Drought Impact on the Morpho-Physiological Parameters of Perennial Rhizomatous Grasses in the Mediterranean Environment

: The selection of non-food crops for bioenergy production in limiting environments is a priority for energy security and climate change mitigation. Therefore, more studies are needed on the interactions between species and environmental factors in speciﬁc sites which allows their selection for biomass production. The objective of this work is to study the impact of drought on the morpho-physiological parameters of perennial rhizomatous grasses Panicum virgatum L., Miscanthus × giganteus , and Arundo donax L. in the Mediterranean environment. Plants were grown on ﬁeld and trials were carried out under support-irrigation and rainfed conditions during two consecutive years. Morpho-physiological parameters were measured in May, June and August, and dry biomass at the end of the experiment. Under rainfed conditions, A. donax presented the highest photosynthesis rate (25, 15 and 10 CO 2 m − 2 s − 1 ), relative water content (85–90%), and dry biomass (~4500 g plant − 1 ) compared with P. virgatum (20, 5 and 5 CO 2 m − 2 s − 1 , 65–85% RWC and ~1400 g plant − 1 ) and Miscanthus (18, 4 and 0 CO 2 m − 2 s − 1 , 80–10% RWC and ~260 g plant − 1 ). It is concluded that A. donax would be the best perennial rhizomatous grass to be used as bioenergy crop under Mediterranean conditions.


Introduction
The increased exploitation of renewable energy resources is expected to contribute to climate change mitigation and improve energy security [1].It is expected that bioenergy will lead the growth of renewable energy in the coming decades [2].In fact, bioenergy use has grown yearly by an average of 7% between 2010 and 2021 and its tendency is expected to increase and thus more efforts are needed to accelerate its use [3].Much of the biomass currently used for bioenergy is still obtained from food crops, such as Saccharum officinarum L. or Zea mays L. However, non-food energy crops, such as Phalaris spp., Salix spp., Populuss spp.Robinia spp., Eucalyptus spp., and Paulownia spp., among others [4], are also expected to play an increasingly important role [5].Therefore, the selection of bioenergy crops with no-food function, is an important priority for biomass production [6].
Three perennial rhizomatous grasses (PRGs), Panicum virgatum L., Miscanthus × giganteus (Greef et Deuter) and Arundo donax L., have been selected as candidates for bioenergy production out of almost 20 tested species [7,8].The P. virgatum is a warm-season C4 photosynthetic metabolism species native to the North American prairies with a diverse geographical distribution and it is adapted to a wide range of climatic and edaphic conditions [9].Miscanthus is also a typical C4 species, native to East Asia, found throughout a wide climatic range [10] and characterized by its rapid growth with low fertilizer requirements and high tolerance to drought, salinity and cold conditions [6]. A. donax has a C3 photosynthetic metabolism unlike the two previous species, but having as high photosynthetic rates as many C4 species [11].This species is native to Asia and has been successfully established in the Mediterranean region and subtropical wetlands [12,13].
The main agronomic traits that define these three bioenergy crops are high biomass productivity, vigor, early growth and regrowth capacity [14].All these species are characterized by a high content of lignin and cellulose that can be used for electricity, liquid fuel, biogas and hydrogen production, and to supply feedstocks for biorefineries [15].Therefore, the cultivation of the PRGs, compared to food crop systems or typical annual crops, represent a viable alternative to minimize competition for land use for food production, and their adverse effects on food security, greenhouse gas emissions (GHG) and the loss of biodiversity [4,16].
Furthermore, these PRGs have a deep and extensive root system which, in the long term, benefits the soil in terms of structure, stability and soil quality by acting as carbon sinks [17].In addition, PRGs have advantages over annual crops in terms of agricultural inputs (less fertilizers, pesticides and herbicides), production costs and harvest [18].Another attribute that makes PGRs appealing for biomass production is their hardiness, which determines their survival during prolonged dry periods, and tolerance to salinity, waterlogging and pests [14].Thus, these species can be grown on marginal lands where food production cannot take place because the land is not productive enough [19].
The most recent Intergovernmental Panel on Climate Change report [20] forecasts that climate change will expose plants to increasing occurrences of combined abiotic stresses, including drought, higher temperatures, and increased atmospheric CO 2 concentrations in the coming decades.This would be especially intense in the Mediterranean basin, which is characterized by 2-6 months of drought periods in summer and shorter drought winter periods from autumn to spring [21].Thus, the crops grown in the semi-arid Mediterranean area could be subjected to multiple stresses that are directly related to climate change.Due to increasing interest in the use of PRGs for bioenergy production, further studies on the interactions between species and environmental factors at specific sites, such as the Mediterranean region, can contribute to the understanding of the adaptive responses of these species to future climate change effects on biomass production [22].
Drought is the main environmental factor affecting plant physiological processes and negatively affecting crop production, from seed germination to adult plant growth and development [23].Morpho-physiological traits play essential roles in growth and development of the plants.However, under limiting water conditions, some species can develop different morpho-physiological adaptation strategies, which give them increased tolerance to drought stress [24][25][26].Thus, the identification of such drought-response traits in nonfood biomass crops candidates is crucial.Leaf gas exchange and chlorophyll fluorescence parameters, relative water content, plant height and above-ground biomass production are among the most studied morpho-physiological traits in candidate biomass species [21,27].We hypothesize that drought modulates the response of the morpho-physiological parameters of the studied species, which allows their selection for biomass production in a specific environment.Therefore, the objective of this work is to study the impact of drought on the morpho-physiological parameters of perennial rhizomatous grasses Panicum virgatum L., Miscanthus × giganteus, and Arundo donax L. in the Mediterranean environment.

Plant Material
Miscanthus × giganteus and Arundo donax L. seedlings were obtained from Piccoplant nursery (Pflanzenvertrieb und Verkauf GmbH, Oldenburg, Germany).Panicum virgatum L. (lowland ecotype, var.Alamo) seeds were donated by the INTA Anguil (La Pampa, Argentina).The seed germination of P. virgatum was performed according to the International Seed Testing Association [28] rules for this species.All the plants were transplanted to the field with four true leaves.

Experimental Design and Water Conditions
The field trial was carried out at the Experimental Fields of the Faculty of Biology, University of Barcelona, Spain (41 • 23 05.7 N 2 • 07 12.9 E) in a typical Calcic Luvisol soil.The physical and chemical characteristics of the experimental soil site have been described by Sauras-Yera et al. [29].In brief, loamy texture classes are observed in this soil with mostly illitic clays.The organic carbon content was low as a result of rapid mineralization of organic matter under semiarid conditions, which decreases with depth.The soil has a basic pH and detectable quantities of calcium carbonate, and therefore are calcium saturated.This soil is derived from colluvial-alluvial sediments which were affected by old pedogenesis that produced the characteristic red-brown color of many Mediterranean and subtropical soils.To prepare the soil, a motor cultivator (Pasquali 956/603) with a tiller was used at a working depth of 0.50 m.
The trials were conducted during 2013 and 2014 in the dry seasons (May-August).The meteorological conditions, such as mean temperature ( • C), precipitation (mm), and relative humidity (%), were registered by the meteorological station of the Department of Astronomy and Meteorology of the Faculty of Physics by the Experimental fields.
The experiment was arranged as a randomized complete block design, replicated three times for each species.The factors were watering levels (WA: watered) and rainfed (NW: non-watered) conditions.During the period May-August, the plants under the WA treatment were irrigated with 50 L m −2 month -1 supplementary water using droppers in an automated watering system.Sixteen plants were planted in 3.24 m 2 plots (1.8 m × 1.8 m) and a 2.7 plants m −2 density (0.6 m between rows, and 0.6 m between plants) based on the procedure stated mentioned in Cosentino et al. [30].At the beginning of spring, all plots were fertilized with 50 kg ha −1 N:P:K (15:15:15) fertilizer.During the two years of trial, no incidents of insects or diseases were observed, therefore, no pesticides or fungicides were applied.Different traits were measured in three times: May (T0), June (T1), and August (T2).

Physiological Measurements 2.3.1. Leaf Gas Exchange Parameters
Leaf-level gas exchange was measured using a Portable Photosynthesis System (Li6400, Li-Cor Inc., Lincoln, NE, USA) equipped with a 2 cm 2 Leaf Chamber Fluorometer (6400-40) and a 10% blue light source.The A/C i curves with chlorophyll fluorescence determinations were conducted in fully expanded leaves from each species at 25 • C leaf temperature, 21% O 2 and a light rate saturated at 1200 µmol mol −1 of PPFD.The curves started at 400 µmol mol −1 CO 2 , decreased stepwise until 250, 100, 50, and 0 µmol mol −1 CO 2 and restarted at 400 and increased stepwise until 750, 1000, and 1200 µmol mol −1 CO 2 .Gas exchange measurements were performed in the middle area of the last unfolded leaf.
Estimations of the maximum Rubisco carboxylation rate (Vc max µmol CO 2 m −2 s −1 ) and the rate of electron transport (Jmax µmol e-m −2 s −1 ) were calculated using equations developed by McMurtrie and Wang [33].
In addition to the response to CO 2 , photosynthetic light-response curves (A/PARi) were performed.The curves started at a maximum light level of 2000 µmol PAR m −2 s −1 and gradually decreased to 1200, 600, 300, 100, 50, and 30 down to 0 µmol PAR m −2 s −1 .Leaf temperatures of 25 • C, 21% O 2 and CO 2 at 400 µmol mol −1 were kept controlled.

Chlorophyll Fluorescence Analysis
Modulated chlorophyll fluorescence measurements were performed simultaneously using the Leaf Chamber Fluorometer coupled to Li6400 (Li-Cor Inc., Lincoln, NE, USA) to estimate the maximum quantum yield of PSII (F v /F m ), the efficiency of excitation energy capture by opened PSII reaction centers (F v '/F m '), the relative quantum yield of photosystem II (Φ PSII ) and the photochemical quenching (qp), which were determined in a completely expanded leaf after 30 min of dark adaptation [34].

Relative Water Content and Leaf Greenness
Relative water content (RWC, %) of the leaves was determined as (FW−DW)/ (TW−DW) × 100, where FW is the fresh weight, DW is the dry weight after drying in an oven at 60 • C until the weight was constant, and TW is the turgid weight of the leaves after equilibration in distilled water for 24 h at 4 • C [35].Leaf greenness was estimated in the last fully expanded leaves using a portable meter (Minolta SPAD 502 Meter, Plainfield, IL, USA) [36].Each measurement is the mean value of five measurements that were performed in the middle of each leaf.

Growth and Biomass Production
Plant height (H) was measured at T0, T1 and T2 from the soil surface to the end of the stem with a measuring tape.For the determination of dry weight (DW), above-ground biomass (leaves and stems) were harvested at the end of the growing period (T2), by cutting at 5 cm above-ground level.The samples were oven-dried at 60 • C until a constant weight expressed in g plant -1 was reached.

Statistical Analysis
The drought impact on leaf gas exchange parameter was assessed using a factorial ANOVA analysis within three studied factors.The modulated fluorescence of chlorophylls, RWC, leaf greenness, H and DW parameters were compared using one-way ANOVA.The means ± standard errors (SE) were calculated for each parameter.When a particular F-test was significant, the means were compared using a Tukey multiple comparison test.The results were considered significant at p < 0.05.Statistical analysis was conducted using the SPSS software package (version 20.0; Inc., Chicago, IL, USA).

Meteorological Conditions
Environmental temperature ( • C), relative humidity (%), and precipitation (mm) data are shown in Figure 1.In the first year, the total rainfall was 555.6 mm y −1 and the annual distribution of precipitation was bimodal, which is characteristic of a Mediterranean environment.The highest amount of rainfall was distributed in two periods: (1) autumn (147.8 mm) and spring (246.3 mm), being the months of April and May where there was more precipitation which favored the implantation of crops; and (2) the driest periods corresponded to winter (95.0 mm) and summer (66.5 mm); in the latter, August was the month with the least rainfall and coincided with the period of greatest crop growth.In the second year, total rainfall was 606.9 mm y −1 ; the highest amount of rainfall was again in autumn (284.5 mm) and spring (123.1 mm), but in this last period, rainfall was higher in April and May.Winter was the driest period (84.3 mm), but summer was wetter (114.9 mm), with higher rainfall in July and August, compared to the first year.The total rainfall received by the crops between May (T0) and August (T2) was 121 mm in the first year and 165 mm in second year, which represents a 36% increase.Plots under WA, that had supplementary irrigation during May-August, received a total of 200 mm extra water (dosed at 50 L m −2 month -1 ).Thus, WA plots received a total of 321 mm for the first year and 365 mm for the second year.
April and May.Winter was the driest period (84.3 mm), but summer was wetter (114.9 mm), with higher rainfall in July and August, compared to the first year.The total rainfall received by the crops between May (T0) and August (T2) was 121 mm in the first year and 165 mm in second year, which represents a 36% increase.Plots under WA, that had supplementary irrigation during May-August, received a total of 200 mm extra water (dosed at 50 l m −2 month -1 ).Thus, WA plots received a total of 321 mm for the first year and 365 mm for the second year.

Physiological Measurements
Significant differences (p = 0.001) were observed between species and treatments throughout the growing season for all physiological parameters and showed similar trends in the second year trial (Table 1, Table S1 and Figure S1).The highest Net CO2 assimilation rate (Asat), stomatal conductance (gs) (Table 1), maximum Rubisco carboxylation rate (Vc max) and the rate of electron transport (Jmax) (Table S1) values were measured at the beginning of the experiment (T0) and under WA conditions.The C4 species, P. virgatum and Miscanthus, did not present significantly different Asat, gs, Vcmax and Jmax values (~18 µmol CO2 m −2 s −1 , ~0.127 mol H2O m −2 s −1 , ~50 µmol CO2 m −2 s −1 and ~100 µmol e-m −2 s −1 , respectively).On the contrary, C3 species A. donax, showed the highest Asat, gs, Vcmax and Jmax values (~25 µmol CO2 m −2 s −1 , ~0.340 mol H2O m −2 s −1 , ~100 µmol CO2 m −2 s −1 and ~240 µmol e-m −2 s −1 , respectively) being even significantly higher than those form the C4 species.These values decreased in T1 and in T2, except for A. donax, which maintained their values similar to T0 under WA conditions in the first year.The three species were affected by drought under rainfed (NW) conditions, showing significantly decreased Asat, gs, Vcmax and Jmax values when compared to WA.In NW P. virgatum, Asat and gs values decreased around 80%, Vcmax decreased 55% in T1 and 85% in T2, Jmax decreased 66% in T1 and 77% in T2, in respect to WA and in both years.

Physiological Measurements
Significant differences (p = 0.001) were observed between species and treatments throughout the growing season for all physiological parameters and showed similar trends in the second year trial (Table 1, Table S1 and Figure S1).The highest Net CO 2 assimilation rate (Asat), stomatal conductance (g s ) (Table 1), maximum Rubisco carboxylation rate (Vc max) and the rate of electron transport (Jmax) (Table S1) values were measured at the beginning of the experiment (T0) and under WA conditions.The C4 species, P. virgatum and Miscanthus, did not present significantly different Asat, g s , Vcmax and Jmax values (~18 µmol CO 2 m −2 s −1 , ~0.127 mol H 2 O m −2 s −1 , ~50 µmol CO 2 m −2 s −1 and ~100 µmol e-m −2 s −1 , respectively).On the contrary, C3 species A. donax, showed the highest Asat, g s , Vcmax and Jmax values (~25 µmol CO 2 m −2 s −1 , ~0.340 mol H 2 O m −2 s −1 , ~100 µmol CO 2 m −2 s −1 and ~240 µmol e-m −2 s −1 , respectively) being even significantly higher than those form the C4 species.These values decreased in T1 and in T2, except for A. donax, which maintained their values similar to T0 under WA conditions in the first year.The three species were affected by drought under rainfed (NW) conditions, showing significantly decreased Asat, g s , Vcmax and Jmax values when compared to WA.In NW P. virgatum, Asat and g s values decreased around 80%, Vcmax decreased 55% in T1 and 85% in T2, Jmax decreased 66% in T1 and 77% in T2, in respect to WA and in both years.
Table 1.Net CO 2 assimilation rate at light saturation (Asat, µmol CO 2 m −2 s −1 ) and stomatal conductance (g s , mol H 2 O m −2 s −1 ) in Panicum virgatum L., Miscanthus × giganteus and Arundo donax L. plants grown under irrigated (WA) and non-water (NW) conditions during the first and second year at T0 (May), T1 (June) and T2 (August).Miscanthus plants exhibited foliar senescence under T2 and NW conditions; thus, these parameters could not be measured during either year of evaluation.Values represent the mean ± SE of nine replicates (n = 9).Different capital letters indicate significant differences (p < 0.05) between years for the same species, time and treatment.Different lowercase letters indicate significant differences (p < 0.05) between treatment for the same species, time and year.In Miscanthus, Asat and g s parameters have shown an 85% decrease during the first year in T1, and 60% during the second year.Vcmax also decreased 84% and Jmax decreased 64% in T1 both years.In contrast, the Miscanthus plants were senescent in T2.In A. donax, the Asat values decreased by 20 and 30% in T1, and 68 and 43% in T2, in the first and second year, respectively, while g s values decreased by about 65% in both times and years (Table 1).Vcmax values in A. donax only showed a significant decrease (~50%) in T2 and in both years, while Jmax values decreased about 10-30% in T0 and T1, and 45% in T1 and T2, in both years.

Panicum virgatum
In relation to intrinsic water use efficiency (WUEi) (Table S1) under WA conditions, significant differences were found between species, where P. virgatum and Miscanthus, showed a greater WUEi (~145 and ~140 µmol mol −1 , respectively) compared to A. donax (~72 µmol mol −1 ) at T0. WUEi values were also affected by rainfed conditions and showed a significant increase under NW compared to WA, except for Miscanthus at T1 and A. donax at T2 during the 1st year.
Concerning the CO 2 assimilation rate in response to the absorbed light (A/PARi) (Figure S1), the highest assimilation rates were observed in WA conditions in the three species and in both years.In turn, it was observed that over time, the assimilation rates decreased within each treatment, being more marked in NW.If we compare the assimilation rate under NW, P. virgatum and Miscanthus presented values below 10 µmol CO 2 m −2 s −1 while A. donax showed rates above this value under high light between 600 and 2000 µmol PAR m −2 s −1 .
In relation to the chlorophyll fluorescence parameters, the maximum quantum yield values of photosystem II (Fv/Fm) were between 0.7 and 0.8 in all species.No significant differences were observed among the species, treatments or times for each species or year of measurement (Figure S2).In contrast, ΦPSII (Figure S3), Fv'/Fm' (Figure S4) and qP (Figure S5) parameters decreased exponentially when increasing the photosynthetically active radiation (PAR) in all species.However, this decrease was greater in P. virgatum and Miscanthus C4 species and under NW conditions.The A. donax values did not show differences between WA and NW conditions.
Under WA conditions, all three species presented a high relative water content (RWC) percentage (~75-95%) throughout the crop growing season in both years.P. virgatum and Miscanthus C4 species decreased the RWC significantly under NW conditions.P. virgatum decreased 22 and 10% in T2 and T0 in the first and second year, respectively (Figure 2A,B).Miscanthus under NW treatment was the most affected species, showing a decreased 20 and 40% RWC in T0 and T1 in the first and second year, respectively and becoming dry at the end of the trial (T2) in both years (Figure 2C,D).In contrast, A. donax maintained RWC values above 80% under NW conditions in both years (Figure 2E,F).At the same time, A. donax presented the highest values of leaf greenness (40-50% relative leaf greenness) under WA and NW when compared to the C4 species (25-35% relative leaf greenness), and similar values were observed in the second year.Again, Miscanthus was the most affected species under NW conditions and showed a decreased leaf greenness between T0 and T1 (20 and 42% in the first and second year, respectively), and plants were completely senescent at T2 (Figure S6).

Agronomic Measurements
In relation to above-ground biomass, the highest DW was reached under WA conditions and for the three species studied.A. donax reached the greatest DW in both years (Figure 3).Under NW conditions, P. virgatum and A. donax presented a reduction in DW of around 30 and 25%, respectively, while Miscanthus was the species that showed the greatest decrease in DW, between 60 and 90% in the first and second year, respectively.The three species showed a significant increase in height under WA conditions throughout the whole growth period (T0, T1 and T2) (Figure S7).At T2, P. virgatum reached a height of approximately 2.00 m in both years (Figure S7A,B), Miscanthus was between 1.75 ± 0.10 and 2.47 ± 0.13 m high in the first and second year, respectively (Figure S7C,D) and A. donax was the tallest species reaching 3.00 ± 0.12 m height in the first year and 5.20 ± 0.25 m in the second year (Figure S7E,F).The C4 species growth was significantly lower in NW conditions.At the end of the trial (T2), the height of P. virgatum decreased between 40 and 30% and Miscanthus decreased between 50 and 80%, at first and second year, respectively.A. donax did not show significant differences between NW and WA treatments, except in T2 of the second year, where it grew more under WA conditions (Figure S7).

Agronomic Measurements
In relation to above-ground biomass, the highest DW was reached under WA conditions and for the three species studied.A. donax reached the greatest DW in both years (Figure 3).Under NW conditions, P. virgatum and A. donax presented a reduction in DW of around 30 and 25%, respectively, while Miscanthus was the species that showed the ± 0.10 and 2.47 ± 0.13 m high in the first and second year, respectively (Figure S7C,D) and A. donax was the tallest species reaching 3.00 ± 0.12 m height in the first year and 5.20 ± 0.25 m in the second year (Figure S7E,F).The C4 species growth was significantly lower in NW conditions.At the end of the trial (T2), the height of P. virgatum decreased between 40 and 30% and Miscanthus decreased between 50 and 80%, at first and second year, respectively.A. donax did not show significant differences between NW and WA treatments, except in T2 of the second year, where it grew more under WA conditions (Figure S7).

Discussion
The variation in the annual distribution of precipitation and water availability (Figure 1) in the period of greatest growth of these species impacted the morpho-physiological parameters, which is in agreement with Gulías et al. [21], Cosentino et al. [37] and Scordia et al. [38].The main differences at the morpho-physiological level between A. donax and the C4 species could be explained by their water requirements: 700-800 mm year −1 are required for Miscanthus, 450-750 mm year −1 for P. virgatum [39,40], and between 300 and 400 mm year −1 for A. donax [41].Thus, only A. donax water requirements were satisfied by rainfall during both years (555.6 and 606.9 mm year −1 , respectively).Therefore, C4 species have severely limited production potential under suboptimal conditions [37].Among the C4 species, Miscanthus was more affected by drought than P. virgatum, possibly due to its high water requirements.In the case of P. virgatum, although the annual precipitation would cover part of the range of water required, the lowland ecotype analyzed in this work is considered more susceptible to water stress than the upland ecotypes [40].
The growth and the biomass production are the result of the interactions between different factors, such as the physiology, water status, nutrient availability and plant morphology [30].Under NW conditions, lower growth and biomass production were observed, similar to other crops in the Mediterranean environment [42].At the same time, the decrease in water content causes the cells to lose turgor pressure [43], and the cells inhibit activities, such as cell division and cell expansion.It also inhibits the supply of organic and inorganic compounds necessary for the synthesis of protoplasm and cell walls, thereby directly affecting the growth of the entire plant [44].
Although A. donax is a C3 photosynthetic metabolism species, it showed the highest Net CO 2 assimilation rate, stomatal conductance, RWC, leaf greenness, height and biomass, even higher than those of C4 P. virgatum and Miscanthus species.The higher Asat and g s values from A. donax agree with previous studies of Rossa et al. [12], Sánchez et al. [13], Ceotto et al. [14], Papazouglou et al. [45] and Nackley et al. [46].Under NW conditions, the CO 2 assimilation rates decreased, mainly due to the decrease in g s , but not because of the decrease in photosynthetic electron transport rate as reported in P. virgatum and A. donax by Sánchez et al. [24].Furthermore, an increased leaf senescence was observed as a symptom of water deficit, mainly in Miscanthus at T2.This decreased photosynthesis response under water deficit could be a result of CO 2 diffusion limitation due to the decreased stomatal conductance, which would correspond to one of the immediate responses of plants under water deficit conditions in order to reduce water loss by evapotranspiration [43].
A. donax maintained a high RWC (>80%) during the growth period (Figure 2).On the contrary, RWC levels below 80% found in P. virgatum and Miscanthus under NW would imply a water potential of about ~1.5 MPa that causes changes in the metabolism and accumulation of proline and abscisic acid (ABA) [47].As a consequence, ABA levels increase rapidly in leaves and induce stomatal closure and a decrease in stomatal conductance (g s ), thus limiting photosynthetic activity and carbon fixation [48].Asat decreased under NW conditions, and the maximum velocity of RuBP carboxylation by Rubisco (Vc, max) and RuBP regeneration (Jmax) values also declined (Table S1).This response would be related to the loss or inactivation of Rubisco and with the reduction in sedoheptulose-1,7bisphosphate, a key regulatory enzyme in the Calvin cycle, as reported by Nogués and Baker [49].
The leaf greenness is also directly related to the photosynthetic activity [50].Our results showed a decreased leaf greenness in the C4 species when grown under NW conditions, as observed in other species under stress conditions [51].Contrarily, A. donax showed the highest leaf greenness and the highest photosynthetic rate (Figure S6).Therefore, it is considered a good indicator of the plant's metabolic conditions.
In general, although C4 plant species showed a higher WUE than C3 plant species because of a lower stomatal conductance, C4 plant species are oversensitive to water stress since they quickly close their stomata and therefore decrease their stomatal conductance, causing a decrease in CO 2 assimilation [52].Hence, in the face of a severe and prolonged water deficit, C4 plant species tend to be less tolerant than C3 plant species.Indeed, the C4 metabolism is an adaptation to increase the WUE under certain water limitations, but not against severe water stress [13,47].Under NW conditions, all Miscanthus plants presented foliar senescence with RWC < 20% at T2 because of the stomatal limitation caused by the strong water deficit they suffered.This difference between the NADP-ME (Miscanthus) and NAD-ME (P.virgatum) C4 photosynthetic metabolism subtypes in relation to their behavior towards the water deficit is due to a better adaptation to drought characterized by the NAD-ME C4 species.Thus, among the C4 species, the NAD-ME species are more tolerant to water deficit than NADP-ME and have a greater WUE [53], as the NADP-ME species are more frequently found in wetter environments [39,40].This feature would indicate that the productivity of Miscanthus in a Mediterranean environment without supplementary irrigation would not be feasible since it has higher water requirements.In other words, the possibility of irrigation in the Mediterranean environment could be beneficial for the cultivation of this species since it responds positively to the increase in the irrigation dose [7], but this would suppose a greater economic cost in the production chain.
Under NW conditions, all three species presented high F v /F m values and there were no significant differences between them under WA treatment (Figure S2).These results would indicate a good photosystem II efficiency or an absence of photoinhibition in the reaction centers of PSII, unlike other plant species that tend to decrease F v /F m as a response to biotic or abiotic stress, which would indicate the existence of photoinhibition damage [44,54,55].The ΦPSII, F v' /F m' , and qp parameters decreased with the increase in light intensity.However, differences were observed in the behavior of each species against NW conditions.In both studied C4 species, there was a difference between the results obtained under WA and in NW, the results being more pronounced in NW conditions.In A. donax there were no differences, which would indicate a greater sensitivity in the quantum efficiency of electron transport than in the C4 species under water deficit conditions [54].Therefore, these results would suggest that these species would be avoiding photoinhibition processes due to excess excitation energy through thermal dissipation processes to maintain the integrity of the photosynthetic membranes [55].
In summary, although rainfed treatment decreased all the parameters, some morphophysiological parameters, such as Asat and g s, showed better performance in A. donax compared to C4 species.This ability to maintain higher rates of CO 2 assimilation under drought conditions from the stomatal regulation would be one of the adaptive strategies of this species, which is in agreement with Cosentino et al. [30].

Conclusions
In the selection of non-food crops for bioenergy production under environmental stresses, the perennial rhizomatous grass emerges as the main candidate.In this regard, choosing the best-adapted species was favored to study the impact of drought on the morpho-physiological parameters.Arundo donax L. showed the best response in all the studied morpho-physiological parameters compared to C4 species.In contrast, Mischanthus × giganteus was less tolerant to these environmental conditions and showed a significant decrease in all physiological parameters resulting in reduced growth, biomass production and an early senescence.Therefore, in the context of climate change, with increasing restriction in water availability, A. donax would be the best perennial rhizomatous grass candidate to be used as a bioenergy crop in the Mediterranean region.
Results achieved in this work are novel since the most of the literature address the study of these species separately and/or under controlled greenhouse conditions.At the same time, the research of these species in this region and at field conditions has been scarcely studied in comparison with other Mediterranean regions, such as central Spain, Greece, northern and southern Italy.On the other hand, since the perennial rhizomatous grasses have not been widely studied at the biochemical and molecular level, unlike other species, more studies are needed to increase productivity and biomass quality under Mediterranean environments.

Figure 1 .
Figure 1.Monthly mean temperature (°C) (solid line), relative humidity (%) (dashed line), and precipitation (filled bars) per month during two consecutive years of the trial.Data were provided by the meteorological station of the Department of Astronomy and Meteorology of the Faculty of Physics at the University of Barcelona.Red arrows indicate the measurement moments (T0, T1 and T2).

Figure 1 .
Figure 1.Monthly mean temperature ( • C) (solid line), relative humidity (%) (dashed line), and precipitation (filled bars) per month during two consecutive years of the trial.Data were provided by the meteorological station of the Department of Astronomy and Meteorology of the Faculty of Physics at the University of Barcelona.Red arrows indicate the measurement moments (T0, T1 and T2).

Figure 2 .
Figure 2. Relative water content (RWC) in leaves of Panicum virgatum L. (A,B), Miscanthus × giganteus (C,D) and Arundo donax L. (E,F) grown under irrigated (WA) and non-water (NW) conditions during the first and second year, at different times (T0: May, T1: June and T2: August).Values represent the mean ± SD of nine replicates (n = 9).Different capital letters indicate significant differences (p < 0.05) between time for the same treatment.Different lowercase letters indicate significant differences (p < 0.05) between treatment for the same time.

Figure 2 .
Figure 2. Relative water content (RWC) in leaves of Panicum virgatum L. (A,B), Miscanthus × giganteus (C,D) and Arundo donax L. (E,F) grown under irrigated (WA) and non-water (NW) conditions during the first and second year, at different times (T0: May, T1: June and T2: August).Values represent the mean ± SD of nine replicates (n = 9).Different capital letters indicate significant differences (p < 0.05) between time for the same treatment.Different lowercase letters indicate significant differences (p < 0.05) between treatment for the same time.

Figure 3 .
Figure 3. Above-ground dry weight (g plant -1 ) from Panicum virgatum L. (A), Miscanthus × giganteus (B) and Arundo donax L. (C) plants grown under irrigated (WA) and non-water (NW) conditions during the first and the second year at time T2 (August).Values represent the mean ± SD of nine replicates (n = 9).Different lowercase letters indicate significant differences (p < 0.05) between treatments for the same year. L.