Use of remote sensing, geographic information systems, and spatial statistics to assess spatio-temporal population dynamics of Heterodera glycines and soybean yield quantity and quality

Abstract

A wide range of soybean [Glycine max (L.) Merr.] growth was generated in field experiments throughout 2002 and 2003. Soybean canopies were assessed periodically with a hand-held, multispectral radiometer measuring percentage reflectance at eight different narrow wavelength bands spaced 50 nm apart from 460 nm to 810 nm. After reflectance data were obtained, soybean plants were removed and green leaf area index (GLAI) determined. Within individual assessment dates, relationships among percentage reflectance of individual wavelength bands and vegetation indices with GLAI were studied using regressions. Percentage reflectance for 660 nm and for 810 nm had the best relationships with GLAI within assessment date and throughout each season. However, regression lines obtained for individual assessment dates were significantly different for the regressions obtained for entire seasons. The indices that best estimated variation in GLAI were radiance ratio (RR), difference vegetation index (DVI), and renormalized difference vegetation index (RDVI). Ground-based RDVI was not affected by environment conditions, and there was no significant difference between the regression lines of RDVI on GLAI obtained for individual assessment dates and those obtained for the entire 2002 and 2003 seasons. Results indicate that caution should be taken when percentage reflectance data obtained at different dates are combined to estimate GLAI. The ability of normalized difference vegetation index (NDVI), transformed vegetation index (TVI), and green normalized difference vegetation index (GNDVI) to estimate GLAI was equivalent throughout seasons. Plant dry weight (PDW) and GLAI were positive and linearly related. Similar relationships between percentage reflectance data and GLAI or PDW were observed.