Contrasting strategies in morphological and physiological response to drought stress among temperate forest understory forbs and graminoids

Abstract

• Drought stress can profoundly affect plant growth and physiological vitality, yet there is a notable scarcity of controlled drought experiments focused on herbaceous species of the forest understorey.

• In this study, we collected seeds from five forb and four graminoid species common in European temperate forests. Seeds were germinated under controlled glasshouse conditions and subjected to moderate drought stress for 5 weeks. We assessed biomass partitioning, stomatal and leaf morphology, leaf gas exchange, minimum leaf conductance (g _min), and chlorophyll fluorescence parameters.

• Comparison of the two ecological guilds revealed that graminoids had a higher R/S, improved WUE, greater carboxylation efficiency, and enhanced non‐photochemical quenching under drought conditions compared to forbs. In contrast, forbs had significantly lower g _min, with higher total biomass and total leaf area. Despite these differences in morpho‐physiological functional traits, both groups experienced a similar relative reduction in biomass after drought stress. Key predictors of biomass accumulation under drought included photochemical quenching, stomatal traits, total leaf area and g _min. A negative correlation between biomass and g _min suggests that plants with lower residual water loss after stomatal closure can accumulate more biomass under drought stress. Additionally, g _min was positively correlated with guard cell length, suggesting that larger stomata contribute to higher residual water loss.

• Contrasting strategies in morpho‐physiological responses to drought define the differences between the two groups. In graminoids, drought resistance suggests greater emphasis on stress tolerance as a survival strategy. In contrast, forbs were able to maintain higher biomass and total leaf area, indicating a competitive strategy for maximizing resource acquisition.

Under drought stress, temperate forest understorey forbs prioritize biomass and leaf area retention, while graminoids enhance water‐use efficiency and photoprotection for survival.

INTRODUCTION

Climate change is expected to exacerbate the impact of drought stress on forest understories, profoundly altering their structure and function (Koelemeijer et al. ²⁰²²; Pozner et al. ²⁰²²; Tng et al. ²⁰²²; Deng et al. ²⁰²³). As global temperatures rise and precipitation patterns become more erratic, the frequency and severity of droughts are anticipated to increase (Cook et al. ²⁰¹⁸; Schuldt et al. ²⁰²⁰; Pokhrel et al. ²⁰²¹). The intensified drought stress can significantly reduce soil moisture levels, limiting water availability for understorey vegetation, which typically comprises shrubs, young trees, and herbaceous plants. These understorey species are crucial for biodiversity, providing habitat and food sources for a variety of wildlife (Cavard et al. ²⁰¹¹; Simonetti et al. ²⁰¹³; Botequim et al. ²⁰²¹). Herbaceous understorey plants are also significant contributors to the ecosystem carbon assimilation budget in natural forests (Petrík et al. ^2024a). However, their root systems, which are generally shallower than those of overstorey trees and some shrubs, could make them particularly vulnerable to water scarcity (Canadell et al. ¹⁹⁹⁶; McLachlan & Bazely ²⁰⁰¹). Prolonged drought conditions can impair photosynthesis, decrease growth, lower reproduction ability and increase mortality rates among understorey plants (Tng et al. ²⁰²²). Consequently, the resilience of the forest understorey to environmental stressors is compromised, potentially leading to long‐term ecological imbalances and decreased forest health and productivity. While a shift in the understorey community toward more adapted species may enhance resilience, it may also lead to a significant biodiversity reduction (Kopecký et al. ²⁰¹³; Chudomelová et al. ²⁰¹⁷).

Plants exposed to drought stress close their stomata to prevent excess water loss (Martin‐StPaul et al. ²⁰¹⁷; Petek‐Petrik et al. ²⁰²³; Alongi et al. ²⁰²⁴). This also prevents CO₂ intake which leads to a reduction in carbon assimilation via photosynthesis (Wilson et al. ²⁰⁰⁰; Gago et al. ²⁰²⁰). Lower assimilation negatively affects their carbon balance and less carbohydrates are available for growth (Attia et al. ²⁰¹⁵). Drought stress also enhances reactive oxygen species (ROS) generation, which can damage the photosynthetic apparatus, further reducing the carbon assimilation potential of plants (Mukarram et al. ²⁰²¹; Torun et al. ²⁰²⁴). Plants can partially mitigate these negative effects via changes in their morphological and physiological traits. Adjustment of morphological or physiological traits that have a positive impact on stress resistance is then called acclimation. For example, plants can change their carbon allocation patterns and invest more carbon into their root system rather than aboveground foliage (Ye et al. ²⁰²¹; Jiang et al. ²⁰²³; Reinelt et al. ²⁰²³). This increases their root‐to‐shoot ratio, which can improve ability to access water in the soil while potentially reducing overall transpiration (Bacher et al. ²⁰²²). Moreover, the development of thicker and smaller leaves with lower specific leaf area (SLA) and higher leaf dry mass content (LDMC) correlates with higher osmotic potential of the leaves (Chelli et al. ²⁰¹⁹; Blumenthal et al. ²⁰²⁰; Bhusal et al. ²⁰²¹) and higher water‐use efficiency (WUE) (Horike et al. ²⁰²¹; Zhong et al. ²⁰²²). Another potential acclimation response to drought stress can be the development of leaves with smaller stomata (Petrík et al. ^{2020, 2022}). Smaller stomata change faster in response to changing environmental conditions (Kardiman & Ræbild ²⁰¹⁸), and guard cell length (GCL) is negatively correlated with WUE (Petrík et al. ^{2023, 2024b}). Stomata size could potentially affect minimum leaf conductance (g _min), which is residual water loss through the cuticle after stomatal closure and incomplete closure of stomata, called ‘leaky stomata’ (Duursma et al. ²⁰¹⁹). Recent studies have identified g _min as a critical trait for drought survival. Species exhibiting lower g _min are better able to maintain water status, thereby delaying hydraulic failure and reducing drought‐induced mortality compared to species with higher g _min (Gleason et al. ²⁰¹⁴; Duursma et al. ²⁰¹⁹; Pereira et al. ²⁰²⁴). Several studies suggest that g _min may be plastic, with plants potentially acclimating to drought stress by reducing g _min (Le Provost et al. ²⁰¹³; Chen et al. ²⁰²⁰). In contrast, other research indicates that g _min is a conservative trait with low phenotypic plasticity and minimal acclimation potential (Schuster et al. ²⁰¹⁷; Slot et al. ²⁰²¹; Wang et al. ²⁰²⁴). Consequently, it remains an open question whether plants can modify g _min in response to drought stress. Furthermore, an increase in non‐photochemical quenching as a relief pathway for excess energy in chloroplasts can be seen as stress acclimation to protect the photosynthetic apparatus from oxidative stress (Vastag et al. ²⁰²⁰; Turc et al. ²⁰²⁴). These acclimation strategies are crucial for maintaining the survival and ecological functions of understorey herbs. However, the effectiveness of these responses might vary among ecological guilds and will be a determining factor in the resilience of forest ecosystems to the escalating impacts of climate change.

Forest understorey graminoids and forbs exhibit distinct drought responses due to their differing morphological and physiological traits (Johnson et al. ²⁰¹¹; van Sundert et al. ²⁰²¹). Graminoids, which include grasses and sedges, generally have narrow, linear leaves and high WUE, as their stomata tend to be smaller and more densely packed, reducing water loss through transpiration (Tobin et al. ¹⁹⁹⁹; Hommel et al. ²⁰¹⁴; Felsmann et al. ²⁰¹⁸). They often have extensive fibrous root systems that allow efficient water absorption from shallow soil layers (Šmilauerová & Šmilauer ²⁰⁰⁶). In contrast, forbs typically have larger leaf surfaces that can lead to higher water loss but also enable greater photosynthetic capacity under favourable conditions (Gobin et al. ²⁰¹⁵; Felsmann et al. ²⁰¹⁸). Forbs may develop deeper taproots, allowing access to deeper soil moisture, and can exhibit greater plasticity in their growth and phenology, adjusting their life cycles to periods of water availability (Zhang & Sun ²⁰²⁰). These differences result in graminoids often being more drought‐resistant due to their conservative water use strategies, while forbs can be more drought‐resilient, capable of rapid recovery following drought periods (Khan et al. ²⁰²²). This divergence in drought response strategies underscores the complex dynamics of forest understorey ecosystems in the face of climate change.

Despite growing awareness of the impact of climate change on forest ecosystems, there remains a significant lack of knowledge about the response of forest understorey herbs to drought (Archaux & Wolters ²⁰⁰⁶). While extensive research has been devoted to the effects of drought on woody plants and overall forest health (Allen et al. ²⁰¹⁰; Brodribb et al. ²⁰²⁰; Mantova et al. ²⁰²²), very few studies have addressed the response of understorey herbs and their communities—a critical component of forest biodiversity and function (cf. Iacopetti et al. ²⁰²¹; Koelemeijer et al. ²⁰²²). This study seeks to compare the drought responses of forbs and graminoids commonly occurring in temperate forest understories and to evaluate their acclimation potential. We anticipate that graminoids will exhibit greater resistance to drought stress compared to forbs, as evidenced by: (i) a lower relative reduction in biomass, (ii) smaller stomata, (iii) lower SLA, (iv) higher WUE, and (v) higher non‐photochemical quenching when exposed to drought stress. We further hypothesize (vi) that drought stress will cause a decrease in g _min in both forbs and graminoids.

MATERIAL AND METHODS

Plant material and seed sampling site

The subjects of our study were nine herbaceous plant species: four grasses—Bromus benekenii (Lange) Trimen, Carex digitata L., Melica nutans L., and Poa nemoralis L.; and five forbs—Geum urbanum L., Impatiens parviflora DC., Lamium maculatum L., Veronica sublobata M. A. Fisch., and Viola mirabilis L. These species are common in European temperate forest understories and, except for the introduced Impatiens parviflora, native in the study region. All are on the EuForPlant list of European forest plant species, mostly with status 1.1—taxa occurring mainly in closed forests (Heinken et al. ²⁰²²).

Seeds were collected directly from wild‐growing populations during the 2021 summer growing season (June–September). Three to five populations of each species were visited on random walks to collect ripe seeds. Between 500 and 1000 seeds were collected from each species, placed in paper bags and stored dry at room temperature (ca. 20°C). All seed material comes from one relatively homogeneous oak–hornbeam forest located on the northeastern slopes of Děvín Hill (550 m a.s.l.) in the southeastern part of the Czech Republic (48.873° N, 16.652° E). The site is characterized by a temperate to sub‐mediterranean climate with an average annual temperature of 9.6 °C and precipitation of 524 mm. In recent years, significant summer droughts have become more frequent, making it suitable for studying the response of understorey plants to drought at various temporal scales (Hédl & Chudomelová ²⁰²⁰).

Seedling germination, experimental design and environment

The seeds collected in the field were initially subjected to a period of cold stratification at 1–3 °C for 5 weeks. Following cold stratification, seeds from the nine studied species were transferred to Petri dishes containing filter paper and 8 ml distilled water. The Petri dishes were then placed in climate chambers (MLR‐352H‐PE) set at 70% relative humidity, with a photoperiod of 8 h light at 30 °C, and 16 h darkness at 20 °C.

Once germination was achieved, the seedlings were placed in a greenhouse (Mendel University, Brno, Czech Republic), from December 2021 to April 2022. The seedlings were potted into 36 plastic trays (40 × 30 × 7 cm) filled with 5 L of a soil mixture (1:1 peat soil:sand) and set in a greenhouse maintained at 20 °C during the day (13 h) and 12 °C at night (11 h). The soil mixture containing silica sand with a grain size of 0.3–0.8 mm and brown peat with a medium degree of decomposition (the AGRO CS company) was used. Each species represented by four trays was divided into two experimental groups: two trays for the control group (well‐watered) and two for the drought treatment group (water deficit). Each tray contained 12 seedlings of one species, in total 48 individuals per species and 24 individuals per treatment.

Initially, all trays were irrigated with 400 ml tap water every other day to maintain soil moisture conditions comparable to those in the Děvín forest. Following plant establishment, the control treatment continued to receive 400 ml tap water every other day for 5 weeks. In contrast, the drought treatment received approximately 150 ml tap water every other day to induce water deficit conditions. The irrigation quantity was adjusted based on continuously measured soil moisture using TMS‐4 probes (TOMST, Prague, Czech Republic). The desired level of volumetric soil moisture was maintained at approximately 20% for control and 10% for drought treatment. This corresponds to −0.3 and −3.0 MPa matric soil water potential, respectively. The volumetric soil moisture was recalculated into the soil matric water potential using the same substrate, 4 TMS probes and a WP4C Soil Water Potential Meter (Meter Group, Pullman, WA, USA). The temporal dynamics of soil water potential differed among control and drought and had similar dynamics among forbs and graminoids (Fig. 1).

Fig. 1.

Temporal dynamics of soil water potential (SWP) plotted against the day of experiment (DOE) for control (blue) and drought (red) treatment. Daily averages are presented for forbs (a) and graminoids (b).

Leaf morphology and biomass

After the 5‐week experimental period, individuals from the control and drought treatment groups were harvested for biomass and leaf morphological analyses. Individuals from each species and treatment group were carefully excavated from the soil and subsequently dissected into leaf, shoot, and root components. Before further analysis, soil was removed from leaves, shoots, and roots using a fine brush. The dissected parts of each individual were weighed with an analytical balance (KERN EWB 620‐2 M) to determine their biomass. All leaves from each individual were scanned using a standard flatbed scanner (Epson DS‐1630, 300 dpi), and total leaf area (TLA) determined using ImageJ software (National Institute of Health, USA). Previously scanned leaves were oven‐dried at 60 °C for 48 h and weighed using an analytical balance to obtain dry biomass. These measurements were used to determine the specific leaf area (SLA), calculated as the ratio of TLA to leaf dry weight. The shoot and leaf biomass of each individual were summed to obtain the aboveground biomass (AGB). The root‐to‐shoot ratio (R:S) was subsequently calculated as the ratio of belowground biomass (BGB) to AGB. Total biomass (B) was calculated as the combined sum of AGB and BGB. For SLA and LDMC, the sample size ranged from 12 to 24 individuals, while biomass metrics were assessed on 15–24 individuals per species and treatment.

Stomatal morphology

Stomatal imprints were collected at the end of the 5‐week drought stress period. The stomatal imprints were taken from fully developed apical leaves using the ‘kollodium method’ (Petrík et al. ²⁰²⁰), in which transparent nail polish was applied to the abaxial and adaxial sides of the leaves and imprints transferred to a microscope slide using transparent tape. The imprints were used for capturing digital photographs using a Levenhuk MED 30 T (Levenhuk, Tampa, USA) microscope equipped with Delta Optical DLT‐Cam Pro 12MPx (Delta Optical, Poland). One photograph of an imprint from each leaf and side was taken at 20 × 10 magnification. The stomatal morphological traits were further analysed with ImageJ software. The length of a guard cell (GCL) was measured on three randomly selected stomatal cells per photograph and then averaged per individual. The stomatal density (SD) was assessed as the total number of stomata per photograph (0.0415 mm²), recalculated to SD per 1 mm². The sample size for both GCL and SD ranged from 10 to 24 individuals per species and treatment. Collecting stomatal imprints was not feasible for some drought‐stressed individuals as their leaves were too small and wilted.

Minimum leaf conductance

The minimum leaf conductance (g _min, mmol m⁻² s⁻¹) was determined on five to six detached leaves for each species and treatment, using the “mass loss of detached leaves” method (Blackman et al. ²⁰¹⁹; Duursma et al. ²⁰¹⁹; Petek‐Petrik et al. ²⁰²³). This method involves monitoring branch weight over time under stable atmospheric conditions as the leaves desiccate. Leaves from the plants were excised and left to saturate in water overnight. The cut was isolated with the tape, leaving the leaf transpiring from both sides, while attached within a controlled‐climate chamber (Fytoscope FS 130, Photon Systems Instruments, Drásov, Czech Republic) to dry out. Within the chamber, relative humidity was maintained at 60% and temperature was set at 20 °C. The climate chamber was placed in an air‐conditioned room with a temperature set at 20 °C to avoid rapid changes in conditions when opening the chamber to conduct the measurements. The loss of leaf mass was measured with an analytical balance with a resolution of 0.1 mg (VWR TA314i, Leuven, Belgium) every 5–10 min for the first hour, and then every 30 min thereafter, until 20 data points were obtained. The g _min was then calculated from the slope of the linear part of the leaf mass loss (mmol) over time (s), as follows:

$$g_{\min }=\frac{\text{slope}\times \frac{\text{atmospheric pressure}}{\mathrm{VPD}}}{\text{leaf area}}$$

Following the measurement of g _min, the projected leaf area was determined using scans (Epson, DS‐1630) and analysed with Image J software to recalculate the g _min values per square meter.

Gas exchange and fluorescence measurements

The light‐saturated net assimilation rate (A), stomatal conductance (g _s), transpiration (E) and the intercellular concentration of CO₂ in the leaves (C _i) were measured with an infrared gas analyser (LI‐6800; LI‐COR, Lincoln, NB, USA). Intrinsic water use efficiency (WUEi) and the instantaneous carboxylation efficiency (A/C _i) were calculated as the ratio of A and g _s and A and C _i, respectively. All measured leaves were enclosed in the chamber and exposed to the following conditions: airflow of 600 μmol s⁻¹, CO₂ concentration of 400 μmol CO₂ mol⁻¹, relative humidity of 50–60% and the saturated photosynthetic photon flux density (PPFD) of 1000 μmol photons m⁻² s⁻¹. The leaf area enclosed into the chamber was 2 cm². Measurements were carried out randomly on 4–24 individuals per species and treatment in the week preceding biomass sampling. We took one measurement per leaf, with a 4‐s averaging period. Due to the small and fragile leaves, gas exchange measurements could not be obtained for drought‐stressed B. benekenii individuals.

The open‐state quantum efficiency of photosystem II (Fv′/Fm′), photochemical quenching (qP) and non‐photochemical quenching (qN) were also measured by infrared gas analyser LI‐6800 using rectangular multiphase flash (MPF) during the same measurement days as gas exchange measurements, on the same leaves. MPF was set to a margin of five points, an outrate of 100 data points per second, a maximum saturating pulse during the flash of 10194 μmol photons m⁻² s⁻¹ and modulation rate of 250000 Hz. The sample size of the fluorescence measurements corresponded to the gas exchange measurements and ranged from 4 to 24 individuals per species and treatment.

Statistical analysis

All statistical analyses were conducted using the R v 4.2.1 software (R Core Team, Vienna, Austria). The species‐level averages of all measured morphological and physiological traits under control and drought treatments, together with 95% confidence intervals, are shown in Table S1. As the sample sizes and variance between the groups were not homogeneous, we applied the Kruskal–Wallis test and Dunn's non‐parametric post‐hoc test to analyse the differences between the treatments and species (Table S2; Figures S1 and S2). We conducted the individual‐level Pearson correlation analysis between the traits separately for control and drought treatment (Figures S3 and S4) using the corrmorant package (Link 2020).

The species were further divided into two ecological guild groups for additional analyses. The forbs group included G. urbanum, I. parviflora, L. maculatum, V. sublobata and V. mirabilis, while the graminoid group comprised B. benekenii, C. digitata, M. nutans and P. nemoralis. Differences between the ecological guild groups and treatments were analysed using the Kruskal–Wallis test (Table S2) and Dunn's non‐parametric post‐hoc test (Figs 2 and 3). Coordination between traits and the distribution of ecological guild groups and control–drought treatments was tested by principal components analysis (PCA) at the individual level (Fig. 4). The representation of traits within the first three principal components of the PCA system was tested by cumulative Cos2 of the variables using the factoextra package (Kassambara & Mundt ²⁰²⁰). The relationship between selected traits at the individual level was tested by linear and logarithmic regression separately for the control and drought treatment across all species (Fig. 5).

Fig. 2.

Ecological guild level boxplots of total biomass (a), aboveground biomass (b), belowground biomass (c), root:shoot ratio (d), total leaf area (e), specific leaf area (f), leaf dry matter content (g), stomatal guard cell length (h) and stomatal density (i).

Fig. 3.

Ecological guild level boxplots of assimilation rate (a), transpiration rate (b), stomatal conductance (c), water use efficiency (d), carboxylation efficiency (e), minimal leaf conductance (f), open state quantum efficiency (g), photochemical quenching (h), non‐photochemical quenching (i).

Fig. 4.

Principal components analysis biplots for the first three principal components illustrating differentiation between control and drought treatments (a, c) and between ecological guilds (b, d). The square cosines for each variable across the three principal components display the representation of the variables in the PCA (e). For explanation of variable abbreviations see Figs 2 and 3.

Fig. 5.

The linear relationships between biomass and minimal leaf conductance (a), guard cell length and minimal leaf conductance (b), stomatal density and minimal leaf conductance (c) and the logarithmic relationship between stomatal density and guard cell length (d), for both control (blue) and drought (red) treatments.

RESULTS

Morphological response to drought stress

The drought stress had a significant impact on biomass and leaf morphology, but a limited impact on stomatal morphology in both forbs and graminoids. The Kruskal–Wallis test revealed that ecological guild, drought treatment and their interaction had a significant effect on total biomass (B), aboveground biomass (AGB), belowground biomass (BGB) and total leaf area (TLA) of the plants (Table S2). Drought stress imposed a reduction in all three biomass metrics and TLA for both ecological guilds (Fig. 2a–c,e). The absolute reduction in these metrics was greater for forbs than graminoids; however, the relative reduction was similar, with B reduced by 28% for forbs and 31% for graminoids. The biomass was consistently reduced between above‐ and belowground parts, resulting in the similar root‐to‐shoot ratio (R:S) of control and drought stress groups. Stomatal density (SD) was significantly higher for forbs, while R:S was significantly higher for graminoids. However, drought stress did not alter these traits (Fig. 2d,i). Specific leaf area (SLA) and leaf dry matter content (LDMC) did not differ between the forbs and graminoids, and drought stress caused a similar reduction in SLA and increase in LDMC in both groups (Fig. 2f,g). Graminoids exhibited significantly higher guard cell length (GCL) compared to forbs. Drought stress significantly reduced GCL in graminoids but had no effect on forbs (Fig. 2h). The species‐specific averages and variability intervals, together with Dunn's post‐hoc tests, can be found in Table S1; Figures S1 and S2.

Physiological response to water deficit

Drought stress slightly decreased intrinsic water use efficiency (WUEi) in forbs and reduced assimilation, transpiration, and photosynthetic efficiency in both forbs and graminoids. Non‐parametric tests revealed that drought treatment significantly affected light‐saturated net assimilation rate (A), transpiration (E), and stomatal conductance (g _s) in both plant groups (Fig. 3a–c). Whole‐plant transpiration (E × TLA) was 80.51 mmol s⁻¹ (±21.34 SE) for control forbs, 3.17 mmol s⁻¹ (±0.73 SE) for droughted forbs, 20.52 mmol s⁻¹ (±8.33 SE) for control graminoids and 0.77 mmol s⁻¹ (±0.38 SE) for droughted graminoids. The reduction in A was notably more pronounced in forbs compared to graminoids, resulting in a significant decrease in WUEi for forbs but not for graminoids (Fig. 3d). Additionally, the instantaneous carboxylation efficiency (A/Ci) under drought condition was significantly higher for graminoids compared to forbs (Fig. 3e). Although drought treatment did not significantly affect minimum leaf conductance (g _min) in either group, graminoids exhibited higher g _min than forbs under drought conditions (Fig. 3f). Moreover, drought treatment significantly reduced the open state quantum efficiency of photosystem II (Fv′/Fm′) in both groups (Fig. 3g). Graminoids consistently had lower Fv′/Fm′ values compared to forbs in both control and drought treatments. The photochemical quenching (qP) was reduced in forbs under stress, whereas no significant change was observed in graminoids (Fig. 3h). Conversely, drought treatment had a positive effect on non‐photochemical quenching (qN) in both forbs and graminoids (Fig. 3i), with graminoids exhibiting higher qN values than forbs in both control and drought conditions.

Coordination and trade‐offs between traits under drought stress

Drought stress led to reductions in biomass, leaf morphological adjustment, and increased stomatal and non‐stomatal limitation of assimilation in both forbs and graminoids (Fig. 4a,c). The differences in stomatal morphological adjustment, qP and qN responses to drought stress among forbs and graminoids resulted in a clear separation between the groups in the PCA (Fig. 4b,d). The first three principal components of the PCA explain all parameters well, except R:S and WUEi (Fig. 4e). The low representation of R:S and WUEi (low cos2 values) is related to their high variability within groups and a weaker influence of drought stress (R:S showed no change and only forbs differed in WUEi). The PCA revealed coordination between plant morphological and physiological traits, with biomass traits (B, AGB, BGB) positively correlated with TLA, SD, gas exchange parameters (A, E, g _s, A/Ci) and Fv′/Fm′, while negatively correlated with GCL, qN and LDMC (Fig. 4a,c).

We performed separate correlation analysis for the control and drought treatment, including both ecological guilts (Figures S3 and S4). The correlation matrix for the control group exhibited generally lower correlation strength and fewer significant relationships (24 at P < 0.05) between traits compared to the drought group (40 at P < 0.05). Specifically, B negatively correlated with g _min under drought conditions but not under control conditions (Fig. 5a). Additionally, g _min showed a positive correlation with GCL and a negative correlation with SD under drought (Fig. 5b,c). The negative relationship between GCL and SD represents a trade‐off because of spatial constraints on the leaf (Fig. 5d). Therefore, plants with lower g _min, lower GCL and higher SD accumulated more B under drought stress (Fig. 5 and Figure S4).

DISCUSSION

Knowledge on the effects of drought stress on the ecophysiological performance of forest understorey herbaceous plants obtained under controlled experimental conditions is extremely scarce, in contrast to the relatively extensive knowledge on woody plants. Our research revealed significant differences in drought responses between forest understorey forbs and graminoids, with graminoids exhibiting a greater capacity for acclimation to drought stress compared to forbs. We elucidated the relationships between key functional morphological and physiological traits, which are potentially critical for assessing drought resistance in forest understorey herbs.

Drought stress impact on morphological traits

Exposure to drought stress affected biomass accumulation, leaf and stomatal morphological traits, with the impact varying according to the ecological guild (Fig. 2). Reduced water availability commonly leads to biomass reduction by limiting photosynthetic assimilation through stomatal closure, which in turn constrains growth (Flexas et al. ²⁰⁰⁸; Pirasteh‐Anosheh et al. ²⁰¹⁶). Under well‐watered conditions, forbs generally exhibit higher aboveground and belowground biomass compared to graminoids (Grime & Hunt ¹⁹⁷⁵). Graminoids typically possess extensive, shallow, yet widely spread fibrous root systems (Bardgett et al. ²⁰¹⁴), whereas forbs often have deeper taproots or a combination of taproots and fibrous roots, allowing them to access moisture from deeper soil layers (Nippert & Knapp ²⁰⁰⁷). Despite this, no significant differences in belowground biomass were observed between forbs and graminoids under drought stress. On the other hand, consistent with the findings of Zwicke et al. (2015), we observed that graminoids maintained a higher R:S under drought stress compared to forbs. However, it is important to note that drought stress had no effect on R:S in either the forbs or the graminoids. A higher R:S typically indicates that plants can access more water relative to their transpiring surface area (Bacher et al. ²⁰²²). Consequently, the higher R:S maintained by graminoids may enhance their drought resistance relative to forbs. In our experiment, the shallow soil trays constrained the development of deeper rooting systems in the herbaceous plants, reflecting conditions similar to those of shallow forest soils. However, our results do not account for root competition with other species, particularly trees, or the influence of microbial and fungal communities present in deeper soils. These factors are critical in natural ecosystems and need further investigation in future research. The relative reductions of the total leaf area (TLA), specific leaf area (SLA) and increase in leaf dry matter content (LDMC) are comparable between the forbs and graminoids. This result contradicts our hypothesis that graminoids would have a smaller SLA in comparison to forbs. Although the reduction of TLA is typically associated with decreased assimilation and growth, it can also be interpreted as an acclimation response aimed at reducing overall transpiration and thereby slowing soil water depletion (Aspelmeier & Leuschner ²⁰⁰⁴; Gebhardt et al. ²⁰²³). SLA is a functional trait often used to characterize the resource acquisition strategy but also the potential stress tolerance of plants (Maire et al. ²⁰¹⁵). Low SLA and high LDMC are generally considered as drought adaptation as they correlate with a higher osmotic potential of leaves (Blumenthal et al. ²⁰²⁰; Bhusal et al. ²⁰²¹). Graminoids have narrow leaves with a high surface area‐to‐volume ratio, which helps reduce water loss through transpiration (Givnish 1978). This adaptation is particularly advantageous in dry or variable environments, allowing graminoids to conserve water and maintain hydration in conditions where it might be limited (Larcher 2003). In contrast, forbs have broader leaves that, while potentially increasing transpiration rates and water loss, also facilitate greater photosynthetic activity when water is available (Sage & Monson ¹⁹⁹⁹). Consequently, the lower TLA observed in graminoids results in reduced whole‐plant transpiration compared to forbs, leading to slower soil water depletion. However, this advantage may be limited in the presence of competing species, which can take up excess available water and potentially offset the benefits of lower transpiration in graminoids.

Forbs exhibited significantly shorter guard cell length (GCL), and higher stomatal density (SD) compared to graminoids, contrary to our hypothesis. Additionally, drought treatment resulted in a reduction of GCL in graminoids, but not in forbs. This indicates that graminoids may possess greater phenotypic plasticity in their stomatal traits compared to forbs (Franks & Beerling ²⁰⁰⁹). Unfortunately, we were only able to obtain the stomatal imprints from two graminoid species, compared to five forbs species, therefore the significant differences at guild level might, in reality, be weaker due to uneven sample size. Both GCL and SD can substantially influence stomatal conductance (g _s) and the overall stomatal responsiveness of plants (Harrison et al. ²⁰²⁰; Murray et al. ²⁰²⁰; Gibbs et al. ²⁰²¹). Furthermore, GCL has been shown to negatively correlate with water use efficiency (WUE), suggesting that plants with smaller stomata are better able to sustain carbon assimilation while conserving water (Petrík et al. ²⁰²³; Pereira et al. ²⁰²⁴). Stomatal morphology exhibits a degree of plasticity and can vary from year to year (Petrík et al. ²⁰²²). The observed reduction in GCL for graminoids in our study can be interpreted as a drought acclimation response (Liu et al. ²⁰²³). In contrast, the absence of GCL adjustment in forbs suggests a stronger genetic control of stomatal traits. These findings imply that graminoids exhibit greater resistance to drought stress in terms of growth and biomass allometry compared to forbs. Additionally, graminoids demonstrate higher plasticity in stomatal morphology, potentially enabling more effective acclimation to drought stress.

Physiological response to drought stress

Drought stress caused reductions in assimilation, transpiration and stomatal conductance in both forbs and graminoids (Fig. 3). Forbs also experienced a significant reduction of intrinsic WUE (WUEi) under drought treatment, compared to control. Conversely, the WUEi of drought‐stressed graminoids was significantly higher than WUEi of drought‐stressed forbs. Generally, high WUEi in grasses (Taylor et al. ²⁰¹⁰) enables them to accumulate more carbon, and also results in higher reproductive capacity compared to forbs under drought conditions. Graminoids exposed to drought stress also had significantly higher instanenous carboxylation efficiency (A/Ci) compared to drought‐stressed forbs. This means that graminoids were able to better mitigate non‐stomatal limitations of photosynthesis, such as Rubisco efficiency or electron transport rate (Alongi et al. ²⁰²⁴). Overall, graminoids showed greater resistance (lower relative reduction) of their assimilation and stomatal conductance than forbs. Similarly, research by Griffin‐Nolan et al. (2019) found that in drought‐stressed environments, graminoids generally experience a smaller reduction in these physiological parameters than forbs, which contribute to their greater drought resistance. Moreover, Lawson & Vialet‐Chabrand (2018) highlighted that graminoids are better adapted to fluctuating water availability, partly related to their ability to modulate stomatal conductance more effectively than forbs. Furthermore, significantly higher WUEi and A/Ci of graminoids under drought stress suggests they can utilize the water and carbon for assimilation more effectively than forbs. These findings underscore the functional differences between plant types and their strategies for coping with environmental stress.

The different strategy of forbs in coping with drought stress could also be seen in the residual water loss from leaves, characterized as minimum leaf conductance (g _min). This trait could play a decisive role in surviving the critical lack of water in soil (Duursma et al. ²⁰¹⁹). There were significant differences in g _min between the control and drought treatment for three out of nine species, but no significant changes in g _min from drought exposure at the ecological guild level for either forbs or graminoids. This supports the current understanding that g _min is a very conservative trait with limited phenotypic plasticity (Slot et al. ²⁰²¹; Wang et al. ²⁰²⁴). Nevertheless, forbs showed significantly lower g _min under drought stress compared to graminoids. Specifically, forbs had 50% lower g _min than graminoids but possess 57% greater TLA. Consequently, the total residual water losses (g _min × TLA) are comparable between the two drought‐stressed groups. Therefore, forbs and grasses employ different strategies to achieve similar total residual water losses at the whole plant level.

Both graminoids and forbs also experienced non‐stomatal limitation of photosynthesis, as apparent from reduction of open state quantum efficiency (Fv′/Fm′) under drought stress. The reduction of Fv′/Fm′ under drought stress is usually caused by oxidative damage to photosystem II (Colom & Vazzana ²⁰⁰³; Konôpková et al. ²⁰¹⁸). Graminoids showed significantly lower Fv′/Fm′ under drought stress compared to forbs. On the other hand, the non‐photochemical quenching (qN) of graminoids was significantly higher than for forbs under drought stress. Therefore, graminoids reduced their photochemical energy pathway in order to increase the qN and increase their photoprotection ability (Turc et al. ²⁰²⁴). There is also some evidence that increasing qN can lead to higher stomatal sensitivity and therefore faster stomatal closure, which can also have a positive impact on water retention of the plants (Głowacka et al. ²⁰¹⁸). The significantly higher qN in graminoids compared to forbs may enable them to dissipate excessive energy more efficiently and better protect their photosynthetic apparatus under drought stress.

Traits correlation

The PCA provided a clear morpho‐physiological division between the control versus drought treatment and between graminoids vs. forbs (Fig. 4). The analysis also revealed both coordination and trade‐offs between the traits. A more detailed examination of the relationships between the traits was then achieved with separate correlation analyses for control and drought treatments (Figures S3 and S4). The photosynthetic trait Fv′/Fm′ showed the strongest correlation with the biomass traits AGB, B and BGB, compared to A or A/Ci. Moreover, SLA was also positively correlated with the Fv′/Fm′ and biomass traits, confirming the general theorem of SLA being a functional trait reflecting the resource acquisition strategy of plants (Wilson et al. ¹⁹⁹⁹; Maire et al. ²⁰¹⁵). The strong correlation between Fv′/Fm′, biomass traits, and SLA suggests that the non‐stomatal limitation better explains biomass accumulation differences under drought stress than stomatal limitation effects (A, A/Ci).

An additional important trait coordinated with biomass accumulation under drought was g _min. We found a negative significant relationship between g _min and AGB, BGB and B, therefore individuals with lower g _min were able to accumulate more biomass under drought than individuals with higher g _min (Fig. 5a). A study by Machado et al. (2021) found a trade‐off between assimilation capacity and residual water losses in plants. However, we found no relationship between A and g _min in our study. Moreover, we found that GCL has a positive and SD a negative impact on g _min of the plants under drought stress (Fig. 5b, c). This suggests that not only the cuticle, but also stomata significantly affect the total g _min (Duursma et al. ²⁰¹⁹). Although we expected that the likelihood of ‘leaky stomata’ increases with higher SD, our results indicate a negative relationship between SD and g _min. This may be related to interspecies variation in stomatal traits, which can influence the expected relationship. For example, ‘leaky stomata’ accounted for 35% of g _min in Hedera helix (Šantruček et al. ²⁰⁰⁴) and for 50–94% of g _min in coniferous tree species (Brodribb et al. ²⁰¹⁴). While we cannot precisely quantify how much of g _min occurs through stomata versus the cuticle, the explanatory power of 39% (GCL) and 42% (SD) in linear regression with g _min theoretically corresponds to the value found experimentally by Šantruček et al. (2004). The g _min is a critical trait to predict the drought mortality timing in plants, as higher g _min means faster depletion of water reserves after stomatal closure (Petek‐Petrik et al. ²⁰²³; Waite et al. ²⁰²⁴; Ziegler et al. ²⁰²⁴). Our study suggests that g _min is also an important trait for biomass accumulation under drought stress, with plants having a lower g _min likely accumulating higher biomass because of better water retention capacity.

CONCLUSION

Distinct adaptive strategies in morphological and physiological traits affect the response of temperate forbs and graminoids to drought. Graminoid species demonstrated greater drought resistance, while forbs displayed potentially higher competitiveness, characterized by the ability to maintain higher total leaf area and biomass. Our findings revealed that minimum leaf conductance plays a significant role in biomass accumulation during drought stress. Specifically, individuals exhibiting lower minimum leaf conductance were able to accumulate more aboveground, belowground, and total biomass compared to those with higher minimum leaf conductance. This relationship emphasizes the significance of water retention mechanisms in relation to drought resilience.

AUTHOR CONTRIBUTIONS

R. H. and J. U. conceived the research idea. R. H., A. P. P., R. P. and J. U. designed the methodology. A. P. P., P. P., M. K., J. U. and M. M. carried out the measurements in the greenhouse and laboratory. M. K. and M. C. collected the seeds in the field. K. H. carried out the seedling germination. P. P. and R. P. analysed the data. A. P. P. and P. P. led the writing of the manuscript. All authors contributed critically to the drafts and gave final approval for publication.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

Supporting information

Table S1. Species‐level averages with 95% confidence intervals for all measured morphological and physiological traits under control and drought treatments.

Table S2. Results of Kruskal–Wallis test for morphological traits and physiological traits with species, ecological guild, water regime treatment, and their interactions as fixed factors (P < 0.05 *, P < 0.01 **, P < 0.001 ***).

Figure S1. Species‐level boxplots of total biomass (a), aboveground biomass (b), belowground biomass (c), root to shoot ratio (d), total leaf area (e), specific leaf area (f), leaf dry matter content (g), stomatal guard cell length (h) and stomatal density (i). BB—Bromus benekenii, CD—Carex digitata, MN—Melica nutants, PN—Poa nemoralis, GU—Geum urbanum, LM—Lamium maculatum, VM—Viola mirabilis, IP—Impatiens parviflora, VS—Veronica sublobata.

Figure S2. Species‐level boxplots of assimilation rate (a), transpiration rate (b), stomatal conductance (c), water use efficiency (d), carboxylation efficiency (e), minimal conductance (f), open state quantum efficiency (g), photochemical quenching (h), non‐photochemical quenching (i). BB—Bromus benekenii, CD—Carex digitata, MN—Melica nutants, PN—Poa nemoralis, GU—Geum urbanum, LM—Lamium maculatum, VM—Viola mirabilis, IP—Impatiens parviflora, VS—Veronica sublobata.

Figure S3. Pearson correlation matrix with linear trends for all evaluated traits under control treatment. Asterisks mark significant correlations at P < 0.05.

Figure S4. Pearson correlation matrix with linear trends for all evaluated traits under drought treatment. Asterisks mark significant correlations at P < 0.05.