Methods for the estimation of sampling sufficiency in urban forest inventories: The case of non-patterned compositions of trees on sidewalks

Sampling inventories are strategies to gather qualified information for managing urban forests, given the scarcity of budgetary resources for a complete inventory and lack of public engagement to reduce costs. However, procedures for testing sampling sufficiency can be unspecified in researches related to urban forest inventories and do not follow any specific pattern. Hence, to determine the sampling sufficiency, we tested different variables related to the trunk, crown, number of trees, and species, focusing on different aims of an inventory of trees on sidewalks. At a level of 10% of the total number of plots, each measuring 50.0 m × 3.0 m, we performed a stratified inventory of a city streetscape whose composition and quality represents most South American cities, with a non-patterned tree compostion. Sampling sufficiency was analyzed considering a limit of error of 10% and 15% by using 12 different variables. The stratification process was necessary for most of the variables analyzed (p > 0.01), with errors ranging from 5.87% to 15.28%. Sampling sufficiency was achieved for 10% of the total population of trees on sidewalks, at a 10% error limit for seven variables: diameter at breast height (DBH), cross-section area, crown diameter, crown area, number of species, and number of species per square meter of sidewalk and per kilometer of the street. However, this result was influenced by the variability of the variables used to estimate sampling sufficiency. As it is not possible to achieve different goals (tree registration, benefits, and diversity) with just one variable like the number of trees per kilometer of street, the sampling sufficiency estimation should be based on the use of at least the DBH, crown diameter, number of trees, and number of species. It would be a better strategy to ensure more reliable data estimations for sampling inventories of trees on sidewalks.