Distinct seasonal dynamics of responses to elevated CO2 in two understorey grass species differing in shade‐tolerance

Abstract Understorey plant communities are crucial to maintain species diversity and ecosystem processes including nutrient cycling and regeneration of overstorey trees. Most studies exploring effects of elevated CO2 concentration ([CO2]) in forests have, however, been done on overstorey trees, while understorey communities received only limited attention. The hypothesis that understorey grass species differ in shade‐tolerance and development dynamics, and temporally exploit different niches under elevated [CO2], was tested during the fourth year of [CO2] treatment. We assumed stimulated carbon gain by elevated [CO2] even at low light conditions in strongly shade‐tolerant Luzula sylvatica, while its stimulation under elevated [CO2] in less shade‐tolerant Calamagrostis arundinacea was expected only in early spring when the tree canopy is not fully developed. We found evidence supporting this hypothesis. While elevated [CO2] stimulated photosynthesis in L. sylvatica mainly in the peak of the growing season (by 55%–57% in July and August), even at low light intensities (50 µmol m−2 s−1), stimulatory effect of [CO2] in C. arundinacea was found mainly under high light intensities (200 µmol m−2 s−1) at the beginning of the growing season (increase by 171% in May) and gradually declined during the season. Elevated [CO2] also substantially stimulated leaf mass area and root‐to‐shoot ratio in L. sylvatica, while only insignificant increases were observed in C. arundinacea. Our physiological and morphological analyses indicate that understorey species, differing in shade‐tolerance, under elevated [CO2] exploit distinct niches in light environment given by the dynamics of the tree canopy.


| INTRODUC TI ON
In order to predict the responses of natural plant communities to future increases in atmospheric CO 2 concentration ([CO 2 ]), it is necessary to understand the different responses of the species and ecosystems to elevated [CO 2 ] and the ability of species to use newly established niches. This is particularly important for the understorey species since the light limitation can strongly affect their response to [CO 2 ]. Previous studies have shown that elevated [CO 2 ] often stimulates growth (e.g., de Graaff, van Groenigen, Six, Hungate, & van Kessel, 2006;Poorter, 1993) and photosynthesis (e.g., Albert et al., 2011), reduces stomatal conductance (Ainsworth & Rogers, 2007), increases water use efficiency (Curtis & Wang, 1998), and increases growth of the root system, particularly root length, and root-toshoot ratio (Anderson et al., 2010;Rogers, Peterson, McCrimmon, & Cure, 1992). Such physiological and anatomical modifications under elevated [CO 2 ] may increase water use efficiency in plants and reduce thus the adverse effects of drought stress (Ainsworth & Rogers, 2007;Tschaplinski, Stewart, Hanson, & Norby, 1995;Wang et al., 2018).
Most of the studies exploring effects of elevated [CO 2 ] in forest ecosystems have, however, been done on dominant overstorey trees under conditions of high light intensities (e.g., Asshoff, Zotz, & Körner, 2006;Norby et al., 2005;Urban et al., 2014), while understorey communities, naturally exposed to very low daily light integrals-photosynthetic photon flux density integrated over a day (DLI), received so far only limited attention (Belote, Weltzin, & Norby, 2004;Dukes et al., 2005;Niklaus & Körner, 2004;Würth, Winter, & Körner, 1998). These communities are, however, crucial to maintain species diversity, the stability of the habitat, and other ecosystem processes including regeneration of overstorey trees and nutrient cycling (Gilliam & Roberts, 2003). Understorey vegetation also plays a crucial role in utilizing new niches arising under changing environmental conditions (Gilbert & Lechowicz, 2004).
Indeed, reports of CO 2 stimulating effects on photosynthesis and related processes under low light intensities are contradictory. Urban et al. (2014) found reduced carbon gain and light use efficiency in temperate beech trees grown under elevated [CO 2 ] during cloudy sky conditions accompanied by low light intensity, low temperature, and high air humidity. In contrary, it has been shown that elevated [CO 2 ] stimulates the rate of photosynthetic CO 2 uptake under the conditions of deep shade and high temperature in the understorey of a tropical rain forest (Würth et al., 1998). Such sensitivity to [CO 2 ] is predicted to be caused by reduced photorespiratory carbon loss, increased apparent quantum efficiency, and accordingly reduced the light compensation irradiance of photosynthesis under elevated [CO 2 ] (Drake, Gonzalez-Meler, & Long, 1997;Farquhar, Caemmerer, & Berry, 1980;Hättenschwiler & Körner, 1996, 2000. All these studies, however, suggest that photosynthetic rate is modulated by combined conditions of elevated [CO 2 ] and low light intensities and may thus potentially alter the carbon balance of understorey plants as well as species composition. A meta-analysis by Kerstiens (2001) (Naumburg & Ellsworth, 2000). Particularly for deciduous and mixed forests, distinct light niches for understorey vegetation are available (Augspurger et al., 2005;Gilbert & Lechowicz, 2004). The first is represented by early spring with an open canopy before leaf out, which can be exploited by species with fast development, ability to utilize higher light intensities, and to survive under later deep shade. The second niche is exploited by typically shade-tolerant species, using mainly the higher DLIs during the summer months.
Such inconsistent results of responses of understorey vegetation to elevated [CO 2 ] may further rise from differences in soil water availability. For example, Belote et al. (2004) observed stimulatory effect of elevated [CO 2 ] on aboveground biomass production of Nepal grass (Microstegium vimineum)-an understorey dominant species in a dry, but not in a wet year.
In the present study, we explored responses of growth and photosynthesis to elevated [CO 2 ] in two grass species with C3 photosynthetic pathway grown in the understorey of an experimental spruce-beech stand. The studied grasses, Calamagrostis arundinacea (L.) Roth and Luzula sylvatica (Huds.) Gaud., represent widespread species of montane forests in Central Europe. Tuft forming C. arundinacea is an expansive and sun-demanding species occurring in the majority of disturbed forests and open deforested areas (Fiala, Tůma, Holub, & Jandák, 2005;Fiala et al., 2001). On the other hand, rhizomatous L. sylvatica is a highly shade-tolerant species, widespread over the temperate zone, and typically occurring in deep forest understories at low DLI (Godefroid, Rucquoij, & Koedam, 2005).
We tested the hypothesis that (a) elevated [CO 2 ] stimulates photosynthesis and growth of understorey plant species under natural low light intensities. More specifically, we have assumed that (b) species differing in shade-tolerance also have a different sensitivity to elevated [CO 2 ] due to a different composition and operation of the photosynthetic apparatus. Finally, we expected that (c) the stimulation effects of elevated [CO 2 ] are changing throughout the growing season following the changes in DLI and development of forest canopies enabling thus the species differing in shade-tolerance to use distinct niches in the light environment.

| Experimental plants and design
At the beginning of the growing season 2007, tillers of C. arundinacea and L. sylvatica were collected from an open area near the experimental station Bílý Kříž (Czech Republic; 49°33′N 18°32′E, 908 m a.s.l.) and subsequently exposed for four growing seasons to ambient (385 µmol CO 2 /mol; AC) and elevated (700 µmol CO 2 /mol; EC) [CO 2 ] using the glass domes at Bílý Kříž (see Figure S1, Šigut et al. (2015) and Urban et al. (2001) Table 1. Plants with comparable biomass and developmental stage were transplanted (data not shown). Plants were grown in the native soil. The geological bedrock is formed by Mesozoic Godula sandstone (flysch type) and is overlain by Ferric Podzols. The total soil nitrogen was found to range between 2.7 and 3.5 mg/g irrespective of [CO 2 ] treatment. Plants within each dome were split into five blocks (replications). Each block consisted of three plants of C. arundinacea and three plants of L. sylvatica. Two plants per block were evaluated, and the average from these two measurements was used for statistical analyses.
The site is characterized by a mean annual temperature of 6.7 ± 1.1°C and precipitation of 1,316 ± 207 mm (average ± standard deviation for the period 1998-2010). The year 2010, in which the measurements were made, was characterized by a mean annual temperature of 6.0°C, the maximal air temperatures in July (35°C), and an annual precipitation of 1,297 mm with the highest amounts of precipitation in mid-May and at the end of August and early September ( Figure 1). Light penetration into the tree understorey amounted to 80% before leaf development (May), while it was only 20% during the peak of the growing season (July-September). The daily maxima of photosynthetically active radiation (PAR) in the forest understorey amounted up to 300 µmol m −2 s −1 in May, but were only 175 µmol m −2 s −1 in October (Figure 2a). Daily light integral (DLI; Figure 2b), mean half-hour PAR values integrated over a day, ranged from 0.1 mol m −2 day −1 (cloudy sky autumn days) up to 14 mol m −2 day −1 (clear sky spring days).

| Gas exchange measurements
Seasonal courses (May 11-12, June 8-9, July 12-14, August 10-11,   Figure   S3) were subsequently modeled as a nonrectangular hyperbolic function of incident PAR using a Nelder-Mead algorithm  to determine values of apparent quantum efficiency

| Morphological and production parameters
Fully developed leaves of C. arundinacea and L. sylvatica, on which the physiological measurements were carried out, were sampled throughout the growing season (May-October) to analyze their dry mass and leaf area. The leaf area was determined by a leaf area meter LI-3000A (Li-Cor) and subsequently dried to constant mass at 60°C for 48 hr. In addition, a destructive sampling of total aboveand belowground biomass of five plants of both grass species was performed in August 2010. Plant parts were dried to constant mass at 60°C for 48 hr. Leaf mass area (LMA; leaf dry mass per leaf area) and the ratio between root and shoot mass (R/S) were calculated.

| Statistical analyses
The data were evaluated by means of an analysis of variance, using the statistical package STATISTICA 12 (StatSoft). Three-way ANOVA analysis was used to test the effect of species (C. arundinacea vs.  (Figures 3-7). The Fisher's LSD post-hoc test was used to evaluate differences between means. For the destructive analysis of above-and belowground biomass, the differences between means were tested using one-sample t tests. Significance levels are reported in the Figure 8 and tables as a significant with *p ≤ .05, **p ≤ .01, and ***p ≤ .001.

| RE SULTS
The three-way ANOVA of the whole dataset (

| Morphological and production parameters
In both grass species, leaf dry mass increased under EC as compared to AC, however, only significantly in July for C. arundinacea and in August for L. sylvatica. While no significant differences in leaf mass per area (LMA) were found in C. arundinacea, a significant increase in LMA, in response to the EC treatment, was observed in L. sylvatica during the whole experimental period, except in October (Figure 8).
Destructive sampling of experimental plants in August showed significant effects of species on shoot dry mass. While no significant response to EC in dry mass accumulation was found in C. arundinacea, a marked increase in root dry mass was observed in L. sylvatica

| D ISCUSS I ON
Climate change may lead to an increase of light intensity in forest understories due to triggered tree die-off and reduction of overstorey canopy (Royer et al., 2011) as well as its reduction when the overstorey leaf area is stimulated by EC conditions (Norby et al., 2005). The

| [CO 2 ] stimulation of photosynthesis at low light intensity
In general, the stimulatory effect of elevated [CO 2 ] on photosynthetic assimilation varies depending on the functional group and interactions with other environmental conditions. Ainsworth and Rogers (2007) concluded that trees are more responsive to elevated [CO 2 ] than other functional groups, including herbaceous understorey species. These conclusions are, however, mainly based on studies where the plants were exposed to high light intensities, while studies conducted on shade-acclimated leaves and understorey vegetation received little attention (Kim, Oren, & Qian, 2016;Valladares, Laanisto, Niinemets, & Zavala, 2016). In the present study, we found evidences supporting the hypothesis that EC substantially stimulates photosynthesis ( Figure 3) and partially also the growth (Figures 8 and 9) of understorey plants naturally exposed to low DLIs (0.1-14 mol m −2 day −1 ), that is, conditions when photosynthesis is limited particularly by an insufficient rate of electron transport and formation of electrochemical potential on thylakoid membrane (Farquhar et al., 1980;von Caemmerer, 2000). However, the analysis of photosynthetic light curves ( Figure S2) shows considerable species-specific differences in EC stimulation in response to light inten- in tropical understorey vegetation (Hättenschwiler & Körner, 1996, 2000Würth et al., 1998) and shade-acclimated shoots of P. abies (Marek et al., 2002). Besides role of carbon source and sink balance and limited feedback regulation of photosynthesis in understorey vegetation, such enhancements are also likely caused by a reduced photorespiration rate due to an increased ratio of intercellular [CO 2 ] to [O 2 ] (Drake et al., 1997;Farquhar et al., 1980;Way et al., 2015). In contrary to DeLucia and Thomas (2000), who observed the proportionately greater stimulation of J max by [CO 2 ], V Cmax to J max ratio remained constant in our study with two understorey grass species.

Mousseau
Moreover, it seems that C sink strength is not reduced in understorey plants as documented by positive [CO 2 ] effect on the growth of aboveground and belowground biomass (Figures 8 and 9). However, it should be emphasized that the degree of [CO 2 ]-induced enhancement of growth may be strongly reduced under the conditions of insufficient nutrient, particularly nitrogen, availability (Kim et al., 2016).

| Responses to elevated [CO 2 ] are speciesspecific
To test the hypothesis that species differing in shade-tolerance also have a different sensitivity to EC, L. sylvatica and C. arundinacea were investigated in this study. Higher values of A 50 and AQE together with lower LCP in L. sylvatica than C. arundinacea under AC conditions (Figures 3 and 5) confirmed that L. sylvatica is a more shadetolerant species than C. arundinacea. We found that EC conditions substantially stimulate the formation of above-and particularly belowground biomass of shade-tolerant L. sylvatica, while only insignificant increases were observed in C. arundinacea plants (Figure 9). This is in accordance with a higher [CO 2 ] stimulation of A 50 , A 200 , and A/R D ratio in L. sylvatica than in C. arundinacea, particularly in summer months. Also Kubiske and Pregitzer (1996)    It is hypothesized that the physiological mechanism behind the stimulatory effect of elevated [CO 2 ] on carbon gain under low light intensities includes an increase of AQE and a reduction of LCP (Osborne et al., 1998). While our study confirmed higher AQE values under EC conditions and particularly in C. arundinacea at the beginning of the vegetation season, the hypothesis of reduced LCP was not supported by our data ( Figure 5). In accordance with DeLucia and Thomas (2000) (Figure 6) leading to an overall shift of the A/PAR curves ( Figure S3). Accordingly, we conclude that increased carbon uptake in understorey plants under EC conditions is primarily caused by increased AQE, that is, reduced photorespiration rate.
[CO 2 ]-induced changes in biomass partitioning between shoots (S) and roots (R) also seems to be species-specific. Although both grass species showed an increase in root biomass and an increase in R/S ratio under EC conditions, these changes were significant only in the shade-tolerant L. sylvatica (Figure 9). Arnone et al. (2000) studied the response of root systems to elevated [CO 2 ] in intact native grassland ecosystems and found one group of plants with no change in the root systems, and the second group with growth increases of 38% in average. Increased root production under elevated [CO 2 ] could, however, be followed by increased root mortality and decomposition rates which may lead to only small changes in root biomass, particularly in high soil moisture conditions (Pendall, Osanai, Williams, & Hovenden, 2003). Differences in root growth stimulation under EC conditions can be explained by variety of mechanisms among which nutrient availability (especially nitrogen) plays a crucial role. Since the carbon investment into the root system is energetically disadvantageous, the plants increase the root system in response to EC only under nitrogen limiting conditions together with improved nutrient uptake by mycorrhiza (Arnone et al., 2000).
More pronounced growth stimulation of shade-tolerant species by elevated [CO 2 ] was confirmed in a meta-analysis by Kerstiens (2001). However, differences between shade-tolerant and shadeintolerant species only occurred at high DLIs (Poorter et al., 2019). Kerstiens (1998)  (e.g., ability to harvest light, water, and nutrients). Highest responses to elevated [CO 2 ] were thus found in species with generally low relative growth rate, low leaf nitrogen content, and high R/S ratio and LMA (Kerstiens, 2001). These are typical traits for L. sylvatica, which showed higher growth stimulation by EC, particularly in summer months with a closed canopy, but slightly increasing DLIs given by longer days and higher incident PAR above the canopy.

| Seasonality of responses to elevated [CO 2 ]
Pronounced seasonal pattern in V Cmax and J max was observed in both understorey grass species studied. In accordance with the study by Xu and Baldocchi (2003), maximum values of V Cmax and J max were recorded in spring after leaf expansion followed by minimal values during hot and dry summer months and partial recovery at the end of summer and autumn. Such seasonal patterns were found under the both [CO 2 ] treatments ( Figure 7).
Our results also support the hypothesis that the stimulatory effect of EC is changing throughout the growing season and is based on species-specific differences in shade-tolerance and developmental dynamics, allowing the two species to exploit different light niches during the season. The existence of two light niches in early spring and during the summer months exploited by typically sun-demanding and shade-tolerant understorey vegetation, respectively, has been proved in our experimental mixed forest (Figure 2b).
For C. arundinacea, the EC conditions led to an increase in A 50 (the most frequent light intensity of a forest understorey; Figure 3) and A 200 /R D ratio (proxy to carbon balance of leaves; Figure 6), particularly at the beginning of the growing season when the leaf area of the overstorey trees was not fully developed. On the other hand, these parameters were substantially stimulated in L. sylvatica by EC in the summer months (July-August) which can be attributed to F I G U R E 9 Mean values (columns) of shoot and root dry mass and root-to-shoot ratio (R/S) of Calamagrostis arundinacea and Luzula sylvatica developed in forest understorey at ambient (AC) and elevated [CO 2 ] (EC). The sampling was done in September 2010, that is, after 4 years of cultivation in AC and EC conditions. Error bars represent standard deviations. A t test was performed to compare differences between means of AC and EC treatments within individual plant species (n.s., non significant; *p ≤ .05; n = 5) significantly lower light saturation intensities, compared to C. arundinacea, and better utilization of low intensities during longer days.
Pronounced stimulation by EC in the summer months can also be associated with lower water availability, which was confirmed by reduced stomatal conductance (Figure 3). The EC generally increases iWUE ( Figure 4) and may thus reduce the negative impact of limited water availability (Valladares et al., 2016 Tschaplinski et al., 1995). In our study, however, the relatively even distribution of precipitation in July and August suggests that peak stimulation by EC during these months was more related to species-specific differences.
Seasonal changes in the gas exchange parameters were in accordance with the seasonal dynamics of leaf dry mass, LMA, and their enhancement ratios (Figure 8). Based on a meta-analysis, Poorter and Navas (2003)  C. arundinacea could be explained by higher production of flowering shoots in comparison with AC in this species (data not shown) and thus lower biomass allocation to vegetative leaves during flowering.
Also Jablonski, Wang, and Curtis (2002) reported significantly enhanced number of flowers and seeds in plants grown under EC in comparison with AC.

| CON CLUS IONS
Our data support the hypothesis that elevated [CO 2 ] increases photosynthetic carbon uptake and stimulates the growth of understorey plant communities. In addition, we confirmed the hypothesis that species with distinct dynamics of development and shade-tolerance utilize different light niches during vegetation season to profit from rising [CO 2 ]. In our study, the elevated [CO 2 ] stimulated particularly growth of shade-tolerant L. sylvatica that was able to sustain [CO 2 ]stimulated photosynthesis at natural light of low intensity during much of the growing season. In contrary, such [CO 2 ]-stimulated photosynthesis in sun-demanding C. arundinacea was found only during the spring months when the tree canopy was not fully developed, and the plants were exposed to relatively high DLI values. Finally, our results imply that understorey vegetation in the future could gain more importance in carbon sequestration and other ecosystem functions as it shows less evidence of photosynthetic downregulation, improved water use efficiency, enhanced amount of carbon accumulated in the biomass, particularly roots, and also high plasticity to changing light conditions given mainly by species-specific differences in the dynamics of development and shade-tolerance.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
P.H., K.K., and O.U. conceived the ideas, designed methodology, and analyzed data; P.H. and K.K. collected data; P.H. and O.U. led the writing of the manuscript assisted by S.L. and K.K. All authors contributed critically to the drafts and gave final approval for publication.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data used in this manuscript are presented in the figures and supporting information.