Effect of Deficit Irrigation and Root-Zone Drying Irrigation Technique under Different Nitrogen Rates on Water Use Efficiency for Potato (Solanum Tuberosum L.) in Semi-arid Conditions (I)

An investigation was carried out at the Technical Center of Potato and Artichoke CTPTA in the region of Saida, located in the lower valley of Medjerda river during the season of 2017. The objective was to evaluate the effects of deficit irrigation (DI) and the root-zone drying irrigation technique (PRD) under different nitrogen rates on total dry matter production (TDM), water consumption (WC) and water use efficiency of potato (Solanum Tuberosum L. VS. Spunta). Three water treatments (T1 = FI = 100% ETC, T2 = DI = 75% ETC and T3 = PRD50) and three nitrogen rates (F1 = N150: 150 kg N ha-1, F2 = N75: 75kg N ha-1, F3 = N0: 0kg N ha-1) were applied since the tuber initiation (55 days after planting) to maturity (100 days after planting). The results showed that the water regime affected negatively the total dry matter accumulation. A decline of 7 and 18.6% was registered in the two treatments T2 and T3 compared to the control T1. The WC decreased during water restriction respectively by 16; 33 and 29% for the T2 and T3 (PRD50 left) and T3 (PRD50 right) compared to T1. For the three nitrogen treatments (F1, F2 and F3) the water restriction has increased the WUE. The best values was recorded in the treatment T2 and then in the treatment T3 from where this increase compared to T1 was equal to (22.6% and 12.9%), (24.1% and 12, 4%) and (21.9% and 15.3%) respectively.


I. INTRODUCTION
In Tunisia, the potato occupies an important place in vegetable production, after tomato, watermelon and melon. It represents 12% of total vegetable production. In fact, the potatoes area cultivation has increased from 16800 ha in 1994 to 26200 ha in 2014, making 16% of the vegetables area, which amounts to about 370 thousand tons cultivated. Potato, have elevated water stores and supplemental irrigation is necessary for successful production. However, water availability for agriculture is being reduced as a consequence of global climate change. Therefore, great emphasis was placed on water management for dry conditions based on plant and crop physiology, with the aim of increasing water use efficiency. Deficit irrigation (DI) and partial root-zone drying (PRD) are two irrigation methods that attempt to decrease the agricultural demand for water. PRD is an irrigation technique where by half of the root zone is irrigated while the other half is allowed to dry out (Posadas et al., 2008). The PRD is quite simple, requiring only the adaptation of irrigation systems to allow alternate wetting and drying of parts of the root zone (Loveys et al., 2000). Numerous studies have shown a clear improvement in the water productivity by deficit irrigation (Intrigliolo and Castel. 2005;Goldhamer et al., 2006). In Tunisia, Ben Nouna et al. (2005) and Zairi et al. (2003) have shown that in the case of low and medium climatic demand, deficit irrigation in tomato and potato crops is possible without causing major yield losses. Sadras (2009), confirmed that the use of PRD irrigation and deficit irrigation have allowed a productivity increase of 82% and 76% respectively compared to conventional irrigation with a non-significant reduction in yield. Yang et al., (2011) studied for three years (2006)(2007)(2008), the water use efficiency of durum wheat under three different irrigation regimes of PRD (30,40 and 50% of the ETM). The results showed a significant improvement in water productivity from 12 to 71.4%. In fact, the use of PRD irrigation as a very ambitious method of water saving in arid regions of China. Liu et al., (2006) studied the effect of PRD irrigation on the yield and productivity of seasonal potatoes during the tuberization stage. The results indicated that PRD irrigation led to a significant reduction in yield compared to control with 37% less irrigation water. Shahnazari et al., (2007) indicated that PRD irrigation has improved by 20% the yield of marketable tubers (size 40 to 50 mm) compared to the control. It was concluded in the framework of this study, that PRD irrigation has saved 30% water irrigation and has improved water productivity in seasonal potatoes in the order of 60%. Similar results obtained on the potato and reported by Jovanovic et al, (2010) showed that PRD irrigation led in two successive years to irrigation water saving of the order of 33 and 42% compared to the witness. Several researchers have studied the water use efficiency of potato under PRD conditions (Liu et al., 2006;Shahnazari et al., 2007;Jovanovic et al, 2010;Ben Nouna et al., 2016). However, only few studies concern the effect under different nitrogen rates. Therefore, this study was undertaken to investigate the effects of deficit irrigation and PRD technique under different nitrogen rates on the water consumption and water use efficiency of potato.

II. MATERIALS AND METHODS Experimental Site
The experiment was carried out at the Technical Centre of Potato situated in the low valley of Medjerda river at Saida, Tunisia (10°EST, 37°N, Alt. 28 m), during the season 2017. The climate is semi arid. The average annual rainfall is about 450 mm, concentrated from December to April with irregular distribution. The soil had a clay-loam texture with 180 mm m-1 total available water and 2 g l -1 water salinity. The bulk density varies from 1.34 to 1.60 from the surface to the depth (Rezig et al., 2013a).

Estimation of Crop Potential Evapotanspiration (ETc)
The Reference evapotranspiration (ET0) was estimated by the software CROPWAT V 8.0 using the FAO-Penman-Monteith approach (Allen et al., 1998). In fact, the climatic data: (1) Daily Minimum and Maximum Temperatures (Tmin and Tmax); (2) Daily Relative Humidity (HR); (3) Wind Speed (V) and (4) Rainfall (P) were registered during the three growing seasons from 2009 to 2011 by automatic agro-meteorological station. The Crop Potential Evapotanspiration (ETc) was determined means the following equation:

ETc = Kc × ET0 (1) Plant Material and Experimental Design
Plant material consisted of one potato variety (Solanum tuberosum cv. Spunta). The potato planting was conducted on 02 March 2017 with a mechanical planter machine. The Planting density was 41667 plants ha -1 . The experiment covered two treatments (T: water regimes and F: nitrogen rates). T consisted of three water regimes (T1 = 100% ETC, T2 = 75% ETC and T3= PRD50). F consisted of three nitrogen rates (F1 = 150 kg N ha -1 , F2 = 75 kg N ha -1 and F3 = 0 kg N ha -1 ). At the beginning of the potato cycle (during the first stages) irrigation and fertilization were started without any difference between the treatments (with the exception of the F3 which did not receive nitrogen from the beginning), from which the crop was given 100% of the water needs and nitrogen requirements in a homogeneous way over the entire plot. The experimental protocol was started 26 April 2017 (55 DAP) at the stage of the initiation of tuberization to potato harvesting and they were irrigated by drip irrigation. The experimental design was Split Plot with 3 replications. The main factor is irrigation regime and the secondary factor is nitrogen rates.

Total Dry Matter Production (TDM) and Tuber Yield (Yd)
The observations were made on above-ground dry matter, tuber dry matter, total dry matter and tuber yield. The sampling was collected for growth analysis at 40, 56, 69, 85 and 96 days after planting Potato (DAPP). At each date, three plants of potato by plot were collected. All material was dried at 85 °C to constant weight and dry weight was measured. The tuber yield was achieved at 96 DAP. The experimental protocol has begun at 55 days after planting Potato (DAPP). For each start date of the protocol, the irrigation treatments (T1 = 100% ETC, T2 = 75% ETC and T3= PRD50) have the same soil moisture conditions. The rainfall observed during the experimental protocol is equal to 2 mm. For the experiment period, we recorded 15 irrigation events. In fact, FI treatment, has entirely received 352 mm. The T2 treatment, has completely received 306 mm. The two treatments T3 (Right side) and T3 (left side), have received the same dose and it was equal to 213 mm. Table 1. Rainfall (mm) and irrigation water depth (mm ha -1 ), for different irrigation treatments (T1 = 100% ETC, T2 = 75% ETC and T3= PRD50)

Theoretical Formulations Estimation of the Daily Water Consumption
The monitoring of soil moisture was carried out on experimental units in order to calculate the water consumption over the entire potato cycle. For the water content measurement, the TDR method was used. For this purpose, we have installed 9 probes in different depths (0-20; 20-40 and 40-60 cm) for each treatments. The initial water stock was determined by the TDR method up to 60 cm for the various experimental units. As well, In each test unit, soil samples were collected every 20 cm to 60 cm deep, and TDR measurements every 20 cm were also carried out to establish the calibration equation.
Water consumption is estimated with soil water balance equation as follows (Hillel, 1998):

WC = P + I + U + R -DW -DS (2)
Where P is the effective precipitation (mm), I is the irrigation (mm), U is the upward capillary flow into the root zone (mm), R is the runoff (mm), Dw was the downward drainage out the root zone (mm) and DS is the change of soil water stored in soil layer of 0-60 cm (mm). The upward and downward flow was estimated using Results indicated that the two items were insignificant at the experimental site. Runoff was also insignificant during the growing season. Soil water content was measured each month with gravimetrically method. Soil water content data were collected for every 20 cm interval in soil depth. Some measurements were added before and after irrigation.

Conversion Efficiency of Water Consumption into Dry Matter Production and Yield (WUE)
WUE of Total Dry Matter (WUETDM) and WUE of Tuber yields (WUEY) were calculated using the following equations:

WUE TDM (kg m -3 ) = TDM / WC (3) WUE GY (kg m -3 ) = GY / WC (4)
Where, WUE is the water use efficiency (kg m -3 ), TDM is the total dry matter production (kg), Y is the tuber yields (kg) and WC is the total water consumption over the whole growing season (m 3 ).

Statistical Analysis
The results were subjected to variance analysis of one factor by General Linear Model (GLM). This analysis was performed using SPSS 20.0 software. The ensemble was completed by multiple comparisons of means with Student Newman Keuls test (S-N-K).

Effect of Partial Root-Zone Drying Irrigation (PRD) and Deficit Irrigation (DI) on the Total dry matter (TDM).
The effect of three irrigation treatments (T1 = 100% ETC, T2 = 75% ETC and T3= PRD50) and the three nitrogen rates (F1 = 150 kg N ha -1 , F2 = 75 kg N ha -1 and F3 = 0 kg N ha -1 ) in the Total Dry Matter (TDM) of potato was given in figure 1. For the three water regimes (T1 = 100% ETC, T2 = 75% ETC and T3 = PRD50), the effects of three nitrogen rates (F1, F2 and F3) on TDM were represented in (Figure 1a,  1b and 1c). The result showed that from the 56 th DAP the TDM of the treatment (F1) was greater than the two treatments (F2 and F3). ANOVA analysis shows that nitrogen application significantly (P < 0.05) increased the TDM. The maximum values of potato TDM were achieved in the treatment F2 and the lowest in F3. In fact, for the three irrigations regimes (T1, T2 and T3), the maximum amount of TDM was recorded in the treatment F2 (958.4; 876.5 and 731.4 g m -2 ) and the lowest in the F3 (689, 683.2 and 666.4 g m -2 ). For the three nitrogen rates (F1 = 150 kg N ha -1 , F2 = 75 kg N ha -1 and F3 = 0 kg N ha -1 ), the effects of three irrigation regimes (T1 = 100% ETC, T2 = 75% ETC and T3 = PRD50) on the total dry matter (TDM) were studied (Figure 1d, 1e and 1f). The results showed that from the 56 th day after potato planting  2) showed that the daily water consumption of potato was variable during cropping season. Also, we observed that irrigation regimes (T1, T2 and T3) have affected the daily water consumption of potato during tuberization stage. The water consumption in the treatment (T1) was higher than that in the two treatments (T2 and T3). For the T1F1 (Figure 2.a), the curve which represents the daily evolution of ETR and ETC followed the same paces and was exactly superimposed. In fact, during the all growing cycle of potato the ETR and ETC was equal to 371 mm. Similarly, for the T2F1 treatment (Figure 2.b), the ETR and the ETC have the same rate until the 56 th DAP, and after this date the daily ETR becomes lower than the daily ETC. In detail, the cumulative ETR in T2F1 regime was 310 mm (from planting to harvest) and the cumulative ETC was 371 mm. In result, the deficit of water needs (ETR compared to ETC) was 61 mm and was observed from 56 DAP to 89 DAP. For the T3F1 treatment (Right side), the ETR and the ETC were equal (Figure 2.c) until the 56 th DAP, and after this date the daily ETR come to pass lower than the daily ETC. In circumstances, the cumulative ETR in T3F1 regime (Right side) was 248 mm (from planting to harvest). The deficit of water needs (ETR compared to ETC) was 123 mm and was registered from 56 to 67 DAP and from 78 to 100 DAP. For the T3F1 treatment (Left side), the ETR and the ETC were equivalent (Figure 2.d) until the 56 th DAP, and after this day the daily ETR becomes inferior to the daily ETC.
In indicate, the cumulative ETR in T3F1 regime (Left side) was 260 mm (from planting to harvest). In effect, the deficit of water needs (ETR compared to ETC) was 111 mm and was marked since the middle of the season (from 50 to 68 DAP and from 78 to 100 DAP). The results obtained showed that the water consumption of the potato crop under water restriction conditions decreased respectively by 16 and 33% and 29% for T2 and T3 (PRD

Effect of Partial Root-Zone Drying Irrigation (PRD) and Deficit Irrigation (DI) on Water Use Efficiency.
The effect of three irrigation treatments (T1, T2 and T3) and the three nitrogen rates (F1, F2 and F3) in the water use efficiency (WUE) of potato was given in Figure 3. The relation between the water consumption and the total dry matter production over all potato growing season and under the nine treatments is given in Figure 3. From these results, we observed toward different treatments that the total dry matter production increased linearly with the cumulative water consumption. The slope of this regression is the conversion efficiency of water consumption into total dry matter production (WUE). Similarly, we noted that, the highest amount of WUE was recorded in the T2 treatment (3.9; 3.7 and 2.9 kg m -3 ) and after that in T3 (3.5; 3.2 and 2.7 kg m -3 ). However, the least was recorded in the T1 treatment (3; 2.8 and 2.3 kg m -3 ). In detail, for the three nitrogen rate (F1, F2 and F3) the water use efficiency in T2 and T3 has illustrated respectively an increased of (23.1; 24.3 and 20.7%) and (14.3; 12.5 and 14.8 %) compared to T1. Statistical analysis showed that the WUE was significantly (P< 0.05) affected by irrigation doses (T1, T2 and T3) for three nitrogen rates. The water use efficiency of total dry matter production at harvest (WUETDM) and the water use efficiency of yield (WUEGY) of the nine treatments were exposed in Table 2.

IV. DISCUSSION
The effect of the deficit irrigation and partial root-zone drying irrigation technique (T1 = 100% ETC, T2 = 75% ETC and T3 = PRD50) under different nitrogen rate (F1 = 150 kg N ha -1 , F2 = 75 kg N ha -1 and F3 = 0 kg N ha -1 ) on the total dry matter production (TDM), the water consumption (WC), the potato yield (Y), the water use efficiency for total dry matter production (WUETDM) and the water use efficiency for potato yield (WUEY) were studied. correlated with the nitrogen application. Indeed, these results are in agreement with those found by Ali et al., (2009), Massignam et al., (2011) and Hamzei (2011), who observed that increased nitrogen improved the biological yield of crops. Similarly, MacDonald (2002) investigated the effect of different nitrogen doses on the yield of several varieties of durum wheat, and observed that TDM at the anthesis stage increased significantly with increasing nitrogen rate. Latiri-Souki, (1994) mentioned that the rate of nitrogen uptake during durum wheat development is linearly related to the development of total dry matter. Moreover, Cheikh M'hamed et al., (2014) showed that the application of nitrogen in N1 (150 kg N ha -1 ) and N2 (100 kg N ha -1 ) significantly (P <0.001) increased the MST compared to N3 (50 kg N ha -1 ) and N4 (0 kg N ha -1 ). Indeed, treatment N1 improved the MST relative to N3 and N4 under the three water regimes D1 (100% ETC), D2 (70% ETC) and D3 (40% ETC). Sanaa (1998) reported that controlled nitrogen fertilization (150 or 300 kg N / ha) increased the dry matter content of leaves and stems compared with the unfertilized control. The same result was reported by Golli (1992) who showed that by increasing nitrogen doses from (0 to 250 kg N ha -1 ), the cumulative aerial dry matter increased from (16.8 to 48.9 g plant -1 ). Also, Ben Ammar (2007) found that the biomass accumulated at the potato beginning cycle in the 3 nitrogen treatments was the same up to 58 JAP. Then, a difference was observed between the treatments N1 (15 t ha -1 ), N2 (30 t ha -1 ) and N0 (without addition of N). Hence, the maximum accumulated biomass was 0.62; 1.06 and 1.32 t ha -1 respectively at the N0, N1 and N2 treatments. Also, according to Shah et al. (2004) MST increased from 1442 to 2131 g m -2 in the treatment with a high nitrogen content compared to the treatments with low nitrogen content. The results (Figure 1d, 1e and 1f) showed that the water regime (T1 = 100% ETC, T2 = 75% ETC and T3 = PRD50) affected negatively the total dry matter accumulation. A reduction of 7 and 18.6% was recorded in the two treatments T2 and T3 compared to the control T1. Indeed, these results are in agreement with those of Rezig et al. (2015 a and b); Cheikh M'hamed et al. (2014 and 2015). These authors studied the effects of deficit irrigation in wheat and potato crops and showed that the water stress affected the accumulation of total dry matter. According to Rousselle et al. (1996), water stress also influenced tuberization by reducing the number of tubers formed. A water deficit during the tuber growth phase can influence vegetative growth and reduce leaf coverage irreversibly (Deunier et al. 1997). In addition, a lack of water causes the closure of stomata, which results in a rise in leaf temperature, a reduction in photosynthetic activity and consequently a decrease in yield (Rousselle et al., 1996).
Similarly, Ben Nouna et al. (2016) reported that for the four water regimes (FI, PRD80, PRD70 and PRD60) the dry matter production increases linearly with the cumulative water consumption. The obtained results in figure 2 showed that the water consumption decreased through water restriction respectively by 16; 33 and 29% for the T2 and T3 (PRD50 left) and T3 (PRD50 right) compared to with the control treatment T1. This feedings are in agreement with those of Rezgui et al. (2005); Rezig et al (2015 a and b) and Cheikh M'hamed et al. (2014 and 2015) for wheat. As well, for potato, Ben Nouna et al (2016) found that the cumulative water consumption was marked a reduction of 12.6; 17.9 and 25.1% respectively for the treatments PRD80, PRD70 and PRD60 compared to the control FI. These authors have shown that the water deficit significantly decreases water consumption. The results (figure 3) showed that the water restriction has led to an increase in the WUE at any rate of the nitrogen supplied (F1, F2 and F3). The best values was recorded in the treatment T2 and then in the treatment T3 from where this increase compared to T1 was equal to (22.6% and 12.9%), (24.1% and 12, 4%) and (21.9% and 15.3%) respectively for the three nitrogen treatments (F1, F2 and F3). Our results consent with those of Ben Nouna et al. (2016) who showed that PRD60 treatment improved the WUE of potato by (10.9 and 10.2%) and (25.8 and 19.7%) compared to full irrigation and PRD80. Also, Saeed et al. (2008) found that PRD irrigation used 29% less water and improved the WUE by 19%. Similarly, Xie et al. (2012) reported that the PRD50 treatment used 50% less water and increased the WUE by 48%. Several studies have indicated that deficit irrigation increased the WUE of wheat, potato, maize and rice. Indeed, moderate water stress induces partial stomatal closure, which would lead to an improvement in WUE (Turner, 1997). Ahmadi et al. (2010b) investigated the effect of PRD irrigation on seasonal potatoes, based on soil texture. They found that PRD irrigation resulted in a significant increase in water productivity in seasonal potatoes of 11% for sandy soil and 36% for sandy loam soil. Zairi et al. (2000) proved for wheat and potato that the better water productivity was obtained for the treatments carried out with the ETM. For potato, the consumption efficiency varied from 11.2 (a severe water stress) to 20.1 kg m -3 (an unstressed treatment). From the results in table 2, the variance analysis showed a significant effect at (5%) of water treatment on the WUETDM at harvest. Indeed, the T3 treatment presented the highest values of the WUETDM for the three nitrogen treatment. In fact, at the level of the T3 water treatment, an increase in the WUETDM of 15.7%, 9.8% and 7% compared to T1 for the three nitrogen treatments F1, F2 and F3 respectively. The water restriction led to an  (Table 2), the T2 showed a significant increase in the WUEGY, from which we have 6.67 kg m -3 compared to 5.49 kg m -3 in the T1. This is can be explained by a decrease of 16.26% in the water consumption in the T2 compared to T1. Although variance analysis showed no significant difference between the T1 and T3 treatments in the WUEGY, we note that, the decrease in water consumption has led to a significant reduction in yield which was equal to 36% compared to T1. Similarly, we observed that at the F2 nitrogen treatment, the T3 had the highest WUEGY, despite the 29.4% decrease in yield compared to T1, which is explained by the decrease in the water consumption from 370 mm in the T1 to 248 mm in the T3. For F3 treatment, variance analysis showed no significant differences at (5%) between water treatments. Indeed, the decrease in water consumption has led to a significant reduction in yield.

V. CONCLUSION
In the conditions of this study, we were able to assume, that the deficit irrigation (DI) and the partial root-zone drying irrigation technique (PRD) were two prospective water-saving irrigation strategies, especially for the drought sensitive crop such as potatoes and with restricted water conditions. During the all cropping potato cycle, the water restriction has improved WUE. The best values was recorded in the treatment T2 and then in the treatment T3. Similarly at harvest, for the two treatments F1 and F2, the highest (WUETDM) was registered under T3 and then in the T2. The lowest was observed under T1. However, for the treatment F3, the maximum WUETDM was registered in the T2 and after that in T3. Even so, for WUEGY statistical analysis showed significant (P < 0.05) difference between the three treatments (T0, T1 and T2) only under F1 and F2. For the treatment F1, the highest WUEGY was recorded under the treatment T2 followed by the treatment T1. Conversely, for the treatment F2, the maximum WUEGY was marked under the treatment T3 after that by the treatment T1. In turn, the use of DI and PRD with 50% of Etc from the initiation of tuberization stage to harvest is advantageous compared to full irrigation in terms of improving the water use efficiency.