Nonlinearity is one of the most important errors in calibrating a microwave radiometer. It cannot be corrected in orbit through direct measurements, and its correction usually uses a set of coefficients derived from prelaunch calibration tests, namely, thermal vacuum test. The main difference between prelaunch and on-board calibration of a microwave radiometer is that of the cold calibration target used in two-point calibration. On board calibration usually uses the cold sky, which is also called cosmic background temperature of 2.7 K, as the lower end in the response of a microwave radiometer, whereas in the prelaunch calibration, a blackbody at a temperature between 80 K and 100 K is used in the calibration practice. Therefore, the applicability of the nonlinear coefficients derived from the prelaunch calibration test for on-board calibration is a long-standing concern that has never been validated. In this study, a special program of prelaunch calibration was designed, adding a validation blackbody target, which works at ambient temperature of 180—200 K in the thermal vacuum chamber, in addition to the traditional cold blackbody target, a warm blackbody target, and a varied-temperature blackbody target. This validation target provides an opportunity to compute nonlinear coefficients directly through three-point calibration and to evaluate the errors in applying the nonlinear coefficients by the traditional calibration method. Furthermore, a novel method is invented through extrapolating the measurements to the cold sky temperature by the equation of three-point calibration during a scan. It serves as a virtual on-board calibration target to calibrate the radiometer and to produce nonlinear coefficients combined with three other calibration targets. The nonlinear coefficients of the traditional and novel methods are compared, and the two groups of nonlinearities are applied to the same data sets. Thus, the errors in applying the nonlinear coefficients by the traditional calibration method can be evaluated. The results show a deviation between the two group of nonlinear coefficients; the greater absolute value of the nonlinear coefficients indicates greater error. The nonlinear coefficients of the traditional and novel methods are compared, and discrepancies are found in applying the two groups of nonlinearities to the same data sets. The finding implies that errors exist in the prelaunch coefficients when using the traditional method. The novel method simplifies the procedure of prelaunch calibration and improves the accuracy of nonlinear coefficients using for on-board calibration.

Microwave radiometer calibration;FY-3E Microwave Humidity Sounder;Prelaunch calibration;Nonlinearity;On-board calibration