Changes in soil chloride following shrub removal and subsequent regrowth

Rates of groundwater recharge are especially unpredictable in semiarid regions, partly because of the role that vegetation plays in soil–plant–water dynamics. Understanding how abiotic and biotic variables affect water movement and associated soil-chloride distribution is essential to estimating groundwater recharge. Converting shrubland to grassland reduces rooting depth and plant cover, at least temporarily. In turn, reduced rooting depth increases leaching of soil chloride from the root zone when evaporative water losses are less in the grassland than in the original shrubland. This study contrasts soil and vegetation effects on soil chloride in semiarid rangelands. We established experimental plots that represent a successional chronosequence. Three native, undisturbed stands of rangeland brush (control) dominated by mature honey mesquite (Prosopis glandulosa) were compared with adjacent land plowed of vegetation at different times (in 1991, 2001, and 2005) and were subsequently allowed to regrow (3 locations × 4 plot types). Soil samples were collected at incremental depths up to 4.2 m and analyzed for physical properties (bulk density, texture, and others), chloride concentration, and root biomass. Aboveground vegetation surveys were also conducted. We expected shallower rooting depth for at least 5 years after conversion of shrubland to grassland and attributed deeper peak soil-chloride concentration to increased deep drainage potentially leading to groundwater recharge. Differences among soil abiotic properties were too minor to change chloride distribution. Rather, differences in chloride distribution were related to changes in vegetation. Shrub cover increased over time after clearing. Grass cover peaked in the 5-year treatment plots and decreased as shrub species regained dominance. Contrary to expectations, root biomass increased for approximately 5 years following the clearing of shrubs, but then it decreased with time. Roots were shallowly distributed in earlier successional plots. Vegetation removal greatly reduced root-zone total chloride in 5- and 15-year treatment plots compared with control plots, suggesting chloride had been flushed from the profile following brush removal, after as short a time as 5 years. In addition, deepest peak soil-chloride concentration occurred in 5-year plots. We created a stepwise multiple regression model based on related soil and vegetation attributes. According to the model, time since treatment accounted for 24% of the variation in the depth to maximum chloride concentration, and root depth accounted for an additional 20% of the variation. Hydrologic changes following brush removal were evident in our study area within the first 5 years, and they are likely to positively influence groundwater recharge.