HR: 10:35h
AN: PP51D-02    [PDF]
TI: Stochastic simulation of large explosive volcanic eruptions and their impacts on climate
AU: * Naveau, P
EM: naveau@colorado.edu
AF: Department of Applied Mathematics, University of Colorado ECOT 231, Boulder, CO 80309-0526 United States
AU: Ammann, C M
EM: ammann@ucar.edu
AF: Climate and Global Dynamics Division, National Center for Atmospheric Research 1850 Table Mesa Drive, Boulder, CO 80307-3000 United States
AU: Oh, H
EM: heeseok@stat.ualberta.ca
AF: Department of Mathematical and Statistical Sciences, University of Alberta, Edmonton, AL T6G 2G1 Canada
AB: Externally forced natural climate variability played an important role over a large part of the 20th century. Unfortunately, projections into the future using a host of different scenarios do not include the natural components but focus on anthropogenic elements only. Although anthropogenic forcing is expected to be dominant, the lack of natural variations is a clear deficiency in regard to uncertainty estimates of future climates. In this talk, we focus on the stochastic modeling of volcanic forcing. A probabilistic model that reproduces the occurrences and the climatic impacts of large explosive volcanic eruptions in the tropics will be presented. To reach this goal, we first introduce an automatic procedure that estimates the climatic impact of strong but short-lived perturbations from large explosive eruptions on a variety of multi-proxy temperature reconstructions and output from a coupled Ocean-Atmosphere General Circulation Model. The extraction method based on a statistical multi-state space model 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 Generalized Extreme Value distribution. Because we can estimate all the parameters of our probabilistic model, we will be able to stochastically simulate the occurrences and the magnitudes of future large explosive volcanic events. Such simulations can be used for adding forcing into future scenario runs in climate models. Currently, this external forcing is completely missing in these scenarios.
DE: 1694 Instruments and techniques
DE: 3309 Climatology (1620)
DE: 3337 Numerical modeling and data assimilation
DE: 3344 Paleoclimatology
DE: 7536 Solar activity cycle (2162)
SC: Paleoceanography and Paleoclimatology [PP]
MN: 2003 Fall Meeting