Monday, November 13, 2006

Remotely Estimating Cotton Defoliation with Reflectance Data.

Glen L. Ritchie, University of Georgia, 1305 Windsor Drive, Tifton, GA 31794 and Craig Bednarz, Texas Tech University, "Plant and Soil Sciences, Box 42122", Lubbock, TX 79409-2122, United States of America.

Cotton harvest readiness based on percent defoliation is usually judged by visual estimates, but these estimates are subjective and may differ from one reviewer to the next. A spectrometric method for quantifying cotton defoliation is proposed.  Leaf area index (LAI) was monitored in multiple environments on 0.91 m sections of row to quantify percentage defoliation and compared with narrow-band spectrometer measurements of reflectance of each plot.  Normalized difference vegetation index (NDVI) models composed of reflectance at all wavelengths were regressed against LAI to determine which wavelengths most accurately estimated changes in LAI. Linear and quadratic models were tested for their usefulness in estimating LAI. The results suggest that reflectance indices based on red edge measurements can offer accurate, consistent defoliation estimates, and could potentially increase defoliation efficiency and decrease costs.