These geological records show that the present is an unusual moment in the
recent history of our planet.  Super-imposed on the global cooling that started some
50 million years (Ma) ago when  polar temperatures were close to10ºC, are 
periodic oscillations that were modest in amplitude up to ~3 million years but
then started to amplify, culminating in dramatic fluctuations between
prolonged ice-ages that persisted for some 100,000 years, and brief,
temperate  interglacials. (Variations in O18 in seafloor cores, the top panel,
are a measure of polar temperatures.) Our species took advantage of the
current interglacial -- it started 10,000 years ago --  to advance rapidly, from
the invention of farming to industrial activities that are causing a rapid rise in
the atmospheric concentration of greenhouse gases. This rise, the vertical red bar
in the lower panel which shows measurements from Antarctic glaciers, is occurring
at a time when those concentrations are at a natural maximum. That it is
also occurring in an era, the past 1 Ma, of great climate sensitivity, follows from
the cause of the recurrent ice-ages: very modest, periodic variations in the
distribution of sunlight because of periodic variations in orbital parameters such as
the tilt of the Earth’s axis. To anticipate what will happen next it will be helpful
to understand what had happened in the past.

Two sets of processes determine the climate variations described above.
One, associated with the drifting of continents (which affects the frequency
of volcanic eruptions and hence the composition of the atmosphere), is
mainly responsible for the long-term global cooling. The other associated with
the Milankovitch cycles, the periodic variations in sunlight. The global
cooling affected the response to relatively constant Milankovitch forcing
by introducing feedbacks at certain times. For example, the global cooling led to
the appearance of northern glaciers around 3 Ma, bringing into play the
ice-albedo feedback. (White glaciers reflect sunlight, thus promoting the growth
of glaciers by depriving the Earth of heat.) This mechanism has been
studied extensively in connection with the ice-ages, but many puzzles remain. One
is the amplification of obliquity but not precession cycles in global ice
volume. Another concerns fluctuations in equatorial sea surface temperatures,
very similar to those in the bottom panel, which lead fluctuations in ice volume
by several thousand years.  This, and other evidence, clearly indicate that the
ice-ages involve more than ice, that the tropics played a role.

Equatorial sea surface temperature patterns depend on the winds and in
turn influence the winds. This implies that interactions between the ocean
and atmosphere amount to  positive feedbacks. Their impact on the global climate
is evident during El Niño episodes when the atmospheric concentration of
the powerful greenhouse gas water vapor increases significantly.  Sea
surface temperature patterns depend critically on the subsurface thermal structure
of the ocean, especially the depth of the thermocline, the interface between
the shallow layer of warm surface waters and the much colder water at depth.
El Niño corresponds to changes in the slope of the equatorial thermocline as in
the sketch on the left. Changes in the spatially averaged depth of the thermocline,
as in the sketch on the right, also alter sea surface temperatures, by means of
entirely different (diabatic) processes  that involve changes in the heat budget of the ocean.  Whereas the oceanic heat gain, in low latitudes where cold water rises to
the surface, depends on oceanic factors, specifically the depth of the
thermocline, the oceanic heat loss in high latitudes depends on atmospheric
factors, the air temperature for example. In a state of equilibrium, heat gain
balances heat loss so that a warm world, with a small loss of heat from the
oceans, must have a small gain and hence a deep tropical thermocline.

Atmospheric conditions in high latitudes can therefore determine the depth of
the tropical thermocline, and the intensity of air-sea interactions. These
connections between low and high latitudes depend on the oceanic
circulation which has two main components: the deep, slow thermohaline circulation, and the shallow, rapid wind-driven circulation. Both effect a
poleward transport of heat by means of meridional overturning; surface waters 
sink in the extra-tropics  and rise back to the surface in lower latitudes. The
sinking depends on the buoyancy of the water and hence on its temperature
and salinity. The heat transport, and consequently the depth of the
tropical thermocline, can therefore be affected by several high latitude factors t
hat include temperature, rainfall, and the melting of snow.

The global cooling that started 50 Ma ago was accompanied by a gradual
elevation of the oceanic thermocline which, around 3 Ma, was so shallow
that tropical air-sea interactions started influencing the global climate. Such
insights, from studies of oceanic connections between low and high latitudes,
are shedding light on phenomena that range from the decadal modulation of  
El Niño, to the perennial (rather than intermittent) El Niño conditions of the
early Pliocene, to the recurrent ice-ages of the past million years.  The theories
help explain the geological record of past  climates, which in turn provide 
valuable checks for computer models of future global warming.