Differences in vegetation indices for simulated Landsat-5 MSS and TM, NOAA-9 AVHRR, and SPOT-1 sensor systems

The objective of this study was to evaluate the effects of the different wavelength bands of the Landsat-5 MSS and TM, NOAA-9 AVHRR, and SPOT-1 sensors on two vegetation indices, the normalized difference (ND) and near IR to red ratio (RATIO). The study also demonstrates how vegetation indices for the Landsat-5 MSS and TM, and SPOT-1 systems may be estimated with NOAA-9 AVHRR data. Agronomic and spectral reflectance measurements of corn (Zea mays L.) canopies were acquired with an Exotech 20C spectroradiometer in field experiments at Purdue University (W. Lafayette, IN). The reflectance factor data were averaged into 10 nm bands over the 400-2400 nm wavelength interval. Each experiment included four rates of nitrogen fertilization (0, 67, 134, and 202 kg/ha) and three replicates. The vegetation indices were computed i) for ground-based sensors by integrating the reflectance factor data over the visible and near-IR bands of the four sensors and ii) for simulated satellite-based sensors by modifying the reflectance factors with the filter response of each sensor, atmospheric transmittance, and solar irradiance at the Earth's surface in each 10 nm waveband. Variability in the RATIO between the four sensor systems was greatest during mid-season when maximum amounts of green vegetation were present. Variability in ND for the four sensors was considerably less than for the RATIO and nearly constant for most of the growing season. Comparisons of predicted agronomic variables indicated that AVHRR data can estimate both of the vegetation indices of the MSS, and subsequently, agronomic variables as effectively as direct use of the vegetation indices of the MSS. The vegetation indices of all four systems were associated with similar amounts of variation in the examined agronomic variables. Thus, under similar viewing conditions, the AVHRR may complement measurements of the other sensor systems for monitoring surface features of the Earth. Studies are encouraged that address the effects of the dissimilar viewing conditions (orbital characteristics, spatial resolution, off nadir views, etc.) of these sensor systems on their respective vegetation indices.