Comparison of Observations of the Spatial Distribution of Water Vapor Derived from Radar and Radiometer Measurements During the DYNAMO Field Campaign

The field campaign of DYNAMO/CINDY2011 took place in the central equatorial Indian Ocean between September 1, 2011 and January 5, 2012. The experiment was primarily designed to improve understanding of the Madden-Julian oscillation (MJO) in that region. Observations of vertical moisture profiles, cloud structure, precipitation processes and the planetary boundary layer are necessary to improve understanding of MJO initiation. A number of remote sensing instruments, including NCAR's S-PolKa (dual-wavelength S- and Ka-band) radar and the University of Miami's microwave radiometers, were deployed to estimate water vapor and cloud structure.

This work focuses on retrieval of integrated water vapor (IWV) from measurements by the K-band radiometer and the S-PolKa radar at various elevation and azimuth angles during DYNAMO. Due to the low elevation angles at which the radiometric data were taken, the consistency and accuracy of the data will be studied to separate the effect of any land contamination from actual meteorological effects in which the azimuth dependence of brightness temperature may reflect water vapor convergence induced by the diurnal heating of the atoll. This will be confirmed through a diurnal cycle analysis, by comparing the radar and radiometer data, as well as through incorporation of the surface meteorological data and the radiosondes released from the DOE ARM site approximately 10 miles away from the SPol-Ka site. Radiometric measurements were performed at 23.8 GHz, which is affected mostly by water vapor, and at 31 GHz, which is primarily sensitive to cloud liquid water. Brightness temperatures at these two frequencies are known to exhibit a linear relationship to IWV during clear sky conditions and a non-linear relationship during cloudy conditions. Based on the linear relationship, IWV during clear sky conditions will be retrieved using a statistical regression method. The estimation technique will then be modified for cloudy conditions. Contributions due to land contamination at low elevation angles will be modeled and removed from brightness temperatures. The estimated slant path water vapor will be projected onto a 2D plane. The retrieved IWV will then be used to estimate the 3D distribution of water vapor density in the volume of the atmosphere contributing 95% of the radiation to a radiometric measurement. This will be done by extending the 2D retrievals corresponding to each azimuth angle into a 3D grid using radiometric weighting functions and geometrical projection. The radiometer-estimated water vapor distributions will then be compared with those estimated by the S-PolKa radar.