Using fPAR as a modelling parameter for tobacco yield estimation in regional cropping areas
Remote sensing has shown its usefulness in many applications for tobacco crop evaluation: using aerial or satellite collected data, mathematical models could be developed to investigate the status of a crop or to prognosticate its yield. Remote sensing is especially useful when dealing with extensive growing areas. The aim of the present paper is to evaluate the efficiency of the fraction of absorbed photosynthetically active radiation (fPAR) as a tobacco yield modelling parameter. fPAR covers the spectral range of 400 to 700 nm that a canopy absorbs and it is a value often supplied by observation satellites. The study area covers 17 municipalities in the State of Rio Grande do Sul, Brazil, with a total area of 7200 square km. The method uses fPAR data collected weekly by NASA’s Modis satellite during the whole tobacco growing period. The comparative analysis of the fPAR profiles during such periods allows the municipalities to be classified into five groups according to evolution of behaviour. The final calibration to obtain the tobacco yield models for every group was done using data from nine consecutive years (2003-2012). Through the regression of fPAR values over the corresponding “Earth true” yield data from the Instituto Brasileiro Geral de Estadisticas (IBGE), linear calibration models were obtained. The evaluation of the method was done by applying the obtained models to the 2013 and 2014 crops of the five municipality groups. The absolute errors relative to the IBGE terrestrial registers resulted in all the cases being in the range of 4 % to 7 %. The results obtained show that fPAR is a satisfactory modelling parameter for tobacco yield evaluation especially for its application to extensive cropping areas.