Soil-vegetation feedbacks in dynamic landscapes: Implications for restoration

State changes from perennial grasslands to shrub-dominated systems characterize much of the arid regions of the world, including the American Southwest. Where this conversion is associated with a change in soil surface texture, near-surface water availability can be increased or reduced with important consequences for restoration of perennial grasses. Loss of soil from one area results in deposition in downslope or downwind areas. Little is known about these spatial connections, and how erosion-deposition patterns influence seedling establishment and recovery of perennial grasses. Our objective was to evaluate how soil erosion-associated changes in surface soil texture influence the probability of seedling establishment of the perennial grass, black grama, across a range of initial soil textures. We used a simulation model of daily soil water dynamics (SOILWAT) to simulate recruitment probabilities of this species for the variation in initial soil texture found at the Jornada Basin ARS-LTER site in southern New Mexico. Vegetation and landform maps combined with soils, vegetation, and weather data were used to parameterize the model. Simulated recruitment probabilities were generally higher where near-surface water holding capacity was higher. However, in some cases deposition of small amounts of coarse-textured material (lower water holding capacity) on a relatively fine-textured (higher water holding capacity) site increased establishment due to reduced evaporative losses. We used a wind model to determine spatial patterns in soil loss and deposition at the landscape scale, and displayed our recruitment results for the Jornada Basin. Locations with the largest changes in soil surface texture as a result of wind had the largest changes in recruitment probabilities. Our results show that wind erosion and deposition patterns have strong influences on patterns in seedling establishment.