Reliably Estimating Street Tree Counts, Species Diversity, and Pest Vulnerability with Random Sampling

Abstract Background Practitioners rely on sample-based estimates of street tree population characteristics when complete inventories are not feasible. Selecting a sample size is a primary consideration when implementing a sample-based inventory, as it involves a tradeoff between costs and data quality. Methods We used street tree inventory data from 16 municipalities in Indiana, USA, to assess how data quality improves with increasing sample size. Specifically, we conducted 1,000 random draws of street segments at increasing sample depths to observe how estimates improved for the number of total trees citywide, species richness, species diversity, and vulnerability to an invasive pest. Results Compared to previous research, our results indicate that a larger percent of sampled street segments is needed to achieve relative standard error values below the heuristic target of 10%. We also calculated reliability thresholds that showed the percent of street segments that would need to be inventoried to achieve estimates within a given margin of the true citywide value in 95% of random draws. Again, relatively large random samples were needed to reliably achieve accurate estimates of street tree characteristics, especially in smaller municipalities. Conclusions This study provides information that practitioners can consider when planning street tree sampling given the community’s size, capacity to inventory trees, and level of data quality needed for planning and management activities. In general, we suggest that municipalities may need to acquire larger samples than previously thought to achieve accurate estimates of citywide street tree characteristics, and smaller municipalities should conduct complete inventories when possible.