Assessment of Quickbird Images to Monitor Early Post-Fire Vegetation Dynamics Applying Spectral Mixture Analysis

The mapping of the spatial development of vegetation cover during the first years after forest fires implies to delineate precise structural parameters of re-growing vegetation by high to very high spatial resolution. This study tested the ability of spectral mixture analysis (SMA), applied to high spatial resolution Quickbird and SPOT data, to produce realistic and meaningful endmember fractions for the study of post-fire vegetation regeneration in the early stage. The analysis was performed on a large burned area in Anchuras, Spain. The affected pre-fire vegetation of the selected burned area was comprised of maquis shrublands and pine woodlands. Special emphasis was put on the pre-processing of the data to retrieve accurate vegetation reflectance values. A linear SMA was applied to all data sets using image derived endmembers, which were identified as green vegetation (GV), soil and shade. Spatio-temporal patterns of vegetation cover changes could be explained on the basis of shade-normalized GV fraction images. Vegetation cover and its change could be successfully detected with high accuracy through the use of the Quickbird and SPOT derived GV fraction images. The preliminary results obtained in this study show promising perspectives on the use of SMA for the analysis of vegetation re-growth after forest fires; however, in order to fully validate the results obtained in this study, it would be necessary to perform this analysis on longer time frame.

REITHMAIER Lucia;  BARBOSA FERREIRA Paulo;  SAN-MIGUEL-AYANZ Jesus; 

2006-11-06

EU Life Project ForestSAFE, National Board of Forestry

JRC31420