Factors Influencing Establishment and Post-Establishment Residential Yard Tree Survivorship in Worcester, Massachusetts

Tree planting in cities and towns to expand existing canopy cover can provide a wide range of benefits (Czaja et al. 2020), including air temperature regulation (Ziter et al. 2019), rainwater interception (Dowtin et al. 2023), increased infiltration (Berland et al. 2017), and pollutant reduction (Grote et al. 2016). While it is generally thought that trees improve air quality in urban areas, they can also trap particulate matter at ground level in areas of high tree canopy cover, potentially increasing instances of respiratory ailments (Setala et al. 2013; Vos et al. 2013; Eisenman et al. 2019). In addition to environmental benefits, urban trees nonetheless have positive impacts on human health (Pataki et al. 2021). Regulating air temperatures in areas prone to extreme heat due to the urban heat island effect can decrease heatstroke, heat related morbidity, and heat related discomfort (Wolf et al. 2020). Urban trees can also increase the cultural and aesthetic value of an area and are associated with increased property values (Pandit et al. 2013). Increased urban canopy cover is associated with better overall health and faster stress recovery within the surrounding community (Jiang et al. 2016; Ulmer et al. 2016). Often, these health benefits are experienced unequally as marginalized, low income, and minority communities have less access to green spaces and experience adverse health outcomes as a result (Jennings and Johnson Gaither 2015).

Tree planting programs are implemented by government or nonprofit organizations to address inequities in urban tree canopy, achieve specific environmental goals (i.e., plant a certain number of trees), or as parts of larger urban greening plans (Nguyen et al. 2017; Riedman et al. 2022; Sousa-Silva et al. 2023). The environmental and health benefits that urban trees provide increase as trees age (Amer et al. 2023). Furthermore, these benefits can only be achieved if trees survive into maturity (Widney et al. 2016). As a result, tree survivorship and health are 2 of the most important metrics for evaluating the effectiveness of tree planting programs, although there has been insufficient monitoring and reporting to date of these metrics from tree planting programs (Roman et al. 2016; Nguyen et al. 2017).

In urban environments, there are many stressors that can impact tree health, and as a result, urban trees have been found to have a mortality rate 2 times higher than that of forest trees in natural succession (Smith et al. 2019). Environmental factors, such as drought, insect pests, and other environmental stressors, as well as human factors including management decisions, tree care, and vandalism, can compound as drivers of tree mortality (Hilbert et al. 2019). The first 1 to 5 years postplanting is known as the establishment phase, and during this period trees are especially vulnerable to mortality; however, there has been limited research comparing what factors influence survivorship in the establishment and post-establishment phases (Nowak et al. 2004; Roman et al. 2014).

Literature concerning factors associated with urban tree survivorship has focused on street trees, with fewer studies examining yard trees due to difficulties collecting data on private properties (Hilbert et al. 2019). Despite the lack of focus in the literature, trees on private property have been found to be integral to urban forest structure and composition and as a result are often targeted for planting in urban tree planting programs (Nguyen et al. 2017; Hutt-Taylor and Ziter 2022). There is great variety in the type of predictor variables included in analyses of urban tree survivorship, as the choice of variables is often tied to data availability and data collection during documentation of the planting process (Hilbert et al. 2019). Annual mortality rates for urban trees ranged between 3% and 7% in selected studies, however some studies do not report annual mortality rates (Table 1). Treelevel variables, such as taxon, species characteristics, and delivery type (e.g., balled and burlap, bare root) were commonly found to be the most important variables, however, site-level variables are important for investigating the broader patterns of tree survivorship.

Socioeconomic predictor variables have been widely used to assess tree survivorship. Commonly used variables include census block group variables such as income, renter proportion, and demographic makeup. These variables are rarely found to be impactful, but Ko et al. (2015) found that trees on properties in Sacramento, California, USA, with relative medium property value had highest survivorship. Roman et al. (2014) notes that neighborhoods with stable homeownership (owner occupied with no house sales) resulted in 4% higher annual survivorship rates, indicating that stable stewardship plays a large role in tree survivorship. While socioeconomic variables have rarely been found to be impactful in predicting planted tree survivorship, it is notable that the distribution of tree planting is often targeted in areas that have low levels of tree canopy resulting from legacies of urban planning (Hilbert et al. 2019; Riedman et al. 2022).

Some studies have also included more direct treecare variables to understand stewardship. Koeser et al. (2014) reported that irrigation increased survivorship by 3% to 24%, depending on species. Construction and renovation activity at the property can decrease survivorship by as much as 2% (Steenberg et al. 2018; Bigelow et al. 2024).

Factors relating to urban form have been used to explore urban tree survivorship as well. Findings from Minnesota indicate that trees near roads had higher survivorship than those in public parks (Wattenhofer and Johnson 2021). Research indicates that single family properties plant more trees than multiunit properties, and that those trees have survivorship rates approximately 1% higher (Steenberg et al. 2018). Additionally, examining tree planting in Massachusetts cities, Geron (2023) and Elmes et al. (2018) reported that trees on properties with buildings built before 1930 had higher survivorship. Both authors speculated that this represented some unmeasured factor about urban form, as urban form has been found to be more impactful than socioeconomic variables in studies on street trees (Pham et al. 2017).

In light of the importance of monitoring urban trees and their survivorship, it is clear that additional research is needed to understand dynamics of survivorship in tree planting programs. Tree planting programs in residential yards are critically important to urban forests but are underrepresented in the literature. Additionally, multiyear follow-up tree inventories are another under-researched topic important for the understanding of urban tree survivorship. This study addresses both issues, reporting on a residential tree survey and examining survivorship in the establishment and post-establishment phases of a tree cohort’s life. The research questions for this study are as follows:

1. How do site-level socioeconomic and biophysical factors impact the survivorship of yard trees?

2. Does the impact of site-level socioeconomic and biophysical factors differ in the establishment and post-establishment phases?

Study Area

The study area for this research project consists of 3 neighborhoods in the City of Worcester, Massachusetts, USA: Burncoat, Greendale, and North Lincoln Street (Figure 1). Worcester has a moist continental midlatitude climate, with an average of 1,219 mm of precipitation annually and average high temperatures of 26.5 °C in July and 0.1 °C in January (National Centers for Environmental Information 2023). Worcester has an average canopy cover of 34.9% as of 2015 due to large urban forests in the city, such as those in Green Hill Park and Broad Meadow Brook wildlife conservation refuge (Elmes et al. 2017).

The neighborhoods in the study are located within the US Department of Agriculture (USDA) regulation zone to control the spread of the Longhorned Beetle (LB; Asian Longhorned Beetle; Anoplophora glabripennis). The LB is a tree-boring insect that is invasive in North America with tree host species including those in the maple, elm, and willow families, where it is known to infest and kill trees (Haack et al. 2010). LB was detected by the USDA in Worcester in 2008, and in the years following over 30,000 trees were removed in a policy effort to eradicate the LB in Worcester County (Danko et al. 2016). Over 25,000 trees were removed from within the study area highlighted in this research (Palmer et al. 2014). As of 2015, the study area has an average canopy cover of 27.6% (Elmes et al. 2017).

• Download figure
• Open in new tab
• Download PowerPoint

Figure 1.

Study area map showing points representing trees included in analysis in the context of the Longhorned Beetle (LB) regulation zone and the state of Massachusetts.

In 2009, the Massachusetts Department of Conservation and Recreation (MA DCR) received a grant from the American Recovery and Reinvestment Act to implement a tree planting program in the LB regulation zone (Danko et al. 2016). The MA DCR’s goal was to restore urban canopy in response to negative resident perceptions of the LB removal policy and to address an immediate gap resulting from the urban canopy loss (Palmer et al. 2014). Trees were planted by MA DCR foresters and contractors on yards in residential, commercial, and public properties across the regulation zone. Because the program sought to plant as many trees as possible, the species makeup of the plantings relied on what was available at nearby nurseries (Massachusetts Department for Conservation and Recreation, personal communication).

For the trees sampled in this study, the 12-year overall mortality was 81% with an annual mortality rate of 3.18%, which was comparable to the median annual mortality rate of similar studies that tracked survivorship of a planted cohort, including street, yard, and park trees, post-establishment: 2.76% to 3.76% (Hilbert et al. 2019). The overall annual survivorship rate is also in line with other MA DCR planting programs (Breger et al. 2019; Geron 2023; Geron et al. 2023).

The building age variable was found to be significant in each of the logistic regression models in predicting tree survivorship. Findings indicate that trees planted on properties with buildings built before 1930 are 2 times more likely to survive than those on properties with buildings built after 1930, and just over 40% of trees inventoried were planted on these properties. This finding may be linked to changes in Massachusetts building codes restricting the construction of 3-story (triple decker) and multi-family housing after the 1930s (Wegmann 2006). The characteristics of these properties, such as amount of available planting space, soil quality, and building density, could impact tree survivorship on a property level. In the study area, buildings built before 1930 have larger lot sizes and planted trees have less distance from canopy, indicating that these properties have a larger footprint and more established canopy than newer properties. Buildings built before 1930 were more likely to be single family detached or multifamily, while buildings built after 1930 were more likely to be single family detached or single family attached. 23% of buildings built before 1930 were renter occupied, compared to 16% of buildings built after 1930. Building age was not found to be significantly related to property value, number of trees, or any other input variable. Other research on Massachusetts tree survivorship echoes these findings, where trees on properties with buildings built before 1930 have higher survivorship than those on properties built after 1930 (Elmes et al. 2018; Geron 2023). Research in Boston found that building style—which corresponds to year built—is related to canopy cover and canopy height in front and back yards, indicating that planted trees have different amounts of competing vegetation based on building age (Ossola et al. 2019). Additionally, findings in Christchurch, New Zealand, indicate that residents on recently redeveloped properties were more likely to remove trees due to development concerns (Guo et al. 2019). Further research is needed to assess the specific factors that influence the importance of this variable and verify if this factor is relevant in other urban areas with different histories of building practices.

The number of trees planted per property was ranked among the highest variable importance values for the establishment phase and overall phase and was significant in predicting tree survivorship in the establishment phase. This result is reflected in the literature as many authors have also found that the number of trees planted on a property is an important variable, where more trees per parcel results in lower survivorship (Ko et al. 2015; Elmes et al. 2018; Geron 2023). This is likely due to the importance of watering and care in the establishment phase and indicates that many residents have limited capacity for maintaining consistent watering of numerous trees. This has to be balanced with the distribution of urban trees in an area, as a sufficient number of trees should be planted in an area to achieve beneficial ecosystem services (Moody et al. 2021).

Distance from buildings was of high importance in the establishment, post-establishment, and overall phases but was only significant in the logistic regression models for the establishment and overall phases. For both phases, increasing distances from buildings resulted in lower survivorship rates, especially over 7 meters from a building. Research indicates that residents strongly value aesthetics of public-facing areas, and as a result residents may be more likely to care for trees in the front yard where they are more visible (Conway 2016; Locke et al. 2018). However, site type was of low importance in each of the CIT models, but trees on front yards did have the highest survivorship in the establishment phase (Figure 2b). Distance from impervious surfaces was found to be of high importance in the post-establishment phase, although it was not significant in the logistic regression model. This indicates that distance from buildings may indicate some combination of the land use, site type, and distance from impervious variables.

There has been very limited research comparing factors influencing survivorship over different phases of a tree cohort’s lifespan with the notable exception of Ko et al. (2015). However, Ko et al. (2015) do not focus on the comparative aspect of factors changing over time. The predictor variable number of trees per property was of highest importance in the establishment phase and only identified as significant in predicting survivorship in the establishment phase. This finding indicates that, as other authors speculated, residents may encounter difficulties caring for many young trees, which may lead to lower survivorship during the establishment phase (Ko et al. 2015; Elmes et al. 2018). This may explain why the number of trees has lower variable importance and is not significant in the post-establishment and overall models. Notably, building age remained at high importance for all phases, indicating that urban form characteristics impact tree survivorship equally as trees age.

For a tree to be found dead, it can either die naturally, of neglect, or be proactively removed by the resident. Motivations that encourage tree removal include fear of damage, poor tree health, and construction, although residents are poorly informed about trees and tree care (Conway 2016; Clark et al. 2020; Bigelow et al. 2024). Homeowners and renters have different levels of agency over properties, where homeowners feel more empowered to make landscaping decisions. Although owner occupied housing was not found to be important in any CIT model, trees on renter occupied housing had slightly higher annual survivorship rates. This indicates that homeowners remove more trees, despite benefiting from increased maintenance and investment a homeowner may provide.

Previous findings in Toronto and Philadelphia found construction on the street and property levels to be related to removal of street and front yard trees, although only statistically significant for street trees (Steenberg et al. 2018; Bigelow et al. 2024). Although presence of one or more constructions permits was not found to be significant or of high importance across any phase, trees on properties with construction permits had lower survivorship. This indicates that the effect of removals due to construction may be less impactful on yard trees than it is on street trees.

The MA DCR has refined its procedures for tree siting and especially tree species selection since the planting cohort examined in this study (Massachusetts Department for Conservation and Recreation, personal communication). More research is needed on monitoring urban tree survivorship as the practices of urban tree planting organizations gain experience and adapt to local circumstances. Future studies would also benefit from more thorough data collection at time of planting, such as recording condition, stock type, and caliper width at planting.

Establishing robust watering programs for tree planting programs on residential properties could mitigate the decrease in survivorship seen when more than 4 trees are planted on one property, as watering commitments have been shown to increase tree survivorship (Mincey and Vogt 2014). While ideal tree placement varies with respect to site specific characteristics, additional research could find how distance from physical attributes impacts survivorship (Wu et al. 2008). More work is also needed to uncover the causal relationship behind the importance of building age with respect to urban tree survivorship in Massachusetts and elsewhere, and how building stock and urban form impact tree survivorship. Continual monitoring of trees could improve knowledge about tree care and gather additional data, allowing for further investigation into the importance of the variables reported here, especially building age, distance from buildings, and number of trees planted on a property.

This study examines the survivorship of a residential yard tree cohort in the establishment and postestablishment phases. The findings indicate that number of trees planted per property is important in predicting survivorship in the establishment phase, and building age is important in predicting survivorship in both the establishment and post-establishment phases. Despite the statistical relationships in the data, it is difficult to ascertain the mechanism that relates site-level variables such as building age to tree survivorship. In light of this ambiguity, more research is needed to relate site-level variables to tree survivorship and understand the mechanisms through which these variables influence survivorship. Additionally, there is a lack of long-term follow-up research on tree planting cohorts, and this study shows the importance of comparing factors influencing survivorship across different phases of a tree cohort’s lifespan.

The authors reported no conflicts of interest.