Spatial characterization of scaled hydraulic conductivity functions in the internal drainage process leading to tropical semiarid soil management

• Scaled hydraulic conductivity functions determined in nested grids were site specific.
• Topographic features could explain 16–40% of the variations in hydraulic conductivity.
• The best fit for the hydraulic conductivity model coefficients was a normal distribution.
• Hydraulic conductivity functions were spatially auto-correlated within 25 and 60 m.
• GIS-map and semivariogram-based management zones could assist soil water management.

Knowledge of water movement is critical to agricultural water use. An internal drainage experiment for assessing downward movement of soil water under the influence of gravity was conducted on a semiarid sandy soil at ICRISAT-Sadore in Niger. The objectives were to estimate unsaturated hydraulic conductivity functions K(θ) at multiple scales using soil water (θ) data determined over the drainage process, and to determine spatial characteristics of K(θ) functions for water management in the semiarid environment. The study site was a naturally rolling field and the soil was a sandy Labucheri soil series. The experimental design consisted of a nested grid with three scales, 1 × 1-m, 5 × 5-m and 20 × 20-m, using 182 neutron access tubes. Soil water movement in space was examined in the 0.15–1.50 m depth throughout the drainage course. Mean soil water downward movement varied between 0.029–1379 mm h−1, depending on water content and soil depth. Elevation features contributed to 16–40% of the variation in hydraulic conductivity functions (K(θ) = e·θf). The best fit for K(θ)-model coefficients was a normal distribution and K(θ) functions were correlated at the three scales (0.81 ≤ R2 ≤ 0.96, P < 0.01). The scaled K(θ) functions were auto-correlated in space within 25–60 m. Management zones based on geostatistical semivariograms and interpolated map patterns were practical for water management planning in the naturally rolling environments.