Occurrence of Cavities and Cavities Suitable for Nesting Wood Ducks Across South Carolina Forest Types
Cindy L. Von Haugg
M.S. Student, Clemson University




Wood ducks (Aix sponsa) are common year-round residents of wetlands throughout the Southeastern United States. Evidence suggests that >90% of the North American wood duck population nests in tree cavities rather than in artificial nest boxes. However, few studies exist on the occurrence of natural tree cavities, particularly across their southern breeding range. We aimed to determine forest and tree characteristics indicative of cavities suitable for nesting wood duck hens to direct future studies to areas most likely to have cavities. We surveyed 20-m radius plots within the five dominant forest types of South Carolina (n = 166) on private and federally managed land and measured and inspected 4,633 trees >22 cm diameter at breast height (DBH) for cavity presence and suitability. We identified 225 potential cavities, 156 cavities, and 31 cavities suitable for nesting wood ducks. We found total cavity and suitable cavity densities (no./ha) at our study sites were greatest in oak (Quercus spp.), gum (Nyssa spp.), and cypress (Taxodium spp.) (19.35 ± 20.27, 4.20 ± 5.12) stands, followed by oak and hickory (Carya spp.) (8.37 ± 11.42, 1.36 ± 2.98), oak and pine (Pinus spp.) (4.90 ± 7.15, 1.17 ± 2.80), loblolly pine (P. taeda) (2.10 ± 4.42, 0.23 ± 1.34), and longleaf pine (P. palustris) (1.40 ± 3.57, 0.23 ± 1.34), which was consistent with densities at both sites individually. The best-fit model for cavity presence showed a significant positive effect for site index, DBH, and stand age and a minor negative effect for tree density. Results for suitability cavity presence also showed a significant positive effect for DBH and stand age, a significant negative effect for tree density, and no effect for basal area. We used an optimized hot spot analysis in ArcGIS to narrow down our total sample area using our findings to areas with 90% (73 ha), 95% (56 ha), and 99% (732 ha) confidence for suitable cavity occurrence, which amounted to 0.001% (73 ha), 0.001% (56 ha), and 0.008% of our total sample area (89,559 ha), respectively. This understanding of the relative abundance of cavities and cavities suitable for wood duck nesting and the identification of tree and stand forest metrics that influence the occurrence of cavities and suitable cavities promotes efficient management of nesting wood ducks, forest, and timber harvest practices.
Figure 1. Densities of cavities (F4, 4624 = 22.83; P < 0.0001) and suitable cavities (F4, 4624 = 4.29; P = 0.0018) for wood ducks in 166 randomly selected 20-m plots by forest type (n = 5). Plot surveys were conducted in stands >50 years old at Hobcaw Barony and Francis Marion National Forest, South Carolina, USA. Error bars indicate 95% confidence intervals, and significant differences are represented with different letters.
Figure 2. Heat map using estimates of relative abundance to develop optimized hot spots for areas with >95% and >99% confidence intervals for highest and lowest cavity occurrence probabilities. The abundance estimates were derived from forest surveys at randomly selected 20-m plots, stratified by forest type, at Francis Marion National Forest (left) in Berkeley and Charleston Counties, and Hobcaw Barony (right) in Georgetown County, South Carolina, USA, from February through July 2022.