Personal tools
You are here: Home Science & Projects Deisa Extreme Computing Initiative Projects 2005 - 2006 Ensemble SimulationS of Extreme weather events under Nonlinear Climate changE

Ensemble SimulationS of Extreme weather events under Nonlinear Climate changE

Project Acronym ESSENCE
Scientific Discipline Meteorology; Climate Variability and Change
Principal Investigator(s) Prof. dr. ir. H. A. Dijkstra
Leading Institution IMAU, Utrecht University
Partner Institution(s) None

Project summary and results

The main aim of the ESSENCE project was to compute an adequate estimate of the statistics of internal climate variability and hence be able to obtain a good signal-to-noise ratio for the forced signal due to the increase of greenhouse gases. In the project, a 17-member ensemble simulation of climate change in response to the SRES-A1b scenario was carried out using the ECHAM5/MPI-OM climate model. The relatively large size of the ensemble indeed enabled us to better distinguish the forced signal from internal variability. We showed that in large parts of the world the observed warming over the last 60 years is statistically indistinguishable from the warming forced by increased greenhouse gas concentrations.

A great advantage of a large ensemble is the large noise reduction that can be achieved by averaging over all ensemble members. We were able to determine the year in which the forced signal (i.e., the trend) in the atmospheric (2 meter) temperature emerges from the noise. (Fig. 1). A student t-test, in which the trend over a particular period is compared with the standard deviation of the noise, was used.


Fig. 1: Year in which the trend (measured from 1980 onwards) of the annual-mean 2-meter temperature emerges from the weather noise at the 95%-significance level.

The earliest detection times are found off the equator in the western parts of the tropical oceans, where the signal emerges as early as around 2000 (and for some regions even earlier) from the noise. In these regions the internal variability is extremely low while the trend is only modest. A second region with an early detection is the Arctic, in which the trend is very large due to the decrease of the sea-ice. The longest detection times are found along the equatorial Pacific where, due to El Nino, the variability is very high, as well as in the Southern Ocean and the North Atlantic, where the trend is very low.

DEISA resources used

The ESSENCE simulations were successfully performed on the DEISA infrastructure, using the NEC SX-8 system of HLRS. The 50 TB dataset of both ensemble simulations, comprising roughly 140-year time series of 80 variables, is a unique source of information for many researchers, and its analysis will provide clear answers on the changes in statistics of extreme weather events. Exploration of these data is facilitated by using high-resolution tiled panel displays at SARA.


Document Actions