Dynamics and ecological consequences of the 2013−2014 koa moth outbreak at Hakalau Forest National Wildlife Refuge.

Abstract

A massive outbreak of the koa moth (Geometridea: Scotorythra paludicola) defoliated more than a third of the koa (Acacia koa) forest on Hawai‘i Island during 2013−2014. This was the largest koa moth outbreak ever recorded and the first on the island since 1953. The outbreak spread to sites distributed widely around the island between 800−2,000 m elevation and in wet rainforest to dry woodland habitats. We monitored the outbreak at two windward forest sites (Laupāhoehoe and Saddle Road Kīpuka) and one leeward forest site (Kona), and we studied the dynamics of the outbreak and its impacts on the forest ecosystem at Hakalau Forest National Wildlife Refuge, our higher elevation windward site. Study sites at Hakalau included two stands of koa that were planted (reforestation stands) in former cattle pastureland about 20 years earlier and two stands of koa that were dominated by ‘ōhi‘a (Metrosideros polymorpha) and that were naturally recovering from cattle grazing (forest stands). We observed one outbreak at Hakalau, multiple outbreaks at the two other windward sites, but no outbreak at the leeward site. Caterpillars at Hakalau reached peak estimated abundances of more than 250,000 per tree and 18,000,000 per hectare, and they removed between 64−93% of the koa canopy in managed forest stands. Defoliation was more extensive in naturally recovering forest, where ‘ōhi‘a dominated and koa was less abundant, compared to the planted stands, where koa density was high. Koa trees were still growing new foliage six months after being defoliated, and leaves were produced in greater proportion to phyllodes, especially by small koa (≤ 8 cm dbh) and by larger trees in forest stands, where light levels may have remained relatively low after defoliation due to the high cover of ‘ōhi‘a. Small branches of many trees apparently died, and canopy regrowth was absent or low in 9% of koa trees and seedlings, which indicates the likely level of mortality. Between 2,000−5,000 kg/ha of frass fell during the defoliation event, resulting in the deposition of up to 200 kg/ha of highly labile nitrogen on the forest floor in less than two months. The deposition of nitrogen was detected as pulses in resin-available nitrogen in the top 5−10 cm of soil at two of three sites. These sites showed elevated soil nitrogen for about seven months. Nitrogen content of understory plant foliage, which is indicative of nitrogen uptake, suggested weak and variable effects of nitrogen deposition in the soil. Foliar nitrogen increased slightly in alien pasture grasses four months after the deposition of frass, although distinctive increases were not detected in native woody species. Birds responded to the abundance of caterpillars by increasing their activity in koa during the buildup of caterpillars and decreasing their use of koa after defoliation. During the outbreak, caterpillars increased in the diets of the two generalist insectivores we examined, and nearly all species gained weight. Bats responded to the abundance of moths by compression of active foraging into the first three hours of darkness each night after presumably having reached a digestive bottleneck. Reduced foraging activity by bats also resulted in lower indices of detectability based upon acoustic monitoring when compared to non-outbreak years. Parasitoid wasps tracked caterpillar abundance, but the low rate at which they attacked caterpillars suggests that they had little influence on the population. The predatory yellowjacket (Vespula pensylvanica) did not respond to the outbreak. Although a single, protracted outbreak occurred at Hakalau, multiple outbreaks and defoliations occurred at lower elevations. Our results provide a broad foundation for evaluating the dynamics and impacts of future Scotorythra outbreaks.