Why don't my split soil samples give the same value?

Monday, April 2, 2012

There are lots of reasons for variation in soil sample results. But it is still the best way to know what nutrients are available in your soil before you plant the crop

by KEITH REID

It's not uncommon to hear reports of someone who wanted to check the performance of a soil test lab by splitting a soil sample and sending it to two labs. Often, widely different results are obtained from each lab. Does this mean soil testing is flawed?

Not at all, but there is a certain level of variability in soil testing that needs to be understood to interpret the results of these differences.

Different test results or different recommendations?
Often, if I ask for more detail about the differences between labs, I will get a response like, "Well, Lab A recommended 20 kilograms of phosphorus (P), and Lab B recommended 50."

While this represents a significant difference in the cost of growing a crop, depending on which recommendation you follow, it does not necessarily mean the sample analysis process was flawed. Because fertilizer recommendations are generally made by matching a range of results to a single fertilizer rate on a table, a small difference in the measured number can make a big difference to the recommendation.

In the Ontario fertilizer recommendations for corn, for example, a measured P test value of 13 has a recommendation of 20 kilograms of phosphate per hectare, while a value of 12 has a recommendation of 50 kilograms, but it would be unreasonable to say that the difference in the measured valued is significant.

This difference occurs even when both labs are using the same fertilizer recommendation systems. But there are also cases where the agronomist at the lab will follow a different philosophy than that which is used for the "official" recommendations. Fertilizer recommendations to replace the nutrients that the crop removes, for example, will tend to be higher than recommendations to maximize the return to fertilizer.

Were the methods the same?
Soil testing tries to extract a portion of the nutrient that is correlated to the amount that plants can use, rather than trying to measure the total amount of a nutrient in a soil sample.

This may seem like a fine distinction but, since many factors can affect how easily nutrients are extracted from the soil, there is much greater variation between methods to measure available nutrients than between methods to measure total nutrients.

Most labs in Ontario should be using the standard extractants (sodium bicarbonate for P, and ammonium acetate for K and Mg), but if you did a comparison between labs in Ontario and Quebec or New York, the extractants would likely be different. The fact that the numbers are different does not mean that one is wrong since, like thermometers that read Celsius or Fahrenheit, they operate on different scales.

This issue gets even more problematic if you look at results for some of the micronutrients or secondary nutrients. There has not been as much effort put into developing and calibrating standard tests for these nutrients, so there is even less consistency from lab to lab than for phosphorus and potassium.

In addition, there is much greater variability within these tests, as they try to measure very small quantities of some nutrients. In other words, don't expect consistency in tests like boron, copper or iron, because it won't exist.

How well was the sample split?
So far, we have been discussing differences in results as if the samples sent to each lab were actually identical, but this is very difficult to do. Unless all of the clods were broken down and the sample thoroughly mixed and subdivided, big differences between subsamples can remain.

When proficiency samples for soil test labs are prepared, the samples are dried, ground, sieved to take out the coarse sand fraction (because it tends to segregate) and then thoroughly homogenized in a machine specifically designed for the job. Even with all this preparation, there are some samples that have too much variability between sub-samples to be used in the program. It is hard to believe that someone mixing moist soil in a bucket at the edge of the field could do a better job.

There are lots of reasons for variation in soil sample results, but that does not mean soil sampling is ineffective. With all its flaws, it is still the best way to know what nutrients are available in your soil before you plant the crop. Do not, however, expect a higher level of precision than the test is capable of producing. BF

Keith Reid is Manager (Eastern Canada), Soil Nutrient and GHG Management, Agriculture and Agri-Food Canada, Guelph.