Nitrogen balance and efficiency as indicators for monitoring the proper use of fertilizers in agricultural and livestock systems

Site description

The experiment was conducted at the Beef Cattle Research Center of the Institute of Animal Science/APTA/SAA, Sertãozinho, São Paulo, Brazil (21°08′16″ S e 47°59′25″ W, average altitude 548 m), during two consecutive years. The climate in this region is Aw according to the Köppen’s classification, characterized as humid tropical, with a rainy season during summer and drought during winter. The meteorological data is reported in Fig. 1. The soil in the experimental area is classified as an Oxisol⁴². Before the experiment, soil samples were collected for chemical characterization (Table 4), which was performed following the methodology described in Van Raij et al.⁴³. Samples were collected in 18 experimental paddocks, at the depths of 0- to 10- and 10- to 20-cm layers, from 10 distinct sampling points in each paddock, in order to create one composite sample per unit, totaling 36 samples analyzed.

The nitrogen total (Nt) content was determined by the micro-Kjeldahl method⁴⁴, and the soil nitrogen stocks (SN) were calculated using the following equation below, according to Veldkamp et al.⁴⁵.

where concentration refers to the Nt concentration at a given depth (g kg⁻¹), BD is the bulk density at a certain depth (average 1.24 kg dm⁻³), and 1 is the layer thickness (cm).

Description of treatments and managements

The experiment was carried out in a 16-ha area, divided into 18 paddocks of 0.89 ha each (Fig. 2), organized in a randomized blocks design with three replicates and six treatments, namely conventional crop system with grain maize production (CROP), conventional livestock system with beef cattle production in pasture using Marandu grass (LS), and four ICLS for the production of intercropped maize grain with beef cattle pasture. All production systems were sowed in December 2015, under a no-tillage system. The fertilization recommendations in the systems were based on the recommendation presented in the Boletim 100⁴⁶.

In the CROP system, the maize Pioneer P2830H was cultivated, sowed in a spacing of 75 cm and sowing density of 70 thousand plants. Applications of 32 kg ha⁻¹ of nitrogen (urea), 112 kg ha⁻¹ of P₂O₅ (single superphosphate) and 64 kg ha⁻¹ of KCl (potassium chloride) were performed. Complementarily, a topdressing fertilization was made using 80 kg ha⁻¹ of nitrogen (urea) and 80 kg ha⁻¹ of KCl. Sowing was carried out for two consecutive years (December 2015 and 2016), providing two harvests of maize grains (May 2016 and 2017), and between one harvest and the other, the soil remained in fallow without any cover crop. The total amount of fertilizer applied in two years was 224 kg ha⁻¹ of nitrogen (urea), 224 kg ha⁻¹ of P₂O₅ (single superphosphate) and 288 kg ha⁻¹ of KCl (potassium chloride).

For the LS treatment, Urochloa brizantha (Hoechst. ex A. Rich) R.D. Webster cv. Marandu (syn. Brachiaria brizantha cv. Marandu) was sowed in a spacing of 37.5 cm, with a density of 5 kg ha⁻¹ of seeds (76% of crop value) for the pasture assemblage. Marandu grass seeds were mixed with the planting fertilizer, applying 32 kg ha⁻¹ of nitrogen (urea), 112 kg ha⁻¹ of P₂O₅ (as single superphosphate) and 64 kg ha⁻¹ of KCl. Applications of 40 kg ha⁻¹ of nitrogen, 10 kg ha⁻¹ of P₂O₅ and 40 kg ha⁻¹ of KCl were also performed as topdressing fertilization in October 2016 and March 2017. 90 days after sowing, the pasture was ready to be grazed (March 2016). Three grazing periods were carried out in continuous stocking systems, with the first period between March and April 2016, the second period between August and October 2016 and the third between November 2016 and December 2017. The total amount for 2 years was 112 kg ha⁻¹ of nitrogen (urea), 132 kg ha⁻¹ of P₂O₅ (single superphosphate) and 144 kg ha⁻¹ of KCl (potassium chloride).

The same cultivar, spacing, sowing density and fertilization rates described in the CROP treatment were used in all ICLS, as well as the same density of Marandu grass seeds and topdressing fertilization adopted in the pasture of the LS treatment. The total amount for two years was 192 kg ha⁻¹ of nitrogen (urea), 132 kg ha⁻¹ of P₂O₅ (single superphosphate) and 224 kg ha⁻¹ of KCl (potassium chloride). In ICLS-1, Marandu grass was sowed in lines simultaneously with maize, while in ICLS-2, the sowing was also simultaneous, but the application of an under-dose of 200 mL of the herbicide Nicosulfuron was used, 20 days after seedlings emergence. In the ICLS-3, Marandu grass seeds were sown the time of topdressing fertilization of maize, thus the grass seeds were mixed with the fertilizer, and sowing was carried out in the interlines of maize, using a minimum cultivator. In ICLS-4, the sowing of Marandu grass was performed simultaneously with maize, but the grass seeds were sowed in both rows and inter-rows of maize, resulting in a spacing of 37.5 cm. In this treatment, the application of 200 mL of the herbicide Nicosulfuron was adopted, 20 days after seedlings emergence.

In all ICLS treatments, maize harvest was carried out in May 2016. Ninety days after harvesting the plants, the pastures were ready to be grazed. Therefore, two grazing periods were made in continuous stocking, being the first period between August and October 2016 and the second period between November 2016 and December 2017. The method for animal stocking in treatments LS and ICLS was continuous with a stocking rate (put and take) being defined according to Mott⁴⁷. Caracu beef cattle with 14 months of age were used at the beginning of the experiment, with an average body weight of 335 ± 30 kg.

Estimations of the nutrient balance (NB) and nutrient use efficiency (NUE)

In this study, the inputs and outputs of N were assessed at the farm level^48,49. The NB was calculated by the equation below^19,45,50.

As for the NUE, this parameter was evaluated as defined by the EU Nitrogen Expert Panel⁵¹, being calculated as the ratio between outputs and inputs of nitrogen.

where NB is the nutrient balance, N is nitrogen, Input is the N concentration in the mineral fertilizer (urea), Output is the nitrogen concentration in export (maize grain and animal tissue), and NUE is the use efficiency of the nutrient.

The amount of N exported in maize grains, the grain production results (Table 2) were multiplied by the mean value of N, consulted in Crampton and Harris⁵².

In order to estimate the amounts of nutrient exported by the animals in their tissues, the values of live weight gain were considered [kg ha^-1 of live weight (PV)] (Table 2), as well as the nitrogen values of the tissue, according to the methodology proposed by Rasmussen et al.²¹. Those authors reported that for animals weighting less than 452 kg/PV, it represents 2.7%, while heavier animals have a 2.4% nitrogen content representation of their body weight.

The inputs and outputs of N in each production system are represented in Figs. 3, 4 and 5. Biological N fixation, atmospheric deposition, denitrification, leaching, rainfall, and volatilization and absorption of ammonia were not considered in the calculation of NB.

Data for animal tissue, animal excreta, and N concentration in grains were obtained from key manuscripts from the scientific literature in order to estimate the N balance.

Calculation of nitrogen quantity and valuation of organic residues

The amount of N in the organic residues was determined as a function of the system (Figs. 3, 4, 5). The residue considered in the CROP was the straw derived from maize, while for LS it was the litter deposited (LD) in the grass Marandu, and animal manure (feces and urine). The ICLS were considered as the straw, LD, and animal manure.

The N concentration in straw and LD was determined following the methods of AOAC (1990). Straw was sampled immediately after maize grain harvest, using a 1-m² frame in the field. The material was collected in two spots of the plot that were chosen randomly. All straw deposited on the soil was sampled, weighted and dried in an oven with air circulation (60 °C) until constant weight, for the determination of dry matter in kg of straw per hectare (Table 2). The LD in the pasture system (Table 2) was analyzed according to Rezende et al.⁵³.

In order to estimate the daily amount of excreta, we considered the stocking rate adopted in the experiment (Table 2) and the values proposed by Haynes and Williams⁵⁴. According to those authors, adult beef cattle can defecate on average 13 times a day and urinate 10 times a day, totaling a daily amount of 28.35 kg of feces and 19 L of urine.

The valuation was calculated based on the mean value of urea for the last 10 years in the fertilizer market^55,56,57, namely $0.28 kg⁻¹ ha⁻¹ of urea, and considering the loss of nitrogen by volatilization, which according to Freney et al.⁵⁸ and Subair et al.⁵⁹ can reach up to 28%.

Statistical analysis

The experiment was assembled in a randomized blocks design. The model adopted for the analysis of all response variables included the block’s and treatments fixed effects (3 blocks and 6 treatments), in addition to the random error. Statistical analysis were carried out by the function “dbc()” of the package “ExpDes.pt” of the software R Development Core Team⁶⁰, and the mean values were compared by the Tukey’s test at a 5% probability level.

Source: Ecology - nature.com