The robust observation record extends back just over 100

Global warming and climate change: the paleo perspective
To predict future climate change scenarios we need to look at the past in order to
understand the boundary conditions under which various climate conditions occurred.
The instrumental record extends back just over 100 years, but only the last sixty years are
reliable. This short record is an inadequate benchmark from which to evaluate the extent
of how the 20th century climate is changing in response to anthropogenic vs natural
forcing alone.
We know from ice cores, sediments, tree rings, speleothems, corals and other archives
that climate has continually changed in the past at various rates and time scales. For
example, that there have been repeated abrupt and regionally extensive climate changes
that are not evident within the instrumental period. The data base of proxy information
contains evidence of repeated large and regionally extensive changes in atmospheric and
oceanic temperatures throughout the Holocene which is the period of Earth’s history that
we live in today. We know from large corals located off La Parguera that during parts of
the Little Ice Age (1450-1880) temperatures in Puerto Rico and the rest of the Caribbean
were three degrees Celsius lower than today. Interestingly, the Holocene is one of the
more stable periods in Earth’s history. During the Ice Ages climate varied much more
than it does today.
Because climatic variability has not recurred during the instrumental record we have
become complacent about the prospects of large and unanticipated climate shifts in the
future as atmospheric greenhouse gas concentrations continue to rise. The influence of
human activities on climate can be thought of as a vast experiment in which we are
changing one of the major controls of climate—the composition of the atmosphere—and
observing the response: ongoing global warming and its impact on the environment.
Recent climate events such as Hurricane Katrina and the rapid changes in sea ice taking
place at high latitudes today underscore the need to improve our understanding of how
the global environment will respond to climate change.
Paleoclimate research that leads to improved estimates of the climate and environmental
behavior is therefore timely and of great societal importance. While it is clear that
Earth’s climate is much more variable than previously thought, the factors that govern the
spatial and temporal patterns of variability are not well understood. Solar variations may
have been a critical factor, but at present there exists only circumstantial evidence for
solar variability influences on Earth’s climate at time scales ranging from decades to
centuries. The meridional overturning circulation of the ocean may also have varied in
the past, but the available evidence indicates that during the Holocene variations it was
relatively modest in comparison to those of the earlier glacial period. Stratospheric
aerosols released from large volcanic eruptions have clearly influenced Earth’s mean
temperature over the past several centuries. But on longer time scales the record of large
eruptions is not well documented, and it has not been possible to quantitatively assess
how important explosive volcanism has been in affecting larger changes in Earth climate.
It is also possible that there has been substantial stochastic variability in the climate
system that has no clear external forcing. Some regions also appear to have been
climatically more variable during the Holocene compared to other regions. Paleoarchives
have revealed, for example, that during the past ten thousand years the high northern
latitudes experienced larger oceanic and atmospheric temperature change than did the
low latitudes. Yet, despite smaller changes in temperature, the low latitude regions (like
Puerto Rico) have experienced substantially greater hydrologic variability.
Modeling efforts are particularly useful in this effort because they provide a globally
integrated picture of the climate system. But even the most sophisticated models have
large uncertainties when projections extend to decades and longer. The best opportunity
we have for advancing our understanding of climate behavior on these time scales is a
research program that utilizes both models and paleoclimate data.