HORTSCIENCE 51(5):608–614. 2016.

Response of Native Aromatic andMedicinal Plant Species toWater Stresson Adaptive Green Roof SystemsIro Kokkinou, Nikolaos Ntoulas, Panayiotis A. Nektarios1,and Dimitra VarelaLaboratory of Floriculture and Landscape Architecture, Department of CropScience, Agricultural University of Athens, Iera Odos 75, Athens 118 55,Greece

Additional index words. drought stress, leaf stomatal resistance, SPAD, deficit irrigation,substrate moisture

Abstract. The aim of this study was to determine the effects of different irrigationregimens on five native aromatic and medicinal species including Ballota acetabulosa(Greek horehound), Helichrysum orientale (helichrysum), Melissa officinalis (lemonbalm), Rosmarinus officinalis (rosemary), and Salvia fruticosa (Greek sage) when grownon adaptive green roof systems. The applied levels of irrigation were 100% (well-wateredcontrol), 75%, 50%, 25%, and 0% (no irrigation) of the daily pan evaporation (Epan).Measurements included the in situ determination of substrate moisture, stomatalresistance, and soil plant analysis development (SPAD) values. It was found that Greekhorehound, helichrysum, and rosemary can sustainably grow at an irrigation of 25%Epan, whereas Greek sage and lemon balm require an irrigation of at least 50% Epan forsustainable growth in shallow adaptive green roof systems.

The rapid development of contemporarycities has caused dramatic changes in urbanlandscape and climate. Increased populationdensity, particularly in city centers, in con-junction with the sealing of major landportions through building and constructionwork, have resulted in the lack of urban openand green spaces (Ferguson, 1998). All theabove have negatively affected the urbanmicroclimate, including air and water qual-ity, and have caused environmental deterio-ration, thus endangering public health, whilealso degrading life quality along with thecomfort and wellbeing of the inhabitants.

Green roofs provide contemporary tech-nical solutions that could increase urbangreen spaces and contribute to the ameliora-tion of environmental problems. Roofs andrecessed penthouses cover a large area of thebuilt urban spaces, especially in city areas

which are characterized by dense buildingnetworks. Several researchers have reported,or forecasted through modeling, that greenroofs could decrease ambient temperature(cooling) during summer, increase relativehumidity, reduce infrared, and diffuse radia-tion (Kumar and Kaushik, 2005; Simmonset al., 2008). Further advantages resultingfrom green roof implementation include ox-ygen production by photosynthesis (Getteret al., 2009), building energy savings forsummer cooling (Kotsiris et al., 2012a), re-duction of air pollutants (Czemiel Berndtsson,2010; Rowe, 2011), regulation of stormwaterrunoff and minimization of flooding events(Czemiel Berndtsson, 2010; Fioretti et al.,2010; Oberndorfer et al., 2007; Simmonset al., 2008; Van Woert et al., 2005), andamelioration of urban heat island effect (Akbariet al., 2001; Getter and Rowe, 2006).

Thus, the general concept is the develop-ment of a green roof networking which cancontribute to the improvement of the micro-climate in multiple ways if green roofs arelargely implemented in urban areas. There-fore, it is necessary to seek ways to constructthem on top of existing buildings. Because ofthe minimum load-bearing capacity of mostexisting buildings, either extensive or adap-tive green roofs (Nektarios et al., 2011, 2015;Ntoulas et al., 2012, 2013b) are appropriatefor green roof implementation. Extensivegreen roofs are characterized by minimalsubstrate depths (2–15 cm) that result in loadsbetween 20 to 120 kg·m–2. They are usuallyplanted with succulent plants and require lowor no maintenance (FLL, 2008). Adaptivegreen roof systems also use minimal sub-strate depth (5–15 cm), but are planted withvarious plant types such as aromatic and

medicinal plants, turfgrasses, and ground-covers. In contrast with the extensive greenroofs, the adaptive ones are accessible andrequire minimal irrigation inputs (Kotsiriset al., 2013; Ntoulas et al., 2012, 2013a,2013b; Ntoulas and Nektarios, 2015). Bothextensive and adaptive green roofs requirethe use of plant species that have adequatewater-stress tolerance and are capable ofgrowing in shallow substrate depths.

Plant species for either extensive or adap-tive green roof systems should be preferablynative and adapted to local environmentalconditions. Dimopoulos et al. (2013) reportedthat the prolific flora of Greece consists of5752 kinds of plant species (1278 endemic),of which many of them are herbs, aromatics,and pharmaceuticals that could be consideredappropriate for sustainable growth on Medi-terranean green roof systems in conjunctionwith minimal water inputs (Benvenuti andBacci, 2010; Kotsiris et al., 2012b, 2013;Nektarios et al., 2011, 2015; Papafotiou et al.,2013; Paraskevopoulou et al., 2015; Tassoulaet al., 2015).

The aim of the present study is to de-termine the irrigation threshold based on Epan

for five medicinal and aromatic plant species.The findings of this research may be valuablein selecting the most appropriate plants forMediterranean green roof systems and inpredicting the minimal required irrigationinputs for sustainable growth.

Materials and Methods

Experimental setupThe study was conducted on the roof of

the campus restaurant at the AgriculturalUniversity of Athens (lat. 37�59#, long.23�42#, 36 a.s.l.). The study period wasbetween 23 July and 9 Sept. 2013 and itwas conducted on 90 rectangular plasticcontainers with internal dimensions of24 cm · 24 cm · 11 cm.Within each container,a green roof system was simulated containingthree layers having a total height of 2.7 cm.The layers were placed in the following orderinitiating from the bottom of the containerstoward the substrate: a) protection and mois-ture retention fabric (VLS 300; Diadem,Landco Ltd., Athens, Greece); b) undulateddrainage (Diadrain 25H; Diadem, LandcoLtd.); and c) geotextile (VLF 150; Diadem,Landco Ltd.). As a growing substrate, a mixof pumice (75% v/v), clinoptilolite zeolite(10% v/v), peat (8% v/v), and compost fromgarden waste and dairy manure (7% v/v) wasplaced on top of the geotextile at a depth of8 cm. The physical and chemical propertiesof the substrate are listed in Table 1 whereasthe mechanical analysis and the water-retentioncapacity are listed in Table 2.

Plant materialThe study included five species of native

aromatic and medicinal plants of the Greekflora: B. acetabulosa (Greek horehound, 15plants),H. orientale (helichrysum, 20 plants),M. officinalis (lemon balm, 15 plants), R.officinalis (rosemary, 20 plants), and S. fruticosa

Received for publication 18 Nov. 2015. Acceptedfor publication 23 Mar. 2016.This article is part of anMSc thesis submitted to theDepartment of Crop Science, Agricultural Univer-sity of Athens, by Iro Kokkinou.The project (Urban BioRoof) with the code number12CHN136 was funded by the Hellenic GeneralSecretariat of Research and Technology under theOperational Programme ‘‘Competitiveness andEntrepreneurship’’ (EPAN II) and by the RegionalOperational Programmes of the five Regions oftransitional support, under the Action ‘‘BilateralResearch and Technological Cooperation betweenGreece and China 2012–2014.’’Themention of a trademark, proprietary product orvendor does not imply endorsement by the authorsnor does it imply approval to the exclusion of otherproducts that may also be suitable.1Corresponding author. E-mail: pan@aua.gr.

608 HORTSCIENCE VOL. 51(5) MAY 2016

(Greek sage, 20 plants). Differences in thenumber of plants used were due to plantlosses during their propagation from the wild.The plants were collected from Euboea(lemon balm), Attiki (rosemary), the islandsof Kythnos (Greek horehound), and Crete(helichrysum and Greek sage). They werecollected as cuttings from their natural hab-itat in May 2012 except for lemon balm thatwas collected in June 2012. Plant propagationwas performed under mist for 5–8 d depend-ing on the species at the Laboratory ofFloriculture and Landscape Architecture ofthe Agricultural University of Athens. Aftertheir propagation, the plants were acclima-tized for 2 weeks and then transferred to theexperimental roof in 10-cm diameter potswhere they grew for 1 year.

On 16 June 2013, a single plant wastransplanted into each experimental containerthat was equipped with green roof layers andsubstrate. At the initiation of the study, theheights of the plants were similar betweenreplications of each species.More specifically,the Greek horehound had a height of 19.2 cm(±1.2 SE), helichrysum 14.2 cm (±0.5 SE),lemon balm 21.2 cm (±0.9 SE), rosemary17.8 cm (±1.1 SE), and Greek sage 25.5 cm(±1.8 SE). After transplanting, all containerswere placed on the building roof under mildshade for 20 d and irrigation was provided asneeded. On 5 July 2013, the plants weretransferred to full sun. On 23 July 2013, irriga-tionwas applied to saturation to produce uniformwet conditions in between treatments at theinitiation of water-stress period.

The water-stress period lasted from 24July to 9 Sept. 2013. During the water-stressperiod, five different irrigation regimenswere implemented based on daily evapora-tion (Epan), which was determined by a ClassA evaporation pan. For each plant species,irrigation was applied daily either at 100%Epan (well-watered control) or at 75%, 50%,25% (as deficit irrigation treatments), or 0%(drought) of Epan. All treatments werewatered by hand at 0800 HR using a wateringcan equipped with perforated nozzle to en-sure uniformity.

Experimental design, treatments, andstatistical analysis

The plot arrangement followed a com-pletely randomized design with helichrysum,rosemary, and Greek sage having four repli-cations and Greek horehound and lemonbalm having three replications per treatment.To determine the irrigation threshold forsustainable growth of each of the five plantspecies, a comparison between the five dif-ferent irrigation regimens (100%, 75%, 50%,25%, and 0% of daily Epan) was performedfor each plant species independently. One-way analysis of variance was performed onthe collected data employing the statisticalanalysis software Statgraphics Centurion,version 15.2.11 (Statpoint Technologies Inc.,Warrenton, VA). Our main research interestfocused on plant response based on the imposedirrigation regimen on each sampling date.Treatment means for all statistical analyses

were separated using the Fisher’s protectedleast significant difference at a 0.05 P level(P < 0.05).

MeasurementsSubstrate moisture content. During the

water-stress period, substrate moisture content

Fig. 1. (A) Daily maximum, minimum, and average air temperatures and (B) Class A-pan evaporation forthe experimental period (24 July to 9 Sept. 2013). No rainfall occurred during the study period.

Table 2. Mechanical analysis and water-retention capacity of the green roof substrate that comprised ofpumice (75% v/v), clinoptilolite zeolite (10% v/v), peat (8% v/v), and compost from garden waste anddairy manure (7% v/v). Values represent the mean values of three replications (±SE).

Mechanical analysis Water-retention capacity

Particle size (mm) Retained (% w/w) Suction (cm) Water content (% v/v)

>10 mm 3.77 (±0.47) 0 32.5 (3.54)10–8 mm 1.54 (±0.26) 10 29.4 (±1.57)8–4 mm 19.30 (±1.01) 20 27.9 (±3.15)4–2 mm 28.08 (±0.33) 30 26.8 (±1.44)2–1 mm 13.28 (±0.39) 40 26.1 (±0.65)1–0.5 mm 8.69 (±0.36) 50 25.7 (±1.95)0.5–0.25 mm 6.61 (±0.30) 60 25.2 (±2.16)0.25–0.1 mm 7.28 (±0.40) 70 24.3 (±0.72)0.1–0.05 mm 4.76 (±0.23) 80 23.1 (±0.62)<0.05 mm 6.69 (±0.95) 90 22.6 (±1.23)

Table 1. Physical and chemical properties of the substrate that comprised of pumice (75% v/v),clinoptilolite zeolite (10% v/v), peat (8% v/v), and compost from garden waste and dairy manure(7% v/v). Values represent the mean values of three replications (±SE).

Measurement Unit Value (±SE)

pH 8.36 (±0.02)Electrical conductivity mS·cm–1 262.7 (±7.80)Weight at saturation g·cm–3 1.30 (±0.05)Weight at maximum field capacity g·cm–3 1.20 (±0.03)Dry weight g·cm–3 0.85 (±0.03)Total porosity % 44.4 (±2.30)Hydraulic conductivity mm·min–1 7.56 (±0.53)

HORTSCIENCE VOL. 51(5) MAY 2016 609

MISCELLANEOUS

was determined every 3 d at 1500 HR, with theuse of aWET-2 (Delta-T Devices, Cambridge,UK) frequency domain dielectric soil moisturesensor having rods of 65 mm in length, andspaced 45 mm apart. The sensor measuredmoisture content from a cylindrical substratevolume of �500 mL, and it was connected toa HH2 handheld moisture meter (Delta-T De-vices). The sensor was calibrated for thespecific substrate used in the study accordingto Kargas et al. (2013).

SPAD. SPAD provides arbitrary values thatindicate the green color of the leaves and itsvalues have been correlated to plant chloro-phyll content (Loh et al., 2002). SPAD mea-surements were performed every 3 d at around1330 HR with a handheld SPAD meter (SPAD-502 Chlorophyll meter; Konica Minolta, Ja-pan). For each measurement, the youngest offully developed leaves were selected.

Leaf stomatal resistance. Leaf stomatalresistance measurements were performedevery 3 d at noon, using an AP4 diffusionporometer (Delta-T Devices). Measurementswere made on the abaxial side of young, fullyexpanded leaves.

Meteorological data. The ambient maxi-mum, minimum, and average temperatureswere recorded by the weather station of theNational Observatory of Athens at the Gaziregion, which is located 885 m from theexperimental site. It should be emphasizedat this point that during the water-stressperiod, no rainfall occurred (Fig. 1).

Results and Discussion

Ballota acetabulosa. In the beginning ofthe study, the substrate moisture content in

Fig. 2. Photographic representation of the five plant species at the initiation (22 July 2013) and at the end of the study (10 Sept. 2013) after the imposition ofdifferent irrigation regimens [100%, 75%, 50%, 25%, and 0% of daily pan evaporation (Epan)].

Fig. 3. (A) Substrate moisture content (% v/v), (B) soil plant analysis development (SPAD) values, and (C)leaf stomatal resistance (s·cm–1) for Ballota acetabulosa under different irrigations regimens [100%,75%, 50%, 25%, and 0% of daily pan evaporation (Epan)]. Bars represent Fisher’s least significantdifference (LSD) at P < 0.05.

610 HORTSCIENCE VOL. 51(5) MAY 2016

plots planted with Greek horehound plantswas similar for 100%, 75%, and 50% of Epan

irrigation regimens. In contrast, irrigationregimens of 25% and 0% Epan exhibited lowersubstrate moisture content compared with theremaining treatments (Fig. 3A). However,differences between 25% and 50%, and 75%and 100% were negated after 11 Aug. 2013.

From macroscopic phenological observa-tions, it was noticed that in Greek horehoundplants, at 0% and 25% Epan irrigation regi-men, the leaves turned yellow and some ofthe leaves dropped (Fig. 2). The nonirrigatedplants (0% Epan) died after 20 d, indicatingthat the reduced leaf area of Greek horehoundin conjunction with its xerophytic morpho-logical and physiological characteristics pro-vided an excellent drought mechanism. Thetime interval of 20 d coincides with thatreported by Bousselot et al. (2011) for theobserved dying of herbaceous plants topgrowth when drought was imposed in green-house conditions.

SPAD results were similar for all treat-ments. At the initiation of the study, SPADvalues were low because Greek horehoundhas a grey leaf color, which is attributable tothe numerous hair found on both sides of theleaf (Fig. 3B).

Leaf stomatal resistance exhibited minimaldifferences between irrigation treatments,according to which the nonirrigated and the25% Epan treatment provided higher resis-tance values (Fig. 3C). The latter illustratesthat the plants of 25% and 0%Epan were morestressed compared with the other irrigationtreatments. Nevertheless, the leaf stomatalresistance values were relatively low for allirrigation treatments and, thus, Greek hore-hound stomata did not close at the deficientirrigation regimens. This observation, in con-junction with the increased substrate moisturecontent, demonstrated that Greek horehoundexhibited other substrate moisture–stress toler-ance mechanisms. It must be noted that Greekhorehound is a common herbaceous, xero-phytic species found in abundance in southernand eastern Greece (Heywood and Richardson,1990). The drought adaptation mechanisms ofxerophytes involve small and thick leaves, withhigh specific dry weight, small volume ofinternal air space, a high percentage of palisademesophyll tissue (Parkhurst and Loucks, 1972;Shields, 1950), and a layer of hairs.

Helichrysum orientale. Similar to Greekhorehound, the moisture content in plotsplanted with helichrysum substrate increasedin the irrigated compared with the nonirri-gated plants (Fig. 4A). However, in this case,the 25% Epan irrigation treatment providedsimilar moisture content compared with theother irrigation treatments during the first15 d of the water-stress period. It was noticedthat in the helichrysum plants irrigated at thelow-irrigation regimen of 25%Epan as well asthe nonirrigated treatment (0% Epan), theleaves reduced their surface area either byreducing their size and/or by becoming twirled(Fig. 2). Helichrysumhas been found to toleraterelatively mild water stress when grown undergreen roof conditions and irrigated to saturation

Fig. 4. (A) Substrate moisture content (% v/v), (B) soil plant analysis development (SPAD) values, and (C)leaf stomatal resistance (s·cm–1) forHelichrysum orientale under different irrigations regimens [100%,75%, 50%, 25%, and 0% of daily pan evaporation (Epan)]. Bars represent Fisher’s least significantdifference (LSD) at P < 0.05.

Fig. 5. (A) Substrate moisture content (% v/v), (B) soil plant analysis development (SPAD) values, and (C)leaf stomatal resistance (s·cm–1) for Melissa officinalis under different irrigations regimens [100%,75%, 50%, 25%, and 0% of daily pan evaporation (Epan)]. Bars represent Fisher’s least significantdifference (LSD) at P < 0.05.

HORTSCIENCE VOL. 51(5) MAY 2016 611

every 5 or 7 d (Papafotiou et al., 2013). As inthe case of Greek horehound, nonirrigatedplants survived for 20 d indicating a gooddrought tolerance plant response.

SPAD values during the water-stress pe-riod had minimal differences and, in thosecases, they were increased for the irrigationtreatments of 50%, 75%, and 100%Epan exceptfor a single date when 25% Epan providedincreased SPAD value (Fig. 4B). SPAD valuesfor all treatments were gradually reduced as thewater-stress study progressed, indicating a re-duction in chlorophyll concentration.

Differences between irrigation treatmentswith regard to leaf stomatal resistance oc-curred only at the initiation of the study whenthe nonirrigated plants reached 12.5 s·cm–1

(Fig. 4C). All other irrigation treatmentsprovided similar low values. From theseresults, it was concluded that, since it keepsits stomata open, helichrysum is capable oftranspiring even at minimum levels of sub-strate moisture. It is likely that anothermechanism, either metabolic or morpholog-ical, is mainly responsible for the increasedsubstrate moisture–stress tolerance exhibitedby helichrysum.

Melissa officinalis. In the case of lemonbalm, the substrate moisture content pre-sented more pronounced differences betweendifferent irrigation regimens in comparisonwith the other plants (Fig. 5A). The high-irrigation regimens (100% and 75% Epan)provided increased substrate moisture con-tent, while irrigation at 50% and 25% Epan

provided lower moisture content. Moreover,the substrates of the nonirrigated plots losttheir moisture content extremely fast, as in-dicated by the abrupt decline of moisturecontent to �1%, which resulted in the loss ofall nonirrigated plants within 7 d after theinitiation of the water-stress treatments. Allthe above indicate that lemon balm hasincreased transpiration and, thus, it is ex-pected to be a less substrate moisture–stresstolerant species.

Differences in SPAD values betweenirrigation treatments were minimal until 26Aug. 2013, when the 25% Epan treatmentreduced its SPAD values compared with the50%, 75%, and 100% Epan treatments, in-dicating a severe reduction of plant chloro-phyll content. Munn�e-Bosch and Alegre(2000) have also reported a decrease inchlorophyll content in lemon balm plantsgrowing under drought conditions. Our find-ings contradict those of Ozturk et al. (2004)who reported that lemon balm plants provedto be drought tolerant when growing inlysimeters (of 62-cm diameter and 22-cmdepth) filled with a sandy loam soil. In ourstudy, substrate depth was only 8 cm in aneffort to simulate a shallow adaptive greenroof system. In the latter cases, plants thathave a drought-avoidance mechanism, suchas the development of deeper root systems,are handicapped on account of the limitedsubstrate depth. In addition, lemon balm hasleaves that are thin and not covered withhairs. It was noticed macroscopically that,lemon balm leaves turn yellow and brown

very fast even from the initial stages of water-stress imposition.

The nonirrigated plants increased theirleaf stomatal resistance and died out whenit reached 4.8 s·cm–1 (Fig. 5C). The 25% Epan

treatment plants also died out 19 d after theinitiation of the water-stress treatments,reaching a leaf stomatal resistance valueof 17.5 s·cm–1. Munn�e-Bosch and Alegre(1999) reported that lemon balm plantssubjected to water-stress conditions de-creased their stomatal conductance by 50%compared with well-watered plants. In ourcase, we also observed a similar increasein stomatal resistance (Fig. 5C). However,this increase was brief since the plants diedsoon after they reached a threshold value of�20 s·cm–1.

Our results are in contrast to those re-ported by Ozturk et al. (2004), who found thatM. officinalis plants were resistant to mildwater-stress conditions. More specifically,they found that the dry weight decrease underdeficit irrigation conditions (irrigation at0%, 12.5%, 25%, 37.5%, and 50% of normalirrigation based on lysimeter moisture-holding capacity) was not significant up tothe level of 25% of deficit irrigation. Fromour study, it was determined that an irrigationof at least 50% of daily Epan is required for

sustainable growth of lemon balm on shallowgreen roof systems.

Munn�e-Bosch and Alegre (1999) founda large decrease of xylematic water potentialbetween irrigated and nonirrigated lemonbalm plants. However, water potential ofthe water-stressed plants remained relativelyconstant for 2 weeks due to the significantreduction of leaf stomatal aperture. In ourcase, the nonirrigated plants were not able tosurvive for 2 weeks which was expected dueto the shallow depth of the green roof sub-strate layer.

Rosmarinus officinalis. The slope of themoisture curve reduction of the nonirrigatedplants was similar to that of helichrysum andGreek horehound and less steep than that oflemon balm and Greek sage, indicating thatrosemary did not consume water very rapidlyunder drought conditions (Fig. 6A). At theinitiation and toward the end of the water-stress period, the 25% Epan irrigation regimenhad a lower substrate moisture content com-pared with all higher irrigation regimens. The100%, 75%, and 50% Epan irrigation regi-mens provided similar substrate moisturecontent throughout the study.

SPAD values were reduced for a singledate in the plants that were nonirrigated(Fig. 6B). In general, SPAD values remained

Fig. 6. (A) Substrate moisture content (% v/v), (B) soil plant analysis development (SPAD) values, and (C)leaf stomatal resistance (s·cm–1) for Rosmarinus officinalis under different irrigations regimens [100%,75%, 50%, 25%, and 0% of daily pan evaporation (Epan)]. Bars represent Fisher’s least significantdifference (LSD) at P < 0.05.

612 HORTSCIENCE VOL. 51(5) MAY 2016

relatively stable for all treatments during thewater-stress period. However, no significantdifferences between the four irrigation regi-mens (25%, 50%, 75%, and 100% Epan) wereobserved throughout the study (Figs. 2 and6B). Since SPAD is an indicator of thechlorophyll content of the leaves, our resultscomply with the findings of Munn�e-Boschet al. (1999). They reported that, drought-stressed rosemary plants grown in the fieldreduced their diurnal CO2 assimilation by80% even though the maximum efficiency ofphotosystem II and chlorophyll content wereunaffected.

The nonirrigated plants of rosemary in-creased their leaf stomatal resistance imme-diately after the imposition of drought stress(Fig. 6C). However, after 29 July 2013 it wasnot possible to determine leaf stomatal re-sistance at the nonirrigated plants althoughthe plants did not die out. It is possible thatthis resulted from the increased stomatalresistance values that exceeded the pore-meter detection capacity. For the remainingtime of the water-stress period, the 25%Epan irrigation regimen provided a trend ofhigher leaf stomatal resistance values, sug-gesting increased water stress for the periodof this treatment. In addition, it indicatedthat stomata closure is one of the mainsubstrate moisture–stress tolerance mecha-nisms of rosemary for withstanding drydown periods.

Several researchers have reported theability of rosemary to exhibit osmotic adjust-ment when grown under water-stress condi-tions. Sanchez-Blanco et al. (2004) reportedthat changes in rosemary plants due to deficitirrigation can be considered as a morpholog-ical adaptation of the species to water stress,thereby reducing the rate of transpiration andminimizing water consumption. In addition,Nogu�es and Baker (2000) and Nogu�es et al.(2001) observed that during severe summerdrought, rosemary significantly decreased itsphotosynthetic rate, probably by means ofstrong stomata closure. Apart from the mor-phological adaptations, Munn�e-Bosch et al.(1999) showed that rosemary is able to suc-cessfully withstand prolonged drought by in-creasing the concentration of a-tocopherol,carnosic acid, and carotenoid concentrationin the leaves. The production of these antiox-idants prevents the oxidative damage in theplants.

Salvia fruticosa. The substrate moisture inplots with nonirrigated plants was reducedabruptly, as illustrated by the curve slope(Fig. 7A), indicating fast water consumptionby the plants. From the initiation of the water-stress period until 14 Aug. 2013, there wasa clear separation between the differentirrigation regimens into two groups. Morespecifically, the 25% and 50% Epan regimensexhibited lower moisture content, while the75% and 100% Epan regimens exhibitedhigher moisture content. From then on anduntil the end of the water-stress period, 100%Epan had increased substrate moisture contentcompared with the other irrigation regimentreatments.

The nonirrigated plants of Greek sagedied out 10 (±2) d after the imposition ofdrought stress. Bettaieb et al. (2009) re-ported that 2 weeks after the imposition ofwater stress, S. officinalis plants exhibiteda negative effect on their morphologicalcharacteristics which became more pro-nounced as water stress progressed. Theyobserved that severely water-deprived plantsproduced thinner stems with fewer and moredehydrated leaves compared with otherplants that were irrigated at higher irriga-tion regimens. In our case, it was observedthat Greek sage plants under water-stressconditions reduced the size of the leaves bytwirling. In addition, the whole leaf area wasreduced by turning the leaves yellow andshedding them (Fig. 2).

The SPAD values of Greek sage increasedas water stress progressed (Fig. 7B). Thiscould have been caused by an increase in cellsolute concentration. Differences in SPADvalues were minimal and, in those cases,plants irrigated at 25% and 50% Epan hadhigher SPAD values compared with the 75%and 100% Epan irrigation regimens.

At the initiation of the water-stress period,the nonirrigated plants had increased leafstomatal resistance (Fig. 7C). After they diedout, the 25% and 50% Epan irrigation regimens

had increased leaf stomatal resistance, whichillustrates their higher stress compared withthe 75% and 100% irrigation regimens.Raimondo et al. (2015) installed S. offici-nalis plants in experimental modules ona green roof and remarked that plants sub-jected to water stress reduced their leaf waterpotential close to the point of turgor lossthrough osmotic adjustment to reduce waterconsumption and to survive throughout thedry down period.

Conclusions

A specific irrigation threshold was de-termined to secure sustainable growth ofeach plant species under adaptive greenroof conditions. More specifically, sustain-able growth of Greek horehound, helichry-sum, and rosemary can be achieved whenthey are irrigated at 25% Epan. By contrast,lemon balm and Greek sage demand anirrigation regimen of at least 50% Epan tosecure sustainable growth on shallow-depthgreen roof systems. It was also establishedthat extreme caution is necessary in rela-tion to the irrigation frequency of the lattertwo species, given that their plants diedwithin 7–10 d of withholding irrigationtreatments.

Fig. 7. (A) Substrate moisture content (% v/v), (B) soil plant analysis development (SPAD) values, and (C)leaf stomatal resistance (s·cm–1) for Salvia fruticosa under different irrigations regimens [100%, 75%,50%, 25%, and 0% of daily pan evaporation (Epan)]. Bars represent Fisher’s least significant difference(LSD) at P < 0.05.

HORTSCIENCE VOL. 51(5) MAY 2016 613

Literature Cited

Akbari, H., M. Pomerrantz, and H. Taha. 2001.Cool surfaces and shade trees to reduce energyuse and improve air quality in urban areas. Sol.Energy 70:295–310.

Benvenuti, S. and D. Bacci. 2010. Initial agro-nomic performances of Mediterranean xero-phytes in simulated dry green roofs. UrbanEcosyst. 13:349–363.

Bettaieb, I., N. Zakhama, W. Aidi Wannes, M.E.Kchouk, and B. Marzouk. 2009. Water deficiteffects on Salvia officinalis fatty acids and essen-tial oils composition. Sci. Hort. 120:271–275.

Bousselot, J.M., J.E. Klett, and R.D. Koski. 2011.Moisture content of extensive green roof sub-strate and growth response of 15 temperateplant species during dry down. HortScience46:518–522.

Czemiel Berndtsson, J. 2010. Green roof perfor-mance towards management of runoff waterquantity and quality: A review. Ecol. Eng.36:351–360.

Dimopoulos, P., T. Raus, E. Bergmeier, T.Constantinidis, G. Iatrou, S. Kokkini, A. Strid,and D. Tzanoudakis. 2013. Vascular plants ofGreece: An annotated checklist. BotanischerGarten und Botanisches Museum (eds.). Berlin-Dahlem (Berlin) and Hellenic Botanical Society(Athens).

Ferguson, B.K. 1998. Introduction to stormwater:Concept, purpose, design. Wiley, New York.

Fioretti, R., A. Palla, L.G. Lanza, and P. Principi.2010. Green roof energy and water related per-formance in the Mediterranean climate. Build.Environ. 45:1890–1904.

Forschungsgesellschaft LandschaftsentwicklungLandschaftsbau (FLL). 2008. Guidelines for theplanning, construction and maintenance of greenroofing: Green roofing guideline. Bonn, Germany.

Getter, K.L., D.B. Rowe, G.P. Robertson, B.M.Cregg, and J.A. Andresen. 2009. Carbon se-questration potential of extensive green roofs.Environ. Sci. Technol. 43:7564–7570.

Getter, L.K. and D.B. Rowe. 2006. The role ofextensive green roofs in sustainable develop-ment. HortScience 41:1276–1285.

Heywood, V.H. and I.B.K. Richardson. 1990.Ballota acetabulosa Benth. In: T.G. Tunin,V.H. Heywood, N.A. Burges, D.M. Moore,D.H. Valentine, S.M. Walters, and D.A. Webb(eds.). Flora europaea. Vol. 3. CambridgeUniversity Press, Cambridge, UK.

Kargas, G., N. Ntoulas, and P.A. Nektarios. 2013.Soil texture and salinity effects on calibrationof TDR300 dielectric moisture sensor. SoilRes. 51:330–340.

Kotsiris, G., A. Androutsopoulos, E. Polychroni,and P.A. Nektarios. 2012a. Dynamic U-valueestimation and energy simulation for greenroofs. Energy Build. 45:240–249.

Kotsiris, G., P.A.Nektarios, andA.T. Paraskevopoulou.2012b. Lavandula angustifolia growth and phys-iology is affected by substrate type and depth

when grown under Mediterranean semi-intensivegreen roof conditions. HortScience 47:311–317.

Kotsiris, G., P.A. Nektarios, N. Ntoulas, and G.Kargas. 2013. An adaptive approach to inten-sive roofs in theMediterranean climatic region.Urban For. Urban Green. 12:380–392.

Kumar, R. and S.C. Kaushik. 2005. Performanceevaluation of green roof and shading for ther-mal protection of buildings. Build. Environ.40:1505–1511.

Loh, F.C.W., J.C. Grabosky, and N.L. Bassuk.2002. Using SPAD 502 meter to assess chlo-rophyll and nitrogen content of benjamin figand cottonwood leaves. HortTechnology 12:682–686.

Munn�e-Bosch, S. and L. Alegre. 1999. Role ofdew on the recovery of water-stressed Melissaofficinalis L. plants. J. Plant Physiol. 154:759–766.

Munn�e-Bosch, S. and L. Alegre. 2000. Changes incarotenoids, tocopherols and diterpenes duringdrought and recovery, and the biological sig-nificance of chlorophyll loss in Rosmarinusofficinalis plants. Planta 210:925–931.

Munn�e-Bosch, S., K. Schwarz, and L. Alegre.1999. Enhanced formation of a-tocopherol andhighly oxidized abietane diterpenes in water-stressed rosemary plants. Plant Physiol. 121:1047–1052.

Nektarios, P.A., I. Amountzias, I. Kokkinou, andN. Ntoulas. 2011. Green roof substrate type anddepth affect the growth of the native speciesDianthus fruticosus under reduced irrigationregimens. HortScience 46:1208–1216.

Nektarios, P.A., N. Ntoulas, E. Nydrioti, I. Kokkinou,E.-M. Bali, and I. Amountzias. 2015. Droughtstress response of Sedum sediforme grown inextensive green roof systems with differentsubstrate types and depths. Sci. Hort. 181:52–61.

Nogu�es, S. and N.R. Baker. 2000. Effects of droughton photosynthesis in Mediterranean plants grownunder enhanced UV-B radiation. J. Expt. Bot.51:1309–1317.

Nogu�es, S., S. Munn�e-Bosch, J. Casades�us, M.L�opez-Carbonell, and L. Alegre. 2001. Dailytime course of whole-shoot gas exchange ratesin two drought exposed Mediterranean shrubs.Tree Physiol. 21:51–58.

Ntoulas, N. and P.A. Nektarios. 2015. Paspalumvaginatum drought tolerance and recovery inadaptive extensive green roof systems. Ecol.Eng. 82:189–200.

Ntoulas, N., P.A. Nektarios, E. Charalambous, andA. Psaroulis. 2013a. Zoysia matrella cover rateand drought tolerance in adaptive extensivegreen roof system. Urban For. Urban Green.12:522–531.

Ntoulas, N., P.A. Nektarios, and E. Nydrioti.2013b. Performance of Zoysia matrella inshallow green roof substrates under moisturedeficit conditions. HortScience 48:929–937.

Ntoulas, N., P.A. Nektarios, K. Spaneas, and N.Kadoglou. 2012. Semi-extensive green roof

substrate type and depth effects on Zoysiamatrella ‘Zeon’ growth and drought toleranceunder different irrigation regimens. Acta Agr.Scand. Sect. B Soil Plant Sci. 62:165–173.

Oberndorfer, E., J. Lundholm, B. Bass, R.R.Coffman, H. Doshi, N. Dunnett, S. Gaffin, M.K€ohler, K.K.Y. Liu, and B. Rowe. 2007. Greenroofs as urban ecosystems: Ecological structures,functions, and services. Bioscience 57:823–833.

Ozturk, A., A. Unlukara, A. Ipek, and B. Gurbuz.2004. Effects of salt stress and water deficit onplant growth and essential oil content of lemonbalm (Melissa officinalis L.). Pak. J. Bot. 36:787–792.

Papafotiou, M., N. Pergialioti, L. Tassoula, I.Massas, and G. Kargas. 2013. Growth of nativearomatic xerophytes in an extensive Mediter-ranean green roof as affected by substrate typeand depth and irrigation frequency. Hort-Science 48:1327–1333.

Paraskevopoulou, A., I. Mitsios, I. Fragakis, P.A.Nektarios, N. Ntoulas, P. Londra, and M.Papafotiou. 2015. The growth of Arthrocne-mum macrostachyum and Halimione portula-coides in an extensive green roof system undertwo watering regimes. Agr. Agricultural Sci.Procedia 4:242–249.

Parkhurst, D.F. and O.L. Loucks. 1972. Optimalleaf size in relation to environment. J. Ecol.60:505–537.

Raimondo, F., P. Trifilo,M.A.L. Gullo, S. Andri, T.Savi, and A. Nardini. 2015. Plant performanceon Mediterranean green roofs: Interaction ofspecies-specific hydraulic strategies and sub-strate water relations. AoB Plants 7:Plv007.

Rowe, D.B. 2011. Green roofs as ameans of pollutionabatement. Environ. Pollut. 159:2100–2110.

Sanchez-Blanco, M.J., F. Ferrandez, M.A. Morales,A. Morte, and J. Alarcon. 2004. Variations inwater status, gas exchange, and growth in Ros-marinus officinalis plants infected with Glomusdeserticola under drought conditions. J. PlantPhysiol. 161:675–682.

Shields, L.M. 1950. Leaf xeromorphy as related tophysiological and structural influences. Bot.Rev. 16:399–447.

Simmons, M.T., B. Gardiner, S. Windhager, and J.Tinsley. 2008. Green roofs are not createdequal: The hydrologic and thermal perfor-mance of six different extensive green roofsand reflective and non-reflective roofs in a sub-tropical climate. Urban Ecosyst. 11:339–348.

Tassoula, L., M. Papafotiou, G. Liakopoulos, andG. Kargas. 2015. Growth of the native xero-phyte Convolvulus cneorum L. on an extensiveMediterranean green roof under different sub-strate types and irrigation regimens. HortScience50:1118–1124.

VanWoert, N.D., D.B. Rowe, J.A. Andresen, C.L.Rugh, and L. Xiao. 2005. Watering regime andgreen roof substrate design affect Sedum plantgrowth. HortScience 40:659–664.

614 HORTSCIENCE VOL. 51(5) MAY 2016