Search

A 1-D variational (1DVAR) retrieval technique has been developed for obtaining temperature and humidity profiles from observations of the Ground-Based Scanning Radiometer (GSR) operating at millimeter-wavelengths. The GSR was deployed in two Arctic experiments held at the Atmospheric Radiation Measurement Program in Barrow, Alaska. Temperature and humidity profiles retrieved with the 1DVAR technique are compared with simultaneous radiosonde observations (RAOBs) during the Radiative Heating in Underexplored Bands Campaign (February-March 2007). Examples and statistical results are presented and discussed to demonstrate the achieved retrieval accuracy and vertical resolution. The 1DVAR retrievals based on GSR observations improve the NWP background up to 5 km, particularly in the lower 3 km. The present implementation achieved an root-mean-square (rms) error with respect to RAOB within 1.5 K for temperature and 0.10 g/kg for humidity profiles of up to 5 km in height, with 2.9 and 2.0 degrees of freedom for signal, respectively. Using the interlevel covariance definition of the vertical resolution, the 1DVAR retrievals showed a < 1-km vertical resolution of up to 5 km for both temperature and humidity profiles. The integrated water vapor obtained from the retrieved humidity profiles showed an rms accuracy within 0.10 kg/m², with small bias (< 0.01 kg/m²) and excellent correlation (0.96). [ABSTRACT FROM AUTHOR]

During the Radiative Heating in Underexplored Bands Campaign (RHUBC), held in February-March 2007, three millimeter-wave radiometers were operated at the Atmospheric Radiation Measurement Program's site in Barrow, Alaska. These radiometers contain several channels located around the strong 183.31-GHz water vapor line, which is crucial for ground-based water-vapor measurements in very dry conditions, typical of the Arctic. Simultaneous radiosonde observations were carried out during conditions with very low integrated-water-vapor (IWV) content (< 2mm). Observations from the three instruments are compared, accounting for their different design characteristics. The overall agreement during RHUBC among the three instruments and between instruments and forward model is discussed quantitatively. In general, the instrument cross-validation performed for sets of channel pairs showed agreement within the total expected uncertainty. The consistency between instruments allows the determination of the IWV to within around 2% for these dry conditions. Comparisons between these data sets and forward-model simulations using radiosondes as input show spectral features in the brightness-temperature residuals, indicating some degree of inconsistency between the instruments and the forward model. The most likely cause of forward-model error is systematic errors in the radiosonde humidity profiles. [ABSTRACT FROM AUTHOR]