Questions

(a) Have Metrosideros polymorpha trees become re‐established in Hawaiian forests previously impacted by canopy dieback in the 1970s? (b) Has canopy dieback expanded since the 1970s? (c) Can spatial patterns from this dieback be correlated with habitat factors to model future dieback in this area?

Study Site

An 83,603 ha study area on the eastern slopes of Mauna Loa and Mauna Kea volcanoes on the island of Hawaii, USA.

Methods

We analyzed very‐high‐resolution imagery to assess status of Metrosideros polymorphaforests across the eastern side of the island of Hawaii. We generated 1,170 virtual vegetation plots with a 100‐m radius; 541 plots in areas mapped in 1977 with trees dead or mostly defoliated (dieback), and 629 plots in adjacent wet forest habitat, previously mapped as non‐dieback condition. In each plot we estimated the frequency of M. polymorpha trees that were dead or mostly defoliated, and the frequency of trees with healthy crowns. These results were combined with habitat data to produce a spatial model depicting probability of canopy dieback within the study area.

Results

Seventy‐nine percent of plots mapped in 1977 in dieback condition recovered their canopy and were now considered in non‐dieback condition. Ninety‐one percent of plots in previous non‐dieback areas were found to still have a healthy M. polymorpha canopy in 2015. A spatial model allowed us to identify areas within the study area with high, medium, and low probability of experiencing this same type of canopy dieback in the future.

Conclusions

Most former dieback areas mapped within the study area in 1977 now show recovery of the tree canopy through growth of new cohorts of young M. polymorpha trees. This suggests these forest communities are resilient to this type of canopy loss and tree death so long as other factors do not disrupt the natural regeneration process.