AO, BO, CO, ...? How to recognize a real teleconnection pattern from a fake

0493297 - ÚFA 2019 DE eng A - Abstrakt
Huth, Radan - Beranová, Romana
AO, BO, CO, ...? How to recognize a real teleconnection pattern from a fake.
EMS Annual Meeting Abstracts, Vol. 15. Berlín: European Meteorological Society, 2018.
[EMS Annual Meeting: European Conference for Applied Meteorology and Climatology 2018. 03.09.2018-07.09.2018, Budapest]
Institucionální podpora: RVO:68378289
Klíčová slova: teleconnection pattern * arctic oscillation (AO) * barents oscillation (BO) * North Atlantic Oscillation (NAO) * principal component analysis (PCA)
Obor OECD: Meteorology and atmospheric sciences
https://meetingorganizer.copernicus.org/EMS2018/EMS2018-233.pdf

AO (Arctic Oscillation) and BO (Barents Oscillation) have been identified as modes of variability (teleconnections)
in sea level pressure (SLP) by unrotated principal component analysis (PCA). AO is typically defined as the
first mode of SLP northward of 20 N, while BO is the second or third mode of such an analysis. AO appears as
a tripole pattern with one centre residing over the Arctic and two other centres of the opposite sign located over
the North Atlantic in the area of the Azores high and over the North Pacific in the area of the Aleutian low.
BO consists of two major centres with opposite signs, one of which is located over northern Eurasia, the other extending from Greenland
across the Canadian Archipelago towards Beaufort Sea. BO has been related e.g. to sea ice transport in the Arctic.
Several studies have shown that unrotated PCA is prone to produce modes (patterns) that do not correspond
to real correlation structures in the data, that is, that are statistical artifacts. Instead, rotated PCA is generally
recommended as a suitable tool for detection of teleconnections.
For a teleconnection pattern to be considered realistic, one must demonstrate that it corresponds to real correlation
structures.
The objective of this contribution is to demonstrate that both AO and BO, if conventionally defined (that is, being
produced by unrotated PCA), are not real atmospheric features (teleconnections) and to contrast their behavior
with the North Atlantic Oscillation (NAO), which is defined from rotated PCA.
Our results clearly indicate that, unlike NAO, which is a real mode of atmospheric variability, AO and BO are not
real teleconnections and should be not interpreted as such.
And what about CO? It has not been identified by anyone yet, but we are afraid that an improper use of PCA
could result in an alleged identification of another spurious, really non-existing mode (teleconnection) in future.

Trvalý link: http://hdl.handle.net/11104/0287048