Abstract:
Soil performance is vital for the survival of human civilizations since it ensures provision of food for the human population. Soil compaction has impacted about 45% of agricultural soil and degraded an estimated 83 Mha of agriculture therefore reducing agricultural productivity. The objective of this study was to model and evaluate the compaction effect of tractor wheel traffic on sub soiling draft for soils in the agricultural farm of Egerton University, Kenya. Tractor wheel traffic experiments were conducted on the selected plots with varying levels of compaction. A dynamometer attached to the subsoiling equipment was used to measure the draft requirements during subsoiling. Soil samples were collected at various depths before and after tractor wheel passes and analyzed for physical properties (bulk density, moisture content, porosity, infiltration rate and saturated hydraulic conductivity) and mechanical properties (penetration resistance, cohesion, angle of internal friction and shear strength) were determined. The study employed a factorial experiment with a Completely Random Block design to look at the effects of five wheel passes (1, 2, 3, 4, 5) on soil properties at depths of 0 - 20, 20 - 30, and 30 - 40 cm with three replications. The wheel passes were equivalent to vertical loads of 26, 51, 77, 102 and 128 kN respectively. Increasing the number of wheel passes and the depth caused a significant increase in the soil's bulk density (from 1256 to 1593 kg m-3), penetration resistance (642 to 1539 kPa), strength (121.20 to 156.97 kPa), internal angle of friction (29 to 35o), and cohesion (6.84 to 8.42 kPa), while decreasing the moisture content (from 41 to 33%), infiltration rate (15.30 to 3.35 mm h-1), porosity (34 to 5%) and saturated hydraulic conductivity (5.63 to 0.54 to mm h-1). The draft requirement for subsoiling increased from 1.40 kN for the no pass at the 0 - 20 cm depth to 10.68 kN for five wheel passes in the 30 – 40 cm depth. Subsoiling draft was modeled as a function of soil depth, bulk density, penetration resistance, shear strength, angle of internal friction and cohesion. The R2 for Multiple Linear Regression (MLR), Dimensional Analysis, Adaptive Neuro Fuzzy Inference System (ANFIS) and Artificial Neural Networks (ANN) were 0.9838, 0.8722, 0.9999 and 0.9941 respectively. The t-test revealed that there was no significand difference between the measured and predicted draft for the MLR (t = 0.13), Dimensional Analysis (t = 0.15), the ANN (t = 0.12) and ANFIS (t = 0.19) models. Conclusions of this study emphasize the significant influence of tractor wheel passes on soil properties and draft requirements. Recommendations include the promotion of optimized subsoiling strategies. This research contributes to promoting sustainable agricultural practices and soil management strategies.
