Hydrology–vegetation interactions in areas of discontinuous flow on a semi-arid bajada, Southern New Mexico

Flows on a bajada surface in the Chihuahuan Desert of New Mexico show a discontinuous pattern, with alternating areas of channelization and deposition. Based on their planform appearance, we have termed the depositional areas ‘beads’. Instrumented catchments demonstrated that in comparison to ‘normal’ dendritic catchments, the beads show net infiltration in all but the largest flow events. Net accumulation of the bead surface appears to occur in years of lower than average rainfall, but the surface is dynamic and suffers net erosion in wetter years. The beads have significantly higher total vegetation covers, and contain higher proportions of grass species (Muhlenbergia porteri) as well as the creosotebush that characterizes the bajada surface in general. Creosotebush in the beads show consistently lower values ofδ13 C, suggesting that they are less moisture stressed than creosotebush elsewhere on the bajada. Values of δ15N are higher in the bead creosotebush, suggesting higher rates of soil–nitrogen transformations and greater loss of nitrogen to the atmosphere by denitrification. Both of these factors are again consistent with higher infiltration and thus moisture contents in the beads. In contrast, xylem-pressure potentials in creosotebush in the bead taken before the summer monsoonal season suggest that plants within the bead are more moisture stressed than those outwith the bead. Four possible explanations are offered for this apparent discrepancy: the greater cover and biomass of plants in the bead leads to higher stresses at the end of the drought period; spatial variability in soil texture means that some areas of the bead retain less moisture; more available water in the bead goes to the grass plants that concentrate their roots near the surface; and there is greater competition in the bead during times of drought stress. Overall, these results are compatible with an ‘islands of fertility’ interpretation of desertification operating in the American Southwest, albeit on a larger spatial scale than usually attributed. The concentrations of nutrient, water and seed resources may mean that these sites are favourable for attempts to reverse the desertification process. However, they appear dynamic with a life cycle of about 30 years.