Keywords

1 Introduction

Many Mediterranean countries, including Egypt, Libya, Tunisia, Algeria, Morocco, Syria, Malta and Lebanon, exhibit water availability below the threshold of 1,000 m3 person−1 year−1. In addition, lower availability than the benchmark of water scarcity is also observed in certain regions within countries such as Spain, Greece and Italy, although on average they exceed the 1,000 m3 person−1 year−1 threshold (UN Population Division 1994). Pressure on the limited water resources in such areas is steadily increasing due to both increasing population and living standards. Higher temperatures and population growth will increase the demands for water in most Mediterranean countries. Higher rates of evaporation would cause rises in salt concentration in surface waterbodies, while rises in sea level would favour sea intrusion into aquifers to coastal areas. Under these conditions, freshwater resources available for agriculture are declining quantitatively and qualitatively (Crescimanno et al. 2004).

Water demands for irrigation are projected to rise, bringing increased competition between agriculture and other users (Crescimanno and Marcum 2009). Therefore, the use of lower-quality water sources, such as saline waters, will inevitably be practised for irrigation purposes in order to maintain an economically viable agriculture (Crescimanno 2001). In order to overcome water scarcity, many countries have adopted the use of marginal, saline water for irrigation (Crescimanno et al. 2004). This coupled with adverse climatic conditions makes the Mediterranean region more vulnerable to salinization and desertification (Szabolcs 1994; Crescimanno and Garofalo 2006b). Salinity acts on plants through non-specific and specific mechanisms. The non-specific effect is due to decreased osmotic potential of soil solution that impedes transpiration and photosynthesis (Shannon and Grieve 1999). Specific effects relate to ion uptake and altered physiological processes resulting from toxicity, deficiency or changes in mineral balance (Hasegawa et al. 2000). Downton et al. (1990) refuted conceptions assuming direct inhibition of photosynthesis by showing that stomatal behaviour altered by salinity sufficiently explains the photosynthetic response.

Quantitative understanding of crop production under deficit irrigation with saline water is generally based on three assumptions. First, an increase in salinity above the crop tolerance level will decrease yield (Maas and Hoffman 1977; Maas 1990); second, biomass production is linearly related to transpiration (de Wit 1958; Shani and Dudley 2001); and third, the effects of salt and water stress on yields are additive (Nimah and Hanks 1973).

The validity of the first two assumptions is well established. The linear dependence of relative dry-matter production (Y actual/Y potential) on relative transpiration (T actual/T potential) under conditions of water deficit has been validated for a variety of climates and crops (De Wit 1958; Shani and Dudley 2001). Under conditions of salt stress (Bresler and Hoffman 1986; Bresler 1987) and Na stress (Shani and Dudley 2001), relative yield and relative transpiration are linearly related.

The validity of the third assumption is less certain. Plants respond to drought by attempting to both decrease transpiration and increase water uptake. Because plants respond to drought induced by limited water or elevated salinity by a similar mechanism, the sum of the matric and osmotic components of the water potential has been used to estimate yield (Nimah and Hanks 1973; Cardon and Letey 1994). However, the complex nature of plant response to salt and water stress may result in a response that is not necessarily equal or additive when the two stress factors are imposed simultaneously.

Grapes have been defined as moderately sensitive to salinity (Maas 1990). Maas reported threshold values for grapevines of ECe 1.5 dS m–1 and a salinity response of 9.6% yield decrease for every subsequent unit (dS m–1) increase in ECe. However, conclusions concerning vine response to salinity are largely based on short-term studies in hydroponic growing conditions or in potting media, and there have been few studies on mature grapevines over time (Walker et al. 2002).

Sicily is a typical Mediterranean country in which conditions of water scarcity and drought as well as increasing use of saline water for irrigation are taking place (Crescimanno 2009). In Sicily, wine grape (Vitis vinifera) is one of the most important crops, both in terms of crop production and economic value. Management options suitable to prevent salinization, while maintaining acceptable levels of crop productivity and/or wine quality in irrigated vineyards, need to be developed in Sicily (Crescimanno and Garofalo 2005, 2006a).

This chapter reports results of a 3-year field investigation carried out in a vineyard located in Sicily (Mazara del Vallo, Trapani) within the framework of the Project Evolution of cropping systems as affected by climate change (CLIMESCO), funded by three Italian Ministries (University, Agriculture and Environment). Research was aimed at investigating soil-plant responses to irrigation with two saline waters having different salinities. Soil hydrological characteristics, soil salinity, crop transpiration and stomatal conductance were measured in experimental plots irrigated by the two different irrigation waters. Effect of water and salinity stress on crop physiology and yield was explored by using the equation proposed by Doorenbos and Kassam (1979) and by the equation proposed by Maas and Hoffman (1977).

2 Materials and Methods

2.1 Field and Irrigation Description

Investigation was carried out at the Foraci Farm (http://www.cantineforaci.com/), a vineyard located in the Mazzaro basin region of Sicily. Two different irrigation treatments were established (L and R) to monitor soil and plant responses to irrigation water salinity (Fig. 29.1).

Fig. 29.1
figure 1

Foraci Farm (Mazzaro basin, Sicily, Italy): location of the two treatments (L and R)

Full size image

Irrigation treatment L used irrigation water from a lake having ECw  =  1.6 dS m−1; irrigation treatment R used water from a well having ECw  =  0.6 dS m−1.

Two vine rows, designated L row and R row, were selected, one in each of the two treatments. Eleven soil sites, corresponding to plant numbers 1, 10, 20, 30, 40, 50, 60, 70, 80, 90 and 100, were selected along each of the two rows for determining the soil hydraulic parameters and plant physiological parameters. Undisturbed soil samples were collected at the selected locations to determine the soil shrinkage curve and the water retention curve.

Irrigation scheduling was established according to a water balance model taking into account climatic data, the amount of soil water available for crops (AWmax) and crop parameters.

Table 29.1 reports the irrigation scheduling applied from 7 June 2007 to 30 July 2009. Irrigation in R and L row plots was performed on subsequent days, and measurements were carried out in plots 1 day after irrigation. Irrigation amount was 15 mm deep per event. Reference evapotranspiration (ETo) was calculated by using the Hargreaves equation (Hargreaves and Samani 1982).

Table 29.1 Irrigation scheduling and measurement dates from 2007 to 2009
Full size table

2.2 Soil Physical and Hydraulic Parameters

Replicated soil cores of different sizes to measure physical and hydraulic characteristics were sampled from different horizons of 11 plot sites selected along each row. The soil shrinkage curve was determined by measuring vertical and horizontal shrinkages (Crescimanno and Provenzano 1999). The water retention curve was determined by the tension method at matric potential at h values between 0 and −150 cm and by the pressure membrane plate apparatus at h values up to −300 cm. Volumetric water content (θ) corresponding to gravimetric water content (U) was determined by using the bulk density (ρ b) values obtained by the shrinkage curve. The equation proposed by Brutsaert (1966) was fitted to the θ and h values. Parameter estimation was performed by fixing the saturated water content, θ s, at the measured value and optimizing the residual water content θ r, together with the Brutsaert α′ and n′ parameters (Crescimanno and Garofalo 2005).

2.3 Soil and Physiological Measurements

Gravimetric water content, U, soil salinity (ECe) and physiological measurements (transpiration, stomatal conductance) were measured 1 day after each irrigation on a total of 11 plants per row going from plant no. 1 to plant no. 100 corresponding to soil sampling plot locations along the L and R rows. U and ρ b (U) were used to calculate volumetric water content θ, which therefore accounted for a variable soil volume. Soil saturated extracts were prepared using the soil collected at 60 cm at each soil-plant sampling plot site, and soil electrical conductivity (ECe) was measured by a conductivity meter (Crison, Micro CM 2002) (Rhoades 1993).

Crop transpiration (T r) and stomatal conductance (G s) measurements were made on three recently matured leaves per plant using a CIRAS-2 portable infrared gas analyzer (PP Systems). Crop water stress index (CWSI) was calculated according to Doorenbos and Kassam (1979):

$$ \text{CSWI}=1-\frac{{T}_{\text{r}}}{{T}_{\text{m}}}$$
(29.1)

with T r  =  actual transpiration (mm) and T m maximum transpiration (mm). T m was calculated as a percentage of maximum evapotranspiration (ETm) (T m  =  0.9 ETm).

2.4 Yield Reduction Due to Water Stress

The effect of water stress on yield can be quantified by relating the relative yield reduction (Y r) to the relative evapotranspiration deficit (1  −  ETe/ETm) through an empirically derived yield response factor (Ky) (Doorenbos and Kassam 1979):

$$ {Y}_{\text{r}}=1-\frac{{Y}_{\text{a}}}{{Y}_{\text{m}}}=\text{Ky}\left(1-\frac{{\text{ET}}_{\text{e}}}{{\text{ET}}_{\text{m}}}\right)=\text{Ky}\left(1-\frac{{T}_{\text{r}}}{{T}_{\text{m}}}\right)$$
(29.2)

where Y a is the actual yield, Y m is the maximum potential yield and Ky is the yield response factor.

2.5 Yield Reduction Due to Salinity

Relative yield (\( {{Y}^{\prime }}_{\text{a}}\)/Y m) under salinity conditions can be predicted by the Maas and Hoffman equation (1977):

$$ \frac{{{Y}^{\prime }}_{\text{a}}}{{Y}_{\text{m}}}=100-b\left({\text{EC}}_{\text{e}}-a\right)$$
(29.3)

where \( {{Y}^{\prime }}_{\text{a}}\)/Y m (%) is the relative yield, ECe (dS m−1) is the electrical conductivity of saturated extract, a (dS m−1) is the salinity threshold value (ECe where \( {{Y}^{\prime }}_{\text{a}}\)  =  Y m) and b (% per dS m−1) is the slope of the regression line. According to Maas (1990), for grapes a is equal to 1.5 dS m−1 and b is equal to 9.6.

Equation 29.3 can be rewritten as follows:

$$ {{Y}^{\prime }}_{\text{r}}=100-b\left({\text{EC}}_{\text{e}}-a\right)$$
(29.4)

where \( {{Y}^{\prime }}_{\text{r}}\)represents the yield reduction due to salinity.

3 Results and Discussion

3.1 Soil Physical and Hydraulic Parameters

Table 29.2 reports soil classification and physical characteristics (particle-size distribution soil texture) of some profiles (E, F, G and H) located along the L and R rows. Table 29.3 reports the Brutsaert water retention parameters obtained for the 11 soil plot locations corresponding to plants from 1 to 100 along the R and L rows.

Table 29.2 Classification and physical properties of the E, F, G and H profiles
Full size table
Table 29.3 Parameters of the water retention curves (Brutsaert) determined for the 11 soil-plant locations along the L and R rows
Full size table

Table 29.4 reports the maximum water available for crops (AWmax) calculated from field capacity and wilting point (θ r) for the 11 L and R plot locations.

Table 29.4 Maximum water available for crops (AWmax) and wilting point (θ r)
Full size table

Statistical analysis (paired t-test) proved that differences between the AWmax values measured in L and R row plots were nonsignificant (P  =  0.01). The same result was obtained with reference to θ r. The two row plots therefore had a similar hydrological behaviour in terms of amount of water retained by the soil and made available to crops. However, the higher θ r values measured in the L plots might be a consequence of the higher salinity, causing swelling of these soils in the course of the laboratory experiments.

3.2 Electrical Conductivity of Saturated Extract (ECe)

Figure 29.2 illustrates ECe values measured in the course of the 2007, 2008 and 2009 irrigation seasons.

Fig. 29.2
figure 2

Electrical conductivity (ECe) measured after irrigation at the depth of 60 cm in the 11 locations along the R and L rows in 2007, 2008 (a) and 2009 (b)

Full size image

Significantly higher ECsat values were measured in the L row, compared to R row plots across all 3 years, consistent with the higher salinity of water used to irrigate the L row plots.

3.3 Transpiration (T r) and Stomatal Conductance (G s)

Figure 29.3 illustrates transpiration (T r) values (mmol m−2 s−1) measured 1 day after irrigation events in the 11 plants located along the L and R rows from 2007 to 2009.

Fig. 29.3
figure 3

Transpiration (T r) values measured after irrigation events in L and R rows (plants from No. 1 to No. 100) in 2007, 2008 (a) and 2009 (b)

Full size image

Paired t-tests proved that significantly higher T r values were measured in R row plots compared to values measured in L row plots, across all 3 sampling years. Higher T r values measured in R plots indicate that greater water uptake occurred in the R plot plants in all 3 sampling years.

Figure 29.4 illustrates stomatal conductance (G s) values measured 1 day after irrigation events from 2007 to 2009 in the 11 plants located along the L and R rows.

Fig. 29.4
figure 4

Stomatal conductance (G s) values measured after irrigation events in L and R rows in 2007, 2008 (a) and 2009 (b)

Full size image

Consistent with the higher T r, statistically significant higher G s values were measured in the R row plots, compared to values measured in the L row plots, across all 3 sampling years.

The results indicated that significantly higher T r and G s were measured in the R row plots compared to those measured in the L row plots in all 3 considered years.

Figure 29.5 illustrates the regression relationships between the T r and G s values measured 1 day after irrigation events from 2007 to 2009 in R and in L row plots.

Fig. 29.5
figure 5

Relationship between T r and G s (R and L rows)

Full size image

As can be seen in Fig. 29.5, both the T r, and G s values measured in the R row plots extended over a much wider range than those measured in the L row plots. For example, G s values reached 1,200 mmol m−2 s−1 in R row plots; however, values measured in L row plots reached a maximum value of only 600 mmol m−2 s−1. This indicates that R row plot plants maintained much higher stomatal conductance and transpiration rates than L row plot plants; in other words, stomatal conductances and associated transpiration rates were highly correlated to irrigation treatment and soil salinity. However, soil solutions in both R row and L row irrigation plots maintained the same hydrological and meteorological conditions throughout the experiment (Table 29.4). Therefore, stomatal closure and resultant decline in transpiration rates occurred in L row plot plants due to the increased salinity and lower osmotic potential of the soil solution in this treatment, which was irrigated with saline lake water, and not due to soil moisture or hydrological differences. Since no significantly different AWmax and θ r values were measured in the two treatments, the higher T r and G s measured in the R row plots can only be attributed to the significantly lower (less saline) ECe measurements in R plots.

3.4 Crop Water Stress Index

Figure 29.6 illustrates crop water stress index (CWSI) values calculated for the R and L row plots during 2007, 2008 and 2009. Plants located in L row plots experienced higher CWSI’s (closer to 1) than those located in R row plots, indicating that L row plants were under greater stress.

Fig. 29.6
figure 6

CWSI determined in the R and L rows during 2007, 2008 (a) and 2009 (b)

Full size image

Statistical analysis (paired t-test) proved that significantly higher CWSI’s occurred during 2007, 2008 and 2009 in the L row plants, indicating stronger stress conditions, compared to those measured in R row plants.

Annual averages for T r, G s, CWSI, ECe, θ r and AWmax values are shown in Table 29.5.

Table 29.5 Annual average values for T r, G s, CWSI, ECe, θ r and AWmax (2007–2009)
Full size table

Table 29.5 indicates that the two irrigation row treatments experienced similar (not significantly different) soil hydrological conditions in terms of maximum water available to crops (AWmax) and of wilting point (θ r). The significantly lower T r and G s and significantly higher CWSI values measured in the L row plots were therefore a consequence of the significantly higher ECe measured in the L plot soils during the course of the 3 years. In addition, ECe value differences between R and L row treatments increased in magnitude from 2007 to 2009, indicating salt accumulation processes occurring in the L plot soil.

3.5 Yield Reduction Due to Water and Salinity stress

Table 29.6 reports yield reduction values (Y r) calculated according to Eq. (29.2) for the R and L row plots using CWSI values averaged over the years and using a Ky factor of 0.85 (Doorenbos and Kassam 1979). As can be seen in Table 29.6, higher Y r values (%), consistent with the higher CWSI values, were predicted for the L row compared to those obtained for the R row over the 3 years. The difference (ΔY r) between the Y r obtained for the L and R row treatments was equal to 18.78% in 2008 and 8.16% in 2009, the years for which yield data (Y meas) was available for R and L row plots.

Table 29.6 Annual averages of CWSI, Y r, Y meas, ΔY meas and ΔY r (2007–2009). Ky  =  0.85
Full size table

Predicted ΔY r were higher than those measured (ΔY meas), which may indicate that a Ky lower than 0.86 would be more appropriate for the local conditions under which these experiments were carried out. Table 29.7 reports recalculated Y r values, again according to Eq. 29.2, but using a Ky  =  0.7.

Table 29.7 Y r, ΔY meas, ΔY r, Y′r and R Y (2007–2009). Ky  =  0.70
Full size table

Table 29.7 shows that ΔY r values predicting a Ky  =  0.70 value were closer to those measured (ΔY meas) both in 2008 and in 2009. These results indicate that accurately predicted Y r can be calculated by Eq. 29.2 using Ky values calibrated under specific soil-plant conditions. Table 29.7 also reports the value of yield reduction due to salinity (Y′r) calculated for the L and the R row treatments according to Eq. 29.4. Higher \( {{Y}^{\prime }}_{\text{r}}\)values, consistent with the higher ECe values (Table 29.5), were found for the L row. However, Table 29.7 shows that the \( {{Y}^{\prime }}_{\text{r}}\)values were always lower than Y r, and lower than ΔY meas. These results can be explained by the fact that in our case, as is always the case where deficit irrigation is practised, concomitant conditions of water stress and salinity stress occurred. In this case transpiration (T r), used to calculate CWSI, included crop response to water and salinity stress, and therefore, Y r calculated by Eq. 29.2 took into account both the effects of water deficit and of the salinity stress.

Table 29.7 also reports the ratio (R Y) between the yield reduction predicted by Eq. 29.4 (\( {{Y}^{\prime }}_{\text{r}}\)) and the yield reduction predicted by Eq. 29.2 (Y r). The R Y obtained indicated that the yield reduction due to salinity represented a percentage of the total yield reduction of up to 3.3% for the R row and up to 11.7% for the L row treatments. These results showed that salinity had a considerable influence on the physiological processes occurring in plants in the L plot compared to the effect of salinity on plants located in the R plot, significantly affecting crop yield.

Figure 29.7 illustrates the CWSI as a function of ECe (mean annual values for 2007, 2008 and 2009) for the L and the R row plots.

Fig. 29.7
figure 7

Regression lines between CWSI and ECe obtained using values reported in Table 29.5

Full size image

The figure shows a different behaviour of the CWSI- ECe relationship obtained for the L row compared to the relationship obtained for the R row plots. The two different regression lines obtained cross close to the value of 1.5 dS m−1. This indicates that a value of ECe 1.5 dS m−1 discriminated plant response to salinity under our experimental conditions and can be considered an indirect confirmation of the threshold value for grapes indicated by Maas (1990).

4 Conclusions

Crop transpiration (T r) and stomatal conductance (G s) measured in grapevines irrigated with water of different salinity proved to be significantly affected by soil salinity conditions, expressed by electrical conductivity of soil saturated extract (ECe). Significant reductions in T r and G s were measured in plants in the treatment irrigated with water having ECw  =  1.6 dS m−1 (L plots) compared to T r and G s values measured in plants irrigated with water having a salinity of 0.6 dS m−1 (R plots).

Significantly higher crop water stress index (CWSI) values were measured in the L plot compared to those measured in the R plot, indicating significantly lower transpiration in the L plot. Validity of the linear relationship between relative yield and relative transpiration was confirmed by this investigation. A value of 0.7 for the yield response factor provided accurate prediction of yield reduction both in 2008 and in 2009. However, although CWSI is generally considered an index of water stress only, when conditions of salinity occur during deficit irrigation, as in our case, CWSI incorporates both water stress and salinity stress.

Yield reductions due to salinity, calculated according to Maas and Hoffman (Eq. 29.4), showed that under our concomitant conditions of water and salinity stress, the yield reduction due to salinity (\( {{Y}^{\prime }}_{\text{r}}\)) represented a percentage of the total yield reduction (Y r) up to 11% in the L treatment plots and up to 3.5% in the R treatment plots. The investigation also indicated that under our conditions a value of ECe  =  1.5 dS m−1, which is the salinity threshold proposed by Maas, discriminated a different plant response to salinity between the L and R plots. Under the irrigation conditions investigated in the Foraci vineyard, it might therefore be advised to implement management strategies aimed at keeping salinity under this threshold value. This objective could be realized by irrigation performed using the lower-salinity water only or by alternating the two irrigation waters.