Reply to comment by Roberto Rondanelli and Richard S. Lindzen on “Variations in convective precipitation fraction and stratiform area with sea surface temperature”

[1] Lindzen et al. [2001] reported that cirrus anvil cloud fraction normalized by deep convective core cloud fraction decreases with cloud-weighted sea surface temperature (SST) by about 20%/K, while Su et al. [2008] showed tropical-mean upper tropospheric (UT) cloud fraction normalized by precipitation decreases with cloud-weighted SST more slowly at around 2%/K. Throughout the text, we use ‘‘%/K’’ to indicate the relative change of normalized cloud fraction per degree change of SST. In their comments to Su et al. [2008], Rondanelli and Lindzen [2009] (hereafter RL09) examined several methodological choices that may contribute to the reported fractional decreases. First, they presented a regression analysis of normalized anvil cloud fraction onto SST using binned cloud fraction data within 0.5 C of SST and 1 1 gridded data. Both binned and gridded regressions show that anvil cloud fraction normalized by convective core cloud fraction decreases with SST at a rate of approximately 20%/K. In their analysis, cirrus anvil fraction is based on brightness temperature (BT) from the Tropical Rainfall Measuring Mission (TRMM) Visible and Infrared Scanner (VIRS) with values between 220 K and 260 K. Their convective core cloud fraction, which served as a measure of convective mass flux, is based on lower brightness temperature (BT 27 C, but decreases with SST 20 mm/d. This ‘‘saturation’’ behavior of cloud fraction with precipitation may contribute partly to the negative slope of precipitation-normalized cloud fraction versus SST. On the other hand, we notice that there is a high occurrence of large cloud fraction when precipitation is zero or very small. Thus, if one performs a linear least squares fit to the full range of cloud fraction versus precipitation, one obtains a nonzero intercept, corresponding to the fact that some high clouds are present regardless of precipitation. These high clouds may not be of convective origin and complicate the relation between the precipitationnormalized cloud fraction and SST. JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 115, D06203, doi:10.1029/2009JD012872, 2010 Click Here for Full Article