HIGH DIMENSIONAL REFLECTANCE ANALYSIS OF SOIL ORGANIC-MATTER

Recent breakthroughs in remote-sensing technology have led to the development of high spectral resolution imaging sensors for observation of earth surface features. This research was conducted to evaluate the effects of organic matter content and composition on narrow-band soil reflectance across the visible and reflective infrared spectral ranges. Organic matter from four Indiana agricultural soils, ranging in organic C content from 0.99 to 1.72%, was extracted, fractionated, and purified. Six components of each soil were isolated and prepared for spectral analysis. Reflectance was measured in 210 narrow (10-nm) bands in the 400- to 2500-nm wavelength range. Statistical analysis of reflectance values indicated the potential of high dimensional reflectance data in specific visible, near-infrared, and middle-infrared bands to provide information about soil organic C content, but not organic matter composition. Although reflectance in the visible bands (425-695 nm) had the highest correlation (r = -0.991 or better) with organic C content among the soils having the same parent material, these bands also responded significantly to Fe- and Mn-oxide content. For soils formed on different parent materials, five long, middle-infrared bands (1955-1965, 2215, 2265, 2285-2295, and 2315-2495 nm) gave the best correlation (r = -0.964 or better) with organic C content. Several wavebands were identified in which the soils were separable, but the reflectance response was dominated by soil factors other than organic matter content, indicating that choice of wavebands should not be based on spectral curve separability alone.