Sowing Date and Genotype Inﬂuence on Yield and Quality of Dual-Purpose Barley in a Salt-Affected Arid Region

: Dual-purpose barley is an alternative approach to producing high-quality forage yield plus an acceptable grain yield in marginal environments of arid regions that are characterized by lack of forage. Field experiment was performed in two consecutive growing seasons at an arid region affected by salinity in irrigation water and soil at Western Sinai Peninsula in Egypt. The study aimed to optimize sowing date and screen salt-tolerant barley genotypes that perform better in terms of forage yield and quality as well as grain and biomass yield production in salt-affected environment. Sowing dates, genotypes, and their interaction signiﬁcantly impacted most of the studied variables such as forage yield, crude protein yield, and grain and biomass yields. The early sowing in late October yielded higher than intermediate sowing in mid-November and late sowing in early December. Some of the tested genotypes performed better than others as indicated by about 50% higher forage yield, 6% crude protein content, 39% grain and 21% biological yields (total aboveground dry matter), suggesting higher adaptation capacity. Interestingly, grain and biological yields did not differ signiﬁcantly between dual-purpose approach and grain-only pattern. In conclusion, dual-purpose barley was found favorable for producing grain and forage production in similar environments under early sowing date.


Introduction
Large agricultural areas in the Mediterranean region are affected by salinity [1]. Indeed, salinity is one of the most important abiotic stresses that destructively affect crop production, particularly in arid and semi-arid regions [2,3]. The arid environments are characterized by water shortage, frequent drought, dry seasons, high climatic variation and different forms of land degradation [4][5][6]. The recent climate change is characterized by decreasing precipitation and increasing temperature, resulting in greater aridity. These conditions increase soil salinity due to salt accumulation in surface layers as a result of higher soil evaporation rates [7].
Barley (Hordeum vulgare L.) is one of the main winter cereals in the Mediterranean region and it has the advantage of growing in marginal environments that are unsuitable for other cereal crops. It ranks fourth in terms of cereal acreage and total production after wheat, maize and rice [8]. Its major uses are livestock or poultry feeding and raw product for malt and starch production as well as for human consumption due to its substantial health benefits and content of dietary fiber [9,10].
Cultivation of cereals for both forage and grain yields during the same growing season is defined as s dual-purpose management system. Plants are grown for grain production but clipped at vegetative stage, then left to redevelop and produce grains [11]. Wheat, barley, triticale, and oat are commonly cultivated for dual-purpose in the Mediterranean basin [12][13][14]. However, barley is advantaged by its tolerance to salinity and therefore producing higher biomass under salinity stress compared with other cereal crops [15][16][17][18]. In addition, it is characterized by rapid re-growing after grazing while maintaining similar yield levels to that of un-grazed ones [19].
Mediterranean regions, especially arid and semi-arid environments, are characterized by low rainfall which makes year-round maintenance fodder production challenging and expensive [12,[20][21][22]. The issue is further intensified in salt-affected areas as most forage crops are sensitive to salinity stress. One of the feasible solutions is the cultivation of dualpurpose barley which can be grazed or harvested at tillering stage for forage production and then re-grow for grain production [23,24].
The success of dual-purpose barley in marginal environments is subject to proper agronomic management practices along with the use of improved genotypes. Adaptation to climate change by adjusting sowing date and using improved genotypes can mitigate the negative effects of climate change on barley production [16,21,22]. Sowing date is considered as one of the major agricultural practices that control barley production, particularly when planted as a dual-purpose crop. Several reports emphasized the importance of optimizing sowing date for dual-purpose barley in order to successfully produce high forage and grain yields [25][26][27]. Another crucial factor is growing site-specific genotypes that are more adapted to the surrounding environment in the context of genotype by management by environment interaction [28][29][30][31]. This was highlighted in previous research that documented different responses of barely genotypes to management practices with respect to their dual-purpose ability in diverse climate conditions [32][33][34][35]. Consequently, it is vital to investigate sowing dates of different cultivars under a dual-purpose system to identify optimal management for producing high forage and grain yields. For that reason, the current study was performed to assess the effect of sowing date on forage yield, forage quality, grain yield, and related traits of various barley genotypes with different genetic background grown in salt-affected sandy soil in the Mediterranean environment.

Experimental Site and Agricultural Practices
Field experiment was performed at Ras-Sudr experimental station, Desert Research Center, Western Sinai Peninsula, Egypt (29 • 35 59 N, 32 • 42 05 E) during 2018-2019 and 2019-2020 growing seasons. The location is described as semiarid with very low precipitation (annual precipitation about 50 mm), unavailable freshwater, while groundwater is largely affected by salinity. The average monthly temperature and monthly cumulative precipitation during the two growing seasons as well as long-term average of 38 years are shown in Table 1. Soil samples were collected to determine soil chemical properties (four samples per replication) before sowing from 0-30 cm soil depth in both growing seasons ( Table 2). The soil is sandy loam throughout the profile (86.95% sand, 8.75% silt and 4.30% clay) with a slightly alkaline pH of 8.15. Electrical conductivity of soil was estimated using EC 1:1 method by adding 100 mL distilled water to 100 g oven-dried soil and the mixture was shaken for 30 min [36,37]. After preparing the extract, the EC was determined using conductivity meter. Average electrical conductivity of soil water extract (1:1) of 7.74 ds m −1 and irrigation water is also affected by salinity (8.35 ds m −1 ).  Phosphorus fertilizer was added during seedbed preparation at rate of 31 kg P ha −1 as super-phosphate (15.5% P 2 O 5 ), potassium fertilizer was applied at rate of 60 kg K ha −1 as potassium sulphate (48% K 2 O) in one dose with the first dose of nitrogen fertilizer. Nitrogen fertilizer was performed with rate of 140 kg N ha −1 as ammonium sulphate (20.5% N).
Dual-purpose production system received additionally 50 kg N ha −1 . Weed control was applied as recommended for barley cultivation in the region [38], using post-emergence herbicides Granstar 75% DF [Tribenuron methyl (Sulfonyl urea) @ 750 g kg −1 ] for broadleaf weed and Axial XL (Pinoxaden and Cloquintocet-mexyl) for grass weed control. There was no relevant pest or disease pressure in the two growing seasons. According to the standard practice for this region furrow irrigation was used every week and it was cutoff two weeks before harvest (about 400 mm in total for growing season).

Plant Material and Experimental Design
Five commercial six-row barley genotypes (Hordeum vulgare L.) were used in this study: G-123, G-126, G-131, G-132, and G-2000 (Table S1). The experimental design was split-split plot in three replicates. The main plots were allocated for sowing date and three dates were tested. Sowing dates were 25 October, 14 November, and 4 December. The subplots were assigned for production systems which were: (i) only-grain (without-cutting) and (ii) dual-purpose with one forage removal at 45 days after sowing (DAS). Sub-sub plots were assigned for the evaluated five barley genotypes. Each experimental unit was 1.2 m wide (six 0.2-m spaced rows) and 3 m long. The seeds were planted at a density of 400 seeds m −2 .

Measured Traits
Plants in dual-purpose pattern were cut once after 45 DAS at 5 cm above soil surface level and fresh forage was weighed. Samples were weighted after oven drying at 72 • C for 48 h to estimate herbage dry yield. No forage removal was performed in only-grain system. Total nitrogen (N) was measured using micro-Kjeldahl method. Crude protein content was estimated by multiplying N percentage by 5.88 [39]. At maturity, plant height was measured as the distance in centimeters from the ground to the top of spike. The number of spikes was counted in 0.5 m 2 . The number of grains spike −1 was measured from ten randomly collected spikes at each plot. 1000-grain weight was recorded as weight of 1000 grains samples. Both grain yield and biological yield (total aboveground dry matter) were estimated by harvesting four central rows from each plot and then converted per hectare. The grain yield was adjusted to 12.5% moisture content.

Statistical Analysis
R statistical software version 3.6. was used for analyzing the data of this study. Differences among sowing date, production system, evaluated genotypes, and their interactions were separated by the least significant difference (LSD) at p ≤ 0.05. Principal component analysis was implemented on the averages of the evaluated traits to study their interrelationship.

Forage and Crude Protein Yields
Analysis of variance displayed highly significant effects of sowing date, genotype performance and their interaction on fresh forage, dry matter and crude protein yields (Table 3). Growing season effect was similar during both years of study. Early sowing on 25 October had the highest forage yield (fresh and dry) and protein content followed by intermediate (14 November) and late sowing (4 December) dates, respectively. Among the tested genotypes; G-123, G-126, and G-2000 possessed the highest forage yield and quality compared to G-131 and G-132. The interaction between sowing dates and evaluated genotypes was highly significant and the averages were presented in Figure 1. The maximum fresh forage yield was produced by G-2000 under early sowing date (16,977 kg ha −1 ) while the lowest fresh forage yield was recorded by G-131 under late sowing date (7452 kg ha −1 ), ( Figure 1A). Similarly, highest dry matter and crude protein yields were obtained by G-2000 followed by G-123 and G-126 under early sowing date whereas the lowest dry matter and crude protein yields were obtained by G-131 and G-132 under late sowing date ( Figure 1B,C).   FFY is fresh forage yield (kg ha −1 ), DMY is dry matter yield (kg ha −1 ), CPY is crude protein yield (kg ha −1 ), NS/m 2 is number of spikes per square meter, NG/S is number of grains per spike, 1000-GW is 1000-grain weight (g), PH is plant height (cm), GY is grain yield (kg ha −1 ) and BY is biological yield (kg ha −1 ). Means followed by different letters under the same factor differ significantly by LSD (p < 0.05).

Grain Yield and Its Components
Grain yield and its components were significantly affected by sowing date, production system, genotypic performance and their interactions ( Table 3). Number of spikes m −2 as one of the major yield components significantly differed in response to sowing date. The uppermost number of spikes was obtained by early sowing date on 25 October. Otherwise, the fewest number of spikes was recorded by late sowing on 4 December. Likewise, production system had significant effect on number of spikes. Obviously, dual-purpose produced higher number of spikes (238.9) than only-grain pattern (227.5) by 5%. The evaluated genotypes exhibited significant differences; G-2000 possessed the copious number of spikes (263.4) followed by G-123 (254.9) while G-131 and G-132 showed the fewest number of spikes (203.1 and 208.4). There was significant three-way interaction effects on grain yield and its component ( Table 3). The significant interaction effects resulted from the varied differences between production systems within sowing dates for different genotypes (Figure 2A). The highest number of spikes was assigned for G-2000 followed by G-123 and G-126 under early and intermediate sowing dates in dual-purpose system. Whereas the lowest number of spikes was presented by G-131 and G-132 under late sowing date in dual-purpose pattern.
Number of grains spike −1 is an important yield component and has a direct effect on barley grain yield, early or intermediate sowing dates produced significantly higher number of grains spike −1 (44.7 and 44.0) than late sowing date (41.8) ( Table 3).
On the other hand, production system did not influence this variable. However, the genotypes showed significant differences in the number of grains spike −1 . The genotype G-126 had highest number of grains spike  Figure 2B). Similarly, 1000-grain weight was significantly influenced by sowing date ( Table 3). The heaviest grains were produced by early sowing (42.3 g) compared to late sowing (40 g). Significant differences were also observed among genotypes in grain index (1000-grain weight). G-123 produced the heaviest grains followed by G-126, while G-131 gave the lightest grain index. The three way interaction results indicated that the heaviest grain index was obtained by G-123 under early sowing date in grain-only pattern (46.9 g), whereas the lightest grains was recorded by G-132 under late sowing date in dual-purpose system (37.6) ( Figure 2C). Grain yield was the result of combined effect for the above-mentioned three yield components. Noticeably, sowing date had significant effect on final barley grain yield (Table 3). Early sowing (25 October) exhibited the highest grain yield followed by intermediate and late sowing date. Production system had no significant difference in grain yield, therefore is strongly recommended for both grain and forage production in the study region and similar environments. The evaluated genotypes displayed greatly significant differences in grain yield; G-123 and G-2000 recorded the highest yield followed by G-126 ( Table 3). The three-way interaction displayed that G-2000 and G-123 exhibited maximum grain yield under early sowing conditions in both production systems. Otherwise, the lowest grain yield was displayed by G-131 and G-132 under late sowing in dual-purpose pattern, (Figure 2D).

Plant Height and Biological Yield
Plant height was significantly impacted by sowing date (Table 3). It was significantly depressed and plant had short vegetative growth under late sowing compared to early ones. Plant height decreased from 78.56 cm at early sowing to 67.05 cm at late sowing. In respect of production system, it had no significant effect on plant height. The evaluated genotypes displayed highly significant variation in plant height showing that G-2000 followed by G-126 and G-123 exhibited tallest plants. The three-way interaction displayed that G-2000 exhibited tallest plants under early sowing in both production systems followed by G-123 and G-126. The shortest plants were observed in G-131 and G-132 under late sowing in dualpurpose pattern ( Figure 3A). Similarly, biological yield significantly differed in response to sowing date ( Table 3). The maximum biological yield (8348 kg ha −1 ) was achieved by early sowing compared to late sowing (7037 kg ha −1 ) ( Table 3). The production system had no significant effect on biological yield. Otherwise, the genotypes displayed significant differences; G-123 and G-2000 recorded the highest biological yield compared to G-131 and G-132. In addition, it is evident from Figure 3B that the highest biological yield was obtained by G-2000 and G-123 under early sowing date in dual-purpose system, while the lowest biological yield was recorded by G-131 and G-132 under late sowing date in dual-purpose pattern ( Figure 3B).

Interrelationship among Evaluated Traits
The interrelationships between evaluated agronomic traits were estimated by principal components under different production systems. The first two principal components presented about 86.75 and 84.92% of variability under only-grain and dual-purpose patterns, respectively. Subsequently, the first two principal components were used to construct the biplot (Figure 4). The traits represented by parallel vectors or vectors close to each other revealed strong positive association. Under only-grain pattern, the evaluated traits could be divided into three groups ( Figure 4A). The first group included only 1000-grain weight. The second group comprised of number of spikes per square meter, grain yield, and biological yield. The third group contained number of grains per spike and plant height. Furthermore, under dual-purpose pattern the evaluated traits could be divided into two groups ( Figure 4B). The first group comprised of number of spikes per square meter, fresh forage yield, grain yield, and biological yield. The second group contained 1000-grain weight, number of grains per spike, crude protein yield, dry matter yield and plant height. In respect of the genotypes; the PCA separated the evaluated genotypes based on the investigated traits under different sowing dates. The genotypes; G-2000, G123, and G126 displayed the highest positive scores on PC1 under early sowing and were located on the far right of the biplot. On the other hand, G-131 and G-132 exhibited the highest negative scores and were placed on the far left of the biplot.  , NS/SM is number of spikes per square meter, NG/S is number of grains per spike, 1000-GW is 1000-grain weight (g), GY is grain yield (kg ha −1 ), BY is biological yield (kg ha −1 ), FFY is fresh forage yield (kg ha −1 ), DMY is dry matter yield (kg ha −1 ), and CPY is crude protein yield (kg ha −1 ).

Discussion
Results of the present research showed the potential of dual-purpose barley to provide high quality forage and grain yields in salt-affected regions. This was evidenced by the non-significantly different grain and biological yields between dual-purpose and grain only management systems (Table 3). Yield attribute variables did not show advantages of grain only management systems, which resulted in a similar yield with dual-purpose barley. These results confirm that dual-purpose management system is highly recommended for fresh production of forage yield and at the same time similar yield level of grain only systems in marginal environments affected by salinity stress. Barley has a short growth cycle that allows to set seed redeveloping and escaping heat and drought stress. In addition, cutting may have stimulated greater tiller numbers which has resulted in greater number of spikes per square meter in dual-purpose barley compared with grain-only system ( Table 3). These mechanisms might explain the successful application of dual-purpose barley in the present study, particularly under early sowing in October for all genotypes (Figure 2). Meanwhile, late sowing in December showed significant yield reduction by dual-purpose than grain only system. This might be due to not receiving the growing degree day (GDD) requirements at early growth stages which reinforced plants to hasten maturity. In addition, dual-purpose system required fifteen more days to reach physiological maturity than grain only irrespective of sowing date. In this context, Juskiw et al. [40] documented the association between GDD requirements of growth stages and grain yield in different barley genotypes.
The results favor the dual-purpose system over the grain-only system due to the invaluable forge yield by the dual-purpose system and, meanwhile, similar grain yield to the grain-only system. Variable findings are reported by previous researches on the impact of forage removal on grain yield of barley. Some have reported no yield reduction or even yield increase by forage removal compared with grain-only system [24,41]. This was attributed to lower incidence of lodging, rapid recovery of leaf area, and moderation of foliar disease. Harrison et al. [24] reported positive impact of forage removal under water deficit conditions to conservation of soil moisture and greater remobilization of stem reserves to grains. On the other hand, significant yield reduction by forage removal was observed which was attributed to the removal of the crop growing points and to a smaller photosynthetic surface after forage removal [42][43][44]. Nonetheless, despite less grain yield, the economic return of dual-purpose barley was considerably greater in a low forage production environment [44].
The findings of present research suggest early sowing in October for successful application of dual-purpose system. Late sowing shortens development phases, which accordingly reduces photosynthetic efficiency and source-sink relationship. In addition, late sowing might expose plants to heat stress in hot environments during the vulnerable growth stage of grain filling and therefore adversely affect yield-related traits and biological yield as reported in the present results. The negative impacts of late sowing were more pronounced in dual-purpose system and negatively affected forage and crude protein yields. These results agree with previous research findings where significantly higher plant height, yield attributes, and grain yield were reported under early sowing in the end of October than later sowing [45][46][47][48]. Negative impacts of late sowing on barley growth and grain yield and its components also were reported by Royo et al. [49]; Singh et al. [50]; Fayed et al. [25]; Farooq et al. [26], and Tahir et al. [27], agreeing with present results. Likewise, forage and crude protein yields tended to be gradually decreased due to delaying of sowing date from 25 October to 14 November or 4 December. Early sowing in late October exhibited maximum forage yield and crude protein content compared with late sowing date in early December. Such a result confirms that climatic conditions were less suitable for growth duration especially temperature degree under late sowing date. These results are in accordance with those reported by Singh et al. [50] and Choudhary and Chaplot [51] showing that early sowing in the end of October or beginning of November resulted in higher protein content and nitrogen accumulation in green fodder, grain and straw over late sowing. Besides, Salama [14] reported that cutting barley at early growth stages (45 and 55 DAS) caused significant increase in forage yield and crude protein content compared to late cutting at 65 DAS.
Barley genotypes respond differently in terms of grain yield, yield attributes, forage yield, and forage protein content, especially under environmental stress [16,30,31,48]. The obtained results revealed highly significant differences among evaluated barley genotypes under salinity stress. The genotypes; G-123, G-126, and G-2000 produced highest forage yield, crude protein, grain yield and its components as well as biological yield. Such differences were due to genetic variation in used genotypes and their interaction with environmental conditions [18,29]. These results are in accordance with those reported by Royo et al. [49]; Kaur et al. [52]; Sharma [32]; Hundal et al. [33]; Choudhary and Chaplot [51]; Hajighasemi et al. [44]; and Kandic et al. [35]. They demonstrated significant differences among dual-purpose barley genotypes in respect of forage yield, protein content, and grain yield.

Conclusions
According to the present findings, implementation of dual-purpose barley is a vital alternative approach to produce high quality forage yield plus acceptable grain yield especially under marginal environments in arid regions. Early sowing on 25 October caused significant increase in forage yield, crude protein, and grain and biological yields of barley compared to intermediate (14 November) and late sowing (4 December). Which outlines avenue for investigating further early sowing dates in future research on dualpurpose barley. Barley genotypes responded differently in terms of grain yield and its components, biological yield, forage yield and forage protein content. G-123, G-126, and G-2000 exhibited uppermost grain yield, yield components, forage yield, and crude protein and biological yield especially under early sowing conditions in dual-purpose approach. Further exploration of barley genotypes can identify other genotypes better adapted to this production system.