ATRIBUTOS QUÍMICOS E FRAÇÕES GRANULOMÉTRICAS DO SOLO EM DIFERENTES MODELOS DE PAISAGEM VISANDO A DIFERENCIAÇÃO DE ZONAS DE MANEJO

Abstract

As soil science’s applied techniques and methods advance, questions regarding spatial variability of soil characteristics arise. Using landscape models could help mapping areas with more homogeneous soils. This article’s goal is to investigate how the soil’s features behave in a hogback model in an agricultural area in the Ponta Grossa region in Paraná. In a 110 hectare area, soil samples from 0.0 – 0.2 m and 0.2 –0.4 m deep were collected – 60 samples per layer. The first model of landscape (TROEH, 1965), area was divided into 3 hogback areas, where the landforms were described as Linear (PF1), Concave – Convex (PF2) and Foothill Deposition (PF3). In the second model (DARLYMPLE, 1968) the area was divided into six units of the hypothetical component, which are described as Top (SV1), Shoulder (SV2), Upper Third of the Slope (SV3), Half-slope (SV4), Lower Third the slope (SV5) and Foothill (SV6).The Chemical analysis showed that pH, H+ Al, Al, Ca, Mg, K and P can change, C e CTCe, CTC, SB, V(%), m(%), relations (Ca + Mg), (Ca/Mg) and ((Ca + Mg)/K) and in Granulometry Physics. Data was analyzed by descriptive statistics and variance analysis with comparison of means by Tukey test at 5% probability. Results indicate that the model of landscape (TROEH, 1965) the greatest variation occurred in comparison with landform PF3 (Foothill Deposition) in both depths. As for the model Darlymple (1968) there is greater variation in variables for all six segments at a depth of 0.0-0.2 mm, compared with the total area. In both models studied, the landforms and segments of the strand was possible to verify interaction with the distribution of the elements studied, especially for the variables K, P, CTC, sand and clay. Thus, we conclude that the models can potentially be used for recommendations of tillage zones.