Retrieval of forest biophysical parameters using physically-based algorithms

Radiative transfer (RT) models provide an improved theoretical and physical basis for deriving biophysical structural information compared with statistically-based empirical methods. It has been used in various studies, mainly in optical remote sensing applications. This study is to explore the potential of applying the physically-based approach to multi-sensor data in forest parameters estimation. Optical reflectance model and radar backscatter models were used to create a look up table by simulation of multi-spectral reflectance at LANDSAT bands and radar backscattering coefficients and the height of scattering phase center in L-band from the forest stands generated by a forest growth model. As a first step, a simulated data set was used to test the look up table method. The results showed that optical reflectance, radar backscatter and interferometric SAR signature have their own advantage in deriving different parameters, and combined use of these data improved the estimation results. In next step of our study, real data, such as LANDSAT data, ALOS PALSAR data will be used in look-up table inversion. The field measurements and parameters derived from lidar data will be used for assessing the accuracy of the forest biophysical parameters from the physically-based algorithms.