Four existing algorithms are used or reconstructed to estimate the IWC and LWC in tropical cyclones. Among those algorithms, TRMM algorithm 2A12 directly gives the output of hydrometeor profile parameters. The 3 other methods have been developed to calculate IWC and LWC from TRMM 2A25 radar reflectivity and rain rate profiles. The first one is by using the empirical Ze-IWC (Black 1990) and Ze-LWC (Gamache et al. 1993) relationships (therefore Z-M method); the second (called 2A25 method) and third (called 2A25_iguchi method) are to calculate LWC and IWC from 2A25 rain rate and fall speed, the latter is obtained by using the empirical reflectivity-fall speed relationship in Marks and Houze 1987. By applying these 4 methods to the 1-year TRMM database, a large variable range of LWC and IWC can be seen among the 4 methods, although 2A12 profiles are close to Z-M profiles except for the lowest 2-3 km. In general, eyewall regions have much more precipitation ice and liquid water mass than inner rainband and outer rainband region; inner band regions have more than outer band regions. In eyewall regions, precipitation liquid water mass is larger than precipitation ice mass. In inner and outer rainband regions, the reverse is true according to some methods, but not all, an issue needing further study.
The shape of radar reflectivity profiles is investigated for the CAMEX-4 hurricane cases by using both EDOP and PR-2 (a precipitation Doppler radar on the DC-8 aircraft) data. By comparing the maximum vertical reflectivity profile in convective cells in hurricanes (e.g. Chantal, Erin, and Humberto 2001) and that in cumulonimbus cloud systems (September 19, 2001), it shows that the slope of the profile above freezing level decreases with height rapidly for both cases. This confirms that weak updrafts are characteristic of many oceanic regimes around the world (Zipser and Lutz 1994). Future work will focus on validating the shape of TRMM radar reflectivity profiles by using EDOP as high-resolution “truth” for the TRMM PR (Heymsfield et al. 2000).
Supplementary URL: