Numerical modeling for yield forecast of flooded rice in the state of Rio Grande do Sul, Brazil

Abstract

The objective of this work was to evaluate a method of yield forecast for flooded rice in the state of Rio Grande do Sul (RS), Brazil, using the SimulArroz rice model and the RegCM4 regional climate model. Daily data of minimum temperature, maximum temperature, and solar radiation, simulated from nine members of the RegCM4 model, were used as input data to the SimulArroz model for rice yield forecast. To test the yield forecast performance, field experiments were carried out during the 2013/2014 growing season, in the municipalities of Restinga Seca and Itaqui, RS, Brazil, where grain yield was evaluated. The observed rice grain yield ranged from 6,898 to 10,272 kg ha^‑1, while the predicted one ranged from 2,853 to 9,636 kg ha^‑1. The rice grain yield forecasts, generated by members 31, 19, 13, and 01, had a root mean square error of 1,218, 1,134, 1,354, and 1,374 kg ha^‑1, respectively. Flooded rice yield forecast for the state of Rio Grande do Sul can be made through the SimulArroz model, using, as input meteorological data, the seasonal climate forecast obtained with the RegCM4 model.