Results/Conclusions Herbicide reduced production of non-native annual grasses by 85% during the first spring following application. Production of non-native and native annual forbs was also reduced (92% and 67%, respectively). Soil seed reserves reflected the lower production of brome grasses (69% reduction in seed densities) and non-native annual forbs (71% reduction), but native annual seed reserves were not affected by herbicide. Herbicide effects persisted during the second spring for non-native annual grasses and forbs (40% and 44% reduction in shoot production, respectively) and were associated with a doubling of native annual forb production. Seed reserves of native perennial grasses, which made up 91% of the seed mixture, were unaffected by herbicide, and four years after seeding treatments, seed densities remained almost three times higher in both seeding treatments compared with the unseeded treatment. Sites that received the highest cumulative rainfall had 1.6 times the seeded perennial plant densities and 3% greater live canopy cover than sites with lower rainfall. Even though the densities of seeded species were reduced by herbicide during the second spring, the live perennial canopy cover was not different between the treatments. Our results demonstrate that management of the soil seed reserves through the combination of seeding and pre-emergent herbicide is a promising strategy for facilitating recovery of burned desert habitats.