1 Introduction

The consideration of sweet sorghum as an alternative crop for bioethanol production requires a thorough study of the water needs in relation to its productivity. This relation is usually expressed by the water use efficiency (WUE), the ratio of yield (y) to evapotranspiration (ET), estimated at various levels of soil moisture availability (Losavio et al. 1994; Tayot et al. 1994; Curt et al. 1995; Mastrorilli et al. 1995, 1999; Cosentino 1996; Dalianis et al. 1996a; Dercas et al. 1996; Dercas and Liakatas 2007; Sakellariou-Makrantonaki et al. 2007; Garofalo et al. 2011).

In addition, the efficient radiation use of a crop is an essential criterion for its selection and promotion as a new energy crop. Crop growth depends on the ability of leaves or roots to receive and use solar radiation, CO2, water and nutrients, with solar radiation providing the energy to drive both CO2 assimilation and water transpiration processes. Since these two processes occur at the same time, at the same surface of exchange with the atmosphere and in part through the same path, plants have developed resource optimization mechanisms (Cowan 1982), so that a strict link between assimilation and transpiration is also established (Hsiao and Bradford 1983).

Consequently, growth-engines (carbon-, water- or solar-driven) based on the mechanistic relationships between the major processes involved (either CO2 assimilation or water transpiration) and growth, or between the driving energy (solar radiation) and growth (Steduto 2003), have solar radiation as the driving source of energy (Albrizio and Steduto 2005).

Radiation interception varies from seedling emergence to crop harvest (e.g., Watiki et al. 1993), largely depending on the canopy leaf area (Biscoe and Gallagher 1977).

In the absence of biotic or abiotic stresses, yield is related to the amount of radiation intercepted by the crop affected by the timing of canopy closure (Ottman and Welch 1989).

In a previous study (Dercas and Liakatas 1999), a model was developed (calibrated and validated using data of the 1994 and 1995 cultivation periods, respectively) in order to estimate the actual evapotranspiration of sweet sorghum. The crop coefficient Kc values were assessed and they were found to be considerably higher than the Kc values suggested by the Food and Agriculture Organization (FAO) (Doorenbos et al. 1986). A second study by Dercas and Liakatas (2007) found a linear relation between water consumption and radiation use efficiency (RUE) for the sweet sorghum. According to this relationship, high water use efficiency values tend to be related with low radiation use efficiency values.

In this paper, we study sweet sorghum canopy (leaf area) development in relation to the radiation capture and the water availability, by examining RUE and WUE for various irrigation and fertilization treatments.

2 Materials and Methods

Data from two experimental sites and four cultivation periods with sweet sorghum were used. Both experimental sites were about 1.5 km apart from each other, chosen independently in the framework of related EU research projects.

2.1 Experimental Site A

During the 1994 and 1995 cultivation periods, the cv Keller was used. The two crops were sown in plots on 10 May 1994 the first and 3 May 1995 the second, with a plant density of ~138,900 plants ha−1. The experimental layout was a randomised complete block design with three replications. The experiment consisted of four irrigation treatments, applying four soil water regimes: (1) highly irrigated (IH) (458 mm and 512 mm in 1994 and 1995, respectively); (2) highly irrigated only until anthesis (IHA) (364 mm and 432 mm in 1994 and 1995, respectively); (3) medium irrigated (IM); and (4) low irrigated (IL). Soil water regimes under the IM and IL treatments were, respectively, 0.56IH and 0.34IH in 1994 and 0.64IH and 0.46IH in 1995 (Dercas and Liakatas 1999). For all treatments, the fertilization rate was 40 kgN/ha.

2.2 Experimental Site B

In 1997, cv MN1500 was sown (~110.000 plants ha−1) on May 16 and in 1998 cv Keller was sown on May 22 at the same plant density. The experimental layout was a randomized complete block design with three replications (Dercas et al. 2000). The experiment consisted of two irrigation treatments (IM = 348 mm and IH = 502 mm in 1997, and IM = 317 mm and IH = 510 mm in 1998), imposing two soil water regimes: (1) highly irrigated (IH); (2) medium irrigated (IM). The medium irrigation rate (IM) was 0.69 and 0.62 of IH in 1997 and 1998, respectively. Three fertilization rates (F0 = 0 kgN ha−1, FM = 60 kgN ha−1, FH = 120 kgN ha−1) were applied.

In both sites, the soil profile was characterized as homogenous: in site A, the soil was clay loam down to 0.7 m and sand clay loam down to 1.8 m, whereas in site B the entire profile was a sand clay loam. In both experimental fields, there was a very deep ground water table that did not affect the root zone.

A drip irrigation system was used for the irrigation of every experimental field. The irrigation system was also used for the nitrogen fertilization.

Soil water content was measured gravimetrically in the depth range of 0.0–0.2 m and with a Neutron moisture meter from 0.2–1.8 m, at an interval of 0.15 m every week as well as one day before and two days after irrigation or rainfall.

An automatic weather station installed near the experimental fields recorded every 10 s and averaged or summed every 15 min the following meteorological parameters: wind speed at 2 and 6 m height, relative humidity and air temperature at 2 m above soil surface, solar and net radiation approximately at 2 m above the crop canopy. These data were supplemented by measurements of rainfall and photosynthetically active radiation (PAR). In addition, solarimeters were installed inside the plantation to estimate the radiation intercepted by the canopy. The monthly variations of the most important meteorological parameters during the four cultivation periods are shown in Table 1.

Table 1 Monthly values and mean/total of meteorological parameters during the 1994, 1995, 1997 and 1998 cultivating seasons of sorghum (Vagia, Central Greece)
Full size table

To evaluate the growth and the biomass production, several harvests were carried out throughout the growing periods. Fresh weights of whole plants and separately of stems, leaves and panicles were measured and the average height of three plants was estimated. A representative subsample from each plant part was then used to determine aerial dry matter. The leaves of 1.0 m row plants were used for LAI evaluation with the help of a leaf area meter.

3 Results and Discussion

3.1 Soil Moisture, Evapotranspiration, Water Use

Figures 1a and b present the variation of the soil moisture (mm of water) in the root zone (0–1.8 m) during 1995 and 1997. At the beginning of the cultivation period, a high (close to field capacity) water content was observed, followed by a progressive decrease, especially in the case of the deficit irrigation rates. The soil moisture associated with the high irrigation treatment, however, remained higher than 50% of the available water capacity during most of the cultivation period. This variation of the soil water content implies that plants having received the highly irrigated treatments did not suffer from any significant water stress. A similar trend was observed during the 1994 and 1998 experiments.

Fig. 1
figure 1

Soil moisture variation during 1995 (a) and 1997 (b) (mm of water in the root zone, 0–180 cm, DOY: Day of Year), with (IH) highly irrigated treatment, (IHA) highly irrigated only until anthesis, (IM) medium irrigated, and (IL) low irrigated

Full size image

Using the data of the soil moisture content, rainfall and irrigation, the actual evapotranspiration was estimated by the water balance method. Then, taking into account the final yield (aerial dry matter) of the various treatments and the cumulative actual evapotranspiration, the water use efficiency for the cultivation period was estimated. Results are presented in Table 2.

Table 2 Total irrigation, cumulative seasonal ET, final yield and WUE for the four cultivation periods
Full size table

The MN1500 (experiment of 1997) is characterized by a high number of tillers, high LAI (up to 8.7) and high biomass production. This led to high WUE values (up to 56.5 kg ha−1 mm−1) for the well watered treatment. WUE was higher by 5 kg ha−1 mm−1 on average, for the less watered treatments, where mean water consumption for seasonal ET was less by about 140 mm, being on average 630 mm for the well watered plants. Net photosynthetic rate and transpiration rate are functions on leaf conductance, which, in turn, depends on soil water availability. Under water stress (soil moisture lower than 50% of maximum available at root zone), the net photosynthetic rate was found by Tingting et al. (2010) to be significantly lower in comparison to non-stressed sweet sorghum plants. Under water stress, stomatal conductance is reduced, whereas WUE is increased (Wand et al. 2014; Enciso et al. 2015). Here, WUE diminishes with seasonal ET increase, positively affecting mean conductance value. Considering the variability of WUE values among cultivation periods and irrigation treatments, Garofalo and Rinaldi (2013) suggested that WUE, normalized with vapour pressure deficit (VPD), is more suitable to different climatic conditions.

Apart from water availability, fertility also affects biomass production. Sawargaonkar et al. (2013) reported that WUE is also a function of nitrogen use, whereas Wang et al. (2014) suggested that improved water and nitrogen use efficiencies under water stress may contribute to the high degree of physiological acclimation of sweet sorghum to drought.

The WUE values for Keller ranged from 43.0 to 57.2 kg ha−1 mm−1 (or from 231.6 to 174.7 mm/kg of dry biomass). These values are very close to those (165.9 to 244.3 mm/kg of dry biomass) obtained for the same variety in Italy reported in Mastrorilli et al. (1995).

Using ETmax (taken by the water balance when soil moisture was at or close to field capacity) and ETref (obtained by the Penman-Monteith method), the crop coefficient Kc values for various crop stages were evaluated. Corresponding values proposed by FAO (Doorenbos et al. 1986) are also presented in Table 3. Evaluated Kc values are close to the FAO Kc values although somewhat higher during the pre-anthesis period and especially during the late-season period. There is no significant difference between cultivars.

Table 3 Kc values for various crop stages
Full size table

3.2 Radiation Regime, Canopy Development, Radiation Use Efficiency

In comparison to all treatments, the least irrigated crop (IL) had the minimum efficient radiation interception canopy permitting, thus, a larger part of the available energy to reach the ground (Fig. 2).

Fig. 2
figure 2

Global solar radiation incident and penetrating the crop canopy (MJ/m2 per day) against time in all irrigation treatments (1994), with (IH) highly irrigated treatment, (IHA) highly irrigated only until anthesis, (IM) medium irrigated, and (IL) low irrigated

Full size image

The total short wave radiation ‘loss’ for the leafage, indicated by the ratio of the radiation passing through the crop canopy to the incident (τ T  = R su /R so , where Rsu global solar radiation beneath the canopy and Rso incident global radiation), starts to decline from 1.0 at LAI values between 1.0 and 2.0 and becomes minimum (~0.1) at or soon after maximum green LAI ~5.5–6.0 (Fig. 3). Similar results were reported by Curt et al. (1997), with Keller variety in experiments in Spain. IL crop resulted in slow vegetative development of a finally poorer LAI by at least 1.0.

Fig. 3
figure 3

Time progress of the ratio of penetrating over incident global solar radiation and LAI, under low and high irrigation treatments (1994), with (IH) highly irrigated treatment and (IL) low irrigated

Full size image

By applying the Beer’s law τ T  = exp(− k T  LAI), the radiation extinction coefficient (k T ) was estimated as the slope of the linear relationship between LAI and the logarithm of τ T (Fig. 4).

Fig. 4
figure 4

Logarithm of the ratio of penetrating over incident global solar radiation against LAI for all treatments in 1994, with (IH) highly irrigated treatment, (IHA) highly irrigated only until anthesis, (IM) medium irrigated, and (IL) low irrigated

Full size image

The values of k T ranged between 0.45 and 0.47 for IM, IH and IHA 1994 plots (R2 = 0.87–0.94), except for IL where extinction of radiation was significantly lower (k T = 0.28, R2 = 0.84). However, by grouping data of the same variety (Keller) from three cultivation periods (1994, 1995 and 1998) according to the total actual evapotranspiration (ET), either between 600 mm and 700 mm or close to 450 mm, k T values were actually not different (Fig. 5a, b), approximately equal to 0.31, though with apparently lower R2 (0.57–0.63). This value is close to that (0.37) obtained by Hammer et al. (2010), for four grain sorghum genotypes and those obtained by Farré and Faci (2006), for a grain sorghum variety under full irrigation (0.42) and moderate water supply (0.32). Such low k T values for global solar radiation are not in contradiction with higher k values reported for intercepted PAR (Ceotto et al. 2013). Goudriaan and van Laar (1994) pointed out that k for near infrared radiation (NIR) is about half of k q for PAR.

Fig. 5
figure 5

Logarithm of the ratio of penetrating over incident global solar radiation against LAI under restricted (a) and high (b) evapotranspiration for all cultivation periods of cv. Keller

Full size image

Taking the estimated value into account, the sorghum leaf absorptivity of quanta (a q ), or photosynthetically active radiation, is approximately 4 times that of the total solar radiation (a T ), considering PAR extinction coefficient k q =0.6 (Varlet-Grancher et al. 1992; Lemaire and Chartier 1996; Fletcher et al. 2013) in equation \( {k}_q/{k}_T={a}_q^{1/2}/{a}_T^{1/2} \) (Monteith and Unsworth 1990). Consequently, considering τ T =0.1 (at LAI~max), the penetration of PAR (or quanta) \( {\tau}_q={\tau}_T^{\left({a}_q^{1/2}/{a}_T^{1/2}\right)} \)(Dercas and Liakatas 2007) is 10 times less (τ q =0.01), meaning that practically all PAR is intercepted by the leafage.

The value of LAImax (τ T = 0.1) of the highly irrigated Keller plants was reached on DOY 205 (Day of the Year - 24th of July), about a month after solar light duration and intensity maximization. It took, however, at least another month for the least watered plants to achieve the same degree of leafage efficiency for radiation capture (Fig. 3). In 1997, when cv MN1500 was cultivated, radiation measurements started rather late, when crop canopy was already (LAI~6.0) an efficient light interceptor (Fig. 6a, b). With this variety, when plants were fully irrigated and sufficiently fertilized (IHFH and IHFM), τ T  = 0.1 was observed at the same (as for Keller) time, but about 3 weeks later when either non-fertilized (IHF0) or poorly irrigated and fully fertilized (IMFH). LAImax, although exceeding 8.0, led to practically no extra radiation gain and no significantly higher above ground dry matter production. Similar results were reported by Lemaire and Chartier (1996) and Ceotto et al. (2013).

Fig. 6
figure 6

Time progress of the ratio of Rsu/Rso and green LAI (a), as well as regression lines of the logarithm of the ratio against LAI (b), for a combination of irrigation and fertilization treatments (MN 1500, 1997)

Full size image

Crop water coefficients (K c  = ETmax/ET ref ) for both varieties were also calculated for soil moisture regimes close to field capacity (within 0.25 of the total soil moisture availability). Values exceed 1.2, with a tendency to decline at the onset of the reproductive stage (Fig. 7a, b). Kc becomes maximum earlier than LAI, rather coinciding with radiation maximization. It seems that transpiration water loss, apart from the water vapor demand of the atmosphere (Cosentino 1996), depends more on the time of stomata opening (day length) and the energy availability for water vaporization rather than on the total transpiring leaf surface. Therefore, crop canopy should be able to receive the largest part of incident radiation at the time of its maximum availability and greatest water demand (end of June). Maximization of leafage development rate (to maximize LAI earlier) is possible by the management of water and fertilizer application.

Fig. 7
figure 7

Kc against LAI variation for all cultivation periods (1994, 1995, 1998) of cv. Keller (a) and one period (1997) of cv. MN 1500 (b). The arrows indicate time of max radiation

Full size image

On the other hand, as the starting time of sorghum plants life is thermally restricted (Chapelle et al. 1996) and leaf growth is temperature dependent (Bell et al. 1992; Andrade et al. 1993; Tayot et al. 1994; Charonnat et al. 1996), changes in thermal conditions, though economically unjustifiable (unless resulting from climatic variation), may also lead to faster canopy development. Tillers formed early, when canopy is incomplete, contribute to radiation capture, despite their early dying (Dalianis et al. 1996a, b) after sufficient leaf area development of the higher canopy layers. If densely sown, high LAI (>6.0) and more tillers producing varieties (like MN1500) may have increased lodging risk due to exceptionally high and thin main stems (Dalianis et al. 1996b), especially if well irrigated (Dercas and Liakatas 1999).

Finally, the radiation use efficiency was estimated for the four cultivation periods (Table 4). The values obtained are 2.4–2.8 g of dry matter/MJ intercepted for the experiments 1994–1995 and 2.75–3.63 g of dry matter/MJ intercepted for the 1997–1998 experiments. These results indicate that the lower plant density during 1997 and 1998 allowed for a better radiation use. The values are close to those obtained by other authors (Varlet-Grancher et al. 1992; Mastrorilli et al. 1995; Cosentino et al. 1997; Rinaldi and Garofalo 2011; Garofalo et al. 2011).

Table 4 RUE for the four cultivation periods
Full size table

Mean RUE for IH treatments was 3.19 and for IM treatments 2.67 g MJ−1. RUE seems to be related to crop water consumption. The mean ratio of RUE to seasonal actual evapotranspiration (ET) (i.e., a combined R-WUE in mg per MJ of PAR intercepted and mm of water used) was 5.1 and 5.6, respectively, for IH and IM treatments. Thus, although these values are not significantly different, it seems that each mm of water added to a stressed sweet sorghum crop will be slightly more productive in comparison to non-stressed crops. In Italy, RUE for biomass sorghum, with an average value 2.91 g MJ−1 (very close to our overall average 2.93 g MJ−1), generally increased by 4.2 mg MJ−1 per mm of water used in the range 454–891 mm (Rinaldi and Garofalo 2011).

It seems that in a Mediterranean environment, high radiation efficiency in biomass production (high yield potential) of sweet sorghum can be expected, provided adequate irrigation water is available during the entire growth cycle (Rinaldi and Garofalo 2011), and when there is no shortage of soil fertility and atmospheric CO2 concentration. Also, early plantation, as it comes out of a research work by Houx III and Fritschi (2015), and fast leaf area development would ensure high RUE values.

4 Conclusions

The following conclusions may be drawn from the present study:

The selective absorptivity of sweet sorghum results in PAR penetration through the crop canopy of only 0.01 when the leaf area index is 5.5–6.0.

Maximum LAI higher than 6.0 leads to no extra PAR capture and no significantly higher biomass production. LAI maximization occurs after maximization of radiation availability and crop water demand. Faster canopy development by rational application of water and fertilizer would attain the highest possible capture and use of radiation at its maximum availability. In comparison to water stressed plants, well watered plants were found to use radiation in a more efficient way. Roughly, 5 mg MJ−1 per mm of water evapotranspired by a sweet sorghum crop may be produced in a Mediterranean environment, where radiation availability is usually sufficient but water availability is not.

For high irrigation treatments, WUE ranges between 43 and 56.5 kg ha−1 mm−1 and for medium irrigation rates from 44 to 61.3 kg ha−1 mm−1. The highest values were obtained for MN1500 characterised by a high number of tillers, high LAI and high biomass production.

Calculated Kc values are mostly higher than the Kc values suggested for sorghum by FAO implying higher crop water requirements, especially during the vegetative period of sweet sorghum.