Guess Less: Types of soil testing and the best method for your goals

If you find yourself answering the following questions, it’s a good idea to arrange for soil sampling this fall:

• Have you taken on a new piece of land that you don’t know much about?
• Did you see any deficiency symptoms in your crops this year that you didn’t have a chance to diagnose in-season?
• Did you apply adequate nutrients, but yield was either way below or way above what you’d thought you’d get? Do you suspect high residual levels of nutrients?
• Are you dealing with known “problem areas” in a field but haven’t diagnosed exactly what the problem is?
• What were protein levels? Not that high in wheat? Too high in malt barley?

A single soil sample taken randomly within a field is incapable of accurately characterizing the nutrient status of that entire field. Yes, it can give an indication of existing nutrient reserves or act as a baseline measurement of more stable properties like texture, pH, or electrical conductivity, but one sample, statistically and practically, is just not enough. At its best, a single soil sample is a guess.

In essence, there are four soil sampling methods: traditional composite, benchmark, grid-based, and landscape directed. Out of these four, what method is going to answer any of the above questions?

Traditional Composite

A suitable method for flat-rate fertilizer application or to get baseline measurements for any new piece of land. Take 30 to 60 random cores per 160 acres in a zig zag pattern and avoid atypical areas (salinity, fence lines, approaches, burn piles, treelines). Bulk them together (composite) and thoroughly mix to take a single subsample for lab analysis. This method doesn’t provide any indication of field variability though and small areas with high nutrient levels may skew the “average” reported fertility level. If the sample result is artificially high, it will lead to under-fertilization and vice versa. There can also be considerable year-to-year variability in measured fertility levels when sampling randomly.

Benchmark Sampling

This method is simply sampling the same location every year. Benchmark locations can be selected based on close observation of crop growth, soil survey information, past experience, yield maps, detailed soil mapping, or remotely sensed images. The average benchmark sample encompasses a quarter acre and is representative of the field or the major soil type within the field. Fifteen to 20 cores are randomly taken in the benchmark area and are composited. Benchmark samples act as the reference area from which fertilizer recommendations for the field are based, so it’s critical to pick the correct areas to sample. This method is less expensive and less time-consuming than grid-based sampling. The year-to-year variations in measured fertility levels are more representative of actual nutrient changes; however, this sampling method assumes that most of the field is represented by the benchmark and therefore doesn’t provide a true indication of field variability unless multiple benchmarks are used.

Grid Soil Sampling

Grid soil sampling uses a systematic approach to soil sampling to reveal different fertility patterns throughout a field. This method assumes there is no logical reason for fertility patterns to vary within a field. A field grid is developed by placing marks at regular intervals in two directions, usually resulting in 2.5 ac grids, then collecting eight to ten soil samples from a 10 to 20’ radius at the intersecting points that are composited. Dense grid sampling is required to effectively reveal fertility patterns resulting in high sampling costs. In fields with complex topography, there is a major risk of missing landscape positions entirely and this method may result in repeatedly sampling only knolls and missing mid-slopes or depressions entirely.

Landscape-directed Soil Sampling

This method assumes that soil variations are not distributed randomly and requires identifying management areas or polygons within a field that have similar soil fertility and moisture conditions. Commonly referred to as zones, the spatial patterns must be defined by prior knowledge of the field, i.e. soil surveys, yield maps, remotely sensed images, or soil mapping. Hilltops, mid-slopes, toe-slopes, or depressions in a field will inherently have different soil characteristics, even if only slightly so. Three to six sample points per zone are identified, then single composite samples from 10-20 cores are taken from each zone. This method gives more accurate fertility results for the different landscape positions, which can then be tailored to variable-rate practices. However, if the zones are based on yield maps, the relationship between crop growth with topography may be completely reversed in years of extreme moisture compared to extreme drought. Past management may also mask the landscape position-soil fertility relationship and on relatively flat, uniform fields with subtle changes in topography, the variability is often more detectable at depth and may require more detailed elevation mapping.

Soil is inherently variable. Macro-variability (> 4’) in soil fertility within a field is associated with differences in parent material, topography, and past management. Considering it took roughly 17,000 years for Prairie soils to develop, soil variability presents the basic challenge in designing an effective soil sampling procedure.

Design a suitable sampling plan based on the level of information needed for your operation. Just know that having a number to base fertility or amendment recommendations on, in combination with yield removal rates or protein percentage, is better than having no number. Guess less.