Spatial patterns of tree recruitment in a montane Hawaiian wet forest after cattle removal and pig population control

Prepared by Thomas Ibanez & Patrick J. Hart

Scattered trees (Metrosideros polymorpha, Myrtaceae) in Hakalau Forest National Wildlife Refuge, Island of Hawaiʻi, U.S. Photo credit: Patrick J. Hart.

As on many remote tropical islands, the natural ecosystems of Hawai’i have greatly suffered from the introduction of ungulates such as cattle, pigs and goats. These ungulates rapidly became feral after they were introduced in the late 1700s and proliferated for more than a century, destroying the native vegetation in many areas. Hawaiian plants are defenceless against these introductions as they evolved for millennia in the complete absence of herbivorous mammals. Hawaiian montane wet forests have been particularly affected by ungulates and a large portion of them have been converted into grasslands. The rapid restoration of montane forests is critical because in the coming decades they may constitute the last refugia for many Hawaiian native forest birds facing avian malaria at lower elevations.

Hakalau Forest National Wildlife Refuge located between 750-2000 m elevation on the windward slope of Maunakea volcano on the island of Hawai’i provides critical habitat for endangered Hawaiian forest birds. The upper part of the refuge was converted to cattle grazing lands by the mid-1800s. In the early 1990s, portions of the upper refuge were fenced, all cattle were removed and the feral pig population was controlled. Monitoring and analysing natural recruitment of forest trees following these passive restoration actions is critical to determining whether or not active restoration is needed and, if so, which restoration action should be conducted to help forest recovery.

We recorded more than 4000 new trees in 200 plots (15-m radius) located on two 100-ha sites over ~17-18 years following cattle removal and pig population control. The spatial distribution of new trees highlights the importance of remnant trees and seed sources for forest recovery. At the landscape scale, our results suggest that the number of new trees per species is limited by the number of trees present at the sites at the beginning of the study. The number of new trees that have been dispersed by birds was higher in areas with lower grass cover, supporting the idea that once grazing pressure removed, competition with grasses is an important limiting factor for tree recruitment. The apparent success of passive restoration in the study sites is likely due to the fact that they haven’t been fully converted into pure grassland with long-lived scattered trees and that relatively intact forest was left nearby. The absence of non-native, invasive tree species at the study sites likely facilitated the recruitment of native trees. Our study highlights the importance of keeping and maintaining scattered trees and forest patches when transforming the forest into agricultural or pastoral lands. Among other benefits for biological conservation and ecosystem functioning, the presence of remnant trees should provide higher resiliency if the decision is made to restore the forest passively.

This is a plain language summary for the paper of Ibanez & Hart published in the Applied Vegetation Science (https://doi.org/10.1111/avsc.12478).