Radiometric features of land surfaces at 37GHz

Summary form only given. Microwave emission from the Earth´s surface was found to be sensitive to land characteristics on both local and global scales. In particular, microwave sensors have shown a high sensitivity to certain features of the soil, as well as those of snow and vegetation, which are of great interest in hydrology, meteorology, climatology and - although to a lesser extent - in agriculture as well. The main parameters influencing microwave emission, and the most important targets of remote sensing of land are soil moisture, vegetation cover, plant water content, snow water equivalent and depth. Theoretical studies and field experiments, carried out by several research groups in order to relate microwave emission from land to physical parameters of soil, snow-pack and vegetation, have revealed different sensitivities of microwave sensors to these land parameters depending on observation parameters (frequency, polarization, incidence angle). The IFAC Microwave Remote Sensing Group has a long-time experience in investigating natural surfaces by using ground based and airborne microwave radiometers in a frequency range between 1.4 and 37 GHz. The research led to determining the radiometric behavior of the basic types of land surfaces and to establishing empirical or semi-empirical rules for classification of land surface and estimation of geophysical parameter corroborated by model analyses. Emission at millimeter-waves is characterized by lower penetration depth inside the observed bodies with respect to the lower frequencies, but the use of sensors operating in this frequency range allows higher ground resolution (or lower antenna sizes), which is a significant advantage from space. Emission from bare soil is essentially determined by the surface characteristics of this medium (moisture and roughness), while radiation from vegetation and snow is emitted by volumetric processes. For example, experimental studies interpreted with a radiative transfer mod- l demonstrated that the optical depth of agricultural crops in the cm and mm wavelength range increases with frequency and is related to the Plant Water Content (PWC) by a nonlinear relationship which shows saturation as the density of vegetation material (in general a sparse low density medium) increases toward the conditions of dense medium. The trend of brightness temperature as a function of PWC shows a maximum at cm wavelengths, denoting prevailing of absorption for low values of PWC and of scattering at the higher values. Instead, in the mm range scattering prevails and emission decreases as soon as the vegetation material increases. A typical dense medium (with density higher than 10%) is represented by terrestrial snow. For dry snow, emission at 37 GHz regards a snow layer of several decimeters depending on snow density and grain size and is much higher at lower frequencies. Thus, the use of a dual frequency system such as the one available in several satellite sensors (19 and 37 GHz) makes it possible estimating snow depth. If snow becomes wet the loss factor increases dramatically and radiation is emitted by a very thin surface layer even at the lower frequencies, so that the estimate of snow water content seems to be practically impossible. Significant studies for discriminating land surfaces and estimating quantitative parameters with microwave passive sensors from space have been conducted using satellite data from SMMR, SSM/I, AMSR-E, and more recently, from AMSR2. For example, data, obtained on different global areas from SSM/I and AMSR-E pointed out a rather flat behavior of the brightness temperature (Tb) spectra from X- to Ka-bands on evergreen forests throughout the whole year, with different levels of brightness only due to surface temperature variations; whereas, on tundra, the spectral trend slightly increases in summertime and sharply decreases in winter. From these results we realized that, although bare and vegetated soils generally show di