Author
KURUM, M - National Aeronautics And Space Administration (NASA) | |
LANG, R - George Washington University | |
O'NEILL, P - National Aeronautics And Space Administration (NASA) | |
JOSEPH, A - National Aeronautics And Space Administration (NASA) | |
Jackson, Thomas | |
Cosh, Michael |
Submitted to: IEEE Transactions on Geoscience and Remote Sensing
Publication Type: Peer Reviewed Journal Publication Acceptance Date: 7/1/2009 Publication Date: 11/10/2009 Citation: Kurum, M., Lang, R., O'Neill, P., Joseph, A., Jackson, T.J., Cosh, M.H. 2009. Estimation of forest canopy attenuation at L-band by a time-domain analysis of radar backscatter response. IEEE Transactions on Geoscience and Remote Sensing. 47:3026-3040. Interpretive Summary: For the remote sensing of soil moisture, the presence of trees causes attenuation to the soil emission underneath. For accurate soil moisture retrievals through a forest canopy, the correction for the vegetation effects is needed. A new technique for determining the canopy attenuation using the measured stepped frequency radar backscatter response was proposed. The technique was applied to truck based L-band stepped frequency radar data collected over stands of deciduous Paulownia trees under various physical conditions. Using this procedure, time-domain analysis for various cases was carried out in order to understand the backscattering sources within a forest canopy and their effects on the transient response of the backscattered field. The technique developed here can be adapted to the airborne or spaceborne geometries using similar instruments. Having the forest attenuation information on a global basis will help to extend accurate soil moisture retrievals from global microwave mission to more areas of Earth’s surface than are currently feasible. Technical Abstract: A new technique for determining the canopy attenuation, using the measured stepped frequency radar backscatter response, is proposed. It employs details found in a transient solution where the canopy (volume scattering) and the tree-ground (double interaction) effects appear at different times in the return signal. The influence that these effects have on the expected time-domain response of a forest stand is characterized through numerical simulations. A coherent scattering model, based on a Monte Carlo simulation, is developed to calculate transient response from distributed scatterers over a flat surface such as a forested terrain. The proposed technique is based on separating the backscattering sources within a forest canopy in the time-domain response. The frequency correlation functions (FCF) of the volume scatter and the double interaction terms are computed by the distorted Born approximation. A ratio of these FCF’s is formed and compared with data at a set of frequencies over the decorrelation bandwidth of the returns. The resulting system of equations only depends on the canopy thickness, the canopy attenuation and a combined parameter involving the forest scattering coefficients and the ground reflectance. A least square method is used to solve for the attenuation and the combined parameter assuming the canopy thickness is known a priori. The technique was used with L-band data collected over deciduous trees to verify the algorithm results match the simulated data. |