1787 Nematode Management using Remote Sensing Technology, Self-Organized Maps and Variable Rate Nematicide Applications

Remote sensing technology is being examined as a field worthy means for predicting reniform nematode population densities in the cotton rhizosphere using hyperspectral reflectance data and self-organized maps. Our objectives were to determine 1) the hyperspectral reflectance of cotton plants affected by different reniform nematode population levels in controlled microplot tests and in fields naturally infested with the reniform nematode, 2) to determine spatial distribution of the nematodes in the field and establish zones for variable rate nematicide applications, and 3) to determine the cotton yield improvements utilizing variable rate nematicide applications. Hyperspectral reflectance data was collected from reniform infected cotton plants growing in controlled microplots and from naturally infested field locations. Hyperspectral signatures have been identified which correspond to reniform nematode colonization and level of infection of cotton plants. These signatures and reniform populations were classified using a supervised artificial neural network Self Organized Map architecyure. Nematode contour maps were developed to demonstrate the spatial distribution of the nematode and to establish zones of differing nematicide rate. Cotton yield results were higher in the site-specific application treatments compared with the conventional single rate applications. The combination of hyperspectral reflectance and VRT applications of nematicides may someday provide our producers with a rapid nematode prediction method and reduce nematicide applications to specific infield sites.