Changes in Pandora Moth Outbreak Dynamics during the Past 622 Years

Episodic outbreaks of pandora moth (Coloradia pandora Blake), a forest insect that defoliates ponderosa pine (Pinus ponderosa Dougl. ex Laws.) and other pine species in the western United States, have recurred several times during the 20th century in forests of south-central Oregon. We collected and analyzed tree-ring samples from stands affected by recent outbreaks of pandora moth to develop a long-term record of outbreaks. Outbreaks were evident in tree-ring series as a characteristic "signature" of sharply reduced latewood width within a ring, followed by reduced ring widths lasting 4-20 yr. We verified that this tree-ring signature was unrelated to drought or other climatic fluctuations by comparing the timing of known and inferred outbreaks with independent climatic data. Using the pandora moth tree-ring signature, we reconstructed a 622-year record of 22 individual outbreaks in 14 old-growth ponderosa pine stands. This is currently the longest regional reconstruction of forest insect outbreak history in North America. Intervals between pandora moth outbreaks were highly variable within individual forest stands, ranging from 9 yr to 156 yr. Spectral analyses of a composite time series from all stands, however, showed more consistent intervals between outbreaks, suggesting quasicyclical population dynamics at regional and decadal scales. Waveforms extracted from the regional outbreak time series had periods ranging over ∼18-24 yr (39.7% variance explained) and ∼37-41 yr (37.3% variance explained). The periods and strengths of these cycles varied across the centuries, with the largest outbreaks occurring when relatively high-amplitude periods of the dominant cycles were in phase. Twentieth-century outbreaks were not more synchronous (extensive), severe, or longer in duration than outbreaks in previous centuries, but there was an unusual 60-yr reduction in regional activity during ∼1920-1980. The changing dynamical behavior of pandora moth populations highlights the need to evaluate historical factors that may have influenced this system, such as climatic variations, forest fires, and human land uses. Although cyclical dynamics in animal populations have most commonly been attributed to endogenous, ecological processes (e.g., "delayed density dependence," predators, pathogens, and parasites) our findings suggest that exogenous processes (e.g., climatic oscillations) may also be involved.