Determination of biophysical variables using remote sensing techniques

Monitoring rangeland characteristics is difficult because of the extensive nature of rangeland and its diversity in terms of vegetation types, soil and topography, weather conditions and management practices. Remotely-sensed images provide an excellent tool to evaluate changes in environment over time, and allows us to construct decision-making programs for different time frames. Empirical vegetation indices (VI) and canopy reflectance models have been used for the past 25 years to extract and interpret information about biophysical variables from remotely-sensed vegetation canopies. This study was conducted at the University of Nebraska Gudmundsen Sandhills Laboratory near Whitman, Nebraska, in June and August 1995 and 1996, to evaluate spectral reflectance curves and vegetation indices for grazed and ungrazed uplands and meadows in the Nebraska Sandhills. Canopy radiance of 128 quadrats (1 m2) located in the upland and meadow sites were measured with a Spectron SE 590 portable radiometer. Standing crop biomass, standing dead material, litter, leaf area index (LAI), percentage foliar cover, and soil moisture content also were estimated for each quadrat. Four vegetation indices, difference vegetation index, ratio vegetation index, normalized difference vegetation index, and soil adjusted vegetation index, were calculated. The main objectives were (a) to determine the spectral characteristics of subirrigated range sites (meadows) and upland pastures dominated by the sands range site; (b) to correlate the biophysical variables with the VI; (c) to identify a `broad-band VI\u27 and/or a `narrow-band VI\u27 for predicting biophysical variables; and (d) to compare green NDVI with red NDVI in predicting biophysical variables. Foliar cover had the highest number of significant correlations (p \u3c 0.05) with the different VI; however, the VI did not predict foliar cover, standing crop biomass, or LAI, reliably and consistently over time or space. None of the VI tested was better than any of the others in predicting the measured values of the biophysical variables. Results of this study showed that GNDVI was more sensitive than RNDVI in estimating standing crop biomass. The use of vegetation indices such as RNDVI and GNDVI have limited value in range management because of the inconsistency of the correlations among the sampling dates