Inter- and Intra-species Variation in Three Crown Condition Indicators for Seven Tree Species in the Southeastern United States

The USDA Forest Service utilizes assessments of tree crowns, specifically crown density, crown dieback, and foliage transparency, to accomplish in part its mission of reporting the long-term status, changes, and trends in forest ecosystem health in the United States. To aid interpretation and provide general guidelines of health across all species the crown condition assessments are classed into categories ranging from “good” to “poor.” The purpose of this research was to evaluate and describe the variation in crown density, crown dieback, and foliage transparency between and within species, and to critique the appropriateness of the current threshold levels. In addition, inter-observer deviation between two assessment crews was evaluated for crown density; however, the attempts to effectively predict between-crew variation were unsuccessful. The seven species included in the analyses were slash pine (Pinus elliottii), loblolly pine (Pinus taeda), Virginia pine (Pinus virginiana), red maple (Acer rubrum), sweetgum (Liquidambar styraciflua), yellow-poplar (Liriodendron tulipifera), and white oak (Quercus alba).

Between- and within-species differences were determined via pair-wise comparisons at the 10h, 25th, 50th, 75th, and 90th percentiles of the empirical distribution function of each crown condition indicator. Random “error” drawn from uniform distributions on the intervals (-2.5, +2.5) and (-7.5, +7.5) was added to the percentile estimates in order to capture the possible within-crew variation in the crown assessments. Bootstrapping was used to compute two-sided 90 percent confidence intervals (CIs) for each percentile with the percentile CI method.

A clear gradient of expected crown conditions was found among the species, but uncertainty in the data made it difficult to confidently pinpoint species-specific differences for the three crown condition indicators. Assuming limited measurement error in the data, the greatest disparity among species was found in crown density. Dissimilarity was apparent between hardwood and softwood crown densities in general, but only scattered differences were found among the species in each group. In terms of foliage transparency, Virginia pine was the most dissimilar overall. No major differences were found among the species in terms of crown dieback. In addition, relatively little variation was found within the two species (loblolly pine and sweetgum) examined for intraspecies variation.

Modifications to the current threshold levels were recommended for all three crown condition indicators. The suggested changes resulted in only small adjustments to the percentage of observations in each category and better reflect the distribution of observations across the range of the crown conditions. The proposed thresholds are:

• crown density: exceptional, 51-100 percent; good, 41-50 percent;moderate, 31-40 percent; and poor, 0-30 percent;

• crown dieback: none, 0-5 percent; light, 6-19 percent; moderate, 20-35 percent; and severe, 36-100 percent; and

• foliage transparency: normal, 0-20 percent; moderate, 21-40 percent; and severe, 41-100 percent.