Crop water productivity and yield response of two greenhouse basil (Ocimum basilicum L.) cultivars to deficit irrigation

Basil (Ocimum basilicum L.), belonging to the Laminaceae family, is one of the most important medicinal plants, which is widely grown for purposes of food and industry (Hartley et al. 2000). Close to 60 species of the basil plant exist across the world and are cultivated in Israel, Hungary, Egypt, the United States, France, Greece, Morocco and Indonesia (Srivasta 1980). The economic organs of the basil plant are seed and leaf (Arabaci & Bayram 2004). Basil with diuretic, stimulant, expectorant and carminative properties has been utilized for stomachache, cough, worms, and headaches in folk medicine (Simon et al. 1990; Arabaci & Bayram 2004). It is known as a source of aroma components and therefore has a range of biological attributes like being a flea and moth repellent, insecticide and anti scorpion, snake and insect bite (Juliani & Simon 2002; Lee et al. 2005; Ekren et al. 2012). Furthermore, the fresh and dry leaves of the basil plant are very popularly consumed in the spice and food industry (Ekren et al. 2012).

The cultivation of medicinal plants has long been common practice to create diversity and sustainability in Iran's agricultural systems. Among the species of Laminaceae family, O. basilicum is cultivated as the most important economic species in most regions of Iran for fresh consumption and rarely for production of essential oils and as the dry product (Hamzezadeh et al. 2011). Water availability is the main challenge for crop production in Iran. The increasing population and rising water and food demand on the one hand, and recent drought and restricted availability of water, on the other hand, have necessitated the suitable use of available water resources (Albaji 2010; Mahjoobi et al. 2010; Hooshmand et al. 2019).

Irrigation scheduling and technologies may be improved for more rational and efficient uses of limited water resources (Abdullatif & Asiri 2012). Deficit irrigation (DI) is a proper method to decrease the amount of applied water, especially in arid and semi-arid regions with scarcity of water (Safi et al. 2019). In DI conditions, the crop faces a particular water stress level either during a special period or the whole growing season (Khalid 2006; Abdullatif & Asiri 2012). However, depending on the locations, DI may have different effects on the same crops (Kresovic et al. 2016) and this is a way of improving crop water productivity (CWP) for higher production per unit of evapotranspirated water (Abdullatif & Asiri 2012). Thus, it is significant to detect the sensitivity of plants for mitigating the effect of DI.

The basil plant has been investigated in different studies to determine yield component, yield, composition of essential oil, essential oil ratio, plant densities and fertilization in diverse climate and soil conditions. However, few experiments have studied the effect of applied irrigation volume. Ekren et al. (2012) and Moeini Alishah et al. (2006) reported that water stress reduced yields and plant height of Purple basil, while it had a positive effect on the essential oil, proline and anthocyanin content. In contrast, Khalid (2006) indicated that irrigation with 75% of field capacity (FC) on both American basil and sweet basil resulted in the highest herb yield and essential oil content compared with other irrigation water levels (50 and 100% of FC). According to Bekhradi et al. (2015), many quality characteristics of Genovese variety and Iranian cultivars (Green and Purple) of basil during storage were unchanged by reduced irrigation. Gao (2015) indicated that DI improved water use efficiency, but had different effects on yield of basil plants. Daily irrigation with 75% of crop evapotranspiration had a higher yield (18% increase), while irrigation with 75% of crop evapotranspiration every six days had a lower yield (12% decrease) compared to irrigation with 100% of crop evapotranspiration.

Sensitivity of the basil plant to irrigation and water stress could be characterized by the CWP and yield response factor (Ky) (Pejic et al. 2017). Defined as the unit increment in crop yield per unit of water consumed, CWP can be calculated if the amounts of actual crop evapotranspiration and actual yield are known (Igbadun et al. 2006). Ky demonstrates the relation between a relative yield reduction (1–Ya/Ym) and a relative evapotranspiration reduction (1–ETa/ETm) (Doorenbos & Kassam 1979). A crop with a Ky value lower than one is known as tolerant to water deficit. In turn, when the Ky value is greater than one, the crop is considered as not tolerant to DI (Alomran et al. 2013).

Efficient use of water resources is needed due to the expected water scarcity in the face of climate change and growing competition for water resources between domestic, industrial and agricultural consumption (Gholami Zali & Ehsanzadeh 2018). Therefore, reducing the volume of irrigation water, while maintaining quality and yield, are considered desirable in regions where water availability is a main restriction. The aims of the present study were (i) to evaluate the effect of different DI levels on yield and CWP of Green and Purple basil cultivars under greenhouse conditions and (ii) to determine the sensitivity of two basil cultivars to water stress by Ky.

The experiment was carried out under greenhouse conditions at the Faculty of Agriculture, Ferdowsi University of Mashhad, Iran, during the spring and summer of 2017 and 2018. The study was located at 36°18′N latitude, 59°38′E longitude and 995m altitude. The culture of the basil plants was conducted in soil in a unit of the greenhouse with 6m width, 10m length and 3m height. LS16 equipment by AP Holland was applied to monitor the agrometeorological parameters. The temperature of the greenhouse was controlled to maintain at 25–30 °C in both study years. The temporal variations of daily and cumulative radiation during the growing season are presented in Figure 1. The soil in the experimental location was silty clay loam. The main physicochemical properties of the soil in the study area are shown in Table 1.

The experiment was a split plot layout as a randomized complete block design with three replications per irrigation treatment and cultivars as the main plot and the subplot, respectively. Each plot consisted of four rows of 90cm length and 25cm distance from each other. A space of 0.5m was considered between plots to minimize irrigation edge effects. In addition, aluminum plates with 2mm thickness were used between plots to prevent water leakage to adjacent plots down to a depth of 0.5m.

Irrigation amounts at DI0 treatment were represented in Table 2. The volume of irrigation water for DI30 and DI60 treatments was obtained by applying coefficients 0.70 and 0.40 to the obtained volume of irrigation water for DI0 treatment, respectively.

The volume of water input to each plot was measured for each irrigation using water volume meters.

Seeds of two basil cultivars, Green and Purple, were sown in seedling trays under control temperature of 25–30 °C on 19 March 2017 and 4 April 2018. The seedlings were transferred to the soil of a controlled greenhouse with temperature between 25 and 30 °C on 17 April 2017 and 1 May 2018. The seedlings in each plot were planted with a space of 10cm within the row and were irrigated without any water deficit until the basil plants reached an average height of 20–25cm and adapted to the soil conditions, and then irrigation treatments were applied.

Before soil tillage, farmyard manure (30 tons ha⁻¹) was applied to the experimental soil and mixed with the soil to a depth of 15–20cm in each experimental year. Considering the initial soil composition, plants were fertilized with a water-soluble 12-12-36 NPK fertilizer (2g per liter of water) about four and eight weeks after seedling stage in 2017 and eight weeks after seedling stage in 2018. Weeds were controlled by hand and no herbicides, pesticides, or chemical fungicides were applied during the growing seasons.

Harvesting of the basil plants was done at the beginning flowering stage. The harvest dates were July 3, August 8, September 6 in 2017 and July 3, August 2 and September 1 in 2018. In each plot after eliminating the side effect (eliminating two side rows of each plot), the basil plants in the two middle rows after eliminating one basil plant from the beginning and end of each plot were selected for the below measurements.

Fresh leaves of the basil plant represent a separate market from green tops and dried herbs. In this study, plant height, fresh herb yield, fresh leaves yield and dry herb yield were measured as follows:

Plant height (cm). Before each harvest, the plant height as vegetative growth indicator was measured from the ground to the tip of the highest leaf.

Fresh herb yield (kg ha⁻¹). The basil plants were harvested 10cm above the ground, and immediately weighed to determine the produced plot yield. Next, the plot yield was upscaled to one hectare land.

Fresh leaves yield (kg ha⁻¹). The leaves of the fresh herbs were isolated from the plant and weighed.

Dry herb yield (kg ha⁻¹). The separated leaves and stems were dried at 70 °C for 48h and thereafter weighed to obtain dry herb weight.

Precipitation amount was not included in the equation due to the experiment was conducted in a controlled conditions without allowing water to penetrate. The capillary rise was assumed to be zero because of the very deep level of groundwater in the study area.

The CWP for the fresh leaves (CWP-FL), fresh herb (CWP-FH), and dry herb (CWP-DH) were estimated by dividing the yield (kg ha⁻¹) to the total actual crop evapotranspiration (m³ ha⁻¹).

All collected data were statistically analyzed by analysis of variance (ANOVA) and the mean differences were compared by Fisher's LSD test (P≤0.05). The relationship between fresh leaves yield and water used by evapotranspiration was evaluated using regression analysis.

Analysis of variance indicated that irrigation treatments had a significant effect on the plant height only for the first harvest (Table 3). There was a significant difference between heights of two cultivars at all harvests except the first one (Table 3). Year and interactions of Y × I, I × C, Y × C and Y × I × C had no significant effects on the height of basil plant at any harvests (Table 3).

Averaged across the two cultivars for the first harvest, the maximum values of the basil height were observed at DI30 treatment as 66.91cm and 72.22cm in 2017 and 2018, respectively (Table 3). Also, for both the study years, DI30 treatment resulted in a maximum average plant height of 69.57cm (Table 4). However, decrease in plant height due to DI on the basil plant has been documented by Rhizopoulou & Diamantoglou (1991), Moeini Alisha et al. (2006) and Ekren et al. (2012).

Averaged across irrigation treatments, Purple basil height in the second and third harvests and average three harvests were more than Green basil height by 4.76%, 16.03% and 8.46% in 2017 and 3.18%, 5.88% and 2.11% in 2018, respectively (Table 3). As seen in Table 4, the average height of Purple basil for the two study years was higher than that of Green basil. The result was consistent with those reported by Goldani (2012).

The irrigation treatments, cultivar and year had significant effects on fresh leaves yield in all harvests except the first harvest for year effect (Table 5). There were significant interactions of Y × I, I × C, Y × C and Y × I × C on fresh leaves yield in all harvests except the third for interaction of Y × C and the first and second harvests for interaction of Y × I × C (Table 5).

Averaged across two cultivars, the highest fresh leaves yield in all harvests and their total in two experimental year except the first and second harvest in 2017 were obtained in DI0 treatment (Table 5). As seen from the fresh leaves yield average for the study years in Table 6, fresh leaves yield decreased with decreasing irrigation water applied; however, there was no significant difference between DI0 and DI30 treatments. Decreasing the amount of irrigation to 40% of the field capacity (DI60) significantly decreased fresh leaves yield compared to full irrigation at all harvests (Table 4). Total values of fresh leaves yield in the DI60 treatment decreased by 31.53% relative to the DI0 treatment (Table 6). Water stress leads to tissue water content losses, which decrease cell turgor pressure, thereby preventing cell elongation and division, causing plant growth reduction (Shao et al. 2007). Similar results for basil leaves yield reduction with decreased water availability, as well as for the harvest periods, were reported by Moeini Alishah et al. (2006) and Jose et al. (2016).

Averaged across irrigation treatments, the Green basil had the highest fresh leaves yield at all harvests in 2017 and 2018 (Table 5). As seen in Table 6, fresh leaves yield average of Green basil was higher than that of Purple basil in all harvests by 24.72%, 33.17%, 20.85% and 26.34% in the first to third harvests and their total, respectively. The fresh leaves yield was found to be higher than the results of Jose et al. (2016). It is known that the fresh yield of the basil plant is controlled by various factors such as plant density, fertilization, irrigation and genotype, and is moderately affected by the environmental factors (Penka 1978; Arabaci & Bayram 2004).

In the water stress conditions, the Purple basil cultivar mostly performed better than the Green basil cultivar. For example, with decreasing irrigation to the 40% of the field capacity, total fresh leaves yield of the Green basil cultivar decreased by 34.79% and 42.89% in 2017 and 2018, whereas the fresh leaves yield of Purple basil decreased by 8.13 and 32% in 2017 and 2018, respectively (Table 5).

The fresh leaves yield in the first experimental year was higher than that in the second year (Table 6). This can be attributed to difference in radiation in the two experimental year. In the first study year, the length of basil growing season was 142 days and the basil plants received 2370.88 MJ m⁻² radiation, while in the second year these values were 123 days and 2048.99 MJ m⁻², respectively (Figure 1). The longer growing season and higher cumulative radiation resulted in a significant increase in fresh leaves yield in 2017 due to increasing photosynthesis and growth.

The irrigation treatments, cultivar and year had significant effects on fresh and dry herb yield at all harvests (Tables 7 and 8). There were significant interactions between Y × I, I × C, Y × C and Y × I × C and on fresh and dry herb yields at all harvests except the first harvest for Y × I and Y × I × C, the first and second harvest for I × C and the third harvest for Y × C (Tables 7 and 8).

The fresh and dry herb yields increased in parallel to the volume of applied irrigation water (Tables 6–8). Averaged across two cultivars at all harvests except the second in 2017 and 2018, the highest values of fresh and dry herb yield were obtained in the DI0 treatment (Tables 7 and 8). For both the study years, DI0 treatment had the highest average values of fresh and dry herb yield at all harvests except the second, although the total value of fresh and dry herb yield in DI30 treatment had no significant difference with DI0 treatment (Table 6). In the second harvest, the basil yield increased with DI (Table 6). The basil dry weight reduced due to the exposure to drought. This could be due to a decrease in the chlorophyll content, and accordingly, photosynthesis efficiency, as stated by Castonguay & Markhart (1991), Viera et al. (1991) and Khalid (2006).

Average across irrigation treatments, the highest fresh and dry herb yield were observed for Green basil (Tables 7 and 8). Also, for both study years, Green basil showed higher yield than Purple basil (Table 6). A similar result was reported by Bekhradi et al. (2015). Various environmental factors such as water stress affect the aromatic plants growth (Burbott & Loomis 1969). In addition to the environmental factors, genetic differences can also affect the basil yield (Ekren et al. 2012).

The effect of DI on fresh and dry herb yield was significantly different depending on the genotype and year. In 2017, the fresh and dry herb yield of Green basil significantly reduced due to 40% DI, probably due to the large vegetative growth, while in the Purple cultivar, irrigation treatments did not have much effect on the yields (Tables 7 and 8). While, in 2018, DI reduced fresh and dry herb yield of both cultivars in I40 treatment compared to the control (Tables 7 and 8).

The fresh and dry herb yields were higher in the first experimental year compared to the second year (Table 6) depending on climate conditions and, next, plant vegetative growth.

Year and irrigation treatment had significant effect on the crop evapotranspiration (Table 9). No significant cultivar and interaction between Y × I × C for crop evapotranspiration were observed; however, there were interactions of I × C and Y × C on crop evapotranspiration (Table 9).

Averaged across two cultivars, crop evapotranspiration in DI30 and DI60 treatments declined by 42.25% and 64.76% in 2017 and 22.68% and 33.95% in 2018, compared to the DI0 treatment (Table 9). As seen from the total crop evapotranspiration average for the study years in Table 9, crop evapotranspiration decreased as the amount of irrigation water applied declined, with the maximum value observed in the DI0 treatment (12,696.71 m³ ha⁻¹).

DI0 treatment for the Purple basil cultivar had the highest crop evapotranspiration in the two years, while the lowest crop evapotranspiration was obtained in the DI60 treatment for Purple basil and the DI60 treatment for Green basil in 2017 and 2018, respectively (Table 9).

The crop evapotranspiration in the first experimental year was more than that in the second year (Table 10). The higher solar radiation in 2017 led to more crop evapotranspiration compared to 2018 (Figure 1).

CWP-FL was significantly affected by irrigation, cultivar, year and their interaction effects, while CWP-FH and CWP-DH was significantly affected by irrigation, cultivar and interactions between Y × I, Y × C and Y × I × C (Table 9). In general, in both years, CWP increased when the irrigation was decreased, although the highest CWP-FL in 2018 was obtained in the DI60 treatment (Table 9). The highest average CWP-FL, CWP-FH and CWP-DH for the study years were obtained in the DI60 treatment (0.84, 12.08 and 1.81kg m⁻³), respectively (Table 10). DI30 and DI60 treatments were in the same statistical group, with the value of a 0.84kg m⁻³ CWP-FL (Table 10). Increasing the CWP value while decreasing the amount of water applied in basil plant was reported by Agami et al. (2016). In contrast, in a previous study by Ekren et al. (2012), the irrigation water use efficiency was not affected by the irrigation water levels because yields obtained for unit irrigation water applied were relatively close to each other.

Regarding the cultivars, the CWP-FL, CWP-FH and CWP-DH were found to be higher in the Green cultivar with 0.87kg m⁻³, 10.31kg m⁻³ and 1.55kg m⁻³ in 2017 and 0.88kg m⁻³, 9.34kg m⁻³ and 1.43kg m⁻³ in 2018, respectively (Table 9). For both study years, Green cultivar clearly indicated the highest CWP-FL, CWP-FH and CWP-DH (Table 10).

In both cultivars, CWP-FL in 2018 increased with DI to 70% of the field capacity and then decreased from DI30 to DI60 treatment, while the CWP-GL in 2017 and CWP-FH and CWP-DH in both experimental years increased with DI in both basil cultivars (Table 9). In 2017, DI increased CWP-FL by 86.44% and 154.54%, CWP-FH by 143.52% and 186.32% and CWP-DH by 142.55% and 185.07% compared to control in I40 Green and Purple cultivars, respectively (Table 10). In 2018, decreasing the amount of irrigation water to 40% of the FC, increased CWP-FL by 18.68% and 9.37%, CWP-FH by 20.76% and 3.07% and CWP-DH by 20.63% and 40.71% compared to control for Green and Purple cultivars, respectively (Table 10). Based on the results, Purple basil was more sensitive to the applied water amount in the first study year. CWP-FL in the second experimental year was higher than the first experimental year, due to lower crop evapotranspiration in 2018 (Table 10).

Evaluation of the effects of crop evapotranspiration on fresh leaves production of two basil cultivars was performed by regression analysis (Figure 2). A linear relationship was observed between crop evapotranspiration and the fresh leaves production of two basil cultivars for two years of the experiment (Figure 2). The coefficients of determination between the fresh leaves yield and crop evapotranspiration were similar for the two basil cultivars in 2018 (r² = 0.87), but the coefficient for Purple basil was higher than Green basil in 2017 (r² = 0.81 and r² = 0.99 for Green and Purple basil, respectively) (Table 11). For the two basil cultivars, both the values of intercepts and the slope of the regression lines varied widely over the two years and a steeper slope of regression line was observed in 2018 (Table 11). The slope, which indicates the increase in fresh leaves production for each unit increase in crop evapotranspiration were 0.29kg m⁻³ and 1.27kg m⁻³ for Green basil and 0.33kg m⁻³ and 0.85kg m⁻³ for Purple basil in 2017 and 2018, respectively (Table 11). The higher slope means a higher efficiency of applied water in fresh leaves yield production of the basil plant. The difference in the slope of regression line between two experimental years is probably due to differences in weather conditions. Our results showed a similarity to the results reported by Payero et al. (2006), who indicated the relationships between maize yield and seasonal crop evapotranspiration are inconsistent and change with location. Combining data for 2017 and 2018, the graph of fresh leaves production versus crop evapotranspiration for Green and Purple cultivars indicated a deviation from linearity to a second order relationship (Figure 2). Namely, crop evapotranspiration maximized fresh leaves yield up to the point when additional crop evapotranspiration did not lead to additional yield, and more irrigation water applied to the basil plant reduced the fresh yield due to the adverse effect of the excess water on the basil plant. The coefficients of determination were 0.82 and 0.84 for Green and Purple cultivars, respectively (Table 11).

For each of the study cultivars, good linear relationships between relative water consumption and fresh leaves production were obtained in both years (Figure 3). According to Doorenbos & Kassam (1979), the slope of the regression line in Figure 3 denotes the Ky. The Ky values were 0.50 and 1.40 for Green basil and 0.65 and 1.08 for Purple basil in 2017 and 2018, respectively. The higher Ky value (i.e. steeper slope of the regression line) leads to a greater reduction of fresh leaves production for a specific reduction in evapotranspiration due to water deficits (Kresovic et al. 2016). The experiment was carried out under controlled conditions in the greenhouse and it was expected that Ky for two experimental years would be close to each other. However, the differences in fresh leaves production between the years (Table 6) were the result of different crop response to water, which in turn relates to irrigation regime and radiation conditions (Figure 1). According to the Ky values for two cultivars (Table 12), in the first study year, Purple basil was more sensitive to the water deficit while in 2018, Green basil was more sensitive. It seems that in favorable radiation conditions, Green basil is the more appropriate option for applying DI than Purple basil, while in radiation conditions similar to 2018, Purple basil performs better. Based on the average Ky value for both years (Table 12), Green and Purple basil showed similar responses to the water stress, with Ky values of 0.70 and 0.76 for Green and Purple basil, respectively. Since the Ky values were lower than one, the basil plant can be considered as a plant tolerant to water deficit. The obtained Ky value is in agreement with those reported by Saeedinia et al. (2019) for Satureja hortensis. However, our results showed lower values than the value reported by Pejic et al. (2017). They obtained a Ky value of 0.22 for basil plants, due to rainy weather conditions during the growing season.

Throughout this study, DI was examined for the two basil cultivars. It was found that by increasing the harvest number, basil production did not show a considerable decrease, and even some increased production was monitored. Therefore, it seems logical to have three harvests of basil plant even under water deficit treatments in greenhouse conditions.

The growth of plant cells is the most important process that is affected by water shortage (Erken et al. 2012). When there is a reduction in the growth of cells, we can recognize the size of plant through the smaller size of leaves or decrease in plant height (Hsiao 1973). In this study, deficit irrigation had a greater effect on the leaves production than the height of the basil plant. Only in the first harvest, the height of basil plant at DI30 treatment was significantly higher than the other irrigation treatments. Basil plant yields decreased and CWP increased depending on the water deficit levels; however, there was no significant difference between fresh leaves production and fresh and dry herb production under DI0 and DI30 treatments. As a result, DI at 70% of the FC was more effective in saving irrigation water, along with a good marketable yield of the basil plant compared to 100% of the FC treatment and furrow irrigation practice by farmers in the greenhouse conditions.

A polynomial relationship was established between fresh leaves production and consumed water, but crop Ky showed a linear relationship between the relative reduction in water applied vs. the relative reduction in yield, with an average value of 0.73. According to the average Ky value for both experimental years, basil plants, especially the Green cultivar, when grown in greenhouse conditions, can be considered as a water-stress tolerant crop.

The results of the study could be used as a good platform for greenhouse basil producers to optimize the use of irrigation water.

This work was supported by the Ferdowsi University of Mashhad, Iran, under Grant number 45280.

The authors declare no conflict of interest.

All relevant data are included in the paper or its Supplementary Information.