Abstract

In order to better understand the characteristics and physical-to-radiative relationships of frozen hydrometeors in hurricane systems, computed brightness temperatures (TB) from 10.7 to 183 +/- 10 GHz were compared with radiometric observations of Hurricane Erin (2001) from the NASA ER-2 aircraft... Show moreIn order to better understand the characteristics and physical-to-radiative relationships of frozen hydrometeors in hurricane systems, computed brightness temperatures (TB) from 10.7 to 183 +/- 10 GHz were compared with radiometric observations of Hurricane Erin (2001) from the NASA ER-2 aircraft. The focus was on the high frequencies (>= 85 GHz) that are particularly sensitive to frozen hydrometeors. In order to initialize the cloud profiles used in the radiative transfer calculations, data from airborne radars, dropsondes, and cloud models were used. Three different ice habit and size parameterizations were used with these cloud profiles to obtain the particle radiative signatures including (1) spherical particles with size distributions derived from in situ observations, (2) spherical "fluffy" snow and graupel particles with modified Marshall-Palmer size distributions, and (3) a non-spherical bullet rosette habit where the radiation attributes (scattering, absorption, and asymmetry properties) were computed using the Discrete Dipole Approximation. In addition, three different reflectivity to ice water content (Z-IWC) relationships were used with the three habit and size parameterizations to provide a measure of the sensitivity of the Z-IWC relationship. This work showed that both the scattering and asymmetry coefficients, along with the ice water content in each layer, play an important role in determining the resultant high frequency brightness temperatures. All low frequency (< 40 GHz) calculations matched the observations with correlation coefficients greater than 0.9. At higher frequencies (> 90 GHz), correlation coefficients ranged from 0.7 to 0.92. Comparing between the three ice habit and size parameterizations showed less than a 0.2 difference in correlation coefficient, while the comparisons between the three Z-IWC relationships caused changes of up to 0.15 in the correlation coefficients, but they had a significant effect on the mean differences between the observations and the calculations. Show less