At

JLa

b

J

a

ARRA

KBLECBT

1

e(siwMpptaY

F

j

h0

Agricultural Water Management 137 (2014) 15–22

Contents lists available at ScienceDirect

Agricultural Water Management

jou rn al hom ep age: www.elsev ier .com/ locate /agwat

ctual evapotranspiration and crop coefficients for five species ofhree-year-old bamboo plants under a tropical climate

ulien Piouceaua,b,∗, Frédéric Panfili a, Grégory Boisa, Matthieu Anastasea,aurent Dufosséb, Véronique Arfia

Phytorem S.A., site d’Areva, chemin de l’autodrome, F-13140 Miramas, FranceUniversité de La Réunion, Laboratoire de Chimie des Substances Naturelles et des Sciences des Aliments, site ESIROI Département Agroalimentaire, 2 rue

oseph Wetzell, F-97490 Sainte-Clotilde, France

r t i c l e i n f o

rticle history:eceived 29 August 2013eceived in revised form 28 January 2014ccepted 4 February 2014

eywords:ambooysimetervapotranspirationrop coefficientiomass

a b s t r a c t

Over the last decade, bamboo plantations have started to be used as vegetation filters for wastewatertreatment. This treatment system can be useful in reducing wastewater discharge into the environment,thus contributing to the preservation of water resources by using the plantation’s evapotranspiration toreduce the rate of water infiltration. The evapotranspiration rates of the bamboo species used is there-fore an important factor. The actual evapotranspiration (ET) and the crop coefficients (kc) for the fivetropical and temperate species of three-year-old bamboo plants, i.e. Bambusa oldhamii, Bambusa mul-tiplex, Bambusa vulgaris, Phyllostachys aurea and Pseudosasa japonica, were studied in lysimeters for aperiod of more than one year under a tropical climate. The average ET rates for the bamboo speciesstudied ranged from 4 to 7 mm day−1 with maximum values of between 10.7 and 17.1 mm day−1 duringthe wet season, and an average kc of 1.1 to 1.9. The ET was correlated to weather parameters, especially

ropical climate minimum temperatures. The differences in ET rates between the bamboo species can be explained bymorphological parameters, in particular the total aboveground biomass. Among the five bamboo speciesstudied, B. oldhamii had the highest ET rate and produced the most biomass. In comparison with otherhigh-biomass-producing plants, the evaporation rates for young bamboo plants were similar to those forwillow and poplar vegetation filters.

© 2014 Elsevier B.V. All rights reserved.

. Introduction

Water resources conservation is an important issue today. Theutrophication process in water especially is a worldwide concernSmith, 2009; Smith et al., 1999). Natural wastewater treatmentystems using phytoremediation principles have been developedn order to provide a useful alternative to more conventional waste-

ater treatment systems, particularly in rural or isolated areas.ost of these systems involve constructed wetland using aquatic

lants (Vymazal, 2011). Another type of system uses terrestriallants such as poplar, willow and – more recently – bamboo, where

he plants are used as vegetation filters (Arfi et al., 2009; Aronssonnd Perttu, 2001; Perttu and Kowalik, 1997; Singh et al., 2008;adav et al., 2010).

∗ Corresponding author at: Phytorem S.A., site d’Areva, chemin de l’autodrome,-13140 Miramas, France. Tel.: +33486260805; fax: +33490581578.

E-mail addresses: julienpiouceau@phytorem.com,.piouceau@hotmail.fr (J. Piouceau).

ttp://dx.doi.org/10.1016/j.agwat.2014.02.004378-3774/© 2014 Elsevier B.V. All rights reserved.

Although bamboo is still rarely used for phytoremediation pur-poses (water or soil treatments), it nevertheless shares a numberof interesting characteristics with other plants used in phytoreme-diation. Among these is bamboo’s apparent tolerance to pollutantssuch as trace metals (Collin et al., 2013; Gui et al., 2011; Shuklaet al., 2011), and to the excessive doses of nutrients that can be con-tained in wastewater (Piouceau et al., 2014). Bamboo also developsa dense root system with a profusion of fine roots and hair roots(Christanty et al., 1996) that favors the hosting of microorganismsand improves the rhizodegradation of organic matter contained inwastewater (Licht and Isebrands, 2005; McCutcheon and Schnoor,2003). Lastly, giant bamboo species are among the most productiveterrestrial plants in the world. The aboveground biomass yield ofmature plantations can reach 25 and 47 t ha−1 yr−1 under temper-ate and tropical climates, respectively (Scurlock et al., 2000). Thesehigh-yield plants are interesting for wastewater treatment pur-

poses because they can help preserve the quality of water resourcesby limiting the eutrophication phenomenon in two complementaryways. First, by storing large amounts of nitrogen and phosphorous(the main elements involved in eutrophication) in their biomass: a

1 ater Management 137 (2014) 15–22

ma(oapc

hctiatIetwtztatas

bKdttt(ttapalkleo(temtlmad

2

2

2oe4

m(

els) with three replicates per species. The bamboo species wereplanted in the lysimeters’ reconstituted soil in May 2008, fourmonths before the start of the experiment, to allow for the bam-boo’s proper rooting and a natural compaction of the soil.

Table 1Average lysimeter soil properties after bamboo plantation.

Particle density (g cm−3) 2.8Bulk density (g cm−3) 0.6Soil water holding capacity (mm) 210pHwater 6.3EC (mS cm−1) 0.3Total nitrogen (g kg−1) 2.7

6 J. Piouceau et al. / Agricultural W

ature bamboo plantation can store 131 to 619 kg ha−1 of nitrogennd 17 to 97 kg ha−1 of phosphorous in its aboveground biomassKleinhenz and Midmore, 2001). Second, by slowing the infiltrationf water through their high evapotranspiration rates; Kleinhenznd Midmore (2002) estimated that the ET rate of a mature bamboolantation can range between 9 and 13 mm day−1 under a tropicallimate.

The development of vegetation filters using terrestrial plantsas been boosted by regulatory pressure to reduce or stop the dis-harge of any treated water into the environment. Indeed, even ifhe treated water complies with environmental quality standards,t still contains small amounts of nitrogen and phosphorus thatre potentially harmful for the water resources. This is particularlyrue for sensitive areas such as the lagoons and streams on Reunionsland (Chazottes et al., 2002; Naim, 1993; Semple, 1997; Tedettit al., 2011), for which a target of zero discharge of any water intohe environment needs to be implemented. In vegetation filters, theastewater is directly spread over the plantation’s soil surface. The

reated wastewater is released through percolation, via the rootone to the water table and into the atmosphere through evapo-ranspiration. This kind of system can be designed and managed tovoid any discharge of any water into the water table. To do this,he evapotranspiration (ET) rates of the plant species used must beccurately known to set the appropriate wastewater volume to bepread on the plantation.

Some studies were carried out on the transpiration of variousamboo species using the sap-flow method (Dierick et al., 2010;omatsu et al., 2010, 2012; Kume et al., 2010), but most of theseid not take into account the evaporation via the soil. Furthermore,hey were conducted in non-irrigated conditions so the experimen-al method did not allow the plants’ full transpiration potentialo be quantified. Another experimental method uses lysimetersAboukhaled et al., 1982). The lysimeter method is well-adaptedo measuring the ET, that is to say both the plant transpiration andhe soil evaporation are measured in field conditions. This methodlso allows the crop coefficient (kc) to be calculated. The kc of alant species is the ratio between the ET rate of the plant studiednd the reference evapotranspiration (ET0) (Allen et al., 1998). Theysimeter method has been successfully used to study the ET andc of some high-biomass-producing plants such as poplar and wil-ow (Guidi et al., 2008; Pauliukonis and Schneider, 2001; Pistocchit al., 2009). To our knowledge, there is currently a lack of datan the evapotranspiration rates of bamboo and other giant grassese.g. switchgrass, Miscanthus sp. or Arundo donax). Consequently,he aim of our study was to determine the ET rates and kc of sev-ral bamboo species and to compare them in order to select theost appropriate species for vegetation filters (i.e. the species with

he highest ET rate). To this end, we put in place percolation-typeysimeters (Allen et al., 2011) to measure ET rates over a period of

ore than one year. Morphological parameters, such as leaf areand total aboveground biomass, were also measured to account forifferences in ET rates between species.

. Materials and methods

.1. Experimental conditions

The experiment was conducted over one year, from August0th, 2008 to October 30th, 2009 (436 days), on Reunion Island, anverseas French department in the south-west Indian Ocean. Thexperimental site was located at Le Guillaume, Saint Paul (21◦03′

8′′S; 55◦19′ 24′′E) at an elevation of 1043 m.Fifteen percolation-type lysimeters were set up at the experi-

ental site (Fig. 1). The lysimeters consisted of 1.2 m3 plastic tanks1.00 m deep, 1.00 × 1.20 m in area). The lysimeters were placed

Fig. 1. Lysimeter set up at the experiment site.

70 cm apart to protect their sides from direct sunlight. The sides ofthe lysimeters were also insulated with a polypropylene membraneto prevent heating.

The lysimeters were filled with 20 cm of gravel at the bottom as adrainage layer. A geotextile (bidim green®, Ten Cate Geosynthetics,France) was laid on top of the gravel before the tanks were filledwith a 3:1:1 (v:v:v) mixture of soil, scoria and sugar cane fibers, thecharacteristics of which are listed in Table 1.

The lysimeters were planted with five different species ofthree-year-old bamboo plants: three tropical species (pachymorphrhizome; Stapleton (1998): Bambusa oldhamii Munro (BO), Bam-busa multiplex (Lour.) Raeusch (BGG) and Bambusa vulgaris Schrad.(BVV)) and two temperate species (leptomorph rhizome; Stapleton(1998): Phyllostachys aurea Rivière & C. Rivière (PA) and Pseudosasajaponica (Steud.) Makino (PJ)). The lysimeters were arranged in acompletely randomized design based on the factor species (five lev-

Total organic carbon (g.100 g dry soil) 3.1Phosphorus Olsen-Dabin (g kg−1) 0.3CEC (meq.100 g) 14.1

Electrical conductivity (EC); Cation exchange capacity (CEC).

ater Management 137 (2014) 15–22 17

(4iioNedutTitadtftlpotitac

p

2

tco

cfo7MPoss

2c

t(ce

E

wtdb(d(

J. Piouceau et al. / Agricultural W

Each lysimeter was irrigated with an automatic drip systemNetafimTM, Israel) using four drippers. Two spikes delivering

× l h−1 per spike were inserted into the soil for each dripper, giv-ng a total flow of 32 × l h−1 for each lysimeter. The lysimeters wererrigated once a week with tap water, with a twice-weekly additionf a fertilizer solution comprising a commercial fertilizer (20:20:20PK soluplant duclos®, Duclos International, France) and a tracelement solution (Oligo Drip 25 Fe, Duclos international, France)iluted with tap water. The fertilizer solution was diluted in-linesing a metering pump (DI 16, Dosatron International, France) withhe resulting nutrient solution going to the drip irrigation system.he dilution was 1% in the dry season (May to October) and 1.6%n the wet season (November to April). The solution concentra-ion was 300 mg l−1 of nitrogen (N), 300 mg l−1 of phosphorus (P)nd 300 mg l−1 of potassium (K) (electric conductivity 1.3 mS cm−1)uring the dry season and 480 mg l−1 of N, P and K (1.8 mS cm−1) inhe wet season. The application of nutrients was designed to makeertilization a non-limiting factor in the bamboo’s growth; duringhe experiment, a total of 672.1 g of N, P and K were applied to eachysimeter, representing a total of 5.6 t ha−1 yr−1 of nitrogen, phos-horus and potassium. The average daily height of water (tap waterr fertilizer solution) added to the lysimeters was 8.9 mm day−1 inhe dry season, and 22.2 mm day−1 in the wet season. The volume ofrrigation water supplied to the lysimeter was adjusted accordingo the soil’s water-holding capacity, 210 mm. This supply of waterllowed the soil’s moisture content to be maintained close to fieldapacity throughout the experiment.

A faucet at the base of each lysimeter was connected to 20-llastic can to collect the drainage water.

.2. Climatic parameters measurements

The air temperature and relative humidity were measured onhe experimental site every hour using data loggers (EL-USB-2, Las-ars electronics, UK) (Fig. 2a and b). Three data loggers were set upn three lysimeters chosen at random.

The rainfall height was measured by the site’s meteorologi-al station and reference evapotranspiration (ET0) was obtainedrom the Météo-France (French meteorological agency) mete-rological station at Bois de Nèfles on Reunion Island, at.656 km from our experimental site. The ET0 was determined byétéo-France (French meteorological agency) using the FAO-56

enman–Monteith equation (Allen et al., 1998), with daily mete-rological data (including air temperature, relative humidity, windpeed and solar radiation) obtained using an automatic weathertation.

.3. Determination of actual evapotranspiration and cropoefficients

Ten-day mean ET for each bamboo species was calculated usinghe simplified water balance equation (Rana and Katerji, 2000) (Eq.1) below), assuming that the soil’s water content was kept at fieldapacity throughout the experiment (Guidi et al., 2008; Pistocchit al., 2009):

T = I + R − D (1)

here I is the amount of water provided by irrigation (mm), Rhe rainfall height (mm) and D the drainage height (mm). Therainage volumes of each lysimeter were measured every two days

y weighing the 20-l plastic cans with a 20 g-precision spring scaleHS-50, Voltcraft®, Germany). The drainage height was obtained byividing the drainage volumes by the surface area of the lysimetersi.e. 1.2 m2).

Fig. 2. Ten-day temperatures (T) and rainfall (a) and relative humidity (RH) (b) dur-ing the experiment from August 2008 to October 2009; DS1: first dry season; WS:wet season; DS2: second dry season.

The mean 10-day crop coefficients (kc) were calculated usingthe following Eq. (2) (Allen et al., 1998):

kc = ETET0

(2)

where ET is the actual evapotranspiration (mm), and ET0 the refer-ence evapotranspiration (mm).

2.4. Growth measurements and plant sampling

At the beginning and end of the experiment, the number ofculms was counted in each lysimeter, their height measured witha tape measure and their basal diameters measured with a digitalcaliper at the middle of the first internode. The initial abovegroundfresh biomass for each bamboo species was calculated from thebasal diameters of culms using allometric equations established byPiouceau et al. (2014). At the end of the experiment, all the culmsin each lysimeter were cut. The total aboveground fresh biomass,total fresh leaf mass and fresh culm mass was weighed with a0.1 g-precision scale (Kern & Sohn GmbH, Germany).

2.5. Leaf area measurements

Three culms were randomly sampled per lysimeter to takesub-samples of leaves. These samples were immediately scanned(Mustek Scanexpress, Mustek Systems Inc., Taiwan) and weighed.The leaf area was calculated using scan images processed using

1 ater Management 137 (2014) 15–22

Actal

2

tirsc

tecWthb

ioprh(ba

3

3

sbOfid

DacsfraSbhi

Ftrn

3

te

Aug-08

Sep-08

Oct-08

Nov-08

Dec-08

Jan-0

9

Feb-09

Mar-09

Apr-09

May-09

Jun-0

9

Jul-0

9

Aug-09

Sep-09

Oct-09

Nov-09

ET

(m

m.d

ay-1

)

0

2

4

6

8

10

12

14

16

18BGG BO BVV PA PJ ET0

Fig. 3. Ten-day average evapotranspiration (ET) for the species Bambusa multiplex

8 J. Piouceau et al. / Agricultural W

dobe Illustrator CS4 software (Adobe Systems Inc., USA). The spe-ific leaf area (SLA), i.e. leaf area per mass unit, was calculated andhe total leaf area per clump was deduced by multiplying the aver-ge specific leaf area of the sub-sample by the total fresh mass ofeaf per clump.

.6. Statistics

We used an univariate repeated ANOVA (SPSS inc., IBM, USA)o study the difference in 10-day ET or kc among species. Spheric-ty was checked with Mauchley’s test; when this assumption wasejected the Greenhouse-Geisser corrections were used for the F-tatistics. The LSD test was used for post-hoc analyses; p < 0.05 washosen to identify significant differences between species.

A one-way ANOVA (SPSS inc., IBM, USA) was performed to testhe effect of the species factor on the number of culms, culm diam-ter and height, total aboveground fresh biomass, average freshulm mass, total fresh leaf mass, total leaf area, and total annual ET.

e performed square root or logarithmic transformations of datao satisfy the ANOVA assumptions. The LSD test was used for post-oc analyses; p < 0.05 was chosen to identify significant differencesetween species.

We conducted a two-tailed Pearson correlation analysis (SPSSnc., IBM, USA) to determine the relationships between ET ratesf each species, ET0 and the weather variables: maximum tem-erature, average temperature, minimum temperature, maximumelative humidity, average relative humidity, minimum relativeumidity and ET0. Another two-tailed Pearson correlation analysisSPSS Inc., IBM, USA) was performed to determine the relationshipsetween the total annual ET and morphological parameters: totalboveground fresh biomass, total fresh leaf mass and total leaf area.

. Results

.1. Meteorological conditions

The temperature, relative humidity and 10-day rainfall arehown in Fig. 2a and b. The climate is tropical with a wet seasonetween November and April, and a dry season between May andctober. During the experiment, the temperatures on site ranged

rom 6 to 38.8 ◦C with an average of 18.6 ◦C, and a relative humid-ty from 33.7% to 99.3%, with an average of 81%. The total rainfalluring the experiment was 1373.9 mm.

Our experiment took place over three seasons: two dry (DS1 andS2) from August to October 2008 and from May to October 2009,nd one wet (WS), from November 2008 to April 2009, the latterorresponding to the main growth period for bamboo. The two dryeasons were characterized by average temperatures that rangedrom 13.5 to 21.5 ◦C, a relative humidity of 61.9% to 90.4%, and lowainfalls (only 63 mm in DS2, the wetter of the two dry seasons). Theverage temperature during DS2 was lower than in DS1 by 2.8 ◦C ineptember and 1.6 ◦C in October. The wet season was characterizedy average temperatures that ranged from 19.2 to 22.3 ◦C, a relativeumidity of 75.8% to 90.4% and abundant rainfall (up to 426.8 mm

n one 10-day period).The abundant rainfall recorded in the first 10-day period of

ebruary 2009 were due to Hurricane Gaël that caused the threeemperature data loggers to break down. The data loggers wereeplaced on March 10th 2009. Consequently, temperatures wereot recorded during this period.

.2. Actual evapotranspiration and crop coefficients

The evapotranspiration patterns are shown in Fig. 3. During DS1,he different bamboo species showed similar ET rates until thend of October 2008. During this period the ET ranged from 0.98

(BGG), Bambusa oldhamii (BO), Bambusa vulgaris (BVV), Phyllostachys aurea (PA) andPseudosasa japonica (PJ); standard error: ±0.2.

to 1.97 mm day−1 for BGG, 1.72 to 2.39 mm day−1 for PA, 1.52 to3.09 mm day−1 for PJ, 1.33 to 3.18 mm day−1 for BVV, and 1.42 to3.65 mm day−1 for the BO species.

At the beginning of the WS, i.e. in November 2008, the BGG andBO species started to show significantly different patterns of ETrates compared to the other species. The BGG species showed thelowest ET rate and BO the highest. In December 2008, the ET ratesincreased rapidly and reached their maximum values in January2009. During the WS, the ET rates showed four peaks for all species.The same differences between species were observed as in DS1:the BGG species showed the lowest ET rate and BO the highest. TheBVV and PA species showed similar ET rates, and the PJ speciesshowed a lower ET rate than the BVV and PA species. BetweenNovember 2008 and March 2009, the ET rates ranged from 2.20to 12.34 mm day−1 for BGG, 2.62 to 12.18 mm day−1 for BVV, 2.69to 12.93 mm day−1 for PA, 2.92 to 13.24 mm day−1 for PJ and 3.83 to17.17 mm day−1 for the BO species. At the end of March 2009, theirrigation system broke down and consequently the ET rates werenot recorded until May.

During DS2, from May 2009 to October 2009, the ET rates ofthe species decreased from 29% to 50% compared to the WS. How-ever, two peaks of ET were recorded in the third 10-day period ofJune and the second 10-day period of August. Once again, the ETrates of the BGG and BO species were lower and higher, respec-tively, than the other species. During DS2, the ET rates were higherthan during DS1, ranging from 1.19 to 6.12 mm day−1 for BGG,1.67 to 8.46 mm day−1 for PJ, 2.46 to 9.87 mm day−1 for PA, 2.73to 10.00 mm day−1 for BVV, and 3.23 to 10.60 mm day−1 for BO.

The crop coefficients closely followed the same trend as the 10-day ET (Fig. 4). During DS1, the different bamboo species showedsimilar kc values: from 0.5 to 1.2 and 0.7 to 1.8 for the BGG and BOspecies, respectively. During the WS the kc ranged from 0.5 to 3.8and 1.0 to 5.7 for the BGG and BO species, respectively, and, duringDS2, from 0.5 to 3.0 and 1.3 to 5.2 for the BGG and BO species,respectively.

The average kc was significantly different for the BGG and BOspecies (p < 0.05) with values of 1.1 and 1.9, respectively (Table 4).The other species showed average kc of 1.4, 1.5 and 1.6 for the PJ,BVV and PA species, respectively.

The ETc rates correlated significantly with weather parameters(Table 2). Average 10-day ET for the species were positively corre-lated with minimum temperature. The ET rates of tropical species,i.e. BGG and BVV species seem to be more strongly correlated with

J. Piouceau et al. / Agricultural Water Management 137 (2014) 15–22 19

Table 2Pearson correlation coefficients (r) for ET rates (mm day−1) of five bamboo species, reference evapotranspiration (ET0) and weather parameters.

Parameters BGG PA PJ BO BVV ET0

Maximum temperature 0.27 0.1 0.24 0.25 0.1 0.60***

Average temperature 0.52*** 0.37* 0.48** 0.52*** 0.36* 0.66***

Minimum temperature 0.72*** 0.63*** 0.66*** 0.74*** 0.61*** 0.57***

Maximum relative humidity 0.34* 0.19 0.33* 0.32 0.16 0.32*

Average relative humidity 0.45** 0.50** 0.43** 0.49** 0.48** 0.05Minimum relative humidity 0.39* 0.47** 0.37* 0.44** 0.46** −0.04ET0 0.44** 0.36* 0.44** 0.46** 0.34* 1

B chys aS

mtrr

3e

obcsaatt

lctc

batgls

F(P

ambusa multiplex (BGG), Bambusa oldhamii (BO), Bambusa vulgaris (BVV), Phyllostaymbols indicate significant correlation: * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001.

inimum temperature (r = 0.72 and r = 0.74, respectively) than theemperate species, i.e., PA and PJ species (r = 0.63 and r = 0.66,espectively). As expected, the ET rates of all the species were cor-elated with average temperature and average relative humidity.

.3. Bamboo morphology and influence on the actualvapotranspiration

Between the beginning and end of the experiment, the numberf culms, culm diameter and height, and the aboveground freshiomass yield increased significantly (p < 0.001). The number ofulms increased by a factor of between 2.1 and 9.4 for the BO and PApecies, respectively, the culm diameter by a factor of between 1.3nd 2.0 for the PA and BGG species, respectively, the culm height by

factor of between 1.1 and 1.7 for the PA and BGG species, respec-ively, and the biomass yield by a factor of between 5.4 and 18.1 forhe BO and BGG species, respectively.

Table 3 shows that the species can be grouped by their morpho-ogical parameters. The BO and BVV species showed fewer, tallerulms with larger diameters compared to the other species. In con-rast, the PJ and BGG species showed a greater number of shorterulms with smaller diameters than the BO and BVV species.

The annual total ET (mm year−1) correlates with the total freshiomass (r = 0.72, p < 0.01), total fresh leaf mass (r = 0.62, p < 0.01)nd total leaf area (r = 0.61, p < 0.05). The BO species, which showedhe highest total ET rate, also produced the highest total above-

round fresh mass (with an average of 39 kg per clump), fresheaf mass (8.6 kg) and total leaf area (117.6 m2). Likewise, the BGGpecies, which showed the lowest ET rate, also produced the lowest

Aug-08

Sep-08

Oct-08

Nov-08

Dec-08

Jan-0

9

Feb-09

Mar-09

Apr-09

May-09

Jun-0

9

Jul-0

9

Aug-09

Sep-09

Oct-09

Nov-09

k c

0

1

2

3

4

5

6BGG BO BVV PA PJ

ig. 4. Ten-day average crop coefficients (kc) for the species: Bambusa multiplexBGG), Bambusa oldhamii (BO), Bambusa vulgaris (BVV), Phyllostachys aurea (PA) andseudosasa japonica (PJ).

urea (PA) and Pseudosasa japonica (PJ).

biomass, fresh leaf mass and total leaf area (14.5 kg per clump, 3 kgand 30 m2, respectively).

4. Discussion

The ET rates varied significantly throughout the year with a highseasonal factor (p < 0.001). All the species showed the same tem-poral pattern, with the highest ET rates recorded during the WS(October 2008 to April 2009), when the temperatures and rainfallreached their highest levels of the year (Fig. 2a). As mentioned pre-viously, the WS is the main growth period for bamboo during whichnew shoots emerge from the soil (Kleinhenz and Midmore, 2001).The high ET rates measured during this season can be ascribedtherefore to the increase in total number of shoots (Table 3), whichleads to an increase in the total leaf area of the bamboo clumps andtherefore to an increase in ET. Moreover, before this WS, the ET rateswere lower and closely followed the reference evapotranspiration’svalues (Fig. 3). We also noticed that ET rates increased between DS1and DS2 despite lower temperatures in DS2, which indicates thatthe increase in biomass during the WS had a significant effect onthe increase in ET in DS2.

During the WS, the ET rates showed four peaks in December,January and March. These peaks may be explained by the increasein average minimum temperatures at the corresponding periods.Indeed, Table 2 shows that ET rates correlate more closely withminimum temperature than any other climate parameter. Manystudies have demonstrated that an increase in ET is closely corre-lated with an increase in minimum air temperatures. Conversely,when the temperature drops below a certain value – depending onthe species –, the stomatal conductance of leaves and the roots’hydraulic conductivity decrease (Jarvis, 1976; Mu et al., 2007),with a consequent decrease in evapotranspiration. Air temperaturealso strongly correlates with photosynthesis (Gratani et al., 2008;Van Goethem et al., 2013), which also impacts transpiration inbamboo (Agata et al., 1985). There are important differences in opti-mum temperature ranges for photosynthesis activity depending onthe bamboo species’ ecological distribution. Gratani et al. (2008)showed that bamboo of the genus Phyllostachys from temperateclimates have a higher sensitivity to high temperature and that,conversely, bamboo of the genus Bambusa from tropical and sub-tropical climates have a higher sensitivity to low temperatures. Forexample, in the case of the genus Bambusa (i.e. tropical species), thenet photosynthesis decreased more than 50% when temperaturesdropped below 16.2 ◦C (Gratani et al., 2008). This physiologicalbehavior and its consequences on the plant’s transpiration are con-sistent with our study, in as much as we have noticed a bettercorrelation between the ET rates and minimum temperatures inthe tropical species (BO and BGG) than in the temperate species(PA and PJ) (Table 2).

The average and total ET, and kc varied significantly depend-ing on the species (p < 0.001). Among the five species studied, theaverage and total ET, and kc of the BO and BGG species were signifi-cantly different to the other three (BVV, PA and PJ; p < 05) (Table 4).

20 J. Piouceau et al. / Agricultural Water Management 137 (2014) 15–22

Table 3Morphological parameters and aboveground fresh biomass for five bamboo species at the beginning (initial) and end (final) of the experiment.

Parameters BGG PA PJ BVV BO

Number of culms Initial 72 ± 12.8d 5.7 ± 2.3a 36.7 ± 2.8c 11.7 ± 3ab 15.7 ± 1.2bFinal 325.3 ± 78.8c 54 ± 8b 195 ± 23c 25.7 ± 3.5a 34.3 ± 3.2ab

Culm diameter (mm) Initial 3.3 ± 0.1a 8.4 ± 0.7bc 7.3 ± 0.3b 11.1 ± 1c 11.2 ± 1cFinal 6.6 ± 0.5a 11.3 ± 0.8b 9.5 ± 0.4b 22.0 ± 2.8c 20.5 ± 1.0c

Height (cm) Initial 115.1 ± 5.4a 205.6 ± 7.4b 170 ± 13.3b 170.9 ± 22.8b 170 ± 13.3bFinal 201.4 ± 13.9a 239.4 ± 9.9ab 208.1 ± 6.8a 259.3 ± 26.9b 274.4 ± 10.0b

Aboveground freshbiomass (kg)

Initial* 0.8 ± 0.1a 1.4 ± 0.3a 3.4 ± 0.5b 2.9 ± 1.1b 7.2 ± 0.7cFinal 14.5 ± 0.3a 18.5 ± 4.3ab 27.0 ± 3.4b 24.6 ± 2.0ab 39 ± 0.1c

V es betP

Inysia

71sesrapmKr1os(eosC2TBcbttp(rt

TM

Ve(

alues are means ±SE. The values with different letters indicate significant differencseudosasa japonica (PJ), Bambusa vulgaris (BVV) and Bambusa oldhamii (BO).

* Estimated biomass using allometric equation.

n addition to the species’ original ecological habit, the main expla-ation for the difference in their ET rates is their different biomassields, as shown in Tables 3 and 4. As observed for the BO and BVVpecies, a high biomass is related to a large total leaf area, whichmplies high ET (Table 4). Indeed, these two species produced tallernd larger culms compared with the three others (Table 3).

For all the species, the average ET ranged between 4 and mm day−1, with maximum values during the WS of 12.3 and7.18 mm day−1, for the BGG and BO species, respectively. Astressed in the introduction, few studies have dealt with thevapotranspiration of bamboo species. By using the sap-flow mea-urement method, Dierick et al. (2010) and Komatsu et al. (2010)eported transpiration rates of 1.4 mm day−1 and of between 3nd 4 mm day−1 for Bambusa blumeana Schult.f. and Phyllostachysubescens J. Houz., respectively. However, unlike in our experi-ent, the soil evaporation in these studies was not included Only

leinhenz and Midmore (2002) have estimated the evapotranspi-ation of B. oldhamii (BO) with ET rates ranging between 9 and3 mm day−1. These values are consistent with those reported inur experiment. However, the latter study was carried out usingmall lysimeters (pots) which can generate experimental artifactsRana and Katerji, 2000). Since there is little data available on thevapotranspiration of bamboo and other giant grasses in literature,ur results can be compared only with other plant species. Forugar cane grown on Reunion Island (the site of our own study),hopart et al. (2007) reported an average annual ET of between.90 and 2.93 mm day−1, with maximum values of 3.50 mm day−1.hese values are lower than those found in our study of the BGG andO species; however, they correspond to a biomass yield for sugarane of 101 t ha−1 of fresh mass (Chopart et al., 2007), whereas theiomass yields in our study ranged between 120 and 325 t ha−1 forhe BGG and BO species, respectively, based on the surface area ofhe lysimeter. Although willow and poplar are trees and temperate

lant species, they, like bamboo are used for wastewater treatmentDimitriou and Rosenqvist, 2011). Indeed, they have a high growthate, biomass yields close to those of temperate bamboo species (10o 15 t ha−1 yr−1 of dry mass) and high water uptakes (Aronsson

able 4orphological parameters and evapotranspiration rates at the end of the experiment afte

Parameters BGG PA

Total aboveground fresh biomass (kg) 14.5 ± 0.3a 18.5 ± 4.3ab

Average fresh culm mass (g) 45.5 ± 5.6a 184.6 ± 25.4cTotal fresh leaf mass (kg) 3.0 ± 0.8a 3.5 ± 0.4a

Total leaf area (m2) 30.0 ± 6.2a 35.6 ± 3.7a

Average ET (mm day−1) 4 ± 0.3a 5.5 ± 0.3b

Average ET0 (mm day−1) 3 ± 0.1 3 ± 0.1

kc 1.1 ± 0.1a 1.6 ± 0.2b

Annual total ET (mm yr−1) 1270.1 ± 67.1a 1605.5 ± 42.7

alues are means ± SE. The values with different letters indicate significant differencevapotranspiration (ET0); crop coefficient (kc); Bambusa multiplex (BGG); Phyllostachys aBO).

ween the species (LSD, p < 0.05). Bambusa multiplex (BGG), Phyllostachys aurea (PA),

and Bergstrom, 2001; Aronsson and Perttu, 2011; Dimitriou andAronsson, 2011; Perttu and Kowalik, 1997). The ET values for theyoung bamboo plants in our study are similar to those reportedfor willow and poplar using the same lysimeter method as in ourstudy. Under a Mediterranean climate, Guidi et al. (2008) reportedan average of between 3.33 and 4.02 mm day−1 (with a maximum of15.46 mm day−1) for poplar and between 3.47 and 6.65 mm day−1

(with a maximum of 21.7 mm day−1) for willow,. Other authorssuch as Pistocchi et al. (2009) reported similar ET values for poplarand willow.

The kc values follow the same pattern as ET and ranged from1.1 to 1.9 for the BGG and BO species, respectively. The valuesrecorded in DS1, are consistent with the values reported by Payetet al. (2009), who reported a kc of between 0.48 and 1.2 for maizecrops on Reunion Island. After the dry season, the kc values recordedfor bamboo were higher than the maximum average kc found in lit-erature for crops which is 1.25 in sugar cane (Allen et al., 1998). Asexpected, the average kc values for the bamboo in our study arecloser to those for willow and poplar in short rotation coppices.Guidi et al. (2008) reported a kc of between 0.33 and 5.30 for wil-low and between 0.36 and 4.28 for poplar. Pistocchi et al. (2009)reported a kc of between 2.54 and 4.03 for both willow and poplarcombined.

The evapotranspiration rates for the young bamboo in our studyappear high, raising the question of possible experimental arti-facts. The main potential source of error with the lysimeter methodcan be drastic differences in atmospheric conditions between thelysimeter and its immediate surroundings; this artifact is knownas the “oasis effect” (Rana and Katerji, 2000). If the surroundingarea consists of bare or poorly vegetated earth, net radiation inexcess of latent heat is converted into sensible heat that is advectedtoward the lysimeter. Such a net input of energy into the lysime-ter canopy results in an overestimation of ET (Rana and Katerji,

2000). As shown by Wegehenkel and Gerke (2013), the oasis effectin lysimeters similar to those used in our study (i.e. 1 m2) can peri-odically occur during the summer when solar radiation reaches itsmaximum intensity. The oasis effect could not be totally excluded

r one year of growth for five bamboo species.

PJ BVV BO

27.0 ± 3.4b 24.6 ± 2.0ab 39 ± 0.1c 95.9 ± 8.3b 514.7 ± 196.7cd 1149.7 ± 102.2d

5.0 ± 0.5b 7.5 ± 0.6c 8.6 ± 0c32.4 ± 4.7a 81.3 ± 4.6b 117.6 ± 2.6c5.1 ± 0.3b 5.4 ± 0.3b 7 ± 0.4c3 ± 0.1 3 ± 0.1 3 ± 0.11.4 ± 0.1b 1.5 ± 0.2b 1.9 ± 0.2c

b 1605.8 ± 98.5b 1704.1 ± 27.9b 2218.7 ± 30.8c

s between the species (LSD, p < 0.05). Actual evapotranspiration (ET); referenceurea (PA); Pseudosasa japonica (PJ); Bambusa vulgaris (BVV) and Bambusa oldhamii

ater M

itedcoooo

bpgtrt

5

lomAbscootbsdbm

A

iddIcc

dbbaaSLc

R

A

A

A

A

J. Piouceau et al. / Agricultural W

n our case, especially during the wet season, which correspondso the period of the year when solar radiation is the highest (Allent al., 1998). However, comparing our results with other studiesealing with evapotranspiration in crops such as maize and sugarane, and in poplar and willow, suggests that the data reported inur study is reliable. Furthermore, in addition to providing new datan bamboo species, our study could provide useful information forptimizing the design for future experiments on ET in bamboo andther giant grasses

Lastly, the ET values measured were those for three-year-oldamboo that is to say for non-mature bamboo. Indeed, a bamboolantation reaches maturity in five to eight years, depending on therowth conditions. This suggests that for a mature bamboo plan-ation of giant bamboo such as the BO species, in which culms caneach 18 m in height at maturity (Castaneda-Mendoza et al., 2005),he ET rates would be even higher.

. Conclusions

The experiment reported here is – to our knowledge – the firstysimeter experiment to investigate the evapotranspiration ratesf different species of bamboo. The results show that seasons andinimum temperatures are the main climatic factors affecting ET.s expected, the differences in ET observed between species cane explained by the aboveground biomass yield and the total leafurface area, the two morphological parameters showing a highorrelation with ET rates. Among the five bamboo species used, B.ldhamii had the highest ET rate and produced the highest amountf biomass during the experiment. In comparison with evapora-ion rates for other high-biomass-producing plants, those for youngamboo plants are similar to those for willow and poplar used inhort rotation coppices. In conclusion, our study provides usefulata for the design and management of vegetation filters usingamboo plantations. The data could also be useful for the wateranagement of commercial bamboo plantations.

cknowledgments

The research was funded by the following French governmentnstitutions: “Fonds Unique Interministériel”, the “Conseil Régionale La Réunion”, the “Conseil Régional de Provence-Alpes-côte’Azur” and the “Département de La Réunion” as part of its “Run

nnovation II” project, supported by the Reunion Island competitiveluster “Qualitropic” and the “Pôle Risques” regional competitiveluster in Aix-en-Provence.

We would like to thank Alexandre Perrusot of “La Bambusaieu Guillaume”, Reunion Island for sharing his knowledge of bam-oo with us and for allowing us to carry out the experiment at hisamboo nursery. We would like also to thank Michel Montevillend Sébastien Séverin for their help in building the experimentnd ensuring the smooth daily running of the irrigation system.pecial thanks go to Laure Prete, Remi Hidouci, Stéphane Maillot,aurent De Fondaumière and Charles-Lee Hoareau for their help inollecting field data and for their technical support.

eferences

boukhaled, A., Alfaro, J., Smith, M., 1982. Lysimeters. Food and Agriculture Organi-zation of the United Nations (FAO), Rome, Italy.

gata, W., Hakoyama, S., Kawamitsu, Y., 1985. Influence of light intensity, temper-ature and humidity on photosynthesis and transpiration of Sasa nipponica andArundinaria pygmaea. The Botanical Magazine 98, 125–135.

llen, R.G., Pereira, L.S., Howell, T.A., Jensen, M.E., 2011. Evapotranspiration informa-

tion reporting: I. Factors governing measurement accuracy. Agricultural WaterManagement 98, 899–920.

llen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration—Guidelines for Computing Crop Water Requirements. Food and Agriculture Orga-nization of the United Nations (FAO), Rome, Italy.

anagement 137 (2014) 15–22 21

Arfi, V., Bagoudou, D., Korboulewsky, N., Bois, G., 2009. Initial efficiency of a bamboogrove-based treatment system for winery wastewater. Desalination 246, 69–77.

Aronsson, P., Perttu, K., 2001. Willow vegetation filters for wastewater treatment andsoil remediation combined with biomass production. The Forestry Chronicle 77,293–299.

Aronsson, P.G., Bergstrom, L.F., 2001. Nitrate leaching from lysimeter-grown short-rotation willow coppice in relation to N-application, irrigation and soil type.Biomass and Bioenergy 21, 155–164.

Aronsson, P., Perttu, K., 2011. Willow vegetation filters for wastewater treatment andsoil remediation combined with biomass production. The Forestry Chronicle 87,797-797.

Castaneda-Mendoza, A., Vargas-Hernandez, J., Gomez-Guerrero, A., Valdez-Hernandez, J.I., Vaquera-Huerta, H., 2005. Carbon accumulation in theaboveground biomass of a Bambusa oldhamii plantation. Agrociencia 39,107–116.

Chazottes, V., Le Campion-Alsumard, T., Peyrot-Clausade, M., Cuet, P., 2002. Theeffects of eutrophication-related alterations to coral reef communities on agentsand rates of bioerosion (Reunion Island, Indian Ocean). Coral Reefs 21, 375–390.

Chopart, J.L., Mézino, M., Aure, F., Le Mézo, L., Mété, M., Vauclin, M., 2007.OSIRI: a simple decision-making tool for monitoring irrigation of small farmsin heterogeneous environments. Agricultural Water Management 87, 128–138.

Christanty, L., Mailly, D., Kimmins, J.P., 1996. Without bamboo, the land dies:biomass, litterfall, and soil organic matter dynamics of a Javanese bambootalun–kebun system. Forest Ecology and Management 87, 75–88.

Collin, B., Doelsch, E., Keller, C., Panfili, F., Meunier, J.-D., 2013. Effects of silicon andcopper on bamboo grown hydroponically. Environmental Science and PollutionResearch, 1–14.

Dierick, D., Hölscher, D., Schwendenmann, L., 2010. Water use characteristics of abamboo species (Bambusa blumeana) in the Philippines. Agricultural and ForestMeteorology 150, 1568–1578.

Dimitriou, I., Aronsson, P., 2011. Wastewater and sewage sludge application to will-ows and poplars grown in lysimeters—plant response and treatment efficiency.Biomass and Bioenergy 35, 161–170.

Dimitriou, I., Rosenqvist, H., 2011. Sewage sludge and wastewater fertilisation ofshort rotation coppice (SRC) for increased bioenergy production: biological andeconomic potential. Biomass and Bioenergy 35, 835–842.

Gratani, L., Crescente, M.F., Varone, L., Fabrini, G., Digiulio, E., 2008. Growth patternand photosynthetic activity of different bamboo species growing in the BotanicalGarden of Rome. Flora—Morphology, Distribution, Functional Ecology of Plants203, 77–84.

Gui, R., Leng, H., Zhuang, S., Zheng, K., Fang, W., 2011. Aluminum tolerance in Mosobamboo (Phyllostachys pubescens). The Botanical Review 77, 214–222.

Guidi, W., Piccioni, E., Bonari, E., 2008. Evapotranspiration and crop coefficient ofpoplar and willow short-rotation coppice used as vegetation filter. BioresourceTechnology 99, 4832–4840.

Jarvis, P.G., 1976. The interpretation of the variations in leaf water potential andstomatal conductance found in canopies in the field. Philosophical Trans-actions of the Royal Society of London. Series B, Biological Sciences 273,593–610.

Kleinhenz, V., Midmore, D.J., 2001. Aspects of Bamboo Agronomy, Advancesin Agronomy. Academic Press, pp. 99–153, http://www.elsevier.com/books/advances-in-agronomy/sparks/978-0-12-000792-9.

Kleinhenz, V., Midmore, D., 2002. Improved management practices for culinary bam-boo shoots. In: Local and Export Markets: A Report for the Rural IndustriesResearch and Development Corporation. Rural Industries Research & Develop-ment Corporation, Kingston, ACT, Australia.

Komatsu, H., Onozawa, Y., Kume, T., Tsuruta, K., Kumagai, T.o., Shinohara, Y.,Otsuki, K., 2010. Stand-scale transpiration estimates in a Moso bamboo forest:II. Comparison with coniferous forests. Forest Ecology and Management 260,1295–1302.

Komatsu, H., Onozawa, Y., Kume, T., Tsuruta, K., Shinohara, Y., Otsuki, K., 2012.Canopy conductance for a Moso bamboo (Phyllostachys pubescens) forest inwestern Japan. Agricultural and Forest Meteorology 156, 111–120.

Kume, T., Onozawa, Y., Komatsu, H., Tsuruta, K., Shinohara, Y., Umebayashi, T.,Otsuki, K., 2010. Stand-scale transpiration estimates in a Moso bamboo forest:(I) Applicability of sap flux measurements. Forest Ecology and Management 260,1287–1294.

Licht, L.A., Isebrands, J.G., 2005. Linking phytoremediated pollutant removal tobiomass economic opportunities. Biomass and Bioenergy 28, 203–218.

McCutcheon, S.C., Schnoor, J.L., 2003. Phytoremediation:Transformation and Controlof Contaminants. John Wiley & Sons, Hoboken, NJ.

Mu, Q., Heinsch, F.A., Zhao, M., Running, S.W., 2007. Development of a global evapo-transpiration algorithm based on MODIS and global meteorology data. RemoteSensing of Environment 111, 519–536.

Naim, O., 1993. Seasonal responses of a fringing reef community to eutrophica-tion (Reunion Island, Western Indian Ocean). Marine Ecology Progress Series99, 137–151.

Pauliukonis, N., Schneider, R., 2001. Temporal patterns in evapotranspiration fromlysimeters with three common wetland plant species in the eastern United

States. Aquatic Botany 71, 35–46.

Payet, N., Findeling, A., Chopart, J.-L., Feder, F., Nicolini, E., Saint Macary, H., Vau-clin, M., 2009. Modelling the fate of nitrogen following pig slurry applicationon a tropical cropped acid soil on the island of Réunion (France). Agriculture,Ecosystems & Environment 134, 218–233.

2 ater M

P

P

P

R

S

S

S

S

2 J. Piouceau et al. / Agricultural W

erttu, K.L., Kowalik, P.J., 1997. Salix vegetation filters for purification of waters andsoils. Biomass and Bioenergy 12, 9–19.

iouceau, J., Bois, G., Panfili, F., Anastase, M., Dufossé, L., Arfi, V., 2014. Effects of HighNutrient Supply on the Growth of Seven Bamboo Species. International Journalof Phytoremediation 16, 1042–1057.

istocchi, C., Guidi, W., Piccioni, E., Bonari, E., 2009. Water requirements of poplar andwillow vegetation filters grown in lysimeter under Mediterranean conditions:results of the second rotation. Desalination 246, 137–146.

ana, G., Katerji, N., 2000. Measurement and estimation of actual evapotranspira-tion in the field under Mediterranean climate: a review. European Journal ofAgronomy 13, 125–153.

curlock, J.M.O., Dayton, D.C., Hames, B., 2000. Bamboo: an overlooked biomassresource? Biomass and Bioenergy 19, 229–244.

emple, S., 1997. Algal growth on two sections of a fringing coral reef subjectto different levels of eutrophication in Réunion Island. Oceanologica Acta 20,851–861.

hukla, O.P., Juwarkar, A.A., Singh, S.K., Khan, S., Rai, U.N., 2011. Growth responses

and metal accumulation capabilities of woody plants during the phytoremedia-tion of tannery sludge. Waste Management 31, 115–123.

ingh, S.K., Juwarkar, A.A., Pandey, R.A., Chakrabarti, T., 2008. Applicability of highrate transpiration system for treatment of biologically treated distillery effluent.Environmental Monitoring and Assessment 141, 201–212.

anagement 137 (2014) 15–22

Smith, V.H., 2009. Eutrophication, Encyclopedia of Inland Waters. Academic Press,Oxford, pp. 61–73.

Smith, V.H., Tilman, G.D., Nekola, J.C., 1999. Eutrophication: impacts of excess nutri-ent inputs on freshwater, marine, and terrestrial ecosystems. EnvironmentalPollution 100, 179–196.

Stapleton, C., 1998. Form and function in the bamboo rhizome. Journal of AmericanBamboo Society 12, 21–29.

Tedetti, M., Cuet, P., Guigue, C., Goutx, M., 2011. Characterization of dissolved organicmatter in a coral reef ecosystem subjected to anthropogenic pressures (La Réu-nion Island, Indian Ocean) using multi-dimensional fluorescence spectroscopy.Science of the Total Environment 409, 2198–2210.

Van Goethem, D., De Smedt, S., Valcke, R., Potters, G., Samson, R., 2013. Seasonal, diur-nal and vertical variation of chlorophyll fluorescence on Phyllostachys humilis inIreland. PLoS One 8, e72145.

Vymazal, J., 2011. Constructed wetlands for wastewater treatment: five decades ofexperience. Environmental Science & Technology 45, 61–69.

Wegehenkel, M., Gerke, H.H., 2013. Comparison of real evapotranspiration measured

by weighing lysimeters with simulations based on the Penman formula and acrop growth model. Journal of Hydrology and Hydromechanics 61, 161–172.

Yadav, S.K., Thawale, P.R., Kulkarni, A.V., Juwarkar, A.A., 2010. Phytoremediationtechnology for wastewater treatment: high rate transpiration system. Interna-tional Journal of Environment and Pollution 43, 117–128.