TY - JOUR
T1 - The distribution of precipitation and the spread in tropical upper tropospheric temperature trends in CMIP5/AMIP simulations
AU - Fueglistaler, Stephan Andreas
AU - Radley, C.
AU - Held, Isaac M.
N1 - Publisher Copyright:
© 2015. American Geophysical Union. All Rights Reserved.
PY - 2015/7/28
Y1 - 2015/7/28
N2 - Reconciling observations and simulations of tropical upper tropospheric temperature trends remains an important problem in climate science. Examining atmospheric models running over observed sea surface temperatures (SSTs), Flannaghan et al. (2014) show that this reconciliation is affected by the SST data set used and that a precipitation-weighted SST (PSST) is valuable in explaining this result. Here we show that even for Coupled Model Intercomparison Project Phase 5 Atmospheric Model Intercomparison Project (CMIP5 AMIP) simulations forced with identical SSTs, tropical upper tropospheric temperature trends across models (and between ensemble members) show a substantial spread (standard deviation ∼10% of the average trend). About 60% of this spread between ensemble means, as well as deviations from the ensemble means, can be explained by PSST calculated from the time-evolving precipitation in each model run. Both PSST and atmospheric temperature trends show statistical evidence for systematic differences between models. We conclude that the response of precipitation patterns to changes in SST patterns is a significant source of uncertainty for tropical temperature trends.
AB - Reconciling observations and simulations of tropical upper tropospheric temperature trends remains an important problem in climate science. Examining atmospheric models running over observed sea surface temperatures (SSTs), Flannaghan et al. (2014) show that this reconciliation is affected by the SST data set used and that a precipitation-weighted SST (PSST) is valuable in explaining this result. Here we show that even for Coupled Model Intercomparison Project Phase 5 Atmospheric Model Intercomparison Project (CMIP5 AMIP) simulations forced with identical SSTs, tropical upper tropospheric temperature trends across models (and between ensemble members) show a substantial spread (standard deviation ∼10% of the average trend). About 60% of this spread between ensemble means, as well as deviations from the ensemble means, can be explained by PSST calculated from the time-evolving precipitation in each model run. Both PSST and atmospheric temperature trends show statistical evidence for systematic differences between models. We conclude that the response of precipitation patterns to changes in SST patterns is a significant source of uncertainty for tropical temperature trends.
KW - convection
KW - sea surface temperatures
KW - tropical temperature trends
UR - http://www.scopus.com/inward/record.url?scp=84938982765&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84938982765&partnerID=8YFLogxK
U2 - 10.1002/2015GL064966
DO - 10.1002/2015GL064966
M3 - Article
AN - SCOPUS:84938982765
SN - 0094-8276
VL - 42
SP - 6000
EP - 6007
JO - Geophysical Research Letters
JF - Geophysical Research Letters
IS - 14
ER -