178 annual mortality rates from planting cohort studies (Table 2) would be around 7 to 11 years, 13 to 18 years, and 33 to 38 years for worse-than-normal, mid- dle-of-the-road, and better-than-normal survivorship scenarios (Figure 5). It is important to note that these curves are not meant for extrapolation to other cities and planting programs, but rather to display a general trend in the survivorship curves derived from reported mortality rates in the published literature. Factors Associated with Mortality Quantitative Associations with Mortality for Field-Based Monitoring Studies The literature cited a variety of statistically significant factors associated with mortality (Table 4 and Appen- dix Table 3). Of the articles that used field-based mon- itoring studies and quantitatively examined factors, the most commonly cited biophysical factors were size/age and taxa (e.g., species or cultivar). The most commonly cited human-related factors associated with urban tree death were stewardship, maintenance, and vandalism. Land use and socioeconomic mea- surements (e.g., net property value, income, and homeownership) were also commonly cited as signif- icant. Trees typically experienced higher mortality when located on properties with unstable homeown- ership (such as rental, foreclosed, and vacant proper- ties) and when located in neighborhoods with lower incomes and property values. Stewardship and main- tenance activities play a positive role in tree survival. Three studies did not find any statistically significant relationships between observed mortality rates and factors examined quantitatively (Thompson et al. 2004; Conway 2016; Martin et al. 2016). The studies that quantitatively analyzed predictors of mortality share many associated factors, but differ on whether some common factors increase or decrease mortality rates, with sometimes contradic- tory results. For example, Roman et al. (2014b) asso- ciated smaller mature tree size with lower mortality, while Ko et al. (2015a) associated it with a higher mortality when compared to medium-sized mature trees, yet both focus on the same residential lawn tree program in Sacramento. It is unclear why these two studies found different results; other issues related to those taxa or planting sites could be more relevant than mature tree stature. Land use, while cited as sta- tistically significant in many studies (Table 4), does not have a clear mechanism of impact on mortality, ©2019 International Society of Arboriculture Hilbert et al: Urban Tree Mortality: A Literature Review and land use categories may also covary with other important factors. For instance, Nowak et al. (2004) reported high mortality for transportation land use, but this land use had a high prevalence of Ailanthus altissima, which the authors speculated might explain that finding. Single-homeowner properties had decreased mortality rates in several studies (Nowak et al. 1990; Lu et al. 2010; Jack-Scott et al. 2013), but the causal explanations for this association are unclear. Various land uses may reflect different plant- ing site conditions and/or maintenance regimes across studies. For example, a tree recorded as single-family residential could alternatively be a yard or street tree, and trees recorded as multi-family residential could be managed by the municipality (such as street trees adjacent to a downtown apartment building) or by landscaping crews hired by an apartment manager (such as lawn trees in a suburban apartment com- plex). In general, it is not clear whether land use is associated with mortality due to biophysical charac- teristics of particular land use categories and planting sites, governance of tree stewardship, or some other phenomena. Most older studies (e.g., Nowak et al. 1990; Miller and Miller 1991) tested for significant influential fac- tors using univariate statistical techniques such as Chi-Square tests, but some more recent studies also relied on univariate techniques (e.g., Nowak et al. 2004; Lu et al. 2010). In contrast, most studies after 2010 (e.g., Staudhammer et al. 2011; Lawrence et al. 2012; Roman et al. 2014a; Koeser et al. 2014; Ko et al. 2015a; Vogt et al. 2015a; van Doorn and McPher- son 2018) used more sophisticated multivariate anal- yses such as logistic regression and non-parametric conditional inference trees. For instance, Roman et al. (2014a) examined the interaction between tree condition and size class using multivariate logistic analysis. When summarizing the papers, no factors stood out as being related to a specific size or life stage in the papers we reviewed, with the possible exception of maintenance or stewardship for recently planted trees. The prevalence of the most common factors used in quantitative analyses could be due to their relative ease of measurement, since such data can be gathered from planting records, quick field evaluations, and the United States census. Multiple studies followed protocols outlined by the United States Forest Ser- vice i-Tree Eco or Forest Inventory and Assessment
September 2019
| 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
