Relating spatial patterns of sugar maple reproductive success and relative deer density in northern New York State

Abstract

Previous studies of sugar maple ( Acer saccharum) reproduction and its potential controlling factors have been performed on fine spatial scales (e.g. <1–10 ha), and the outcomes of these studies have often been extrapolated to regional scales (e.g. 1000–100,000 km 2). The resulting public and scientific perceptions are that sugar maple reproduction is declining over large areas in northern forests, and that white-tailed deer ( Odocoileus virginianus) herbivory is an important causal factor in this decline. We examined the validity of these perceptions in northern New York State (42,000 km 2) by performing a geostatistical analysis of sugar maple reproduction based on USDA Forest Service inventory data and relative deer density data constructed from hunter harvest records. We indexed sugar maple reproductive success (SMRS) as the number of sugar maple seedlings and saplings per saw log. Regression analysis indicated that global trends were present for both SMRS ( n=290, multiple r 2=0.085, P<0.001) and relative deer densities ( n=200, multiple r 2=0.17, P<0.001). Autocorrelation analysis indicated that SMRS had weak but significant spatial structuring at scales from 2400 to 40,000 m, but that most of the variability in SMRS probably occurred at scales finer than the data sources were able to detect (<2400 m). Relative deer densities demonstrated spatial patterning at scales from 10,000 to 82,000 m. At the township scale (i.e. approximate spatial grain of 10,000 m), SMRS and relative deer density observations were not correlated ( r 2=0.04, P=0.61, n=143). Additionally, the global trends in SMRS and relative deer density did not appear to be related. Within the range of deer densities observed in northern New York and at spatial scales >10,000 m, high deer densities were not necessarily associated with poor maple reproductive success. Lack of a broad-scale relationship between deer densities and SMRS does not preclude relationships at fine spatial scales (<10,000 m). Potential explanations for the lack of relationship between SMRS and deer density at broad scales include: (1) the generally low deer densities across the study area which may be below critical thresholds to impact reproduction; (2) the importance of functional compared to numerical response of deer in determining deer impact; and (3) the importance of other factors and multiple-factor interactions in affecting spatial patterns of reproductive success.