Strategies for Improved Yield and Water Use Efficiency of Lettuce (Lactuca sativa L.) through Simplified Soilless Cultivation under Semi-Arid Climate

Simplified soilless cultivation (SSC) systems have globally spread as growing solutions for low fertility soil regions, low availability of water irrigation, small areas and polluted environments. In the present study, four independent experiments were conducted for assessing the applicability of SSC in the northeast of Brazil (NE-Brazil) and the central dry zone of Myanmar (CDZ-Myanmar). In the first two experiments, the potentiality for lettuce crop production and water use efficiency (WUE) in an SSC system compared to traditional on-soil cultivation was addressed. Then, the definition of how main crop features (cultivar, nutrient solution concentration, system orientation and crop position) within the SSC system affect productivity was evidenced. The adoption of SSC improved yield (+35% and +72%, in NE-Brazil and CDZ-Myanmar) and WUE (7.7 and 2.7 times higher, in NE-Brazil and CDZ-Myanmar) as compared to traditional on-soil cultivation. In NE-Brazil, an eastern orientation of the system enabled achievement of higher yield for some selected lettuce cultivars. Furthermore, in both the considered contexts, a lower concentration of the nutrient solution (1.2 vs. 1.8 dS m−1) and an upper plant position within the SSC system enabled achievement of higher yield and WUE. The experiments validate the applicability of SSC technologies for lettuce cultivation in tropical areas.


Introduction
The detrimental effects of climate change are resulting in dramatic environmental, economic, and social consequences across the world [1]. Current projections show an overall increase in temperatures, with rainfall being irregularly distributed and characterized by heavy downpours [2][3][4][5]. Erratic climate can negatively affect natural resources availability (e.g., water and agricultural land), as well as posing severe risks on both ecosystems and human health [2,6]. Many developing • Experiment 1 (NE-Brazil): The trial considered a curly green lettuce (cv Isabela). Conventional on-soil cultivation and SSC system were compared. Plants were transplanted on 24 June 2009 and were harvested when reaching full maturity, which occurred at 40 and 31 days after transplanting (DAT) for traditional on-soil cultivation and in the SSC system, respectively. The experimental design was a strip block design with two treatments and three replicates.

Simplified Soilless Cultivation System
The SSC system used was the so-called Bottles system (Figure 1), developed and tested in the northeast of Brazil since 2005 [18]. It is composed of a wooden/bamboo frame and a gravity-flow system, where nutrient solution drains from a tank of 310 L volume placed above the system at 2 m height. Hydraulic pipes with an emitter flow rate of 2 L h −1 direct the flux into the declined garden with a slope of 24%, which is composed by connecting plastic drinking bottles that host both substrate (rice husk in all experiments) and plants. The excess nutrient solution is then directed through a drainage pipe system to another tank placed below. A 50% shading net was placed above the system, to reduce light intensity. In NE-Brazil, the system used for the experiments was 6 m long and 3 m wide (18 m 2 ) and accounted for 20 lines of 2 L plastic bottles (8 bottles line −1 ). Each bottle had two holes for hosting plants. Therefore, at full regime, the system could accommodate 320 plants. In CDZ-Myanmar, the system was tailored to the local context to meet the vegetables production needs of individual households. Accordingly, the system size was reduced (5 m long and 2 m wide, resulting in a garden surface of 10 m 2 ), and a smaller tank for the nutrient solution (100 L) was adopted. Each module hosted 240 plants. When also considering the surrounding paths allowing for garden access (about half a meter on each side and in internal paths), the net planting density was of 26 plants m −2 in both NE-Brazil and CDZ-Myanmar.
In NE-Brazil, a nutrient solution (NS 1.6 ) previously adopted for local SSC cultivation was used [18,19]. The NS 1.6 was prepared with locally available simple mineral salts and soluble fertilizers and was characterized by an electrical conductivity (EC) of 1.6 dS m −1 and a pH of 6.5. In CDZ-Myanmar, for both experiments, the NS was prepared by using locally available NPK fertilizer (15-15-15). During experiment 2, the adopted nutrient solution presented an EC of 1.2 dS m −1 and a pH of 7.7, while in experiment 4 the nutrient solution was prepared at two concentrations, respectively, 0.6 g L −1 in NS 1.2 (EC = 1.2 dS m −1 , pH = 7.3) and 0.8 g L −1 in NS 1.8 (EC = 1.8 dS m −1 , pH = 7.5).
Details on macronutrient and micronutrient concentrations of nutrient solutions are reported in Tables 1 and 2. Agronomy 2019, 9, x FOR PEER REVIEW 4 of 14 Myanmar, the system was tailored to the local context to meet the vegetables production needs of individual households. Accordingly, the system size was reduced (5 m long and 2 m wide, resulting in a garden surface of 10 m 2 ), and a smaller tank for the nutrient solution (100 L) was adopted. Each module hosted 240 plants. When also considering the surrounding paths allowing for garden access (about half a meter on each side and in internal paths), the net planting density was of 26 plants m −2 in both NE-Brazil and CDZ-Myanmar. In NE-Brazil, a nutrient solution (NS1.6) previously adopted for local SSC cultivation was used [18,19]. The NS1.6 was prepared with locally available simple mineral salts and soluble fertilizers and was characterized by an electrical conductivity (EC) of 1.6 dS m −1 and a pH of 6.5. In CDZ-Myanmar, for both experiments, the NS was prepared by using locally available NPK fertilizer (15-15-15). During experiment 2, the adopted nutrient solution presented an EC of 1.2 dS m −1 and a pH of 7.7, while in experiment 4 the nutrient solution was prepared at two concentrations, respectively, 0.

Traditional on-Soil Cultivation
The soil of the two regions had a loamy sand texture with similar hydrological soil parameters (wilting point and field capacity at 6% v:v and 13% v:v, respectively). The physical and chemical characteristics of the soil in the two locations are described in Table 3. In both NE-Brazil and CDZ-Myanmar, the soil was overturned and dug with a hoe prior to cultivation. Soil fertilization provided a supply of 1.5 kg m −2 of cattle manure and 3.75 g m −2 of N, P, and K (mineral fertilizer 10-10-10 and Nitrophoska 15-15-15 in NE-Brazil and CDZ-Myanmar, respectively). Fertilizer was manually applied three days before transplanting. No additional fertilizer was applied during the crop cycles. Due to low soil pH in NE-Brazil, 0.15 kg m −2 of dolomitic limestone was added into the soil. The plots were raised by 20 cm and a trapezoid shape was developed, ensuring a base 1.2 m wide and a top 1.0 m wide. Finally, each plot was adjusted with a rake. Between the experimental plots, a space of approximately 0.7 m was left to facilitate maintenance, data collection and harvesting process.

Irrigation Management
In the SSC system, nutrient solution flux started early in the morning (at 7:00 am) and continued until dusk (6:00 pm). Three times per day (at 7:00 am, 11:00 am, and 3:00 pm), the drained nutrient solution was moved back to the upper tank. The daily nutrient solution consumption was calculated by the difference between the nutrient solution volume between the upper tank (at the beginning of the day) and the bottom tank (at the end of the day). The nutrient solution in the system was refreshed every day by adding new nutrient solution to a set level When plants were grown on the soil-based system, the irrigation management was different in NE-Brazil and CDZ-Myanmar experiments. In NE-Brazil, irrigation management was carried out based on the traditional local habit of the farmers by using manual irrigation. Water was distributed across experimental plots through manual labor, and a 12 L watering bucket was used. The amount of water distributed in a plot was based on farmers' experience. In CDZ-Myanmar, the irrigation management of soil-based treatments was based on crop evapotranspiration (ETc), restoring 100% of crop ETc by means of a drip irrigation system ETc was calculated by using the following equation (Equation (1)) where ET c (mm day −1 ) is the calculated crop evapotranspiration, ET 0 (mm day −1 ) is the reference evapotranspiration, and K c is the FAO crop coefficient for lettuce [20].
For the estimation of the reference evapotranspiration (ET 0 ), the Hargreaves-Samani (HS) equation (Equation (2)) was used, where ET 0 (mm day −1 ) is the reference evapotranspiration rate, T mean , T max and T min are the mean, maximum and minimum temperature ( • C) of the day, respectively, and R a (W m −2 day −1 ) is the extraterrestrial solar radiation [20]. The meteorological data for the determination of the reference evapotranspiration were daily downloaded from the website of the Agro-Meteorological Department of Yezin Agriculture University (http://www.yau.edu.mm/), located inside the university campus, excluding extraterrestrial radiation Ra that was calculated according to Duffie and Beckman [21].
The amount of water used for each irrigation was calculated based on plant water balance considering soil properties, root depth, and climate data (including rainfall, if any). Daily ETc was estimated considering the FAO crop coefficient for lettuce crop growth stages. Lettuce cycles were divided into three growth stages, and the Kc used was 0.7, 1.0 and 0.95, respectively. The time of irrigation was determined when readily available soil water (50% available soil water) was depleted.
Sixteen mm diameter drip pipes were used. Drippers had a flow rate of approximately 1.3 L h −1 , and each plant was supplied with a single dripper. A flow rate test and calculation of distribution uniformity (DU) were carried out before transplanting. The DU was calculated following the indications from Baum et al. [22]. Irrigation management (time and rate) was manually performed.

Plant Measurements
At harvest, plants were weighed to determine the fresh weight (g plant −1 ). Yield (kg m −2 ) was assessed by excluding external leaves which appeared damaged or wilted. Leaf number was also counted. Water use efficiency (WUE) was determined as the ratio between fresh weight and the volume of water used and was expressed as g FW L −1 H 2 O. In experiment 4, leaf stomatal conductance was also measured using a handheld photosynthesis measurement system model CI-340 (Camas, WA, USA), equipped with 6.25 cm 2 cuvette. Measurements were made at 27 DAT on the upper surface of the canopy on three leaves per each plant from 10:00 to 14:00 taking approximately one hour to complete each replication. All plants were measured on a single day. In the system, EC and pH were constantly monitored using a Combo pH/EC/TDS/Temp tester Model HI98130 (HANNA ® , Villafranca Padovana (PD), Italy). In experiments 2 and 4, the nutrient solution temperature was also monitored twice a week.

Statistical Analysis
Data were collected on 12 plants from the central part of each plot. Data from experiments 1 and 2 were analyzed using one-way ANOVA. Data from experiments 3 and 4 were analyzed by using twoand three-way ANOVA, respectively. Means were separated using the Tukey HSD test at p ≤ 0.05. Before the analysis, all data were checked for normality and homogeneity of variance. Averages and standard errors (SE) were calculated. Statistical analysis was carried out using R statistical software (version 3.3.2, package "emmeans" and "car").

NE-Brazil
During experiment 1, maximum air temperature ranged between 31.4 and 34.7 • C with an average of 33.0 • C. Minimum temperature ranged between 17.8 and 22.4 • C with an average of 19.8 • C. The daily relative humidity (RH) ranged between a minimum of 57% and a maximum of 97% (Table 4). Furthermore, 20.3 mm of effective rainfall occurred. During experiment 3, maximum air temperature ranged between 31.7 and 35.4 • C with an average maximum temperature of 34.0 • C. Minimum temperature ranged between 17.8 and 22.4 • C with an average minimum temperature of 19.8 • C. The maximum relative humidity (RH) was 97% and the minimum RH was 55% ( Table 4). The growing degree days (GDD) from transplanting to harvest ranged from 710 • C (experiment 3) to 920 • C (experiment 1).

CDZ-Myanmar
During the experiment 2, maximum air temperature ranged between 25.0 and 34.0 • C with an average of 31.5 • C. Minimum temperature ranged between 14.4 and 20.4 • C, with an average of 16.7 • C. The daily relative humidity (RH) ranged between a minimum of 48% and a maximum of 73% (Table 4). During experiment 4, maximum air temperature ranged between 30.7 and 38.6 • C with an average maximum temperature of 35.6 • C. Minimum temperature ranged between 16.0 and 23.0 • C with an average minimum temperature of 19.5 • C. The maximum relative humidity (RH) was 59% and the minimum RH was 30%. No rainfall occurred during the experiments. The growing degree days (GDD) from transplanting to harvest ranged from 662 • C (experiment 2) to 731 • C (experiment 4) ( Table 4).

Experiment 1-NE-Brazil
Lettuce yield was higher (+35%) in the SSC system, with a mean value of 2.3 kg m −2 , as compared to 1.7 kg m −2 achieved on soil (Figure 2a). This was mainly due to larger size of the leaves (data not shown), while leaf number was higher in plants grown on soil (Figure 2b). The increased yield was obtained with a daily water use (L m −2 d −1 ) approximately four times lower in the SSC system, as compared to conventional on-soil cultivation (1.8 vs. 7.5 L m −2 d −1 ) (data not shown). As a consequence, WUE in SSC system was 7.7 times higher, as compared to the conventional on-soil system, with mean values of 43.7 and 5.6 g L −1 H 2 O, respectively (Figure 2c).

Experiment 2-CDZ-Myanmar
During experiment 2 the average minimum temperature of the nutrient solution was 19.6 ± 1.64 • C while the average maximum temperature was 29.2 ± 1.69 • C. The average pH was 7.7, ranging from 7.4 to 8.1. The average EC was 1.28, ranging from 1.12 to 1.46 dS m −1 .
Yield (kg m −2 ) was increased by 72% (3.1 vs. 1.8 kg m −2 ) in SSC in comparison to the soil treatment ( Figure 2d) and the leaf number was significantly higher in soil-grown lettuce compared to soilless-grown plants (Figure 2e). Daily water use (L m −2 d −1 ) was approximately two times lower in the SSC system (2.66 L m −2 d −1 ) as compared to conventional on soil production (4.07 L m −2 d −1 ). WUE in the SSC system was found to be 2.7 times higher than that obtained with conventional cultivation, with average values of 37.1 and 13.7 g L −1 H 2 O, respectively (Figure 2f). consequence, WUE in SSC system was 7.7 times higher, as compared to the conventional on-soil system, with mean values of 43.7 and 5.6 g L −1 H2O, respectively (Figure 2c).

Experiment 2-CDZ-Myanmar
During experiment 2 the average minimum temperature of the nutrient solution was 19.6 ± 1.64 °C while the average maximum temperature was 29.2 ± 1.69 °C. The average pH was 7.7, ranging from 7.4 to 8.1. The average EC was 1.28, ranging from 1.12 to 1.46 dS m −1 .
Yield (kg m −2 ) was increased by 72% (3.1 vs. 1.8 kg m −2 ) in SSC in comparison to the soil treatment ( Figure 2d) and the leaf number was significantly higher in soil-grown lettuce compared to soillessgrown plants (Figure 2e). Daily water use (L m −2 d −1 ) was approximately two times lower in the SSC system (2.66 L m −2 d −1 ) as compared to conventional on soil production (4.07 L m −2 d −1 ). WUE in the SSC system was found to be 2.7 times higher than that obtained with conventional cultivation, with average values of 37.1 and 13.7 g L −1 H2O, respectively (Figure 2f).

Experiment 3-NE-Brazil
Considering the system orientation, significant differences for yield were found only in Veronica and Banchu cultivars, for which the west-oriented system showed a reduction in yield of 10 and 44%, respectively, as compared to the east-oriented one. In contrast, yields of cv Isabela and cv Mimosa

Experiment 3-NE-Brazil
Considering the system orientation, significant differences for yield were found only in Veronica and Banchu cultivars, for which the west-oriented system showed a reduction in yield of 10 and 44%, respectively, as compared to the east-oriented one. In contrast, yields of cv Isabela and cv Mimosa were not affected by the SSC system orientation (Figure 3). Daily water use was about 1.8 L m −2 d −1 , as for experiment 1. . Yield response to the simplified soilless system orientation (east, grey columns; west, white columns) in four lettuce cultivars (Isabela, Veronica, Banchu, and Mimosa). Vertical bars represent standard errors. Significant differences at p ≤ 0.01 (**), ns = not significant differences.

Experiment 4-CDZ-Myanmar
The average minimum temperature of the nutrient solution was 20.7 ± 1.1 °C while the average maximum temperature was 39.5 ± 0.87 °C. Daily water use was 2.50 L m −2 d −1 for NS1.2 and 2.23 L m −2 d −1 for NS1.8. The average pH was 7.3 and 7.5 for NS1.2 and NS1.8, respectively. pH ranged from 6.4-8.7 for the former, and 6.6-8.9 for the latter. Average EC was 1.25 and 1.83 dS m −1 for NS1.2 and NS1.8, respectively, ranging from 1.14-1.48 dS m −1 for solution NS1.2 and 1.59-2.06 dS m −1 for solution NS1.8 (data not shown). Results of analysis of variance in Table 5 show that the EC of the nutrient solution (EC), lettuce cultivar (Cv), and plants position (P) significantly affected plant morphological and productive parameters, as well as WUE and the crop physiological response.
Yield, stomatal conductance and WUE were affected by EC, Cv, and P-wherein a significant interaction between the three factors was noted-while leaf number was only affected by EC and Cv, with a significant interaction between the two factors ( Table 5). Yield of plants placed in the lower position (LP) was not affected by Cv and EC, while for both cultivars the plants in the upper position (UP) yielded more when NS1.2 was used ( Table 6). The yield of plants belonging to cv Thai and grown by using NS1.2 was four times higher, as compared to yield of Thai lettuce supplied with NS1.8 and placed in the same position within the system ( Table 6). The increased yield was mainly due to leaf number, as Thai plant grown adopting NS1.2 showed the highest number of leaves (12.9 leaves plant −1 ) while no differences were observed between the other treatments (data not shown). Stomatal conductance was highest (212 mmol m −2 s −1 ) in Thai lettuce grown on the upper part of the system by using NS1.2 (Table 6). For plants grown in the lower part of the system (LP), stomatal conductance was only affected by CV, and was higher in cv Thai for both considered EC (Table 6). Leaf temperature was only affected by the position (P) in the system (data not shown), and was lowest in plants grown on the top of the system (28.8 °C compared to 29.8 °C measured in plants grown at the bottom of the system). In cv Thai, WUE was highest in plants fed with NS1.2 and grown in the upper position (UP) of the system, while for cultivar Rapido 344 the only statistically significant difference was evidenced between plants grown on the upper position and fed with NS1.2 and plants grown in the lower position of the system and fed with NS1.8 (Table 6). Table 5. Results from the ANOVA on experiment 4 (CDZ-Myanmar). Effect of EC of the nutrient solution (EC), cultivar (Cv), and position within the garden (P) on lettuce yield, leaf number, and water use efficiency (WUE). Significant differences at p ≤ 0.05 (*), p ≤ 0.01 (**) and p ≤ 0.001 (***), ns = not significant differences.

Experiment 4-CDZ-Myanmar
The average minimum temperature of the nutrient solution was 20.7 ± 1.1 • C while the average maximum temperature was 39.5 ± 0.87 • C. Daily water use was 2.50 L m −2 d −1 for NS 1.2 and 2.23 L m −2 d −1 for NS 1.8. The average pH was 7.3 and 7.5 for NS 1.2 and NS 1.8 , respectively. pH ranged from 6.4-8.7 for the former, and 6.6-8.9 for the latter. Average EC was 1.25 and 1.83 dS m −1 for NS 1.2 and NS 1.8 , respectively, ranging from 1.14-1.48 dS m −1 for solution NS 1.2 and 1.59-2.06 dS m −1 for solution NS 1.8 (data not shown). Results of analysis of variance in Table 5 show that the EC of the nutrient solution (EC), lettuce cultivar (Cv), and plants position (P) significantly affected plant morphological and productive parameters, as well as WUE and the crop physiological response. Table 5. Results from the ANOVA on experiment 4 (CDZ-Myanmar). Effect of EC of the nutrient solution (EC), cultivar (Cv), and position within the garden (P) on lettuce yield, leaf number, and water use efficiency (WUE). Significant differences at p ≤ 0.05 (*), p ≤ 0.01 (**) and p ≤ 0.001 (***), ns = not significant differences.

EC of nutrient solution (EC)
*** *** *** *** Cultivar (Cv) ** *** *** * Position (P) *** ns *** *** EC × Cv * ** ** * EC × P *** * ** *** Cv × P ns ns ns ns EC × Cv × P ** ns * ** Yield, stomatal conductance and WUE were affected by EC, Cv, and P-wherein a significant interaction between the three factors was noted-while leaf number was only affected by EC and Cv, with a significant interaction between the two factors ( Table 5). Yield of plants placed in the lower position (LP) was not affected by Cv and EC, while for both cultivars the plants in the upper position (UP) yielded more when NS 1.2 was used ( Table 6). The yield of plants belonging to cv Thai and grown by using NS 1.2 was four times higher, as compared to yield of Thai lettuce supplied with NS 1.8 and placed in the same position within the system ( Table 6). The increased yield was mainly due to leaf number, as Thai plant grown adopting NS 1.2 showed the highest number of leaves (12.9 leaves plant −1 ) while no differences were observed between the other treatments (data not shown). Stomatal conductance was highest (212 mmol m −2 s −1 ) in Thai lettuce grown on the upper part of the system by using NS 1.2 (Table 6). For plants grown in the lower part of the system (LP), stomatal conductance was only affected by CV, and was higher in cv Thai for both considered EC (Table 6). Leaf temperature was only affected by the position (P) in the system (data not shown), and was lowest in plants grown on the top of the system (28.8 • C compared to 29.8 • C measured in plants grown at the bottom of the system). In cv Thai, WUE was highest in plants fed with NS 1.2 and grown in the upper position (UP) of the system, while for cultivar Rapido 344 the only statistically significant difference was evidenced between plants grown on the upper position and fed with NS 1.2 and plants grown in the lower position of the system and fed with NS 1.8 (Table 6). Table 6. Results from experiment 4 (CDZ-Myanmar). Effects of factorial combination of EC of the nutrient solution (EC, 1.2 vs. 1.8 dS m −1 ), cultivar (Cv, Thai vs EW) and position (P, upper position, UP vs lower position, LP) within the garden on lettuce yield, stomatal conductance (g s )and water use efficiency (WUE). Different letters indicate significant differences at p ≤ 0.05.

Discussion
The application of different cropping systems significantly affected yield, physiological response and water use efficiency of lettuce grown in both NE-Brazil and CDZ-Myanmar.
Water availability is one of the major constraints for agricultural development and food production. The first and second experiments aimed to determinate whether SSC lettuce production is a suitable and sustainable alternative to conventional on-soil production in both locations. Barbosa et al. [23], when comparing commercial (high-tech) hydroponic greenhouses against on-soil lettuce production, found that hydroponics could increase yield by 11-folds, thanks to improved nutrition and environmental control. According to our results obtained in both experiments, the use of a simplified (low tech) soilless system allowed increase in the yield of lettuce but to a lesser extent (+35% in NE-Brazil and +72% in CDZ-Myanmar, (Figure 2a,d)). Yield increase can be the result of higher planting density (26 vs. 13 plants m −2 , on SSC and on-soil cultivation respectively), fast plant growth and precocity of production (31 vs 40 DAT according to experiment 1) and the improved environmental conditions maintained within the SSC system, including plant nutrition, uniform and constant irrigation, as well as the shading cover integrated in the SSC system. According to Zhao et al., the adoption of a shading net as a cover for lettuce production in the summer season in Kansas led to a slightly lower daily maximum air temperature relative to the open field, with an average reduction of ≈0.4 • C [24]. Moreover, Zhao et al. reported that the shading net has a significant impact on soil temperature and leaf temperature [24]. Indeed, in comparison with open field conditions, when shading net is adopted, a considerable reduction of leaf surface temperature, by 1.5 to 2.5 • C, was observed [24], thus affecting the plants' capacity to absorb water and nutrients [25]. In the SSC system, the higher fertigation frequency probably affected production capacity. Silber et al. experimented on the effect of fertigation frequency on yield, water and nutrient uptake of lettuce [26], finding that high fertigation frequency (from 2 to 10 events a day) induced a significant increase (13-15%) in lettuce fresh weight (FW) [26].
Furthermore, SSC systems are also considered water-saving technologies, thanks to the capability to deliver water directly to the plant root [27,28]. Despite limited soil exploration by the shallow rooting system of lettuce, in NE-Brazil, when a conventional growing system was adopted, irrigation water was applied by means of a bucket or can on the entire soil surface and consequently a significant amount of water is lost through evaporation and percolation into the sandy soil. Increase in the use of low-flow and more targeted irrigation techniques, such as the adoption of a drip irrigation system, could lower the overall water use of conventional farming [23]. As a matter of fact, drip irrigation was used as a control treatment in the CDZ-Myanmar experiments. Accordingly, the adoption of an SSC system enabled a reduction of water use by 76% and 59% in NE-Brazil and in CDZ-Myanmar, respectively, as compared to on-soil production. The observed water savings are consistent with previous literature, e.g., when a SSC system was adopted in Colombia, water use was reduced by 90% as compared to the traditional on-soil cropping system [29].
A consequence of higher yield and lower water use was an increased WUE in the SSC systems. In NE-Brazil and CDZ-Myanmar, WUE was, respectively, 7.7 and 2.7-fold higher, as compared to conventional on-soil production (Figure 2c,f). Similarly, WUE for lettuce in hydroponics was previously found in the range of 2.9 g of dry mass per L −1 H 2 O [30], or 41 g of fresh mass per L −1 H 2 O [31]. Lettuce grown in high-tech hydroponic conditions showed a reduction in water use by 13-fold, as compared with traditional on-soil cultivation [23]. Under the expected climate change scenarios and water limitation for agriculture, SSC systems could be a valuable strategy to sustain highly productive agriculture where the adoption of high-technology systems is not affordable [3,5,28,32].
In the third experiment, the adaptability of four lettuce cultivars to two different garden orientations was addressed to have a deeper understanding of the SSC system management. It emerged that the response of plant growth to the garden orientation was cultivar dependent (Figure 3). Accordingly, Veronica and Banchu achieved a higher yield with the eastern garden exposure (Figure 3), which received a lower amount of solar radiation in the afternoon mitigating the high air temperatures. Wheeler et al. [33] indicated 23 • C as the optimal daily temperature for growing lettuce, a condition that is far below the mean temperatures observed in NE-Brazil during the experiment. Heat stress could also result in a greater osmotic stress caused by the nutrient solution, resulting in lower water uptake and reduced plant growth [27]. Moreover, due to elevated temperatures in the root zone environment, hypoxia may occur, inhibiting root respiration, mineral uptake and water movement into the roots [34].
In the fourth experiment, the growth response of plants grown in different positions within the system was addressed as a function of both plant cultivar and the nutrient solution concentration (Table 5). Accordingly, the highest yield was associated with Thai cultivar supplied with NS 1.2 (Table 6). Possibly, under the local climate, plants preferred a nutrient solution with lower EC, and the Thai cultivar better responded to reduced osmotic stress. Furthermore, yield was the highest when Thai cultivar was grown in the upper part of the system. ( Table 6). As reported by Gianquinto et al. [14], this aspect could be due to the increased temperature reached by the nutrient solution during the flow between the top and the bottom tank. It would suggest that by the time the nutrient solution reached the lower section of the SSC, it was significantly warmed up by irradiance in the plastic bottles, although this statement should further be confirmed by determination of nutrient solution and substrate temperatures in the different positions. Thompson et al. [35] showed that a 24 • C root temperature in hydroponic systems is the ideal temperature whereby lettuce growth can be maximized, even with elevated air temperature. Different studies also reported that high nutrient solution temperature depresses water and nutrient uptake through reduced oxygen availability, also affecting physiological processes such as root browning and active transport in membranes [36]. Moreover, it was also observed that high solution temperature might decrease nutrient concentration (particularly of N, K and Ca) in the root, which may ultimately decrease crop growth [25]. It should be further studied, however, whether this may be associated with increased nutrient solution temperature as water flows through the system, or with selective absorption of specific nutrients from those plants that receive the nutrient solution first. In this regard, it could also be considered to add additional hydraulic pipes with emitters in the middle of the system.

Conclusions
The study addressed the application of SSC technologies [37][38][39] for lettuce production in tropical wet and dry climates. Elevated potentialities, in terms of both yield increase and improved water use efficiency in comparison with traditional on-soil cultivation technologies, were evidenced in both locations. Furthermore, the study explored alternative crop management strategies evidencing differences in cultivar adaptability and potential productivity. For instance, garden orientation was shown to affect crop productivity on a cultivar-dependent basis. Finally, under the elevated temperatures that are locally experienced, it is advisable to reduce the concentration of the nutrient solution (with EC of 1.2 dS m −1 providing better results than EC of 1.8 dS m −1 ). Interestingly, the yield was also improved when plants were located in the upper positions of the garden. Government and local support services could influence the future of soilless farming, as subsidies could be used to offset the relatively high initial cost of SSC infrastructure. We conclude that a simplified soilless system could become one of the efficient strategies for contributing to sustainably feeding the world's growing population, especially in challenging areas such as the north east of Brazil and the central dry zone of Myanmar.