Soil resistance to root penetration is an important physical characteristic for defining soil management techniques. The purpose of this study was to estimate the mechanical soil penetration resistance of an Oxisol under perennial crop, at three gravimetrical moisture contents, based on bulk density, moisture, and soil clay and organic matter content. The soil penetration resistance was evaluated by an impact penetrometer and the impact data in dm-1 were converted into dynamic resistance (MPa). At the moment of the determination of the mechanical soil resistance, samples (30 samples for each layer (0-0.10; 0.1-0.3; 0.3-0.5 m)) were collected to determine three different gravimetrical moisture contents (Ug) (Ug1 < 0.20 kg kg-1 , 0.20 < Ug2 < 0.30 kg kg-1 and Ug3 > 0.30 kg kg-1 ), clay texture (Arg), organic matter (MO) and soil bulk density (Ds). Tests of normal distribution were applied to select the participant variables of the models by the W test. Thereafter, analyses of simple correlation between the variable and STEPWISE multiple regression analyses were performed. To fit the models an analysis was performed with four independent variables, represented by: X1 = Ug, X2 = Ds, X3 = MO, X4 = Arg and the dependent variable (y) was represented by the soil penetration resistance (RP). After establishing the final models, the effects of multicollinearity were tested by VIF (Variance Increase Factor), where the VIF values < 10 indicate that the effect of multicollinearity does not influence the results of regression. A significance level of 15 % probability of the F value of the variable was used for the inclusion or exclusion of a variable from the model. Results led to the conclusion that the equations that best estimated soil penetration resistance were: for Ug < 0.20 kg kg-1: RPE = 1.12 + 2.04Ds + 0.17MO (R² = 0.92; R = 0.96); for (0.20 < Ug < 0.30 kg kg-1): RPE = 2.93-17.85Ug + 2.98Ds (R² = 0.95, R = 0.97); for Ug > 0.30 kg kg-1 :RPE = 3.20 - 13.78Ug + 4.43Ds (R² = 0.97, R = 0.98). In all studied situations the coefficients of variation (CV) varied from high (situation UG1) to very high (situation UG2 and Ug3). This high spatial variability of the RP means that care must be taken in the use of the models selected since points estimated by the models are dispersed in relation to the measured values.
compaction; management systems; bulk density; gravimetric moisture; impact penetrometry; model
