Modelagem da perda de produtividade da cultura do milho em função da deficiência hídrica

• Ano de defesa: 2017
• Autor(a) principal: Buske, Taise Cristine
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• dARK ID: ark:/26339/0013000016cqd
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Brasil; Engenharia Agrícola; UFSM; Programa de Pós-Graduação em Engenharia Agrícola; Centro de Ciências Rurais
• Palavras-chave em Português: Déficit hídrico; Modelos agrometeorológicos; Quebra de rendimento; Water deficit; Agrometeorological models; Break of yield; CNPQ::CIENCIAS AGRARIAS::ENGENHARIA AGRICOLA
• Link de acesso: http://repositorio.ufsm.br/handle/1/13528

The Northwest mesoregion of Rio Grande do Sul is characterized by the predominance of spring-summer crops, with maize being an important item in the production system. Among the limiting components, the water deficit is one that affects agricultural production with greater frequency and intensity. In order to characterize the effects of climatic variations on grain yield, agrometeorological models have been used, which are very useful because they require very little input data. Therefore, this study aimed to evaluate the yield losses of maize cultivated under rainfed conditions in relation to the potential productivity in the Cruz Alta micro region of the state of Rio Grande do Sul. The study was developed using daily meteorological data, in the period of 1993-2014. Maize yield data were obtained from the IBGE website. In order to estimate the actual productivity of the crop the agrometeorological models of Jensen (1968), Minhas, Parikh and Srinivasan (1974) and, Doorenbos and Kassan (1979) were adopted. The reference evapotranspiration was estimated by the Penmam-Montheit method, as for the estimation of the actual evapotranspiration was used the water balance method according to Thornthwaite and Mather (1955). In the determination of potential productivity the Wageningen and Agroecological Zone methods were used. Also, the planting combinations were simulated in September, October, November and December. Initially, the analyzes were carried out with parameters of the agrometeorological models recommended in the literature, later an adjustment of the same was done. The accuracy of the estimation of each agrometeorological model was determined from the linear regression analysis, performed between the annual values of observed and estimated real productivity. It is noted that MWa tends to overestimate the potential productivity, whereas the MZA method better tracked the fluctuations of the results. In general, the agrometeorological models tested in the different conjunctions, with parameters recommended in the literature, presented an unsatisfactory coefficient of determination and the performance ranged from poor to medium. After adjusting the parameters of the models, the improvement in the coefficient of determination was evident, except for Doorenbos and Kassan. The performance of the different combinations ranged from poor to very good, and Jensen's model was rated very good in October and November, a result that was also found for the model of Minhas, Parikh and Srinivasan in November. The reccomended coefficients are -0.768, 0.699, 0.374 and -0.330 for the Jensen model, and -1.438, 1.078, 0.439 and -0.442 for the Minhas, Parikh and Srinivasan model, according to the phenological stage I, II, II and IV, respectively. It was also observed a drop in yield in most of the studied years, notoriously in the bands greater than 30% of productivity loss, being able to reach relative frequency of 30% for October, November and December. Even in the range with less than 10% of losses, in any evaluated period, losses were observed in 15% of the years. It was verified that the maize crop is affected by the water deficit in the spring-summer period, causing risks of obtaining grain yield below the expectation.