Earth Observation Seminar
Wednesday, November 1, 2006
(3:30 p.m. - Stewart Center G52)
Potential for Measuring Deep and Undercanopy Soil Moisture: a Tower-Based Radar Prototype System and Measurement Results
Dr. Mahta Moghaddam
Electrical Engineering and Computer Science, University of Michigan, Ann Arbor
Knowledge of subsurface characteristics such as permittivity variations and layering structure could provide a breakthrough in many terrestrial and planetary science disciplines. For Earth science, knowledge of subsurface and subcanopy soil moisture layers can enable the estimation of vertical flow in the soil column linking surface hydrologic processes with that in the subsurface. Soil moisture in the root zone can also provide a link between the carbon cycle and the water and energy cycles. For planetary science, determining the existence of subsurface water and ice is regarded as one of the most critical information needs for the study of the origins of the solar system.
The subsurface in general can be described as several near-parallel layers with rough interfaces. Each homogenous rough layer can be defined by its average thickness, permittivity, and RMS interface roughness assuming a known surface spectral distribution. As the number and depth of layers increases, the number of measurements needed to invert for the layer unknowns also increases, and deeper penetration capability would be required.
To nondestructively calculate the characteristics of the rough layers, a multifrequency polarimetric radar backscattering approach can be used. One such system is that we have developed for data prototyping of the Microwave Observatory of Subcanopy and Subsurface (MOSS) mission concept. A tower-mounted radar makes backscattering measurements at VHF, UHF, and L-band frequencies. The radar is a pulsed CW system, which uses the same wideband antenna to transmit and receive the signals at all three frequencies. Numerical results will be shown using actual radar data acquired with the MOSS tower radar system in Arizona and Oregon, and compared with in-situ measurements. It is shown that deep soil moisture can be obtained with good accuracy when the number of soil layers assumed in the model is consistent with that actually present in the ground down to the penetration depth of the signals.