Two-Level Forest Model Inversion of Interferometric TanDEM-X Data

Paper in proceeding, 2014

In this work, a two-level model is inverted and estimates of ground-to-vegetation ratio and level distance are obtained. Interferometric TanDEM-X data acquired over Remningstorp, a hemi-boreal test site situated in southern Sweden are used. Interferograms are flattened through the removal of ground phase, computed from a highresolution digital terrain model. 32 circular, 40-metre radius field plots are used to show that the two estimated model parameters are well-correlated with the lidar-derived parameters vegetation ratio (VR, a canopy density measure) and 95 th -percentile forest height (H95), with the respective Pearson correlation coefficients -0.83 and 0.99. Two new quantitites, IVR and IFH95, are introduced as interferometry-based estimates of VR and H95, with RMSE values 0.68 m and 4.50 percentage points, respectively.