Background/Question/Methods   Arid and semi-arid rangelands have suffered widespread degradation globally and remain difficult to rehabilitate. The development of resilient rehabilitated ecosystems which are self-sustaining over the long term requires the re-establishment of ecological function, interactions and processes. We investigated whether artificial water retention structures could retain resources and re-establish the soil-water-plant feedbacks that support biological activity in these variable ecosystems. Restored areas behind water retention dikes (earthen berms 37-45 cm tall, 50-150 m long, in sets of 4-5 dikes, at 3 sites) constructed 1975-1981 on Jornada Experimental Range NM were compared to nearby unmanipulated reference areas. Accumulated changes 25-30 years after construction of dikes were measured in vegetation species composition and percent cover, soil texture, aggregate stability and saturated single-ring infiltration. We investigated the current resilience to the major drivers: rainfall, grazing and nutrients. Rainfall was monitored with a tipping bucket recording rain gauge at each site. Herbivory by terrestrial large grazers (cattle, oryx, pronghorn) or small herbivores (> 0.6 cm X 0.6 cm) was manipulated by fencing plots at each site. A complete mixture of macro- and micro-nutrients, including 200 kg N/ha, was added annually to subplots from 2005-2007. Soil moisture was monitored with shallow (5 cm) and deep (40 cm) ECH2O Probes at 2 locations within each dike and reference area. Species composition and percent cover were measured pre-treatment and annually from 2005-2008, net primary production and nitrogen uptake pre-treatment and 2008.

Results/Conclusions   Percent cover of vegetation, litter and grass were greater in the dike areas than in the reference areas. Plant species richness was 2-5 times greater in the dike areas than in the reference areas. Stability of the soil crust was greater in the dike areas. At the two finer textured sites, the rate of infiltration was not significantly different in the dike and reference areas, but at the sandy site, the infiltration rate in the dike area was slower, 55% of the infiltration rate in the reference area. The dike areas responded more strongly to rainfall, exclusion of large and small grazers and increased availability of nutrients. Artificial water retention structures reduced surface water flow rates, allowing more time for water to infiltrate into the soil, initiated the formation of resource islands and re-established some soil-water-plant interactions. In this long-term restoration, some ecological functions have been re-established and resilience has improved.