Drivers of seedling establishment success in dryland restoration efforts.

Restoration of degraded drylands is urgently needed to mitigate climate change, reverse desertification and secure livelihoods for the two billion people who live in these areas. Bold global targets have been set for dryland restoration to restore millions of hectares of degraded land. These targets have been questioned as overly ambitious, but without a global evaluation of successes and failures it is impossible to gauge feasibility. Here we examine restoration seeding outcomes across 174 sites on six continents, encompassing 594,065 observations of 671 plant species. Our findings suggest reasons for optimism. Seeding had a positive impact on species presence: in almost a third of all treatments, 100% of species seeded were growing at first monitoring. However, dryland restoration is risky: 17% of projects failed, with no establishment of any seeded species, and consistent declines were found in seeded species as projects matured. Across projects, higher seeding rates and larger seed sizes resulted in a greater probability of recruitment, with further influences on species success including site aridity, taxonomic identity and species life form. Our findings suggest that investigations examining these predictive factors will yield more effective and informed restoration decision-making.

Patch-scale effects of equine disturbance on arthropod assemblages and vegetation structure in subalpine wetlands.

Assessments of vertebrate disturbance to plant and animal assemblages often contrast grazed versus ungrazed meadows or other larger areas of usage, and this approach can be powerful. Random sampling of such habitats carries the potential, however, for smaller, more intensely affected patches to be missed and for other responses that are only revealed at smaller scales to also escape detection. We instead sampled arthropod assemblages and vegetation structure at the patch scale (400-900 m(2) patches) within subalpine wet meadows of Yosemite National Park (USA), with the goal of determining if there were fine-scale differences in magnitude and directionality of response at three levels of grazing intensity. Effects were both stronger and more nuanced than effects evidenced by previous random sampling of paired grazed and ungrazed meadows: (a) greater negative effects on vegetation structure and fauna in heavily used patches, but (b) some positive effects on fauna in lightly grazed patches, suggested by trends for mean richness and total and population abundances. Although assessment of disturbance at either patch or landscape scales should be appropriate, depending on the management question at hand, our patch-scale work demonstrated that there can be strong local effects on the ecology of these wetlands that may not be detected by comparing larger scale habitats.