Comparing classification systems for ponderosa pine snags in northern Arizona

The degree to which snag classification systems predict actual snag '' age '' (time since tree death) is rarely known. We used dendrochronology to determine the death dates of 79 large (more than 40 cm dbh, more than 3 m in height) ponderoso pine snags in northern Arizona and compared the snag ages to those predicted by a commonly used simulation model and to five classification systems. The methods are limited to trees for which intact tree-ring samples could he extracted, but the data provided evidence of some snags standing for long periods (to 141 years) and correlation of snag traits with age. The simulation model underestimated snag age in the first three categories. Mean ages of snags in the lost category were estimated accurately but with low precision. (omporison of snag ages to their classification in five systems showed that bole breakage was unrelated to snag age, that limb loss occurs at divergent rates between individual trees, and that wood decay does not necessarily proceed sequentially. A system based on the attrition of bark from snag boles proved useful for distinguishing snags by physical traits, but different classification systems may be needed to meet specific needs for assessing past forest structure and wildlife habitat.