Effect of Weed Treatment on Cereal Yield in Direct Seeding: A Challenge Between Soil Pollution and Seeds Quality 

Food security was synonymous with the supply of high-calorie staples such as cereals and tubers to resolve problems of protein-energy malnutrition (Sage 2019). During the first decade of the twenty-first century, cereal prices rose to their highest levels in real terms since the early 1970s, reaching a peak in 2008. In Algeria, wheat durum represents 46% of grain crops (Benbelkacem and Kellou 2000). Moreover, the peak of cereal imports reached 7.4 million tons in 2011 and 6.9 million tons in 2012 (Touchan et al., 2016). The adoption of conservation agriculture worldwide as a sustainable cultivation system is a challenge to increase productivity (Hobbs et al., 2008). Sustainable agriculture involves optimizing agricultural resources and at the same time maintaining the quality of environment and sustaining natural resources (Kumari Aruna et al., 2018). In India, direct seeding played a greater role to improve rice yield (Kumari et al., 2017).  It is considered as common practice before green revolution due to its potential to save water and labour (Gupta et al., 2006). In the other hand, 85% of the Brazilian soybean crop area was cultivated with no-tillage system to the expansion of soybean cultivation and for food security (Bohm et al., 2014). However, a rhythm of direct seeding adoption in Algeria is still very slow.
       
According to Rouabhi et al., (2018), no adoption of direct seeding is linked to technical and agronomic constraint as weeds control and proliferation of bromus. sp. Indeed, in the less developed areas of the world, the need for substantial increase in agricultural production is an urgent problem. On the other hand, direct seeding needs the use of agrochemical, so the increase in agrochemical use can be foreseen (Kumari Aruna et al., 2018). In direct seeding, the use of herbicides as “glyphosate” is the active matter; it will be imperative operation during first years of system adoption (Labad and Hartani 2016).
       
In the other ways, it was found that the use of glyphosate promoted high residual levels in soil and seeds (Bohm et al., 2014; Aruna Kumari 2018). Also, the residues of glyphosate were found in surface waters, suspended sediments and bottom sediments in Argentina (Aparicio et al., 2013; Lupi et al., 2015; Giaccio et al., 2016; Primost et al., 2017), rivers, rainwater and groundwater in the United States (Battaglin and al., 2014), soil and groundwater in Europe (Poiger et al., 2017; Silva et al., 2018) and even fish in Brazil (Moura et al., 2017). Hence, the present study was an attempt to direct seeding productivity along with herbicide accumulation in soil and seeds. The herbicide residues were present during three crop years (2014/2015- 2015/2016- 2017/2018).

A field experiment was conducted at the Technical Institute of Cereals (ITGC- Setif) during three crops seasons (S₁= 2014/2015, S₂= 2015/2016, S₃= 2017/2018). S₁ was considered as “reference season” to test protocol and study its feasibility.
       
The climate of studied region is semi- arid with a cold rainy winter and a hot dry summer. A climate variability recorded during experiment period is mentioned in Fig 1.
 

        
The experimental site which is characterised by moderately deep soil was cultivated under direct seeding for ten years before (Table1). Thus, it was active biologically (Labad et al., 2018). The experiment comprising of five weed treatments based on “glyphosate” as active matter. T₁: weed treatment by D₁ = 1080g ha^-1, T₂: weed treatment by D₂= 900g ha^-1, T₃: weed treatment by D₃= 720 g ha^-1, T₄: weed treatment by D₄= 540 g ha^-1 and T₅ is the control plot where no weed treatment was applied.  The barley was sown one week after weed treatment for all experiment seasons, using seeds rate of 120 kg ha^-1. A series of soil sampling were done to determinate variation of herbicide concentrations in H₁ and H₂. In addition, the amount of glyphosate in barley seeds was determined after harvesting and grinding as flour (Table 2). Yield was calculated by the weight of the seeds obtained in the plot and extrapolated to hectare. The glyphosate as phosphonomethyl glycine herbicide is unstable in the environment, so derivation step with FMOC-Cl is peremptory, before using HPLC- UV method (Peruzzo et al., 2008). The protocol was adapted in our laboratory conditions (Software, temperature, pressure...).
 

 

Barley yield variation
 
According to our results, the yield recorded in S₃ is more important than S₂ (Fig 2). Hence, the yield values depend on the herbicide doses applied before seeding for weed treatment. Raunet et al., (1998) found that the use of herbicides in direct seeding involves weed control, especially before crops seeding and at the beginning of its cycle. Under control soil sample and the lowest dose applied (D₄= 540 g.ha^-1), a significant decrease in yield was recorded (p< 0.05). Soil was affected by weed development. Singh et al., (2014) were reported that weeds are a serious constraint to the productivity causing 100 per cent yield loss under uncontrolled conditions. On the other hand, the highest yield responses to the highest dose of herbicide applied (1080g ha^-1) during S₂ and S₃. The average yield obtained during two crops seasons (S₂ and S₃) is 2.1t.ha^-1. Similar results were reported by Obour et al., (2016), where they recorded an increase in soybean yield applying a highest dose of glyphosate (840 g. ha^-1). Moreover, the yield variation was significant using D₁ and D₄ (P<0.05). These confirm that all yield parameters were affected by weed control treatment (Singh et al., 2015). 
 

 
Herbicide kinetics in the soil
 
Soil analyses done during S₁ showed that fractions of glyphosate reached soil during weed treatment by D₁ applied on December, 2014. After 319 days, herbicide was not totally degraded and concentration recorded in H₁ was 0.380 µg.kg^-1. The follow up of this concentration have continued in S₂ as control soil. Thus, four soil sampling were done on 373 days, 436 days, 476 days and 506 days. In control soil, herbicide concentrations decline to 0.267 µg.kg^-1 over a period of 506 days in H₁ (Fig 3). Otherwise, in H₂, the amount of glyphosate was under LQ (LQ= 0.264 µg.kg^-1).
 

        
The variability of glyphosate concentrations in H₁ during S₂ and S₃ from December to May (140 days) were given in Fig 4 (a/b). Glyphosate dynamic in soils depends on soil physical chemical and biological characteristics (Giesy et al., 2000; Duke et al., 2012).
 

        
The results showed that herbicide residues were more important in S₃ than S₂ and depend on the doses applied. Kinetics dissipation showed significant decline of herbicide concentration linked to high values of DT₅₀ (Table 3). DT₅₀ values calculated through SFO kinetics explain the persistence of molecule in the soil even using lowest doses, well half-live values of glyphosate found under field conditions were high compared to the results of literature (Grunewald et al., 2001).
 

        
On the other hand, a significant effect of rainfall on glyphosate in soil deep layer was observed analysing results of H₂ presented in Fig 5 (a/b). Herbicide concentrations transferred in soil deep layer via soil structure were more important in S₃ than S₂, when 442 mm of rainfall were recorded. Borggaard and Gimsing (2008), mentioned that soil with high macro porosity may increase the leaching risk, but only when a large precipitation occurs close to the application. Similar results were reported by Peruzzo et al., (2008) about significant effect of rainfall on glyphosate dissipation in the soil.
 

 
Seeds quality
 
The analyses of grains after harvesting showed significant negative relationship between doses applied and herbicide accumulation in grains in two crops seasons (Fig 6 a/b). It was found that the accumulation of glyphosate in barley grains is more important applying highest doses (D₁ and D₂). On other hand, glyphosate concentration was under LQ in grains harvested in soil sample without treatment in S₃. It is important to highlight that analyses of soil sample without weed treatment showed the values under LQ in S₃. On the other hand, the concentration of glyphosate in the soil has a significant effect on herbicide accumulation in grains (P< 0.05). High quantity accumulated varied between: 15.6 µg.kg^-1, 13.8 µg.kg^-1 for S₂ and 18.22 µg.kg^-1, 17.08 µg.kg^-1 for S₃. These results partially agree with Bohm et al., (2008), when the high residual levels of glyphosate were detected in soybean seeds after applying the recommended rate. Many authors explain high residual levels of glyphosate in grains by multiple factors as: soil and crop conditions, doses applied and season when glyphosate applications were performed (Busse et al., 2001; Araújo et al., 2003a; Duke et al., 2003; Reddy et al., 2004; Zablotowicz and Reddy 2007). On the other hand, Duke and Powles (2008) have explained accumulation of glyphosate on wheat seeds in relation with its systemic characteristics. When glyphosate is applied on the leaf surface, it will be relocated to the roots, stems and seeds. Seeds physiological quality is an essential factor for crop performance in the field.
 

During the study crop cycles and under field conditions, the following findings were tired:
 
•  The use of herbicides in pre- direct seeding for weeds management is indispensable to save cereal yields. 
•  Glyphosate as total herbicide used can reduce weed development even with low dose applied (D₄).  Nevertheless, to enhance productivity highest doses are required.
•  Highest doses applied involve important level of residues in soil surface, which they transferred in soil deep layer and accumulate in cereal seeds.
               
In addition, it is clear via our findings that low DT₅₀ value corresponds to the highest doses, because glyphosate can be degraded biologically, but transfer and accumulation phenomena persisted. For these reasons, further investigations are needed to manage weed treatment in direct seeding for safety environment.