Location of the study
Four experiments were carried out with the soybean crop in three locations in the state of Rio Grande do Sul, Brazil. The climate of the region where the experiments were conducted, according to the Köppen classification, is Cfa, humid subtropical without a defined dry season (Alvares et al. 2013).
The first experiment was conducted in the experimental area of the Federal University of Fronteira Sul during the 2018/2019 crop season, located in the municipality of Cerro Largo (27° 08' S; 54°45'W; altitude: 258 m). During the 2019/2020 crop season and in the 2021 off-season two experiments were carried out in the municipality of Entre-Ijuís (28° 23' S; 54° 16' W; altitude: 273 m). In addition, another experiment was conducted during the 2020/2021 crop season in the area experimental at the University of Passo Fundo located in Passo Fundo (28º 15' 41” S; 52º 24' 45” W; altitude: 709 m).
Experimental Design
The experimental design used in in Cerro Largo was completely randomized with three replications. In the other experiments, it was used a randomized complete block design with four replications, except for the study carried out in the crop season in the municipality of Entre-Íjuis, which had three replications. All experimental units were 2.25 m wide and 5 m long, totaling an area of 11.25 m².
Sowings were carried out on November 28th, 2018 (Cerro Largo), December 8th, 2020 (Passo Fundo) and January 11th, 2021 (off-season, Entre-Ijuís), using row spacing of 0.45 m and density of 300,000 plants per hectare. In the crop season in Entre-Ijuís, sown occurred on December 6th, 2019, a row spacing of 0.45 m was used and two sowing densities were used, totaling 246,666 and 451,111 plants per hectare.
Forecasting system and treatments
The Cerro Largo experiment was conducted in a factorial scheme in which the treatments consisted of a combination of two soybean cultivars (SYN 1561 IPRO and TMG 7363 RR) and the chemical control strategies based on the prediction system from ASR. In the 2019/2020 crop season in Entre-Ijuís, the cultivars BMX Compacta IPRO, M5838 IPRO and TMG 7262 RR were used. In the off-season (2021), two plant densities were tested and the cultivar that was used was BMX Compacta IPRO. In the Passo Fundo, only one soybean cultivar (BMX Ativa RR) was used in the experiment.
The chemical control strategies for ASR were defined according to the calculated severity values (CSV) obtained from the proposed forecasting system. The CSV were calculated daily, obtained by the product of the response functions to leaf wetness duration and air temperature during the wetness period, both varying from 0 to 1. The presence of leaf wetness was considered when air relative humidity was higher than 85%. The response function to leaf wetness duration (f(W)) is described below:
where: W corresponds to leaf wetness duration, in hours. The response function to air temperature (f(T)) is described as follows:
the values 42, 22.75, and 8 indicate, respectively, the maximum, optimum, and minimum air temperatures for disease development. To determine the CSV data were obtained using an automatic meteorological station installed close to each experimental area, which were stored at 10-minute intervals.
In Cerro Largo, fungicide applications were carried out every 3, 5, 7 and 9 CSV, accumulated from the fifth fully expanded trifoliate leaf (V5). In Entre-Íjuis, considering the crop season experiment, applications were carried out every 7, 9, 11, 13 and 15 CSV and in the off-season every 6, 9, 12 and 13 CSV. In Passo Fundo, applications were made every 7, 9 and 11 CSV. In addition, all experiments had an absolute control where no fungicide was applied and a control with calendar application every 14 days, starting from the seventh expanded trifoliate leaf (V7) (Fehr and Caviness 1977).
Fungicide applications were performed differently in each experiment. In Cerro Largo and Entre-Ijuís, the first two applications in the treatments were carried out with the fungicide containing azoxystrobin (90 g ha-1 a.i.) + benzovindiflupyr (45 g ha-1 a.i.) and the other applications with the fungicide with trifloxystrobin (60 g ha-1 a.i.) + prothioconazole (70 g ha-1 a.i.). In Passo Fundo, the first and third applications were performed with fluxapyroxad (50 g ha-1 a.i.) + pyraclostrobin (100 g ha-1 a.i.), the second with bixafem (62.5 g ha-1 a.i.) + trifloxystrobin (75 g ha-1 a.i.) + prothioconazole (87.5 g ha-1 a.i.) and the last two with fenprofimorph (225 g ha-1 a.i.).
The severity (%) of ASR was estimated weekly in the Cerro Largo experiment, evaluating leaves from the lower and upper thirds of ten random plants in the plot. In Passo Fundo, the evaluations were performed biweekly, and the collection of the leaves were made without distinction of thirs of the plant. Disease evaluations ocurred from the observation of the first symptoms with the aid of the diagrammatic scale (Godoy et al. 2006). In Entre-Ijuí, no disease severity assessments were carried out. The area under the disease progress curve (AUDPC) was determined by calculating the integral of the disease progress curve, according to Campbell and Madden (1990).
Soybean yield
In the experiments located in Cerro Largo and Entre-Íjuis, soybean plants were harvest in three meters from the two central rows of each plot to determine the grain yield and thousand grain mass considering grain moisture content at 13%. In Passo Fundo, soybean plants were harvest in five meters in the three central rows of the plot.
Data analysis
The data were subjected to analysis of variance using the F test and the means compared using the Tukey’s test at a 5% error probability level. The analysis was performed separately for each experiment. Data on severity of ASR were submitted to the arcsen transformation (0.01x)0.5 for statistical analysis. The analyzes were performed in the R software (R Core Team 2022).