Identifying Fertilizer Requirements

Table of contents

    Important tips for best management

    • Most soil sampling in the Prairies is now conducted by private or industry agronomists who are equipped with modern sampling equipment.  Nevertheless it is important that they follow accepted sampling protocols and sample handling practices to ensure accurate results that will lead to good recommendations.
    • Submitting one composite sample per field can provide a general impression of soil nutrient levels. Take 15-20 sub-samples from the most representative areas of the field (not hill tops, not low spots, avoid saline areas) to provide a more accurate composite. Growers should ensure anyone doing custom sampling for them is aware of any unusual areas of the field that may skew results.
    • Submit samples for soil depths of 0-15 cm (0-6”), 15-30 cm (6-12”) and 30-60 cm (12-24”), or 0-15 cm (0-6”) and 15-60 cm (6-24”). Incremental depths can give a better indication of nutrient stratification. If submitting only one depth stick with 0-15 cm (0-6”).
    • The most accurate sampling time is just prior to seeding. However, this isn’t practical because time is needed to purchase and perhaps place the fertilizer before seeding. Options are: (1) Early spring sampling done after the soil has thawed and is no longer saturated from snow melt. Some sampling can be done by well equipped agronomists prior to frost melt. (2) Late fall sampling once the soil has cooled to at least 5 to 7°C or cooler. This helps reduce nutrient content changes due to microbial activity that may occur prior to seeding next spring.

    Soil Testing for Nutrient Content

    Nutrient content in soil varies over years, between fields and even within fields that appear uniform. Soil sampling and analysis methods were developed to assess the fertility level and to predict crop response to applied fertilizer or manure. Soil testing is not an exact science due to nutrient spatial variability inherent in most fields and the inability to predict growing season weather. Although soil testing is not exact, it can help estimate soil fertility and give reasonable guidelines for profitable fertilizer application.

    Spatial nutrient variability in fields creates problems for soil testing and fertilizer application. The variability makes it difficult to obtain representative soil samples. Using single fertilizer rates across variable fields results in over-fertilized and under-fertilized areas within the field. Although variable rate fertilization is being researched and developed, most fields still are successfully fertilized with a single rate. In addition, since fertilizer response calibrations were developed from research sites with low variability, they may under-predict the optimum fertilizer rate for larger farm fields with more variability.

    Proper soil sampling is necessary for meaningful soil test results. A sampling error in the field is usually much greater than the analytical error in the lab. Ensure soil samples accurately reflect the overall field.

    Collecting one composite sample for the whole field remains a common practice. Accuracy of a composite soil sample increases with the number of sub-samples taken. Accuracy refers to how similar the soil sample value is to the true field average. Is the sample representative? Precision describes how often the same value can be obtained when repeating the procedure (reproducibility). The common recommendation to sample a field 20 times and mix all the samples into one composite produces an accuracy of about ±17% for nitrate (NO3-), assuming 80% precision. This is an acceptable level of accuracy and precision for fertilizer recommendation purposes in most cases.

    In addition to adequate sample numbers, proper soil sampling techniques include:

    • where and how to sample each field (sampling plan)
    • proper equipment
    • proper sampling time
    • correct depth
    • proper sample handling

    Four soil sampling strategies

    Sample individual fields separately. The first step is to assess the field variability and identify representative areas. The type and level of variability can influence the choice of sampling plan. Consider these four basic sampling plans:

    1. Random soil sampling uses a random pattern across a field, targeting the midslopes by generally avoiding unusual or problem areas such as hilltops, potholes and saline areas. Bulk together 20 soil cores into one composite sample then send to a soil test lab. This common method is adequate for smaller, relatively uniform fields.
    2. Zone or topographic sampling involves separating sets of samples based on soil differences, topography or other criteria such as elevation, yield maps, electrical conductivity, aerial images or combinations. Identify the dominant topographic features such as hilltop, midslope and bottom slope, and take 20 core samples for each zone. Bulk sub-samples from each type into one composite sample for each landscape zone. Send several samples to the testing lab. This method can provide more meaningful results for variable fields, but at additional expense and labour. To improve profitability and from this added cost you must be willing and able to apply different fertilizer rates in the separate landscape areas.
    3. Benchmark sampling is based on selecting small areas of the field that are representative of either the whole field or large zones within the field. Benchmark sampling areas are much smaller than the whole field, and samples are taken from the same spot within these benchmark areas. Sampling in the same spot from year to year is assumed to reduce variability of the test results, and allows tracking of changes in residual nutrient levels over time. This information can be used to validate whether fertilizer application rates are generally matching crop use over the crop rotation. If the benchmarks are carefully selected to represent the majority of the field, then good soil test results can be obtained.
    4. Grid or systematic sampling follows an organized grid pattern, perhaps every 0.2 to 2 ha (0.5 to 5 acres). This method can reveal field nutrient variability and allow for variable rate fertilization and precision farming techniques. However, the sampling and analysis costs are high and generally not economic with current field crop prices on the Prairies.

    Proper Soil Sampling Equipment

    Soil sampling to a 60 cm (24”) depth can be done with a probe or auger. Do not use flight or screw sampling augers if samples need to be separated by depth since mixing between layers will occur. If testing for micronutrients, ensure the sampling tool is chrome plated or stainless steel and rust-free. Do not bulk samples in metal pails. Clean plastic pails labeled for location and depth work well. Information sheets, sample bags and shipping boxes are available from soil testing labs and most fertilizer dealers.

    Proper Sampling Date

    The most accurate sampling time is just prior to seeding. However, this isn’t practical because time is needed to purchase and perhaps place the fertilizer before seeding. Sampling is commonly done earlier in spring or in the fall. Hand sampling in the spring is done after the soil has thawed and is no longer saturated from snow melt. Mechanical sampling may proceed with some frost still remaining in the soil.

    Late fall soil sampling should occur after soil has cooled to at least 5-7°C. Cooler soils reduce nutrient content changes due to microbial activity. Sampling at this late fall stage can provide a close reflection of nitrogen levels in the spring prior to planting. Studies have shown that N availability can increase over the winter; however most of that gain is lost in the spring. Therefore, the net change from late fall to late spring was minimal —as long as soils stay frozen. In the southern Prairies prone to warm Chinook winds in the winter, significant changes can occur in mild winters.

    Late fall sampling tends to more accurately reflect spring nitrate (NO3-) contents than early fall sampling, especially for Black soils. Alberta research in the 1980s compared soil samples taken in the fall (early October and early November) to spring samples for 26 stubble fields. Early fall samples averaged 34 kg/ha (30 lb./ac.) less nitrate N than spring samples, while late fall samples averaged only about 17 kg/ha (15 lb./ac.) less. The early fall samples were also more variable in relation to spring samples. Overall, late fall samples more accurately predicted spring nitrate contents and grain yields than early fall. Spring samples were slightly better than late fall samples for predicting grain yield and N uptake.

    In contrast, research in Manitoba measured very little change in soil nitrate levels in cereal stubble from early September to freeze-up. North and South Dakota extension soil scientistsrecommend early fall sampling in view of time constraints, but suggest to reduce the N recommendation by 0.2 kg (0.5 lb) for each sampling day prior to September 15.

    Another experiment in central Alberta compared the effects of sample timing on phosphorus (P) soil test values. On average over 27 sites, the extractable P increased from 28 kg/ha (25 lb./ac.) in early October to 49 kg/ha (44 lb./ac.) by early November, and to 50 kg/ha (45 lb./ac.) by late April to early May. The relationship between early fall and spring P values was not close and, therefore, it was not possible to simply correct the early fall values. In contrast, research on a Brown soil in Saskatchewan over 24 years found both overwinter increases and decreases in soil P tests, but relatively few were significant. An experiment on irrigated alfalfa on a Dark Brown soil near Lethbridge, Alberta found significant overwinter increases in organic P. The conflicting results may be related to the differences in soil organic matter content and biological activity, and, therefore, potential for microbial changes to the plant available P pool.

    Occasionally early fall sampling can create higher than necessary fertilizer N and P recommendations due to an underestimation of spring nitrate N and available P, especially in the Black and Gray soil zones. Fertilizer response curves have been calibrated only against spring nutrient contents.

    One disadvantage to late fall sampling after soil has cooled to 5-7 C is that fall fertilizer banding opportunities become more limited. By the time the samples are taken, dried, sent to the lab, analyzed and results returned, the soil may have become frozen or covered with snow. On average over the Prairies, soil cools by 1 C every 5 days in the fall.

    However, sampling immediately after combining would be the least accurate for measuring nutrient needs for next spring’s crop, and is not recommended for this purpose. This method may be able to show what nutrients were historically deficient, if any, for the current year’s crop, but changes in nitrogen levels may occur after early fall sampling if rainfall encourages mineralization or if nitrogen is lost to leaching, denitrification or immobilization.

    One key advantage to sampling immediately after harvest is that proper depths of sampling are maintained. For growers who use tillage, sampling on tilled fields is more difficult since the base soil level has been disrupted with soil loosening.

    Correct Sampling Depth

    The common current recommendation is to provide a 0-15 cm (0-6”) depth sample, and also provide a 15-60 cm (6-24”) sample or separate 15-30 cm (6-12”) and 30-60 cm (12-24”) samples. With most sampling done by commercial agronomists, the grower can ask for the right depth and method.

    For immobile nutrients such as phosphorus (P) and potassium (K) and most micronutrients, the ideal depth is 0-15 cm (0-6”) because fertilizer response calibrations were developed based on that depth. If only the 0-30 cm (0-12”) depth is sampled, the soil testing labs must use a correction factor to estimate the value for the 0-15 cm (0-6”) depth. Since these nutrients are relatively immobile, they tend to remain at the fertilizer application depth. Therefore, the 0-30 cm (0-12”) depth may underestimate the concentration of these nutrients and lead to high fertilizer recommendations.

    Proper Sample Handling

    Handle samples carefully to prevent accidental mixing and contamination. Put samples in the fridge overnight if it needs to be cool and send as soon as possible to the lab. Many send by courier.

    Consistency of Soil Test Lab Recommendations

    Soil sampling, analysis and interpretation is not an exact science. However, reasonable precision and accuracy is needed in order to make efficiently manage fertilizer inputs and maximize returns on that investment. Over the years, various growers, agencies or companies have sent duplicate samples to different labs to compare their analysis and recommendations. Growers may be frustrated by this because values will certainly vary, as will recommendations. This is why it is so important to track soil test levels from year to year, and stick with the same lab. 

    Ask for actual soil residual nutrient measurements as well as the lab’s recommendations. If recommendations are not in line with fertilizer rates you’d use based on actual soil residual levels, ask the lab how it calculates its recommendations. This will help to better understand how your specific production practices might lead to differences in recommendations.

    Two fundamental reasons contribute to differences in results. First, labs may be using different analytical procedures to measure soil nutrient content. For example, there are several methods to extract soil phosphorus. Or the technique may differ slightly when using a particular method.

    Variations in soil test recommendations arise mainly due to the differences in each lab’s interpretation. The recommended rates of fertilizer application for the same soil test level can vary significantly from lab to lab. This may be due to using:

    • different critical (deficient) soil analysis levels
    • regional fertilizer response (calibration) data but modifying recommendations to fit a particular strategy for fertilizer use or economic payback
    • recommendations from other regions or countries
    • a unique system of fertilizer recommendation not based on regional calibration data or economics
    • systems employing short-term economic response versus long-term maintenance of soil test level.

    Look for labs that participate in the voluntary North American Proficiency Testing Program (NAPT), which circulates soil samples of known nutrient content. Labs test the circulated sample and then are informed as to their performance.  Participating labs are listed on line

    Overall, soil testing is a useful agronomic practice. Use labs that base fertilizer recommendations on economics using regional calibration data. Be prepared to question unusual recommendations based on experience and the local knowledge of qualified agronomists. Keep in mind that the accuracy of fertilizer recommendations will always be limited by sampling challenges and the inability to predict the weather of the upcoming growing season.

    Soil Analysis Methods

    Make sure the soil analysis lab is using a nutrient extraction method that best suits the soil type. Phosphorus (P) analysis methods are a good example of how methods vary from lab to lab. Prairie labs use the Modified Kelowna or Olsen methods for phosphorus (and in the case of the Modified Kelowna method, for nitrate-nitrogen and potassium as well.) Some labs outside the Prairies use the Bray method, an acid extractant method that is not suitable for alkaline soils and has not been calibrated for western Canadian soils.

    Here are methods used across the Prairies. Note, this list does not include “root simulator technology,” an alternative method for determining nutrient deficiencies.

    • Nitrogen (Nitrate-N): Alberta labs typically use Modified Kelowna (acetic flouride). Others use the calcium chloride test.
    • Phosphorus: Modified Kelowna (acetic flouride) is common in Alberta and Saskatchewan. Olsen (sodium bicarbonate) is well suited to alkaline soils and is the common method in Manitoba.
    • Potassium: Modified Kelowna (acetic flouride) or ammonium acetate methods are the two in common use on the Prairies.
    • Sulphur (Sulphate-S): Calcium chloride method is standard.
    • Copper, iron, manganese, and zinc: Diethylene triamine pentaacetic acid (DTPA) method is standard.
    • Boron: Hot water extraction is standard.

    Plant and Tissue Testing

    Plant and tissue analyses can supplement, but not replace, soil testing. Tissue sampling can be used to diagnose crop problems that may be nutritionally related and to identify any nutrients that may be limiting yields, but it does have serious limitations. Tissue nutrient levels vary significantly depending on stage of plant growth, plant parts sampled, and the time when samples are collected (e.g. time of day, timing relative to environmental stresses).

    No reliable interpretative criteria exist for nutrient ranges in seedling canola. And because nutrient contents usually differ greatly between different plant parts and ages, the proper part must be sampled at the proper growth stage.

    Each tissue testing lab will have slightly different requirements for taking and submitting samples. In general, avoid unusual, dead or stressed plants, as well as those covered with soil or recent sprays. Cut samples with a clean, rust-free knife or scissors.

    Send separate samples from good and poor areas within a field. Make sure the plants in each area are at the same growth stage. Comparing the plant analysis results from areas of a field that differ visibly in growth can be difficult to interpret because nutrient content differences can be confounded by growth differences. If the two areas differ mainly in deficiency symptoms, then comparative sampling can be useful. In this case, collect the samples soon after the symptoms appear and before major differences in growth and maturity occur.

    Plant and tissue analyses measure the nutrient content of above ground plant parts during growth. The values are compared to established ranges for inadequate, adequate and excess levels.  The following table shows sufficiency levels for most plant nutrients in flowering canola.

    Table 3. Plant Tissue Analysis Interpretative Criteria for Canola (whole above ground plant at flowering)
    NutrientSufficiency Level
    Nitrogen (N) % > 2.4
    Phosphorus (P) % > 0.24
    Potassium (K) % > 1.4
    Sulphur (S) % > 0.24
    Calcium (Ca) % > 0.49
    Magnesium (Mg) % > 0.19
    Zinc (Zn) ppm > 14
    Copper (Cu) ppm > 2.6
    Iron (Fe) ppm > 19
    Manganese (Mn) ppm > 14
    Boron (B) ppm > 29
    Molybdenum (Mo) ppm > 0.02

    How to Diagnose Nutrient Deficiency Symptoms

    A systematic diagnosis of visible symptoms is needed to correctly identify limiting nutrients. Symptoms usually appear on either old or young leaves depending on the mobility of the nutrient in question. Chlorosis (yellowing, loss of green colour) and necrosis (death of plant tissue, often leading to white or brown colour) are important visible symptoms. Diagnosis under field situations can be complicated by high field variability, multiple deficiencies, and other causes such as weather, pests and herbicide injury. For example, a sulphur deficiency can easily be confused with Group 2 herbicide injury due to similar symptoms.

    Proper diagnosis of nutrient deficiency should use most of the following tools:

    1. Soil test. Are any obvious shortages evident? Other soil test parameters such as texture, pH and electrical conductivity may also provide clues in the diagnosis.
    2. Fertilizer history. In past years, what rates and source products have been applied on that field? Include crop yields and consider whether rates have been adequate to match removal.
    3. Tissue test. Do not use this alone. A tissue test may show that the plant is deficient in phosphorus or calcium, for example, but heavy rains and saturated soils may have stopped nutrient uptake. The nutrients may be in the soil at adequate rates, but the plant simply can’t access them. So even if the tissue test shows deficiency, a rescue application of these products will not help. Plants may recover on their own when soils dry out again.
    4. Herbicide history. Is there any chance of carryover from products applied one or two years ago? In dry conditions or very wet conditions, herbicides can take longer than expected to break down to safe recropping levels for canola.
    5. Environment. Cold, wet, hot and dry can all stress canola plants, creating symptoms that may look like nutrient deficiencies. If neighbors have similar symptoms, the cause is probably environmental — frost, excess moisture, etc.
    6. History of the land. Recently broken forage land is likely to be depleted in a lot of nutrients. Canola seeded into a field that was in long-term alfalfa, for example, is one case where you may see severe crop stunting and delayed maturity as a result of phosphorus deficiency. Otherwise it is rare to see severe phosphorus deficiency in canola, but it can occur with high yields and tight rotations that will start to mine the soil of its reserves.
    7. Look at other fields for similar symptoms. When diagnosing for a specific nutrient, target the crop that tends to be most sensitive to that nutrient. If your farm is depleted of copper for example, this deficiency is likely to show up in wheat before any other crop.

    Using optical sensors (e.g. Greenseeker) is another method to measure and correct nitrogen deficiencies in a field. These optical sensors estimate crop biomass, yield potential and crop nitrogen requirements. The sensors can be attached to a nitrogen applicator and provide continuous and instantaneous readings for variable rate applications.


    Malhi, S.S., and M. Nyborg, 1986, “Increase in mineral N in soils during winter and loss of mineral N during early spring in North Central Alberta,” Canadian Journal of Soil Science 66(3): 397-409. Ten field experiments were conducted on cultivated soils in north-central Alberta to determine any change in mineral N content of soils during winter, and during early spring after the soils had thawed. Soil samples were taken periodically from fall to spring to a depth of 120 (or 90) cm and were analyzed for NH4-N and for NO3-N. Mineral N changes occurred primarily in the top 60 cm. Between fall and late winter, there was an increase of 48 kg N ha−1 of mineral N (range of 27–83) in the 60-cm depth of eight experiments set on stubble and the value increased only to 55 kg N ha−1 when the sampling depth was extended to 120 (or 90) cm. There was a loss of mineral N during early spring of 44 kg N ha−1 (range of 18–71). The two experiments on summerfallow had more over-winter accumulation of mineral N and more loss in early spring compared to the stubble experiments.

    Dormaar, J.F., 1972  “Seasonal pattern of soil organic phosphorus” Can. J. Soil Sci. 52: 107-112

    Ross McKenzie, researcher with AARD in Lethbridge. Personal communication.

    This is based on information from Ross McKenzie, researcher with AARD in Lethbridge, Alberta; John Heard, soil fertility specialist with MAFRI in Carman, Manitoba; and Dianne Kamppainnen with ALS Environmental soil test lab in Saskatoon. Manitoba’s methods are printed in the Manitoba Soil Fertility Guide: Detailed explanations of techniques are outlined in the NCR-13 handbook: