Estimation of sugar cane productivity using time series of AVHRR/NOAA-17 images and a spectral-phenological model

Abstract

Despite the progress and economic and social importance of agribusiness in Brazil, there is still a great difficulty in estimating the accuracy and anticipation necessary to harvest the main crop of the country, which justifies the study and development of methods employing the sensing remote for this purpose. This study aims to evaluate a methodology based on the use of a spectral-phenological model proposed by Pellegrino (2001) and time series of NDVI values of the final estimate of the productivity culture. The tests were performed in four seasons (2003/2004, 2004/2005, 2005/2006 and 2006/2007) and six municipalities located in the São Paulo state (Araraquara, Cravinhos, Jaboticabal, Pitangueiras, Ribeirao Preto and Sertãozinho). The reference values for the yield were provided by IBGE (National Institute of Geography and Statistics). The methodology is useful for indicating the operational support and official systems of forecasting/monitoring of sugar cane, providing errors less than 6% in advance from 81 to 142 days before harvest.