We propose to develop and implement a comprehensive Level 2 inter-calibration procedure between all spaceborne microwave radiometers that could reasonably be included in the Global Precipitation Measurement (GPM) Constellation. Our approach will be to adapt the vicarious calibration techniques that we have developed for previous spaceborne radiometers (SSM/I, TOPEX, GEOSAT Follow-On, Jason and WindSat) and apply them to the suite of radiometers expected to comprise the GPM constellation (SSM/I, AMSR, WindSat, CMIS and others). The vicarious calibration technique provides a means to transfer main beam brightness temperature calibration standards between spaceborne radiometers that operate at different frequencies, incidence angles, polarizations and/or orbit geometries. Using the technique, absolute calibration of the sensors has been demonstrated at the ~ 1K level. Relative calibration between sensors, as well as stabilization of a sensor's calibration over time periods of ten years or more, has been demonstrated at the ~ 0.3 K level.
In summary, the vicarious calibration technique consists of three principle steps. First, calibration algorithms are coded up for each sensor. These algorithms convert raw (Level 0) radiometer digital counts into Level 1 radiometric antenna temperatures - by correcting for on board calibration and other instrumental effects - and then into Level 2 main beam averaged brightness temperatures - by correcting for antenna pattern and spacecraft attitude effects. The algorithms will at first be based on the Algorithm Theoretical Basis Documents (ATBDs) that were developed by the cognizant instrument developers and will be adapted and corrected as needed. Step 2 involves the assembly of external, vicarious calibration measurements. This process is explained in detail below. To summarize, certain statistical properties of the brightness temperature measurements made over specific classes of radiometrically cold ocean and warm land regions can be predicted with high accuracy and extremely high precision. The vicarious calibration data allows for the identification of errors in sensor calibration. The third step involves characterization of the calibration errors, typically by appropriate sorting and binning of the results of Step 2, followed by an iterative refinement of the Level 1 and Level 2 algorithms to remove the errors. Adequate characterization of the errors is critical in order to determine which part(s) of the algorithms should be adjusted.
