommended. Because N-based application of by-product P sources often adds more P than is removed in the harvested portion of the crop, regular soil testing is encouraged to ensure that soil P does not build to values that may are agronomically unnecessary and of environmental concern.



View larger version (35K):
[in this window]
[in a new window]
  		Fig. 7. Frequency distribution of degree of P saturation via the oxalate method (DPSox) (a, b, c) and Mehlich 3 [P/(Al+Fe)] (d, e, f) in the below optimum, optimum, and above optimum Mehlich-3-P ranges. Data are for agricultural topsoils only (n = 335). Median, mean, and standard errors apply to DPSox and Mehlich 3 [P/(Al+Fe)] data in each interpretive category.

 


View larger version (34K):
[in this window]
[in a new window]
  		Fig. 8. Frequency distribution of FeO-P (a, b, c) and water soluble P (WSP) (d, e, f) in the below optimum, optimum, and above optimum Mehlich-3-P ranges. Data are for agricultural topsoils only (n = 335). Median, mean, and standard errors apply to FeO-P and WSP data in each interpretive category.

 
Optimum.

This is the soil test category where "The nutrient is considered adequate and will probably not limit crop growth. There is a low probability of an economic yield response to adding the nutrient. If soils are tested annually no nutrient additions are needed for the current crop. For other than annual soil testing nutrient applications are often recommended to maintain the soil in the optimum range" (Beegle et al., 1998). The optimum soil test category used in Delaware is a M3-P range from 51 to 100 mg kg-1 (M3 [P/(Al+Fe)] = 0.06–0.11). For our data, median values for DPSox, M3 [P/(Al+Fe)], FeO-P, and WSP for soils in the optimum category were 25%, 0.08, 14 mg kg-1, and 3.0 mg kg-1, respectively (Fig. 7a, 7d, 8a, and 8d). The frequency distribution of the values for these four parameters showed that 52% of the soils in the optimum category had DPSox values >25% and that 4% of the samples had M3 [P/(Al+Fe)] > 0.15. Because the probability of a profitable crop response to applications of any form of P is considered to be very low when soil test P is in the optimum range, applications of fertilizer P are rarely recommended. Exceptions can include recommendations for low rates of P in starter fertilizers or small amounts of P for crops with high P requirements (e.g., alfalfa, vegetables). Applications of animal manures and other by-product P sources are usually not recommended because of the difficulty in accurately applying these materials at the low rates of P that might be recommended for some crops in the optimum category.

Above Optimum.

Beegle et al. (1998) defined "above optimum" as the soil test category where "The nutrient is considered more than adequate and will not limit crop yield. There is a very low probability of an economic yield response to adding the nutrient. No nutrient additions are recommended. At very high levels remedial action may be required." In Delaware, soils with M3-P values >100 mg kg-1 (M3 [P/(Al+Fe)] > 0.11) are considered above optimum. Based on our data, soils in this category had median values of 58%, 0.22, 44 mg kg-1, and 7.6 mg kg-1 for DPSox, M3 [P/(Al+Fe)], FeO-P, and WSP respectively (Fig. 7a, 7d, 8a, and 8d). For soils in the above optimum range, 97% had DPSox values >25 and 78% had M3 [P/(Al+Fe)] > 0.15. Thus, based on past research on the relationship between soil P saturation and the potential for P loss to water (Chardon and van Faassen, 1999; Khiari et al., 2000), soils in the above optimum category will be of significant environmental concern, should processes exist to transport particulate and dissolved P to surface and shallow ground waters. A more detailed breakdown of the properties of soils in the above optimum category is provided in Table 5. Clearly, as soils become increasingly enriched with P the potential for the loss of soluble P and particulate P increases as well. In Delaware, because of the low probability of a profitable crop response and the potential for negative impacts on water quality, applications of P from any source are not recommended for soils in the above optimum category.


View this table:
[in this window]
[in a new window]
  		Table 5. Phosphorus characteristics of agricultural topsoils in the above optimum soil test category.{dagger}

 
Environmental.

The complexity of developing environmental interpretations for soil P tests is reflected in the definition for this category proposed by Beegle et al. (1998): "There is the possibility that soils testing above this level may result in environmental degradation. This soil test level is independent of crop response categories and may be above or even below the optimum level based on crop response. This level may vary depending upon several other site specific characteristics (e.g., slope, hydrology)." Such a definition clearly implies that the most effective approach to identify fields requiring improved P-based management practices, where water quality protection is the main concern, would be to include soil test P values in a comprehensive site evaluation process, such as the PSI (Gburek et al., 2000; Lemunyon and Gilbert, 1993). The PSI characterizes the relative risk of P loss to water from a field based on both site characteristics (erosion, runoff, drainage, proximity to water, sensitivity of water body to P) and nutrient management practices (soil test P, application rate, method and timing for fertilizer and organic P sources). However, resources and time will often limit our ability to do this type of P loss assessment on all fields, on a farm, or in a watershed. Thus, at least as an interim step, there is a need (or mandate) to define a soil test P value, usually referred to as an environmental threshold, where the implementation of improved P management practices should be a high priority. At the same time it is important to recognize, as noted by Beegle et al. (1998), that erosion of soils with soil test P values below this threshold can also be of serious concern for water quality and thus erosion control practices should be an integral part of any P-based management plan that is developed. It is also important to consider the impact on water quality of direct losses of P from fertilizers and manures that are applied to the soil surface, such as with manure applications to pastures. In these situations P loss directly from the source may be more important than P loss from the underlying soil.

Selection of any environmental threshold for soil P is very likely to be based on a combination of science and policy and the best professional judgment of those involved in the decision. Given this, it is not our intent to recommend a universal environmental threshold value for M3-P (or for M3 [P/(Al+Fe)]), only to point out some of the factors that should be considered when making such a decision. For example, it could be argued, based on our data, that improved P-management practices should be implemented for all Delaware soils in the above optimum category (M3-P > 100 mg kg-1; M3 [P/(Al+Fe)] > 0.11) because: (i) there is a high probability that these soils will be sufficiently saturated in P to result in increased P losses by runoff and leaching (Fig. 5, 7, and 8); and (ii) there is no agronomic need for P, thus farmers who plan to apply manure or other by-product P sources should receive high priority for technical assistance to develop alternatives to land application. If the above optimum category were to be used as an environmental threshold, we estimate, based on recent soil test summaries of the UDSTP, that 58% of the soils in Delaware and 66% of the soils in Sussex County, Delaware (site of the state's highly concentrated poultry industry) would have M3-P values >100 mg kg-1. Because of the costs required to implement improved P management practices at such a large scale and the lack of field data relating soil P to P losses to water, others have argued for higher environmental thresholds for soil test P or for use of the PSI instead. This issue has been debated, and at least temporarily resolved, in Delaware and Maryland where state nutrient management laws require that some form of soil P testing be used to identify fields where more intensive P management practices will be required. In both states a M3-P value of 150 mg kg-1 (M3 [P/(Al+Fe)] > 0.15) was recently adopted by state nutrient management advisory boards, based on consultations with university scientists, NRCS, state environmental and agricultural agencies, farmers, nutrient consultants, and others, as the criteria to identify fields where some form of P-based management must be implemented. Based on our data soils with M3-P > 150 mg kg-1 clearly merit serious consideration for improved P management as the median values for DPSox, M3 [P/(Al+Fe)], FeO-P, and WSP were 67%, 0.27, 50 mg kg-1, and 9.8 mg kg-1 respectively. All soils in this category had DPSox values >25% and M3 [P/(Al+Fe)] > 0.15 In Maryland, if M3-P is >150 mg kg-1 a PSI assessment must be conducted; based on the results of this assessment management practices are recommended to reduce the potential for P transport (erosion control, buffer strips, etc.) or alter the rate, method, and timing of P applications in fertilizers or organic P sources. In Delaware, if M3-P is >150 mg kg-1, P applications in fertilizers and manures cannot exceed the amount of P removed in the harvested portion of the crop unless a PSI assessment indicates these applications will have minimal impact on water quality. In both states, and throughout the Mid-Atlantic region, research and extension efforts are underway to assess the costs and value of adopting approaches such as these to identify areas where "P-based" management will be required.

Natural Resource Conservation. Finally, we suggest that there is a need to consider identifying upper limits for soil test P where no P should be applied for natural resource conservation reasons. Since the global supply of P for use in fertilizers and animal feeds is finite, applying P to soils that are highly overfertilized with P represents an inefficient use of a valuable natural resource. Consider, for example, soils with M3-P > 200 mg kg-1 (M3 [P/(Al+Fe)] > 0.20). In our study, median values for DPSox, M3 [P/(Al+Fe)], FeO-P, and WSP in this range were 73%, 0.31, 56 mg kg-1, and 12.4 mg kg-1. All samples in this range had DPSox values >25 and 80% were >50% P-saturated; more than 95% of the samples had M3 [P/(Al+Fe)] ratios >0.15 and 58% had ratios >0.30 (data not shown). However, in some situations the potential for P loss from soils such as these by erosion, runoff, or leaching can be very low because site characteristics are not conducive to P transport to surface waters or shallow ground waters. In this case, should applications of any form of P still be avoided to conserve a finite natural resource? While establishing natural resource conservation thresholds for soil P has received little debate or discussion to date, we argue that it should be a component of any holistic approach to improve agricultural P management.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 NOTES
 CONCLUSIONS
REFERENCES
 
The advent of P-based laws, regulations, and guidelines in the Mid-Atlantic USA has intensified the need to identify those agricultural settings that are most at risk for P losses to water. A clear consensus is emerging in the USA today that the most effective means to assess this risk is through a comprehensive approach, such as the PSI, that integrates soil P testing with site properties (e.g., erosion, runoff, drainage), water body characteristics (sensitivity to P inputs, proximity to agricultural fields), and nutrient management practices (properties of the P source used and the rate, method, and timing of application) (Coale, et al., 2002; Gburek et al., 2000; Sims and Leytem, 2001). The use of a soil P test alone, particularly agronomic soil tests, is considered to be less desirable because soil tests do not provide information on the other parameters known to influence P transport and thus the impacts of soil P on water quality. Having said this, most advocates of the PSI approach also recognize that the time and costs of conducting PSI assessments of all agricultural fields in a watershed, state, or region are formidable and will, at best, slow our ability to use the PSI to prioritize the implementation of P-based management practices. Individuals in state and federal regulatory and advisory agencies recognize this and are pressing for simpler, quicker approaches to identify areas where P-based management should be required or recommended in the near future.

We conclude from the results of our study that routine agronomic soil tests, such as the Mehlich 3, can be an effective interim approach to guide environmentally based P recommendations for fertilizers, manures, biosolids, and other P sources. Our data showed that M3-P could accurately segregate soil P into categories of increasing environmental risk based on WSP, FeO-P, and soil P saturation. Higher risks are clearly associated with M3-P values that are above the concentrations needed for economically optimum crop yields. Results of our leaching and runoff studies also showed that the M3 [P/(Al+Fe)] ratio was more accurate than M3-P alone at identifying soils of higher risk for P losses by runoff or leaching. We suggest that future research in this area should evaluate whether the M3 [P/(Al+Fe)] should be used in place of agronomic soil tests in PSI-based risk assessments of the impact of P on water quality.


   NOTES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 NOTES
CONCLUSIONS
 REFERENCES
 
1 Note: the median value for M3-P for these 18 000 samples was estimated from Mehlich 1 (M1) P, the soil test used by the UDSTP during 1993 through 1999 using the linear regression equation M3-P = (2.04 x M1 P) + 7.3, developed by Gartley et al. 2002. Back

Received for publication August 22, 2001.


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