Dry matter accumulation and nutrient cycling by soil cover plants in an intensive corn silage production system

This study aimed to evaluate dry matter accumulation and nutrient cycling by soil cover plants in an intensive corn silage production system. At the beginning of October, a study consisting of five soil cover plants, plus a control treatment (fallow area vegetation) was installed. The following cover plants were used: brachiaria brizantha, brachiaria ruziziensis, crotalaria juncea, jack bean and velvet bean. At the beginning of February of the year following sowing, we evaluated dry matter and nutrient accumulation in plant shoots. There was a significant effect at 0.1% of the cover plants on dry matter and nutrient accumulation. The average values for dry matter accumulation in plant shoots were 19.88; 18.06; 16.38; 9.66; 8.96 and 4.26 t ha -1 for brachiaria brizantha, crotalaria juncea, brachiaria ruziziensis, jack bean, velvet bean and fallow area vegetation, respectively. The highest nitrogen accumulation was observed in crotalaria juncea (343 kg ha -1 ), followed by jack bean (269 kg ha -1 ). For brachiaria brizantha and ruziziensis, N accumulation of 247 and 216 kg ha -1 were found, respectively. Brachiaria brizantha had the highest accumulation of phosphorus and potassium, with average values of 39 and 362 kg ha -1 . In the fallow area vegetation, the accumulations of N, P and K were 42; 9.0 and 62 kg ha -1 , respectively. sulfur in of cover phosphorus accumulation in the shoot kg ha Brachiaria brizantha by crotalaria juncea brachiaria ruziziensis. the fallow legumes brachiaria cover plants increases of 220 430% in phosphorus cycling. percentages of potassium cycling were for brachiaria brizantha, crotalaria juncea and brachiaria ruziziensis the fallow area Pittelkow et al. be plants capacity to cycle potassium. potassium decumbens


Introduction
Corn silage has been the most used forage for animal feed in dairy farms with medium to high animal productivity (Oliveira et al., 2017b). In recent decades several technologies have been developed by research centers, universities and private companies aiming to increase the yield of corn crops, reduce the cost of production, recycle waste from dairy production, improve silage quality and preserve the environment (Costa et al., 2014;Freitas et al., 2021 andOliveira et al., 2021b).
In areas of intensive production of corn for silage, very little shoot plant material remains on the soil surface after the harvest of forage. In addition to the soil in these areas being unprotected, the intense traffic of machines also contributes to the degradation of the physical properties of the terrain, especially the sealing of the soil surface and the increase in soil density (Oliveira et al., 2017a;Oliveira et al., 2021b andSilva et al., 2021). In several crops intended for silage production, corn is sown without tilling and mulching with straw is one of the factors responsible for the success in direct seeding, especially in seedling emergence and in the early stages of corn development. Mulch regulates soil temperature, reduces erosion and soil sealing, increasing rainwater infiltration and water availability, thus facilitating the emergence of corn seedlings, resulting in more uniform crops with greater yield potential (Pittelkow et al., 2012;Oliveira et al., 2021b andThomé et al., 2021).
Numerous plants have been used for soil cover. Grasses stand out in terms of biomass yield and legumes because of biological nitrogen fixation from the atmosphere. Among the most common grasses used as mulch for soil cover are brachiaria, millet, Mombasa grass and forage sorghum. Among the most used legumes are crotalaria juncea, pigeon pea, jack bean, and velvet and black velvet beans (Haskel et al., 2020;Oliveira et al., 2021a;Oliveira et al., 2021b andSantos et al., 2021). Therefore, the aim of this study was to evaluate dry matter accumulation and nutrient cycling by five soil cover plants and a control treatment (fallow area vegetation) in an intensive corn silage production system in the spring and summer.

Methodology
The study was conducted in a rural property which intensively uses forage to feed dairy cows. The property is in the city of Mercês (Latitude 21º11'39"S and Longitude 43º20'29"W), located in the Zona da Mata de Minas Gerais, Brazil ( Figure   1). According to Köppen classification, the climate of the study area is tropical of altitude with rainfall in summer, and average annual temperature of 18 °C, with variations between 24 °C and 13.8 °C. The average annual rainfall is approximately 1,200 mm, with a water surplus from October to April. The relief varies from flat to gently undulating. The soil was classified as a medium-textured Dystrophic Red-Yellow Latosol. In September, soil samples were collected in the 0-20 cm and from 20-40 cm layers to evaluate the need for lime and gypsum application. Due to the area use history, especially to previous applications of dolomitic limestone and gypsum, the soil had a base saturation of 63% at 0-20 cm and low aluminum saturation at 20-40 cm (Table 1), requiring no correction of soil acidity or gypsum application, according to the recommendation of Oliveira et al. (2007), Raij (2011) and Oliveira et al. (2021a).
At the beginning of October, the study was installed shortly after the first rainfall events. It consisted of five soil cover plants and a control treatment (fallow area vegetation). The following plants were used for soil cover: brachiaria brizantha, brachiaria ruziziensis, crotalaria juncea, jack bean and velvet bean. The experimental design was randomized blocks with five replications, and the plots consisted of five furrows of five meters long, spaced 0.50 meters apart.  The sowing densities were 15 kg of viable pure seeds per hectare for brachiaria, and 25, 100 and 70 kg ha -1 for crotalaria juncea, jack bean and velvet bean, respectively. Green manure seeds were not inoculated as studies by Chada & De-Polli (1988), Barretto & Fernandes (2001), and  in rural properties, sugar mills and research areas of federal universities showed that inoculation of legume seeds with nitrogen-fixing bacteria did not increase nitrogen intake in the soil-plant system.
Chemical control of leafcutter ants was necessary during the growth of soil cover plants. At the beginning of February of the following year, dry matter accumulation was evaluated in the shoots of the soil cover plants and fallow area vegetation.
The plants were cut close to the soil surface, weighed and a subsample of each plot was put through forage chopper.
Subsubsamples of these plants were dried in a forced ventilation oven at 65 °C to constant weight. These subsubsamples were passed in a stainless-steel mill and analyzed for nitrogen, phosphorus, potassium, calcium, magnesium and sulfur contents, following procedures described by Malavolta et al. (1997) and Silva & Queiroz (2002).
Based on the values of the samples and chemical analysis, the accumulation of dry matter and nutrients in shoot biomass of the soil cover plants and spontaneous vegetation were calculated. The results of nutrient concentration in the shoots and dry matter and nutrient accumulation were submitted to analysis of variance and the means compared by the Scott-Knott test at 5% probability (Ferreira, 2011).

Results and Discussion
Firstly, we present and discuss the results of nutrient concentration in the shoots of the soil cover plants and fallow area vegetation. Table 2 shows the analysis of variance for nutrient concentration in the shoots of the cover plants and the vegetation of the fallow area. There was a significant effect of plants, at 0.1%, on the concentration of all the nutrients assessed in our study. The highest coefficients of variation were observed for nitrogen and magnesium concentrations. Several factors can influence the concentration of nitrogen, whose uptake and metabolism are strongly influenced by other elements, such as phosphorus and chlorine (Malavolta et al., 1997) and (Oliveira et al., 2021a). However, in addition to this factor, there was possibly high variability of organic matter or even nitrate contents in the area. Oliveira et al. (2018) evaluated the variability of soil fertility, nutritional status and yield of a sugarcane field managed homogeneously. In the phase of maximum growth of the sugarcane and at the time of its harvest, the authors performed visual evaluations of the uniformity of sugarcane, aiming to implement the study in the first regrowth cycle. At this time, ten georeferenced samples of the soil, leaf +3 and yield were taken in the center of the plot of about 5.0 hectares. For soil organic matter, the coefficient of variation was 31.93%, with values ranging from 8.0 to 28.0 g per dm 3 of soil, with an average of 19.8 g per dm -3 . Table 1 -Mean squares of analysis of variance and coefficient of variation (CV) for concentrations of nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg) and sulfur (S) in the shoots of the soil cover plants and fallow area vegetation.

Source of
Variation GL Still regarding the evaluations carried out by Oliveira et al. (2018), large variations in soil fertility, nutritional status and yield were found even with the sugarcane regrowth being visually uniform and having been managed homogeneously. In this study, the variation of Mg content in the soil was also high: 34.95%, with minimum, average and maximum values of 0.80; 1.16 and 2.20 cmolc dm -3 , respectively. According to Montezano et al. (2006), the spatial variability of some soil properties, expressed by the coefficient of variation, may be less than 10%, while for others, it exceeds 1,000%. Warrick and Nielsen (1980) classified variability into three levels: low (CV lower than 12%, average (12 to 52%), and high (above 52%).    Lima et al. (2010), the coefficient of variation for nitrogen contents in the shoot biomass was 8.6%, slightly lower than those obtained in this study (  Costa et al. (2014) also reported potassium concentration values of 13.0 g kg -1 for brachiaria ruziziensis, which is practically the same concentration found in this study.

Accumulation of dry matter and nutrients in the shoots of soil cover plants
The chemical composition of the cover plants, especially nitrogen contents, carbon-to-nitrogen ratio (C:N ratio), lignin, cellulose, hemicellulose and polyphenol contents, associated with environmental factors such as temperature, humidity and aeration, are variables that interfere in the mineralization of plant material (Alovisi et al., 2021;Oliveira et al., 2021a). In situations where soil cover plants are intended to be a quick source of nutrients for the succeeding crop, faster mineralization is desirable. However, in most areas of direct seeding of corn, a longer time of permanence of straw on the soil is recommended.
In this case, more important than nutrient concentration is the mass of recycled matter and nutrients or accumulation in plant shoots (Oliveira et al., 2021b;Silva et al., 2021 andOliveira et al., 2022). Table 4 presents the analysis of variance for the accumulation of dry matter and nutrients in the shoots of the soil cover plants and fallow area vegetation.
Another interesting study related to brachiaria ruziziensis is that of Oliveira et al. (2017a) who evaluated the production and chemical composition of brachiaria ruziziensis cultivated after silage of corn from the first crop, in place of another sowing of corn. The authors found the average shoot biomass accumulation of brachiaria ruziziensis was 11.84 t per hectare, with a coefficient of variation of 11.34%, even with the water and thermal restrictions of the second crop period. This is because rainfall from February to June (growing season of brachiaria ruziziensis) was 282 mm, and the rainfall events of March and April equaled approximately 60% of the total. Thus, the results of Costa et al. (2014), Oliveira et al. (2021b) and Santos et al. (2021) support the need for local studies to evaluate the yield potential of soil cover plants.
Juncea crotalaria also stood out in terms of shoot dry matter accumulation, with an average of 18.06 t per hectare.
When crotalaria juncea is sown in early spring, shoot dry matter accumulation ranges from 14 to 21 t per hectare (Mascarenhas et al., 1994;Lima et al., 2010;Teodoro et al., 2011 andOliveira et al., 2021a). However, crotalaria juncea is highly sensitive to nyctoperiod (Haskel et al., 2020 andOliveira et al., 2021a) and toxic soil aluminum, making it demanding in terms of soil fertility, and soil acidity should be corrected before cultivation (Ernani et al., 2001;Meda, 2003;Rast et al., 2010 andOliveira et al., 2021a). Oliveira et al. (2021a) compared the accumulation of dry matter and nutrients in juncea crotalaria cultivated in soils with medium fertility to those of low soil fertility. The average shoot dry matter accumulation of juncea crotalaria was 14.5 t per hectare in plots with medium soil fertility, while only 5.7 t per hectare in those with low fertility soil. Thus, in the soil of medium fertility, the accumulation of dry matter in the shoots of crotalaria juncea was equivalent to 254% of that found in the soil of low fertility. In soils of medium fertility, both the absence of exchangeable aluminum and the availability of phosphorus and basic cations contributed to the greater accumulation of dry matter in the shoots of crotalaria juncea, supporting the citations of Ernani et al. (2001) and Meda et al. (2003), which classified crotalaria juncea as one of the legumes most sensitive to exchangeable aluminum.
In the group of plants with the third highest shoot dry matter accumulation were the jack bean and velvet bean, with an average yield of 9.3 t per hectare, surpassing the vegetation of the fallow area by about 220%. In one of the studies cited by Oliveira et al. (2021a), yield potential of six green manures (crotalaria juncea, pigeon pea, jack bean, velvet bean, black velvet bean and spontaneous vegetation) was evaluated. Crotalaria juncea had higher accumulation of dry matter and nutrient cycling compared to the other green manures. In the average of the two years of study, crotalaria juncea accumulated about 15 t of dry matter per hectare in the shoots, statistically higher than the others. These results support the observations of Oliveira et al.
(2021a), who reported higher biomass yield of crotalaria juncea, compared to other green manures. Pigeon pea had the second highest dry matter accumulation, on average 10.5 t ha -1 . Jack bean and velvet bean accumulated close to 8 t ha -1 , not differing from one another. For the vegetation of the fallow area, there was an average accumulation of dry matter close to 5 t per hectare.
In relation to nitrogen accumulation, jack bean had the second highest accumulation, being surpassed only by crotalaria juncea, which had an average accumulation of 350 kg per hectare, values close to those reported by Oliveira et al. (2021a). For brachiaria ruziziensis, velvet bean and brachiaria brizantha, nitrogen accumulations in shoot biomass ranged from 216 to 247 kg per hectare, surpassing the vegetation of the fallow area by about 200 kg per hectare (Table 5).
Nitrogen accumulation in the shoots of brachiaria is possibly largely cycling of soil nitrogen, perhaps even from deeper layers. However, for legumes, biological nitrogen fixation from the atmosphere is responsible for a large part of the nitrogen accumulated in the plant. Oliveira et al. (2021a) mentioned that for sowing carried out from early October to early November, nitrogen accumulation in the shoot biomass of crotalaria juncea oscillated around 300 kg ha -1 , confirming the observations of Perin et al. (2004) and Duarte Júnior and Coelho (2008). Of the total nitrogen accumulated in the shoot biomass of crotalaria, about 60% to 70% originated from the symbiotic associations of legume roots with the atmospheric N2fixing bacteria, resulting in the contribution of significant amounts of this nutrient to the soil-plant system (Perin et al., 2004), thus contributing to greater sustainability of the subsequent crop (Mascarenhas et al., 1994;Lima et al., 2010 andHaskel et al., 2020). For comparative purposes, ammonium sulfate, one of the most used nitrogen fertilizers is cited. In 100 kg of this fertilizer, there is 20 kg of N. So, to obtain 200 kg of N, we would need to use 1,000 kg of ammonium sulfate.
The average phosphorus accumulation in the shoot biomass ranged from 9.0 to 39.0 kg ha -1 . Brachiaria brizantha had the highest, followed by crotalaria juncea and brachiaria ruziziensis. Compared to the vegetation of the fallow area, the inclusion of legumes or brachiaria as soil cover plants resulted in increases of 220 to 430% in phosphorus cycling. High percentages of potassium cycling were also observed for brachiaria brizantha, crotalaria juncea and brachiaria ruziziensis compared to the fallow area vegetation. According to Pittelkow et al. (2012) and Oliveira et al. (2021b), brachiaria have shown to be plants with high capacity to cycle potassium. In a study conducted in soil with potassium content of 98 mg dm -3 , Santos et al. (2021) found K accumulation by brachiaria decumbens of 250 kg ha -1 60 days after emergence, reaching 630 kg ha -1 at 105 DAE. In medium fertility soil, Oliveira et al. (2021b) found potassium accumulation in shoot biomass of brizantha brachiaria ranging from 385 to 468 kg ha -1 , with an average value of 418 kg of potassium per hectare.
No plant stood out in isolation in terms of calcium, magnesium and sulfur accumulation. Jack bean and velvet bean accumulated calcium the most, with average values of 70 kg per hectare. For magnesium, there was emphasis on crotalaria juncea, which was also the plant that accumulated sulfur the most. Compared to the vegetation of the fallow area, the brachiaria and legumes increased nutrient cycling by at least 300% for calcium, 220% for magnesium and 240% for sulfur. In crops of the corn hybrid BM3066, Oliveira et al. (2017b) found average forage production of 59.5 t of natural matter, about 20 t of dry matter, per hectare. In these crops, the average accumulation of nitrogen, phosphorus, potassium, calcium, magnesium and sulfur were 245; 36; 195; 31; 25 and 24 kg ha -1 , respectively. According to Oliveira et al. (2021b), unlike the other elements, potassium has a high release rate from straw or mulch. However, only in a situation of financial difficulty should the farmer deduct potassium recycled by the soil cover plants in fertilization, which emphasizes that return fertilization or fertilization based on the yield expectation has provided excellent financial and agronomic results in the production of corn for silage.

Final Considerations
In the edaphoclimatic conditions in which this study was conducted, and based on results of a single agricultural year, brachiaria brizantha, crotalaria juncea and brachiaria ruziziensis showed the highest accumulation of dry matter in plant shoots.
Crotalaria juncea had the highest accumulation of nitrogen, followed by jack bean, with values that exceeded 260 kg ha -1 , though much of this nitrogen originates from the biological nitrogen fixation from the atmosphere.
Brachiaria brizantha accumulated the most phosphorus and potassium, with average values of 39 and 362 kg ha -1 . The lowest values of nutrient accumulation were found in the plants of the fallow area.
In comparison to the fallow area vegetation, the brachiaria and legumes increased nutrient cycling by at least 500% for nitrogen, 220% for phosphorus, 160% for potassium, 300% for calcium, 220% for magnesium and 240% for sulfur.
The soil cover plants with the highest potential for straw formation and nutrient cycling were brachiaria brizantha, brachiaria ruziziensis and crotalaria juncea.
There is a need for long-term studies to quantify the effects of soil cover plants as an exclusive spring-summer crop in intensive corn silage production systems in the zona da Mata Mineira region.