How many soil samples should you take?

Figuring out the optimum number of soil samples and how best to divide your field is the key to obtaining the best net yield

by KEITH REID

It is a given that soil samples are the best way to determine how much fertilizer you need. But how should you collect those samples to get the most accurate information without overspending?

Ideally, a soil sample would tell us what parts of each field need fertilizer, and how much to apply for optimum crop growth and maximum profit. The reality is that soil fertility varies from point to point within a field, so no single soil sample will give the correct answers for everywhere in the field. The art of soil sampling is in taking enough samples from the right locations to identify the major divisions within the field.

Figure 1 shows a typical pattern of phosphorus soil test results if someone had collected samples once every rod (five metres) the length of a 100-acre field on a livestock farm. There are higher soil tests at the front of the farm because of past manure use, as well as natural variation from differences in soil types and a random component. The average of all the soil tests is 38 milligrams of phosphorus per litre which would indicate that no phosphorus fertilizer is required to grow any field crop. 

Looking at the graph, however, it is obvious that there are significant areas within thefield which have a high probability of responding to fertilizer. Given that we won’t know ahead of time where those areas are, how is the best way to divide the field?

Figure 2 shows the impact of dividing this field into two, three or four smaller sections, and using the average of all the values to predict the optimum fertilizer recommendation. Any subdivision is better than none, but it is not until this 100-acre parcel is divided into four that the higher recommendation for the back part of the field appears.

The true test, however, is how well each sampling scheme performs economically. Figure 3 shows the expected value of increased yield for fertilizer applied according to each sampling scheme, along with the costs for fertilizer and sampling.

In this example, the greatest net benefit to soil testing came from dividing the field into four subsections of 25 acres each. Interestingly, this also matches up with the original field boundaries. Subdividing into 40 fields produced higher yields, but the fertilizer and sampling costs were also higher so there was no additional benefit to the more intensive sampling. Larger field sizes did not generate as much net benefit because they did not identify the responsive areas of the field.

This example is artificial, and the numbers were chosen on purpose to show a range of soil tests from highly responsive to non-responsive. Fields that test predominantly low or high will show less economic benefit from dividing the field into sections. In low-testing fields, the best investment is in fertilizer rather than in more precise testing. In high-testing fields, the need for precision is in identifying areas of potential environmental risk, since the cost of intensive sampling to find the parts of the field that might be responsive to fertilizer will be larger than the value of any increased yield.

Keith Reid is soil fertility specialist with the Ontario Ministry of Agriculture, Food and Rural Affairs based in Stratford. keith.reid@ontario.ca