Use of Radarsat-2 and ALOS-PALSAR SAR images for wetland mapping in New Brunswick

Armand Larocque, University of New Brunswick
Brigitte Leblon, University of New Brunswick
Renata Woodward, Nature Trust of New Brunswick
Michael Mordini, Michigan Technological University
Laura Bourgeau-Chavez, Michigan Technological University
Antony Landon, Michigan Technological University
Nancy French, Michigan Technological University
Jessica McCarty, Nature Trust of New Brunswick
Tom Huntington, United States Geological Survey
Phil Camill, Bowdoin College

Abstract

Our study tests the use of dual-polarized (HH, HV) RADARSAT-2 C-band and ALOS-PALSAR L-band SAR images for mapping wetland areas in New Brunswick. The study also uses LANDSAT-5 TM and DEM data. The resulting maps were compared to GPS field data as well as to two wetland maps currently in use by the Province of New Brunswick. Overall the Random Forests classifier gave better classification accuracies than the maximum likelihood classifier. The comparison with the 146 wetland truth sites shows that 73.3% are correctly identified over the LANDSAT-5 TM classified image. For the SAR-based classified images, the number of correctly identified wetland ground truth sites is higher when the image acquired during the flooding is considered, the difference being higher with the ALOS-PALSAR images than with the RADARSAT-2 images. The number of correctly identified sites is the highest when both the ALOS-PALSAR images and RADARSAT-2 images are used (98.6%). These percentages of correctly identified wetland sites are well above of those computed using the DNR wetland and forested wetland maps (44.5 %).