Agronomic Thresholds for Soil Phosphorus in Alberta: Basis for Fertilizer Recommendations in Alberta

Soil test P results ultimately form the basis for crop response and fertilizer recommendations. Agricultural soil analysis techniques have been primarily used to identify total P (Pt ) or plant-available P, which is correlated with soil test P (STP). Crop fertility status and fertilizer requirements are based on STP; therefore, the test is widely used in the industry for routine soil analysis. Organic P (Po) forms are important when assessing soil P transformations, which can be applied to fertility, especially when considering manure. Other P fractions, including biologically available P (BAP), dissolved P (Pd) and particulate P (Pp), have become important factors when assessing the environmental impacts of soil P contributions; however, they have not generally been part of routine soil analysis.

Soil test P methods are based on using chemicals to dissolve enough P from the soil to estimate the amount of P the soil is capable of supplying to the crop. The main purpose of a soil P test is to determine an index of the amount of supplemental P required to prevent loss of crop value because of P deficiency (Fixen and Grove 1990). Plant uptake results in a decrease in solution P, which is replenished by dissolution of mineral P or release of adsorbed P. Soil conditions, especially pH and P sorption capacity, influence this process. The STP analysis uses chemical extractants to create solutions that dissolve P in amounts consistent enough to be correlated with crop P use. The amount of dissolved P can change for different soils and different pH levels.

Soil tests for plant-available P

Analysis of soil to detect P levels follows the general method of dissolving some or all of the soil P in an extractant, then determining the concentration of P in the extractant. A typical extraction procedure involves the shaking of a known weight of soil (e.g. 5 g) with a known volume of extractant (e.g. 50 mL) for a fixed period of time (e.g. 30 min.) at room temperature. The mixture is then filtered, and the filtrate is analyzed for extractable P (PO4-3) content. Measurement of plant-available P in filtered extractant is determined using an inductively coupled plasma (ICP) spectrophotometer, or colorimetrically using phosphomolybdic blue complex reduced with ascorbic acid (Watanabe and Olsen 1965). The ICP probe will respond to other forms of P as well as PO4-3, and can give some slightly higher results for some extracts (Miller and Kotubny-Amacher, in press).

The purpose of the test and the type of soils being tested determine the type of extraction and the method of soil sampling. For STP, soil tests are based on a cored or augered sample to 0.15 m depth. The results of the analysis are then correlated to crop growth or nutrient uptake. In cases where a particular soil P test does not correlate to crop response, either by research or field experience, it can eventually be replaced by a test in which the users have more confidence. For example, the Miller-Axley test, which has been used in the Alberta provincial laboratory for several years, has recently been reported as underestimating soil P availability (McKenzie et al. 1998). Most STP analyses in Alberta are now performed using modifications of the Kelowna test.
Bray and Kurtz (1945) and Olsen et al. (1954) used weak extractants, which correlated well with plant uptake. The Bray and Kurtz (1945) extraction method, referred to as Bray-1, initially used 0.03N NH4F and 0.025N HCl for determining "adsorbed-P". Fluoride was included in the extractant because it had a high capability to displace adsorbed P. Surface adsorbed P has a greater influence on P availability because its surface area is exposed to the chemical reactions, therefore, it more readily replaces the P in soil solution as solution P is removed by plant uptake (Thomas and Peaslee 1973). Since adsorbed and acid-soluble forms of P co-exist in soils, Bray and Kurtz (1945) proposed a stronger extractant using 0.1 N HCl instead of 0.025N HCl for determining both forms of P.

Relationships between crop response and the Bray-extractable P were not consistent across several soil types. Free carbonates in soils neutralize the acid in the extractant resulting in precipitation of dissolved P (Syers et al. 1987). Soils at pH values greater than 7.2 frequently contain enough free carbonates to at least partially neutralize the HCl in the Bray extractant (van Lierop 1988). Numerous extraction methods have been developed, all with advantages and limitations depending upon the soil type for which they have targeted. Their suitability relies on long-term correlation studies that have established relationships between the extraction method, the soil and crop response. Therefore some extraction methods are favored more in some parts of the world than in others. Soil pH, mineralogy and organic matter content are major factors in accuracy. The tests are differentiated by the extractant used. A summary of the more common tests is presented in Table 2.1.

In Alberta, the Miller-Axley and the Bray were the main extractions for available P, but both were unsuitable for the highly calcareous, high pH soils of southern Alberta because the free carbonates neutralized the weak acid in the extractant (McKenzie and Kryzanowski 1990). During the 1970s and 1980s, a second extraction using the Olsen (NaHCO3) test was performed in Alberta and British Columbia provincial soil laboratories to determine the available P in high pH soils (van Lierop 1988; McKenzie and Kryzanowski 1990). The Miller-Axley test was developed in the 1970s, before large amounts of commercial fertilizer P were added to soils. At that time, the chief sources of P were from soil minerals and organic sources (McKenzie et al. 1998).

The Kelowna extraction was proposed by van Lierop (1988) as a single test that would apply to a wide range of soil conditions. It could replace the Bray and Olsen tests. The extractant used acetic acid in place of the weak HCl used in the Bray and H2SO4 used in the Miller-Axley extractants. While the Kelowna extraction performs similar to the Olsen test at high soil pH levels, it does not require charcoal filtration and does not evolve CO2 (Qian et al. 1994).

In the Canadian prairies, modifications of the Kelowna test were tested against crop response in greenhouse trials (Ashworth and Mrazek 1989; McKenzie et al. 1989; Qian et al. 1994) and field trials (Ashworth and Mrazek 1989, 1995). The testing validated the Kelowna and modifications of the Kelowna for a variety of soil types. Addition of ammonium acetate to the Modified Kelowna extractant resulted in an improved prediction of available potassium (K) compared to the Kelowna (Qian et al. 1994). The Kelowna and its modifications are considered accurate for Canadian prairie soils (Havlin et al. 1999).

Table 1 : Summary of P analysis methods (modified from Sibbesen and Sharpley 1997, and Howard et al. 1999).

Analysis method	Extractant	Comments
Olsen	0.5M NaHCO3 @ pH 8.5	-best suited for neutral and calcareous soils (Qian et al. 1994) -process of maintaining pH level, driving off CO2, and filtering extractant through activated charcoal makes the procedure awkward (Qian et al. 1994)
Mehlich-3	0.2M CH3COOH 0.25M NH4NO3 0.015M NH4F 0.013M HNO3 0.001M EDTAz	-common method for assessing crop- available P in U.S.
Bray-1	0.03N NH4F 0.025N HCl @ pH 3.5	-designed for neutral - acidic soils -not suited for calcareous soils
Bray (strong)	0.03N NH4F 0.1N HCl @ pH 1.0
Miller-Axley	0.03N NH4F 0.03N H2SO4	-not suited for calcareous soils
Modified Kelownay	0.015M NH4F 0.25M ammonium acetate 0.25M acetic acid	-best method for a wide range of soil pH levels in the prairie provinces -measures available P and K
Modified Kelownax	0.015M NH4F 1.0M ammonium acetate 0.5M acetic acid	-best method for a wide range of soil pH levels in the prairie provinces -measures available P and K
Kelowna	0.015M NH4F 0.25M acetic acid	-suitable for a wide range of soil pH levels

Extraction method (y)	Regression equation	r2
Enviro-Test Modified Kelownaz	y = 4.038 + 0.941x	0.88
Miller-Axley	y = 3.258 + 0.717x	0.57
Olsen	y = -3.319 + 0.886x	0.76
Kelowna	y = 1.915 + 1.137x	0.86
Resin	y = -16.23 + 1.938x	0.51
Bray-1y	y = -4.08 + 1.103x
Mehlich-3x	y = 13.004 + 1.037x	0.97

• adding enough P to build the soil to the desired level for the most economic production for several years (e.g. a rate of 60 kg ha-1 for a corn crop on a soil with low P levels).
• adding enough P to give the maximum net return on the immediate crop, neglecting residual benefits (e.g. a rate of 30 kg ha-1 for a corn crop on a soil with low P levels).
• fertilizing for a period of 3 to 4 years at a uniform annual rate that will give the highest net return for the period (e.g. an annual rate of 20 kg ha-1 for the period, for corn crop on a soil with low P levels).
• fertilizing for a period of 3 to 6 years by calculating the total requirement for the period and then applying the P fertilizer at the most effective times to realize the greatest net return for the period (e.g. a rate of 45 kg ha-1 initially for a corn crop on a soil with low P levels, followed by perhaps one other application during the period).

• higher crop yields than were obtained in the original calibration studies,
• increased use of minimum tillage,
• increased use of equipment and practices that result in increased soil compaction,
• earlier planting dates into cooler soils, and
• inadequate attention to proper soil pH levels for optimum nutrient efficiency.

Wheat response to starter P fertilizer with increasing soil test P levels in a fallow -wheat - wheat rotation at Swift Current, Saskatchewan (Roberts et al. 1999).

Relationship between plant-available P and water quality risk

Optimum crop production requires additional input of soil P, and a small buildup of soil P is desirable. Since applications of starter fertilizer can even increase soil P levels with time, care must be taken to ensure that levels do not become high enough to be an unacceptable risk to water quality. Farm managers will be expected to take greater responsibility for maintaining soil P levels for good crop production without unacceptable risk to water quality. However, do farmers have applicable means for assessing risk on their farms?

Ekert et al. (2000) presented criteria which farmers can use to assess and quantify environmental risk. The criteria are based on using the production potential of the land most efficiently, but at the same time, keeping the impacts on soil, water, biota and air within tolerable limits. These criteria include establishing scientifically-acceptable tolerability ranges for risk, development of convincing and feasible actions and goals, and are relatively simple, reproducible and economically acceptable. The concept is holistic and points out that P should not be looked at in isolation of the other fertility factors.

The potential for soil P to impact water quality has been extensively reviewed by Howard et al. (1999). Methods to assess soil P risk, including the Soil P Index, the single limit approach, use of percent saturation, and nutrient management planning were all discussed and evaluated as potential tools for managing P to minimize the risk of degrading water quality. However, finding a practical means to measure soil P and relate its risk to water quality has yet to be agreed upon in the literature.

Because STP is widely used in crop management, a considerable knowledge base exists in understanding how the chemical extraction processes relate to P in different soils. Commercial labs are also set up to perform routine STP analysis at a relatively low cost. As we begin to address environmental issues, will STP be a suitable tool for identifying risk? If producers are to monitor their soils for environmental risk, it would be practical to have environmental soil P levels determined by STP methods, since producers would require only one set of soil samples and one type of analysis. Versions of the Soil P Index use STP as the measurement for soil P in assessing the vulnerability of a site to contribute P to the water system (Sharpley et al. 1999; Hilborn and Stone 2000).

While it may be practical and convenient to use STP as a test for water quality risk, care must be taken to ensure other factors affecting the potential for P transport in surface runoff are considered. First of all, P transport sensitivity requires a different approach to sample collection than that required for soil fertility analysis. For most agricultural soils, samples collected to a depth of 20 to 40 mm would accurately define the effective depth for interaction between surface soil and runoff for rainfall intensities (Coale 1998). Second, some STP tests, such as the Mehlich-3 test which uses strong acids, remove more P forms than just the readily available P fraction (Self-Davis et al. 1998). These tests would, therefore, tend to overestimate the dissolved P (Pd) fraction in runoff. Because of the wide variety of STP tests, care must be taken to ensure any regional environmental guidelines or standards provide correlation for several soil tests that may be used, or have been used, in the region. The third consideration is that environmental risk is not strictly related to dissolved P. While dissolved P is an important water quality parameter, there is a considerable amount of particulate P in runoff that would be undetected (Daniel et al. 1998). This fraction can potentially become available for aquatic plant growth.

Sharpley et al. (1996) suggested that for an environmental test, distilled water or a dilute salt solution (e.g. 0.01M CaCl2) would be better suited to assess the Pd fraction. They stated that such methods would need further evaluation before they could be recommended as a standard procedure.

Other Documents in the Series