On May 27 2008,the first satellite in the new Chinese Fengyun polar-orbiting satellite series,FY-3A,was successfully launched into a circular,sun-synchronous,near-polar and morning-configured(10:00 LST,mean equator crossing local solar time of the descending node) orbit,with an altitude of 836 km above the Earth's surface,and an inclination angle of 98.75°.The Microwave Temperature Sounder(MWTS) on board the FY-3A has four channels in the oxygen band,at frequencies ranging from 50.3 to 57.3 GHz,for temperature sounding from the surface,to about 1 hPa,due to the absorption and emission of microwave radiation by atmospheric oxygen.MWTS provides a total of only 15 field of views(FOVs) along each scan line.While the calibration accuracy of the MWTS's brightness temperature was well characterized during its prelaunch period using a two-point calibration algorithm,which converts the Earth scene,warm target and cold target counts into Earth scene radiances,the MWTS's on-orbit performance remains important after the launch of each satellite.This is a relative calibration method,and satisfies weather forecasting application,but it is not sufficiently suitable for climate application,as the requirements on satellite data calibration for climate studies are different from weather forecasting applications.Issues such as variable calibration accuracy associated with each satellite instrument and bias changes with respect to time due to the satellite orbital drift,warm target calibration problems and pass-band frequency shifts must be resolved,as they may be mistakenly interpreted as climate influences.In this study,Global Position System Radio Occultation(GPS RO) data are used to absolutely calibrate the MWTS data throughout all of 2010,based on two methods.The first method is to calculate the scan-dependent global mean differences between the MWTS upper air sounding channels 2-4 measurements and GPS RO simulations over ocean under clear-sky conditions.This can be expressed as a scan-dependent function μ₁^o(α)=T_b,MWTS^obs-T_b,1^sim,where α is the scan angle,T_b,MWTS^obs represents the MWTS brightness temperature,and T_b,1^sim represents the CRTM brightness temperature simulations with input GPS RO profiles.The second method is similar to the first,except for the fact that the simulated brightness temperature T_b,2^sim is derived through the linear regression of MWTS measurements and GPS RO simulations at each scan angle.The absolute calibration of the MWTS data is to subtract either μ₁^o(α) or μ₂^o(α).The effects of the above two calibration methods are examined by comparing the biases of the MWTS observations with respect to the National Centers for Environmental Prediction(NCEP) Global Forecast System(GFS) simulations,with and without calibration.A unique feature of a cross-track scanning radiometer instrument is that the impacts of side-lobe effects and spacecraft radiation on the MWTS observations are higher at larger scan angles,and these effects may be difficult to take into a full account in the calibration process.In short,the biases are channel dependent,and all three of the channels 2-4 are of negative biases.It can also be seen that the bias at channel 3 is the largest.Moreover,the biases are all seasonal dependent.It is also found that the MWTS global biases and standard deviations are both reduced.The channel and time dependences of the MWTS biases are eliminated as well.It is well know that a single satellite has a limited life span,varying from 5 to 13 years.However,long-term climate microwave temperature sounder data records may be obtained if the measurements from different satellites are linked together.The absolute calibration using GPS RO data is definitely one of the more effective methods.This paper also represents a preliminary study toward establishing a refined satellite microwave temperature sounding instruments fundamental climate data records (FCDR).Further work includes diurnal correction using GPS RO observations to remove the impact of orbital drift.The impact of the first guess and the vertical error covariance matrix on the quality of the temperature and water vapor profiles retrieved by a 1D-Var approach require assessment.