Sampling soil effectively

SESL Australia

A spade in a very deep soil profileIdeally, you want to know the physical and chemical characteristics of every cubic centimetre of your soil. But testing the whole paddock would be impossible. So you have to take samples.

Those samples must be representative of the whole paddock or the results will be useless. So what is the best way to collect samples that you can be reasonably sure are representative of the whole paddock?

  • Start with a plan. Why are you sampling? What do you need to know? Which areas give you most concern? Give yourself enough lead time if you’re preparing for planting.
  • Use stainless steel tools wherever possible, or at least carbon steel. Don’t use any galvanised equipment, such as buckets, as these will contaminate the samples with zinc. Tools to use include an auger, a shovel or spade, a trowel, a soil corer and a ruler.
  • Put the samples in clean plastic containers or bags that can be sealed.
  • For best results in the long term, sample every year. The results will give you a complete history of your soil management and show you trends over time.
  • Sample at the same time every year. Otherwise, results cannot be reliably compared, as weather, temperature, fertiliser regime and crop growth stage will alter them.
  • Identify any areas that particularly concern you, such as wet sites or low-yield sites, and sample these separately: they might need different treatment.
  • Record the sampling locations so you can use them again next time for comparison. A GPS receiver is ideal, but a property map and a pencil will do. An aerial photo will help you to identify sites that show different growth and that might therefore need different treatment.
  • Because natural soils are not uniform (they are heterogeneous), you must collect enough samples to cover the range of variations in soil properties. Collecting not enough samples will miss some of the variation and will lead to over- or underestimation of results. So the more samples you can collect and combine into one representative sample for the paddock, the better. Note that we are talking about analysing only one sample per paddock (or area), but that this sample is a composite of several samples that you have mixed to get an “average” sample. The actual number of samples to collect and combine depends on the size and uniformity of your paddock and your willingness to dig them all. The more the better, but we suggest 10 to 20 per paddock. Aim to collect samples that represent the different features of the site, such as slope, aspect, elevation and soil moisture availability.
  • If it is clear that your paddock has more than one soil type in it, then collect separate samples for each soil type. Don’t combine soil types.
  • At each site, collect soil from three depth ranges: 0–100 mm, 100–300 mm and 300–600 mm. Contrary to common practice, SESL advises clients to pay attention to the nutrients (and moisture) in the subsoil, where most plant roots extend, in addition to the topsoil. Don’t mix the samples from the different depths.
  • Certain sites should not be sampled. These include bands where was fertiliser was applied, paddock boundaries and fence lines, cattle camps, former material stockpile sites, dip sites and sites where old buildings once stood. If you think that a site is a one-off exception in the paddock, skip it.
  • When combining the 10 to 20 subsamples per paddock into one composite sample (per depth), mix them thoroughly to make the soil as uniform as possible. Then take about 500 g of that composite to send off for analysis. Seal it well so it doesn’t spill during transport.

In summary, you should end up with three samples per paddock (or special area), one for each sampling depth. Send them to your preferred lab with full details: name, address, date, paddock name, soil depth, crop or pasture, and your reasons for wanting an analysis, such as fertiliser program, nutrient deficiency or concern about salinity. Allow plenty of lead time for both the analysis and your follow-up actions.