Monitoring the state of soil compaction periodically by assessing soil penetration resistance is a practical way of evaluating the effects of different management systems on the soil structure and crop root development. This study aimed to evaluate the variation of soil penetration resistance in response to the number of replications (sample population) of different field sampling forms of an Oxisol under three management types: non-tillage (PD), perennial crop (CP) and conventional tillage (PC). The experiment was carried out in the Northern Paraná State, Brazil. Samples were collected in three sub-areas of 1ha to determine soil penetration resistance at different depths (0-0.10; 0.10-0.20; 0.20-0.40; 0.40-0.60 m). Sampling was carried out as follows: systematic sampling (grid points, spaced 25 m apart) and completely randomized sampling, with 1, 3, 5, 10, 15, 20, 30, 40, and 50 replications. For all points and depths it was determined the average value of penetration resistance (MPa), the confidence interval and estimation accuracy (D) of the penetrometer measurements through classical statistical theory based on the number of samples (n ) and standard deviation of the sample (S), at a significance level of 0.05. For the given experimental conditions (sub-areas of 1ha Oxisol under three different managements), results indicated that the number of representative samples to determine soil penetration resistance did not vary significantly with the sampling type. However, the confidence intervals were always lower when it was adopted systematic sampling and the number of representative samples to determine soil penetration resistance varied with management system and sampling depth. The mean soil penetration resistance was more representative for n > 15 (DP and CP) or n > 20 (0-0.10 m) and 15 (0.20-0.60 m) for PC; the population sample n 10 in the 0-0.60 m layer by both sampling types and soil managements studied enabled the obtaining a high accuracy of the data and the statistical parameters more reliable, with a certain homogeneity and linearity of the results presented in curves of the sample populations, considering a sample error of 10 %.
management; sampling; soil penetration resistance; soil compaction
