Effects of shadowing types on ground-measured visible and near-infrared shadow reflectances

We analyzed the effect on the visible and near-infrared reflectance of shadows cast by a building, a conifer, and a hardwood on three different surfaces (grass, bare soil, and asphalt). Ground-measured radiances were collected with a radiometer measuring the SPOT-HRV bands and with a spectroradiometer measuring in the SPOT-HRV and LANDSAT-TM bands as well as giving the whole spectrum from 400 nm to 900 nm. We considered two types of reflectances: the ''sunlit reflectance'' and the ''shadow reflectance,'' which is calculated, using as irradiance, the radiance of the sunlit reference panel and of the shadowed reference panel, respectively. An analysis of variance showed that visible and near-infrared reflectances of shadows are significantly affected by the surface type and the shadow type. F-test values are higher for sunlit reflectances than for shadow reflectances as well as for near-infrared reflectances than for visible reflectances. Reflectances of hardwood shadows are generally higher and more variable than the ones of conifer shadows. Building shadows generally have lower reflectances than hardwood shadows and similar reflectances as conifer shadows. These reflectances were used to calculated two vegetation indices (NDVI and ARVI). They are also affected by the surface and shadow types. Whatever the object casting the shadows, spectra measured in the shadow are typical of the surface type (grass, bare soil, or asphalt). Further studies are needed on multiple shadows and penumbra effects, which are particularly important in the case of forest canopies. (C) Elsevier Science Inc., 1996.