Title 
The Role of Nature in Future Climate Change: Spatio-temporal statistical
analysis of volcanic eruptions and
solar irradiance changes and their impacts on past and future climate.

Authors: 
C.M. Ammann, P. Naveau, H.-S. Oh

Abstract:
Climate reconstructions of the recent past have improved tremendously over the
last 10 years. Estimates of surface temperature and hydrologic
variables (e.g. precipitation, soil moisture) are now available for many parts
of the globe. This richness of data can be used to evaluate causes of
climate variations on the global or hemispheric scale. 
Using indirect proxy-climate measures from the
real world as well as output from different climate models, one can show that
solar and volcanic forcings have left a clear fingerprint in past
climate variations. Due to improved spatial coverage of the proxy-based climate reconstructions,
we can now also search for spatial characteristics of past climate
changes and compare them to physical processes operating in the climate system
that translate the different forcings into a surface response. 
Selected
scenarios of future changes in atmospheric greenhouse gas emissions and
concentrations in the atmosphere are often used to predict what climate might
look like in the near future (e.g. Intergovernmental Panel on Climate
Change). These predictions use exclusively the anthropogenic changes but
don't include any of the natural forcings. 

We present a statistical approach on how the past record of both solar
irradiance changes and volcanic eruptions can be used to augment the future
climate change scenarios. Appropriate statistical methodologies have to be
applied to capture the intrinsically very different temporal
characteristics of the more smoothly changing solar and the intermittent
appearing volcanic forcing. This work highlights the advantages of
statistical modeling of natural processes both for the detection of past
influences on climate as well as for generating statistically based future
uncertainty boundaries for possible forcing range.

Using non- decimated descrete wavelet transform, the spatial features of solar
modulated climate variations are extracted from proxy records as well as from
two experiments with a coupled General Circulation Model which were forced
with solar irradiance variability and explosive volcanism. We discuss
important aspects of coherency and stationarity for the known solar variations
at 11/22-year, 80-88-year and ~200-year cycles and quasi-cycles. 

To extract the climate response to an explosive volcanic eruption, we use
 extraction method based on a statistical multi-state
space model. This approach  provides an accurate estimator of the timing and duration
of the climate response to an eruption.  This will not only allow for a more
objective estimation of the associated peak amplitude (cooling) and the
subsequent
time evolution of the signal, but at the same
time it will provide a measure of confidence through the posterior
probability for each cooling event.
Secondly, the distribution of the derived magnitudes from these largest
volcanic coolings  will be  shown to follow a
a Generalized Extreme Value distribution.

Finally, we
present a statistical framework which can help identify mechanisms and
feedbacks involved in translating a solar modulated perturbation into the
Earth's climate system.