Desempenho agronômico de aveia-branca, avaliado por sensoriamento remoto, sob aplicação de efluente de esgoto tratado e níveis de irrigação
Carregando...
Arquivos
Data
2019-02-19
Autores
Coelho, Anderson Prates [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
A utilização e o reúso da água na agricultura são uma das maiores preocupações mundiais atualmente. Alternativas e práticas que otimizem esse recurso são sempre temas de debate e atenção por parte de pesquisadores e governantes. Para o aumento da eficiência e otimização do uso e reúso da água, o sensoriamento remoto de áreas agrícolas tem se mostrado excelente alternativa na definição de áreas de manejo específico. Dessa maneira, objetivou-se avaliar os efeitos da aplicação de efluente de estação de tratamento de esgoto (EETE) no cultivo de aveia-branca (Avena sativa L.), definindo-se zonas de manejo a partir de sensoriamento remoto com o índice de vegetação NDVI e avaliar a capacidade de generalização de modelos de estimava da produtividade de aveia-branca em função de índices de vegetação (IRVI e NDVI). O estudo foi constituído de dois experimentos, sendo um realizado na safra de inverno de 2017, com aplicação de níveis de EETE na lâmina de irrigação, apresentando cinco tratamentos (11%, 31%, 60%, 87% e 100% de EETE) e o segundo nas safras de inverno de 2017 e 2018, constituído de cinco níveis de aplicação de água na cultura da aveia-branca (11%, 31%, 60%, 87% e 100% da evapotranspiração da cultura). Para o experimento de apenas uma safra (2017), os índices médios de NDVI na cultura obtidos por sensor terrestre ativo (GreenSeeker) foram correlacionados os componentes de crescimento e produtividade da cultura. Para a definição dos limites críticos de NDVI foram geradas equações dos valores de NDVI em função da produtividade de biomassa relativa (PBR), a partir da produtividade de biomassa média, em quatro estádios fenológicos. Foram considerados como críticos os valores de NDVI referentes a 90% e 110% da PBR. No experimento de duas safras (2017 e 2018), foram calibrados modelos para a estimativa da produtividade de grãos (PG) e de biomassa (PB) da aveia-branca em função dos índices de vegetação IRVI e NDVI em quatro estádios fenológicos, obtendo-se equações para cada situação. A calibração foi realizada com os dados experimentais de 2017. Para a validação dos modelos calibrados em 2017, foram utilizados os dados experimentais de 2018. A máxima PG (5.775 kg ha-1) e PB (20.731 kg ha-1) de aveia-branca foram atingidas com doses de nitrogênio aplicadas via EETE de 45 kg ha-1 e 88 kg ha-1, respectivamente. Leituras de NDVI apresentam correlações superiores a 0,85 com os atributos de crescimento e rendimento da aveia-branca. Os modelos apresentaram capacidade de generalização para a estimativa do rendimento da aveia-branca, especialmente para a PB. Modelos parametrizados em estádios fenológicos mais avançados e com o índice IRVI apresentaram menor erro médio de estimativa.
The use and reuse of water in agriculture is one of the world's major concerns today. Alternatives and practices that optimize this resource are always topics of debate and attention on the part of researchers and rulers. In order to increase efficiency and optimize the use and reuse of water, remote sensing of agricultural areas has proven to be an excellent alternative in the definition of specific management areas. The aim of this study was to evaluate the effects of treated sewage effluent (TSE) application on white oats (Avena sativa L.), defining management zones from remote sensing with vegetation index NDVI and to evaluate the generalization capacity of models estimating the yield of white oats as a function of vegetation indices (IRVI and NDVI). The study consisted of two experiments, one of which was performed in the winter crop of 2017, with application of TSE levels in the irrigation depth, presenting five treatments (11%, 31%, 60%, 87% and 100% of TSE) and the second in the winter crops of 2017 and 2018, consisting of five levels of water application in the white oats crop (11%, 31%, 60%, 87% and 100% of crop evapotranspiration). For the first experiment, the mean NDVI indices in the crop were obtained by active terrestrial sensor (GreenSeeker), which correlated the components of growth and crop yield. In order to define the critical limits of NDVI, values were generated as a function of relative biomass yield (RBY), from mean biomass yield, in four phenological stages, considering as critical the NDVI values referring to 90% and 110% of RBY. In the second experiment, models for estimation of grain yield (GY) and biomass (BY) of white oats were calibrated as a function of IRVI and NDVI vegetation indexes in four phenological stages, obtaining equations for each situation. Calibration was performed with the experimental data of 2017. For the validation of the calibrated models in 2017, the experimental data of 2018 were used. The maximum GY (5,775 kg ha-1) and BY (20,731 kg ha-1) of white oats were reached with nitrogen rates applied by TSE of 45 kg ha-1 and 88 kg ha-1, respectively. NDVI readings have correlations higher than 0.85 with the attributes of growth and yield of white oats. The models presented generalization capacity for estimating the yield of white oats, especially for BY. Parametrized models in more advanced phenological stages and with IRVI index presented smaller mean error of prediction.
The use and reuse of water in agriculture is one of the world's major concerns today. Alternatives and practices that optimize this resource are always topics of debate and attention on the part of researchers and rulers. In order to increase efficiency and optimize the use and reuse of water, remote sensing of agricultural areas has proven to be an excellent alternative in the definition of specific management areas. The aim of this study was to evaluate the effects of treated sewage effluent (TSE) application on white oats (Avena sativa L.), defining management zones from remote sensing with vegetation index NDVI and to evaluate the generalization capacity of models estimating the yield of white oats as a function of vegetation indices (IRVI and NDVI). The study consisted of two experiments, one of which was performed in the winter crop of 2017, with application of TSE levels in the irrigation depth, presenting five treatments (11%, 31%, 60%, 87% and 100% of TSE) and the second in the winter crops of 2017 and 2018, consisting of five levels of water application in the white oats crop (11%, 31%, 60%, 87% and 100% of crop evapotranspiration). For the first experiment, the mean NDVI indices in the crop were obtained by active terrestrial sensor (GreenSeeker), which correlated the components of growth and crop yield. In order to define the critical limits of NDVI, values were generated as a function of relative biomass yield (RBY), from mean biomass yield, in four phenological stages, considering as critical the NDVI values referring to 90% and 110% of RBY. In the second experiment, models for estimation of grain yield (GY) and biomass (BY) of white oats were calibrated as a function of IRVI and NDVI vegetation indexes in four phenological stages, obtaining equations for each situation. Calibration was performed with the experimental data of 2017. For the validation of the calibrated models in 2017, the experimental data of 2018 were used. The maximum GY (5,775 kg ha-1) and BY (20,731 kg ha-1) of white oats were reached with nitrogen rates applied by TSE of 45 kg ha-1 and 88 kg ha-1, respectively. NDVI readings have correlations higher than 0.85 with the attributes of growth and yield of white oats. The models presented generalization capacity for estimating the yield of white oats, especially for BY. Parametrized models in more advanced phenological stages and with IRVI index presented smaller mean error of prediction.
Descrição
Palavras-chave
Acurácia, IRVI, Lâminas de irrigação, Modelagem, NDVI, Produtividade, Accuracy, Irrigation depths, Modeling, Grain yield