Stimulated by the infrared iris hypothesis proposed by Lindzen et al. (2001; LCH2001) and following critical works, we approached the issue by statistically examining the variations of tropical rain characteristics associated with SST. By utilizing TRMM Precipitation Radar (PR) data, not only we can directly classify convective rain and the stratiform rain, but also we can know the precipitation top heights (PTH). It is assumed that deep stratiform cloud cover would be in a positive correlation with the deep stratiform rain (PTH>5km; DSR) cover.

All path data of PR2a25 version 5 for the period from January 1998 to December 2000, except for January 1999 and September to November 1999 where data were missing, were utilized in a pixel by pixel basis. Classification between the convective rain and the stratiform rain depends basically on the rain flag provided by the PR2a23 data set. Warm isolated rain pixels were reclassified from the stratiform to convective rain, following the suggestion by Schumacher and Houze (2003). PTH for each pixel was identified with a threshold of 0.3mm/hr. Before the statistics, daily gridded rain data were made into 2.5deg x 2.5deg longitude-latitude grids. Rain pixels over the ocean and over the land are separately counted utilizing the surface data for each pixel. In order to obtain the information for the local sea surface temperature where the rain was observed, we utilized the 3-day running mean gridded sea surface temperature estimated form the TRMM Microwave Imager data with Shibata's algorithm. Next, we binned the daily 2.5degx2.5deg rain data into 1degC SST bins for the same day and same grid. Then we made statistics of the coverage, amount, and intensity of deep convective rain (PTH>8km; DCR) and DSR over ocean in relation to SST for four tropical regions(Region B: 30N-30S, 130E-170W, GMS region in LCH2001, Region D: 15N-15S, 125E-90W, North Pacific ITCZ, Region E: 5N-10S, 60E-100E, equatorial Indian Ocean, Region F: 20N-20S, 0-360E, entire tropics). Counted rainy 2.5deg grid numbers are 212599, 115650, 48888, 646099, for regions B, D, E, F, respectively.

Primary results are as follows: Statistically significant results are obtained over the warmer water (SST>25C). The coverage and the amount of DCR and DSR both significantly increase with SST (see Figure), except for over the Indian Ocean, where the DSR amount shows a negative correlation with SST. On the other hand, intensities of DCR and DSR have a tendency to decrease with SST. Over the colder SST in the GMS region (region B here), their relation differs completely. The negative correlation of the anvil cloud cover with the cloud weighted mean SST suggested by LCH2001 may be an artifact by averaging over the GMS region mixing properties over the subtropics and the deep tropics, as already pointed out by Hartmann and Michelsen (2002).

However, when we examine the ratio of DSR cover to DCR cover or amount (Figure), we find that these values have significant negative correlation with warmer (>26C) SST. The decrease rate of DSR-cover/DCR-cover and DSR-cover/DCR-amount for SST>26C are, 12-25%/1degreeC and 9-22%/1degreeC, respectively, in good agreement with what suggested by LCH2001. The ratio of DSR-amount versus DCR-amount also significantly decreases with SST. It is notable that these DSR properties normalized by DCR show very similar behavior in relation to SST independent of the regions.

In conclusion, over the tropical warm ocean (SST>25C), while DCR and DSR intensity both slightly decrease with SST, DSR cover mostly increase with an exception over the Indian Ocean. DCR and DSR amount, as a result, have positive correlation with SST. Therefore, although DSR cover normalized by DCR cover or by DCR amount both decrease with SST, it does not result in the reduction of DSR coverage and most probably the high stratiform cloud amount. Instead, an increase of stratiform cloud cover with SST is statistically suggested. However, as shown with the Indian Ocean case, the relationship between deep stratiform cloud and SST are also controlled by factors other than convective enhancement, such as surface solar irradiance, surface winds and the oceanic mixed layer depths. In order to capture the cloud-SST relationship accurately, we have to understand the air-sea interaction. On the other hand, very robust relationship of DSR properties normalized by DCR values, free from regional differences, is suggested. These relations may become a clue to the variation of rain properties associated with the climate change in future studies.

References:

Hartmann, D. L., and M. L. Michelsen, 2002: No evidence for iris. Bull. Amer. Meteor. Soc., 83, 249-254.

Lindzen, R. S., M. D. Chou, and A. Y. Hou, 2001: Does the earth have an adaptive infrared iris? Bull. Amer. Meteor. Soc., 82. 417-432.

Schumacher, C., and R. A. Houze, Jr.: The TRMM precipitation radar's view of shallow, isolated ran, J. Appl. Meteor., 42, 1519-1524.