The Microwave Radiation Imager (MWRI) is a main payload of the second-generation Chinese polar-orbiting meteorological satellite Fengyun-3. The MWRI observes the Earth radiation at 10.65 GHz, 18.7 GHz, 23.8 GHz, 36.5 GHz, and 89 GHz with dual polarization. Sensitivity is the minimum transformation amount of the brightness temperature that can be detected by the radiometer, that is, the equivalent noise temperature of the radiometer system; it is expressed as the Root Mean Square (RMS) standard deviation of the output temperatures when the radiometer observes a target with a fixed bright temperature. Therefore, sensitivity is a key parameter for evaluating the performance of the MWRI, and the long-term stability of this parameter could directly affect the application of remote sensing data. At present, the brightness temperature data of the hot load (black body) equipped on the MWRI are used to evaluate the on-orbit sensitivity of the instrument. However, when the spaceborne microwave imager is in flight, the hot load measurements could be affected by environmental noise, such as receiver temperature and ground Radio Frequency Interference (RFI), resulting in errors in the on-orbit sensitivity obtained by the RMS method. The validity of the Allan standard deviation method is studied to calculate the sensitivity and accurately evaluate the on-orbit sensitivity of the spaceborne microwave imager, and then the Allan method is used to evaluate the long-term stability of the on-orbit sensitivity of the MWRIs on board FY-3 B, C, and D satellites.Combining the ground and on-orbit observations of the MWRI, two methods of root mean square standard deviation and Allan standard deviation were used to calculate the sensitivities of the imager. The comparison found that: (1) The deviation between the results obtained by Allan method and the ground test is less than 0.03 K, indicating that under the vacuum calibration test, the Allan method can effectively calculate the sensitivity parameters of the radiometers. (2) When the observation area of the space-borne microwave imager is different, the on-orbit sensitivity obtained by Allan method is unchanged, indicating that the Allan method can eliminate the interference of RFI. (3) The on-orbit sensitivity obtained by Allan standard deviation is unaffected by the receiver’s ambient temperature change. Therefore, the Allan method could be used to calculate the on-orbit sensitivity of a space-borne microwave imager.The Allan method was used to analyze the long-term stability of the on-orbit sensitivity of MWRIs on board FY-3B, C, and D satellites, and concluded that: (1) The on-orbit sensitivities of the three microwave imagers were stable, and the stability of FY-3C and FY-3D MWRIs is better than that of FY-3B MWRI. The standard deviation of the on-orbit sensitivity of the FY-3B MWRI is ≤ 0.015 K, and the standard deviation of the on-orbit sensitivity of the FY-3C and FY-3D MWRIs is ≤ 0.01 K. (2) FY-3C and FY-3D MWRI sensitivities are the same, and the sensitivities of the 89 GHz receiving channels of the FY-3C and FY-3D MWRIs are significantly better than that of the FY-3B MWRI. The sensitivity of each receiving channel of the FY-3C and FY-3D MWRIs is better than 0.5 K, and that of the FY-3B MWRI is better than 0.6 K.The comparison of the ground test data and on-orbit observations confirmed that the Allan method can effectively calculate the sensitivity of the spaceborne radiometers. In addition, the calculation results of the FY-3 MWRIs show that the on-orbit sensitivity of the MWRIs is good, and the long-term on-orbit working state is stable.

FY-3 MWRI;Sensitivity;RMS method;Allan method;long-term stability