Spatial variability of soil hydraulic conductivity along a tropical rainforest catena

We investigate the spatial variability of saturated hydraulic conductivity (Ksat) along a tropical rainforest catena, which consists of a Typic Kandiudult (Latossolo Vermelho Distrofico argissolico) and Plinthic Hapludox (Latossolo Amarelo Distrofico tipico). The former soil type extends from midslope to interfluve and has a high clay content throughout the profile, whereas the latter soil type extends from the midslope to valley bottom and only has a high clay content at depths >90 cm. We measured Ksat along a transect at depths of 20, 30, 50 and 90 cm with a compact, constant-head permeameter. Ksat is not significantly different (α=0.10) between these soil types at any depth except for 90 cm. Ordinary and robust linear regression models with distance from the interfluve as in the independent variable showed no significant change in Ksat as a function of topography. Loess regression showed some dependency of Ksat on topography only at the 90-cm depth. Semivariograms showed no apparent spatial structure in Ksat at distances ≥25 m for all depths. We conclude that the strong topography dependence of soil types along this catena and, hence, primary soil attributes, is not reflected in a similar dependence of Ksat, and tentatively attribute this lack of dependence to the overriding influence of bioturbation-controlled macroporosity. We outline a framework for detecting possible Ksat–topography relationships based on the distinction between bioturbation-controlled nonstructural macroporosity and mineralogy-controlled structural macroporosity.