L-Band Active and Passive Sensing of Soil Moisture through Forests

Remote sensing of soil moisture under a mature forest canopy by active and passive techniques is discussed. Emphasis is placed on examining the sensitivity of the backscattering coefficient and the brightness temperature to soil moisture variations. The effects of the underlying surface roughness of the forest floor on these sensitivities are addressed. The backscattering coefficient from the forest is modeled using the distorted Born approximation. Using this method the backscatter response is decomposed into a direct or volume scattering component, an interaction component between the vegetation and the average surface, an interaction component between the vegetation and the surface fluctuations and finally direct backscatter from the surface attenuated by the vegetation. Peak's method is used to determine the brightness temperature. The models are validated using ground truth and remote sensing data taken during the Forest Ecosystem Dynamics (FED) experiment conducted near Rowland, Maine, USA in 1990. During this period, the AIRSAR synthetic aperture radar and the Push Broom Microwave Radiometer overflew the Rowland area collecting data. Once the model predictions are compared with the passive and active data, the results of a sensitivity analysis involving soil moisture and surface roughness will be presented