Padrões de variabilidade espacial e temporal da produção em um pomar jovem de laranja Hamlin

Abstract

Yield data of 14 71 young Hamlin trees in a commercial grave were collected in 2 seasons, 2000/2001 and 2001/2002, for identifying the pattems of spatial and temporal yield variability. Geostatistical tools and fuzzy clustering were used for the analyses. Also, the prajected crown area coverage, volume, height, foliage density, DLLAI, total leaf area, and spectral response data of 52 trees distributed along 2 transects were collected for studying the relationships between crawn parameters, spectral response, and yield. Multispectral digital images of the grave area were captured by a high resolution vídeo camera in 3 distinct phenological stages: during winter dormancy, beginning of the fruit maturation and just after the flowering period of the next season. NDVI, SA VI, red and infrared raw DN data were used in regression analyses between spectral response and variables. The results showed there was spatial dependence between pairs of yield data, especially when the climatic conditions were poor. Crass-sernivariogram showed the yield was consistently spatially dependent throughout the 2 seasons, despite the great climatic fluctuations. Fuzzy classification has identified general pattems of between-season variation within a field. But, these pattems correspond to regions poorly spatially coherent. Despite such behavior, spectral response Xll of the trees was significantlyrelated to these regions. Regression analyses showed poor relations between yield and projected crown area coverage, volume, height, foliage density, DLLAI, and total leaf area of the trees. But, spectral response was significantlyrelated to crown parameters and yield. NDVI was found to relate better with the varYield data of 14 71 young Hamlin trees in a commercial grave were collected in 2 seasons, 2000/2001 and 2001/2002, for identifying the pattems of spatial and temporal yield variability. Geostatistical tools and fuzzy clustering were used for the analyses. Also, the prajected crown area coverage, volume, height, foliage density, DLLAI, total leaf area, and spectral response data of 52 trees distributed along 2 transects were collected for studying the relationships between crawn parameters, spectral response, and yield. Multispectral digital images of the grave area were captured by a high resolution vídeo camera in 3 distinct phenological stages: during winter dormancy, beginning of the fruit maturation and just after the flowering period of the next season. NDVI, SA VI, red and infrared raw DN data were used in regression analyses between spectral response and variables. The results showed there was spatial dependence between pairs of yield data, especially when the climatic conditions were poor. Crass-sernivariogram showed the yield was consistently spatially dependent throughout the 2 seasons, despite the great climatic fluctuations. Fuzzy classification has identified general pattems of between-season variation within a field. But, these pattems correspond to regions poorly spatially coherent. Despite such behavior, spectral response Xll of the trees was significantlyrelated to these regions. Regression analyses showed poor relations between yield and projected crown area coverage, volume, height, foliage density, DLLAI, and total leaf area of the trees. But, spectral response was significantly related to crown parameters and yield. NDVI was found to relate better with the variables. Despite the significant relations between yield and spectral response, particularly when the climatic conditions are favorable, it is suggested the regression analyses should be separate by soil type to improve the relations. Taken multispectral digital images before or after the vegetative growth period were similar performed.iables. Despite the significant relations between yield and spectral response,particularly when the climatic conditions are favorable, it is suggested the regression analyses should be separate by soil type to improve the relations. Taken multispectral digital images before or after the vegetative growth period were similar performed.