138 Ordóñez et al.: Influence of Abiotic Factors in a Commercial-Retail Streetscape the differences in proportions between discrete variables (e.g., bed and pit planters) using contingency coef- ficients (e.g., Pearson’s χ2 Contingency analysis was employed to explore tree condition ratings between ) to determine if there was a difference (Jutras et al. 2010). The mag- nitude of this difference was then measured using Phi (φ) and Cramer’s measures, for both binomial and multinomial factors, respectively. Given the limited availability of data for some variables, correlation, MANOVA, and contin- gency analyses were applied either to the sub- sample (n = 57), or the full data set (n = 133), depending on available data. Modelling tech- niques, such as logistic regression, were tested but later discarded due to small sample size and inconclusive results. All data in these analyses were tested for normality with the Shapiro-Wilk’s and Kolmogorov–Smirnov tests, and for homo- geneity of variance with the Levene’s Median test. Follow-up univariate ANOVAs, t-tests, and non-parametric Mann-Whitney tests were per- formed on the continuous variables that were identified as significant in MANOVA. Comparison of means between groups in multinomial variables (i.e., tree condition) was performed by the Tukey’s HSD post hoc and non-parametric Kruskal- Wallis tests. Given the limited availability of data for some variables, these tests were performed either on the full (n = 133) or the subsampled (n = 57) data sets. All statistical analyses were carried out using a 95% confidence level (critical P-value < 0.05) in R (v. 3.3.2) and IBM® SPSS® (v. 23) soſtware. Only significant results are reported. RESULTS The 133 trees were planted in soils of a sandy- clay-loam texture (averaging: sand = 65%, silt = 14%, clay = 21%, by weight). The soil in the pit planters had moderate levels of compaction, never exceeding 2 MPa. The distance from the planting locations to the nearest intersection ranged be- tween 6.2 and 129 m (mean = 54.3 m, SE = 2.9), with 45 trees located in closest proximity to a ma- jor intersection, and 88 to a minor one. There were 71 alive and 62 dead trees at the time of removal in 2015 (i.e., over four years, a mortality rate of 46.6%), whereas 64.5% of the dead trees were already as- sessed as dead in 2014. The majority (79%) of trees removed in 2015 displayed some form of human- caused damage. Soil at the site had generally high pH (mean = 8.2, SE = 0.1) and elevated Na (mean = 685.3 ppm, SE = 72.6), but a low EC content (mean = 0.61 dS/m; SE = 0.04). Finally, the site displayed variable sunlight availability, receiving more sun- light hours in the summer solstice (mean = 5.3 hr, SE = 0.2) compared to the spring equinox (mean = 1.6 hr, SE = 0.1). The growing season’s range of sunlight per site varied between 0 and 8 hours (mean = 3.7 hr, SE = 0.2), and averaged seasonally between 0.5 and 7 hours (mean = 3.4 hr, SE = 0.1). Analyses indicated that there were strong or moderate correlations between some soil indica- tors and tree metrics, particularly between Na, EC, and Mg; and Ca, pH, and DBH (Table 1). There were statistically significant differences in the distribution of dead and alive trees and tree condition and tree damage, with a moderate or strong association (φ or Cramer measure >0.2; Table 2). However, there were no differences between mortality and type of planter (Tables 2). The variations in Mg, Na, EC, and DBH were statistically significant according to tree mortal- ity (Table 3) and tree condition patterns (Table 4), as were variations in Mg, Na, pH, and EC in bed and pit planters (Table 5). Finally, there were statistically significant differences in the way sunlight was distributed across tree planters on the north and south sides of the street (Table 6). As noted before, the temporal analyses between yearly tree-condition ratings and climatic data were inconclusive and are not reported. Table 1. Results from the correlation analysis of tree and soil characteristics of the Bloor Street trees, in Toronto, Ontario, Canada, indicating correlation coefficients, with significant values in bold (analysis performed on subsample, n = 57). Variable Correlation coefficient Mg Na Ca DBH EC Mg Na pHz 0.18 0.29 -0.58 n/a -0.63 -0.27 -0.15 -0.45 0.78 -0.63 n/a 0.31 Ca n/a 0.38 0.08 0.18 -0.15 0.49 pHz 0.49 0.08 0.05 -0.27 0.31 n/a EC 0.08 -0.32 n/a -0.58 0.78 0.05 DBH 0.38 n/a -0.32 0.29 -0.45 0.08 z Coefficients refer to Spearman’s Rho, since the variable is not normally dis- tributed (i.e., Shapiro-Wilk’s and Kolmogorov–Smirnov tests P-value < 0.05) ©2018 International Society of Arboriculture
May 2018
| Title Name |
Pages |
Delete |
Url |
| Empty |
Search Text Block
Page #page_num
#doc_title
Hi $receivername|$receiveremail,
$sendername|$senderemail wrote these comments for you:
$message
$sendername|$senderemail would like for you to view the following digital edition.
Please click on the page below to be directed to the digital edition:
$thumbnail$pagenum
$link$pagenum
Your form submission was a success. You will be contacted by Washington Gas with follow-up information regarding your request.
This process might take longer please wait
