Future prospects - Ice Ages and Interglacials: Measurements, Interpretation, and Models - page 337

According to Lindzen, global climate models are not good at estimating the
coupling between tropical and extratropical regions and, therefore, do not allocate
the global distribution of water vapor accurately; this has a profound effect on the
putative heating effect of increased CO 2 . It seems possible that global warming
might decrease the humidity of air descending above the desert areas of the Earth
and, since these regions are by far the most sensitive to changes in humidity, they
would counterbalance the smaller heating effect of increased humidity in regions
where the humidity is higher. The regional effects of changes in humidity far out-
weigh the effects of changes in net global humidity. The expansion of already dry
regions is more important than net humidifying of the globe. Net moistening of
the Earth could have a negative water vapor feedback if most of that moistening
occurs in already moist regions. Climate models that employ average humidity for
the whole Earth are overly simplistic.
The effect of water vapor feedback in amplifying global warming produced by
the CO 2 concentration increasing requires an understanding of the distribution of
humidity changes resulting from warming; global average humidity change is not
good enough. Of course, one must still cope with the problem of changes in
cloudiness, despite being armed with a thorough understanding of the regional
dependence of humidity change. Lindzen et al. (2001) and Lindzen (2007) studied
this and came up with four major points: (1) cloud and water vapor feedbacks are
intimately connected; (2) feedbacks are primarily associated with changing areas of
moist and cloudy regions vs. regions that are dry and cloud free (as opposed to
mean humidity); (3) models must have spatial and temporal scales (5-10 km and
hours) characteristic of clouds in order to evaluate feedbacks; and (4) the effect of
cumulus activity must be included. A simplistic model that merely treats humidity
as a global average which increases when surface temperatures rise, which ignores
regional changes in humidity, and which treats clouds crudely will always over-
estimate the temperature rise due to increased CO 2 .
While most climate models deal with such elements as clear sky humidity,
average humidity, or differences between regions of high and low humidity,
Lindzen and co-workers studied feedback involving changes in the relative areas
of high and low humidity and cloudiness. Their results suggest that cloudy moist
regions contract when the surface warms and expand when the surface cools. In
each case the change acts to oppose surface change and, thus, presents a strong
negative feedback to climate change like a sort of Le Chatelier's principle. They
concluded that the relevant feedbacks are negative rather than positive and very
large in magnitude. Spencer et al. (2007) studied the effect of changes in clouds on
changes in temperature in tropical regions and found a negative feedback of
6W/m 2 per degree of temperature rise. This provides some support for Lindzen's
hypothesis.
Dessler et al. (2008b) attempted to derive water feedback sensitivity by
comparing data on global temperature and humidity during the winter months
of 2006-2007 and 2007-2008. However, Douglass and Christy (2009) demon-
strated a strong correlation between global temperature and an El Nin˜ o index
since 1978, and particularly for 2006-2008 (see Figure 11.5 ) . Dessler et al. (2008b)