Introduction

Site specific management accepts that variability occurs within fields. Understanding the variability allows fields to be divided into relatively uniform units, which can be managed using techniques such as variable-rate fertilization or spraying. Soil sampling for nutrient levels is a critical step, both in identifying the variability within the field and in providing appropriate fertilizer recommendations.

In most instances, new soil sampling strategies are needed for maximizing site-specific management techniques. Many of these new methods rely upon the Global Positioning System (GPS) technology that allows targeting of initial soil sampling and the ability to return to the same spots in the future for fertilizer application, scouting or re-sampling. Differential correction is essential for GPS sampling in order to maintain an accuracy of 1 to 3 meters, since the uncorrected GPS signal is accurate only to 100 meters.

Traditional Composite Sampling

This soil testing method has generally served agriculture well over the past several decades. It involves randomly taking cores throughout a field, bulking them, then thoroughly mixing and sub-sampling for submission as a single sample.

Following are the basic steps in traditional whole-field sampling:

• Evaluate the field to determine representative areas for sampling. Level fields appear relatively easy to sample. For hilly fields with knolls, slopes or depressions, take samples from mid-slope positions to get "average" results.
• Avoid sampling obvious sources of unusual variability, such as saline areas, eroded knolls, old manure piles, burnpiles, haystacks, corrals, fencerows or old farmstead sites, on headlands, within 50 feet of field borders or shelterbelts, within 150 feet of built-up roads.
• Sample at 0 to 6 and to 24-inch depths and keep depths separate. Samples for pH, P, K, organic matter and micronutrients should be analyzed from the 0 to 6" depth. N and S should be analyzed for both depths as these mobile nutrients may be deficient in the surface 0 to 6" yet adequate in the 6-24" depth.
• A composite soil sample should include at least 15 to 20 sample sites per field, with a minimum of one sampling site for every 2.5 to 4 acres.

With the use of GPS, traditional random sampling can be located at the same co-ordinates each year to reduce year-to-year sampling variability.

Limitations:

• Does not provide any indication of field variability.
• Small areas of very high nutrient levels that are probed and included in composite samples may cause the "average" reported value to be artificially high, resulting in much of the field being underfertilized. This frequently occurs with the nutrient sulphur, but may occur with other nutrients also.
• This system does not offer the potential for variable-rate application.

In most cases whole fields have been represented by a single soil sample. However, university, government and industry soil-sampling guidelines have indicated for some time that "major areas within fields having distinctly different soil properties such as texture should be sampled and fertilized as separate fields because of differences in nutrient requirements that will occur." In the past this was rarely done, since farmers had limited options for "variable rate fertilization." Some of the newer soil sampling techniques employ this guideline in their sampling plan.

Benchmark Soil Sampling

The basic principle of benchmark sampling is continued sampling at the same location from year to year. Each benchmark is an area of approximately 1/4 acre that is chosen as typifying the field or a majority soil type within the field. In this benchmark area, 15 - 20 samples are randomly collected and composited. This technique assumes that the benchmark area is less variable than the entire field, because it is much smaller. Year after year that same benchmark location and method is used, which should minimize sampling errors. It is treated as a reference area from which all fertilizer recommendations for that field are based.

More than one benchmark site per field may be chosen if complex soil types or variable landscape occurs, or if variable-rate fertilization is an option.

The critical part of this method is in benchmark site selection. Select representative sites by close observation of the crop, particularly during early growth stages when fertility differences are most evident, past grower experience, yield maps, soil surveys, and remote sensed images.

Features:

• Less expensive and time-consuming than grid soil sampling.
• Year-to-year variations better reflect actual nutrient changes.
• May provide information for variable-rate application when different benchmark sites are selected to represent different areas of the field.

Limitations:

• Does not provide a full indication of field variability, but assumes that the rest of the field will behave similarly to the benchmark area.

Grid Soil Sampling

The grid sampling system uses a systematic method of soil sampling to reveal fertility patterns and assumes there is no logical reason for fertility patterns to vary within a field.

The first step in grid sampling is to divide the field into small areas or blocks. The second step has been to identify a sample location within the grid, usually the point at the centre of the grid cell, which is referred to as grid point sampling. Each sample represents a composite of some 3 to 10 sample cores taken in a 10 to 20-foot radius of the centre location of that grid. However, grid point sampling may cause bias because of the regular row and column sample alignment. Other regularly spaced patterns, such as tillage, drainage tiles and ditches or fertilizer spreading may cause a repeating pattern that if aligned with the sample rows, will seriously bias results. Modifications to the sampling pattern, such as staggering of sample points or randomized placement within the grid, may be used to overcome this concern.

 

The greater the sampling intensity, the greater the likelihood of identifying fertility patterns using grid sampling. Some U.S. states recommend a sampling density of one sample per acre in order to obtain representative soil phosphorus, potassium and pH data. Sampling at wider sample spacing may still provide useful information on the magnitude of field variability, but may be too inaccurate for variable-rate management.

Features:

• Grid sampling is well integrated into commercial GPS-based soil sampling and nutrient-mapping GIS programs.

Limitations:

• The dense grid sampling required to effectively reveal fertility patterns can be quite expensive, especially for the lower-value grain and oilseed crops grown in the prairies.
• There is no soil-landscape rationale for grid size. In fields with complex landscapes, there is a risk of missing some soil units with a large grid size, and commercial grid spacing is often too large.

A field may be a candidate for grid sampling if:

• the field history is unknown.
• the field has a history of manure.
• smaller fields have been merged into larger ones.
• high rates of fertilizers/limestone have recently been applied.
• it is important to consider non-mobile nutrients, such as phosphorus and certain micronutrients.

Landscape Directed Soil Sampling

Landscape-directed sampling is based on spatial patterns defined by some prior knowledge or observation about a field and assumes that fertility patterns exist for logical reasons. For example, soil survey information, detailed elevation mapping, aerial photographs, satellite imagery, combine yield monitor maps or grower experience may indicate a pattern in soil and crop variation. This type of variation is called systematic variation because it follows a system or pattern and is predictable and manageable if that pattern is understood. The recent use of yield monitors in Manitoba has shown that yield variation is often related to topography, although this variation may be due more to differences in drainage or available water and weed pressure than nutrient variation.

Soil development and productivity are largely a function of water flow, which is in turn controlled by landscape properties such as slope gradient and length, slope curvature and relative elevation. The combination of these factors determines the location of soil types in the landscape and their inherent productivity. Nutrient levels, particularly the mobile nutrients nitrogen and sulphur, have shown consistent relationships to landscape or topography in recent studies. Soil-moisture relationships largely control nitrogen availability in the landscape through processes of denitrification, leaching, mineralization of organic matter and crop uptake. Elevation measurements may be used to initially develop topographic management zones, but it is actually landscape structure or slope position that influences nutrient relationships.

This system requires the identification of areas (polygons) with similar soil and hydrologic conditions. Properly identified, there will be less variability within each polygon than between polygons.

Research has not firmly established the required sampling density or pattern for landscape sampling. If landscape units were totally homogeneous, one sample would characterize the entire unit, but in reality these units are not homogeneous. Options are to take several (3 to 6) point samples per landscape unit or take a composite sample of 10 to 20 cores for each area. The boundaries of management zones for topography-based sampling may be delineated with the aid of elevation maps, yield maps, or remotely sensed images. Boundaries may be adjusted with further data and experience with the system.

Features:

• Has potential to reduce the number of soil samples required, compared with intensive grid sampling.
• Nutrient distribution and management-unit boundary designation is often superior to grid sampling, especially for nitrogen.

Limitations:

• Requires previous knowledge of crop performance within the field and an ability to discern slight topographic features and soil changes within the field.
• Crop growth and yield relationships with topography may be completely reversed in years of extreme wetness versus years of extreme dryness.
• For fields with subtle changes in topography, a digital elevation map may be needed to select sample sites. Such elevation mapping is available as a commercial service in Manitoba.
• Past management, such as heavy fertilization or manuring, may mask the landscape-nutrient relationships and reduce the usefulness of this method.

Research and field experiences suggest a field may be a candidate for topographic sampling if:

• Remote sensing or yield mapping reveals some relationship between crop growth and landscape properties.
• The field does not have a history of manure application.
• Relatively low rates of nutrients have been applied.
• It's important to know mobile nutrient patterns, especially nitrate-N.

Fertilizer Recommendations

Traditional fertilizer recommendations are based on regionally developed response curves. These response curves served well when managing fields as a single management unit. But ongoing research clearly demonstrates that fertilizer response may vary across landscapes within the same field, even when soil nutrient levels are the same. These traditional response curves may be too simplistic for a site-specific system. Assigning target yields and expected moisture availability by soil management unit may aid in refining recommendations. Field research continues to determine appropriate recommendation strategies.

Knowing your soil nutrient variability may be useful even when variable-rate application is not being considered. Directed bulk spreading of nutrients may be well justified to areas of the field identified as requiring more nutrients than those supplied through the standard whole-field fertilizer program. An example might be greater potassium requirements on lighter textured (sandier) soils. Likewise, sampling may detect areas where excessive nutrients are accumulating. This may be due to poor crop growth caused by salinity or lack of available soil moisture. Reducing fertilizer applications to these areas will reduce costs and prevent loading of nutrients that may be harmful to the environment.

Evaluating Your Soil Sampling Strategy

To assess a new soil sampling and variable-rate fertilization strategy, incorporate several constant rates strips into the field. Base the rates for these strips on a whole-field, composite approach. Several side-by-side strips spaced across the field will produce a more reliable evaluation than only splitting a field once.

Remain flexible and prepared to adjust your strategy. Emerging technology will further refine these sampling systems. In the future, soil nutrients sensors may be developed to map field nutrient levels, and satellite imagery may identify specific in-field nutrient stresses. And a growing database including yield maps, strip test evaluations, satellite images and aerial photographs will aid in defining appropriate management units.