A gronomy J our n al • Volu me 10 0 , I s sue 4 • 2 0 0 8 electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.T here is a growing library of work published on spectral refl ectance measurements in the visible and near infrared wavelengths to assess plant nutrient and water status, which follows decades of work based on the Normalized Diff erence Vegetation Index (NDVI) (Rouse et al., 1974) and related indices. To date, the use of indices that exploit refl ectance measurements in narrow waveband intervals, especially those involving the chlorophyll red-edge, have produced impressive results in characterizing plant nutrient status (e.g., Haboudane et al., 2002;Gitelson et al., 2003;Strachan et al., 2002). Key questions remain about application of these indices to detect crop stress. One is whether it is possible to detect plant stress for areas smaller than the GSD, or nominal sample size, of a given sensor. In the case of imagery, this is oft en referred to as subpixel detection. Surprisingly, while there has been substantial research based on linear mixing models (e.g., Goel et al., 2003;Okin et al., 2001;and Bannari et al., 2006) there has been little evaluation of the ability of vegetation indices to detect stress areas that are smaller than the GSD. Accurate subpixel stress detection would certainly benefi t and extend applications of space-born imaging spectrometers to crop assessment.Another concern in the application of the spectral refl ectance indices is how well they are able to diff erentiate water and nutrient stress. Some research indicates they do this quite well. Barnes et al. (2000) evaluated the use of refl ectance measurements at key wavelengths of 555 nm, 670 nm, 720 nm, and 790 nm. Among their fi ndings, they demonstrated that the use of one chlorophyll red-edge-based index (the Canopy Chlorophyll Content Index [CCCI], Fitzgerald et al., 2006) diff erentiated the low water/high N treatments from the low water/low N and the high water/low N treatments in cotton. Fitzgerald et al. (2006) also demonstrated signifi cant relationships between CCCI and leaf chlorophyll concentration or shoot N concentration in wheat that were independent of water stress. However, other work suggests that the spectral refl ectance indices may not always reliably separate nutrient and water stress. Clay et al. (2006) designed a factorial N and water experiment in corn. Th ey used canopy refl ectance to calculate NDVI, the Normalized Diff erence Water Index (NDWI) (Gao, 1996), and red-edge indices from Gitelson et al. (2005). Th e ANOVA results for the V8-V9 development stage showed that both NDVI and the red-edge indices were signifi cant for N treatment; the red-edge indices were also signifi cant for water. At the later V11-VT growth stage, red-edge indices were less signifi cant than NDVI for N treatment. At both stages, the NDWI was not signifi cant for water, and none of the indices evaluated were signifi ca...