Abstract
In October 2011, a snowstorm in the northeastern USA caused many branch failures of many tree species commonly planted in urbanized settings. Immediately following the storm, we assessed 1,764 trees for possible snow-induced damage and factors affecting it on the campus of the University of Massachusetts in Amherst, MA, USA. Nearly all failures were of branches, most of which were not defective. We used logistic regression to assess whether the probability of branch failure differed among species, diameter at breast height (DBH) and the presence of a defect or leaves increased for different species. We also measured branch morphology of (i) branches that did and did not fail for one angiosperm species and (ii) all branches on a sub-sample (stratified by DBH) of three individuals of seven other angiosperm species. Probability of branch failure differed among species. It also increased with greater DBH in eight of ten species studied, decreased when defects were present in four of ten species, and increased in one species when leaves were present. The relationship between branch failure and DBH appeared to be due to the correlation between DBH and branch morphology, which was mostly similar among species. As DBH increased, so did the mean diameter and length of primary branches, and the cumulative diameter of secondary branches. In contrast, branch slenderness decreased with increasing DBH. Combined, these factors presumably expedited the accumulation of snow on branches due to greater surface area and less flexibility. This explained why most failed branches were not defective. Since the frequency of intense storms is predicted to increase with global climate change, urban foresters should consider the timing of leaf senescence when selecting deciduous trees, to reduce the likelihood of failure of open-grown, deciduous trees in urbanized areas.