Tweeters,

Some people dispute whether wintering Snowy Owl numbers
show population cycles with a peak every 3 or 4 years. The paper
referred to by Mike Patterson (Kerlinger et al., 1985) claims to
have disproved the idea of a 3-4 year cycle. I haven't seen
the entire paper, and would like to do so before debating their
claim in detail. However, the fact is that in some areas at least, there
is strong evidence for a 3-4 year cycle. Kerlinger et al. admit that
Snowy Owl numbers are "irruptive in eastern and western North America".
While the difference between "irruptive" and "cyclic" may be semantic,
"cyclic" implies that population fluctuation are somewhat regular and,
while "irruptive" is irregular and unpredictable.

In the Vancouver, BC area, since the mid-1960s, peak Snowy Owl
numbers have occurred in the following winters: 1966-67,
1970-71, 1973-74, 1977-78, 1980-81, 1984-85, 1992-93, 1996-97,
and 2004-05. In some cases, there was an "echo" flight as large
as, or nearly as large as, this peak flight the following year (e.g.
1971-72, 1993-94, 2005-06). The expected invasions of 1988 and 2000
did not materialize; there was a small flight in 2000, but no birds remained
over the winter near Vancouver, although there were a few in the Puget
Sound area.

This looks suspiciously like a 3-4 year cycle to me.

There is a tendency for every 3rd invasion or so to be much larger than the
others. The biggest flights in the last 40 years were those of
1973, 1984, and 1996. (No, 2005 does not compare with the
numbers seen in those years.) Mike Patterson is correct when he says
that he sees evidence of a 9- to 12-year cycle in numbers (not a lack
of cycles, as Kerlinger et al. claim). However, part of the reason he
fails to recognize the less dramatic 3-4 year cycle is that in these
flight years, few or no birds get as far south as Oregon, and in
some of them, none get south of the Canada/US border on the west coast.
It may be that 1988 and 2000 were also flight years, but that in
those years, few owls got even as far south as Vancouver.

>From 1968 through 2004, here are the combined counts of Snowy Owls from
2 Christmas Counts in SW BC-- Ladner and Vancouver. (About 75% of
these were seen on the Ladner count.)


1968 2
1969 3
1970 12
1971 13
1972 6
1973 134
1974 22
1975 19
1976 3
1977 26
1978 21
1979 1
1980 8
1981 1
1982 3
1983 0
1984 43
1985 5
1986 5
1987 2
1988 1
1989 1
1990 1
1991 1
1992 15
1993 13
1994 0
1995 0
1996 62
1997 1
1998 0
1999 0
2000 0
2001 0
2002 0
2003 0
2004 8

This looks like pretty convincing evidence to me of a population
periodicity with a peak every 3-4 years, except that the 1988 and 2000
peaks did not appear. It also shows good evidence of a mega-peak every
10-12 years. Perhaps these can be tested in a statistical way, but if so,
I'm not familiar with the methodology.

As I said, I haven't seen the entire Kerlinger et al. paper. However,
the portion posted by Mike seems to contain some serious flaws in analysis
(lumping areas that are much too large in their comparisons) and in logic.
For one example, they set up a straw man only to knock it down: no one
ever claimed that Snowy Owl cycles and invasions occur in the same
winters all across North America. They are regional phenomena, and
the ones I am familiar with are those on the Pacific Coast. In
any case, the Kerlinger et al. paper is 20 years old and badly in need
of re-analysis.


If one is prepared to accept that there is a 3-4 year periodicity to
Snowy Owl wintering numbers at least in SW BC and the Puget
Sound area (I could have included numbers from Skagit County
as well), the causes of these fluctuations in numbers are another
matter. Lemmings (2 species) in the Arctic tundra are known to have 3-4
year population cycles, and Snowy Owl breeding success is closely
related to breeding densities of lemmings on their nesting grounds
(although they do eat some other prey species). All researchers who
have studied Snowy Owls seem to agree on this. Some people
have claimed that Snowy Owl invasion years correspond to lemming
population crashes; others have claimed that they correspond instead
to lemming population peaks. To me at least, it seems highly probable
that the timing of southward invasions is tied somehow to prey densities
on the breeding grounds, and no one has yet disproved this hypothesis.
Perhaps weather is involved as well, and may be ultimately related
to lemming numbers. Does the El Nino cycle (yes, a 3-4 year cycle, on
average) affect weather in a predictable way as far north as the
Arctic tundra? Perhaps. However, I'm not aware that anyone has
shown convincingly that it does. Maybe someone will take up the
challenge soon.

Wayne C. Weber
Delta, BC
contopus at telus.net