We investigated changes in vegetation composition of different grass buffer strips in a fragmented coastal agricultural landscape to evaluate the potential for native grass restoration of sites that receive agricultural runoff. Vegetative buffers bordering Elkhorn Slough, draining into Monterey Bay, California, were either seeded with a non-native annual grass (Hordeum vulgare) or with a mix of native perennial grasses (Bromus carinatus, Deschampsia cespitosa, Nassella pulchra), and above-ground biomass and cover of vegetation were measured over a 4-yr period. Based on preliminary results, we initiated a second, smaller-scale experiment to test establishment of native perennial grasses (Bromus carinatus, Elymus glaucus, Hordeum brachyantherum) at different seeding densities with combinations of non-native annual grasses (H. vulgare or Lolium multiflorum and Vulpia myuros) to optimize erosion control.
In the first experiment, plots seeded with non-native annual grasses had greater biomass than native perennial plots in the first year. Biomass and cover of seeded annual grass decreased each year, which resulted in these plots being dominated by unseeded non-native species by the third year. In contrast, seeded native perennial grasses increased in both biomass and cover by the second year, with little cover of non-native species; but, in the third year cover of non-native species increased. By the fourth year, unseeded non-native species provided nearly all plant biomass and cover in all treatments. In the second experiment, native perennial grass cover was low, but was greater when seeded alone compared to when seeded with non-native annuals. The seeded annual grass V. myuros invaded and provided the majority of cover in most plots by the second year. Our results suggest that some species of native perennial grass can establish on former agricultural lands, but long-term survival is difficult without extensive management.