Carbon assimilation and water use efficiency of a perennial bioenergy crop (Cynara cardunculus L.) in Mediterranean environment

Agricultural and Forest Meteorology 217 (2016) 137–150

Contents lists available at ScienceDirect

Agricultural and Forest Meteorology journal homepage: www.elsevier.com/locate/agrformet

Carbon assimilation and water use efficiency of a perennial bioenergy crop (Cynara cardunculus L.) in Mediterranean environment G. Rana a , R.M. Ferrara a,∗ , D. Vitale b , L. D’Andrea a , A.D. Palumbo a a Consiglio per la Ricerca in Agricoltura e l’Analisi dell’Economia Agraria (CREA)—Research Unit for Cropping Systems in Dry Environments (CREA-SCA), via C. Ulpiani, 5, Bari 70125, Italy b DIBAF, Dipartimento per l’Innovazione nei sistemi Biologici, Agroalimentari e Forestali, University of Tuscia, via C. de Lellis, Viterbo 01100, Italy

a r t i c l e

i n f o

Article history: Received 4 February 2015 Received in revised form 23 November 2015 Accepted 29 November 2015 Available online 11 December 2015 Keywords: Eddy covariance Gross primary production Net ecosystem exchange Respiration Crop actual evapotranspiration

a b s t r a c t Here we investigate how cardoon (Cynara cardunculus L.), energetic crop cultivated under Mediterranean climate in rainfed conditions, is adapted to the environment. Two main resources used for producing biomass are analysed in detail: water (H2 O) and carbon dioxide (CO2 ). Following micrometeorological approach, the eddy covariance technique has been used for monitoring H2 O and CO2 exchanges between canopy and atmosphere in order to investigate the dynamics of the cardoon growth at field level and to compute the Gross Primary Production (GPP). The dynamics of canopy CO2 assimilation in terms of GPP, evapotranspiration (ET) and water use efficiency (WUEGPP , as ratio between seasonal GPP and seasonal ET and WUEagro as ratio between yield and seasonal ET) were analysed during three successive growth seasons in a south Italy site. The environmental drivers of CO2 assimilation and ET were analysed at instantaneous scale. The crop showed increasing resource use efficiency along the three seasons of experiment for all considered resources: in particular, for the last two seasons cumulated GPP increased and cumulated ET decreased. It seemed to require a season for its establishment to the environment, improving the use of water and CO2 assimilation in the second and third season. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The European Community perceives bioenergy as a mean to reduce its dependence on external energy supplies and to mitigate pressing issues of climate change. In fact, directive 2009/28/EC of 23 April 2009 on the promotion of the use of energy from renewable sources (http://ec.europa.eu/energy/renewables/biofuels/biofuels) established mandatory national targets consistent with a 20% share of energy from renewable sources in energy consumption by 2020. Among energy from renewable source, the increasing biomass and biodiesel uses are key tools proposed to reduce the dependence on imported oil and oil products, thus improving the security of energy supply in the medium and long term, also through the cultivation of dedicated species, the so called energy crops. Among these, Cynara cardunculus L. var altilis D.C. was proposed as bioenergy crop, especially in the biofuel and lignocellulosic chains for heating or power production. Rather, cultivated cardoon is considered as one of the most promising first generation bioenergy crops, also in marginal rural locations with limited resources. In fact, it

∗ Corresponding author. Tel.: +39 0805475007; fax: +39 0805475023. E-mail address: [email protected] (R.M. Ferrara). http://dx.doi.org/10.1016/j.agrformet.2015.11.025 0168-1923/© 2015 Elsevier B.V. All rights reserved.

is a drought tolerant herbaceous perennial plant (Curt et al., 2002) and it can be easily propagated by seed with important advantages for crop management. Thus, it can be cultivated without irrigation (rainfed), hence it is particularly adapted to the Mediterranean environments, where the major limiting production factor is just water. Cardoon was proved to carry out high yields (Foti et al., 1999; Piscioneri et al., 1999; Grammelis et al., 2008; Fernández et al., 2006; Ierna et al., 2012). Several works were published on its agronomics in Mediterranean region, demonstrating the strong potentiality of this crop as energetic source, both as biomass and biofuel from seeds (Raccuia and Melilli, 2007; Ierna and Mauromicale, 2010; Gominho et al., 2011) or, more marginally, for industrial and animal use (Christaki et al., 2012). After a long trial in Mediterranean region, Angelini et al. (2009) found that the average biomass yield of two cardoon cultivars obtained with 860 mm mean annual rainfall was 14 and 15 Mg ha−1 year−1 . In this case the maximum yield was reached at the third year for both cultivars. Gherbin et al. (2001) reported a decreasing cardoon yield from the first to the last (fifth in this case) season for all the 17 cultivars grown under Mediterranean climate. In this work the authors found that the yield of the different cardoon cultivars had a very large variability, ranging between 5 and 14 Mg ha−1 year−1 , with

138

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

500 mm mean annual rainfall. In all these previous studies the cardoon crop was cultivated in small plots (about 100 m2 ), the typical ones for agronomic purpose investigations. Actually, there are no surfaces specifically dedicated to cardoon for energy in Italy. However, to evaluate the adaptability of a crop to an environment, yield and its partitioning analysis is not enough (Fernández et al., 2005; Richardson et al., 2012). Conversely, it is essential to give information about the use of natural resources in forming biomass to assess agronomic benefit and environmental impact in a changing world. Long-term studies on CO2 and water vapour exchange in agricultural ecosystems can improve the understanding on the biophysical factors affecting changes in the functionalities of this kind of cropping system (Beer et al., 2009; Suyker and Verma, 2010). For bioenergy crops, similar recent studies were carried out on reed canary grass in Finland (Shurpali et al., 2009, 2013), on young switchgrass in the northeastern USA (Skinner and Adler, 2010), on switchgrass in Oklahoma, USA (Wagle and Kakani, 2014), on miscanthus and switchgrass in Central Illinois, USA (Zeri et al., 2011). Furthermore, present climate changes and climate variability impose that experimental investigations would be carried out in order to actualise the biological and physiological relationships among natural resources, climatic variables and agricultural crops (Fischer et al., 2007; Giannakopoulos et al., 2009; Espadafor et al., 2011). Non-invasive and continuous measurements of energy and mass flow between a crop and its environment are essential for understanding crop growth and its interactions with the environment over time, and to compare and evaluate cropping systems (i.e. Asseng and Hsiao, 2000). The micrometeorological method eddy covariance (EC) is considered the most direct approach for measuring canopy gas exchange at ecosystem scale and, thus, the results can be more easily generalised. This study was conducted in a field submitted to Mediterranean semi-arid climate, in order to understand how cardoon spends and uses water and CO2 during three successive growth seasons. In particular, the analysis has been carried out by detailing the dynamics of canopy CO2 assimilation, of evapotranspiration (ET) and water use efficiency (WUE), from emergence up to harvest, using an EC system for measuring H2 O and CO2 fluxes at hourly scale. In our knowledge this is the first study on the topic for a bioenergy crop in Mediterranean climate.

The cardoon, selected at the Polytechnic University of Madrid for the adaptability to semi-arid environments, was sown on 31th October 2009 on an area of about 2 ha, with a plant density of 30,000 plants ha−1 and an inter-row spacing of 1.2 m. Main tillage was conducted in the autumn 2009 as medium-depth ploughing (0.3 m). Seed bed preparation was conducted immediately before sowing, by a pass with a disk harrow. Pre-plant fertiliser was distributed at a rate of 100 kgP2 O5 ha−1 (triple superphosphate). Starting from the second growing season, the crop was managed with low inputs, by avoiding soil tillage and mineral fertilisation. No water was applied as irrigation so that the cardoon regrew immediately after the first rain in autumn. During the seasonal growing cycle, plants were picked up every month with randomised samples. On each sample, the fresh and dry weight of biomass, the leaf area index (LAI) and the height were measured. The aboveground biomass was harvested in summer (August 2010, 2011 and 2012) when the crop was completely dry (about 70–80% of dry matter (DM)). The plants were placed in a thermoventilated oven at 65 ◦ C until constant weight. The production was finalised to extract the oil for energy purposes. The percentage of C in the harvested biomass content was determined in triplicate using an elemental analyzer (EA-1108 model, Fisons Instruments, Crawley, UK). The phenological stages of the cardoon crop were registered according to Archontoulis et al. (2010), by recording the first day of each phase. In the spring of the fourth vegetation cycle (2013), the crop was fully destroyed by the insect Cassida deflorata Suffr. (order Coleoptera) which compromised once and for all the leaves and stems, interrupting the photosynthetic function of the plants. As a consequence, neither harvest occurred in the summer 2013 nor data could be collected. For this reason the experiment was limited to three successive crop growth seasons. The results here presented are for 1009 days, from the sowing to the third cut and harvest of cardoon crop. In this work three growth seasons were considered as following: (1) first season, from sowing up to first harvest (1 November 2009–19 August 2010); (2) second season, from the first day after the first harvest up to second harvest (20 August 2010–9 August 2011); (3) third season, from the first day after the second harvest up to third harvest (10 August 2011–6 August 2012). 2.2. The setup

2. Materials and methods 2.1. The site and the crop The field site for monitoring carbon (C) and water fluxes of cardoon was the CREA-SCA Research Unit experimental farm located in southern Italy (Rutigliano-Bari, 41◦ 01 N, 17◦ 01 E, altitude 147 m a.s.l). The site is characterised by typical Mediterranean semi-arid climate, with mild winters and warm-dry summers. On average, the rain is 535 mm year−1 mainly distributed between the autumn and the late winter. The annual water deficit (reference evapotranspiration (ET0 )—Rain) is 560 mm (Campi et al., 2009). The soil, classified as “Lithic Rhodoxeralf”, is characterised by clay texture, stable structure, shallow profile (0.6–1 m), because of a cracked limestone subsoil, and fast drainage. On average, its Total Organic Carbon (TOC) content is 12.0 g kg−1 . Such a low value is due to high temperatures reached during the spring–summer seasons. The field capacity and the permanent wilting point water contents on dry soil weight are equal to 30 and 18%, respectively. Therefore, with a bulk density of 1.15 Mg m−3 , the available soil water ranges from 80 to 140 mm because the soil profile depth is affected by high spatial variability (De Benedetto et al., 2012).

The EC flux tower was established in the field in November 2009 for making continuous measurements and the setup consisted of a three-dimensional sonic anemometer (USA-1, Metek GmbH, Germany) for acquiring the three wind component and sonic temperature and in a fast-responding open-path infrared gas analyser (IRGA, LI-7500, Li-COR Inc., Lincoln, NE, USA) in order to measure atmospheric CO2 and H2 O concentrations. The sonic anemometer was positioned at 1 m above the top of the crop and was adjusted to follow its growth, the maximum height of the EC sensors was 2.8 m; the gas analyser was placed at 0.3 m to the side of the anemometer in order to ensure measurement of the same air particles and to avoid airflow distortions. Micrometeorological measurements were recorded every 0.1 s (10 Hz) on an Industrial Computer (Advantech, USA) by a resident software (MeteoFlux, Servizi Territorio, S.n.c., Cinisello Balsamo, Italy) and were collected on a continuous hourly basis. Net radiation, Rn in W m−2 , was measured by means of two net radiometers (model Q*6, REBS, USA) 1 m above the top of the canopy. Soil heat flux, G in W m−2 , was measured by means of two heat flux plates (FP-1, Campbell Sci., USA), placed at 0.1 m depth into the soil. Rn and G were collected by a CR10X (Campbell Sci., USA) data logger at a frequency of 0.1 Hz and stored for 1 h on average. Soil

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

temperature in ◦ C were measured with two thermo-resistances (PT100) one 2 cm below and one 2 cm above each soil heat flux plate. At the same time continuous hourly meteorological data of air temperature, relative humidity, precipitation, global (Rg ) and photosynthetically active (PAR) radiation were recorded from an automatic weather station by standard sensors. In this work also the absorbed PAR was taken into consideration, it was calculated as APAR = PAR (1 − exp (−k × LAI))

(1)

with k extinction coefficient, that for the cardoon crop is equal to 0.773 (Archontoulis et al., 2011). Part of the data (1.6%, 2.9%, 7.8% and 8.8% for air temperature, VPD, PAR and precipitation, respectively) missed due to sensors malfunction or maintenance. The soil water content (SWC, volume of water in volume of soil, in m3 m−3 ) was measured at hourly scale, in two points close to the EC tower, at 20 and 40 cm depth, by Time-Domain Reflectometry technique (sensor CS605, system TDR100 Reflectometer, multiplexer SDM8X50 Coax, datalogger CR1000, all equipment by Campbell, Utah, USA). 3% (672 on 23,767) of SWC data were discarded due to sensors’ malfunctioning. 2.3. Eddy covariance flux calculation The EC method is based on estimating the transport of a scalar by turbulent motions of upward and downward moving air contained in the atmosphere (e.g. Kaimal and Finnigan, 1994; Lee et al., 2004). The turbulent fluxes of CO2 (Fc , ␮mol m−2 s−1 ), H2 O (mmol m−2 s−1 ), sensible heat (H, W m−2 ), latent heat (LE, W m−2 ) were calculated from the covariances between the respective scalar and the vertical wind speed (m s−1 ). The EddyPro® software (http://www.licor.com/eddypro) was used, employing the statistical test as described in Vickers and Mahrt (1997), the double coordinate rotation, 60 min block averaging, the maximum cross-covariance method and the spectral corrections proposed by Moncrieff et al. (1997). Finally, a density correction (the traditional WPL terms by Webb et al., 1980) was applied. By integrating the corrected Fc flux with the CO2 storage (i.e. the flux accumulated below the measurement height and not accounted by the EC system during periods of low turbulence) it could be defined the Net Ecosystem Exchange (NEE) of CO2 . Following Moureaux et al. (2012) and the references therein, in short ecosystems like croplands, the storage term is expected to be small and it could be computed from the single point method (Aubinet et al., 2000), conscious of the limits of this approach. 2.3.1. Flux quality control and filtering Fluxes were filtered to remove data corresponding to technical troubles, inaccurate meteorological conditions for the applicability of EC method, low spatial representativeness and violation of theoretical basis of the method. The following criteria were adopted. Quality flag for all fluxes was elaborated, applying tests for steady state conditions and well-developed turbulence (integral turbulence characteristics test) after Foken and Wichura (1996) in the version proposed by Mauder and Foken (2004) flagging policy. Additional low quality fluxes not yet filtered, i.e. those outside the physical plausible range and/or measured during rainy events that contaminate the open-path gas analyser optical sensor and/or identified as spike following the method proposed by Papale et al. (2006), were identified and filtered. The footprint analysis was conducted with the Lagrangian stochastic particle dispersion model described by Kljun et al. (2004). The winds mainly come from the NNW during daytime and

139

S during nighttime as the effect of sea and land breezes characterizing the geographical area. The estimated peak value of the footprint function typically stays at around 18 m distance from the tower and the fetch including 90% of the flux was around 50 m, exceeding the edge of the plot in cases of W and E wind directions. Hourly fluxes coming from these wind directions were filtered because not sufficiently representative of the plot (less than 10%), while, the fluxes along the prevailing wind direction since surrounding fields contained crops with similar phenology were not discarded. A friction velocity (u*) criteria was used to determine periods within low turbulence regime when fluxes are systematically underestimated by EC measurements. The automatic method described in Reichstein et al. (2005) was considered. This procedure was applied separately for each season to data subsets aggregated as a function of the phenological phases. The u* threshold values for the site were in the range 0.10–0.13 m s−1 . According to the above-described strategy, 46%, 44% and 55% of hourly EC data were classified as good to excellent quality, for each season, respectively. In particular, this percentages derived by summing 21%, 32% and 13% of messiness, 10%, 6% and 7% discarded values because of low quality, 4%, 4% and 8% identified as spikes, and finally, 19%, 14% and 17% deleted through the u* filtering procedure. Some of messiness was related to long-gaps: in particular November 2009 and September and October 2010 were completely missing due to EC system maintenance and instrument problems. As in these months the crops was completely absent or in emergency state, and since it is not recommended to fill long-gaps, NEE missing values were calculated through estimation from the soil respiration model proposed by Reichstein et al. (2003) (see Section 2.3.4). 2.3.2. Uncertainty analysis Among several methods that have been proposed in literature to the total random flux uncertainty quantification (see Richardson et al., 2012; reference therein), the 24-h differencing approach developed by Hollinger and Richardson (2005) was applied. The base assumption is that two flux measurements made at a single tower exactly 24 h apart (to minimise the diurnal effects) and under similar environmental conditions, are considered as a repeated and independent measurement of the same quantity, thus difference between them can be largely be attributed to random error rather than environmental forcing. Under this assumption, for each flux (f), the statistical properties of the differentiated time series (εf ) provide uncertainty quantification of the total random flux measurement errors ((εf )). Relationships between (εf ) and Fc , H and LE fluxes were established for 50 flux bins with the same number of data (N) in each bin, considering the high-quality filtered dataset relative to the three growing seasons. Previous researches have shown that (εf ) increases with the magnitude of the flux and that the error distribution is not Gaussian, but approximately it follows a Laplace distribution (see Richardson therein). et al., 2012 and reference  For these reasons, after estimatn  ing  (εt ) = 2ˇ with ˇ = f − f¯  /N and f¯ the median in each i=1 i bin, the following relation can be written:  (εt ) = a + bf¯

(2)

where a represents the underlying base uncertainty. For Fc , flux uncertainty increases by a relatively constant 0.09 (0.01) ␮mol m−2 s−1 per 1 ␮mol m−2 s−1 increase of flux from a minimum uncertainty of about 0.94 (0.1) ␮mol m−2 s−1 . For H, flux uncertainty increases by a relatively constant 0.11 W m−2 per 1 W m−2 increase of flux from a minimum uncertainty of about 14 W m−2 . For LE, the minimum uncertainty is around 11 W m−2 , and the rate of increase is about 0.21 W m−2 . Because the increase in uncertainty with an increase in flux magnitude is less than unity, the relative

140

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

 



uncertainty = (εt )/ f¯  decreases with the absolute value of the flux, resulting in about 20% of the measured flux according to Mauder et al. (2013). 2.3.3. Systematic error and the energy balance closure Energy balance closure has been historically accepted as quality test of eddy covariance data (see Wilson et al., 2002, among many other, for an extensive review on this issue). More precisely, the lack of surface energy balance closure can be considered as a quantification of the systematic error in EC flux measurements (Mauder et al., 2013), once the available energy at the surface (Rn − G—storage term J) is accurately quantified. Since the cardoon crop is supposed to have a huge biomass, in this work for the storage term the complete expression was used: J = StrC + StrG + StrP

(3)

where StrC is the canopy heat storage in biomass and water content, StrG is ground heat storage above the layer where the soil heat flux was measured and StrP is the energy flux for photosynthesis; all terms in W m−2 . The storage terms were calculated according to Meyers and Hollinger (2008); to calculate the term StrP it has been assumed that a canopy assimilation rate of 2.5 mgCO2 m−2 s−1 equates to an energy flux of 28 W m−2 (Nobel, 1974; Meyers and Hollinger, 2008). Following Mauder et al. (2013), the energy balance ratio, EBR, was determined for daily values as

K EBR =

K

i=1

(Hi + LEi )

+ Gi + Ji ) i=1 (Rni

(4)

with K equal to all available filtered data. In the present case we found EBR = 0.946, which was on line or even better than other reported results (Mauder et al., 2013), particularly for biomass crops (Wagle and Kakani, 2014; Zeri et al., 2011). As usually, a threshold of Rg > 20 W m−2 was accounted to distinguish the day by the night data. 2.3.4. Flux gap filling and NEE partitioning Gap filling of quality checked and filtered hourly flux time series was performed by the Marginal Distribution Sampling technique described in Reichstein et al. (2005) which has been demonstrated one of the best performing method (Moffat et al., 2007). According to the algorithm, missed hourly fluxes were filled with the average values calculated during similar meteorological conditions defined when Rg , Tair and VPD do not deviate by more than 50 W m−2 , 2.5 ◦ C and 5 hPa, respectively in a moving time window of 7 or 14 days to taken into account the different regimes of crop phenology. After gap filling was achieved, NEE was partitioned into the Gross Primary Production (GPP), i.e. the gross CO2 assimilation by photosynthesis of vegetation, and the ecosystem Respiration (Reco ), i.e. the CO2 release from the ecosystem to the atmosphere, which is the sum of heterotrophic (by soil micro-organisms and animals) and autotrophic respiration (by leaves, stems, branches, inflorescences, infructescences, roots and rhizosphere). In the present study, the ecological sign convention adopted was such that NEE = Reco − GPP, with Reco and GPP both defined as positive values. Atmospheric convention was used in this paper with negative fluxes moving downward from the atmosphere to the ecosystem and positive flux moving upward. The approach proposed by Reichstein et al. (2005) was adopted for partitioning NEE into GPP and Reco . The algorithm fits a respiration model to the measured night-time NEE and extrapolates the optimised model to daytime periods using the respective temperature observation of the day. Diurnal GPP was finally calculated as the difference between the diurnal gap filled NEE and the diurnal calculated Reco .

2.3.5. Ecosystem carbon balance Following Chapin et al. (2006), net ecosystem carbon balance (NECB) is the net rate of C accumulation in (or loss from, negative sign) ecosystems. NECB represents the overall ecosystem C balance from all sources and sinks—physical biological, and anthropogenic. In our case we used a simplified approach, by only adding to NEE the C removed by harvesting. As for NEE, negative values of NECB indicate uptake from the atmosphere to the ecosystem (C-sink) and positive values indicate loss from the ecosystem to the atmosphere (C-source). Since in this work we omit anthropogenic part, so then NECB is reduced to NBE, Net Biome Exchange. 2.4. The water use efficiency The WUE of crop production has been the focus of attention in agronomical research since the beginning of the 20th century (Steiner and Hatfield, 2008). Many definitions of WUE have been proposed depending on the time and spatial scales considered and the objectives of the experiment. Traditionally, the standard agronomic WUEagro (Katerji et al., 2008), in kgDM kgH2 O−1 , is defined as the ratio between the crop final harvested production (yield of dry biomass, in kgDM m−2 ) and the cumulated seasonal evapotranspiration (ET in kgH2 O m−2 ) calculated by summing the hourly values when Rg > 20 W m−2 : WUEagro =

yield ET

(5)

From the physiological point of view, the WUE is defined as the ratio between photosynthesis and transpiration (Baldocchi, 1994; Katerji et al., 2008). Actually, direct measurements of carbon and water exchange between canopy and the atmosphere, as previously described, provide the opportunity to examine water use efficiency at the ecosystem level (Beer et al., 2009). So, at field and daily scales a simple WUE (gC kgH2 O−1 ) relation based on GPP is used (Beer et al., 2009; Suyker and Verma, 2010; Tian et al., 2010; Cabral et al., 2013) as WUEGPP =

GPP ET

(6)

where ET is the daily evapotranspiration (kgH2 O d−1 ) calculated as sum of hourly values when Rg > 20 W m−2 . This WUEGPP has been used for studying the adaptability of crops to the environment (e.g. Bhardwaj et al., 2011; VanLoocke et al., 2012). The equivalence of WUEGPP and the ratio between above-ground biomass and ET was underlined by Suyker and Verma (2010) for field crops. Moreover, the link between GPP and vegetation greenness was first demonstrated by Monteith (1972, 1977) through relationships between crop growth and light and PAR (Flanagan et al., 2002). On the other hand, Suyker and Verma (2010) well demonstrated strong linkage between LAI and GPP for maize and soybean crops in eastern Nebraska (USA) under both rainfed and irrigation conditions. Nevertheless, several authors (Baldocchi, 1994; VanLoocke et al., 2012; Zeri et al., 2013; Li et al., 2015) define the WUE (gC kgH2 O−1 ) at ecosystem scale as the ratio between NEE and ET, as WUENEE =

NEE ET

(7)

Even if many authors recommend the use of daytime data for computing WUENEE , the difference between the above two definitions (6) and (7) of water use efficiency at ecosystem level is that WUENEE considers the soil and plant respirations as integral part of the canopy assimilation of CO2 , while WUEGPP considers the perfect equivalence between assimilation and photosynthesis. As the two efficiencies differ for the amount of CO2 exchanged between the system soil-canopy and atmosphere by respiration, in the present work we are presenting and calculating only the WUEGPP after a

300 days 30 October 20days 10 October (3) 2011–2012

27 April y 180 days

3 June 37 days

20 June 17 days

11 July 13 days

30 July 19 days

6 August 7 days

311 days 15 September 14 days 1 September (2) 2010–2011

28 April 195 days

25 May 27 days

12 June 17 days

28 June 16 days

27 July 29 days

9 August 13 days

273 days 6 July 16 days 20 June 5 days 15 June 35 days 15 February 41 days 5 January 48 days (1) 2009–2010

18 November

11 May 85 days

>50% of the heads completely brown–yellow End of flowering

4 August 29 days

18 August 14 days

Total

First flower petals visible

Weather conditions during the study are summarised in Fig. 1 at daily scale from the first day after sowing (DAS) and in Table 1 at seasonal scale. Air daily mean temperature followed the standard pattern of the area, ranging between 0.7 and 31.1 ◦ C (Fig. 1a), mean annual values of the three seasons were greater than the reference value for the period 1984–2011 (16.0 ◦ C). The first was the coldest season (16.7 ◦ C on mean), while the second one was the warmest (17.0 ◦ C). VPD followed the same dynamics for all seasons but had crescent maximum values for the three seasons, being 2.49, 2.96 and 3.24 kPa, on 11 June 2010, 14 July 2011 and 16 July 2012, respectively (Fig. 1b); for this variable the mean annual values for the three seasons were slightly lower than the one for the reference period (0.81 kPa). PAR followed the standard pattern (Fig. 1c), exactly as well as the global solar radiation (data not shown). The total values of precipitation were similar for the first (435 mm) and third season (490 mm) and very high for the second one (690 mm). In the second season, in fact, 190 mm of rain precipitated in only two weeks (a violent storm occurred on 27 October 2010 without causing damages to the plants, because at that time only a few plants started to emerge after the harvesting). However, although no serious damage occurred to sensors during the raining period in October 2010, the EC tower was recovered to permit a detailed maintenance to prevent sensors’ malfunction. The final part of the third season was characterised by a moderate scarcity of water from April to June 2012. The phenological phases of the cardoon crop are reported in Table 2. The dynamics of the first season was quite different from that of the successive seasons, in accordance to the description of the phenological stages by Archontoulis et al. (2010). In our case, the whole growth cycle of the first season (273 days from germination until harvest) is sensibly lower, almost one month, than the duration of the other two seasons (311 and 300 days from the full emergence until harvest for the second and the third season, respectively). This difference is due to the fact that the duration of first cycle is calculated by the germination after sowing (November) and the successive growth cycle durations are calculated from the emergence after harvesting (September). The duration of the initial growth phases from the leaf development until the beginning of “inflorescence emergence and development” is almost the same for the second and the third season (209 and 200 days, respectively), while it is markedly lower for the first season (126 days). The huge amount of water precipitated in October 2010 and the scarcity of rain in the period April–June 2012 (see the above description of weather) did not seem to have strongly influenced the completion of the crop cycle. In fact, the harvest, carried out exactly when the seeds were completely filled, occurred in the same period for all seasons, with a few days ahead of time in the second (9 days with respect the first cycle) and third (3 day with respect to the second cycle) season.

90% of the buds formed

3.1. Weather conditions and crop development

Main inflorescence buds visible

3. Results and discussion

Five leaves visible

discussion about the weight of the respiration in the ecosystem WUE values (see Section 3.3).

Three leaves visible (100% of plants emerged)

435 690 490 535

Only first cycle

0.70 0.76 0.75 0.81

Harvest

16.7 17.0 16.8 16.0

Capitulum and seed ripening

P (mm)

01/11–19/08 20/08/–09/08 10/08–6/08 1984–2009

Flowering and capitulum formation

VPD (kPa)

(1) 2009–2010 (2) 2010–2011 (3) 2011–2012 Reference

Inflorescence emergence and development

Tair (◦ C)

Leaf development

Period

141

Germination

Season

Season

Table 1 Annual mean air temperature (Tair ), vapour pressure deficit (VPD) and annual total precipitation (P) at the experimental site for the three growth seasons, together with the multi-annual values in the period 1984–2009 as reference.

Table 2 Date of the beginning of each phenological phase (first line) for each growth season of cardoon, in the second line the number of days from the previous phase is also indicated, in the latter column the duration of the whole growth season is reported.

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

142

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

DAS 0

100

200

300

400

500

600

700

800

900

1000

Air temperature (°C)

40

(a) 30 20 10 0

(b) VPD (kPa)

3 2 1

PAR (MJ m-2 d -1 )

0

(c) 8

4

Rain (mm d -1)

0

(d) 40

20

0 Nov-09 Jan-10 Apr-10 Jul-10 Oct-10 Jan-11 Apr-11 Jul-11 Oct-11 Jan-12 Apr-12 Jul-12

Date (mmm-yy) Fig. 1. Daily air temperature (a), vapour pressure deficit, VPD (b), photosynthetically active radiation, PAR (c) and rain (d) during the three growing seasons. DAS is days after sowing.

The measured green biomass (GB) and LAI are reported in Fig. 2 (panel (a) and (b), respectively). From this figure it is clear that in the first season, the plants emerged and started to growth in January 2010 (see also the phenological phases in Table 2). The growth in this case is very slow and reached a maximum of about 400 g plant−1 on early March 2010, afterwards it started to slowly decline. In the second season the GB started to grow at the end of September 2010, about one month and half after the harvest, then the crop grew rapidly and in one month, from early November to early December 2010, reached the maximum value of about 1700 g plant−1 ; afterwards it rapidly declined until January 2011, when it begun to newly grow very slowly until middle May 2011. The growth during the third season was very similar to the second one, with the first maximum value reached on middle January 2012 and the second maximum on middle June 2012. The maximum height of the plants was 120 ± 19, 175 ± 38 and 172 ± 39 cm in the first, second and third crop cycle, respectively and was reached at the maximum crop development. The LAI followed in some extent the same pattern of the GB, low values in the first growth season, high values and similar pattern in the successive seasons. In the second and third season the maximum value was reached on middle May

2011 and middle June 2012, respectively, before the senescence phase. The crop covered completely the soil from early April, early December and end December for the first, second and third season, respectively; thus the evaporation from the soil was negligible in the most part of the experimental period. The soil water content in m3 m−3 is almost the same for the three seasons (Fig. 2c, the values are at monthly scale), either in terms of dynamics or in terms of absolute value for both measurement depths (20 and 40 cm). The mean value of the soil water content at 40 cm was similar for the three seasons, being 0.35, 0.31 and 0.33 m3 m−3 , for the first, second and third season, respectively. The pattern and the values of the soil water content in the experimental period show that there was no strong impact due to the huge precipitations in October 2010 and the scarcity of rain in the last part of the experiment (see above weather description). This can be explained by the particular structure of the soil in the experimental plot, which is shallow with a cracked limestone subsoil, which permits a fast drainage of the water in excess (i.e. Katerji et al., 2010; Rana and Katerji, 2000), maintaining in the same time good water reservoir for the absence of soil evaporation during most part of the growth cycles.

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

143

DAS 100

150

200

250

300

350

400

450

500

550

600

650

700

750

800

850

900

950

1000

2000 1500

a)

1000 500 0

Water content (m3 m-3)

4

b)

2

0 40 cm 20 cm

c) 0.4

0.3

0.2

3000

GDD (°C)

50

2500

LAI (m2 m-2)

Green biomass (g)

0

d)

2000 1000 0 Nov-09

Jan-10

Apr-10

Jul-10

Oct-10

Jan-11

Apr-11

Jul-11

Oct-11

Jan-12

Apr-12

Jul-12

Date (mmm-yy) Fig. 2. Patterns of green biomass, in g per plant, with standard deviation (panel (a)), LAI with standard deviation (panel (b)), soil water content at two depths in the soil (20 and 40 cm, panel (c)) and growing degree days (GDD, panel (d)) during the three cardoon crop seasons. DAS is days after sowing.

In order to partially cleanse the cardoon development dynamics from the weather and the soil water content, the relationships among crop growth and growing degree days (GDD) are here analysed. GDD is defined as the number of temperature degrees above a certain threshold base temperature, which varies among crop species. The threshold value is fixed for each crop species and is equal to 7 ◦ C for the cardoon (Bidini et al., 2007). The pattern of the GDD is shown in panel d) of Fig. 2 and completes the description of the cardoon development. In fact, from this figure it is clear that, due to the decreasing of the air temperature in the autumn and winter months, deducible from the gentle increase of GDD from November to April of the second and third season, the crop suffered a decreasing of the development with a general dormancy of physiological activities. The development of the crop in the first season, above described, did not show the same dormancy (Archontoulis et al., 2010) because the plants emerged in the middle winter and grew slowly and monotonically up to middle spring. Furthermore, the value of the cumulated GDD in the second season until harvest (3031 ◦ C) is quite similar the value of GDD in the third season (3206 ◦ C), with similar dynamics, being much lower in the first season from seeding until harvest (2129 ◦ C). In conclusion, the first cardoon growth season is very different from the other ones, being shorter and with specific dynamics of development, in terms of both GB and LAI. This pattern of GB and LAI, together with the duration of each phenological phase, can be explained by the specific functioning of a perennial crop, which spends one year, the first one after sowing, for its establishment, by investing mostly in developing the storage organs of root system, rather than in the development of above ground biomass (i.e.

Gominho et al., 2011; Anderson-Teixeira et al., 2009; Dohleman et al., 2012; Kahle et al., 2001; Neukirchen et al., 1999). The same behaviour was also reported by Angelini et al. (2009) in a trial with two cardoon cultivars. Raccuia and Melilli (2010), for the same crop cultivated in two successive seasons, reported that by the end of the first year, root biomass was 35% of the total biomass and 338 days after planting the re-sprouting roots contributed 70% of the total biomass. Thus the strong development of the root system supports the development of the green biomass in the seasons following the first one, when the cardoon crop showed similar physiological functioning, both in terms of phenological phases, green biomass and LAI development. 3.2. CO2 and H2 O fluxes In general, the diurnal patterns of NEE are characterised by daytime CO2 uptake and nighttime CO2 release as the effect of interaction between meteorological dynamics and ecophysiological processes (i.e. Steduto et al., 1997; Shurpali et al., 2013). In the present case, NEE (gC m−2 d−1 ) pattern in function of the DAS is quite similar for the three growth seasons, but with substantial differences (Fig. 3, panel a). Just after sowing in the first season (November 2009–January 2010) and during the periods between harvest and emergence (August–October 2010 and 2011) the NEE is almost zero or slightly positive, because the soil is almost bare, thus the soil–canopy–air system is a source of CO2 towards atmosphere. After emergence, during the development period, the crop uptakes CO2 for the growth, therefore NEE become negative by decreasing its values, as well as the crop increases its biomass, by reaching

144

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

DAS -2 -1 -2 -1 -2 -1 ET (kgH2O m-2 d-1) Reco (gC m d ) GPP (gC m d ) NEE (gC m d )

0

50

100

150

200

250

300

350

400

450

500

550

600

650

700

750

800

850

900

950

1000

8 4

a)

0 -4 -8

-12 16 12

b)

8 4 0

16

c)

12 8 4 0 6

d)

4 2 0 Nov-09

Jan-10

Apr-10

Jul-10

Oct-10

Jan-11

Apr-11

Jul-11

Oct-11

Jan-12

Apr-12

Jul-12

Date (mmm-yy) Fig. 3. Patterns of net ecosystem exchange (NEE, panel (a)), gross primary production (GPP, panel (b)), ecosystem respiration (Reco panel (c)) and actual crop evapotranspiration (ET, panel (d)), during the experiment. DAS is days after sowing.

the minimum value during the flowering stage. From the flowering stage the NEE values increase as well as the leaf area decrease up to the beginning of senescence stage, when the plants start to dry the green components (see Fig. 2); the dates of this inversion (from negative to positive NEE, i.e. from crop CO2 uptake to release into atmosphere) are 4th July, 7th and 18th June for the three seasons, respectively. So, the period of CO2 cardoon uptake was of 141, 218 and 243 days for the first, second and third season, respectively. The duration of the CO2 uptake periods are clearly linked to the duration of the growth cycles which was shortest in the first season. The GPP (gC m−2 d−1 ) pattern is shown in Fig. 3, panel (b). During the first season, GPP is close to 0–1.5 gC m−2 d−1 during the first three months after sowing (the values in the first days after sowing are slightly greater than 0 since they were simulated and not directly measured), because the crop emerged slowly as already seen in the description of the crop development dynamics. It started to grow around the DAS 135, corresponding to middle of March 2010, when it rapidly increased up to reach its maximum around the end of May 2010 (the maximum GPP value was recorded on 31st, 211 DAS). After that it decreased when the plants stopped the growth with the formation of the floral heads and the ripening of the seeds up to harvest. After several days, the cardoon plants restarted to emerge, GPP increased in a few days, remaining stable from October 2010 until end January 2011 around 3 gC m−2 d−1 . Starting from February 2011, the crop rapidly increased the GPP and the production of biomass. In this case the maximum GPP was reached on mid-May 2011 (the maximum value was recorded on 15th, 560 DAS), then GPP started to decrease, following the same dynamics as in the previous season up to the harvest. Once again, the crop remained quiescent for a couple of weeks and restarted to rapidly increase its GPP on the end of September 2011, when the plants re-emerged. In two months the GPP value reached

7 gC m−2 d−1 on 7th November 2011 (736 DAS). From now GPP remained almost stable, or slightly decreasing, around the value of 6 gC m−2 d−1 up to the end of February 2012 when, as for the previous season, it rapidly increased. In the third season the cardoon crop reached its maximum GPP (19 gC m−2 d−1 ) at the beginning of May (the maximum values was recorded on 1st May 2012, 912 DAS). The maximum value of GPP was reached two weeks before the date of the second season which, in its turn, was reached two weeks before the date of the first season. Moreover, during the second and the third season, the GPP pattern presented a sort of step and following gentle decreasing of almost four months (October–January) around a value which increased with the number of growth seasons after the first one. This pattern of GPP values along the three analysed seasons of cardoon crop can be explainable both with the weather pattern and the plant physiology of this perennial crop. Of course, the weather played a fundamental role, particularly the air temperature and radiation, which permitted the photosynthesis by the crop which has to be well developed for improving its CO2 assimilation. Further, after the first year of cultivation, in the successive growing seasons the increasing of the whole biomass of cardoon (hypogea and aboveground) was permitted by an increasing of CO2 assimilation (Ierna and Mauromicale, 2010; Ierna et al., 2012). The Reco values during the whole experimental period is presented always in Fig. 3, panel c and, of course, reflect the previous GPP pattern: the maximum positive values are reached in the same dates as GPP, showing an increasing values along the three development seasons. The respiration trend presented a step in the second season and another step, greater than the first, in the third season, in the period from harvest dates to the beginning of March for both seasons. This huge amount of released CO2 towards the atmosphere can be attributed to the effect of leaving the crop residues

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

in the field after cutting which might have led to higher microbial respiration rate, resulting in higher Reco (Verburg et al., 2004). During the 1st growing season the Reco increased slowly up to end of May, then decreased up to harvest date. During the middle of June, there was a peak more likely related to leaves decomposition in the field. During the second season, after cutting and when the 1st autumn rainfalls occurred, the respiration increase up to the middle of November, afterwards it decreased very gently up to the end of February. From the beginning of March the respiration started to increase quite rapidly and following a monotonic trend until its maximum at the end of May. The same trend was shown in the third season, with the same time milestones, but with different values, being greater in absolute values in the last case. The rapid increase and the gentle decrease of Reco component reflected more likely the air temperature dynamics and the fact that most part of fresh biomass was available to the decomposition plus respiration of standing biomass in the 1st days after harvest (Suyker et al., 2004; Verburg et al., 2004; Reichstein et al., 2005). In some extent the pattern of CO2 assimilation (GPP) during the three seasons of cardoon cultivation was followed by the crop actual evapotranspiration (Fig. 3, panel d). In fact, ET (in kgH2 O d−1 ) showed the same step plus plateau periods during the winters of the second and third growth seasons; the low but significant values of ET during the initial part of the first season is essentially due to the evaporation after the first water supply by precipitation, which allowed a suitable seed germination and plant installation. 3.3. The weight of Reco in the evaluation of NEE GPP differs from NEE for the value of the respiration term, thus they are differently related to crop ET (Kuglitsch et al., 2008). The daily value of any flux (NEE, GPP, Reco , ET) depends on the crop physiology, water conditions and weather (Baldocchi, 1994; Grelle et al., 1997), mainly temperature, radiation and rain. On the other hand, the respiration may obscure the coupling of canopy carbon and water fluxes, therefore, GPP should be preferred to NEE for calculating WUE at ecosystem level (Kuglitsch et al., 2008). A way to specify the relation among GPP, NEE and Reco is to evaluate what is the weight of Reco term on NEE value along the crop growth season (Jans et al., 2010; Cabral et al., 2013). GPP, Reco and their ratios (GPP/Reco ) are presented in Fig. 4 at monthly scale, according to Falge et al. (2002). During October 2010 both GPP and Reco were estimated, so the uncertainty for each term is very high (Falge et al., 2001), hence we preferred to skip this value. Reco values are greater than or equal to GPP during the establishment phase of the crop in the first cycle (December 2009–January 2010), indicating a predominance of the heterotrophic respiration of the carbon fluxes due to the reduced activity of the plants. Further, during the senescence phase of each cycle (about July and August of 2010, 2011 and 2012), during the restarting phase of the second cycle (August 2010) and of the third cycle (September–October 2011) Reco exceeds the GPP value, indicating the lowering of photosynthetic activity. While, during the crop development phase the assimilation exceeded the respiration during all three cycles. However, the GPP/Reco ratios were close to, or above to two for most of the development phase. The ratio between GPP and Reco greater than two implies that NEE is almost equal to Reco during the most efficient assimilation period (crop development), thus indicating that the carbon fluxes are mainly driven by the plants autotrophic respiration, as already observed for other agricultural crops (Falge et al., 2002; Cabral et al., 2013). Summarizing, also from Fig. 4 can be argued that our perennial ecosystem followed the same seasonal trend for all growth seasons which concerns the phenology of CO2 uptake and release. In particular, it presents a maximum accumulated C uptake at the end development phase followed by a decrease due to the dominance

145

of soil respiration over photosynthesis after plant senescence, emissions from bare soil following harvest or decomposition of aboveground biomass remaining in the field (Zeri et al., 2013). Furthermore, GPP was much more greater than Reco during the development phase, only in May during the first season and from March to May for the other two seasons, perfectly on line with the pattern of the green biomass and LAI (see Fig. 2). This indicates a greater photosynthetic activity in the seasons successive to the first one, due to the larger aboveground biomass supported by well developed roots occurred in the first season. 3.4. Environmental control of CO2 assimilation and evapotranspiration In order to analyze the relationships among crop growth and climatic drivers, the GPP and ET were related to VPD and absorbed PAR at hourly scale (Cabral et al., 2013). The results are depicted in Fig. 5. The presented data are relative to a short period of three clear successive days close to the maximum physiological activity, corresponding to the maximum values of GPP (in particular the presented data are relative to the days 25–27 May 2010, 9–11 May 2011 and 28–30 April 2012, for the three growth seasons, respectively). During these periods of three days the cardoon was in the same phenological phase (“main inflorescence buds visible”, see Table 2), having, however, different values of LAI (1.1, 4.0 and 3.8, for the first, second and third season, respectively) and soil water content (0.32, 0.30 and 0.40 m3 m−3 ). In the upper panels of Fig. 5 the GPP normalised to the maximum value, GPPmax (Tallec et al., 2013) and the ET were related to the VPD at hourly scale. For both variables, logarithmic functions well described these relationships, showing asymptotic values from a threshold around 0.4 kPa of VPD. The bottom panels of Fig. 5 show the relationships among GPP and ET and absorbed PAR (APAR). Also in this case the relationship is clearly logarithmic for GPP, showing a saturation value of APAR around 1 MJ m−2 s−1 . The relation between ET and APAR is monotonically increasing, slightly logarithmic without any clear saturation threshold. From these figures it is clear that the cardoon crop has similar physiological response to climatic drivers VPD and APAR in the three seasons. Furthermore, these responses are mathematically similar for both GPP and ET in the three seasons during the same growth stage, without any reference to the values of LAI and soil water content. 3.5. Water use efficiencies and net ecosystem carbon balance Since GPP and ET showed similar relationships with respect to the drivers of climate (VPD and APAR, see previous section), it is worth to investigate about the water use efficiency at instantaneous (hourly) scale, in order to explore the water and assimilation behaviour of the cardoon crop in the three successive seasons. According to Kuglitsch et al. (2008) we related the hourly GPP values (in gC h−1 ) to the hourly evapotranspiration ET (kgH2 O h−1 ) during sample days in the development growth phase. These results are shown in Fig. 6, and are relative to the days when the GPP reached its maximum for the three cardoon seasons (31st May 2010, 15th May 2011 and 1st May 2012, respectively). This figure reported hourly diurnal values of GPP vs ET, when the global radiation was positive. From the figure it is clear that a strong linear relationship exists between instantaneous GPP and ET for all the seasons and that the instantaneous water use efficiency (the slopes of the linear regressions forced intercept to 0) increases from the first to the third cardoon crop growth season, being 2.6, 3.2 and 4.8 gC kgH2 O−1 , indicating an improvement in water use. The WUEGPP pattern at daily scale for the whole experimental period is shown in Fig. 7. In general the values increased with the

146

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

5

4

GPP/Reco

300 3 200 2 100 1

0

GPP Reco (gC m-2 month-1)

400 GPP/Reco Reco GPP

0 Mar-10 Jun-10 Sep-10 Dec-10 Mar-11 Jun-11 Sep-11 Dec-11 Mar-12 Jun-12 Sep-12

Month Fig. 4. Monthly total of gross primary production (GPP) and ecosystem respiration (Reco) , together with the ratio GPP/Reco observed during the three cycles of cardoon crop.

season of cultivation, being in mean 1.8 for the first year, 2.6 for the second one and 3.3 gC kgH2 O−1 for the third one. The WUEGPP pattern showed a high variability in the growth stage for all seasons, the highest values of WUE corresponded to clear and sunny days after rainy days. WUEGPP is very low during the establishment phase in the first cultivation cycle, showing the slight development of biomass during the germination phase (Archontoulis et al., 2010). The same for the successive seasons: WUEGPP is low during sprouting phase after harvesting in the period August–October before sharp increasing during the winter months of second and third seasons, as already found for GPP and ET patterns (see Fig. 3). Apart the above mentioned variability, during the second and third season any particular abnormal value of WUEGPP is clearly visible as a

consequence of the huge precipitation (autumn 2010) or the scarcity of precipitation (summer 2012). The yield, the total gross primary production and the total evapotranspiration are reported in Table 3 for the three seasons of our cardoon crop cultivation. From this data it is evident that the yield in the second season is sensibly higher than that in first year (+85%), while it is quite stable in the third growth season (+5%). The same trend of cardoon yield was found by Raccuia and Melilli (2007, 2010), Gominho et al. (2011) and Ierna et al. (2012). So, the first year of cultivation saw the plants establishing in the environment, while from the first year to the second year we can observe a very rapid yield increase in aboveground biomass, later, the production level was stabilised (Raccuia and Melilli, 2007).

0.1

0.1

GPP/GPPmax=0.11 ln(VPD) + 0.75 r2=0.41

ET=0.08 ln(VPD) + 0.39 r2=0.31 0.01

First season LAI=1.1 SWC=0.32 m3 m-3 Second season LAI=4.0 SWC=0.30 m3 m-3 Third season LAI=3.8 SWC=0.40 m3 m-3 0.01

1 0

0.4

0.8

1.2

1.6

2

0

0.4

0.8

1.2

1.6

2

VPD (kPa) Hourly GPP

hourly ET (mm h-1)

1

0.8

40

0.6

0.4 20 0.2

GPP=8.74 ln(APAR) +32.2 r2=0.80

hourly ET (mm h-1)

Hourly GPP/GPPmax

1

ET=0.14 ln(APAR) + 0.44 r2=0.62 0

0 0

0.4

0.8

1.2

1.6

0

0.4

0.8

1.2

1.6

APAR (MJ m-2 s-1) Fig. 5. Gross primary production (GPP, left panels) and evapotranspiration (ET, right panels) vs vapor pressure deficit (VPD, upper panels) and absorbed photosynthetically active radiation (APAR, bottom panels) at hourly scale, during the days of maximum physiological activity.

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

147

Table 3 Gross primary production (GPP), actual evapotranspiration (ET), yield, water use efficiencies at ecosystem level (WUEGPP ) and agronomic level (WUEagro ), at seasonal scale. Season

GPP (gC m−2 )

Yield (kgDM m−2 )

ET (kgH2 O m−2 )

WUEGPP (gC kgH2 O−1 )

WUEagro (gDM kgH2 O−1 )

1 2 3

955 1758 1798

0.9 1.7 1.8

517 677 605

1.9 2.6 3.0

1.74 2.51 2.97

GPP=4.76 ET r =0.55

3

GPP=3.15 ET r =0.87

First season Second season Third season

GPP (gC m-2 h-1)

2

1

GPP=2.63 ET r =0.90

0 0

0.2

0.4

0.6

0.8

ET (kgH2O m-2 h-1) Fig. 6. Relationship between gross primary production (GPP) and evapotranspiration (ET), at hourly scale during the day of maximum CO2 assimilation in the three growing seasons.

GPP showed clear increase from the first to the third season in our Mediterranean environment, while ET increased from the first to the second season and decreased from the second to the third season. The first large total ET increasing (+23%) is due to the huge increasing of the green biomass (see Fig. 2) in the second growth season, moreover the duration of the growth cycle was lower in the first season (273 days) than the second one (311). The decreasing of total ET from the second to the third season is quite little (−11%, from 677 to 605 kgH2 O m−2 ) and did not depend on the soil water availability (see Fig. 2) and the weather (see Fig. 1). As a consequence of the above values, the WUEGPP at ecosystem level, calculated as the ratio between the sum of daily GPP and the sum of daily ET increased from the first to the third season (see Table 3). In

the same Table 3 also the standard agronomic water use efficiency at seasonal scale is reported, showing exactly the same trend from the first to the third season as for the other considered water use efficiencies. In this case, the agronomic WUEagro showed a great increase from the first to the second growth season (+44%) and a sensibly lower increase in the third season (+18%). Also this result is linked to the crop functioning. In fact, as previously mentioned, the cardoon crop spent the first year of growth for developing the storage organ in the root system, by, consequently, reducing the above-ground biomass development. In the following, the aboveground biomass production greatly increased from the first to the second year and became quite stable in the third year (Ierna and Mauromicale, 2010). All terms of the net ecosystem balance, NBE, in gC m−2 season−1 are reported in Fig. 8, for the three cardoon successive growth seasons analysed in this work. The cumulated annual NEE values was the lowest in absolute value (270 gC m−2 season−1 ) during the first growing season (plant stage), the highest during the second one (660 gC m−2 season−1 ). The cumulated NEE during the third growing season (390 gC m−2 season−1 ) showed a decrease respect to the previous year. This reduction was due to increased respiration processes (Reco ) influenced by climate dynamics (Flanagan et al., 2002; Ammann et al., 2007; Jans et al., 2010; Shurpali et al., 2013) and decomposition of crop residues left in the field after cutting, even if other causes could be related to the physiology of the perennial crop under study (Angelini et al., 2009) and to the duration of each phenological stage (Raccuia and Melilli, 2007). In fact, the huge increase of the respiration of the second season with respect to the first one (+60%) can be mainly attributed to the autotrophic respiration, supported by the huge development of the root from the first to the second growth season (see details above). Concerning the measured carbon content in aboveground harvested dry biomass (272, 501, 573 gC m−2 season−1 for the three successive growing seasons, respectively), increasing variations of +84% and +1.4% respect to previous year value were registered. Similar dynamics were found in GPP and can be linked to the adaptability of the crop to the Mediterranean environments, in agreement with the fact that cultivated cardoon needs one year for its establishment and two years before a stabilisation of the

DAS 0

50

100

150

200

250

300

350

400

450

500

550

600

650

700

750

800

850

900

950

1000

14

WUEGPP (g Kg-1)

12 10 8 6 4 2 0 Nov-09

Jan-10

Apr-10

Jul-10

Oct-10

Jan-11

Apr-11

Jul-11

Oct-11

Date (mmm-yy) Fig. 7. Daily water use efficiency for GPP, WUEGPP . DAS is days after sowing.

Jan-12

Apr-12

Jul-12

148

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150 Table 4 Mean water use efficiencies at ecosystem level (WUEGPP ) for other biomass crops in different locations. Crop

Site

WUEGPP (gC kgH2 O−1 )

Reference

Coniferous

Nebraska, USA

3.3

Maize

Nebraska (USA)

3.2

soyabean

Nebraska (USA)

2.1

Sugarcane

Brazil

4.2

Switchgrass

3.3–4.1

Switchgrass

Pennsylvania, USA Oklahoma, USA

Reed canary grass

Finland

2.5

Cynara cardunculus

South Italy

2.5 (1.9–3.0)

Eggemeyer et al. (2006) Suyker and Verma (2010) Suyker and Verma (2010) Cabral et al. (2013) Skinner and Adler (2010) Wagle and Kakani (2014) Shurpali et al. (2013) Present work

3.2

with the environment, showing increasing ability to use water and carbon dioxide. 4. Conclusions

Fig. 8. Carbon balance of the cardoon crops, at seasonal scale, during the three studied growth seasons.

biomass amount (Angelini et al., 2009; Raccuia and Melilli, 2010). Finally, NBE, which represents the net rate of C accumulation in (or loss from) ecosystems (Chapin et al., 2006) was C-neutral for the first year, where NEE and C in the harvested biomass were approximately equal. Then this cardoon ecosystem was C-sink during the second season and C-source during the third season. By considering the cumulated values of NBE in all the observed seasons, this cardoon crop under Mediterranean environment was C-source (+29 gC m−2 ). Comparing our results with other works, focusing attention on bioenergy crops, we found that Shurpali et al. (2009), for another perennial bioenergy crop (reed canary grass) reported that it was C-sink during the first year and C-source during the successive three years. Zeri et al. (2011) reported that miscanthus was a C-sink during the first 2.5 years of establishment, but the lower amount of harvested biomass was determinant to that balance. Cabral et al. (2013) for a sugarcane crop cultivated over two years in Brazil found that it was C-neutral in the first growth cycle and C-source in the second year. Skinner and Adler (2010) reported that their switchgrass field in Pennsylvania (USA) was a net C-sink for the first three years, but became a source the final year. Finally, Wagle and Kakani (2014) found that a switchgrass crop grown in Oklahoma (USA) was C-source in both years of experiment. Therefore, from the carbon balance point of view, to state that a bioenergy crop can be considered as a sink of carbon, it is necessary to study the ecosystem over a long period, surely greater than two years (Zeri et al., 2011). A comparison of our results with other values published in peer reviewed papers (see Table 4 for details), concerning WUE as ratio between total GPP and total evapotranspiration at seasonal scale, showed that cardoon under rainfed conditions has values of WUEGPP greater than those of soybean in Nebraska (USA) and comparable to those of other biomass crops like reed canary grass in Finland and switchgrass in Oklahoma. Sugarcane in Brazil has the best WUEGPP , while maize in Nebraska (USA), switchgrass in Pennsylvania (USA) and Oklahoma (USA) show WUEGPP values slightly greater than the cardoon one. In conclusion, the cardoon seems to be a crop which, under rainfed Mediterranean conditions, needs a couple of growth cycles to establish the best relation and settlement

In this work the behaviour of a bioenergy crop (cardoon, C. cardunculus L.) was analysed with respect to the use of natural resources (water and CO2 ) in Mediterranean region (Apulia, southern Italy) under rainfed conditions, for three successive growth seasons for a total of more than one thousand days. The weather of the second season was characterised by huge precipitation in Autumn and the third season presented a scarcity of precipitation during spring. However, the soil water content showed similar values during the whole experimental period, both as trend and in absolute values. The air temperature, VPD and PAR showed quite similar patterns in the three seasons. This perennial crop had the first growth cycle shorter than the other two successive cycles. The duration of the phenological stages is different for the first season with respect to the other two successive seasons, indicating that the crop needed almost one year (the first one) as period of establishment to our semi-arid Mediterranean environment. However, in any season, the CO2 assimilation in terms of GPP, and the water consuming in terms of actual ET, showed similar relationships with the climate (VPD and APAR), at instantaneous hourly scale, at least when the crop is at the maximum development stage corresponding to the maximum photosynthetic activity (maximum GPP), also with different LAI and soil water content values. Following the analysis at monthly scale about the weight of the soil–crop respiration on the net ecosystem exchange, this work confirmed and stressed that a study on the crop water use at ecosystem level should be realised by investigating the ratio between GPP and ET. This cardoon crop also showed increasing water use efficiency along the three seasons of experiment, in terms of WUEGPP at hourly, daily and seasonal time scales, by indicating an improved adaptation of the crop to the environment with respect to the water use-assimilation coupling. According to Kuglitsch et al. (2008), highest WUE when yield is highest indicates that these high carbon uptake rates are not only caused by stomatal regulation, affecting both carbon and water fluxes, but also increases carboxylation efficiency. Both total GPP and yield increased from the first to the third season, while the total evapotranspiration increased from the first to the second season and then decreased from the second to the third season. Therefore, the WUE increased both at ecosystem and agronomic level from the first to the third year of cultivation. The

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

cardoon water use in our semi-arid environments is quite efficient if compared with other biomass crops in other environments. Anyway, agronomical features (i.e. pests, diseases, weeds) could be limiting factors to its development in the field and need to be further investigated. From the carbon balance point of view, this cardoon perennial crop cultivated for three years under semi-arid climate and with low inputs, can be considered a C-source with respect to the atmosphere. Therefore, precautions should be taken into consideration, from the environmental point of view, when this crop is cultivated to supply “green” energy.

Acknowledgments The research was carried out under the National Research Project ‘Optimisation of existing bioenergy chains for economic and environmental sustainability’ (BIOSEA). Funding: Ministry of Agriculture, Food and Forestry Policies (MiPAAF), Italy (D.M. 16916/7303/10, 23 July 2010). The soil water content data were kindly supplied by Dr. Pasquale Campi (CRA-SCA, Bari, Italy). The authors thank Nicola Martinelli for management of experimental devices and Vito Casulli, Vincenzo Cesareo, Michele Introna and Nicola Sanitate for the field work.

References Angelini, L.G., Ceccarini, L., Nassi o Di Nasso, N., Bonari, E., 2009. Long-term evaluation of biomass production and quality of two cardoon (Cynara cardunculus L.) cultivars for energy use. Biomass Bioenergy 33, 810–816. Anderson-Teixeira, K.J., Davis, S.C., Masters, M.D., Delucia, E.H., 2009. Changes in soil organic carbon under biofuel crops. GCB Bioenergy 1, 75–96. Ammann, C., Flechard, C.R., Leifeld, J., Neftel, A., Fuhrer, J., 2007. The carbon budget of newly established temperate grassland depends on management intensity. Agric. Ecosyst. Environ. 121, 5–20. Archontoulis, S.V., Struik, P.C., Vos, J., Danalatos, N.G., 2010. Phenological growth stages of Cynara cardunculus: codification and description according to the BBCH scale. Ann. Appl. Biol. 156 (2), 253–270. Archontoulis, S.V., Vosa, J., Yina, X., Bastiaans, L., Danalatos, N.G., Struik, P.C., 2011. Temporal dynamics of light and nitrogen vertical distributions in canopies of sunflower, kenaf and cynara. Field Crops Res. 122, 186–198. Asseng, S., Hsiao, T.C., 2000. Canopy CO2 assimilation, energy balance, and water use efficiency of an alfalfa crop before and after cutting. Field Crops Res. 67, 191–200. Aubinet, M., Grelle, A., Ibrom, A., Rannik, Ü., Moncrieff, J., Foken, T., Kowalski, A.S., Martin, P.H., Berbigier, P., Bernhofer, C., Clement, R., Elbers, J., Granier, A., Grunwald, T., Morgenstern, K., Pilegaard, K., Rebmann, C., Snijders, W., Valentini, R., Vesala, T., 2000. Estimates of the annual net carbon and water exchange of forests: the EUROFLUX methodology. Adv. Ecol. Res. 30, 113–175. Baldocchi, D., 1994. A comparative study of mass and energy exchange rates over a closed C3 (wheat) and an open C4 (corn) crop: II. CO2 exchange and water use efficiency. Agric. For. Meteorol. 67, 291–321. Beer, C., Ciais, P., Reichstein, M., Baldocchi, D., Law, B.E., Papale, D., Soussana, J.F., Ammann, C., Buchmann, N., Frank, D., Gianelle, D., Janssens, I.A., Knohl, A., Köstner, B., Moors, E., Roupsard, O., Verbeeck, H., Vesala, T., Williams, C.A., Wohlfahrt, G., 2009. Temporal and among-site variability of inherent water use efficiency at the ecosystem level. Global Biogeochem. Cycles 23, http://dx. doi.org/10.1029/2008GB003233. Bhardwaj, A.K., Hardwaj, J., Zenone, T., Jasrotia, P., Robertson, G.P., Chen, J., Hamilton, S.K., 2011. Water and energy footprints of bioenergy crop production on marginal lands. Global Change Biol. Bioenergy 3, 208–222. Bidini, G., Fantozzi, F., Buratti, C., Bartocci, P., 2007. Most suitable areas for the cultivation of the herbaceous energy crop in Umbria region (Italy) and biomass production evaluation. In: Proceeding of 15th European Biomass Conference and Exhibition, 7–11 May, 2007, Berlin, Germany, pp. 392–400. Cabral, O.M.R., Rocha, H.R., Gash, J.H., Ligo, M.A.V., Ramos, N.P., Packer, A.P., Batista, E.R., 2013. Fluxes of CO2 above a sugarcane plantation in Brazil. Agric. For. Meteorol. 182–183, 54–66. Campi, P., Palumbo, A.D., Mastrorilli, M., 2009. Effects of tree windbreak on microclimate and wheat productivity in a Mediterranean environment. Eur. J. Agron. 30, 220–227. Chapin, F., Woodwell, G., Randerson, J., Rastetter, E., Lovett, G., Baldocchi, D., Clark, D., Harmon, M., Schimel, D., Valentini, R., Wirth, C., Aber, J., Cole, J., Goulden, M., Harden, J., Heimann, M., Howarth, R., Matson, P., McGuire, A., Melillo, J., Mooney, H., Ne, J.R. H., Pace, M., Ryan, M., Running, S., Sala, O., Schlesinger, W., Schulze, E., 2006. Reconciling carbon-cycle concepts, terminology, and methods. Ecosystems 9, 1041–1050.

149

Christaki, E., Bonos, E., Florou-Paneri, P., 2012. Nutritional and functional properties of Cynara Crops (Globe Artichoke and Cardoon) and their potential applications: a review. Int. J. Appl. Sci. Technol. 2 (2), 64–70. Curt, M.D., Sánchez, G., Fernández, J., 2002. The potential of Cynara cardunculus L. for seed oil production in a perennial cultivation system. Biomass Bioenergy 23, 33–46. De Benedetto, D., Castrignano, A., Sollitto, D., Modugno, F., Buttafuoco, G., lo Papa, G., 2012. Integrating geophysical and geostatistical techniques to map the spatial variation of clay. Geoderma 171–172, 53–63. Dohleman, F.G., Heaton, E.A., Arundale, R.A., Long, S.P., 2012. Seasonal dynamics of above- and below-ground biomass and nitrogen partitioning in miscanthus giganteus and Panicum virgatum across three growing seasons. GCB Bioenergy, http://dx.doi.org/10.1111/j.1757-1707.2011.01153.x. Eggemeyer, K.D., Awada, T., Wedin, D.A., Harvey, F.E., Zhou, X., 2006. Ecophysiology of two native invasive woody species and two dominant warm-season grasses in the semiarid grasslands of the Nebraska Sandhills. Int. J. Plant Sci. 167, 991–999. Espadafor, M., Lorite, I.J., Gavilàn, P., Berengena, J., 2011. An analysis of the tendency of reference evapotranspiration estimates and other climate variables during the last 45 years in Southern Spain. Agric. Water Manage. 98 (6), 1045–1061. Falge, E., Baldocchi, D., Tenhunen, J., Aubinet, M., Bakwin, P., Berbigier, P., Bernhofer, C., Burba, G., Clement, R., Davis, K.J., Elbers, J.A., Goldstein, A.H., Grelle, A., Granier, A., Guomundsson, J., Hollinger, D., Kowalski, A.S., Katul, G., Law, B.E., Malhi, Y., Meyers, T., Monson, R.K., Munger, J.W., Oechel, W., Paw, U., Pilegaard, K.T., Rannik, Ü., Rebmann, C., Suyker, A., Valentini, R., Wilson, K., Wofsy, S., 2002. Seasonality of ecosystem respiration and gross primary production as derived from FLUXNET measurements. Agric. For. Meteorol. 113, 53–74. Falge, E., Baldocchi, D., Olson, R., Anthoni, P., Aubinet, M., Bernhofer, C., Burba, G., Ceulemans, R., Clement, R., Dolman, H., et al., 2001. Gap-filling strategies for defensible annual sums of net ecosystem exchange. Agric. For. Meteorol. 107 (1), 43–69. Fernández, J., Hidalgo, M., Del Monte, J.P., Curt, M.D., 2005. Cynara cardunculus L. as a perennial crop for non irrigated lands: yields and applications. Acta Hortic. 681, 109–115. Fernández, J., Curt, M.D., Aguado, P.L., 2006. Industrial application of Cynara cardunculus L. for energy and other uses. Ind. Crops Prod. 24, 222–229. Fischer, T., Tubiello, F.N., von Velthuisen, A., Wiberg, D.A., 2007. Climate change impacts on irrigation water requirements: effects of mitigation 1990–2080. Technol. Forcast. Soc. Change 74, 1083–1107. Flanagan, L., Wever, L., Carlson, P.J., 2002. Seasonal and interannual variation in carbon dioxide exchange and carbon balance in a northern temperate grassland. Global Change Biol. 8, 599–615. Foken, T., Wichura, B., 1996. Tools for quality assessment of surface-based flux measurements. Agric. For. Meteorol. 78, 83–105. Foti, S., Mauromicale, G., Raccuia, S.A., Fallico, B., Fanella, F., Maccarone, E., 1999. Possible alternative utilization of Cynara spp, Biomass, grain yield and chemical composition of grain. Ind. Crops Prod. 10, 219–228. Gherbin, P., Monteleone, M., Tarantino, E., 2001. Five year evaluation on cardoon (Cynara cardunculus L. var. altilis) biomass Production in a Mediterranean Environment. Italian J. Agron. 5 (1–2), 11–19. Giannakopoulos, C., Le Sager, P., Bindi, M., Moriondo, M., Kostopoulou, E., Goodess, C.M., 2009. Climatic changes and associated impacts in the Mediterranean resulting from a 2 ◦ C global warming. Global Planet. Change 68 (3), 209–224. Gominho, J., Lourenc¸o, A., Palma, P., Lourenc¸o, M.E., Curt, M.D., Fernández, J., Pereira, H., 2011. Large scale cultivation of Cynara cardunculus L. for biomass production—a case study. Ind. Crops Prod. 33, 1–6. Grammelis, P., Malliopoulou, A., Basinas, P., Danalatos, N.G., 2008. Cultivation and characterization of Cynara cardunculus for solid biofuels production in the Mediterranean region. Int. J. Mol. Sci. 9, 1241–1258. Grelle, A., Lundberg, A., Lindroth, A., Moren, A.-S., Cienciala, E., 1997. Evaporation components of a boreal forest: variations during the growing season. J. Hydrol. 197, 70–87. Hollinger, D.Y., Richardson, A.D., 2005. Uncertainty in eddy covariance measurements and its application to physiological models. Tree Physiol. 25 (7), 873–885. Ierna, A., Mauro, R.P., Mauromicale, G., 2012. Biomass, grain and energy yield in Cynara cardunculus L. as affected by fertilization, genotype and harvest time. Biomass Bioenergy 36, 404–410. Ierna, A., Mauromicale, G., 2010. Cynara cardunculus L. genotypes as a crop for energy purposes in a Mediterranean environment. Biomass Bioenergy 34, 754–760. Jans, W.W.P., Jacobs, C.M.J., Kruijt, B., Elbers, J.A., Barends, S., Moors, E.J., 2010. Carbon exchange of a maize (Zea mays L.) crop: influence of phenology. Agric. Ecosyst. Environ. 139, 316–324. Kaimal, J.C., Finnigan, J.J., 1994. Atmospheric Boundary Layer Flows—Their Structure and Measurements. Oxford University Press, Oxford, pp. 304. Kahle, P., Beuch, S., Boelcke, B., Leinweber, P., Schulten, H.R., 2001. Cropping of miscanthus in Central Europe: biomass production and influence on nutrients and soil organic matter. Eur. J. Agron. 15, 171–184. Katerji, N., Mastrorilli, M., Rana, G., 2008. Water use efficiency of crops cultivated in Mediterranean region: review and analysis. Eur. J. Agron. 28 (4), 493–507. Katerji, N., Mastrorilli, M., Cherni, H.E., 2010. Effects of corn deficit irrigation and soil properties on water use efficiency. A 25-year analysis of a Mediterranean environment using the STICS model. Eur. J. Agron. 32, 177–185.

150

G. Rana et al. / Agricultural and Forest Meteorology 217 (2016) 137–150

Kljun, N., Calanca, P., Rotach, M., Schmid, H., 2004. A simple parameterisation for flux footprint predictions. Boundary Layer Meteorol. 112, 503–523. Kuglitsch, F.G., Reichstein, M., Beer, C., et al., 2008. Characterisation of ecosystem water-use efficiency of European forests from eddy covariance measurements. Biogeosci. Discuss. 5 (6), 4481–4519. Lee, X., Massman, W., Law, B.E., 2004. Handbook of micrometeorology: a guide for surface flux measurement and analysis. In: Atmospheric and Oceanographic Sciences Library. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 250. Li, S., Kang, S., Zhang, L., Du, T., Tong, L., Ding, R., Guo, W., Zhao, P., Chen, X., Xiao, H., 2015. Ecosystem water use efficiency for a sparse vineyard in arid northwest China. Agric. Water Manage. 148, 24–33. Mauder, M., Foken, T., 2004. Documentation and instruction manual of the eddy covariance software package tk2. Arbeitsergebn. Univ. Bayreuth Abt. Mikrometeorol. 1614–8916, 26–42. Mauder, M., Cuntz, M., Drüec, C., Graf, A., Rebmann, C., Schmid, H.P., Schmidt, M., Steinbrecher, R., 2013. A strategy for quality and uncertainty assessment of long-term eddy-covariance measurements. Agric. For. Meteorol. 169, 122–135. Meyers, T.P., Hollinger, S.E., 2008. An assessment of storage terms in the surface energy balance of maize and soybean. Agric. For. Meteorol. 125, 105–115. Moncrieff, J.B., Massheder, J.M., de Bruin, H., Elbers, J., Friborg, T., Heusinkveld, B., Kabat, P., Scott, S., Soegaard, H., Verhoef, A., 1997. A system to measure surface fluxes of momentum, sensible heat, water vapour and carbon dioxide. J. Hydrol. 188–189, 589–611. Moffat, A.M., Papale, D., Reichstein, M., Hollinger, D.Y., Richardson, A.D., Barr, A.G., Beckstein, C., Braswell, B.H., Churkina, G., Desai, A.R., Falge, E., Govec, J.H., Heimann, M., Hui, D., Jarvis, A.J., Kattge, J., Noormets, A., Stauch, V., 2007. Comprehensive comparison of gap-filling techniques for eddy covariance net carbon fluxes. Agric. For. Meteorol. 147 (3–4), 209–232. Monteith, J.L., 1972. Solar radiation and productivity in tropical ecosystems. J. Appl. Ecol. 9, 744–766. Monteith, J.L., 1977. Climate and the efficiency of crop production in Britain. Philos. Trans. R. Soc. London 281, 277–294. Moureaux, C., Ceschia, E., Arriga, N., Beziat, P., Eugster, W., Kutsch, W.L., Pattey, E., 2012. In: Aubinet, M., Vesala, T., Papale, D. (Eds.), Eddy Covariance: A Practical Guide to Measurement and Data Analysis. Springer, Dordrecht, p. 438. Neukirchen, D., Himken, M., Lammel, J., Czypionka-Krause, U., Olfs, H.-W., 1999. Spatial and temporal distribution of the root system and root nutrient content of an established Miscanthus crop. Eur.J. Agron. 11, 301–309. Nobel, P.S., 1974. Introduction to Biophysical Plant Physiology. Freeman, New York, NY. Papale, D., Reichstein, M., Aubinet, M., Canfora, E., Bernhofer, C., Kutsch, W., Longdoz, B., Rambal, S., Valentini, R., Vesala, T., Yakir, D., 2006. Towards a standardized processing of net ecosystem exchange measured with eddy covariance technique: algorithms and uncertainty estimation. Biogeosciences 3, 571–583. Piscioneri, I., Sharma, N., Baviello, G., Orlandini, S., 1999. Promising industrial energy crop, Cynara cardunculus: a potential source for biomass production and alternative energy. Energy Convers. Manage. 41, 1091–1105. Raccuia, S.A., Melilli, M.G., 2007. Biomass and grain oil yields in Cynara cardunculus L. genotypesgrown in a Mediterranean environment. Field Crops Res. 101, 187–197. Raccuia, S.A., Melilli, M.G., 2010. Seasonal dynamics of biomass, inulin, and water-soluble sugars in roots of Cynara cardunculus L. Field Crops Res. 116, 147–153. Rana, G., Katerji, N., 2000. Measurement and estimation of actual evapotranspiration in the field under Mediterranean climate: a review. Eur. J. Agron. 13 (2–3), 125–153. Reichstein, M., Rey, A., Freibauer, A., Tenhunen, J., Valentini, R., Banza, J., Casals, P., Cheng, Y., Grünzweig, J.M., Irvine, J., Joffre, R., Law, B.E., Loustau, D., Miglietta, F., Oechel, W., Ourcival, J.-M., Pereira, J.S., Peressotti, A., Ponti, F., Qi, Y., Rambal, S., Rayment, M., Romanya, J., Rossi, F., Tedeschi, V., Tirone, G., Xu, M., Yakir, D., 2003. Modeling temporal and large-scale spatial variability of soil respiration from soil water availability, temperature and vegetation productivity indices. Global Biochem. Cycles 17 (4), 1104, http://dx.doi.org/10.1029/2003GB002035.

Reichstein, M., Falge, E., Baldocchi, D., Papale, D., 2005. On the separation of net ecosystem exchange into assimilation and ecosystem respiration: review and improved algorithm. Global Change Biol. 11, 1–16. Richardson, A.D., Anderson, R.S., Arain, M.A., Barr, A.G., Bohrer, G., Chen, G., Chen, J., Ciais, P., Davis, K.J., Desai, A.R., Dietze, M.C., Dragoni, D., Aarrity, S.R., Gough, C.M., Grant, R., Hollinger, D., Margolis, H.A., McCaughey, H., Migliavacca, M., Monson, R.K., Munger, J.W., Poulter, B., Raczka, B.M., Ricciuto, D.M., Sahooa, A.K., Schaffer, K., Tian, H., Vargas, R., Verbeeck, H., Jiao, J., Xue, Y., 2012. Terrestrial biosphere models need better representation of vegetation phenology: results from the North American Carbon Program Site Synthesis. Gobal Change Biol. 18, 566–584. Shurpali, N.J., Hyvönen, N.P., Huttunen, J.T., Clement, R.J., Reichstein, M., Nykänen, H., Biasi, C., Martikainen, P.J., 2009. Cultivation of a perennial grass for bioenergy on a boreal organic soil–carbon sink or source? GCB Bioenergy 1, 35–50. Shurpali, N.J., Biasi, C., Jokinen, S., Hyvönen, N.P., Martikainen, P.J., 2013. Linking water vapor and CO2 exchange from a perennial bioenergy crop on a drained organic soil in eastern Finland. Agric. For. Meteorol. 168, 47–58. Skinner, R.H., Adler, P.R., 2010. Carbon dioxide and water fluxes from switchgrass managed for bioenergy production. Agric. Ecosyst. Environ. 138, 257– 264. Steduto, P., Katerji, N., Puertos-Molina, H., Unlu, M., Mastrorilli, M., Rana, G., 1997. Water use efficiency of sweet sorghum under water stress conditions. Gas-exchange investigations at leaf and canopy scales. Field Crops Res. 54, 221–234. Steiner, J.L., Hatfield, J.L., 2008. Winds of change: a century of agro-climate research. Agron. J. 100 (Suppl. 3), 132–152. Suyker, A.E., Verma, S.B., Burba, G.G., Arkebauer, T.J., Walters, D.T., Hubbard, K.G., 2004. Growing season carbon dioxide exchange in irrigated and rainfed maize. Agric. For. Meteorol. 124, 1–13. Suyker, A.E., Verma, S.B., 2010. Coupling of carbon dioxide and water vapor exchanges of irrigated and rainfed maize–soybean cropping systems and water productivity. Agric. For. Meteorol. 150 (4), 553–563. Tallec, T., Beziat, P., Jarosz, N., Rivalland, V., Ceschia, E., 2013. Crops’ water use efficiencies in temperate climate: comparison of stand, ecosystem and agronomical approaches. Agric. For. Meteorol. 25, 284–307. Tian, H., Chen, G., Liu, M., Zhang, C., Sun, G., Lu, C., Xu, X., Ren, W., Pan, S., Chappelka, A., 2010. Model estimates of net primary productivity, evapotranspiration, and water use efficiency in the terrestrial ecosystems of the southern United States during 1895–2007. For. Ecol. Manage. 259, 1311–1327. VanLoocke, A., Twineb, T.E., Zeri, M., Bernacchi, C.J., 2012. A regional comparison of water use efficiency for miscanthus, switchgrass and maize. Agric. For. Meteorol 164, 82–95. Verburg, P.S.J., Arnone III, J., Obrist, D., Schorran, D.E., Evans, R.D., Leroux-Swarthout, D., Johnson, D.W., Luo, Y., Colemans, J.S., 2004. Net ecosystem carbon exchange in two experimental grassland ecosystems. Global Change Biol. 10, 498–508. Vickers, D., Mahrt, L., 1997. Quality Control and Flux Sampling Problems for Tower and Aircraft Data. J. Atmos. Oceanic Technol 14 (3), 515–526. Wagle, P., Kakani, V.G., 2014. Growing season variability in evapotranspiration, ecosystem water use efficiency, and energy partitioning in switchgrass. Ecohydrology 7, 64–72. Wilson, K., Goldstein, A., Falge, E., Aubinet, M., Baldocchi, D., Berbigier, P., Bernhofer, C., Ceulemans, R., Dolmanh, H., Field, C., Grelle, A., Ibrom, A., Law, K.B.E., Kowalski, A., Meyers, T., Moncrieff, J., Monsonn, R., Oechel, W., Tenhunen, J., Valentini, R., Verma, S., 2002. Energy balance closure at FLUXNET sites. Agric. For. Meteorol. 113, 223–243. Webb, E.K., Pearman, G., Leuning, R., 1980. Correction of flux measurements for density effects due to heat and water vapour transfer. Q. J. R. Meteorol. Soc. 106, 85–100. Zeri, M., Anderson-Teixeira, K., Hickman, G., Masters, M., DeLucia, E., Bernacchi, C.J., 2011. Carbon exchange by establishing biofuel crops in Central Illinois. Agric. Ecosyst. Environ. 144, 319–329. Zeri, M., Hussain, M.Z., Anderson, D., Teixeira, K.J., DeLucia, E., Bernacchi, C.J., 2013. Water use efficiency of perennial and annual bioenergy crops in central Illinois. J. Geophys. Res.: Biogeosci. 118 (2), 581–589.