Antenna spacing and pattern differences: Their impact in MIRAS reconstruction error

After 6 years in orbit, the European Space Agency (ESA) Soil Moisture and Ocean Salinity mission (SMOS) continues to provide high-quality, global maps of soil moisture and ocean surface salinity. Its payload MIRAS (Microwave Imaging Radiometer with Aperture Synthesis) is an L-band interferometric radiometer which achieves unprecedented resolution. It was successfully launched in November 2009 under the European Space Agency Earth Explorers program, and has since been acquiring high resolution images whose accuracy, precision and stability have been steadily improved. The present work analyzes in detail the observed reconstruction error in SMOS brightness temperature images, and its relationship with the antenna pattern differences between receivers. Such reconstruction error is defined as the spatial ripple that remains after an image reconstruction process, assuming perfect knowledge of the instrument and no calibration errors. The error so defined is therefore completely independent of the radiometric noise or the presence of modeling errors, being then solely related with the reconstruction process. Antenna pattern differences, up to 5 % of the nominal peak value, have a noticeable impact in the observed reconstruction error. Moreover, this impact is more significant when the spacing between elements is above the Nyquist alias-free separation, as is the case in SMOS. In this paper the behavior of the reconstruction error is parametrized in terms of antenna differences and element spacing. All simulations have been performed with the best possible model of the instrument in full polarimetric mode. Although the analysis here is done for MIRAS antennas, it is intended to provide a guidance for a future SMOS follow on mission.