Hotter and drier climates globally, coupled with periodic drought, often necessitate large quantities of irrigation water to maintain visual quality, growth, and development of landscape plants (Mee et al., 2003). Approximately 60–90% of household water is used for urban landscape irrigation in the western United States (Sovocool et al., 2006; Center for Water-Efficient Landscaping, 2020; Arizona Department of Water Resources, 2022). However, due to the increasing water demand of a growing population, designing landscapes with drought-tolerant adaptive plants or plants native to arid and semiarid areas is important for long-term water conservation in the western United States. In addition, landscape plants are threatened by increasingly common droughts and heatwaves in the western United States because they are largely reliant on irrigation (Miller et al., 2020). A recent drought caused urban vegetation coverage in downtown Santa Barbara, California, to decline from 45 to 35% (Miller et al., 2020). Hence, landscape plants characterized by morphological and physiological plasticity, which can better acclimate to water and heat stresses, are desirable for future landscapes. Unfortunately, drought responses of landscape plants are seldom investigated, and drought tolerance studies have largely been conducted based on local precipitation rates, rather than well-controlled inputs (Meyer et al., 2009).

Reduction in soil water availability causes cell dehydration, resulting in leaf wilting and degrading aesthetic appearance (Zollinger et al., 2006). Cell dehydration then prevents chlorophyll production and photosynthesis, which reduces leaf greenness and plant growth (Ahluwalia et al., 2021). For instance, Orthosiphon aristatus (cat’s whiskers) exhibited wilted leaves and reduced leaf and root biomass when no irrigation was applied (Kjelgren et al., 2009). Water stress also inhibits leaf expansion, reducing light-capture area (Dale, 1988) and may indirectly induce heat stress in plants because of reduced transpirational cooling to counter absorbed radiation (Nobel, 2009). Gaillardia aristata (blanketflower) and Penstemon barbatus (golden-beard penstemon), for example, showed over 50% of the leaves burned when water was limited (Zollinger et al., 2006). High temperatures may disrupt plant metabolism and protein stability, leading to leaf burn and necrosis (Taiz et al., 2015).

Plant acclimation involves changes in morphology and physiology without genetic modification (Taiz et al., 2015). Under drought conditions, plants may acclimate to drought by decreasing water loss and reducing heat load and leaf temperature (Mee et al., 2003). Root growth may be promoted to increase water uptake, leading to a greater root-to-shoot ratio (Ahluwalia et al., 2021). Water loss may be minimized via stomatal closure, leaf senescence, and reduced leaf size (Zollinger et al., 2006). For instance, Stromberg (2013) found that xeric species growing in the southern United States have greater root-to-shoot ratios, but smaller leaves, than mesic species. In hot and arid environments, plants gradually reduced their stomatal conductance and transpiration along with increasing leaf temperatures and higher leaf-to-air vapor pressure deficit (VPD) to prevent excessive water loss (Roessler and Monson, 1985). Minimizing stomatal conductance when solar radiation and air temperature are greatest at midday can protect plants from xylem dysfunction and maintain water status (Zhang et al., 2013).

Plant leaf temperature may be regulated by adjusted leaf size, orientation, and trichome density (Ehleringer et al., 1976; Nilsen, 1987; Nobel, 2009). For example, small leaves are advantageous for increasing sensible heat loss. The leaves of native plants in the western United States, such as Artemisia tridentata (big sagebrush) and Cercocarpus montanus (alder leaf mountain mahogany), are less than 2.5 cm wide, helping to reduce plant heat load more efficiently (Mee et al., 2003). Leigh et al. (2017) reported that plants in hot and dry environments of Australia, such as Banksia grandis (bull banksia), Grevillea agrifolia (blue grevillea), and Telopea speciosissima (waratah) have leaves covered by dense trichomes and vertical leaf orientation, which reduces the interception of solar radiation.

Trichome density has been found to be affected by soil water content, air temperature, and VPD (Ehleringer, 1982; Banon et al., 2004; Bickford, 2016; Shibuya et al., 2016). For instance, the trichome density of Lotus creticus (cretan trefoil) increased when the amount of irrigation water decreased by 70% (Banon et al., 2004). Shibuya et al. (2009) discovered that Cucumis sativus (cucumber) had 255 trichomes per cm² of leaf area at a vapor pressure deficit of 0.4 kPa which increased to 463 trichomes per cm² at 3.8 kPa. Ehleringer (1982) observed that trichome density of Encelia farinosa (brittlebush) grown in California positively correlated to the mean maximum air temperature of the growing habitat. However, the effect of water stress on plant trichome development has not been widely studied. Early research suggested leaf trichome production was promoted under water deficit (Quarrie and Jones, 1977). However, this finding contradicts the fact that plant cell division is inhibited under drought stress conditions (Dale, 1988). Brodribb et al. (2013) reported that changes in cell size provided a substantial means to modify leaf function without disturbing other tissue/organ functions. Murphy et al. (2014) found that epidermal cell expansion facilitated the decrease of stomatal density under shade, where large leaves had low stomatal density. Stomata and trichomes are both epidermal appendages and their development occurs prior to cell expansion. Hence, changes in cell size may modify trichome density under water stress.

Shepherdia ×utahensis ‘Torrey’ (hybrid buffaloberry) is an interspecific hybrid between Shepherdia argentea (silver buffaloberry) and Shepherdia roundifolia (roundleaf buffaloberry). Shepherdia argentea tolerates a wide range of growing conditions from wet to dry soil (Mee et al., 2003), while S. roundifolia is extremely resistant to hot and arid conditions (Sriladda et al., 2016). Xeric S. roundifolia has denser leaf trichomes as compared to riparian S. argentea (Sriladda et al., 2016), which indicates trichome density of Shepherdia species may be influenced by water availability. Shepherdia ×utahensis has leaf trichomes (Sriladda et al., 2016) and grows well in a variety of substrates (Chen et al., 2021). However, the effects of soil moisture level on trichome density have rarely been investigated. The hypotheses of this research are (1) the morphology and physiology of S. ×utahensis change at different substrate water contents, and (2) leaf trichome density is affected by cell size under drought. To test these hypotheses, the objectives of this research were (1) to evaluate the morphological and physiological responses of S. ×utahensis under various substrate volumetric water contents in a greenhouse and (2) to quantify the relationship between trichome density and water deficit.

Plant morphology and physiology in this study changed along with decreasing substrate matric potential that resulted from reduced substrate volumetric water contents (Table 1). As substrate volumetric water content decreased, S. ×utahensis leaves and stems dehydrated, and the proportion of visibly wilted leaves increased (Table 2). In addition, plant growth indices, relative chlorophyll content (SPAD reading), numbers of shoots and leaves, total leaf area and dry weight, and photosynthesis were impaired (Tables 2–4). These results are in line with previous studies that reported negative effects of water stress on aesthetic appearance, plant growth, and net assimilation rate of ornamental plants (Niu et al., 2006; Cai et al., 2012; Zhou et al., 2021). In this case, decreased stem water potential is best interpreted as a passive response resulting from the effects of decreased soil water potential and higher leaf evaporative demand (Table 2). Similarly, Rosa ×hybrid (rose) and Nerium oleander (oleander) decreased stem water potential in response to low substrate or soil water potential under drought conditions (Molz, 1981; Niu et al., 2006; Cai et al., 2012). Decreased substrate volumetric water contents also inhibited nodule formation in S. ×utahensis (Table 3), which suggested that infection of symbiotic actinobacteria was affected by water availability. Actinobacteria move with water in the soil, and the process of reaching and infecting the roots of host plants slows down when soil water content decreases (Huss-Danell, 1997).

Plant morphological and physiological acclimations were observed in this study. In response to drought, S. ×utahensis reduced midday stomatal conductance to a value close to 0 when substrate volumetric water content decreased (Table 4). Midday stomatal conductance is positively correlated to stomatal opening and plant water status (Zhang et al., 2013). When S. ×utahensis plants dehydrated as a result of decreasing substrate volumetric water contents, plants closed their stomata to reduce transpiration and stomatal conductance as a drought acclimation to maintain plant water status and prevent water losses and further dehydration (Martinez-Vilalta and Garcia-Forner, 2017). Although CO₂ uptake is limited when stomata are closed and stomatal conductance reduced (Aroca, 2012), plants had a lower proportion of visibly wilted leaves in this study or better aesthetic quality under drought conditions in other reports. These plants were considered drought tolerant in ornamental plant evaluations in semiarid regions in Australia and the United States (Zollinger et al., 2006; Kjelgren et al., 2009; Reid et al., 2017). Shepherdia ×utahensis reduced its midday stomatal conductance at lower water availability and can be considered as a low water-use landscape plant. Plants have the capacity of regulating stomatal conductance that is related to their habitat aridity. Kjelgren et al. (2009) reported that plants native to arid regions, such as Dianella revoluta ‘Breeze’ (‘Breeze’ blueberry lily) and Ptilotus nobilis (yellow tails), showed greater reduction of stomatal conductance compared with those from humid areas.

Because of restricted transpiration, plants with acclimation capability may reduce leaf size to enhance convective heat loss to mitigate heat stress that causes high leaf-to-air VPD and leaf wilting (Tables 3, 4; Nobel, 2009; Devi et al., 2015). The fact that leaf-to-air VPD increased when substrate volumetric water content decreased (Table 4) is likely a direct consequence of increased leaf temperature because leaf vapor pressure is estimated by leaf temperature. To avoid heat stress, leaf energy is balanced primarily using sensible heat loss under drought (Bowen, 1926). The efficacy of sensible heat loss relates to boundary layer resistance, which is positively correlated to leaf width (Nobel, 2009). Under drought conditions, cell division and leaf expansion are limited (Dale, 1988), and smaller leaves are beneficial for dissipating heat through convection and conduction to maintain leaf temperature close to air temperature (Nobel, 2009). In this study, S. ×utahensis produced smaller leaves under water stress (Table 3) and leaf size of plants grown at the substrate volumetric water content of 0.05 m³·m⁻³ was 51% smaller than those at 0.40 m³·m⁻³. This result is in line with previous studies that consistently reported reductions in leaf size under water stress for drought-tolerant ornamental plants (Mee et al., 2003; Zollinger et al., 2006). For instance, Zollinger et al. (2006) suggested that small leaves allow Lavandula angustifolia (English lavender) and Penstemon ×mexicali ‘Red Rocks’ (‘Red Rocks’ penstemon) to reduce water loss when irrigation intervals were increased from 1 week to 4 weeks. Toscano et al. (2014) also found that leaf size of Viburnum tinus ‘Lucidum’ (‘Lucidum’ viburnum) decreased by 19% to acclimate to drought stress.

Shepherdia ×utahensis decreased total leaf area under water stress as a result of reductions in leaf number and size (Table 3). However, plants with decreased total leaf area have fewer stomata and less light interception, which controls transpiration and leaf temperature, respectively (Zollinger et al., 2006; Toscano et al., 2014). Reduced total leaf area has been reported as a means of avoiding drought stress in ornamental plants such as Lavandula angustifolia, Pittosporum tobira (pittosporum), and Viburnum tinus ‘Lucidum’ (Zollinger et al., 2006; Toscano et al., 2014). The root growth of S. ×utahensis was enhanced at low substrate volumetric water content, while shoot growth was inhibited, resulting in a higher root-to-shoot ratio (Table 3), which helps plants to obtain water more efficiently. Rosa hybrida ‘Ferdy’ (‘Ferdy’ rose) and Populus cathayana (poplar) have been observed to increase root growth to maintain water status under water stress (Henderson et al., 1991; Yin et al., 2004). Drought-tolerant plants native to the western United States also produce small leaves and deep roots to reduce water demand and loss and increase water uptake (Mee et al., 2003).

In this study, as substrate volumetric water content decreased, leaves of S. ×utahensis curled as stem water potential became more negative. At the substrate volumetric water content of 0.05 m³·m⁻³, the leaf curling index was 0.17, suggesting that the light interception area was 83% that of flattened leaves. Similarly, Dianella revoluta ‘Breeze’ and Ctenanthe setosa (prayer plant) have been shown to minimize sunlight exposure through leaf curling under water deficit (Kjelgren et al., 2009; Nar et al., 2009). Although light-harvesting efficiency is reduced, leaf curling limits water loss from transpiration and protects plants from overheating to sustain photosystem functions and other biochemical/physiological processes (Hook and Hanna, 1994; Fitter and Hay, 2002). In addition, as the rooting substrate became drier in this study, specific leaf area decreased, indicating that leaves became thicker (Table 3), which prevented leaves from overheating. Plants may decrease specific leaf area to acclimate to water stress as reported in Ptilotus nobilis (Kjelgren et al., 2009).

The trichome density of S. ×utahensis in this study was affected by substrate water availability and plant water status (Table 5). Water-stressed S. ×utahensis produced densely packed trichomes, resulting in a silvery appearance, while well-watered plants had fewer trichomes to cover epidermal cells and exhibited a greener color (Table 5; Figure 4). Trichomes promote leaf reflectance (Figure 6), which helps balance energy and reduce heat stress (Ehleringer et al., 1976). Positive effects of trichomes on leaf reflectance of visible light have been reported on Verbascum thapsus (common mullein) and Salix commutata (undergreen willow; Wuenscher, 1970; Mershon et al., 2015). However, because trichomes are broad-spectrum reflectors (Bickford, 2016), the reflectance of PAR, blue, green, and red light is proportional to the trichome density (Figure 6). When substrate volumetric water content decreased, the reflectance of green light (530 nm) did not increase as much as blue light (450 nm) and red light (660 nm) due to the chlorophyll in the epidermal cells (Table 5). Increased leaf reflectance has been shown to sacrifice the efficacy of light-harvesting pigments (Bickford, 2016) and reduce the net assimilation rate when plants are grown in drier conditions. Previous research also suggested that trichomes improved the reflectance of near-infrared light (Ehleringer, 1988). However, in this study, denser trichomes produced in drier substrate did not affect near-infrared light reflectance of S. ×utahensis (Supplementary Figure S4). Slaton et al. (2001) reported similar results that near-infrared light reflectance was not affected by increased trichome density in 48 species. More studies are needed to evaluate the effects of trichomes on near-infrared reflectance.

Increased trichome density has smaller effects on decreasing gas exchange coed with the effects on leaf reflectance (Jordan et al., 2005; Moles et al., 2020). However, densely packed trichomes covering the stomata of S. ×utahensis may increase resistance to transpiration and reduce water loss (Table 5; Figure 4D). Leaf trichomes also increase leaf roughness and increase the laminar boundary layer to restrict air movement across leaf surfaces to reduce transpiration (Grammatikopoulos and Manetas, 1994). Eriogonum corymbosum (crisp-leaf buckwheat) and S. rotundifolia produce leaf trichomes for better protection from wind and to maintain water status (Miller, 2011). Densely packed trichomes add an atmospheric boundary layer that imposes additional resistance to water vapor diffusion (Nobel, 2009). However, CO₂ influx is also limited by the boundary layer resistance, decreasing the net assimilation rate (Ehleringer et al., 1976). Although trichome-induced boundary layer resistance has a smaller effect on transpiration than stomatal conductance (Parkhurst, 1976), it still provides an advantage for desert plants to survive in dry and hot conditions.

Trichome density changes genetically (adaptation) and environmentally (acclimation). The genetic regulation of trichome density of Caragana korshinskii (Korshinsk pea shrub) has been reported by Ning et al. (2016). However, it is unclear how xeric plants change their trichome density to acclimate to drought conditions. A negative correlation between leaf trichome density and leaf size or epidermal cell size occurred in this study (Figure 5A), which suggests that cell expansion may control trichome density. Low trichome coverage fraction, which was related to greater space between trichomes, showed when epidermal cell density decreased, indicating cell expansion may coordinate trichome density. Ascensão and Pais (1987) reported the number of trichomes is determined during leaf lifespan, and leaf cell differentiation does not affect trichome number. Similar results showed in our research that plants had similar total numbers of trichomes per leaf at different substrate volumetric water contents. This may indicate that S. ×utahensis develops trichomes independent of leaf development. In fact, trichomes develop at the early stage of leaf development and often earlier than stomatal development (Werker, 2000). For instance, trichomes of Inula viscosa (false yellowhead) are fully developed and reach mature size when leaves are 2 mm long; however, a mature leaf is 6–8 cm long (Werker and Fahn, 1981). Ocimum basilicum (basil) forms trichomes at an early stage of leaf development and trichomes then grow independently (Werker et al., 1993). In the same study, trichomes covered young leaves but became more widely spaced when leaf cells started to expand (Werker et al., 1993). In our study, the total number of epidermal cells per leaf was similar on plants at different substrate volumetric water contents, which indicates cell differentiation might have minor effects on regulating trichome density. In contrast, cell expansion might be the main factor for regulating trichome density because leaf size, epidermal cell size, and the space among trichomes changed along with substrate volumetric water contents and correlated significantly with trichome density of S. ×utahensis (Figures 5A,B). Ehleringer (1982) found a negative correlation between leaf size and trichome density of Encelia farinosa, but cell size was not determined. Cell enlargement at high soil moisture levels amplified leaf size and the space among trichomes, reducing the trichome density on the S. ×utahensis leaves in this study. The relationships between trichome density, epidermal cell size and density, and leaf reflectance might indicate changes in cell size predominantly controls trichome density to modify leaf reflectance.

Modifying leaf reflectance via the change in cell size helps rapidly acclimate to environmental change without compromising whole leaf function (Brodribb et al., 2013). Cell-expansion-driven leaf anatomic change has been widely reported on adjusting stomatal density (Xu and Zhou, 2008; Murphy et al., 2016). For instance, Murphy et al. (2016) observed that cell expansion was the predominant factor for coordinating vein and stomata density of eight angiosperm species under sun and shade. Stomatal density decreases and the size of guard cells increases when leaf water potential increases (Xu and Zhou, 2008), suggesting cell expansion not only enlarges the distance between epidermal appendages but also increases their size.

Environmental factors also promote leaf trichome density. Such factors include increased leaf-to-air VPD (Ehleringer, 1982; Shibuya et al., 2009, 2016) and drought (Quarrie and Jones, 1977; Banon et al., 2004), all of which negatively affect plant water status. For instance, high leaf-to-air VPD may increase water loss via transpiration, leading to plant dehydration. Leaf trichome density of Cucumis sativus increased when air humidity decreased from 90 to 20% at 28°C, causing leaf-to-air VPD to increase from 0.4 to 3.0 kPa (Shibuya et al., 2016). Shibuya et al. (2016) did not investigate cell or leaf expansion of C. sativus, but increased leaf-to-air VPD may promote trichome density because rising leaf-to-air VPDs reduces cell size (Murphy et al., 2014), making space between trichomes smaller. In this study, higher leaf-to-air VPD and smaller leaves were observed when S. ×utahensis plants grew at the lower θ_t and the smaller epidermal cell size resulted in greater trichome density. Therefore, because increased leaf-to-air VPD and drought led to a reduction in cell enlargement and denser trichomes in S. ×utahensis, leaf trichome density was regulated using turgor-pressure-driven cell expansion to acclimate to drought conditions.

As the soilless substrate became drier, S. ×utahensis exhibited poorer visual quality due to wilted foliage. Water stress also imposed negative effects on plant growth and gas exchange. When substrate water levels decreased, S. ×utahensis increased root growth to increase the ability to uptake water. Stressed plants also lowered total leaf area and stomatal conductance to reduce water loss via transpiration. Leaf temperature was regulated through smaller leaves, curled leaves, and densely packed trichomes. Substrate volumetric water content and stem water potential negatively affected trichome density, which helped reflect a broad spectrum of visible light under drought. An increase in cell size and leaf expansion may have regulated the trichome density. Under water stress, dense trichomes resulted from the limited cell expansion and small space between trichomes. In contrast, greater water availability increased cell size which promoted cell/leaf expansion and enlarged trichome size and the space between trichomes, leading to lower trichome density and improve light-harvesting efficiency.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

J-JC and YS conceptualized and designed the study and drafted the manuscript. J-JC performed the experiment and data analysis. KK, LO, SJ, and LH reviewed and edited the manuscript. All authors contributed to the article and approved the submitted version.

This research was supported in part by the United States Department of Agriculture (USDA) National Institute of Food and Agriculture (NIFA) Hatch project UTA01381, New Faculty Start-Up Funds from the Office of Research and Graduate Studies, the Center for Water-Efficient Landscaping, and the Utah Agricultural Experiment Station, Utah State University, and approved as journal paper number 9568. In addition, funding for this project was made possible by the USDA’s Agricultural Marketing Service through grant AM190200XXXXG005 [19-1044-110-SF] and grant AM170100XXXXG022 [190905].

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.