USING THE NORMALIZED DIFFERENCE VEGETATION INDEX AND A CROP SIMULATION MODEL TO PREDICT SOIL SPATIAL VARIABILITY

Improved understanding and better techniques to measure spectral reflected radiation must translate into practical applications to better manage cropping systems. This study developed a simple procedure to predict the unknown underlying patterns of soil variation using an airborne multispectral image. A crop simulation model (CERES–Maize) was used to establish the “base line” effects of genotype and population on crop growth, and on the reflected radiation from maize (Zea mays L.) canopies. The normalized difference vegetation index (NDVI) was calculated from an aerial spectral image obtained at silking. The NDVI–based image analysis indicated the areas of the field supporting crop growth above or below the “base line” for each treatment, thereby revealing the spatial patterns of soil variability. The relative growth map obtained by NDVI analysis compared well with a relative growth map derived from field measurements of leaf area index at silking. The procedure offers potential for target–oriented soil and crop sampling for spatial models, site–specific management, and also identifying patterns of crop–limiting factors not related with the soil, such as pests or diseases.