Sorghum–Grass Intercropping Systems under Varying Planting Densities in a Semi-Arid Region: Focusing on Soil Carbon and Grain Yield in the Conservation Systems

A major challenge in sorghum intercropping systems is maintaining their yields compared to the yields of the solo crops, especially in arid and semi-arid regions. This study aimed to test the hypothesis that intercropping systems using sorghum ( Sorghum bicolor (L.) Moench.) and Brachiaria sp. are viable means to increase sorghum production and soil carbon in the conservation systems. Field trials were conducted in the semi-arid region of Minas Gerais, Brazil, during two crop cycles of sorghum associated with different grasses ( Andropogon gayanus —AG; Cenchrus ciliaris cv. Aridus —CCA; Cenchrus ciliaris cv. 131—CC; Brachiaria decumbents —BD; Brachiaria brizantha —BB; Brachiaria ruziziensis —BR; Panicum maximum —PM), using row spacings of 0.4 and 0.8 m. Panicles of sorghum (yield) and grass dry matter were collected to determine yields. Results showed that the addition of grasses in systems decreased the grain yield in all systems, except in the systems using sorghum with CCA in 0.4 m, AG in 0.8 m, or BR in 0.8 m. In the 0.4 m row spacing, the sorghum associations with CC, BB, or PM are greater alternatives to increase soil carbon. However, when the row spacing was increased, the sole sorghum was the best alternative to increase the carbon. In machine learning, sorghum systems with CCA and AG are better alternatives to increase the yields, while sorghum with CC, PM, BR, and BB increases the grass dry matter in soil.