Variability of spectral reflectance and vegetation indices in tallgrass prairie: spatio-temporal analysis using semivariograms and close-range remote sensing

The effects of fire and other disturbances on spatial patterning of vegetation in tallgrass prairie has been extensively studied using remote sensing. Data used in these studies is frequently typically collected from operational remote sensing satellites with spatial resolution coarse relative to the grassland canopy. The authors use semivariograms calculated from close range remote sensing instruments to investigate the spatial structure of spectral reflectance and vegetation indices in tallgrass prairie at finer spatial resolution more appropriate to the developing canopy. Data were collected at the Konza Prairie Research Natural Area near Manhattan, KS. Two watersheds were surveyed; one unburned the other burned just prior to data collection. Spatial dependence and anisotropy for three spectral bands (TM3, TM4, and TM5) and two vegetation indices (NDVI, NDVIc) were calculated for three dates distributed throughout the 1997 growing season. Results show that both anisotropy and dominant scale of spatial dependence vary seasonally and across spectral region