Physiological Response of Trees to Seasonal Variation Reveals Native Species Are Suitable for Urban Parks That Resemble Natural Sites

Urban trees may suffer water stress due to urban heat island effects, and this problem is expected to increase with climate change (Chapman et al. 2017; Burley et al. 2019). Urban heat island effects could become particularly acute when overlain with late-summer water stress, which may be one factor leading to the decline in the urban canopy (Nowak and Greenfield 2018). Urban foresters thus face the daunting task of choosing tree species that can tolerate the stress of urban environments while maintaining aesthetic value and presenting limited maintenance issues. The choice of urban tree species has been subjective, however, the result of heuristic decision-making, nursery availability, and neighborhood resident preferences (Avolio et al. 2018). In the past decade, there has been growing interest in using plant physiology to define species suitability for urban plantings. Intensive, comparative studies of tree physiology can accurately quantify stress tolerance, but most previous research on urban tree physiology focuses on a narrow set of drought tolerance traits for a few commonly planted species or cultivars (Gillner et al. 2015; Gillner et al. 2017; Sjöman et al. 2018). In addition, physiology is difficult and time-consuming to measure, so studies that relate dynamic physiological responses to quick-to-measure traits are needed to speed the discovery of suitable species for use in urban forestry.

Comparing spring versus summer water relations can provide insight into how trees manage water stress, because drought stress increases predictably over summer as temperatures increase. Maximum leaf hydraulic conductance (K_leaf), or the rate of leaf laminar water flow at a given water potential gradient, determines the capacity for leaf water loss and thereby sets limits on stomatal conductance (Sack and Scoffoni 2012; Scoffoni et al. 2016). Similarly, the water potential at leaf turgor loss point (Ψ_TLP) sets limits on cellular function and predicts stomatal closure and wilting (Rodriguez-Dominguez et al. 2016; Maréchaux et al. 2018). If K_leaf declines from spring to summer, growth limitations arise due to increased resistance in the leaf hydraulic pathway (Scoffoni et al. 2016), which reduces the capacity of the leaf to replace water lost due to evaporation, leading to stomatal closure (Brodribb and Holbrook 2003). On the other hand, when Ψ_TLP declines from spring to summer, increased cell solute concentration can allow stomata to remain open as leaf water potential declines (Maréchaux et al. 2018). Thus, changes in Ψ_TLP could mitigate risk of K_leaf decline, allowing photosynthesis to continue even as drought stress increases over summer.

Potentially exacerbating heat and drought stress in urban areas, urban soil is prone to compaction due to human activities such as increased foot traffic and heavy machinery use, leading to reduced porosity and water-holding capacity. Urban street trees and those in planters are likely to experience more intense drought effects than urban park trees due to small soil volumes (Kjelgren and Clark 1993). Urban park trees are an interesting context in which to explore the performance of trees, because even with larger soil volumes compared to street trees, urban park soil can also be extremely compacted, which should increase the probability of seasonal water stress. The extent of soil compaction in urban parks, however, is likely to vary substantially based on park history (Scharenbroch et al. 2005). Species also vary in their ability to tolerate poor quality or compacted soil; for example, tree species that have wide geographic distributions, those which can tolerate flooding, or those which tend to grow in disturbed soils may have a stronger ability to acclimate and thereby maintain their performance across seasons and different soil types (Sheth and Angert 2014; Puglielli et al. 2023).

Soil is an important component that interacts with precipitation to determine physiological outcomes. Trees experiencing lower precipitation and more heavily compacted soil may exhibit the largest K_leaf decline over the summer, because compact soil prevents water infiltration and limits root growth (Scharenbroch et al. 2005). Faster growing species should show the largest K_leaf decline due to their need to maintain higher stomatal conductance to support higher photosynthetic capacity, unless they can respond to stressful conditions with larger seasonal changes in Ψ_TLP (Bartlett et al. 2012). Pyrus calleryana is commonly used in urban settings in part because it is viewed as fast-growing (Rahman et al. 2015), but fast growth is likely one reason this species has escaped cultivation and demonstrated potential to be invasive (Culley and Hardiman 2007). In addition, underlying physiological trade-offs may result in lower mechanical strength of fast-growing species (Reich 2014), so this might be one reason why P. calleryana is prone to breakage (Dirr 1990). Slower-growing species should allocate more to carbon-rich structural tissues (Wright et al. 2004), making their leaves and wood more suitable to provide ecosystem services like carbon sequestration over the long term. While fast growth might provide an advantage in some urban settings, we lack information about the extent of stress tolerance required for long-term survival of urban trees in different types of growing conditions.

In 2023, the Ohio Department of Natural Resources banned the planting of Pyrus calleryana, meaning that now alternatives which can perform under similar urban conditions are urgently needed. Tree species selection for urban plantings has long involved debate over the relative value of native versus non-native species. Although native species are thought to be better adapted to regional conditions, this assumption may not hold in highly modified urban environments where abiotic stressors can differ substantially from those in surrounding natural areas (Kendle and Rose 2000). Consequently, trait-based and physiological assessment may provide a more mechanistic framework for informing site-specific urban forestry decisions. Thus, providing more information on the suitability of native species for urban plantings is critical. For instances when locally native species’ suitability is limited, consideration of southern species may be appropriate. Under climate change models, suitable habitats are projected to shift northward, potentially shifting species distributions more northward (Prasad et al. 2025). Species that are currently distributed farther south of a particular urban area may therefore be useful candidates for urban plantings. This hypothesis is based on the idea that plants occurring at warmer or drier locations within a species range may have greater heat or drought tolerance; however, this idea should be evaluated physiologically and at the site scale because tree responses to climate change are strongly trait-dependent (Fei et al. 2017).

To investigate the suitability of deciduous hardwood species native to the Northeastern United States for urban park plantings in Northeast Ohio, we conducted a study of urban park trees growing at Lakeview Cemetery, located in the heart of Cleveland, Ohio, and Secrest Arboretum, located in Wooster, Ohio, a small city situated approximately 40 km away from the Akron-Canton metropolitan area. For both locations, Mixed Oak Forest was the native plant community prior to urbanization. These are nonirrigated parks characterized by planted, open-grown trees surrounded by turf grass with some paved utility and walking paths. We compared 3 tree species native to Ohio (Acer rubrum, Acer saccharum, and Nyssa sylvatica) with a Southeastern United States native species, Liquidambar styraciflua, and East Asian Pyrus calleryana. We measured seasonal changes in K_leaf and Ψ_TLP to characterize tree responses to late-summer water stress, along with photosynthetic capacity (maximum electron transport rate [J_max] and maximum carboxylation rate [Vc_max]), and gas exchange rates (net photosynthesis [A], stomatal conductance [g_s], and instantaneous water use efficiency [WUE]). We compared these physiological rates across parks and to previously published data on soil texture and bulk density of samples collected from each tree planting location. Lastly, to understand potential growth-stress trade-offs and identify quick-to-measure physiological proxies, we asked how these dynamic physiological rates measured in our study compared to previously published data on relative growth rate and carbon allocation (aboveground carbon sequestration, diameter at breast height, specific leaf area, leaf dry matter content [LDMC], and leaf carbon:nitrogen ratio [CN]) collected in the same parks on the same plants examined in our study by Simovic et al. (2024).

Specifically, we asked the following:

1. Do tree physiological rates differ across parks with different weather patterns or according to soil properties of the planting location?

2. Do tree species native to Northeast Ohio exhibit physiological rates consistent with late-summer water stress, including a larger change in K_leaf and a smaller seasonal change in Ψ_TLP, compared with the Southeastern US native Liquidambar styraciflua and East Asian Pyrus calleryana?

3. Are faster physiological rates and greater seasonal change in water relations associated with quick-to-measure traits reflecting faster relative growth rates and/or lower carbon allocation to leaves and wood?

We measured physiological responses to seasonal variation for 3 Ohio native tree species (Acer rubrum, Acer saccharum, and Nyssa sylvatica), along with the Southeastern US species Liquidambar styraciflua and East Asian Pyrus calleryana, at two urban parks in Northeast Ohio. We found that nearly all measured physiological rates were higher at Lakeview, indicating faster and more efficient growth, compared to Secrest. Physiological rates do not account for factors such as differences in tissue construction costs (Kruger and Volin 2006), which may be one reason why we did not see correspondingly higher carbon sequestration at Lakeview in our study. Trees at Lakeview did have a net-positive change in maximum leaf specific hydraulic conductance (ΔK_leaf) from early to late summer, indicating that growth was not hindered by evaporative demand, which our data suggests did not exceed the ability of the plant to supply water to the leaf (Sack and Scoffoni 2012; Scoffoni et al. 2016). At Secrest, declines in K_leaf and Ψ_TLP over the season suggest that trees at this site did experience mild seasonal water stress (Brodribb and Holbrook 2003; Maréchaux et al. 2018). But overall, we saw physiological adjustments in K_leaf and Ψ_TLP from early to late summer at both parks, which should mitigate drought stress and contribute to the relatively robust photosynthetic rates we observed during late summer. Compared to P. calleryana, these seasonal adjustments were similar in magnitude for the other study species, demonstrating their ability to comfortably manage late-summer water stress in an urban setting.

Secrest was drier than Lakeview during the period of this study, and long-term climate records indicate that trees at Secrest experienced lower-than-average summer temperatures, whereas the summer was hotter than usual at Lakeview. These weather patterns are influenced by the proximity of Lakeview to Lake Erie, which results in higher cloud cover, precipitation, and humidity compared to Secrest. In addition, Lakeview is a larger and more densely planted park compared to Secrest, which could have large impacts on micro-scale climate experienced by individual trees. Our weather data should be viewed as characterizing only the prevailing weather conditions at each park, but in addition to the larger park size and higher planting density at Lakeview, the combined evidence suggests that trees at this park experienced conditions more similar to surrounding native forest compared to Secrest.

While our climate analysis does not capture climate variation at the scale of individual trees, our soil analysis does provide insight into tree-scale variation across the parks. Planting soils at Lakeview had a higher proportion of clay and lower silt compared to those at Secrest, suggesting they could have slower drainage, although there was substantial variation in soil texture within parks as well. We found that K_leaf in late summer as well as the efficiency of gas exchange (water use efficiency, maximum electron transport, and carboxylation rates) were all higher with increasing clay percentage and higher bulk density. While higher performance at higher bulk density may seem counterintuitive, the range of bulk densities at our study sites was on the lighter side for urban areas, though some planting-site soils approached what is considered the upper limit for unrestricted root growth (1.6 g cm⁻³)(Scharenbroch et al. 2005). If higher bulk density is driven by higher clay contents, rather than compaction, the higher clay percentage could improve soil water-holding capacity or cation-exchange capacity. Interestingly, planting-site soils at Lakeview and Secrest parks had bulk density values that were 3× to 4× higher and sand percentages that were 2× higher compared to soils in a nearby forested area which includes the Ohio native species we studied here (personal communication Katie Stuble). The percentage clay in nearby forested areas was similar to that measured at Lakeview, but approximately 2× higher than percentage clay measured at Secrest (personal communication Katie Stuble). This similarity between nearby forests and Lakeview soils could contribute to the higher physiological rates we observed at Lakeview.

Another potential factor contributing to higher physiological rates at Lakeview is the fact that the trees were smaller compared to those at Secrest (Table 4), and water transport and photosynthetic capacity typically decline with tree size (Bond 2000). In support of this idea, our pairwise correlations did show that photosynthetic efficiency was positively correlated with relative growth rate but negatively correlated with tree size. Tree size has rarely been considered in studies of urban tree physiology (but see Simovic et al. 2024), and our work shows that this could be important to clarify the sources of physiological variation within urban spaces. It is also important to note here that we found variation in both tree size and soil properties across the two parks we investigated. Our use of park as a random factor in our soil analysis does provide confidence that soil does affect physiology independent of other factors that differ across parks, but the full extent to which soil versus size drives our results cannot be determined based on the current study. These findings highlight the complexity of the urban environment and the need for physiological proxy measures in urban tree selection, since the time-consuming nature of the seasonal physiological measurements limited the number of parks and trees we could include in our study.

Within this context, our investigation of seasonal responses at each site provides valuable insight because this analysis is independent of tree size. We hypothesized that the seasonal physiological response of trees could depend on whether the species is fast- or slow-growing (Wright et al. 2004), a trait which foresters may use to select suitable species (Rahman et al. 2015). We did find some support for this hypothesis, as our pairwise correlations did show a significant negative relationship of CN to ΔK_leaf. Plants with lower CN are expected to have a more acquisitive physiology (Reich 2014), and our data show that these plants had a larger change in K_leaf across seasons. In addition, our principal component analysis showed that the change in water potential at turgor loss point from early to late summer (ΔΨ_TLP) was positively correlated with relative growth rate and negatively correlated with carbon sequestration. In this analysis, LDMC was also positively correlated with ΔΨ_TLP, with a larger change indicating higher resistance to loss of turgor pressure in late summer. Pairwise correlations showed no significant correlations between these variables however, suggesting that other traits moderate the relationship of LDMC to ΔΨ_TLP. A previous study of the same trees suggested that higher LDMC might alleviate water stress, if it was correlated with a higher resistance to loss of turgor pressure and/or xylem conductivity (Simovic et al. 2024). Our work supports this hypothesis and further suggests that LDMC and CN, which are both extremely quick-to-measure traits, should be further explored as proxies for determining the growth and stress responses of species. Interestingly, P. calleryana in our study had the highest growth rate at Lakeview and the lowest growth rate at Secrest, such that the mean across the two parks was intermediate, and this species was not statistically distinguishable from others in this analysis.

The two Acer species we examined were highly similar in their responses, although we observed key contrasts with relevance for urban plantings. Acer rubrum is a pioneer, bottomland species which is commonly found on disturbed soil and can tolerate a wide range of soil types (Burns et al. 1990). Acer rubrum has a strong ability to acclimate to stressful urban conditions (McDermot et al. 2020) such that its photosynthetic rates can be higher in urban compared to rural areas (Lahr et al. 2018). Our data show that this species had a high Ψ_TLP, which did not change across seasons, leading to the steepest seasonal decline in K_leaf and the lowest late-summer K_leaf compared to all other species. This is consistent with previous work showing a decline in gas exchange rates of A. rubrum from early to late summer (Lahr et al. 2018). Our principal component analysis adds to this by showing that A. rubrum trees are more different from one another than they are compared to trees of other species (widely dispersed along PC1 and PC2), perhaps reflecting their strong ability to acclimate to different growing conditions (McDermot et al. 2020) or higher intraspecific genetic variation for A. rubrum. Acclimation ability in particular may explain high variation in A. rubrum performance across different types of conditions (Burns et al. 1990), because acclimating to optimize growth under current site conditions can compromise future survival if conditions change (Medeiros 2025).

Our study sites are within the center of the native range of A. rubrum, but the other Acer species we examined, A. saccharum, has a more northerly distribution (Burns et al. 1990). This is a climax species which is restricted to cool, moist upland habitats, and it does best on well drained loam soil (Burns et al. 1990), suggesting that A. saccharum should have poor performance in urban sites. Instead, we found this species performed very similarly to A. rubrum, including having low K_leaf and high Ψ_TLP, which did not change appreciably across seasons. One reason for their similar performance may be the ability of A. saccharum to tolerate compacted soils (Leasia et al. 2012). Acer saccharum was shown to be the second most commonly planted urban tree in the Midwest United States, while A. rubrum was the sixth most common (Schoon 1993). Our study shows that A. saccharum trees had the lowest spring K_leaf at Lakeview but the highest Ψ_TLP at Secrest, and individual trees exhibited strong similarity to one another in our principal component analysis. This is consistent with the idea that this species has lower acclimation ability and implies that it may reach its turgor loss point earlier and more severely compared to A. rubrum but could also reflect lower within-species genetic variation for A. saccharum.

In terms of carbon allocation, the two Acer species had similarly high LDMC values compared to other species, but A. rubrum trees were the largest trees at both parks in terms of DBH and height. Compared to trees at Lakeview, all those at Secrest were larger and they grew more slowly, but A. rubrum showed the largest contrast in size across parks, again reflecting high variation across growing condition for this species. Simovic et al. (2024) also showed that trees at Secrest were older compared to those at Lakeview, and that younger age explains the smaller sizes at Lakeview rather than differences in park environmental conditions. Our physiology data provides additional support for this conclusion by showing that all measures of physiological performance were higher at Lakeview compared to Secrest, which is consistent with data on physiological variation between older and younger trees (Bond 2000).

We also investigated East Asian Pyrus calleryana because this species has been considered better suited to urban conditions compared to other tree species (Rahman et al. 2015). Further, the 2023 ban on planting this species in Northeast Ohio necessitates identification of other species which can be successful in sites where P. calleryana is currently planted. In our study, P. calleryana had a slightly larger ΔΨ_TLP from early to late summer compared to other species, but we did not see discrepancies in ΔΨ_TLP among species, seasons, or parks. This may be in part because the lowest Ψ_TLP values we measured were less negative compared to those measured in previous studies of forest trees (Table 7), suggesting that trees at our study parks did not experience water stress outside what they would experience in their native habitats. This is perhaps surprising given that urban trees are often expected to be more drought-stressed compared to nonurban counterparts (Sjöman et al. 2015; Sjöman et al. 2018) and it highlights that our assumption of higher drought stress for urban trees will not be universally true, especially in urban parks. Carbon allocation of P. calleryana was strikingly different from all other study species, with lower carbon sequestration rates in aboveground woody biomass and low CN at both parks. Low CN indicates faster growth rate (Wright et al. 2004), which is commonly associated with low wood strength (Poorter et al. 2010), providing support for the idea that this species is less structurally sound than native tree species (Dirr 1990).

Interestingly, Ψ_TLP values for Liquidambar styraciflua and Nyssa sylvatica were not significantly different from Pyrus calleryana at Lakeview and they were lower compared to the two Acer species. We also found that L. styraciflua and N. sylvatica, along with P. calleryana, had the highest values of K_leaf across both sites and seasons. Liquidambar styraciflua, native to the Southeastern United States and the southern part of Ohio, has previously been shown to acclimate to drought and salt stress (Baraldi et al. 2019) and different types of urban planting conditions (Kjelgren and Clark 1993). We found that L. styraciflua had carbon sequestration that was statistically indistinguishable from P. calleryana at Secrest, our more stressful site, and the DBH of these two species was not statistically distinguishable at either site. Lastly, L. styraciflua has the lowest LDMC value of any species, and combined with the growth patterns, our data suggest this species would perform similarly to P. calleryana in Northeastern Ohio park plantings. Yet, L. styraciflua was ranked 22nd in terms of presence in urban areas (Schoon 1993), indicating this species is underutilized.

In contrast, virtually nothing was previously known about the physiology of Nyssa sylvatica in urban settings, which was ranked 97th in terms of commonality in urban spaces (Schoon 1993). This species prefers well-drained light-textured soils found in creek bottoms, and trees growing on dry upland sites tend to be smaller (Burns et al. 1990), which may be one reason why they seem to be passed over in choosing urban trees. Our study parks are at the northern edge of N. sylvatica range (Burns et al. 1990), so the shorter growing season could explain the fact that trees of this species were the smallest in our study. Nyssa sylvatica also has higher growth rates in the presence of deep, periodic flooding (Keeland and Sharitz 1995), but our study suggests that they are highly suitable trees for urban parks in Northeast Ohio, potentially due to their ability to tolerate swings from drought to waterlogging that can occur in compacted urban soil (Scharenbroch et al. 2005).

Given the high heat and water stress reportedly experienced by urban street trees, it makes sense that urban foresters have sought to determine drought-tolerant species suitable for these extreme conditions (McCarthy et al. 2011; Cregg et al. 2023). Our study highlights the fact that not all urban environments are characterized by drought, and parks may be particularly good choices to increase the diversity of urban tree species. For example, L. styraciflua trees planted in urban parks experience less water and nutrient stress compared to those planted in areas with a greater amount of adjacent impervious surface (Kjelgren and Clark 1993). Isolated trees in parks should have minimal light limitation, potentially increasing the negative impacts of water limitation (Simovic et al. 2024), but this may be counteracted by relatively larger soil volumes and minimal competition, allowing for better root growth and water relations compared to plaza or street trees.

Here, we have presented perhaps the most detailed seasonal comparison to date of native and non-native tree physiology in urban parks. By combining leaf hydraulic conductance, turgor loss point, and photosynthetic CO₂ response curves and harnessing the interpretive power of 2 seasonal time points, we offer unique insight into how park trees manage environmental conditions. Importantly, our study shows that species native to Ohio or the Eastern United States performed comparably to East Asian P. calleryana and thus might make good alternatives in Northeastern Ohio parks that previously included P. calleryana. The assessment of tree species for Northeastern Ohio parks is far from complete, however, as the Ohio Department of Natural Resources includes 99 species native to the state, and each urban area has its own nearby local populations of native species for consideration. Thus, our results should be considered exploratory, opening the door for new hypothesis testing, rather than an exhaustive assessment of species suitability. These findings do provide a foundation for city planning and future research, particularly in Northeast Ohio. Specifically, our identification of LDMC and leaf C:N ratio as potential quick-to-measure proxies for physiology has practical implications, and future work should consider these traits for a larger number of individual plants, for more species, and across sites with different urban land-use types to better assess the suitability of diverse tree species for use in urban forestry settings.

The authors reported no conflicts of interest.