Calibração e avaliação do modelo CSM-Ceres-Maize para cultivares crioulas e melhoradas de milho

Abstract

Maize is an important crop of cereals around the world and one of the most mathematical models used to simulate the growth and development of maize is the CSM-Ceres-Maize model. However, the use of this model for open pollination varieties (landraces and improved) was not found in the literature. The objectives of this dissertation were (a) Calibrate the CSM-Ceres-Maize model available on the DSSAT v4.5 platform (Decision Support System for Agrotechnology Transfer) for two landrace maize cultivars ('Cinquentinha' and 'Bico de Ouro ), an improved variety ('BRS Missões') and a modern hybrid ('AS 1573PRO') and (b) Apply the CSM-Ceres-Maize model in Rio Grande do Sul State to simulate the kernel yield of the four cultivars for the recommended period and compare the agronomic performance of the open pollinated varieties with the hybrid in the condition without water restriction. The calibration of the CSM-Ceres-Maize model on the DSSAT platform was performed using the trial and error approach, obtaining a total of 7 parameters calibrated with a field experiment conducted in Santa Maria, in the 2013/14 growing seasons (Sowing date: 11/04/2013). The test of the model was performed with independent data obtained from the field experiment in Santa Maria, RS, with sowings dates in 08/20/2013, 02/03/2014 and 08/15/2014, and also the model was run for 15 locations in Rio Grande do Sul State, with sowing date range in a 5-day time step to analyze the kernel yield between cultivars. The model was able to simulate the phenology and productivity of four cultivars, capturing the evolution of each one during your cycle. Also, it was possible to prove the stability of potential kernel yield of the landrace maize cultivars in comparison with the hybrid.