HR: 16:15h
AN: A44A-04 [Abstracts]
TI: Measuring the potential utility of seasonal climate predictions
AU: * Tippett, M K
EM: tippett@iri.columbia.edu
AF: International Research Institute for Climate Prediction, The Earth Institute at Columbia University,
Lamont Campus / 61 Route 9W, Palisades, NY 10964 United States
AU: Kleeman, R
EM: kleeman@cims.nyu
AF: Courant Institute of Mathematical Sciences, New York University, 251 Mercer Street, New York, NY 10012 United States
AU: Tang, Y
EM: ytang@cims.nyu.edu
AF: Courant Institute of Mathematical Sciences, New York University, 251 Mercer Street, New York, NY 10012 United States
AB:
Variation of sea surface temperature (SST) on seasonal-to-interannual time-scales leads to changes in the distribution of
seasonal climate anomalies. Relative entropy, an information theory measure of utility, is used to quantify the impact of SST variation on seasonal precipitation compared to natural variability. Ensemble simulations from two general circulation
model (GCMs) are used to estimate relative entropy in three regions where tropical SST has a large impact on precipitation:
South Florida, the Nordeste of Brazil and Kenya. The impact is statistically significant about half of the years. Yearly
variation of relative entropy is strongly correlated with shifts in ensemble mean precipitation and weakly correlated with
ensemble variance. Further analysis using relative entropy as a metric indicates only modest useful and detectable
year-to-year variation of higher order distribution moments. Relative entropy is also found to be related to measures of the ability of the GCMs to reproduce observations.
DE: 3354 Precipitation (1854)
SC: Atmospheric Sciences [A]
MN: 2005 Joint Assembly