Title: The ENSO Signal in Stratospheric Temperatures from Radiosonde Observations
1The ENSO Signal in Stratospheric Temperatures
from Radiosonde Observations
- Melissa Free
- NOAA Air Resources Lab
- Silver Spring
2Recent work confirms ENSO effects on winter
surface climate in Europe
slp
Observed surface climate response to ENSO From
Ineson Scaife 2008
Surface T
3This effect may operate through changes in the
Arctic stratosphere
Weaker polar vortex Warmer arctic stratosphere
El Nino
North Atlantic Oscillation -less zonal flow
Colder European winter
4Weak polar vortex is associated with cooling at
surface
Warm Cold
Surface T From Thompson et al. 2002
5Previous work 1. Tropics Several works note
cooling in stratosphere during El Ninos, but
others found little or no effect in zonal mean
satellite temperature. 2. Arctic Models find
warming T and weakening vortex for El Nino
winters Sassi et al. 2004 Taguchi and Hartmann
2005 Manzini et al. 2006 (but no signal for La
Nina) Ineson and Scaife 2008 Data No
significant relation Hamilton 1993 and other
papers Significant only during solar
minima Kryjov and Park 2007 Significant
warming at 10 mb Garfinckel and Hartmann 2007
Significant warming from 10 to 50 mb
Camp and Tung 2007 Depends on phase of
Quasi-biennial Oscillation (QBO)?
6Data Radiosonde datasets used (1958-2004) Radios
onde Atmospheric Temperature Products for
Assessing Climate RATPAC (NOAA) 85
stations Hadley Center Atmospheric
Temperatures HadAT2 (UKMO) 676 Iterative
Universal Kriging IUK (Yale)
527 Integrated Global Radiosonde Archive IGRA
(NCDC) gt1000 Indices ENSO index Nino3.4
SSTs QBO index 50 mb Singapore winds
7RATPAC
HadAT2
IUK
IGRA
8Methods Problem distinguishing QBO and
volcanic effects from ENSO effects (limited
number of cases) 1. Linear regression using ENSO
and QBO indices Remove QBO first (stepwise
multiple r.) Temperatures lagged 4 months behind
SSTs 2. Compositing of El Nino versus La Nina
years 3. Correlation coefficients (Skip 3
post-volcano years)
9ENSO temperature signal from regression using all
months
Pressure (mb)
K
latitude
10ENSO signal from regression using DJF mean
temperatures
Pressure (mb)
T (K)
11Arctic signal is strongest in February
Pressure (mb)
IUK
12DJF temperature
Compositing warm and cold ENSO events 1.
Compositing for all QBO phases
K
Pressure (mb)
13Compositing warm and cold ENSO events-- 2.
Compositing for West QBO only
Pressure (mb)
14Winter mean T at 100 mb for 70-90N with SST3.4
ENSO index
K
Correlations and explained variance70-90N
15- Conclusions
- 1. Significant ENSO temperature signal in the
tropical stratosphere around 50 mb. - 2. Significant ENSO signal, up to 5 K, in the
winter Arctic stratosphere extending into the
troposphere, primarily in February. (Consistent
with previous work but located lower in
atmosphere) - 3. Arctic winter signal is strong for West QBO
and absent for East QBO (as in reanalysis
studies). - 4. Level of maximum Arctic warming depends on
dataset and method of analysis. - Variance in Arctic winter temperature in
stratosphere explained by ENSO is 20 for most
datasets.
16 Remaining issues Mechanism Why QBO
effect? Is the effect limited to sudden
stratospheric warmings, or is there a general
seasonal warming? Why no cold European winter in
1997-98?
17Regression signal, DJF