Development of microwave vegetation index from multi-sensor observations

Microwave Vegetation Indices(MVIs) are newly developed vegetation indices using microwave remote sensing data (SMOS, WindSat, and AMSR-E, respectively). SMOS operates at L-band, in dual polarization and a range of viewing angles. WindSat is a satellite-based multi-frequency and multi-angle polarimetric microwave radiometer with the same transit time as SMOS. Establishing a synergetic vegetation index with WindSat and SMOS data will provide multi-angle and multi-frequency information in vegetation monitoring. As the deducations of previous indices are based on single senor configuration, it is necessary to demonstrate the applicability of the synergy for these two sensors. From the previous research, we found that bare soil emissivities at adjacent frequencies or angles exhibited a good linear relationship under Gaussian correlation function. Whether this relationship exists for L-band and C-band and for Exponential correlation function has to be deeply explored. With this objective, we have built a simulation database using the Advanced Integral Equation Model (AIEM) at L-band (SMOS, 1.4GHz, multi-angle) and C-band ( WindSat, 6.8GHz, 53.5°) in dual-polarization (v and h), respectively. The relationships of the surface emissivities are significantly different, while the fresnel reflectivity exhibits linear relationship. Hence, we consider using fresnel reflectivity to establish the bridge between different sensors. The radiation of vegetated land surface can be described by a 0th-order radiative transfer model. The total emission from ground can be rearranged into a linear model with the vegetation emission component as the intercept and the vegetation transmission component as the slope. With this linear relationship, the Fresnel reflectivity at one frequency can be expressed as a linear function of that at another frequency. And the intercept and the slope are the newly derived vegetation indices. They are mainly affected by vegetation properties and have - consistent spatial pattern with NDVI.