Deriving stand age distribution in boreal forests using SPOT VEGETATION and NOAA AVHRR imagery

Stand age distribution in boreal forests reflects past disturbances, which is a major determinant of the carbon strength (sink or source). We used mapped and dated fire scars as a reference and developed a remote sensing algorithm for mapping boreal forest stand age distribution that accounts for other stand-replacing disturbances such as harvesting. Considering its strong discriminatory capacity on vegetation type and stand age, a short-wave vegetation index (SWVI) (Remote Sensing of Environment 82 (2002) 95), calculated from the SPOT VEGETATION (VGT) near-infrared and short-wave infrared bands, was utilized to quantify forest regeneration following fire. Stand age of a pixel was then determined by its SWVI value in comparison with the known ones from the dated fire scars. A change indicator (Latifovic, R. (2001). VGT and AVHRR Manager User manual. Canada Centre for Remote Sensing, Natural Resources Canada. Ottawa, Ontario, Canada), which considered interannual changes in the Normalized Difference Vegetation Index (NDVI) from sequential NOAA Advanced Very-High Resolution Radiometer (AVHRR) imagery (1993-1998), was employed to differentiate areas with lower SWVI values resulting from biophysical factors other than disturbances. We evaluated the stand age distribution product using Landsat TM imagery and discussed the limits and constraints of the remote sensing approach of deriving stand age distribution in the boreal forest region. (C) 2004 Elsevier Inc. All rights reserved.