|
Climate Dynamics
Climate dynamics is the field that studies the processes that
control climate and how it evolves. Since climate is
determined by the interaction of the atmosphere, the oceans,
the land surfaces, the cryosphere and the biosphere, virtually
all of the disciplinary research carried out by PAOC faculty
contributes to our knowledge of climate dynamics. However, at
the core of the field of climate dynamics are studies that
deal with the coupling between the different components of the
climate system. These studies necessarily involve the creation
of coupled models of the climate system, or of portions of it,
and using these models to elucidate the feedbacks between the
different components, to prioritize the factors that are
important in determining climate sensitivity, to explain past
climate changes, to make projections of future climate change,
and to assess to what extent climate change is predictable.
The three faculty members most involved with this kind of work
are Professors Prinn, Stone, and Marshall. Much of this work
is being carried out as part of two programs, the
Joint Program on the Science
and Policy of Global Change, and the Climate
Modeling Initiative. These programs are engaged in
constructing and applying highly sophisticated models of the
climate system. Details of this work can be found at the two
web sites listed above.
However the most sophisticated models require large
amounts of computer time and are almost as difficult to
understand and analyze as the real climate system.
Consequently simplified models also play a very important role
in studies of climate dynamics. Indeed, most of our
understanding of how the climate system works has come from
studies with simplified models. Such models have an important
role in the work on climate dynamics in PAOC. An example of a
study using such a model is one currently underway as part of
the Joint Program on the Science and Policy of Global Change.
This study is using a two-dimensional model of the climate
system, one which only represents the latitudinal and vertical
structure of the system, to simulate the climate changes that
one would expect to have occurred over the last 100 years as a
result of anthropogenic factors. The model simulations depend
on a number of things which are uncertain. One is how
sensitive climate is to changes in the radiative forcing. In
more sophisticated three-dimensional models there is a wide
range of sensitivities, because of uncertainty in how clouds
will change when climate changes. This sensitivity can be
measured by the increase in global mean surface temperature
when the atmospheric concentrations of CO2 is doubled, and the
climate system is allowed to come into equilibrium with the
doubled amount of CO2. Conventional estimates of this
sensitivity range from 1.5 to 4.5 C. Another uncertainty is
how rapidly heat penetrates into the deep ocean. This can be
measured by an effective diffusivity, and model calculations
give estimates of this diffusivity ranging from zero to 40 cm2/s. The attached figure shows how the MIT model's
simulation of changes in upper-air temperatures over the
period 1961-1995 depends on these two uncertainties, and how
the simulations compare with observations. Because there are
other uncertainties that also affect these simulations, it is
necessary to carry out many simulations to assess the impact
of all the uncertainties, and this would not be possible with
the more sophisticated models because of their computational
requirements. In this study the comparison with observations
is being used to place joint constraints on all the
uncertainties, and these constraints are then being used to
place objective constraints on projections of possible climate
changes over the next century.
Fig. 1. Latitude-height pattern of
temperature change for 1986-1995 minus 1961-1980 periods from
radiosonde observations (upper left) and model simulations
with
(Kv [cm2/s], S[K]) = {0.16, 4.5}, {2.5,
3.0}, and {40.0, 1.6}. The model is forced by changes in
greenhouse gas, sulfate aerosol, and ozone concentrations.
The model data are shown on the model grid without the
observational mask. Negative temperature changes are shown
in blue.
|
