A Mystery of the Inconsistency Between Microwave Polarimetric Observations and Radiative Transfer Simulations

A factor of 2 difference is found between microwave brightness temperature (TB) observations and simulations from vector radiative transfer models. A physical explanation of this difference is related to the calibration process. In principle, for the calibration of polarimetric instruments of a total power radiometer, the radiation should be only half of the values derived from Planck function but in the actual calibration process, a full value is used. When this full value is converted to a TB it is approximately, but not exactly, twice the TB simulated by a vector radiative transfer model. This inconsistency is verified by simulations of three conically scanning microwave imagers (WindSat, AMSR2 and MWRI) and one cross-track scanning microwave sounder (AMSU-A). This difference cannot be compensated simply by dividing a factor of 2 to the processed TBs, especially at high frequencies or at low physical temperatures due to the nonlinear property of Planck function.