Water fluxes between inter-patches and vegetated mounds in flat semiarid landscapes

Abstract

It has been assumed that bare soil (BS) inter-patches in semi arid spotted vegetation behave as sources of water to near vegetated soil (VS) patches. However, little evidence has been gained from direct measurements of overland and infiltration water fluxes between bare soil inter-patches and shrub mounds at a scale compatible with available high resolution imagery and hydrological modeling techniques. The objective of this study is to address the thin scale internal redistribution of water between BS inter-patches and vegetated mounds at relatively flat spotted semiarid landscapes. The relation between plant cover, topography and runoff was inspected with non-parametric association coefficients based on high resolution remotely sensed imagery, ground truth topographic elevation and spatial-explicit field data on potential runoff. Measurements of advective flows at the same spatial scale were carried out at micro-plots of BS and shrub mounds. Water fluxes between BS inter-patch and a shrub mound were simulated under varying typical Patagonian rainfall scenarios with an hydrological model. Results obtained revealed that the soil properties, infiltration and overland flow metrics at the mounds and inter-patches exhibit spatially and dynamic variable hydraulic properties. High micro-topographic roughness and depression storage thickened overland flow depth at VS patches. At BS inter-patches prevailing low slopes and depression storage were found to be important variables attenuating the surface runoff. At both rainfall scenarios simulated, the soil under the shrub mound accumulated more moisture (from direct rain) and reached saturation long before this occurred in BS nearby inter-patch area. Overland flow at the inter-patch was attenuated as it reached the border of the patch, diverging from the latter as it followed the (small) topographic gradient. The overland flow generated inside the vegetated mound was effectively retained at the typical Summer rainfall scenario; while several threads of runoff were routed outside the mound at the typical Winter rainfall scenario. The results here shown fail to detect: (a) enough runoff momentum that could route runon onto the vegetated mounds and (b) a contrast in infiltration rates between BS and the vegetated mound enough to lead to a free-surface gradient in ponded water that could inundate the mound.