Testing the utility of multi-angle spectral data for reducing the effects of background spectral variations in forest reflectance model inversion

This investigation uses multiple-view-angle spectral data to estimate forest characteristics by inverting a geometric-optical reflectance model. The emphasis of the work was to assess to what extent multi-angle data can be used to help reduce the difficulties in model inversion, caused by variations in the background spectral reflectances. Inversions were tested with both simulated reflectance data and multi-angle Advanced Solid-state Array Spectrometer (ASAS) data. The background reflectances were included in the inversion process, with the assumptions that these are independent of viewing geometry, and the multi-angle data were acquired on a single date, with a constant solar zenith angle. Inversions using multi-angle data gave more accurate retrieval of stand characteristics than inversions using only nadir data; multi-angle data helped to reduce the problem of local minima in the inversion by providing additional information, with uncorrelated samples having the greatest utility. The utility of inversion was dependent on specification of the correct general range of background reflectances as constraints in the inversion process. Evaluation of the effects of errors in both the reflectance data and the ancillary data highlighted some important difficulties encountered in practice. The results indicate that multi-angle data have potential for improving the accuracy of forest characteristics derived by inversion, but further testing seems essential to better understand the limits of the method in practice. (C)Elsevier Science Inc., 2000.