Shallow Subsurface Flow in Semiarid Shrublands on Karst Terrain

We know comparatively little about runoff generation in semiarid landscapes—especially those overlying karst geology, where flow pathways are exceedingly complex. Another layer of complexity is added by the phenomenon often described as woody plant encroachment (increasing coverage by woody plants). In the Edwards Plateau region of Central Texas, where Ashe juniper (Juniperus ashei) has been spreading, a commonly expressed concern is that streamflows and groundwater recharge are declining as a result. With the aim of better understanding the hydrological implications of Ashe juniper expansion in environments having an underlying geology of predominantly Cretaceous limestones and dolomite, we have been conducting large- scale rainfall simulation studies at a number of locations across the Edwards Plateau. Simulated rainfall has been applied on hillslopes ranging from 12 to 26 m in length in (1) areas with dense juniper canopy, (2) intercanopy areas, and (3) areas recently cleared of juniper. At the base of each hillslope, a trench was excavated for monitoring shallow subsurface flow, or interflow. In addition, for each experiment we monitored Horton overland flow, stemflow for the trees, and throughfall. Our results were consistent: for the canopy plots—both before and after removal of trees—almost all the runoff was shallow subsurface flow, which accounted for anywhere from 50% to 100% of the water applied. In contrast, for the intercanopy plots most of the runoff occurred as overland flow (around 30% of the water applied), with lesser amounts of interflow. Removing the juniper had little or no effect on runoff processes, in that interflow remained the dominant process. However, canopy removal had a large effect on how much water reached the surface. We estimate that even under the high rainfall application rates, the canopy was intercepting and storing only about 20% of the water applied. Runoff generation in these landscapes occurs as a patchwork of both Horton overland flow and interflow; which process dominates in a given area is probably controlled by the permeability of the underlying parent materials.