Abstract
Background Use of vegetation in urban areas for climate change adaptation is becoming increasingly important; however, urban vegetation is itself vulnerable to the effects of climate change. Better understanding of which species will survive and thrive in urban areas with projected climate change will increase confidence in choosing climate-ready species for resilient urban greening outcomes. Plant selector tools based on the suitability of species for future climates, however, are lacking.
Methods The Which Plant Where plant selector webtool (www.whichplantwhere.com.au) was created by combining sophisticated species distribution models and trait and environmental tolerance data from a variety of sources to allow users to select appropriate species which are climatically suitable for Australian urban environments for 3 different time periods (2030, 2050, and 2070). The tool allows users to calculate co-benefits afforded by planting palettes and offers suggestions for alternative species based on climate suitability to help diversify plantings and provide options where substitutions may have to be made.
Results The tool contains information for over 2,500 unique plant entries (encompassing species, subspecies, cultivars, varieties, and hybrids) from 9 different growth forms (trees, shrubs, palms, ferns, cycads, climbers, succulents, grass, and herbs). The tool contains many resources to design and maintain resilient urban green spaces, from the planning stage up to monitoring and maintenance.
Conclusion Which Plant Where was designed to allow practitioners and urban forest managers to confidently identify climate-ready species now to ensure urban green spaces remain diverse and resilient into the future.
Introduction
The world’s climate is changing at an unprecedented rate, with accelerated change predicted in the near future (IPCC 2022). The effects of climate change threaten natural ecosystems as well as urban centres, where 68% of the world’s population is expected to reside by 2050 (DESA 2018). A common adaptation strategy to mitigate the effects of climate change in many cities around the world is to increase the amount of urban tree canopy cover (e.g., doubling tree canopy, planting 5 million trees)(City of Toronto 2008; City of Melbourne 2013; Bristol One City 2019; New South Wales Government 2022). This is because urban vegetation can help to increase thermal comfort through the provision of shade and evapotranspiration (Dimoudi and Nikolopoulou 2003). However, the fact that urban vegetation is vulnerable to the effects of climate change has more recently come to light (Ordóñez and Duinker 2015; Brandt et al. 2016; Burley et al. 2019; Khan and Conway 2020; Esperon-Rodriguez et al. 2021). Urban landscapes are often disproportionately affected by climate change due to the urban heat island effect and exacerbated effects of drought and heatwaves due to impervious surfaces, small root space, soil compaction, and poor management (Sæbø et al. 2003; Gillner et al. 2017). As climate change continues to intensify, urban landscapes may become increasingly unsuited for the current species palette. It is, therefore, imperative that cities begin to adapt their planting palettes to species that will be able to thrive in predicted future climates.
Various studies have predicted that climate change may have a deleterious effect on the suitability of urban areas for the survival of plant species currently found there. For example, Yang (2009) found that the future climate in Philadelphia, PA, USA would be unsuitable for 10 out of 60 tree species assessed. Similarly, Burley et al. (2019) predicted that suitable urban habitat might decline for 73% of native tree species planted across Australia’s urban areas by 2070. On a global scale, more than 13,000 species will be at risk from projected changes in temperature by the year 2050 (Esperon-Rodriguez et al. 2021). However, the deleterious effects of climate change on urban vegetation are already beginning to be observed. For example, Zhang and Brack (2021) found that only 37% of street trees surveyed in Canberra, Australia showed no visible stress symptoms (e.g., dead branches, crown dieback, etc.) compared with 80% of trees surveyed in 1999. Other street tree surveys have also found significant damage to foliage following increases in the intensity and frequency of droughts and heatwaves in urban areas associated with climate change (Tabassum et al. 2021a; Haase and Hellwig 2022). Despite these negative impacts, some species are expected to benefit by gaining climatically suitable habitats in urban areas in the future (Yang 2009; Burley et al. 2019). Identifying which species are likely to thrive in future climates is paramount to increasing the resilience and sustainability of urban green spaces.
Although choosing climate-ready species is important for increasing the resilience of urban green spaces to threats such as climate change, ensuring that we plant a diverse range of climate-ready species is also critically important (Kendal et al. 2014). However, most urban green spaces are characterised by low diversity (Nowak 1994; Bourne and Conway 2014; Tabassum et al. 2020). Urban green spaces with low plant diversity are more vulnerable to climate change due to a reduced capacity for adaptation and an increased vulnerability to species-specific threats such as pests and pathogens. For example, the popularity of the American elm (Ulmus americana) as a street tree in several regions of the USA and Canada led to significant canopy loss when Dutch elm disease (Ophiostoma ulmi) spread throughout the region during the 20th century, with canopy cover in some regions only recovering to pre-disease levels after 40 years (Roman et al. 2018). Climate change is predicted to impact the distribution and abundance of many pests and pathogens (Tubby and Weber 2010; Ramsfield et al. 2016), potentially making devastating losses of urban vegetation cover more prevalent. Despite advocating for change to how urban green spaces are designed, there are still many challenges to overcome such as limited nursery stock and a tendency to rely on a handful of well-known species (Sjöman and Nielson 2010; Conway and Vecht 2015; Khan and Conway 2020).
Many urban forestry practitioners have identified climate change as an important challenge for the survival of urban green spaces (Živojinović and Wolfslehner 2015). However, there is often uncertainty regarding which species will thrive under future climates (Khan and Conway 2020). There may also be a mismatch between what practitioners would like to plant and what is available for purchase at nurseries (Sydnor et al. 2010; Khan and Conway 2020). This is because nurseries may need 5 to 10 years to produce stock to be sold, therefore, favouring species which are already in demand (D’Amato et al. 2002). Even if practitioners took a proactive approach to planting species suited to future climates, this may not be possible due to the limited supply of plants available in the present market. Many species which are planted in urban areas now can survive for decades into the future and potentially be exposed to a significantly different climate. Therefore, it is imperative that we predict which species may be able to survive in future climates, with a high degree of confidence. Current urban planning and management decisions need to consider future climates to ensure that urban green spaces remain functional and healthy, delivering the desired benefits for years to come.
Numerous stakeholder workshops were conducted before the creation of Which Plant Where, to identify functionality that would be most beneficial for end users. Participants at these workshops identified that the end users of this tool would be Australian urban planners, practitioners, and specifiers (e.g., professionals working in local and state government, roads and transport services, the nursery industry, landscape architecture and development). During these workshops, barriers and knowledge gaps for species selection for urban planting projects were identified. These include identifying climate-ready species for urban landscapes, the ability to make informed decisions regarding species substitutions, the ability to quantify co-benefits to assess the success of urban plantings, the need to create a central hub for information on plant traits and tolerances, as well as designing best-practice, urban-greening guidelines.
Many plant selection tools have been created to aid practitioners in choosing species for urban settings, such as i-Tree (Nowak et al. 2018), Citree (Vogt et al. 2017), and Tree Species Selection for Green Infrastructure: A Guide for Specifiers (Hirons and Sjöman 2019). These tools contain in-depth information regarding species’ appearance, environmental tolerances, planting requirements, risks, and benefits. However, plant selection tools based on the suitability of species for future climates are generally lacking, with the few that exist using only 1 or 2 climate variables to match species’ climate tolerances with future projected climates (Yang 2009; Kendal et al. 2017; Brandt et al. 2021). Here, we describe the creation of Which Plant Where, a comprehensive plant selection tool for supporting resilient, climate-ready, urban green spaces in Australia. The tool allows practitioners to: (1) select species for urban greening projects that have been projected by species distribution models to tolerate climates for the periods 2030, 2050, and 2070; (2) diversify their planting palettes by browsing similar climatically suitable species; and (3) assess the benefits afforded by selecting palettes of plant species.
Materials and Methods
Species Selection
To create Which Plant Where, we compiled a comprehensive list of species, encompassing both common species in the horticultural trade as well as less common or underused species. Species were selected from multiple sources, including a list of species most commonly grown in Australian nurseries (M. Plummer, personal communication) and feedback from extensive stakeholder engagement regarding species types and underutilised species. Following stakeholder feedback, we placed a greater emphasis on native (two-thirds of the species list) and woody (two-thirds of the species list) species. We included species, subspecies, cultivars, varieties, genus cultivars, hybrids, and hybrid cultivars from various plant growth forms including trees, shrubs, climbers, herbaceous species, and graminoids sold and grown throughout Australia and planted in urban landscapes. To standardise the taxonomy, the species list was first checked against the backbone taxonomy of the Global Biodiversity Information Facility (GBIF) and then against The Plant List (TPL) using the Taxonstand package in R v.3.6.2 (R Core Team 2019).
Modelling
For Which Plant Where, we utilised 2 approaches to estimate the climate suitability of species at the postcode level (n = 2648). Postcodes in Australia largely correspond to suburbs, and through stakeholder engagement events, it was agreed that the postcode was the smallest geographic space over which most of the users of the tool would be operating. For most species (n = 1377), we fitted species distribution models (SDMs). For species for which SDMs either could not be fitted, due to lack of occurrence data, or had poor accuracy, we used a simplified approach (the niche method)(n = 463). Models were fitted at the species level only, due to the lack of robust occurrence data for cultivars, hybrids, and subspecies. All analyses were undertaken in R v.3.4.4 (R Core Team 2019). A full description of modelling methods can be found in Burley et al. (2019) and are briefly described below.
SDM Method: Occurrence Records
For each species, global occurrence records were downloaded from GBIF and the Atlas of Living Australia (ALA). These records were cleaned to remove duplicates and errors.
SDM Method: Climate Data
We obtained climate data from CHELSA v.1.2 (Karger et al. 2017). CHELSA contains 19 climatic variables (Appendix) summarised for the baseline (climate data averaged from 1979 to 2015) and 2 future time periods (2050 and 2070). These data were at a spatial resolution of approximately 1 km. We selected 6 of the 19 variables with correlations below 0.7 to fit models: annual mean temperature, temperature seasonality, maximum temperature of the warmest month, annual precipitation, precipitation of the driest month, and precipitation seasonality. However, due to poor model performance for 312 species (found mainly in Southwest and Western Australia), we employed an alternative set of bioclimatic predictors to maximise model performance (annual mean temperature, maximum temperature of the warmest month, mean temperature of the coldest quarter, annual precipitation, and precipitation of the driest month). To assess the impacts of climate change, data from multiple global climate models with low interdependence should be used to improve accuracy of model predictions (Beaumont et al. 2008; Taylor et al. 2012; Baumgartner et al. 2018). Hence, we used 10 global climate models identified as per Sanderson et al. (2015)(ACCESS10, CESM1BGC, CESM1CAM5, CMCCCM, FIOESM, GISSE2H, INMCM4, IPSLCM5AMR, MIROC5, and MPIESMMR) for Representative Concentration Pathway 8.5 (RCP 8.5). The RCP 8.5 trajectory was chosen to provide the worstcase scenario when modelling. CHELSA does not provide data for the year 2030; hence, we generated these for each global climate model through a linear interpolation based on the baseline and 2050 data. As such, the 3 future time periods we used are centred in 2030, 2050, and 2070. All data were reprojected to a Mollweide equal-areas projection (ESRI: 54009)(1 km × 1 km).
SDM Method: Modelling Approach
Our approach to model fitting and validation procedures can be found in Burley et al. (2019). For each species, we modelled the distribution of suitable climates for 2030, 2050, and 2070 using a machine learning approach, MaxEnt (Phillips et al. 2006; Elith et al. 2011), which has been shown to be more effective compared with other models when true absence data is unavailable (Elith et al. 2006). In addition to occurrence records, MaxEnt requires user-defined background locations for which it can compare the climate to that where the species is present. For each species, 50,000 random background locations were extracted from within 500 km of the target species’ occurrence records.
Following Burley et al. (2019), linear, product, and quadratic features were used to fit models in addition to default values for other parameters. Models were deemed to be of acceptable quality if values for the average test AUC (area under the receiver operating curve) and TSS (true skill statistic) were ≤ 0.7 and 0.5, respectively. Models were projected onto climate data representing current and future time periods. Additionally, Multivariate Environmental Similarity Surfaces (MESS) maps were generated to identify areas where the model was required to extrapolate into non-analogue climates (Elith et al. 2010; Di Cola et al. 2017). Regions of model extrapolation were then removed from maps of projected climate suitability to reduce uncertainty.
While AUC and TSS values are useful for identifying poor quality models, models with acceptable values of these variables can still produce ecologically unrealistic maps of the distribution of a species’ suitable climate. Hence, 3 plant experts independently validated each map of current climate suitability across Australia for each species, as well as current suitability across the native range for exotic species (i.e., those not native to Australia). Experts classified each model as acceptable or unacceptable.
SDM Method: Climate Suitability at the Postcode Level
We used the Postal Areas ASGS Ed 2016 Digital Boundaries from the Australian Statistical Geography Standard (Australian Bureau of Statistics 2011.) to extract climatic suitability at the postcode level (n = 2648) for each species. To do so, we calculated the median value of the 10 climate suitability maps for each species for each time period, thereby resulting in 3 maps per species (i.e., for 2030, 2050, 2070). Next, by overlaying the climate suitability maps with postcode geospatial data, we extracted the average suitability score for each postcode. This score (ranging from 0 to 1) was then converted to categorical data using values of 2 species-specific thresholds commonly used to convert MaxEnt output from continuous maps to binary maps: (1) fixed cumulative value 5 logistic threshold and (2) fixed cumulative value 10 logistic threshold. When applied to maps, these thresholds result in binary surfaces that include all but 5% and 10%, respectively, of the training samples as presences. By applying these thresholds to the average suitability score for each postcode, we classified postcodes as unsuitable for the species if the postcode’s suitability value lay between 0 and the value of the first threshold; marginal if the postcode’s value was between the first and second threshold, and suitable if it was above the second threshold.
Niche Method
For some species, there was insufficient data to create SDMs (i.e., fewer than 30 occurrence records) or the SDM did not pass the model validation step (i.e., poor AUC or TSS value or unrealistic map of current climatically suitable habitat). This was typical of exotic species with few records in Australia from GBIF/ALA. For these species, we used a simplified approach whereby we obtained CHELSA data for average maximum temperature of the warmest month and average precipitation of the driest month, for each occurrence. We then calculated the minimum, maximum, 5% and 95% values across these species’ occurrences for the baseline period. This climate niche represents the conditions for which we know the species can likely tolerate.
We also calculated the median value for both variables across each postcode for the 3 future periods (2030, 2050, and 2070) from the set of 10 global climate models. Each species was then classified as unsuitable, marginal, or suitable for each postcode and year based on the median values of the 2 climate variables for the postcode relative to the species’ climate niche. Species were classified as unsuitable if the median value of either climate variable for the postcode was greater or less than the minimum or maximum values of the species’ climate niche. This indicated that the postcode would likely be too cool/dry or hot/wet for the species. Species were classified as marginal if the median values of the postcode were within the 0% to 5% or 95% to 100% of the species climate niche for either variable. Species were classified as suitable if the median values of the postcode lie within the 5% to 95% values of the species climate niche for both variables.
Trait Collection
We conducted an extensive literature review to collect trait information for the species in the tool from a variety of nursery (domestic Australian and international), government, university, botanical garden, landscaping, and horticultural websites. The collected trait data were refined iteratively following consultation with key stakeholders from the nursery and landscaping industries. These key stakeholders decided on a core set of traits which were essential to display for each entry in the plant selection tool: common name(s), minimum and maximum height and width, origin, flower colour, flower period, leaf loss, shade tolerance, urban space type, and use. Other traits that were collected include information on environmental tolerances, soil conditions, and biodiversity benefits (see Table 1 for a full list of traits).
List of all traits and their classes found in the Which Plant Where plant-selector tool. Trait information was obtained from a variety of nursery (domestic Australian and international), government, university, botanical garden, landscaping, and horticultural websites.
Most of the traits that were collected had either continuous values (e.g., growth dimensions) or nominal values (e.g., flower colour) which were relatively easy to interpret. However, ordinal values, such as tolerance for drought, frost, and coastal conditions, were more ambiguous. For example, drought tolerance in horticultural literature is described using a variety of terms such as “tolerates very high water deficit,” “occasional drought,” and “prefers regular access to moisture.” For these traits, we used the methodology described in Tabassum et al. (2021b), where terms were tallied and classified as high tolerance, moderate tolerance, or no tolerance, using a consensus approach. Finally, an extensive assessment of the species and trait matrix was performed by an industry expert (Gwilym Griffiths) to ensure accuracy.
As well as traits collected from open sources, Which Plant Where contains additional information on heat and drought tolerance for a subset of 113 species and cultivars obtained from glasshouse experiments. This selection comprised a mixture of species both native and exotic to Australia, from a variety of different growth forms, and were species commonly available in the Australian horticultural market, as well as uncommon species less readily available commercially and infrequently planted in urban areas. These plants were subjected to experimental heatwaves and droughts in climate-controlled glasshouses to measure their physiological responses. Detailed explanations of how plants were droughted and subsequently ranked as drought tolerators and drought avoiders can be found in Marchin et al. (2020) and Tabassum et al. (2021b) respectively. An explanation of how plants were subjected to an experimental heatwave can be found in the Appendix.
Co-Benefit Calculations
Which Plant Where provides the opportunity for users to create multiple planting palettes to help design planting lists for urban greening projects. In addition to organising planting lists for specific projects, planting palettes calculate key co-benefits that urban plantings can provide. Average co-benefits are automatically calculated when 10 or more plants are selected to form a palette. The co-benefits that are calculated by the tool are shade value, carbon value, biodiversity value, planting diversity, and total canopy cover (Figure 1). Co-benefit calculations may be particularly useful for end users to assess whether their urban greening project meets standards such as tree canopy targets. An explanation of how these co-benefits are calculated within the tool can be found in the Appendix.
Screenshot of co-benefits calculated by the tool for a planting palette. Practitioners can use this function to modify species palettes based on specific ecosystem service targets, such as shade, carbon, and biodiversity.
Alternative Species
When plants are selected for planting palettes, Which Plant Where provides alternatives to the selected species based on similarity in climate suitability, growth form, height, exotic/native status, and leaf loss (Figure 2). This is to ensure that if substitutions need to be made for planting palettes, alternative species will still be climatically suitable.
Screenshot of Which Plant Where webpage for Backhousia citriodora, displaying common names, photos, traits, climate suitability, and occurrence maps. Species that were tested for heat and drought tolerance in the glasshouse are clearly marked. Users can input their desired location into the climatic suitability box to obtain information on the climatic suitability of the species in 2030, 2050, and 2070.
Users may view alternative species that are similar in terms of climate suitability and growth traits. Species which are declared environmental weeds in any Australian state or territory have a warning.
Maps showing climate suitability for Australia as well as local (Australian) and global occurrence records. On the climate suitability map, users can toggle between the periods 2030, 2050, and 2070 to visually assess how climate suitability for the species will change within Australia.
Results
Tool Properties
Which Plant Where contains 2,508 unique plant entries (encompassing species, subspecies, cultivars, varieties, genus cultivars, hybrids, and hybrid cultivars) from 150 different families and 662 different genera (Table 2). The tool consists of 1,548 plants which are native to Australia and 960 plants which are exotic to Australia, across 9 different growth forms (Table 3).
Breakdown of different types of taxa found in the Which Plant Where plant-selector tool. Note that genus cultivars refer to plants with only a genus name and a cultivar/variety name (e.g., Correa ‘Adorabell’). Similarly, hybrid cultivars refer to plants with only a hybrid name and a cultivar/variety name (e.g., Citrus × paradisi ‘Marsh’).
Summary of different growth forms found in the Which Plant Where plant-selector tool. Note that the total number is not equal to the total number of unique plants in the tool, as some plants may be classified as multiple growth forms (i.e., a plant could be both a shrub and a tree).
The traits in Which Plant Where are divided into 3 categories: plant form, site conditions, and performance. These main categories contain several subcategories such as soil properties and environmental tolerances, which help users to identify species for particular urban contexts (Table 1). The common denominator for all plants in the tool is the scientific species name. Subspecies, cultivars, and varieties are listed at the bottom of species pages. Conversely, subspecies, cultivars, and varieties have a link to their parent species at the bottom of their pages. Similarly, for hybrids, hybrid cultivars, and genus cultivars, links for up to 4 parent species occur at the bottom of their pages. This page structure enables users to easily navigate between linked plants.
Tool Functionality
Users can search for species for their urban greening projects in 1 of 2 ways: (1) searching for individual species (scientific or common name), or (2) searching through location (postcode or suburb name). Individual species searches will direct users to the species page, which displays information on species’ traits and maps showing climate suitability and occurrence records (Figure 2). Users have the option to input their desired location via postcode or suburb name directly onto the species page to retrieve that species’ climate suitability for 2030, 2050, and 2070. When conducting a location search, users can input their desired postcode or suburb name to obtain a list of species which are climatically suitable for the area. The species output is ordered alphabetically according to climate suitability, with species that are suitable for all time periods listed first, followed by species that are climatically suitable for 2030 and 2050 and marginally suitable for 2070, and so on. Users can then filter search results using a variety of categories including growth form, size, and co-benefits (Figure 3).
Screenshot of filters available to users when a location-based search is performed.
Users can choose to add their selected species to their palette. The palette page displays information about the number of taxa (species, genera, and families) and growth forms in the palette, enabling easy assessment of planting diversity. If there are 10 or more different plants selected for a palette, the tool will automatically calculate various co-benefits provided by this selection of plants (Figure 1). Palettes can be exported as a comma-separated values (CSV) file, containing information on species name, climate suitability, and co-benefits. This function is designed to assist in planning and specification of urban greening sites. Within the tool, users can create and manage multiple palettes for various urban greening projects.
Discussion
Climate change has had significant measurable effects in many ecosystems around the world, including urban ecosystems (IPCC 2022). However, the effects of climate change may be intensified in urban areas due to compounding effects of the urban heat island and impervious surfaces that reduce water infiltration (Sæbø et al. 2003). To ensure that urban green spaces survive and thrive into the future, information about which species will tolerate future climate conditions is needed now. Which Plant Where aims to provide a comprehensive guide to supporting resilient climate-ready urban green spaces by providing the tools and resources needed through the entire process, from planning to implementation and maintenance. It contains articles and best practice guidelines covering topics such as plant selection, plant procurement, and urban forest management. The tool contains information for over 2,500 native and exotic plant species, cultivars, varieties, and hybrids from a range of different growth forms. Users of the tool can easily select climatically suitable species for their urban greening projects based on specific locations. The tool also offers alternative species based on climate suitability to help diversify plantings and provide options where substitutions may have to be made. Lastly, Which Plant Where allows users to quantify various co-benefits that urban plantings can provide. To the best of our knowledge, this is the first plant selection tool in the world to utilise sophisticated species distribution models for such a large variety of species.
Limitations and Future Improvements for the Tool
Which Plant Where models climate suitability. However, other factors which are important for plant establishment and growth, such as soil nutrients and pest and disease distribution, have not been considered. The modelling approach used for Which Plant Where assumes that the occurrence records from the native range of species represent the breadth of environmental conditions that species can tolerate. However, forestry and horticultural studies have found that this assumption can often be false, as factors other than climate may directly limit the species native distribution, preventing it from occurring in other regions for which it is climatically adapted (Booth 2017; Kendal et al. 2018). Our species distribution models were fitted with long-term climate data and do not incorporate extreme weather events, such as heatwaves and droughts. Yet, these events have already had, and will continue to have, negative consequences for urban plants (Tabassum et al. 2021a; Haase and Hellwig 2022). Further, our calculations of climate suitability for each period were based on median values across 10 climate scenarios, highlighting that there is uncertainty in the magnitude of future climate change. It is anticipated that the modelling input for the tool will be updated on a regular basis to ensure that the climate suitability information is current. Future iterations of the modelling output could be improved by incorporating occurrence records from urban plantings to improve modelling outcomes.
Urban landscapes also contain a variety of microclimates which could not be accounted for in our modelling approach. Thus, a plant may survive in a climatically unsuitable postcode if the microclimate (e.g., as controlled by aspect) is favourable. Similarly, plants may survive outside of their optimal climate conditions due to human-mediated actions such as supplemental watering (Yang 2009). On the other hand, species planted in a particularly warm microclimate (e.g., against the western wall of a dark coloured building) in an otherwise climatically suitable postcode may not survive. To account for microclimatic effects when selecting species for urban plantings, more information about urban-specific stressors would provide additional functionality for Which Plant Where. For example, the woody species selection tool for urban spaces, Citree (Vogt et al. 2017), contains additional filters for site characteristics such as soil depth, soil compaction, and waterlogging risk. It is anticipated that future versions of Which Plant Where will include additional components such as soil volume requirements, locally indigenous information at postcode level and useful life expectancy information.
Although there are over 2,500 entries in the tool, sub-species, cultivars, varieties, and hybrids could not be modelled, resulting in 27% (668 entries) of plants not having associated climate suitability information. Without climate suitability information, the plant selector tool is not able to recommend a plant. Consequently, they will only appear if users search directly for them or if they are accessed through the parent species pages. A recommendation from this work is that more data be collected and made available on the performance of sub-species, cultivars, varieties, and hybrids, as they may perform very differently in response to environmental conditions, such as heat stress and low water availability, than the parent species they are derived from.
High climate suitability is an important contributor but does not guarantee good plant performance post-planting. Many factors can contribute to the failure of urban plantings, including poor quality nursery stock, inadequate site preparation, and inadequate maintenance during the establishment period (Hilbert et al. 2019). Sharing knowledge about successful and unsuccessful plantings can help urban forest managers identify good practice for species planting and how to best allocate resources to urban greening projects. It is intended that future versions of Which Plant Where be utilised as a platform to allow users to document planting outcomes through moderated discussion boards. This will ultimately allow for increased collaboration between different organisations (e.g., local councils, arborists, land managers, etc.) and help inform future trait collection for the tool.
Which Plant Where identifies environmental weed species that have been declared on state- and territory-level weed lists in Australia, meaning they are either prohibited from being brought into the state/territory or their presence in the state/territory is strictly controlled. However, many garden plants which are not currently problematic have the potential to become naturalised (form self-sustaining populations) and even invasive (spread beyond their point of introduction) in the future. The nursery industry represents a major avenue for new plant introductions, with over 70% of naturalised species around the world originally introduced as garden ornamentals (van Kleunen et al. 2018). Invasive weed species already have a major impact on Australia’s natural ecosystems, costing billions of dollars annually to control their spread (Hoffman and Broadhurst 2016). In response to this, the Gardening Responsibly program (Gardening Responsibly 2021) was launched in September 2022. It provides an assessment of invasive risk of ornamental garden plants in Australia, using a risk-assessment tool specifically developed for the program. Through this risk assessment approach, plants are categorised as having a high-invasive risk if they possess traits such as vegetative reproduction, long-distance dispersal, high seed output or are known to be invasive in other parts of the world, all features commonly found in invasive species (Hayes and Barry 2007; Pyšek and Richardson 2008; van Kleunen et al. 2010). In the next iteration of Which Plant Where, it is intended that the output from the Gardening Responsibly program will be aligned with the tool to give users confidence that their plant selection will be resilient under future climate and not be a risk for biodiversity.
Conclusion
Which Plant Where provides users with the means to select climatically suitable species to use in their urban greening projects and is underpinned by sophisticated species distribution models. To the best of our knowledge, this is the first urban plant selection tool to offer this feature. The tool also provides information regarding species’ growth characteristics, environmental tolerances, and ecosystem services to assist users in developing well designed planting palettes. We envisage Which Plant Where to be the ultimate resource for the urban greening industry in Australia, not just for climate-ready plant selection but as a focal point for best practice information. The ultimate goal of the tool is to be self-sustaining, enabling regular updates, improvement, and expansion. Future improvements to the tool will aim to incorporate the experiences of practitioners and urban forest managers in documenting success and failures in urban settings and improve confidence in selecting species with low-invasive potential. To ensure that our urban green spaces continue to thrive under projected climate change, tools to help select climate-ready species at local-regional scales will become increasingly important. As the climate data underpinning the plant selector tool is free to access, it is possible for other countries around the world to develop similar tools based on this approach.
Conflicts of Interest
The authors reported no conflicts of interest.
Acknowledgements
We thank Tim Maher, Malin Höppner, and David Coleman for assistance with the trait database; Hugh Burley, Shawn Laffan, and Anikó B. Tóth for assistance with the species distribution modelling and database structure; Sally Power and Paul Rymer for assistance with the co-benefit information; David Ellsworth and Renée Marchin for leading the glasshouse experiments; and Paul Rymer and Manual Esperón-Rodríguez for contributing to the tool resources. This manuscript is a contribution of the Which Plant Where project, funded by the Green Cities Fund, as part of the Hort Frontiers Strategic Partnership Initiative, with co-investment from Macquarie University, Western Sydney University, the New South Wales Department of Planning and Environment, and funds from the Australian Government.
Appendix
Climate Variables Obtained from CHELSA (v. 1.2) for the Species Distribution Modelling
Mean annual air temperature
Mean diurnal air temperature range
Isothermality
Temperature seasonality
Mean daily maximum air temperature of the warmest month
Mean daily minimum air temperature of the coldest month
Annual range of air temperature
Mean daily mean air temperatures of the wettest quarter
Mean daily mean air temperatures of the driest quarter
Mean daily mean air temperatures of the warmest quarter
Mean daily mean air temperatures of the coldest quarter
Annual precipitation amount
Precipitation amount of the wettest month
Precipitation amount of the driest month
Precipitation seasonality
Mean monthly precipitation amount of the wettest quarter
Mean monthly precipitation amount of the driest quarter
Mean monthly precipitation amount of the warmest quarter
Mean monthly precipitation amount of the coldest quarter
Methods for the Glasshouse Heatwave Treatments and Heat Tolerance Rankings
Glasshouse Experiment
A total of 113 perennial plant species and cultivars, including both native and exotic to Australia and across 5 different growth forms (trees, shrubs, climbers, herbs, graminoids), were selected for the experiments. These species were a mixture of species commonly available in the Australian horticultural market, as well as uncommon species less readily available commercially and infrequently planted in urban areas.
All plants used in the glasshouse experiments were obtained from commercial Australian nurseries as either tubestock, 140-mm, or 200-mm pot size. The original soil/media from around the roots of each plant was carefully removed, and plants were re-potted into 6-L pots (or 9-L pots for larger individuals) containing native potting mix (< 30% sand/coir, > 70% screened composted pine bark)(Australian Growing Solutions, Tyabb, VIC, Australia), 38 g of controlled-release native plant fertiliser (Scotts Australia Osmocote Slow Release, Bella Vista, NSW, Australia), and a 1.25-g systemic insecticide and fertiliser tablet (Yates Confidor, Padstow, NSW, Australia). All plants were well watered using drip irrigation and grown for 6 to 7 weeks in the glasshouses to allow for acclimation and the formation of new leaves. This reduced the chance that cultural practises in source nurseries would cause unexplained variation in the results. Plants were maintained at an average temperature of 27 °C, with a daytime range of 21 to 34 °C. These temperatures were chosen to mimic warm, urban, summer conditions that could occur in the major centres of Australia but were not excessive or stressful temperatures for plant species in this study.
After approximately 2 months of acclimation, plants were assigned to 1 of 4 treatments: (1) control (plants remained well-watered and grown at average summer temperatures); (2) drought (plants were subjected to a controlled drought while growing at average summer temperatures)(Marchin et al. 2020); (3) heatwave (plants were subjected to elevated temperatures while being well watered); and (4) a combination of drought and heatwave (plants were subjected to drought conditions and elevated temperatures).
Heatwave Treatment
Plants in the heatwave treatment were exposed to a 6-day heatwave after the drought period. During the heatwave, the average temperature of the glasshouse was 35 °C (8 °C warmer on average than the control temperature)(Figure S1), reaching a maximum of 41 °C. Air humidity was kept high during the heatwave regime to avoid the confounding effects of vapour pressure deficit and heat on plants.
Daily temperature profile of the growing conditions (left panel) and heat wave conditions (right panel) for the glasshouse experiments.
Trait Measurements
Visual Assessment of Damage
One of the most noticeable signs that a plant is suffering from heat stress is the visible browning and desiccation of leaves as cells begin to die. We assessed the effect of the heatwave on plants by visually estimating the percentage of leaf desiccation at the whole plant level after the final day of the heatwave.
Dark-Adapted Chlorophyll Fluorescence (Fv/Fm)
Fv/Fm represents a measure of how “healthy” green leaves are. During photosynthesis, light enters the chloroplasts in leaves, but not all of the light energy can be used by the plant. This unused light energy is released back as red light, and this is known as fluorescence. This fluorescence can be measured using a pulse amplitude modulated (PAM) fluorometer, which emits a high-intensity beam of light onto a leaf surface and measures the amount of light reflected (fluorescence). To take this measurement, plants were first dark adapted for 30 minutes to allow all of the reaction centres in the chloroplasts to be ready to receive light. Fo is the background level of fluorescence when the plant has been kept in the dark (Figure S2). A strong pulse of light is then fired onto the leaf by the PAM machine, saturating all the reaction centres. Fm, or maximal fluorescence, is then measured when all the reaction centres have been saturated and the excess light energy is emitted. Fv/Fm is then calculated as variable fluorescence (which is Fm – Fo) divided by Fm (Figure S2). High values of Fv/Fm indicate that the plant is healthy with healthy photosynthetic machinery. Low values of Fv/Fm, however, indicate that a plant may be stressed, as the photosynthetic machinery may be damaged.
Diagram showing a typical fluorescence curve and the relationship between Fm, Fo, and Fv.
Leaf Critical Temperature (Tcrit)
Tcrit marks the temperature where the photosynthetic machinery of a leaf starts to fail and cell death occurs. We measured Tcrit on 5 individuals of all species by placing a mature leaf on top of a thermoelectric heater and raising the leaf temperatures from 30 to 60 °C at the rate of 1 °C per minute. Fluorescence of the leaf was measured with the PAM machine, with the exact temperature at which leaves began to die marked by the sharp increase in fluorescence as photosynthetic machinery is damaged (Figure S3). Tcrit can be used to rank species based on thermal tolerance, with plants from warmer regions found to have higher Tcrit values.
Graph showing the point where Tcrit is reached and (insert) the aftermath of Tcrit testing on a leaf. Graph modified from Knight and Ackerly 2002.
Heat Tolerance Rankings
Heat tolerance rankings were calculated by weighting the heat tolerance traits (60% visual damage, 20% Tcrit, 20% Fv/Fm), which were then used to classify species into 4 categories (sensitive, intermediate, intermediate-tolerant, tolerant). These categories were determined based on the distribution of values for the heat tolerance traits. Species that fell into the sensitive category generally had at least 5% foliage damage in the heat-only treatment, decreases in Fv/Fm values (compared to no heat treatment), and low Tcrit values. Species that fell into the tolerant category had no perceptible foliage damage in the heat-only treatment, a maximum of 7% decrease in Fv/Fm values (compared to no heat treatment), and high Tcrit values. Species that fell into the intermediate category had 2% to 5% foliage damage in the heat-only treatment and moderate Tcrit values, while species in the intermediate-tolerant category had 0% to 2% foliage damage and moderate Tcrit values.
Methods for Co-Benefit Calculations for Species Palettes in the Which Plant Where Webtool
Shade Value
Shade values for planting palettes were calculated by first determining the shade value at maturity for each tree species in the Which Plant Where webtool. The shade value at maturity was calculated from the maximum canopy width (CWmax) and the maximum tree height (THmax) that a species can reach at maturity on average. The shade value at maturity is an average of (1) the average vertical canopy shade projection at midday 
that depends on tree canopy width; and (2) the lateral canopy shade projection at sunset/sunrise (THmax × CWmax) that depends on both canopy width and tree height. The differentiation of these 2 values has been introduced, recognising that some tree species might have the same vertical shade projection on the ground at midday but vastly different shade casting when the sun is not at the zenith. Please note it is assumed that tree canopies are circular when the shade footprint is projected onto the ground.
The shade value at maturity for each tree species was then converted into a shade index value ranging from 1 to 4, based on the distribution of values of all the Which Plant Where tree species:
1 (very low): first quartile of distribution (i.e., from minimum value to 25th percentile)
2 (low): second quartile distribution (i.e., 25th to 50th percentile)
3 (medium): third quartile of distribution (i.e., 50th to 75th percentile)
4 (high): fourth quartile of distribution (i.e., 75th percentile to maximum value)
To calculate the shade value at the palette level, the shade index values for the trees in the palette were tallied and expressed as a percentage of the maximum shade index value obtainable for the same number of species in a palette. This value, ranging from 0 to 100, was then converted to the shade value as follows:
Low shade: values between 0 and 25
Moderate shade: values between 26 and 50
High shade: values between 51 and 75
Very high shade: values between 76 and 100
Carbon Value
Carbon values for planting palettes were calculated by first determining the carbon value at maturity for each tree species in Which Plant Where. The carbon value at maturity was calculated from the maximum height that a tree species can reach on average (THmax).
THmax values were then converted into a carbon index value ranging from 1 to 3 based on the distribution of values of all the Which Plant Where tree species:
1 (low): first third of distribution (i.e., minimum to 33rd percentile)
2 (medium): second third of distribution (i.e., 34th to 66th percentile)
3 (high): last third of distribution (i.e., 67th percentile to maximum value)
To calculate the carbon value at the palette level, the carbon index values for the trees in the palette were tallied and expressed as a percentage of the maximum carbon index value obtainable for the same number of species in a palette. This value, ranging from 0 to 100, was then converted to the carbon value as follows:
Low carbon: values between 0 and 25
Moderate carbon: values between 26 and 50
High carbon: values between 51 and 75
Very high carbon: values between 76 and 100
Biodiversity Value
Biodiversity values for planting palettes were calculated by first determining the biodiversity index for each plant in Which Plant Where. Each plant was assigned a baseline value of 1 for habitat. Information on whether plants provided specific habitat and/or food resources for 4 different types of fauna (pollinators [i.e., bees and butterflies], insects [other non-pollinator insects], lizards/mammals, and birds) was obtained from a variety of sources, with each plant being assigned a value of 1 for each faunal group it provides resources for. The biodiversity index for each plant was calculated as the sum of all the singular biodiversity scores for each group, with a maximum value of 5.
To calculate the biodiversity value at the palette level, the biodiversity index values for plants in the palette were tallied and expressed as a percentage of the maximum biodiversity index value obtainable for the same number of species in a palette. This value, ranging from 0 to 100, was then converted to the biodiversity value as follows:
Low biodiversity: values between 0 and 25
Moderate biodiversity: values between 26 and 50
High biodiversity: values between 51 and 75
Very high biodiversity: values between 76 and 100
Planting Diversity
To estimate planting diversity at the palette level, we utilised the 10:20:30 rule from urban forestry (Santamour 1990). This rule is composed of 3 criteria:
Species palettes should contain at least 10 different species (note that this is a slight modification from the original 10:20:30 rule proposed in Santamour 1990).
No more than 20% of species from a given palette are from the same genus.
No more than 30% of species from a given palette are from the same family.
Based on these criteria, each palette is given a diversity score based on the following criteria:
High diversity: the palette list fulfils all 3 criteria for the 10:20:30 rule.
Moderate diversity: the palette list fulfils any
2 of the 3 criteria for the 10:20:30 rule.
Low diversity: the palette list fulfils only one criterion for the 10:20:30 rule.
Total Canopy Area
Total canopy area for planting palettes were calculated by first determining the canopy cover at maturity (CCmat) for each tree species in Which Plant Where. CCmat was calculated from the maximum canopy width that a species can reach at maturity (CWmax) using the following equation: 
. Please note it is assumed that tree canopies are circular when projected onto the ground.
To calculate total canopy area of a palette, canopy covers for all trees in the palette are summed. If duplicates of particular tree species are specified, the total canopy area is automatically adjusted to reflect this.
- © 2023 International Society of Arboriculture
Literature Cited
Jump to section
Related Articles
Cited By...
- No citing articles found.