Factors related to the recovery of subalpine woodland on Mauna Kea, Hawaii
We measured mature tree and sapling density, tree associations, crown size, age structure, recovery from ungulate browsing, and grass cover at four study sites in two types of subalpine woodland on Mauna Kea volcano, island of Hawaii. Beginning in 1981, introduced ungulates were reduced in number to allow regeneration of Sophora chrysophylla (mamane) in habitat supporting the endangered Hawaiian finch, Loxioides bailleui (palila). We found Sophora regeneration at all four study sites, but regeneration was higher in mixed species woodland with codominant Myoporum sandwicense (naio) than in areas where Sophora dominated. Regeneration of Myoporum was uniformly very low in comparison. Invasive grass cover, which suppresses Sophora germination, was highest in mid-elevation woodland where Sophora dominated. The distribution of mature and sapling Sophora were both related to study site, reflecting previous ungulate browsing and uneven recovery due to grasses. Densities of Sophora snags were not different among any of the sites, suggesting a more even distribution in the past. Selective browsing before ungulate reduction may have favored Myoporum over Sophora , leading to high densities of mature Myoporum in codominant woodland. After ungulate reduction, however, we found no pattern of competitive inhibition by Myoporum on regeneration of Sophora . Reduction of Myoporum is not likely to enhance habitat for Loxioides as much as supplemental planting of Sophora , grass control, and continued ungulate eradication. Mid-elevation Sophora woodland areas, where Loxioides forage and nest in high densities, would benefit the most from these management actions.