Weed Suppression by Cover Plants in the Amazonian

Leandro Amorim Damasceno, José Eduardo Borges Carvalho, Francisco Alisson Xavier, Ansselmo Ferreira dos Santos, Gerlândio Suassuna Gonçalves, Alan Ferreira Leite de Lima, Wildson Benedito Mendes Brito, Cláudio Luiz Leone Azevedo & José Ferreira da Silva 1 Federal Instituto of Amazonas, Eurinepé, AM, Brazil 2 Embrapa Fruticultura e Mandioca, Cruz das Almas, BA, Brazil 3 Federal Instituto of Mato Grosso, Juína, MT, Brazil 4 Federal University of Amazonas, Itacoatiara, AM, Brazil 5 Federal University of Amazonas, Manaus, AM, Brazil


Introduction
Brazil became the largest orange producer in 2017 ahead of the United States, in second place, and China, in third place.The production of this crop in December 2017 reached 18.7 million tons, produced in 629.8 thousand hectares (IBGE (Instituto Brasileiro de Geografia e Estatística), 2017).The main producing states are São Paulo and Minas Gerais.
However, among the Brazilian states with the greatest potential for citrus cultivation is the Amazonas, with a harvested area of 3,265 hectares and an average yield of 71,830 tons of orange (IBGE, 2017).This activity is favored by compensating prices of citrus fruits and suitable climate conditions for over the year.However, most of the production comes from small orchards, with little use of cultivation technologies, which contributes to a low average productivity.
Technological limitations and inadequate management of orchards are threats to the sustainability of the crop.Currently, the conventional system of citrus production in the Amazonas adopts practices that imply a high production cost and environmental risk (Coelho & Nascimento, 2004), mainly in tropical and subtropical conditions, in which the use of harrowing and herbicides applied in pre-emergence are frequent, resulting in soil exposure to the direct action of rain and wind.Compaction and erosion have motivated some field studies to assess weed management alternatives in the crop interrow, such as the use of cover plants to minimize losses, exploiting more rationally the natural resources (Carvalho, Paes, & Menegucci, 2001;Carvalho, Montebeller, Franco, Valcarcel, & Bertol, 2005).crop rotation, conventional tillage at sowing, and the use of competitive cultivars (Welch, Behnke, Davis, Masiunas, & Villamil, 2016).
However, many of these methods require a long time in the field, heavy machinery traffic, and intensive soil tillage, which can lead to soil compaction, erosion, and changes in soil nutrient and water dynamics (Welch et al., 2016).Thus, studies that aim at efficient weed control methods in a way that integrates ecological and economic concepts and that involve a rationalization of resources are needed.
One of these methods that have been shown to be effective is weed suppression by the use of cover plants.Sturm, Peteinatos, and Gerhards (2018) clarify that the ability of cover plants to suppress weeds is characterized by the competition for light, water, space, and nutrients.Their correct use provides ecological and economic benefits in the field, including nutrient recycling and soil erosion reduction.
The effects of the practice of using cover plants on crop productivity are numerous due to the improvement of soil physical, chemical, and biological conditions, reduction of pest and pathogen incidence, biological fixation of atmospheric nitrogen, and reduction of weed population (Castro & Lombardi Neto, 1992;Gelmini, Trani, Sales, & Victoria Filho, 1994;Gravena et al., 1998).
Recently, many studies have demonstrated the benefits of weed suppression, such as suppression in an effective and sustainable way and with a reduced use of herbicides (Uchino et al., 2012;Dorn & Van Der Heijden, 2015).In addition, it assists in increasing integrated weed management strategies and brings economic benefits to the producer (Eshel et al., 2015;Gfeller, Herrera, Tschuy, & Mirth, 2018).
Although the benefits of using cover plants depend on soil and climate conditions, in Brazil, a large part of the researches that deal with studies with cover plants is restricted to the South and Southeast regions, where climate conditions are very different from those found in the northern region.In addition, when considering only citrus cultivation itself, only one single research carried out by Lineares et al. (2008) with organic citrus grown in Florida, USA, could be found in the literature.These authors observed that cover plants such as radish, winter rye, and crimson clover provide and excellent weed suppression, with a more efficient control than other methods, including the conventional cultivation.
Therefore, studies that investigate the ability and adaptability of different cover plants on weed suppression aiming at its control in areas with citrus orchards in northern Brazil are needed, mainly in the Amazonas State, which has a small-scale cultivation and little investment in technological infrastructure.Considering that, this study aimed to assess the effect of different soil cover plants on weed suppression in an area cultivated with orange in Rio Preto da Eva, AM, Brazil.

Material and Methods
The experiment was carried out in a citrus commercial orchard located in Rio Preto da Eva, AM, Brazil, at the geographical coordinates 02°42′24.1″S and 59°26′02.6″W (Figure 1).The experiment was performed in an orchard of Pera orange (Citrus sinensis (L.) Osbeck) grafted on Rangpur lime (Citrus limonia (L.) Osbeck) with four years of age and planted at a spacing of 7.0 × 4.0 m. jas.ccsenet.Before sowing the species, the herbicide Paraquat (400 g ha -1 ) was applied in the orchard interrow with a boom sprayer.Subsequently, liming was carried out by applying 3.6 t ha -1 of dolomitic limestone with a 91% relative power of total neutralization and incorporating it with a harrow.Only the application of herbicide and liming was carried out in the control treatment.
Cover plant seeds were randomly distributed to the experimental unit at the beginning of the rainy season (January/2012) at the following densities: 120 kg ha -1 of jack bean, 20 kg ha -1 of forage turnip, 20 kg ha -1 of millet, and 16 kg ha -1 of brachiaria.For the treatment with millet + jack bean, 10 + 60 kg ha -1 of seeds were used.Each cover plant species was mixed with 4.16 kg of single superphosphate, being broadcast applied in the citrus interrow, followed by incorporation to the soil with a rake.In the control treatment, a mowing was carried out at the same time as the sowing of the cover plants in order to provide the same condition for all treatments.
For plant sampling, a sampler of 1 m 2 was randomly placed 2 times in the useful area of each plot at 90 days after sowing (DAS), the period in which all the cover plants were at the point considered adequate to be cut.The plant shoot inside the sampler was cut close to the soil, counted, identified, and taken to the laboratory.This procedure was also performed for the control treatment.After measuring the leaf area of the cover plants with a LI-COR 3050A area meter equipment, the plants were dried in a forced air circulation oven at 75 °C until constant weight.Density was determined in this same sampling, and dry matter of the weeds present in the treatments was determined using the cover plants.
The leaf area ratio (LAR), specific leaf area (SLA), and leaf area index (LAI) were calculated according to the data of leaf area and dry matter, as in Cairo, Oliveira, and Mesquita (2008).
The degradation rate of cover plants in the field was quantified by the litterbag method by using polyethylene bags.
The bags had a 5-mm mesh, which allowed the colonization by microorganisms and invertebrates in the cover material.An amount of 20 g of dry matter of cover plants was placed in each litterbag.These bags were then randomly distributed in each experimental unit on the ground of the interrow of the citrus orchard.The periods in which the litterbags remained in the field were 7, 14, 28, 56, 84, 112, and 140 days.At the end of each period, a litterbag was taken per plot and the residual cover material was dried in a forced air circulation oven at 75 °C until constant weight and then weighed.
The difference between the initial dry weight (20 g) and that obtained at the end of each period was used to quantify the loss of dry matter by decomposition over the assessed period.The decomposition rate of plant residues was estimated by the exponential mathematical model described by Thomas and Asakawa (1993), according to Equation 1: where, X is the remaining amount of dry matter after the period of time t (days), X 0 is the initial amount of dry matter, and k is the decomposition constant of the residue.
To determine the decomposition constant (k) of each treatment, the following neperian logarithm (In) was applied (Equation 2): The half-life time (T 1/2 ) of the remaining residues was calculated from the k value by using the expression proposed by Paul and Clark (1996), which expresses the time required for 50% decomposition to occur, as described in Equation 3: To quantify the effect of cover plants on the suppression of weed growth, the relative neighbor effect (RNE) was calculated according to Smith, Atwood, Pollnac, and Warren (2015) (Equation 4): where, P control is the weed biomass in the control treatment, P mixture is the weed biomass in the presence of a cover plant, and X is the P control if P control > P mixture or the P mixture if P mixture > P control .
Thus, values of RNE = 1 indicate a complete weed suppression by cover plants, RNE = 0 indicate a complete weed suppression by cover plants, and RNE = -1 indicate the facilitation of the weed by the cover plant.
The results were submitted to analysis of variance and the means of the treatments were compared by the Tukey's test (p ≤ 0.05) using the software Assistat (Silva, 2012).

Results and Discussion
All the assessed variables of the cover plants were significant (Table 2).The treatments with brachiaria, millet, and jack bean + millet did not differ from each other as they presented a higher stem weight, followed by jack bean and forage turnip.
Regarding leaf dry matter, jack bean presented a higher weight in relation to the other cover plants, with a value of 2.44 t ha -1 .The leaf dry matter of treatments with brachiaria, millet and jack bean + millet did not differ from each other at a 5% probability.The forage turnip presented the lowest weight of leaf dry matter.This result is similar to that found by Reis, Azevedo, Albuquerque, and Junior (2013), who also observed a low leaf dry matter production for forage turnip, when compared to the other species.This may be related to its growth cycle, which was early in this research and reached 50% flowering at 38 days after sowing and presented total senescence of leaves at 90 days.
The total dry matter production for brachiaria was 6.35 t ha -1 at 90 days (70.55 kg day -1 ha -1 ) after sowing, and did not differ from the other treatments, except for forage turnip.This value is lower than that obtained by Montanari, Carvalho, Teixeira Filho, and Dalchiavon (2013), who observed a value of 23.6 t ha -1 at 107 days (220.56 kg day -1 ha -1 ).The forage turnip presented the lowest total dry matter weight, with a value of 1.08 t ha -1 .This difference in growth per day may be due to the length of day, which was lower than the critical photoperiod, reducing the period of dry matter accumulation of forage turnip and brachiaria.
Faversani, Cassol, Piva, Minato, and Rocha (2014) assessed the soil cover rate of winter plants at 15, 30, 45, 60, 75, and 90 days after emergence in 2011 and 2012, in Pato Branco, PR, Brazil.The cover plants presented different performances.Forage turnip was the species that covered the soil faster, being superior in all the assessed dates, followed by vetch (Vicia villosa Roth), black oat (Avena strigosa Schreber), and forage pea (Pisum sativum L.).These results were due to the light exposure above the critical photoperiod, unlike the response found in our experiment.Halde, Gulden, and Entz (2014) conducted an experiment at the Morrison Research Station in Carman, Manitoba, in the Northern Great Plains of Canada, with six plant species: vetch (Vicia villosa Roth), forage pea (Pisum sativum L.), forage turnip (Raphanus sativus L.), sunflower (Helianthus annuus L.), flax (Linum usitatissimum L.), and barley (Hordeum vulgare L.).The authors found that the remaining biomass of the forage turnip presented the lowest value among all the plant species.The treatments jack bean, millet, and the mixture presented values of dry matter production of 5, 6.11, and 5.18 t ha -1 , respectively.The values obtained for the decomposition constant (k), half-life time (T 1/2 ), equations, and determination coefficients obtained from the regression analysis are shown in Table 3.The cover plants jack bean and millet underwent 50% decomposition of residues (T 1/2 ) at 44 days after the placement of bags in the field.Brachiaria presented a T 1/2 at 37 days.In a study conducted in Uberaba, MG, Brazil, during the dry season, Torres, Silva, Cunha, Valle, and Pereira (2014) observed higher T 1/2 for jack bean (52 days), millet (93 days), and brachiaria (70 days).These results show that climate directly affects the decomposition of cover crop residues on the soil.
Considering the spontaneous vegetation of the experimental area, the control treatment presented a much lower T 1/2 of 39 days when calculated up to 112 days after the management, since in this period the treatment still had the remaining dry matter in the field.
Jack bean + millet presented an intermediate half-life of 42 days.For the soil cover maintenance function, organic residues derived from grasses may present lower decomposition rates, and they are considered to be more interesting in the management plan.On the other hand, materials from legumes can recycle nutrients more quickly (Teixeira et al., 2009).In this research, the grass/legume combination can be considered an interesting option for the production system since it demonstrated a pattern of intermediate decomposition to single crops, while at the same time promoting the functions of soil protection and nutrient recycling in the medium term.
The treatment forage turnip presented a half-life of 28 days.The straw of forage turnip has a low persistence in relation to grasses.The degradation rate is directly related to humidity and temperature conditions that affect the activity of decomposing organisms, i.e., the higher the temperature and humidity are, the higher the degraded phytomass fraction (Hentz, Carvalho, Luz, & Barcellos, 2014).In addition, other factors such as the development stage of the species in which the management and the chemical composition of the cultivated plant material were carried out justify the observed decomposition rate (Vilanova et al., 2014;Hentz et al., 2014).
According to Cardoso, Tsai, and Neves (1992), most plant residues are composed of more or less complex polymers, which initially undergo enzymatic hydrolysis with an essential performance of microbial exoenzymes.
According to the ease of decomposition, some organic substances have a very short life in the soil, while others remain almost unchanged for long periods, such as high molecular weight phenolic compounds, which contain nitrogen and carbohydrates in their structure.Thus, succulent materials of young plants, such as cover plants, can undergo decomposition in a few weeks, as verified with some studied species.
Several authors, such as Teixeira et al. (2012) and Torres et al. (2014), state that climate conditions act directly on the soil microbial biomass, promoting the decomposition of plant residues and altering their decomposition rate, thus decreasing T 1/2 .

Conclusion
Brachiaria and jack bean presented the highest growth rates.
Cover plants suppressed weed growth in relation to the producer management, except for the forage turnip.
The highest decomposition rates occurred in the first 28 days, mainly for the forage turnip, which presented large losses in the initial phase and, after this period, a stabilization was observed in the decomposition rate of the remaining phytomass.
Brachiaria, jack beans, millet and their mixture presented longer half-life times, showing to be good alternatives for soil cover.

Figure
Figure 1.C

Table 1 .
Soil chemical characterization of the orchard in Rio Preto da Eva,AM, Brazil, 2012

Table 3 .
Decomposition constant (k), half-life time (T 1/2 ) in days, regression analysis, and determination coefficient of dry matter of cover plants in the orange orchard.Rio Preto da Eva,AM, Brazil, 2012