Professor William Hsieh
Dept. of Earth & Ocean Sciences, Univ. of British Columbia
The use of neural network models to nonlinearly project the El Nino-
Southern Oscillation (ENSO) index to the winter N. Hemisphere atmospheric
anomaly field (500 hPa geopotential height, sea level pressure, surface
air temperature or precipiation) has led to the extraction of not only the
classical linear teleconnection pattern but also a nonlinear
teleconnection pattern, which is quadratically associated with the ENSO
index. Similar calculations are performed using the Arctic Oscillation
(AO) index instead. In both the ENSO and AO cases, the nonlinear
teleconnection extends further from the source region than the linear
teleconnection. Implications for seasonal climate prediction will be
discussed.
A simple theory is presented on how nonlinear teleconnection modifies the
classical linear teleconnection under climate change. This theory is
tested with the CCCma coupled model runs, including the increasing carbon
dioxide scenario runs, with good agreement.