Lagged Correlations of the Antarctic Stratosphere with Tropical Pacific SST Anomalies

Abstract:

Coupling between the central tropical Pacific (CTP) sea surface temperature (SST) variability and the Antarctic stratosphere known from previous studies is analysed using a lagged correlation method. Monthly mean total ozone from the Merged Ozone Data Set and sea surface and stratospheric temperatures from the NCEP–NCAR reanalysis over the 1979–2011 period are used. Indices for their interannual variability were created and sequential lagged correlation ‘index – temperature field’ was used to both identify spatial patterns of the coupled regions and determine time lag of maximum correlation. It has been revealed that interannual SST variations in a westward-oriented V-shaped pattern in the CTP region are associated with statistically significant positive signals in zonally asymmetric region of the Antarctic stratosphere. The CTP index (CTPI) in this work characterizes the CTP SST variations in a domain 160–220°E, 30°N–20°S, which is close to the standard Niño-4 domain by longitude but is five times wider by latitude. The strongest correlation between the CTPI in June and total ozone/stratospheric temperature is found in October (r = 0.7, see figure). In figure the lagged cross correlation between the T50- and TOC-indices in October (solid and dashed curves, respectively) and the CTPI with a time lag from 0 to –12 months) are shown. Vertical line is at the month of correlation peak (CTPI in June, –4 months) and horizontal line marks the 95% confidence level at r= 0.37. Concurrent teleconnection in winter and spring between the identified regions is negligible and, thus, the results suggest about four month delay in signal propagation between the CTP and Antarctic stratosphere. The importance of the off-equatorial tropical SST variability in the V-pattern (i.e. poleward of the standard Niño-4 domain) for the delayed response of the Antarctic stratosphere is emphasized. Possible causes of the V-pattern formation and related response delay are discussed briefly.

Acknowledgements. This work contributes to the Polar FORCeS Project 4012 of the Australian Antarctic Science program.