Features Agronomy Fertility and Nutrients
Making soil tests work

Soil testing has long been considered an important tool for determining fertilizer
rates, yet a perception exists that the soil test can provide an exact fertilization
recommendation for an individual field. The recommendation, however, depends
on the quality of the sample submitted; recommendation philosophy of the laboratory
or agronomist using the soil analysis; the integration of the farmer's yield
goal with management ability; and local environmental conditions and expectations.

"It is important to remember that soil testing is not an end in itself,
but simply a means to an end. It's one of many tools used to assure profitable
crop production," says Adrian Johnston, Northern Great Plains Director
with the Potash and Phosphate Institute of Canada (PPIC) at Saskatoon, Saskatchewan.

Taking the soil sample
Soil tests do have limitations, and understanding those limitations helps to
make soil test recommendations more useful and meaningful. A major limitation
of soil testing comes from sampling. The soil test result is only as good as
the sample collected for analysis. That might sound like a broken record, but
clearly, the more cores taken per sample, the better the results.

Current recommendations call for 15 to 20 core samples per field, regardless
of field size. Taking cores that are representative of the field will help to
ensure the overall sample reflects the field's overall fertility. If one core
out of 15 or 20 does not represent the average of the field, such as an old
yard site, saline border of slough, old fencelines, or eroded knolls, then there
is less potential that the sample will be spoiled, compared to when only six
to eight samples are taken.

Johnston recommends farmers consult with custom samplers to help ensure that
unrepresentative sites are avoided when sampling. However, even sampling between
or on the previous crop row can have an impact, especially with immobile nutrients
like phosphorous (P). Research shows that hitting a residual band of fertilizer
P can have an impact on the soil P estimate. For example, when 40 pounds of
P205 was seed applied to the previous
crop, the subsequent soil test found 27 pounds of available P per acre on a
sandy loam soil compared to only 12 pounds of available P from samples taken
between the row. In practical terms, though, when P is banded separately from
the seed, avoiding these bands is difficult, which makes soil testing over the
long-term more valuable, as it provides a more complete picture of long-term
trends in soil fertility.

Field variability can skew results
Field variability also comes into play. The goal of taking 15 to 20 cores is
to get a composite sample that reflects the overall field average. Research
results from a grid sampling project illustrate the variability found when intensively
sampling a field with rolling topography. The researchers sampled a 40 acre
hummocky field on a one acre grid over a three year period. The soil analysis
looked at nitrogen (N), P, potassium (K) and sulphur (S). During the three year
period, in all cases the average soil test value was greater than the mode (most
common value), indicating that a few large maximum values skewed the results
to a higher average value.

A good example of this variation is found looking at the K analysis in the
research. Approximately 37 percent of the 40 samples were greater than 143ppm,
a level above which no recommendation would be made in Alberta. However, 30
percent tested between 101ppm and 143ppm K, and may need supplemental K fertilizer.
Johnston says that the results of this research show the importance of sampling
in a random pattern. "Sampling requires far more attention to selecting
the areas of the field based on production history and simply using your head
to avoid problem areas."

Laboratory analysis techniques vary
Analysis of the soil sample is generally the most accurate part of soil test
recommendations. However, different laboratories use different tests for some
nutrients. Nitrate and sulphate are water soluble and use water to extract these
nutrients for a measurement that reflects plant available supply. Phosphorus
and K, though, use extraction processes that measure the total supply in the
soil sample, not just readily available plant supply. In this case, the soil
test laboratory makes an estimate of how much nutrient will become plant available
over the growing season, based on the soil type and environmental conditions.
Johnston says that it is important to select a laboratory that uses procedures
and philosophies that are appropriate for the soil properties in the area.

Results must be correlated to field response
Soil test results give a nutrient concentration in ppm, which have absolutely
no use unless those concentrations have been correlated to crop response in
the field. "A common misconception is that nutrient extraction equals nutrient
availability," says Rigas Karamanos, manager of agronomy with Westco at
Calgary, Alberta. He explains that a soil analysis provides a nutrient inventory
in the soil and it does not reflect nutrient availability.

After the soil chemical analysis, the resulting values are correlated to crop
response to nutrients to make fertilizer recommendations based on relevant research
for the crop. The research uses a large number of field trials where crop response
to applied fertilizer, across a wide range of soil nutrient values, is measured.

Johnston says it is important to remember that the resulting recommendations
are based on the average response from the combined locations/years and environments
sampled. Almost all laboratories have some sort of research data they use to
correlate soil sample results with, and provide some sort of calibration of
the data set. However, a laboratory based in the southern US will likely use
data from that region, which may often not be relevant to western Canadian farms.

Karamanos says that most of the calibrations were conducted back in the 1970s,
and no calibrations have been done since because of the expense of generating
the large amount of site years. "The calibrations may not be appropriate
to the changes in farming systems, such as zero-till or the new higher yielding
varieties, but nobody has the resources to redo them."

Some laboratories use crop removal data to calibrate soil test results and
fertilizer recommendations, not worrying about field response data. This is
especially true in areas where leaching between growing seasons results in the
loss of mobile nutrients like nitrate and sulphate.

"Ask your soil test lab where the data set comes from, how they calibrate
the soil test results and if the calibration is relevant to your farm,"
says Johnston. "It can make a large difference in the recommendations."

What are the recommendation philosophies?
So how do fertilizer companies get from ppm in the soil to pounds of fertilizer
per acre recommendations? That depends on their philosophy. Most companies use
a sufficiency approach that is based on yield response curves to applied fertilizer.
The sufficiency strategy recommends fertilizer application only when there is
a good chance that a profitable yield response can be obtained. An important
question to ask when using this approach is when and how the fertilizer is applied.
Depending on the nutrient, this also impacts on how much fertilizer should be
applied, as fertilizer use efficiency varies with time and method of placement.

The other approach used is the build and maintenance strategy. For immobile
nutrients like P and K, the fertilizer is applied in excess of crop removal
as a means of increasing the soil test level to the non-responsive range. Once
at this level, the nutrient is maintained by applying fertilizer to meet crop
removal. With tight economics, this approach is not generally used in western
Canada.

Another important consideration is whether the laboratory's recommendations
reflect nitrogen mineralization from soil organic matter. Few laboratories have
an estimate of the amount of nitrogen that may be made available to the plant
throughout the growing year. Karamanos recommends that growers and agronomists
ask whether the recommendations take mineralization into consideration and if
they do not, can they estimate it on their own? Mineralization varies depending
on the amount of organic matter in the soil and the soil moisture and temperature
during the growing season, so estimating the N released in pounds per acre can
be difficult.

Johnston also recommends that farmers work with their agronomists to develop
the fertilizer recommendations. Many other field specific factors must be considered,
such as yield goals, soil moisture conditions, fertilizer prices, previous crop
grown and previous year harvested yield and protein content, past fertilization
history including manure applications, tillage practices, previous soil test
records and observed production problems in the field.

"Bringing together current and historical information, with a good measure
of common sense, will help to build a good nutrient recommendation," says
Johnston. "So the next time someone hands you a soil test report, start
by asking questions that will help you make an informed decision."