Sugarcane yield gap in Brazil: a crop modelling approach

• Ano de defesa: 2015
• Autor(a) principal: Monteiro, Leonardo Amaral
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Saccharum spp; Agricultural management; Attainable yield; Average yield; Déficit hídrico; Manejo agrícola; Potential yield; Produtividade média; Produtividade atingível; Produtividade potencial; Sustainability; Sustentabilidade; Water deficit
• Link de acesso: http://www.teses.usp.br/teses/disponiveis/11/11152/tde-08032016-142721/

Currently, the cropping area is around 10 million hectares, in which the sugarcane fields are expanding for marginal regions, mainly where grains and pasture were previously cultivated. From that, the objectives of this study were: to calibrate and evaluate a sugarcane yield model using data from 12 fields conducted under high technology field conditions; to evaluate the performance of a gridded system (NASA/POWER) to increase the spatial density of the weather stations in Brazil, to be employed as input data of crop simulation models; to map, in micro-region scale, the potential (Yp), the best farmer\'s (Ybf) and average actual (Yavg) sugarcane yields in Brazil, in order to determine the sugarcane yield gaps by water deficit (YGWD) and by crop management (YGCM), and to define strategies for a most sustainable sugarcane crop production. The yield model showed a good performance in the yield simulation, during the calibration and validation phases. The estimated yield in the calibration phase was 81.9 Mg ha-1 while the observed one was 82.3 Mg ha-1. In the validation phase, the estimated yield was 82.9 Mg ha-1 and the observed was 86.9 Mg ha-1. These results suggested that this kind of model can be used for yield estimation, mainly for agricultural planning purposes, at regional and national scales. The NASA/POWER weather data showed a reasonable performance when compared to observed data that control Yp (solar radiation and air temperature). On the other hand, although the annual average rainfall were very similar in all locations evaluated, this variable presented unsatisfactory statistical coefficients (R2 = 0.60 and MAPE = 233.4%), being suggested, therefore, to replacement of rainfall data from the gridded system by the ones from local rainfall stations (ANA). In the majority of the locations, the percentage errors of Yp were ±15%, while the attainable yield was overestimated by 14% when estimated without replace the rainfall data by the ANA\'s data. Otherwise, when the rainfall data were modified by the ones from ANA, a better adjustment was obtained, revealing an overestimation of only 5%. Finally, 259 virtual weather stations were generated with NASA/POWER data and rainfall from ANA database to estimate yields. The yield types were spatialized through software ArcGis 9.3® at micro-region level. The yield gaps by water deficit and crop management were determined. It was observed that the sugarcane yield losses in Brazil are mainly caused by water deficit (74% of total yield gap), while 26% was due crop management. These results contribute for a better understanding about the factors that control sugarcane production and, therefore, they can be used to define strategies, such use of drought tolerant cultivars, irrigation, and soil decompaction, to make sugarcane production in Brazil more efficient and sustainable.