Conceptualizing the human drivers of low tree diversity in planted urban landscapes

Despite the abundance of tree diversity in the natural world, and generally high tree species richness in urban areas, urban forests continue to be dominated by a limited number of species. As socio-ecological systems, urban forests are shaped by historical and current management efforts and decision-making across a wide range of human actors. Drawing on past research, we offer a conceptual framework for describing the complex interactions among tree producers and consumers as trees are selected, grown, specified, and planted in private and public urban areas. We illustrate how multiple layers of selection criteria filter down the entirety of potential local tree diversity to a handful of commonly used and accepted tree species. We detail the actors and decision-makers who impact tree composition and diversity across several land types. Finally, we identify research, education, and outreach needs as they relate to creating more diverse and resilient urban forest ecosystems.


INTRODUCTION
As urban areas expand in size and as human populations continue to concentrate within them (Seto et al. 2017;United Nations 2019), scholars are increasingly interested in how urban ecosystems develop and function (Sukopp 1997;McDonnell 2015;Jones and Instone 2016;McPhearson et al. 2016;Pickett et al. 2020).Urban ecosystems, including urban forest systems, are shaped by current human actions, as well as the legacies of past decisions (Roman et al. 2018) and can be described as socio-ecological systems (Vogt 2020).To provide humans with ongoing ecosystem services, these dynamic urban landscapes must be resilient to biotic and abiotic stresses (McPhearson et al. 2015).Resilience is an ecosystem's ability to recover from or adjust to disturbances (Holling 1973), and biodiversity is often viewed as key requirement of resilience (Alvey 2006;Ordo ´n ˜ez and Duinker 2012;McPhearson et al. 2015;Steenberg et al. 2019;Huff et al. 2020).
In this article, we focus on planted landscapes that exist in many urbanized contexts in cities, suburbs, and small towns (i.e., not just in downtown urban cores), which can be broken down into several different land types (Table 1).Specifically, we focused on the planted urban landscape in which tree selection and population demographics are anthropogenically controlled (Roman et al. 2016), and we outline the human actors who participate in tree selection and procurement.When approaching urban forestry from a management perspective, as we do in this manuscript, it is useful to expand the conception of urban forests beyond trees to include ''the associated biotic and abiotic components, including people, institutions, and infrastructure'' as part of an integrated socio-ecological system (Ostrom 2009;Steenberg et al. 2019;Vogt 2020).To develop management strategies to ensure the resilience of urban forest systems, research is needed on the relationships between humans and tree species (Loreau et al. 2001;Pickett et al. 2011;Pett et al. 2016).Specifically, there is a basic management need for information regarding the subset of tree species that are suitable for stressful urban growing conditions, produce maximal desired ecosystem services, limit disservices (Roman et al. 2020), and are compatible with changing climatic conditions in a given locale (Esperon-Rodriguez et al. 2022).Legacies of past monocultures have made urban forests vulnerable to tremendous losses from pests and pathogens (Campanella 2003;Poland and McCullogh 2006), yet urban forest managers continue to face challenges in bolstering diversification of planting palettes (Hilbert et al. 2023).Although urban foresters have recognized for decades that taxonomic diversity of planted species boosts urban forest resilience (Santamour 1990;Ball and Tyo 2016), most urban communities rely on a small number of species which dominate the total count of public trees (Lohr et al. 2016;Ma et al. 2020;Galle et al. 2021).With increasing attention on urban tree diversity and compositional patterns from both scholars and practitioners (e.g., Ordo ´n ˜ez and Duinker 2012; Jenerette et al. 2016;Lohr et al. 2016;Nitoslawski et al. 2016;Steenberg et al. 2017), there is a need to advance conceptual understandings of the actors and drivers shaping urban forests.
We draw on a range of sources from the fields of horticulture, planning, design, urban forestry, and urban ecology to identify the key human actors whose decisions influence tree composition and diversity in planted urban landscapes.We present a conceptual framework to explain how these actors' tree selection decisions perpetuate low urban forest species diversity.Our framework is largely drawn from the context of the continental United States (US) and Canada, although past literature clearly shows that taxonomic diversity challenges exist in cities throughout the world.

THE PROBLEM OF LOW URBAN TREE SPECIES DIVERSITY
Urban forests are comprised of all the publicly and privately owned trees within an urban area (Nowak et al. 2010) and are shaped by past and present biophysical and social processes (Roman et al. 2018;Vogt 2020).Tree species diversity includes the total amount of different tree species in an urban area (i.e., species richness) and the relative proportions of those different species (i.e., evenness).Composition (i.e., the identities of species and other taxa) is also important for resilience, particularly in light of pests and diseases impacting multiple hosts from unrelated taxa (Lac ¸an and McBride 2008).Many urban areas have high tree species richness due to an abundance of nonnative species (Aronson et al. 2015;Gillespie et al. 2017).This is demonstrated in studies comparing tree species diversity in city centers to adjacent peri-urban and rural areas (Ku ¨hn et al. 2004;Blood et al. 2016;Jha et al. 2019).However, the evenness of urban tree ecological communities is often low due to the dominance of a few species within a given city or neighborhood (Lohr et al. 2016;Wang and Zhang 2022)-this over-reliance on a few species is the problem of low urban tree species diversity that we address in this paper.The overuse of planted tree species poses a challenge to managers seeking to improve overall taxonomic diversity of the urban forest.Furthermore, while urban tree species diversity is generally discussed as a means of increasing resiliency to emerging biotic and abiotic threats, there are additional benefits associated with ecosystem functioning and provisioning for wildlife that must also be considered (Ordo ´n ˜ez and Duinker.2012).
The challenge of low urban tree diversity due to an overreliance on a few species threatens urban forests across the globe.In a global assessment of 108 urban tree inventories from around the world, on average, a single species made up 20% of a given city's tree population (Lohr et al. 2016).In Helsinki (Finland) and Bangkok (Thailand), 40% of the urban forest was represented by one tree species, common linden (Tilia 9 europaea) and angsana (Pterocarpus indicus), respectively (Lohr et al. 2016).Galle et al. (2021) noted that tree species diversity appears to be the most limited near the most heavily developed cores of cities.In Amsterdam (The Netherlands) for example, elms (Ulmus spp.) accounted for nearly half of the trees (47%) in the city center, compared to 11% of the trees in more suburban quarters of the city (Galle et al. 2021).A similar association between taxonomic diversity increasing with increasing distance from city center was found in Beijing (China; Jiao et al. 2021).While there are some instances of species diversity increasing over time (Nitoslawski and Duinker 2016;Cowett and Bassuk 2021), it is also the case that some communities have experienced declining diversity (Sjo ¨man et al. 2012a).
A limited number of tree species and genera dominate urban forests in the US and Canada, as well (Lohr et al. 2016;Cowett and Bassuk 2017;Ma et al. 2020;Galle et al. 2021).Historically, uniform planting of a single species through entire urban neighborhoods was viewed as desirable both esthetically and in terms of management convenience, with taxa such as American elm (Ulmus americana) and London planetree (Platanus 9 hispanica) dominant in street tree plantings in northeastern US cities (Du ¨mpelmann 2019; Roman and Eisenman 2022).The legacies of past monocultures create time-lagged vulnerabilities to pest and disease outbreaks (Greene and Millward 2016).
Overall, low tree species evenness increases the potential for greater losses due to an over-reliance on a few taxa.Furthermore, species-level diversity assessment can be problematic.When species diversity is the primary metric of diversity and higher taxonomic relationships are not considered, or are considered to a lesser degree, this exacerbates the potential threats to the urban forest because many of the most damaging introduced pests and diseases impact plants at the genus or family level (Morgenroth et al. 2016).For example, emerald ash borer (Agrilus planipennis) typifies a genus-level threat.Following the widespread loss of American elm to Dutch elm disease (Ophiostoma ulmi and Ophiostoma novo-ulmi), ash trees (Fraxinus spp.) were planted throughout temperate urban landscapes in the US and Canada.Between 2009 and 2019, emerald ash borer affected an estimated 37.9 million ash trees in urban communities in the eastern US, with an estimated cost of US$10.7 billion in management expenses (Kovacs et al. 2010).Financial losses from emerald ash borer extended to nursery producers as well.In Michigan, nurseries experienced US$11.6 million in damages and restricted sales due to reduced demand for the trees as a result of the invasive pest (Herms et al. 2004).This is just one example of a contemporary pest disaster resulting from low urban forest species diversity, and accompanying challenges for the nursery trade.
The Dutch elm disease crisis around the 1970s, which decimated American elm populations in cities throughout the US and Canada (Campanella 2003), actually spurred the development of the ''urban forest'' as a concept (Dean 2008;Roman et al. 2018), in that managing a collection of urban trees for pathogens required a system-wide consideration of the entire forest, as opposed to single-tree arboricultural treatments.Since the late twentieth century, urban forest researchers and managers have proposed various guidelines for managing urban forest taxonomic diversity.For instance, Santamour's (1990) oft-cited 10-20-30 rule suggests that if managers want to limit deforestation due to pests and disease, a given community of trees should be comprised of no more than 10% of a single species, 20% of a single genus, or 30% of a single family by stem count.Despite widespread recognition of the importance of species diversity to forestalling major tree losses, a few species dominate the urban forests of many cities (Ma et al. 2020).For example, across 188 communities throughout the continental US, only six species accounted for the majority (61.5%) of a given city's street trees, and the single most common species in a given region had a mean abundance of 14% to 23% (Ma et al. 2020).These results mirror the findings of a study of 275 urban tree inventories in New York, Pennsylvania, and New Jersey, which found that Norway maple (Acer platanoides), a known invasive in this region, accounted for over 16% of street trees (Cowett and Bassuk 2017).Moreover, maples (Acer spp.) accounted for nearly 39% of the aggregated population across the tree inventories in those tree states (Cowett and Bassuk 2017).Maples also dominate in Toronto, Ontario, and other Canadian cities, due in part to political symbolism of maples in Canada (Vander Vecht and Conway 2015;Roman et al. 2018).With multi-host pests such as Asian long-horned beetle (Anoplophora glabripennis) and shothole borer (Scolytus rugulosus) threatening other regions of the US and Canada (Berland and Hopton 2016;Rabaglia et al. 2019), strategies to diversify at multiple taxonomic levels, and invest in underutilized species, have become even more important to reduce urban forest vulnerabilities (Lac ¸an and McBride 2008;Hilbert et al. 2022).
The limited urban tree species diversity revealed in the aforementioned studies is a world-wide problem and is the product of a long chain of ecological, social, and economic constraints associated with tree biology, nursery production, site-specific demands, and final adoption by endusers, that is, actors selecting trees for public spaces (Conway and Vander Vecht 2015;Kabrel 2016;Nitoslawski et al. 2016).To remain competitive, tree growers are constrained by biology; the realities of producing a quality, marketable product of slow-growing plants; and the complexities of consumer demand (Thompson et al. 2021;Hilbert et al. 2023), which do not fully reflect the risks associated with continued reliance on over-used species in the planted landscape.High profitability and low economic risk motivate tree producers to favor fast-growing, easy-to-manage, and high-demand trees (Hilbert et al. 2023).These business realities diminish the palette of available urban trees, resulting in production systems and landscapes that are less diverse and thus less resilient to pests and pathogens, as well as to abiotic pressures (e.g., climate change) (Lohr et al. 2016).As most trees require three to ten years to reach marketable size, this relatively long-term investment is a financial burden for many producers (Warren 1990;Burcham and Lyons 2013).Consequently, there exists little economic reward for experimenting with species perceived as less familiar, slower growing, or for which current market demand is low or questionable.At the end of the process, major purchasers of trees for urban tree planting (e.g., municipal arborists, landscape designers working with developers) are left with limited options that often do not reflect the taxonomic diversity that is ecologically possible for a given region (Hilbert et al. 2023).

THE PLANTED URBAN LANDSCAPE
Trees in the planted urban landscape (as opposed to trees in urban forest natural areas) are subject to intensive human control over species composition and community structure, with trees typically arising from planting decisions with minimal natural regeneration (Roman et al. 2014(Roman et al. , 2022)).The planted urban landscape includes trees along streets, in parking lots, and other hardscape settings, as well as trees in lawns and other manicured ground cover at neighborhood parks, institutional settings, and residential landscapes (van Doorn et al. 2020).The arboricultural best practices for planted tree care include pruning, mulching, weeding, plant health care interventions, and planned removals.It can be particularly challenging to grow trees in highly developed landscapes because of conditions like altered soils, drought stress, pollutants, and disruptions to nutrient and water availability (Miller et al. 2015;Roloff 2016).In this article, we draw on multidisciplinary sources to propose a conceptual framework that demonstrates how human decisions and associated limitations influence tree composition and species diversity in planted urban landscapes.Our discussion focuses primarily on the process of selecting trees through nursery production and eventual planting in the built landscape, but acknowledges the indirect effects some actors, such as urban planners, can have on the built landscape and resulting taxonomic composition.

CONCEPTUAL FRAMEWORK: HOW HUMANS PRODUCE URBAN TREE SPECIES DIVERSITY
Urban tree selection, both with regard to what is produced in the nursery and what is planted in the landscape, is a long process involving multiple decision-makers.The incorporation of a tree species into a planted landscape is typically preceded by a long process that begins with local, regional, and global plant exploration and the testing of the viability of the species as a commercial product (Stage 1: Germplasm and Propagule Supply) (Sjo ¨man et al. 2012b;Jones 2016), then leads to nursery production (Stage 2: Market Availability) (Avolio et al. 2018), inclusion in landscape design plans and specifications (Stage 3: Site Design) (Conway and Vander Vecht 2015;Thompson et al. 2021), and the eventual planting of the tree (Stage 4: Planted Trees) (Fig. 1).The different stages in this process are driven by the decisions made by different actor groups, each limited to varying degrees by the decisions made by those in preceding stages (Fig. 1, Table 2).Furthermore, there is a feedback loop that occurs when those who purchase and plant trees (Stage 4) send a consumer signal to the tree growers (Stage 2) who in turn base stock decisions on sales.Those at the end of the process are greatly limited by availability, and even if they want a diverse set of species to select from, their purchasing may be directed toward the available tree species, as opposed to the desired alternatives.''Stage 3: Site Design,'' refers not only to landscape architects/designers, but the planners and managers who are responsible for creating landscaping codes, developing planting lists, and making tree recommendations.Some actors in Stage 3 may also operate within Stage 4, and vice versa, but our conceptual framework organizes them based on the stages within which they predominately operate.
Key actors' decisions also affect tree species diversity at different spatial scales in the urban landscape (Thompson et al. 2003(Thompson et al. , 2021;;Cook et al. 2012;Shakeel and Conway 2014;Steenberg et al. 2015;Yang et al. 2015).As outlined in Table 2, this can include household, neighborhood, municipal, and regional levels (Table 2).In addition to acting on different stages in the selection process and different spatial scales, human decision-makers also impact different built-up land types-either directly, indirectly, or both, depending on their role in the actual selection, procurement, or installation of trees (Table 3).The actors listed toward the top of Table 3 tend to have indirect effects on the resulting tree species diversity within all the land types, since they are farther removed from the actual decision of ''what tree gets planted here.''Those involved with landscape design and the creation of tree planting lists and recommendations can have more complex relationships with the diversity of land types since their species recommendations may or may not be implemented in the landscape when the time to plant comes (i.e., what was planted was not what was recommended, or recommended lists were not used).Arborists, landscape contractors, urban foresters, and homeowners active in tree procurement and planting have the most direct interactions with the resulting observed tree species diversity.

SUMMARY
The focus of this framework was on the actors involved in tree selection, procurement, and planting, and on the limitations and decision-making processes connected to the various actors.As demonstrated by the conceptual framework and our summary of the supporting literature, urban forest species diversity is undermined by the established system of tree production and procurement.This is concerning given that taxonomic diversity is a key facet of resiliency in the face of threats from introduced pests and climate change (Berland and Elliot 2014;Brandt et al. 2016).By focusing on the highly managed urban landscape-where new trees are by and large the result of human planting decisions-we created a focused framework that can guide future research and transdisciplinary collaborations regarding urban tree species diversity in a particularly challenging landscape for trees.We suggest the following future research topics to address the different stages within the framework: • Stage 1: Germplasm and propagule supply: discovery and trialing of underutilized species for urban use, Research findings from each of these topical areas should be paired with industry-aimed outreach and education efforts (i.e., extension).Extension efforts should ideally bring together the various industry groups Fig. 1 The selection funnel demonstrating how cumulative limitation factors and decisions by human actor groups at each stage in the tree supply chain (Stages 1-4) reduce the diversity of trees that are perceived as suitable for urban areas from a broader pool of possible locally or globally sourced tree taxa (adapted from Krabel 2016).Furthermore, a consumer demand feedback loop occurs when tree purchasers (Stage 4) select from the available stock, signaling growers (Stage 2) to continue to produce that subset of tree species that sell well  (2020), Cubino et al. (2020) associated tree production and purchasing but reaching each audience through their existing networks individually is also important.Conceptual frameworks such as the one we lay our here are meant to be revised and expanded upon as new knowledge and scenarios are discovered (Jabareen 2009).As such, this framework should guide future work by researchers examining different ecological, economic, and governance contexts to bring more understanding to why we have low urban tree species diversity in planted landscapes and what managers can do to intervene in order to make urban forests more resilient to emerging abiotic and biotic threats.
Table 3 The authors' interpretation of the different built-up urban land types and the human actors whose decisions impact tree species selection and resulting urban forest species diversity.Actors with a direct role in tree species selection actively decide which species will be used.Actors with an indirect role in tree species selection impact species selection through other means (e.g., education, policy creation/enforcement, site design/modification)

Residential
particularly with respect to climate change (e.g.,McPherson et al. 2018;Hilbert et al. 2022).•Stage 2: Market availability: nursery production studies to bring underutilized species to market; economic policy studies on incentives that reduce the risks growers face in introducing new stock; case studies of successful procurement arrangements (e.g., Stephens 2010).• Stage 3: Site design: social studies on the knowledge, priorities, and concerns of the different actor groups in relation to tree species selection and urban tree diversity; development of urban landscape management interventions that reduce site limitations and positively influence tree survival and health.• Stage 4: Planted trees: social studies on the knowledge, priorities, and concerns of the different actor groups in relation to tree species selection and urban tree diversity (e.g., Cubino et al. 2020).
d = direct role in tree selection, = indirect role in tree selection; = either direct or indirect role depending on situation; -= no role

Table 1
Built-up urban land types on which trees are actively planted and managed (adapted fromNitoslawski et al. 2016).Based on urban forest systems in the continental US and Canada

Table 2
Key human actors whose decisions impact tree selection and resulting urban forest species diversity.The stages correspond to the planted tree diversity conceptual framework (Fig.1).Spatial scale refers to the scale at which the actors predominately operate