Leaf Area Index Estimation of Boreal and Subarctic Forests Using VV/HH ENVISAT/ASAR Data of Various Swaths

This paper demonstrates the potential of dual polarization synthetic aperture radar (SAR) images in the estimation of the leaf area index (LAI) of boreal forests. The SAR data do not suffer from the low sun elevation and frequent cloud cover, which often complicate the use of optical wavelengths for LAI retrieval in the area. The analysis was based on a large number of environmental satellite (ENVISAT) advanced synthetic aperture radar (ASAR) alternating polarization vertical polarization (VV)/horizontal polarization (HH) single-look-complex images covering several test sites, boreal and subarctic, during summers 2003-2006. The swath range from IS1 to IS7 was studied. In all test sites, a linear regression with the VV/HH backscattering ratio as the independent variable could typically be used for the estimation of LAI. All swaths could be used for the estimation, but larger incidence angles generally performed better. The best results were obtained for swath IS6. The swaths could be used also together, but better results were obtained using the diverse swaths individually. The LAI estimation error decreased essentially exponentially with the number of pixels averaged to give one backscattering value. The LAI estimation accuracy for a set of average quality ASAR images of swath IS6 reached 0.1 when the averaging number of pixels was 33 150, which would correspond to an area of about 2.2 km2 for images with no overlap. The spatial LAI estimation error decreased with the number of images covering the same area.