Modelos agrometeorológicos estatísticos de previsão de produtividade e qualidade para cana-de-açúcar

• Ano de defesa: 2015
• Autor(a) principal: Marcari, Marcos Antonio [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Balanço hidrológico; Cana-de-açúcar; Climatologia agricola; Meteorologia agricola; Produtividade agrícola; Meteorology
• Link de acesso: http://hdl.handle.net/11449/128085; http://www.athena.biblioteca.unesp.br/exlibris/bd/cathedra/18-09-2015/000849756.pdf

The climate is the important factor of influence for sugarcane and its study is fundamental for understanding the climatic requirements of the crop. The estimate is made with historical data and is a current condition, as the forecast is the estimate made for the future, ie, with current available data simulate a future state. The present work aimed developing regional agro-meteorological models to make the yield forecasting of tons of sugarcane per hectare (TCH) and quality in relation to total recoverable sugar (ATR) of sugarcane in a monthly scale. We used monthly climatological data (air temperature, Precipitation, Water Deficiency and Surplus, Potential and Actual evapotranspiration, Soil Water Storage, Solar Global irradiation) of the previously year to forecast TCH and ATR of the next year using multiple linear regression. The combination of monthly climatological data was made searching a small mean absolute percentage error possible with p-value less than 0.05, and models with greater possible anticipation. Data of 12 years of Jaboticabal, a major sugarcane producer in the State of São Paulo, were used for analysis, being the period from 2002 to 2009 used for calibration and from 2010 to 2013 for validation. We observed that all models calibrated were significant and accurate, because the higher values of mean absolute percentage error (MAPE) were of 4.06% in the forecasting of the TCH(C) of July. The model calibrated for November had the presence of water deficit variable in every environment, showing the importance of this variable in the crop. Monthly models tested in this work showed significant performances in their forecasting's. For example, the forecast of the TCHMAY in the AB environment (MAPE = 1.89% and R2 adj = 0.90) considering an average value of 90.6 t ha-1 in the region, the model misses about 1.7 t ha-1.In this case the anticipation for forecasting TCHMAY was eight months because the last climatological ...