A test of detecting spring leaf ̄ ush within the Alaskan boreal forest using ERS-2 and Radarsat SAR data

This study tests whether ERS-2 and Radarsat Standard Beam SAR data could be used to detect spring green-up of broadleaf forests on ̄ oodplains in interior Alaska. SAR data were acquired for preand post-budburst dates during the spring of 1997, 1998 and 1999. The images were radiometrically calibrated todB and then geometrically co-registered to a colour infrared Landsat TM image. Ten broadleaf and ten conifer ̄ oodplain forest polygons were then delineated from the Landsat TM image for the extent of each SAR image. The mean dB from each broadleaf and conifer stand was computed from each SAR image to test whether there was a consistent change in meandB associated with spring leaf-out of broadleaf trees. There was no evidence that the 1997 or 1998 ERS-2dB SAR data were useful for detecting spring leaf ̄ ush. The 1998 Radarsat broadleaf stands all increased in dB after budburst, but this was attributed to an increase in surface moisture from rain, rather than due to spring leaf ̄ ush. The 1999 Radarsat stands all decreased in SARdB after budburst, while control conifer stands increased and decreased in dB during the same time period. Confounding factors such as changing surface moisture, and possibly freezing/ thawing eVects may mask any change in ERS-2 or Radarsat SARdB due to spring leaf ̄ ush.