How large is large? Urban tree allometrics from 16 U.S. climate regions

Background/Question/Methods

Planning the appropriate space for mature trees in growing cities, projecting costs for managing urban tree populations through time, modeling carbon storage, air pollutant uptake, energy savings, rainfall interception, and heat island reduction associated with urban trees all depend upon tree growth and size data for a wide range of native and introduced species. And yet, internationally, very little data are available to predict relationships between tree age and diameter-at-breast height (DBH), or DBH and other dimensional parameters. We collected data on over 16,000 trees representing 171 unique species growing in 22 U.S. cities across 16 climate regions. A stratified random sample of street trees was drawn from each city’s municipal tree inventory and measured to establish relations between tree age (number of years after planting) and DBH; DBH and tree height, crown height, average crown diameter, and leaf area; and average crown diameter and DBH. Five models were tested at four weights for all parameters. The Akaike Information Criterion (AIC), correlation and MSE were used to select best models. Like species were compared within and across climate zones to assess trends and evaluate potential causes for size difference and effects of tree size on provisioning of ecosystem services.

Results/Conclusions

The analysis produced 2,562 models for 171 species. Over 80 percent showed correlations higher than 0.80 (alpha=0.05). Models with weaker correlation typically predicted crown height and the greatest variability associated with small trees, often subjected to severe pruning. Using DBH to predict tree height and crown diameter showed the strongest correlations. Comparisons for like species in Fort Collins, Colorado and Cheyenne, Wyoming, cities less than an hour’s drive apart, predicted significantly different growth rates. Cheyenne species were 15-25% smaller in height and crown diameter over the same growth period. Slower growth reduced the production of ecosystem services. For example, the monetized value of ecosystem services for Acer saccharinum in Cheyenne were 40 percent less than the same species growing in Fort Collins over 40 years. Climate and soils may be major causal agents in Cheyenne, whereas rigorous pruning in San Francisco was responsible for relative crown reduction compared to like species in other cities. The database for this 12-yr long study and the resultant models will form the basis for what will become an international tree growth database. Researchers and practitioners can contribute new data and access existing data for tree growth and longevity analyses.