Spatial and temporal variability of soil saturated hydraulic conductivity in gradients of disturbance

Tropical montane rain forests are subject to both natural and anthropogenic disturbances, such as shallow landslides and forest-to-pasture conversion. Vegetation regrowth is rapid upon attaining hillslope stability and pasture abandonment, respectively, and apt to affect soil hydrology via changes in soil structure, a sensitive indicator of which is soil saturated hydraulic conductivity (Ks). Our objective was to quantify the influence of these regionally widespread and important disturbances on Ks and the subsequent recovery of Ks, and to describe the resulting spatial patterns. In a 2 km2 large research area in southern Ecuador, we used a mixed design- and model-based sampling strategy for measuring Ks in situ at soil depths of 12.5, 20, and 50 cm (n = 30–150/depth) under landslides of different ages (2 and 8 years), under actively grazed pasture, fallows following pasture abandonment (2–25 years of age), and under natural forest, and for elucidating its spatial patterns. Global means of soil permeability generally decrease with increasing soil depth. Ks does not differ among landslides and in comparison to the natural forest, which suggests a marginal effect of the regional landslide activity on soil hydrology. In contrast, results from the human-induced disturbance regime show a permeability decrease of two orders of magnitude after forest conversion to pasture at shallow soil depths, and a slow regeneration after pasture abandonment that requires a recovery time of at least one decade. Disturbances affect the Ks spatial structure, in particular the correlation length, in the topsoil. The largest differences in the covariance parameters, however, are found for the subsoil Ks, where the spatial structure is independent of land cover. This case study suggests a rather disparate soil hydraulic response to regionally important disturbances. Cattle grazing strongly affects the spatial mean of Ks, whereas landslides do not, and both the processes affect the spatial structure of Ks in the topsoil.