Representative sampling size and number of required samples for soil testing in direct seeding fields

Non-Peer ReviewedDirect seeding is widely utilized in the southern Canadian prairies. The associated band application of fertilizers makes conventional soil testing problematic. Strip sampling was suggested in direct seeding fields, but little is known about the optimum strip length. The objective of this study was (1) to identify the representative sampling size (RSS) of a sampling strip and (2) to determine the number of required samples (NRS) in a field in terms of point-based random sampling. Soil samples of 0-10 cm and 10-20 cm were collected from two 10 acre farm fields near Central Butte in the Brown soil zone of Saskatchewan. For strip sampling, five 160 cm long and 10 cm wide sampling strips were selected in these two fields. Different samples with sampling lengths ranging from 5 cm to 155 cm were obtained. For random sampling, 30 and 45 samples using a 4 cm diameter tubular probe were collected in these two fields. The results showed that RSSs differed with fields, nutrient types and soil layers. On average, the RSSs were 60 cm, 65 cm and 35 cm, respectively for testing NO3--N, NH4+-N and extractable P. The NRSs differed with sampling fields but not with nutrient types and soil layers. With a confidence level of 95%, about 30 and 80 random samples are needed in these two fields, respectively, to achieve mean estimate of soil nutrients with a relative error of 10%. This study provided a reference of soil sampling for soil nutrient tests in direct seeding fields