Abstract
Diversity in tree populations is viewed as essential for protecting the public investment in urban trees and for preserving the environmental, social, and economic benefits that these trees provide. It is therefore crucial for officials responsible for the management of municipal trees to know the diversity of their municipal tree populations and whether their efforts to increase diversity have been effective or should be modified. We assessed street tree diversity in New York State, USA by analyzing municipal street tree inventory data from two data sets, the first comprised of 75 inventories collated from municipalities, and the second comprised of 32 sets of inventories conducted at multiple points in time. This analysis builds on two previous papers containing similar assessments by analyzing more current data and by calculating diversity index statistics and relative abundance percentages for prevalent street tree species and genera. Findings indicate that there has been substantial progress to increase street tree diversity in New York State. This progress is correlated with reductions in the dominance of Norway maple (Acer platanoides), the state’s most prevalent street tree species (17% of street trees statewide), and in the dominance of maple (Acer), the state’s most prevalent street tree genus (35% of street trees statewide). Work remains to be done to further increase species and genus diversity so as to meet the challenges posed to municipal street tree populations by invasive pests and climate change. Strategies are proposed for accomplishing this.
INTRODUCTION
Biodiversity contributes to the stability of biological systems in the face of environmental change (Ives and Carpenter 2007). In the field of urban forestry, and particularly with respect to municipally managed street and park trees, diversity in tree populations is viewed as fundamental to the planning and management required to protect the public investment in urban trees and to preserve the environmental, social, and economic benefits trees provide (Morgenroth et al. 2016). Tree diversity was not always emphasized as a management imperative in the United States. The City Beautiful movement at the beginning of the 20th century advocated for uniform plantings of the same tree species along streets to increase a city’s “stateliness, impressiveness, and charm” (Robinson 1901). Such planting practices, which date back to the garden allées of 16th century Europe (Lawrence 1988), may have been aesthetically pleasing, but, by encouraging the overplanting of certain tree species, they rendered large proportions of municipally managed trees vulnerable to an insect or disease (Raupp et al. 2006). This vulnerability was demonstrated by the devastation wrought, beginning in the 1930s, by Dutch elm disease (DED, Ophiostoma ulmi) on the American elm (Ulmus americana), which had been planted in large numbers as a street tree in many municipalities (Campanella 2003). More recently, the emerald ash borer (EAB, Agrilus planipennis) has decimated many tree-lined streets, especially in the Midwest, where ash (Fraxinus spp.) were often planted to replace American elms (Poland and McCullough 2006); the Asian longhorned beetle (ALB, Anoplophora glabripennis) has threatened trees in the Northeast, where maple (Acer spp.), one of its favorite hosts, is the most prevalent street tree genus (Cowett and Bassuk 2017); the spotted lanternfly (SLF, Lycorma delicatula) prefers feeding on tree-of-heaven (Ailanthus altissima), but feeds as well on maple, cherry (Prunus spp.), and apple (Malus spp.)(Urban 2020); and oak wilt (Bretziella fagacearum) kills many species of oak (Quercus spp.), with trees in the red oak family especially susceptible (Juzwik et al. 2011). Climate change is expected to additionally endanger urban trees by increasing temperatures (Lohr et al. 2016) and by changing the patterns and intensity of precipitation and drought (Easterling et al. 2000; David et al. 2018). Reducing reliance on overly abundant tree species susceptible to drought or soil inundation could mitigate the risk of future tree loss due to climate change (McPherson et al. 2018).
Diversity is not a panacea for enhancing the survivability of urban tree populations. Street trees in particular are vulnerable to a host of environmental stressors (Quigley 2004), and diversity does not insure against the mortality associated with development, vandalism, and harmful maintenance practices (Steenberg et al. 2017). Moreover, some tree species are more resistant to harsh urban conditions and better candidates for survival (Roloff et al. 2009), which explains in part their abundance relative to other tree species and limits to some extent the choices available for increasing diversity (Richards 1993; Watson 2017). Finally, planting decisions made today for increasing diversity may prove unsuccessful in the future with the onset of a new invasive pest or disease (Greene and Millward 2016). Notwithstanding these caveats, officials tasked with setting urban forest policy and priorities on national, state, and municipal levels have recognized the importance of diversity in municipal tree populations and promoted increased diversity where populations are insufficiently diverse. For example, the National Urban and Community Forestry Advisory Council (NUCFAC 2015) has associated increased diversity with greater resilience in urban and community tree populations and supported the use of site-appropriate species adaptable to climate change and resistant to insect and disease damage; the Maryland Department of Natural Resources (2015) has encouraged the “expansion of [the] species palette” to improve urban forest health and survival; and the City of Portland, Oregon (2020) has advocated for planting a diversity of trees as “an important step toward creating a healthier and more resilient urban forest.” Moreover, in Vancouver, Canada, a study of urban forestry practitioners found that “urban tree diversity” and “physical access to nature” were rated as the two most significant indicators of a healthy and resilient urban forest (Barron et al. 2016).
If the health and stability of the urban forest are correlated, at least in part, with tree diversity, then it is crucial for officials responsible for the management of municipal trees to know the diversity of those tree populations. In addition, if efforts have been made to increase diversity where tree populations were found insufficiently diverse, it would be helpful to compare diversity metrics at multiple points in time in order to understand the extent to which these efforts have been effective and if strategies need to be changed. In the United States, many studies have analyzed municipal tree diversity at a single point in time for a particular municipality (Sanders 1981; Maco and McPherson 2003), and some studies have done the same for multiple municipalities (Raupp et al. 2006; Subburayalu and Sydnor 2012) or on a statewide basis (Ball et al. 2007; Cumming et al. 2008; McPherson et al. 2016). Fewer studies have analyzed municipal tree diversity at multiple points in time for one or more municipalities (Dawson and Khawaja 1985; Lockwood and Berland 2019), still fewer have done this on a statewide basis (Gartner et al. 2002), and, to the best of our knowledge, only Ma et al. (2020) have done so on a national basis in the United States. We examine street tree diversity on a statewide basis in New York State by analyzing street tree inventory data obtained from municipalities in the state. This analysis has two components. The first is a statewide assessment that builds on two previous statewide assessments (Cowett and Bassuk 2014; Cowett and Bassuk 2017) by analyzing more current data. The second is an assessment limited to those municipalities in the state for which data collected at multiple points in time are available and directly comparable. Based on these assessments, we evaluate diversity in the state’s street tree population and consider whether diversity appears sufficient to sustain the environmental, social, and economic benefits that these trees provide.
MATERIALS AND METHODS
Study Area
New York State is located in the Northeast of the United States (Figure 1). It is the 30th largest state with a land area of 122,056 km2 (47,126 square miles), the 4th most populous state with an estimated population of 19,453,561, and the 7th most densely populated state with a population of 158.76 per km2 (411.2 per square mile)(United States Census Bureau 2019). The state is divided into 57 counties which are subdivided into cities, towns, and Indian Reservations; the 5 boroughs of New York City (Bronx, Brooklyn, Manhattan, Queens, Staten Island) each comprise an additional county (NYSDOS 2018). Towns may contain villages and hamlets; a village is an incorporated named place with defined boundaries, and a hamlet is an unincorporated named place without defined boundaries (NYSDOS 2018). Most of the state is located in the Eastern Temperate Forest ecoregion, and a smaller portion, containing the Adirondack and Appalachian Mountains, is located in the Northern Forests ecoregion; both ecoregions are characterized by a severe, mid-latitude, humid continental climate with warm summers and cold, snowy winters (CEC 2009). Average annual precipitation is over 1,016 mm (40 inches), but this varies statewide, with mountainous regions receiving more than 1,270 mm (50 inches), and the western part of the state receiving less than 1,016 mm (40 inches) (NOAA 2016). USDA Plant Hardiness Zones range from Zone 3B (−37.2 to −34.5 °C, −35 to −30 °F) in the Adirondack Mountains in the north, to Zone 7B (−14.9 to −12.3 °C, 5 to 10 °F) in New York City and Long Island in the south (USDA 2012). Climate change is projected, on its current trajectory, to increase average annual temperatures across the state by 1.12 to 1.91 °C (2.0 to 3.4 °F) in the 2020s, 2.30 to 3.81 °C (4.1 to 6.8 °F) by the 2050s, and 2.97 to 5.66 °C (5.3 to 10.1 °F) by the 2080s; temperature increases are expected to be accompanied by increases in late-summer short-duration droughts and in the strength and frequency of intense precipitation events (NYSERDA 2014).
Location of New York State, USA.
Data Sets
Municipal tree inventories were obtained from 71 municipalities to facilitate a statewide assessment of street tree diversity. None of these inventories were included in the 2014 and 2017 statewide street tree assessments (Cowett and Bassuk 2014; Cowett and Bassuk 2017). However, 28 of the 71 municipalities were represented in the 2014 and 2017 statewide assessments by inventories conducted at earlier dates. These 28 municipalities are therefore associated with newer and older inventories conducted at multiple points in time. The mean and median year of the newer inventories is 2017, and the mean and median year of the older inventories is 2005. These inventories are directly comparable and comprise a data set from which the trend of street tree diversity can be ascertained for each municipality.
Between the 71 municipalities in the statewide assessment, 22 inventories were obtained from cities, 13 inventories were obtained from towns, and 36 inventories were obtained from villages. The New York City street tree inventory was apportioned to the city’s 5 boroughs (Bronx, Brooklyn, Manhattan, Queens, Staten Island), predicated in part on a series of diversity t-tests (Hutcheson 1970), which found statistically significant differences (P < 0.05) for street tree species and genus diversity between all boroughs, with the lone exception of genus diversity between Brooklyn and Queens. The portion of New York City street trees contained in each borough has been treated as a unique inventory. Therefore, the number of municipal tree inventories in the statewide assessment is 75 (Figure 2). Between the 28 municipalities with multiple inventories, 12 municipalities are cities and 16 municipalities are villages, including New York City. Because, as stated above, each New York City borough has been treated as a unique inventory, there are 32 municipalities with newer and older inventories for which the trend of street tree diversity can be ascertained (Figure 2).
Street tree inventories obtained in New York State.
Most inventories obtained for this paper were conducted by professional arborists employed by tree inventory firms. Some inventories were conducted by soil and water conservation district staff, and some were conducted by nonprofessional volunteers, such as citizen scientists and students. Inventory data collected by persons with different levels of expertise and experience have been found to vary in accuracy due to observer error. Bloniarz and Ryan (1996) found that nonprofessional inventory volunteers agreed with professional arborists 94% on tree genus identification and 80% on tree species identification; Ball et al. (2007) found that master gardeners and other community volunteers misidentified 1.2% of tree species compared to professional foresters; and Roman et al. (2017) found that citizen scientists agreed with experts (i.e., urban forest researchers and certified arborists with prior tree inventory experience) 90.7% on tree genus identification and 84.8% on tree species identification. Concern with tree data accuracy has led some urban forest researchers to modify their methods and analysis. For example, Lockwood and Berland (2019) had sufficient concern with possible tree species misidentification that it contributed to a decision to limit results to the genus level.
In addition to variability in observer expertise and experience, data obtained for this paper are not standardized, not only between those inventories conducted by the tree inventory firms and those conducted by others, but also between the inventories conducted by tree inventory firms. In particular, there are differences among the inventories in tree classification at the species level. For example, some inventories classify all species in a genus, such as Amelanchier, as Amelanchier spp.; some inventories classify all members of a genus subset, such as ornamental cherry trees, as the species of that genus (e.g., Prunus spp.); and some inventories classify all trees by species. Variability may exist as well in the geographic extent and completeness of the inventory. For example, state highways may or may not be inventoried, maintenance responsibilities for some streets may change between governmental agencies, and some areas may not be surveyed due to budget or time constraints. Data sets frequently specify latitude and longitude GPS coordinates for each tree, which enables the extent for most, but not all, inventories to be known. The great majority of the inventories obtained for this paper were complete inventories with GPS coordinates. Finally, some inventories include not only street trees, but park trees. Previous studies have found the dynamics and population structures of street trees and park trees to be significantly different, including their species composition and diversity (Welch 1994; Nielsen et al. 2007; North et al. 2018). This suggests that street tree diversity is best assessed independently of park tree diversity. Accordingly, for municipalities where street trees and park trees were both inventoried, and where it was practicable to do so, street trees were differentiated from park trees and analyzed separately. Park trees have not been analyzed.
Ideally, all inventories obtained for this study would have been conducted by the same personnel utilizing the same inventory collection methods and protocols, especially for those municipalities for which data were collected at multiple points in time. Doing so would have maximized data accuracy, made data sets fully comparable, and improved the validity of findings. That this did not happen is unfortunate, but not surprising, since it is common in urban tree inventories that different approaches are used at different times by different personnel (Morgenroth and Östberg 2017). In other words, everybody does things differently. However, where we had sufficient concern about data accuracy and variability potentially impacting the validity of findings, especially for municipalities with multiple inventories, those data were excluded from assessment.
Data Analysis
Whereas the municipalities with newer and older inventories are directly comparable, the 75 inventories in the statewide diversity assessment comprise a non-random data sample with the potential for selection bias that could reduce the accuracy of any findings, especially in comparison to previous statewide assessments. The 2014 and 2017 assessments (Cowett and Bassuk 2014; Cowett and Bassuk 2017) both employed post-stratification of data and weighting with auxiliary information to correct for selection bias (Bethlehem 2010). Data were stratified by USDA Plant Hardiness Zones and weighted by measures of street length contained within those zones. The 2014 assessment stratified data with the 1990 version of the zones and created weighted measures from New York State ALIS (Accident Location Information System) street centerline files. The 2017 assessment stratified data with the 2012 version of the zones and created weighted measures from United States Census Bureau TIGER-Line All Roads files. This statewide assessment used the same post-stratification and weighting steps as the 2017 assessment and the same weighted measures (Table 1). Inventories were assigned to a 2012 Plant Hardiness Zone based on location of the municipality’s inner centroid (i.e., a geometrically calculated center point within municipal boundaries). Zone 3 is sparsely populated and not associated with any municipal tree data, and tree species and genera hardy in Zone 3 are not meaningfully different from those hardy in Zone 4 (Dirr 1998). Accordingly, Zones 3 and 4 were aggregated, and this is reflected in the weighting of data. Between the 75 inventories in this statewide assessment, 3 inventories are located in Zones 3 and 4, 25 inventories are located in Zone 5, 36 inventories are located in Zone 6, and 11 inventories are located in Zone 7 (Figure 2).
Summed selected street length contained within the 2012 USDA Plant Hardiness Zones in New York State.
In assessing the 32 municipalities with multiple inventories, data were not stratified and weighted with auxiliary information to correct for selection bias. Nevertheless, the distribution of these municipalities with respect to the 2012 Plant Hardiness Zones is worth noting. Between the 32 municipalities with newer and older inventories, 2 sets of inventories are located in Zones 3 and 4, 10 sets of inventories are located in Zone 5, 14 sets of inventories are located in Zone 6, and 6 sets of inventories are located in Zone 7 (Figure 2).
Relative abundance percentages of street tree species and genera in relation to the street tree population as a whole (i.e., species and genus composition) were calculated for each municipal tree inventory. Such percentages are commonly utilized as a metric for diversity. Santamour (1990) proposed that no tree species should exceed 10%, no tree genus should exceed 20%, and no tree family should exceed 30% of a municipal tree population in order to guard against large-scale devastation by a pest or disease. Santamour’s 10-20-30 rule is not without critics, particularly since it offers less protection against a polyphagous pest, such as the ALB, that attacks a host range wider than a single tree species (Watson 2017), but it has achieved wide acceptance by urban forest managers for providing benchmarks for diversity (Kendal et al. 2014). For this statewide assessment, relative abundance percentages for each municipality were allocated to the municipality’s 2012 Plant Hardiness Zone. Group means for tree species and genera percentages were calculated for each zone. These means were then weighted by the measures derived from United States Census Bureau TIGER-Line All Roads files to create weighted statewide percentages according to the formula:
where m1, m2, m3, and m4 denote the group means (i.e., means for species and genus relative abundance percentages in the 2012 USDA Plant Hardiness Zones) and w1, w2, w3, and w4 denote the different weights for each group (i.e., the relative percentages of statewide street length contained within those zones). For the 75 inventories in this statewide assessment, comparisons were then made between these weighted statewide relative abundance percentages and those reported in the 2014 and 2017 statewide assessments. Comparisons were also made for the 32 municipalities with multiple inventories to ascertain any trends in relative abundance percentages for prevalent street tree species and genera. As stated above, data for this assessment were not stratified and weighted with auxiliary information to correct for selection bias.
In addition to relative abundance percentages, diversity index statistics were calculated for each municipal street tree population. These indices, which are frequently used to make comparisons between biological populations, typically consider more than relative abundance and include additional factors, such as population size and the number of species and genera in that population, in their calculation. Simpson’s Diversity Index (SDI)(Simpson 1949) and the Shannon-Wiener Diversity Index (Shannon 1948) are two diversity indices often utilized in urban forest research. The SDI is sometimes preferred because it is more sensitive to population evenness (i.e., how evenly the members of a population are distributed between all the species and genera in that population) and gives less weight to rare species and genera; the Shannon-Wiener Diversity Index is more sensitive to species and genera richness (i.e., the number of species and genera in a population) and to sample size (Colwell 2009). Because the SDI measures dominance (i.e., the greater the SDI statistic, the greater the dominance level), the Inverse SDI (1/SDI) is sometimes preferred to the SDI as a measure of diversity (i.e., the greater the Inverse SDI statistic, the greater the diversity level)(Sun 1992). Additionally, because the Shannon-Wiener Diversity Index is logarithmic, effective diversity, which is the exponential of the Shannon-Wiener statistic, or eH where H is the Shannon-Wiener statistic, has also been utilized in urban forest research, since it produces statistics that are not logarithmic and are therefore more directly comparable (Jost 2006).
Statistics for the Inverse SDI, the Shannon-Wiener Diversity Index, and effective diversity were calculated as measures of tree species and genera diversity. Separate statistics for population evenness (Buzas and Gibson 1969) were also calculated due to the importance of evenness in more fully understanding the diversity of tree populations. For example, even if Population A has a greater number of tree species and genera than Population B, if a sufficient number of the species or genera in Population A are represented by only a few trees, then the most prevalent species and genera in Population A can be more dominant than the most prevalent species and genera in Population B, and Population B would be more diverse. Table 2 summarizes these diversity statistics, which were calculated with PAST (PAleontological STatistics) software version 4.2 (Hammer et al. 2001). Comparisons were made between diversity statistics for the 75 inventories in this statewide assessment and those reported in the 2014 and 2017 statewide assessments (Cowett and Bassuk 2014; Cowett and Bassuk 2017). Statistics for evenness and effective diversity were not reported in the 2014 assessment, and statistics for effective diversity were not reported in the 2017 assessment, but these statistics were calculated from the data associated with those papers. For the 32 municipalities with multiple inventories, a diversity t-test (Hutcheson 1970) assessed the statistical significance of change (P < 0.05) for Shannon-Wiener Diversity Index values between newer and older inventories.
Measures of tree species and genera diversity.
RESULTS
Species and Genus Composition
For this statewide assessment, Norway maple (Acer platanoides) was found to be the most prevalent street tree species, with a weighted statewide percentage of 17.64% (21.38% unweighted), and maple (Acer spp.) was found to be the most prevalent street tree genus, with a weighted statewide percentage of 35.35% (41.82% unweighted)(Table 3). These findings are similar to those in the 2014 and 2017 statewide assessments in which Norway maple and maple were also found to be the most prevalent street tree species and genus (Cowett and Bassuk 2014; Cowett and Bassuk 2017). However, the percentage of Norway maple relative to all street tree species was found to be 14.58% less prevalent compared to its prevalence in 2014, and the percentage of maple relative to all street tree genera was found to be 19.90% less prevalent compared to its prevalence in 2014 (Table 3). Sugar maple (Acer saccharum), silver maple (Acer saccharinum), and London planetree (Platanus × acerifolia) were street tree species also found to be less prevalent between this statewide assessment and the 2014 assessment, and planetree (Platanus spp.) and ash (Fraxinus spp.) were street tree genera also found to be less prevalent between this statewide assessment and the 2014 assessment. Conversely, honeylocust (Gleditsia triacanthos), Callery pear (Pyrus calleryana), crabapple (Malus spp.), and littleleaf linden (Tilia cordata) were street tree species found to be more prevalent between this statewide assessment and the 2014 assessment, and oak (Quercus spp.), pear (Pyrus spp.), linden (Tilia spp.), and cherry (Prunus spp.) were street tree genera found to be more prevalent between this statewide assessment and the 2014 assessment. The 6 most prevalent street tree species statewide accounted for 43.19% of all street trees in this statewide assessment as compared to 52.57% of all street trees in the 2014 assessment, and the 6 most prevalent street tree genera statewide accounted for 62.58% of all street trees in this assessment as compared to 71.37% of all street trees in the 2014 assessment.
Statewide relative abundance percentages (weighted) for street tree species and genera in New York State.
Consistent with these results are results for the 32 municipalities with inventories conducted at multiple points in time (Table 4). In these municipalities, for both the newer and older inventories, Norway maple was found to be the most prevalent street tree species, and maple was found to be the most prevalent street tree genus. Additionally, between the newer and older inventories, the mean percentage of Norway maple declined from 21.97% to 17.10% (a 22.14% reduction), and the mean percentage of maple declined from 47.54% to 35.86% (a 24.57% reduction). Other street tree species exhibiting declines in these municipalities include sugar maple, silver maple, and green ash (Fraxinus pennsylvanica), and other street tree genera exhibiting declines include planetree and ash. Conversely, street tree species exhibiting increases include honeylocust, Callery pear, crabapple, and northern red oak (Quercus rubra), and street tree genera exhibiting increases include oak, cherry, spruce (Picea spp.), and pear. The 6 most prevalent street tree species accounted for 55.83% of all street trees in the newer inventories and 67.59% in the older inventories (a 17.40% reduction). The 6 most prevalent street tree genera accounted for 72.49% of all street trees in the newer inventories and 80.87% in the older inventories (a 10.36% reduction).
Relative abundance percentages for street tree species and genera in municipalities with multiple inventories.
Despite less prevalence found for Norway maple and maple in the 75 inventories comprising this statewide assessment as compared to the 2014 statewide assessment, the weighted statewide percentage of Norway maple for the 75 inventories exceeded Santamour’s 10% rule for species, and the weighted statewide percentage of maple exceeded his 20% rule for genus. Of the 75 inventories, 73 (97.33%) exceeded the 10% rule for species, and 72 (96.00%) exceeded the 20% rule for genus. In nearly all cases, this was attributable to the abundance of Norway maple and maple. However, in a few cases, honeylocust and sugar maple exceeded 10% of all street tree species, and honeylocust, oak, and planetree exceeded 20% of all street tree genera. Similarly, for the 32 municipalities with inventories conducted at multiple points in time, despite reductions found for the prevalence of Norway maple and maple, the mean percentage of Norway maple and maple for the newer inventories exceeded Santamour’s 10% rule for species and his 20% rule for genus, and, in nearly all cases, this was attributable to the abundance of Norway maple and maple.
Diversity Statistics
Statistics were generated for street trees at species and genus levels for Simpson’s Diversity Index, the inverse of the Simpson’s Diversity Index (Inverse SDI), the Shannon-Wiener Diversity Index, and Jost’s effective diversity. For the 75 inventories comprising the statewide assessment, the mean Inverse SDI was found to be 12.36 for species and 5.54 for genus; the mean Shannon-Wiener Diversity Index was found to be 3.03 for species and 2.26 for genus; and mean effective diversity was found to be 22.93 for species and 10.54 for genus (Table 5). These are all larger values signifying greater diversity than the values reported for or calculated from the data analyzed for the 2014 and 2017 statewide assessments (Cowett and Bassuk 2014; Cowett and Bassuk 2017). For the 32 municipalities with multiple inventories, for the newer inventories, the mean Inverse SDI was found to be 13.95 for species and 6.68 for genus; the mean Shannon-Wiener Diversity Index was found to be 3.17 for species and 2.41 for genus; and mean effective diversity was found to be 25.62 for species and 11.98 for genus (Table 5). These are all larger values signifying greater diversity than the values found for the older inventories. A diversity index t-test (Hutcheson 1970) found the differences in the Shannon-Wiener Diversity Index to be statistically significant (P < 0.05), with the exception of one village for species and genus and one village for species alone.
Diversity statistics for street tree species and genera.
Statistics were generated as well for the evenness of street tree species and genus distributions. For the 75 inventories comprising the statewide assessment, evenness values (Buzas and Gibson 1969) for the species and genus distributions were found to be less than the values reported for or calculated from the data analyzed for the 2014 and 2017 statewide assessments (Table 5). In addition, for these inventories, the Shannon-Wiener Diversity Index and effective diversity at the species level were found to be correlated more with the number of species than with evenness, and the Inverse SDI, Shannon-Wiener Diversity Index, and effective diversity at the genus level and the Inverse SDI at the species level were found to be correlated more with evenness than with the number of species and genera (Pearson’s R, P < 0.05)(Table 6). For the 32 municipalities with multiple inventories, evenness values (Buzas and Gibson 1969) were found to have increased for the species and genus distributions between the newer and older inventories (Table 5). In addition, for the newer inventories, the Inverse SDI at the species level and Inverse SDI and Shannon-Wiener Diversity Index at the genus level were found to be correlated more with evenness than with the number of species and genera, and the Shannon-Wiener Diversity Index at the species level and effective diversity at both the species and genus levels were found to be correlated more with the number of species and genera than with evenness (Pearson’s R, P < 0.05)(Table 6).
Correlations for Inverse SDI (Inverse of Simpson’s Diversity Index), Shannon-Wiener Diversity Index, and effective diversity (eH), and number of species and genera, number of trees, and distribution evenness for street trees (Pearson’s R, P < 0.05).
Finally, diversity statistics were correlated with relative abundance percentages for prevalent street tree species and genera. At the species level, for the 75 inventories comprising the statewide assessment, and for the newer inventories from the 32 municipalities with multiple inventories, the percentages of Norway maple were found to be negatively correlated with the Inverse SDI, the Shannon-Wiener Diversity Index, effective diversity, and evenness (Pearson’s R, P < 0.05)(Table 7). Conversely, for the 75 inventories comprising the statewide assessment, the percentages of London planetree and pin oak (Quercus palustris) at the species level were found to be positively correlated with the Inverse SDI, effective diversity, and evenness; for the newer inventories from the 32 municipalities with multiple inventories, the percentage of London planetree was found to be positively correlated with the Inverse SDI, effective diversity, and evenness, and the percentage of Japanese tree lilac (Syringa reticulata) was found to be positively correlated with the Inverse SDI, the Shannon-Wiener Diversity Index, effective diversity, and evenness. At the genus level, for the 75 inventories comprising the statewide assessment, and for the newer inventories from the 32 municipalities with multiple inventories, the percentages of maple were found to be negatively correlated with the Inverse SDI, the Shannon-Wiener Diversity Index, effective diversity, and evenness (Pearson’s R, P < 0.05)(Table 7). Conversely, for the 75 inventories comprising the statewide assessment, and for the newer inventories from the 32 municipalities with multiple inventories, the percentages of oak and cherry were found to be positively correlated with the Inverse SDI, the Shannon-Wiener Diversity Index, effective diversity, and evenness.
Correlations for Inverse SDI (Inverse of Simpson’s Diversity Index), Shannon-Wiener Diversity Index, effective diversity (eH), and evenness with relative abundance percentages for prevalent street tree species and genera (Pearson’s R, P < 0.05).
DISCUSSION
Statewide relative abundance percentages between this statewide assessment and the 2014 statewide assessment (Cowett and Bassuk 2014) indicate fewer Norway, sugar, and silver maples relative to other street tree species, and fewer maples relative to other street tree genera. Caution should be exercised in characterizing these findings as declines, since there are meaningful differences between the statewide assessments: a majority of the municipalities are not common to both assessments, the auxiliary information used to stratify and weight data (i.e., Plant Hardiness Zones and statewide road files) is not precisely the same, and there are fewer inventories in this assessment, both overall and in Zones 3 and 4 and Zone 7, compared to the 2014 assessment. However, these differences do not apply to the municipalities with multiple inventories, and comparisons made between the newer and older inventories yield findings similar to the statewide findings, with 12 years on average between the inventories. Therefore, between the statewide assessments and the municipalities with multiple inventories, it does appear that the percentages of Norway, sugar, and silver maples relative to other street tree species and the percentage of maples relative to other street tree genera have declined in New York State. Moreover, for the municipalities with multiple inventories, the number of Norway, sugar, and silver maple street trees declined, as did maple street trees more generally, while the overall number of street trees increased.
Statewide statistics for the Inverse SDI, the Shannon-Wiener Diversity Index, and effective diversity indicate that the values calculated from the inventories obtained for this statewide assessment are greater than the values calculated from the inventories associated with the 2014 and 2017 statewide assessments (Cowett and Bassuk 2014; Cowett and Bassuk 2017). For some of the same reasons applicable to the statewide relative abundance percentages, caution should be exercised in characterizing these findings as increases. However, for the municipalities with multiple inventories, values for the Inverse SDI, the Shannon-Wiener Diversity Index, and effective diversity are greater for the newer inventories than for the older inventories. Therefore, between the statewide assessments and the municipalities with multiple inventories, it does appear that diversity values for street tree species and genera have increased in New York State. Given that Norway maple and maple are the most prevalent street tree species and genus, and that their relative abundance percentages were found to be negatively correlated with statistics for the Inverse SDI, the Shannon-Wiener Diversity Index, and effective diversity (Table 7), it is reasonable to attribute at least part of the increase in diversity values to the decline in the percentages of Norway maple relative to other street tree species and to the decline in maple relative to other street tree genera. However, the mixed results for evenness (Table 6), particularly at the species level, where some diversity statistics were found to be correlated more with the number of species and genera than with the evenness of the species or genera distribution, suggest that, although there have been increases in the number of street tree species and genera in street tree populations (Table 5), these increases do not represent sufficiently large numbers of trees. As a result, Norway maple continues to be the dominant street tree species, and maple continues to be the dominant street tree genus (Tables 3 and 4), and nearly all the 75 inventories comprising the statewide assessment and the 32 sets of newer and older inventories fail to meet Santamour’s 10% and 20% benchmarks for species and genera diversity.
Therefore, more work appears to be required to further increase diversity. Should those state and municipal officials responsible for street tree management keep doing what they are doing, or should they be doing some things differently? The annual rate of decrease for Norway maple found in the municipalities with multiple inventories suggests that, if that rate remains constant, the percentage of Norway maple for these municipalities could decline to 10% in 18 years. Similarly, the annual rate of decrease for maple found in the municipalities with multiple inventories suggests that, if that rate remains constant, the percentage of maple for these municipalities will decline to 20% in 16 years. However, these estimates may be optimistic, since they are based on a subset of municipalities in the state comprising a nonrandom data sample with the potential for selection bias that could reduce their accuracy. For example, since the prevalence of maple and Norway maple was found previously to vary geographically within the state (Cowett and Bassuk 2014), these estimates may or may not accurately reflect this geographic variability. Furthermore, Santamour’s benchmarks for species and genus diversity may be widely used, but there is no scientific basis confirming their efficacy (Kendal et al. 2014), and, since polyphagous pests such as the ALB and SLF attack multiple tree species and genera rather than a single species or genus, they may not be stringent enough (Laćan and McBride 2008). Accordingly, Bassuk et al. (2009) have recommended limiting any one tree species to between 5% and 10% of an urban tree population; Ball and Tyo (2016) have proposed a 5% rule for tree genera and placed greater emphasis on genus diversity than on species diversity, since many pests function at the genus level; and Simons (2018) has suggested not only lower limits on genera percentages, but also tailoring these percentages to the USDA Plant Hardiness Zones, with colder zones having greater limits (e.g., 10% for genera in Zone 3) than warmer zones (e.g., 3% to 4% for genera in Zone 7).
In addition to the issues described above, there are practical difficulties to be faced in continuing to diversify New York State’s street tree population. As stated previously, not all tree species are good candidates to be street trees, since some tree species are less resilient to streetscape conditions and environmental stressors (Roloff et al. 2009). Some less commonly used tree species might be resilient to streetscape conditions and environmental stressors, and would therefore be good street tree candidates, but sourcing these trees can be a problem due to limited nursery availability (Sydnor et al. 2010; Conway and Vecht 2015). The choice of street tree species will be narrowed further if emphasis is placed on planting native trees to the exclusion of non-native trees (Sjöman et al. 2016). Notwithstanding the findings made by Barron et al. (2016) cited earlier, the importance of tree diversity is not universally recognized by municipal and nursery personnel (Polakowski et al. 2011; Lohr 2013), factors other than diversity may be prioritized when making planting decisions (Conway and Vecht 2015), and preferences for commonly planted tree species are persistent (Simons and Hauer 2014). Finally, smaller-sized municipalities have a knowledge and resource gap relative to larger-sized municipalities, and this gap can contribute to the continuation of established planting practices (Doroski et al. 2020).
Given the complexities involved, continuing the diversification of New York State street trees, if not increasing it to levels more ambitious than those suggested by Santamour, will require not just intensified effort, but coordinated action due to the multiple actors (e.g., property owners, arborists, tree boards, public utilities, nurseries, state and local officials) and geographic scales (e.g., parcel, block, neighborhood, municipality, region, state) involved in street tree management (Clark et al. 1997; Mincey et al. 2013). For example, Hilbert et al. (2020) have called for state and Extension officials to work with individual municipalities to identify underutilized tree species (i.e., less than 1% of the total tree population by stem count) that are readily available and appropriate choices based on a matrix of tree and site characteristics; Avolio et al. (2018) have encouraged ecologists who are working with planners, nurseries, and the public to build support for tree diversity by coupling resident preferences for key plant attributes, such as shade provision, with scientific information when making planting recommendations; Hirons et al. (2020) would like arboreta and botanic gardens to collaborate with nurseries and the plant-user community to research and provide new trees appropriate for urban landscapes; and Doroski et al. (2020) have proposed establishment of networks and funding programs targeted at smaller-sized municipalities to remedy planting inequities and knowledge gaps relative to larger-sized municipalities.
We also see opportunities for coordinated action specific to New York State and in particular the urban and community forestry program administered by the Department of Environmental Conservation (NYSDEC). Norway maple is categorized by the NYSDEC as a regulated invasive species, meaning it cannot be knowingly introduced into a “free-living state” (i.e., public lands or lands connected to public lands, natural areas, and public waters or waters connected to public waters), but its sale and use as a street tree is still allowed (NYSDEC 2014). Categorizing Norway maple as a prohibited invasive species and phasing out future sales in the state, as has been done in Massachusetts (MDAR 2021), would expedite its decline as a prevalent street tree. Additionally, the NYSDEC provides funding for municipal urban and community forestry projects, which includes grants for a community forest management plan (CFMP) as well as for tree planting as long as a street tree inventory has previously been conducted. The request for applications (RFA) for the 2019 round of these grants contained a single reference to increasing tree species diversity, which was buried in an appendix for a CFMP sample work plan (NYSDEC 2019). A comparable grant program in New Jersey stipulated that tree planting should include a diversity of tree species and follow Santamour’s 10-20-30 rule for the long-term health and stability of community trees (NJDEP 2017). The New York State grant program should adopt comparable language emphasizing street tree diversity and should integrate tree diversity into the scoring criteria for consideration of future grant proposals. It could also follow the example of the Tennessee Community Tree Planting Program by calling out tree species deemed invasive, undesirable, or vulnerable to pests and disease, and explicitly deny funding for the planting of these species (Tennessee Department of Agriculture 2020). Finally, North (2018) has suggested that municipalities enter into contract growing agreements with nurseries to ensure availability of less commonly planted tree species, as was done in New York City (Stephens 2010). However, such agreements typically require a level of arboricultural and planning expertise exceeding the resources and abilities of smaller municipalities. Assistance with such expertise and planning could be provided by the state’s regional foresters and by its Cooperating Forester program. Also, since it already operates a nursery to produce tree and shrub seedlings for conservation plantings on public and private lands, the NYSDEC could possibly perform some contract growing itself.
CONCLUSION
Street tree diversity in New York State was evaluated by analyzing street tree inventory data obtained from municipalities in the state. This analysis had two components: the first, a statewide assessment comprised of 75 inventories that built on two previous statewide assessments (Cowett and Bassuk 2014; Cowett and Bassuk 2017), and, the second, an assessment limited to 32 municipalities where inventory data had been collected at multiple points in time (i.e., newer and older inventories). Based on these assessments, we believe that progress has been made to increase the diversity of street tree populations in New York State and that this progress is likely to continue with the continued support of national, state, and municipal officials. However, there is no guarantee that diversity will increase to targeted levels, and it is unlikely that diversity alone will be sufficient to protect street tree populations from the challenges posed by polyphagous pests and climate change (Berland and Hopton 2016; Dale and Frank 2017). Nevertheless, we believe that increasing street tree diversity is essential to protect the public investment in these trees and to sustain the environmental, social, and economic benefits they provide. Recommendations have been made to accomplish this. Additionally, there is a need for periodic and systematic monitoring designed to provide the information necessary for state and local officials responsible for municipal tree management to assess on a statewide basis not simply diversity, but the health of tree populations, and to more fully understand the extent to which current management efforts are effective or if these efforts need to be changed. Finally, some thought should be given to the aesthetic consequences of increased tree diversity on municipal streetscapes. The City Beautiful movement may have erred from an ecological perspective in advocating uniform street tree plantings, but visual aesthetics remain a significant factor in community attitudes and decisions concerning street trees. For example, “beauty” was the reason most cited by the residents of 6 Australian cities for planting trees (Kirkpatrick et al. 2012), “pleasing to the eye” was the highest-rated reason for valuing urban trees given by residents of New Haven, Connecticut (Locke et al. 2015), and Australian municipal tree managers gave more weight to visual and aesthetic benefits than to environmental benefits in selecting street tree species (Roy 2017). Integrating increased street tree diversity with the aesthetics of tree selection is worthy of consideration so as to provide visual coherence and avoid the “menagerie effect” (Cregg 2014).
Footnotes
Conflicts of Interest:
The authors reported no conflicts of interest.
- © 2021, International Society of Arboriculture. All rights reserved.
LITERATURE CITED
Jump to section
Related Articles
Cited By...
- No citing articles found.