Acclimatization of Mediterranean Native Sages ( Salvia spp.) and Interspeciﬁc Hybrids in an Urban Green Roof under Regular and Reduced Irrigation

: Native-to-Greece sage species, namely, Salvia fruticosa , S. ofﬁcinalis , S. pomifera ssp. pomifera , S. ringens , S. tomentosa and interspeciﬁc hybrids, were evaluated for their acclimatization in an extensive Mediterranean green roof during summer under regular and reduced irrigation (every 2–3 days with substrate moisture 16–22% v / v and 4–5 days with substrate moisture 7–11% v / v , respectively). A substrate (grape-marc compost:perlite:pumice, 3:3:4, v / v ) that was 10 cm deep was used. Regardless of the irrigation frequency, S. pomifera ssp. pomifera × S. ringens and S. ofﬁcinalis × S. pomifera ssp. pomifera showed the highest survival of all hybrids and species, along with satisfactory growth, while S. fruticosa showed the lowest survival. Reduced irrigation resulted in the reduction of aboveground and root biomass, with no damage to the photosynthetic apparatus. S. fruticosa showed the highest (53%) aboveground biomass reduction and S. ofﬁcinalis, S. ofﬁcinalis × S. ringens and S. pomifera ssp. pomifera × S. ringens showed the lowest (28, 23 and 3%, respectively), while S. ofﬁcinalis × S. pomifera ssp. pomifera and S. pomifera ssp. pomifera × S. ringens showed the lowest reduction in root biomass (13 and 16%, respectively). With a reservation for S. fruticosa , Greek Salvia spp. and their interspeciﬁc hybrids studied in the present work are recommended for sustainable exploitation in extensive green roofs in arid regions and generally in xeriscaping.


Introduction
Green roofs (GRs) are one type of green infrastructure that can be applied to city buildings and provide significant environmental, social and economical advantages to the urban environment that mitigate the adverse effects of urbanization and make cities more safe, sustainable and resilient to the climate crisis [1][2][3][4]. GRs not only contribute to biodiversity increase [5] but also constitute the missing link between the built and the natural environment, which is required for ecological living in cities. GRs are a sustainable alternative to conventional roofs that provide multiple ecosystem services, including improved stormwater management, CO 2 sequestration and reduction in the urban heat island effect [6,7]. Apart from ecosystem services, GRs could increase the aesthetic value of buildings and the city overall, contribute to the socialization of the multistory building residents, support the psychosomatic health of the elderly by volunteering in GR maintenance and provide environmental education [8].
A factorial experiment with two factors, i.e., Salvia type (10 species and hybrids) and irrigation frequency (normal, sparse), was utilized. Therefore, 20 treatments were applied (10 plant types × 2 irrigation frequencies), and in each treatment, six containers were used, with two plants of the same type per container. The number of repetitions (n) is shown in each data table.

Irrigation Scheduling
For the first 10 days after planting, the plants were watered every second day for the plants to overcome transplant stress. On 14 April 2021, the plants were irrigated and then exposed to a preliminary drought experiment in order to determine the number of days that the plants could withstand without irrigation. Moisture (% v/v) of the substrate was recorded daily (three measurements from each container at 1900 to 2000 HR) using a handheld moisture meter (HH2; Delta-T devices, Cambridge, UK), with a soil moisture dielectric sensor (WET-2; Delta-T devices) inserted from the surface that measured 65 mm in depth and 45 mm in width. Most plants showed wilting symptoms 5 days after irrigation. On this day, the mean substrate moisture measured was 7-11% v/v. Therefore, this was decided to be the "sparse" irrigation frequency. The "normal" irrigation frequency was decided to be when the moisture of the substrate was 16-22% v/v and this was measured on day 3. Substrate moisture tests were carried out regularly and based on these, on 5 June 2021, the irrigation frequency was reset to every 2 days (normal) and 4 days (sparse) in order to have substrate moisture before an irrigation event similar to the previous experimental period. This irrigation schedule was kept until the end of the experiment.
Automatic drip irrigation on the substrate surface was applied before sunrise by two drippers placed at equal distances from the center of the container and the plants (dripper supply 4 L·h −1 , irrigation period: 60 min, adequate to allow water to drain off the container).

Meteorological Data
The ambient average, maximum and minimum daily air temperature; daily rainfall; and average daily wind speed during the experimental period (April to September 2021) are presented in Figure 1 (http://meteosearch.meteo.gr/, 23 February 2022 date of access). Regarding insolation, the monthly total radiation from April to September varied from 0.24 (September) to 0.32-0.37 KW m −2 (June to August). These data were recorded from a meteorological cage in the proximity of the experimental roof, however small differences between these data and the actual temperature and wind speed on the roof may have been present.
Automatic drip irrigation on the substrate surface was applied before sunrise by tw drippers placed at equal distances from the center of the container and the plants (dripp supply 4 L·h -1 , irrigation period: 60 min, adequate to allow water to drain off the co tainer).

Meteorological Data
The ambient average, maximum and minimum daily air temperature; daily rainfa and average daily wind speed during the experimental period (April to September 202 are presented in Figure 1 (http://meteosearch.meteo.gr/, 23 February 2022 date of acces Regarding insolation, the monthly total radiation from April to September varied fro 0.24 (September) to 0.32-0.37 KW m −2 (June to August). These data were recorded from meteorological cage in the proximity of the experimental roof, however small differenc between these data and the actual temperature and wind speed on the roof may have be present.

Plant Growth and Physiological Parameters Evaluation
In August (21st and 22nd), the day before and one day after irrigation, the leaf s matal resistance (Rleaf) and the maximum quantum yield of photosystem II (ΦPSIIo) we recorded.
Rleaf was measured with an AP4 Porometer (Delta-T devices) in two fully develop young leaves of each plant (the average value was recorded, i.e., n = 12) from 11.00-13 HR, as defined by the daily fluctuation of Rleaf.

Plant Growth and Physiological Parameters Evaluation
In August (21st and 22nd), the day before and one day after irrigation, the leaf stomatal resistance (R leaf ) and the maximum quantum yield of photosystem II (Φ PSIIo ) were recorded. R leaf was measured with an AP4 Porometer (Delta-T devices) in two fully developed young leaves of each plant (the average value was recorded, i.e., n = 12) from 11.00-13.00 HR, as defined by the daily fluctuation of R leaf .
The Φ PSIIo was measured before sunrise with a Photosynthesis Yield Analyzer (MINI-PAM, Portable Fluorometer, Walz, Effeltrich, Germany). One measurement per plant was taken, as described by Tassoula et al. [13], in eight randomly selected plants of each treatment (n = 8). The intensity of the measuring light of the MINI-PAM was set once so that the chlorophyll fluorescence yield base levels (Fo) were within the limits set by the manufacturer and held constant thereafter. Maximum fluorescence yield (Fm) was recorded by applying a saturation pulse of 12,000 mmol quanta·m −2 ·s −1 for a 0.8 s duration and Φ PSIIo was calculated as (Fm − Fo)/Fm.
At the end of the experiment (10 September 2021), it was not possible to separate the root system of each plant in the container; therefore, in the statistical analyses, the average values of aboveground and root biomass of the two plants of each container were used. The aboveground part of each plant was collected, its fresh weight was measured, placed in an oven at 70 • C for 7 days to dry and the dry weight was measured. In addition, the root systems of the plants per container were excised from the substrate, rinsed under running tap water in a sieve and their fresh and dry weights were measured.

Statistical Analysis
The significance of the results was tested using one-and two-way analysis of variance (ANOVA) (F test, discrete variables followed the normal distribution). The treatment means were compared using Student's t-test at p ≤ 0.05 (JMP 13.0 software, SAS Institute Inc., Cary, NC, USA, 2013).

Plant Growth and Survival
Five months (early April to early September 2021) after planting in an extensive green roof, the survival rates of Salvia spp. and their hybrids were affected by both the plant genotype and the frequency of irrigation ( Figure 2). Plant losses occurred sporadically during the June-September period, with most losses in July and August (monthly data not shown), which were the hottest, driest and most windy months (Figure 1a-c). S. fruticosa presented a low survival rate (42-50%), regardless of irrigation frequency, while its hybrid with S. ringens, although under normal irrigation, also showed a low survival rate; in sparse irrigation, its survival rate was among the highest (92%). S. ringens was the species with the highest survival rate (100%) under normal irrigation, but under sparse irrigation, its survival rate was significantly reduced to 67%. However, all three S. ringens hybrids showed a high survival rate under sparse irrigation (83-100%). The hybrids S. pomifera ssp. pomifera × S. ringens and S. officinalis × S. pomifera ssp. pomifera were the only Salvia types that showed 100% survival under sparse irrigation ( Figure 2).
The experimental factors (Salvia type and irrigation frequency) significantly affected the aboveground fresh and dry weights and the root dry weight, as well as the ratio of root/aboveground dry weight of the plants (Table 1). Sparse irrigation resulted in lower aboveground and root biomasses and root/aboveground dry weight ratio compared to normal irrigation. As shown in Figure 1b, during the experimental period, there were only three rain incidents, one in April and two close to each other on 12 and 14 June; therefore, the irrigation treatments were not affected by rain incidents. Concerning plant type, the highest aboveground biomass was recorded for the S. officinalis × S. ringens hybrid with no statistical difference from S. tomentosa. The same hybrid also showed one of the highest values of root biomass, along with S. officinalis × S. pomifera ssp. pomifera and S. officinalis, while the latter showed the highest of all species' and hybrids' root/aboveground dry weight ratio as well (Table 1, Figure 3).
during the June-September period, with most losses in July and August (monthly data not shown), which were the hottest, driest and most windy months (Figure 1a-c). S. fruticosa presented a low survival rate (42-50%), regardless of irrigation frequency, while its hybrid with S. ringens, although under normal irrigation, also showed a low survival rate; in sparse irrigation, its survival rate was among the highest (92%). S. ringens was the species with the highest survival rate (100%) under normal irrigation, but under sparse irrigation, its survival rate was significantly reduced to 67%. However, all three S. ringens hybrids showed a high survival rate under sparse irrigation (83-100%). The hybrids S. pomifera ssp. pomifera × S. ringens and S. officinalis × S. pomifera ssp. pomifera were the only Salvia types that showed 100% survival under sparse irrigation ( Figure 2). Under normal irrigation, the species S. fruticosa, S. tomentosa and S. pomifera ssp. pomifera and the hybrids S. officinalis × S. pomifera ssp. pomifera and S. officinalis × S. ringens developed the highest aboveground biomass compared to all other species and hybrids, followed by S. ringens, S. fruticosa × S. ringens and S. officinalis × S. tomentosa (Table 2, Figure 3). As for the root biomass, this was highest in S. fruticosa, S. tomentosa and S. officinalis and the hybrids S. officinalis × S. pomifera ssp. pomifera and S. officinalis × S. ringens, while S. officinalis presented the highest ratio of root/aboveground biomass of all (Table 2). Table 1. The effects of experimental factors, i.e., Salvia type (S. fruticosa, S. officinalis, S. pomifera ssp. pomifera, S. ringens, S. tomentosa, S. fruticosa × S. ringens, S. officinalis × S. pomifera ssp. pomifera, S. officinalis × S. ringens, S. officinalis × S. tomentosa, S. pomifera ssp. pomifera × S. ringens) and irrigation frequency (normal, sparse) on aboveground and root system growth parameters of sage species and interspecific hybrids after five months of growth (April-September 2021) in an urban Mediterranean green roof.    Significance § F one-way ANOVA ** ** ** ** ** * † Mean values (n = 6) (±SE) in each column followed by the same lowercase letter (a-f) did not differ significantly at p ≤ 0.05 using Student's t-test. § NS or * or **, non-significant at p ≤ 0.05 or significant at p ≤ 0.05 or p ≤ 0.01, respectively.

2-Way
Under sparse irrigation, the hybrid S. officinalis × S. ringens presented the highest aboveground biomass of all hybrids and species and the highest root biomass, similar to S. officinalis and the hybrid S. officinalis × S. pomifera ssp. pomifera, while S. officinalis presented the highest root/aboveground biomass ratio (Table 2, Figure 3).
Low irrigation frequency reduced both the aboveground and root system biomasses to varying degrees between species and hybrids ( Figure 4). Regarding species, the highest percentage of aboveground dry matter reduction using sparse irrigation was observed in S. fruticosa and the lowest in S. officinalis; the former showed the highest aboveground biomass reduction of all species and hybrids. As for the hybrids, S. officinalis × S. ringens and S. pomifera ssp. Pomifera × S. ringens showed the lowest reduction in aboveground dry matter, especially the latter, whose aboveground biomass was almost unaffected. The highest percentage of root dry matter reduction using sparse irrigation was observed in S. fruticosa and the lowest in species S. officinalis and S. ringens and in hybrids S. officinalis × S. pomifera ssp. pomifera and S. pomifera ssp. pomifera × S. ringens (Figure 4).  S. officinalis had the highest root/aboveground fresh and dry weight ratios under both irrigation frequencies compared to all other species and hybrids, which did not change using sparse irrigation, while S. fruticosa and S. fruticosa × S. ringens were the only ones to have a significant reduction in the root/aboveground fresh weight ratio under sparse irrigation ( Table 2).

Physiological Parameters
The day before irrigation, all species and hybrids had increased R leaf compared to R leaf one day after irrigation, and under sparse irrigation, the R leaf was significantly increased compared to normal irrigation in S. pomifera ssp. pomifera and S. ringens and in all hybrids, except S. pomifera ssp. pomifera × S. ringens, indicating water stress ( Figure 5). Figure 5. Effect of Salvia type and irrigation frequency on plants' Rleaf (s cm −1 ) the day before (a) and one day after (b) an irrigation even values (n = 12) in each figure followed by the same lowercase letter (a-j) did not differ significantly at p ≤ 0.05 using Student's t-test. § N p ≤ 0.05 or significant at p ≤ 0.01, respectively. Figure 5. Effect of Salvia type and irrigation frequency on plants' R leaf (s cm −1 ) the day before (a) and one day after (b) an irrigation event in August 2021. † Mean values (n = 12) in each figure followed by the same lowercase letter (a-j) did not differ significantly at p ≤ 0.05 using Student's t-test. § NS or **, non-significant at p ≤ 0.05 or significant at p ≤ 0.01, respectively.

ER REVIEW
All species and hybrids showed Φ PSIIo values above 0.8 under both irrigation frequencies, indicating normal operation of the photosynthetic apparatus ( Figure 6). Before irrigation, there were no differences in Φ PSIIo values between plants under normal and sparse irrigation (Figure 6a). Interestingly, one day after an irrigation event, the plants under the sparse irrigation regime showed increased Φ PSIIo values (significant in the two-way ANOVA). S. pomifera ssp. pomifera, S. tomentosa and S. officinalis × S. pomifera ssp. pomifera had higher Φ PSIIo values under sparse irrigation compared to those under normal irrigation (Figure 6b). Figure 6. Effect of Salvia type and irrigation frequency on plants' ΦPSIIο the day before (a) and one day after (b) an irrigation event in August 2 (n = 8) in each figure followed by the same lowercase letter (a-g) did not differ significantly at p ≤ 0.05 by Student's t-test. § NS or **, non-sign significant at p ≤ 0.01, respectively. Figure 6. Effect of Salvia type and irrigation frequency on plants' Φ PSIIo the day before (a) and one day after (b) an irrigation event in August 2021. † Mean values (n = 8) in each figure followed by the same lowercase letter (a-g) did not differ significantly at p ≤ 0.05 by Student's t-test. § NS or **, non-significant at p ≤ 0.05 or significant at p ≤ 0.01, respectively.

Discussion
The evaluation of native species with minimal water requirements under diverse growth conditions is considered a crucial stage in developing plants that are suitable for sustainable green roofs in semi-arid Mediterranean regions [9]. Apart from addressing the problem of limited water availability in such regions, biodiversity, including concern for pollinator reduction worldwide [11], and the preservation of the local character can be met via the use of native plants in landscaping and green roofs in particular [12,13].
Greek flora and especially macchia vegetation is rich in species with potential use as green roof plants and Salvia spp. are among them. S. officinalis and S. fruticosa were tested previously with promising results, especially S. officinalis, in terms of growth in extensive green roofs [21][22][23][24][25], while S. pomifera ssp. pomifera, S. ringens and S. tomentosa were exploited for the first time as landscape plants in the present study. Interspecific hybrids between these native-to-Greece Salvia spp. that were developed to serve the horticulture industry's need for new ornamental species were shown to grow successfully under limited water supply [20] and were also tested in this work as green roof plants.
The growth and survival of all five sage species and five hybrids in an extensive green roof during the hot and dry season of the Eastern Mediterranean were affected by both the plant genotype and the frequency of irrigation. S. fruticosa showed the lowest survival rate of all sage types under both normal and sparse irrigation. This response appeared to have been inherited in the S. fruticosa × S. ringens hybrid when grown under normal irrigation but not under sparse irrigation. S. fruticosa × S. ringens resembles S. fruticosa in height and branching (tall with few lateral shoots), but it has intensely hairy segmented leaves like S. ringens [20]. Thus, leaf morphology may be one of the reasons the hybrid had a higher survival rate under reduced irrigation compared to S. fruticosa.
In general, the hybrids acclimatized more efficiently to the green roof conditions compared to their parents and the hybrids of S. pomifera ssp. pomifera, with the latter being either a seed parent to S. ringens or a pollen parent to S. officinalis, showed the highest survival of all species and hybrids at both irrigation frequencies. Regarding hybrid parents, S. officinalis is considered suitable for use in green roofs [21][22][23], as supported by the present work as well, whereas S. fruticosa showed fast water consumption and irrigation requirements of at least 50% of the daily pan evaporation [25], where this partly explained the low survival rate of this species in the present work. Furthermore, all hybrids used in the present study, especially S. officinalis × S. pomifera and S. officinalis × S. tomentosa, survived drought stress better than S. fruticosa in greenhouse conditions [20].
Significant differences were observed in the growth of Salvia hybrids in the green roof, which seemed to have been determined by the characteristics inherited from each of their parents, as described by Papafotiou et al. [20]. For instance, the hybrids S. officinalis × S. pomifera ssp. pomifera and S. officinalis × S. ringens produced the greatest aboveground and root biomasses of all the hybrids. The high value of their aboveground biomass seemed to have been inherited from their pollen parent, while that of the root from their seed parent S. officinalis. The vigorous canopy growth, in combination with the rich root system of these two hybrids, probably contributed to their higher survival rate. This is reinforced by the fact that the root system is the main plant organ for adaptation to drought stress conditions [33,34], while large biomass allocation into the root system likely allows for the higher accumulation of reserves for sustaining post-drought recovery [29].
All species and hybrids, except for S. pomifera ssp. pomifera × S. ringens, showed a reduction in aboveground plant biomass under water stress, which is an avoidance mechanism caused by a dehydration process [35], leading to water loss reduction. In several plant species, under drought conditions, the root biomass is reduced less than the aboveground biomass, resulting in a higher root/aboveground ratio [34], which optimizes water uptake [36]. In the present work, all species and hybrids, except S. officinalis, S. ringens and S. officinalis × S. pomifera ssp. pomifera, showed a greater reduction in root biomass than aboveground biomass under sparse irrigation. Thus, S. officinalis, S. ringens and S. officinalis × S. pomifera ssp. pomifera were the only ones to show an increase (not statistically significant) in the root/aboveground biomass ratio, but this was not associated with a higher survival rate compared to the other species. The root system was found to play a key role in plant drought resistance in S. officinalis [30], a result that was also supported by our research since S. officinalis showed the largest root/aboveground fresh and dry matter under both normal and sparse irrigation.
Moreover, S. officinalis had the lowest reduction rate of all species in both aboveground and root dry matter under sparse irrigation. The opposite was observed in S. fruticosa and its hybrid S. fruticosa × S. ringens, whose aboveground and root dry matter, particularly that of the root, was much reduced under sparse irrigation. This led to a lower root/aboveground biomass ratio under deficit irrigation that was statistically significant in the case of fresh matter and observed only in these two plant types. However, although the S. fruticosa × S. ringens hybrid showed high plant mortality at the end of the hot and dry season under normal irrigation, similar to S. fruticosa, under sparse irrigation, in contrast to S. fruticosa, its plant mortality was among the lowest. Plant mortality in a green roof with low substrate depth (extensive type green roof) is affected by both drought and heat and is possibly affected to a larger extent by substrate temperature than drought per se [9]. The important role of temperature is further supported by a preliminary experiment (data not presented) we conducted the previous year (2020) with these five species of Salvia in the same green roof. In that year, S. fruticosa survived at much higher rates (83-91%) similar to those of the other species, probably because the maximum air temperatures in 2020 in the period June-August were 3-4 • C lower than in 2021.
Of all sages tested in the present study, S. officinalis, which is widely used worldwide as a medicinal and ornamental plant, was studied in terms of its drought resistance by many researchers [26][27][28]37]. Unlike other Mediterranean xerophytes, such as Origanum dictamnus, which are characterized by defoliation during the hot and dry season [13], which is a mechanism of adaptation to drought, sages did not show leaf loss under water stress in the present study. It was reported for several plant species that when exhibiting over 50% leaf desiccation in a green roof during drought, they were unable to survive [38]. In the present study, sage plants, regardless of genotype, that did not survive the end of summer did not show significant defoliation.
Chlorophyll a fluorescence is widely used in studying plant response and adaptation to stressful environments [39]. A slight decrease in Φ PSIIo values often accompanies mild water stress conditions, despite the significant reduction in the gas exchange parameters, while severe water stress causes pronounced effects on the Φ PSIIo parameter [40]. R leaf values one day before irrigation were increased, indicating water stress and Φ PSIIo values were reduced, possibly due to limited CO 2 supply at the carboxylation centers. However, the values of Φ PSIIo under water stress indicated that the Φ PSII photochemistry was functional. In addition, the recovery of Φ PSIIo values after irrigation at optimal levels indicated that no permanent photoinhibition had developed. This was consistent with previous findings for other Mediterranean xerophytes grown in extensive green roofs [13,41]. In the driest period (August), sage plants of all species and hybrids under both irrigation frequencies bore several light green-yellow leaves; however, the Φ PSIIo values were above 0.7, even in these leaves, indicating that there was no critical damage to the photosynthetic apparatus.
Prior to irrigation, there were no differences in Φ PSIIo values between plants under normal and sparse irrigation. However, the day after irrigation, the plants under sparse irrigation had higher Φ PSIIo values compared to those under normal irrigation. In August, after 4 months in the green roof, plants under sparse irrigation were smaller and possibly better acclimatized to green roof conditions compared to those under normal irrigation; it was argued that water stress can change the gene expression, morphology and plant physiology and improve photosynthesis [42,43].
Native-to-Greece Salvia spp. and even more so the interspecific hybrids, with a reservation for S. fruticosa, were shown to be suitable for use in extensive green roofs in the Eastern Mediterranean, even under deficit irrigation. In addition, as non-invasive species, they are recommended for sustainable exploitation in green roofs and xeriscaping in other regions with a semi-arid/arid climate as well.

Conclusions
The acclimatization of five native-to-Greece sage species, namely, Salvia fruticosa, S. officinalis, S. pomifera ssp. pomifera, S. ringens and S. tomentosa, and five interspecific hybrids in a Mediterranean extensive green roof was affected by both the plant genotype and the frequency of irrigation.
The hybrids S. pomifera ssp. Pomifera × S. ringens and S. officinalis × S. pomifera ssp. pomifera showed the highest survival rate of all Salvia spp. and hybrids tested, along with satisfactory growth both under regular and reduced irrigation frequency.
S. fruticosa showed the lowest plant survival rate under both normal and sparse irrigation and the S. fruticosa × S. ringens hybrid showed a low survival rate under normal irrigation.
All hybrids, except S. officinalis × S. tomentosa, grew more efficiently in a green roof under water deficiency compared to their parental species.
Greek Salvia spp. and their interspecific hybrids, with a reservation for S. fruticosa, are recommended for sustainable exploitation in green roofs in regions with a Mediterranean climate and generally in xeriscaping.