Comment on “Partitioning tropical oceanic convective and stratiform rains by draft strength” by David Atlas et al.

Abstract

[1] The crux of this comment is to clarify attribution. If someone looks at a published graph and has a different interpretation from that of the original authors, that new interpretation should not be attributed to the original authors. [2] The references given by Atlas et al. [2000] regarding distinct convective and stratiform reflectivity Z to rain rate R relations attributed to Yuter and Houze [1997] are incorrect. The convective and stratiform Z-R relations labeled as those of Yuter and Houze in Table 3 (incorrectly referred to there as 1996) and mentioned in the text on pp. 2262, 2264, and 2265 of Atlas et al. [2000] do not appear anywhere in the Yuter and Houze [1997] paper. The Yuter and Houze [1997] study used two-dimensional (2-D) particle probe drop spectra collected by the National Center for Atmospheric Research Electra aircraft during the Tropical Ocean-Global Atmosphere (TOGA) Coupled Ocean-Atmosphere Response Experiment (COARE) to compute Z and R values over 6 s intervals of flight track ( 1400 L sample volumes). These data were classified into convective and stratiform subsets using radar data obtained by the National Oceanic and Atmospheric Administration WP-3D airborne radar. Segments of the Electra flight track were categorized by noting when the Electra flew within regions of convective and stratiform precipitation classified according to a radar reflectivity texture algorithm based on the paper by Steiner et al. [1995]. The main conclusion of the Yuter and Houze [1997] study was that the populations of radarclassified convective and stratiform Z-R points overlapped in dbZ-log R space and did not form two statistically distinct populations. [3] The previous statements do not rule out the possibility that different physical processes can yield different drop size distributions (DSD) [e.g., Braun and Houze, 1994, pp. 2749–2750]. However, in the large areas of radar-classified convective and stratiform precipitation examined in this study, the physical processes mixed, exhibited natural variation, and were subject to sampling error to such a degree that the Z and Rvalues associated with the DSD samples in the convective and stratiform populations did not form distinct populations. A classification method that can truly distinguish between dominant precipitation growth by vapor deposition versus accretion (riming and collection/coalescence) would have the property of yielding distinct populations in lowrain-rate regions of dbZ-log R space where both processes are known to occur. 2. Follow-up