Maximizing Dryland Seed-Based Restoration Success Via Retrospective Analyses of Species-Use Patterns and Field Testing of Seed Mix Composition

Due to harsh and variable environmental conditions, successful plant establishment in "dryland" ecosystems – areas with scarce precipitation – is a major challenge for restoration. One of the most important decisions influencing restoration outcomes is choosing which species to plant (i.e., species selection). Here I explored two opportunities to improve restoration success related to species selection. First, I examined recent restoration projects across Utah, USA to understand species selection patterns by site conditions (e.g., precipitation, elevation) and seed costs, and to identify tools for matching species to sites. Second, I tested how commonly used non-native species – those not originally from the region, but sometimes used for their quick establishment and high seed production – affect native species abundance and community diversity, and whether adding more natives to a mix (increasing richness) and/or applying more seeds (increasing seeding rate) improved a) seed mix performance (e.g., weed suppression, grass establishment, diversity) in native-only mixes or b) native abundance and community diversity when co-seeded with non-natives. From the first study, I found that species choices generally followed planting guidelines (e.g., appropriate elevation) and seed supplier recommendations (e.g., precipitation requirements) and that more expensive seeds were selected less often – even when ecologically valuable. Species selection aligned with the Resistance and Resilience (R&R) framework, which classifies sites by their potential to resist invasion and recover from stress. A species' suitability for a site was prioritized over its origin (native vs. non-native), and seed mix composition varied according to the R&R framework. In the field experiment, high-performing non-native grasses reduced native species (but not overall diversity), though increasing native richness helped offset this. Native-only mixes suppressed weeds as effectively as mixes with both natives and non-natives but yielded lower overall grass abundance. However, increasing seeding rates and richness in native-only mixes helped compensate for this difference. These findings highlight opportunities to a) prioritize research and development for ecologically-valuable species to reduce costs, b) test R&R as a tool for seed mix design, c) co-seed natives and non-natives while balancing multiple goals, and d) reduce reliance on non-natives by increasing the use of native plant species.