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

Abstract 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.