"Graham Davis" wrote in message
...
Alastair McDonald wrote:



I am not joking, but I have been reading "Weather Cycles" by
Burroughs ISBN 0 521 52822 4, and he points out that as soon
as you identiy a weather cycle it stops working!

A quick search with www.google.com will soon find you plenty on
the QBO. Note that it is quasi, and so it varies and its average
length is 27 months. That would account for 1976 being hot, as
well as 2003. They are exactly 12 QBO cycles apart.

AFAIK there is no explanation for the QBO, so I cannot explain
why it should produce its effects. Moreover the solar cycle (sun
spots) and the ENSO will also have an effect.

Cheers, Alastair.



Soon after I started working in the Ice Section of the Met Office
(1965), I noticed that pattern of spring breakup of the ice in Hudson's
Bay followed a biennial pattern. Every even year, the ice broke from the
eastern shore and, in odd years, from the west. As far as I remember,
the oscillation had started by 1960 or earlier. I know it continued to
1972 but I left the section before seeing the 1973 breakup.

Against the odds, this continued after I started using it for
forecasting but perhaps Murphy's Law worked in a different way, because
the knowledge was next to useless for me to help our customers. What was
the point of knowing which side of Hudson's Bay would break first if you
couldn't tell when you could penetrate the ice along the route into the Bay.

The oscillation didn't even give me a clue as to whether the breakup
would be late or early, because an odd year could have persistent, warm
SW winds or cold winds from the NW. Similarly, even years could have
early or late breakups.

It is interesting that it was sea ice that first caught your eye as being
biennial. Many annual average temperature records also show periods
of annual reversals. But it was the plots below of ice extent and ice area
for the Arctic which finally sent me searching on the web for a biennial
oscillation, and I found the QBO. The plots to which I am referring are;
http://iup.physik.uni-bremen.de:8084/decade.html
There are two plots on that diagram. The light grey dots are ice
extent and the black dots are ice area. The extent, which is probably
more sensitive to the polar vortex shows the most obvious bieniality,
from 1992 to 2000. What interests me is the linearity of the alternative
years 1996, 1998, 2000, & 2002. It will be interesting to see if this
year 2004 follows that trend.

The fact that the bieniality breaks down is probably due to the fact
that the cycle is not exactly two years. With a period of 2.25 years,
it will be 180 degress out of phase with the seasonal cycle after
4 cycles viz 10 years.

Cheers, Alastair.