Lindzen's new GRL paper on climate sensitivity

Skip to first unread message


Sep 11, 2009, 3:05:17 PM9/11/09
to globalchange
Many skeptic blogs are trumpeting the recent GRL paper by Lindzen and
Choi that estimates climate sensitivity at about 0.5 C per CO2
doubling based on the claimed predominance of negative feedbacks to
the original forcing. The main mechanism is asserted to be an
increase in outgoing shortwave (SW) radiation. The estimate, based on
observational (ERBE) data, is stated to conflict with the high
sensitivity and positive feedbacks predicted from climate models.

A number of commenters have crtiticized the use of AMIP models for
comparison purposes in the paper, since these atmospheric models are
not designed to simulate real world outcomes involving a coupled
atmosphere/ocean system. There are also concerns about the paper's
exclusion of temperature data below a 0.1 C threshhold, although the
importance of this concern is hard to judge. However, it seems to me
that the conclusions in the paper are subject to a more fundamental
error than these. In essence, the paper attempts to extrapolate
climate sensitivity data from observed changes in SST (mainly ENSO-
related) to a sensitivity estimate that would characterize CO2
forcing. I would argue that this involves an assumption that is
unproven and probably wrong.

The assumption is that climate sensitivity, if broadly defined as an
equilibrium temperature change expected from a given W/m^2
perturbation of the climate system will be the same regardless of the
cause of the perturbation or its location. This may be roughly
correct for perturbations originating in the atmosphere (e.g., changes
in greenhouse gases vs changes in aerosols), but I would submit that
it is unwarranted when comparing atmospheric perturbations with those
originating in the ocean. What follows is an example of why these two
types of sensitivities might differ radically. The example may or may
not explain the Lindzen data, and is almost certainly overly
simplistic, but illustrates the principle that perturbations
originating in the atmosphere and ocean may not involve comparable

When heat is added to the atmosphere (e.g., by a rise in CO2
concentrations), the direct immediate consequence is atmospheric
warming, which by the Clausius/Clapeyron equation tends to reduce
relative humidity (RH). This is compensated by the effect of the
atmospheric warming on the ocean, causing more water to evaporate and
restore RH toward the previous value. The additional water exerts its
own greenhouse effect, thereby contributing a positive feedback.
However, there is some evidence that over decades of increasing CO2,
RH has tended to decline slightly, implying that the evaporation has
not quite kept pace with the warming. A decline in RH leads to a
reduction in low cloud cover and a consequent increase in SW
absorption by the oceans, which would contribute its own positive
feedback to the system. The recent Science paper by Clements et al
provides evidence that this in fact may have been happening; those
authors have suggested that their results imply a greater climate
sensitivity than the typical 3 C value typically used as an estimate.

Consider, however, what might happen, if an El Nino event caused SST
to rise in the presence of an unwarmed atmosphere (atmospheric
responses typically lag SST ENSO changes by months). In this case,
the warmer ocean will cause more water to evaporate, but unlike the
CO2 forcing scenario, RH will rise. Among the consequences might be
increased precipitation, reducing the extent of positive feedback from
the extra water, a reduced lapse rate imposing a negative feedback,
but most relevant to the GRL paper, an increase in cloud cover from
the increased RH. These extra clouds would be consistent with the
increased loss of SW radiation to space reported in the GRL paper. To
summarize this concept, warming originating in the atmosphere might be
expected to reduce RH, permit more evaporated water to remain aloft,
and reduce cloud cover - all phenomena that would mediate positive
feedbacks. Warming originating in the oceans in the presence of an
unwarmed atmosphere might reduce or reverse these proesses, and in
particular, by increasing RH, lead to increased cloud cover and a
strong negative feedback from increased cloud albedo.

ENSO-related changes in SST also involve changes in atmospheric and
ocean circulation patterns, and there are many more variables that
render the comparisons with CO2-mediated forcing questionable beyond
the source of the extra heat. Rather than claim to have explained all
of Lindzen's results by my simple example, I will limit myself to
suggesting that his explanation does not deserve to be accepted at
this point as anything more than a demonstration that heat added to
the climate system from the oceans may possibly generate negative
feedbacks, and that the magnitude and direction of feedbacks from CO2
have not yet been shown to be derivable from that conclusion.
Reply all
Reply to author
0 new messages