Careful analysis of data reveals no shrinkage of tropical cloud anvil area with increasing SST

FEBRUARY 2002 AMERICAN METEOROLOGICAL SOCIETY | L indzen et al. (2001, hereafter LCH) present observational analysis suggesting that the area coverage of anvil clouds associated with tropical convection is less extensive when sea surface temperatures (SSTs) are higher. This conclusion is based on a negative correlation between the fractional area coverage by high clouds and the average SST under the clouds. LCH observe this correlation in about 20 months of Geostationary Meteorological Satellite (GMS) radiance data for the oceanic regions in the domain between 30°S–30°N and 130°E–170°W. LCH then used this correlation to hypothesize that increased SST leads to reduced area of anvil cloud associated with convective cores in the Tropics. They then concluded that reduced anvil cloud would give rise to a negative feedback through reduced upper-tropospheric water vapor and diminished greenhouse effect. We have obtained the cloud data from LCH and have investigated the geographical patterns of cloud fraction variation that are associated with variations in the cloud-weighted SST. These patterns show that the changes in cloud-weighted SST for clouds with emission temperatures less than 260 K are dominated by cloud variations in subtropical latitudes. These variations are geographically separated from the variations in the deepest convective clouds (those with emission temperatures less than 220 K), which occur only near the equator where surface temperatures are highest. LCH assume that the coldest clouds are representative of the convective cores that feed anvil clouds. In this article we show that the deep convective cores are separated by more than 1000 km from the clouds that are associated with most of the variation in cloud-weighted SST. The explanation for the correspondence between cloud-weighted SST and cloud area is thus a shift in the latitude or longitude of the cloudiness and not a change in the relation between deep convective and associated tropical anvil cloud amounts. These latitude and longitude shifts are associated with meteorological forcing and not with SST forcing. Much of the meteorological forcing seems to originate in the extratropics and is probably unrelated to tropical SSTs. If the SSTs and the nearequatorial clouds remain fixed and constant, any variation in subtropical cloud fraction will produce the negative correlation on which LCH base their negative feedback hypothesis.