Can mast history be inferred from radial growth? A test using five species of California oaks

Abstract

Quantifying masting behavior—the highly variable, yet synchronized, production of seeds by forest trees and other plants—is of considerable importance to ecosystem function and forest management, yet typically requires years of study to acquire. In contrast, measuring radial growth by mean of tree-ring cores potentially yields decades or more of data relatively quickly. Given the widespread existence of a negative correlation between growth and reproduction, can radial growth be used to infer historical masting behavior and detect long-term changes in reproductive behavior? Here we test this hypothesis in five species of California oaks (genus Quercus) for which we have long-term weather and 40 years of acorn production data at Hastings Reservation in central coastal California. Radial growth was measured for the three deciduous species using tree-ring analysis between 1980 and 1994 and for all five species using dendrometers between 1996 and 2019. Success was judged based on four criteria focused on the proportion of variance in annual acorn production explained, whether values using radial growth predicted very large and very small crop years, and whether five basic metrics of radial growth quantifying variance, autocorrelation, and synchrony matched values based on acorn production. Of the criteria, none was met using tree-rings with the exception of one of the metrics (mean pairwise synchrony). Dendrometers performed better but still failed two of the four criteria entirely and only met the other two for at most three of the five species. We conclude that radial growth does a generally poor job of estimating historical acorn production in these oaks. Although future advances may help resolve these shortcomings, demonstration that the relationship between radial growth and masting meet appropriate criteria based on years of overlapping data is currently necessary prior to any use of tree-ring growth to infer masting patterns.