Like Dennis, I have been curious about possible correlations of Snowy Owl
migration irruptions with climate and prey fluctuations.
I first downloaded the Audubon CBC data for Washington and for British
Columbia for the last 40 years. Comparable numbers are reported each year
for these two areas, and ups and downs are strongly correlated between WA and
BC. I then looked at correlations with climate indices such as the ElNino
Southern Oscillation Index (ENSO), Pacific Decadal Oscillation, and an
Arctic Oscillation index, and found no correlation. (For example, some seabirds
and Whooping Crane populations exhibit correlations with various of these
indices).
As has been mentioned, lemmings exhibit a fairly regular 4-year cycle,
although the peaks in the cycle are not geographically correlated as one looks
longitudinally across the arctic. One can look for a regular 4-year cycle
for snowy owls by making a spectral density plot (plot of frequency of a
particular time interval versus time interval) as had been done by Kerlinger
et al (1985) using earlier data. This approach should show a peak at 4
years for either of Dennis?s two suggested ways, as it does not depend on
whether due to a minimum or a maximum in the lemming population. Neither the BC
+ WA or WA snowy owl count data exhibited a peak at 4 years. I tried this
for several thresholds of how many owls constituted an irruption. Thus
lemming populations don?t seem determinative in generating irruptions, either
directly or a lag effect.
I am therefore left with no explanation for the timing of owl irruptions.
Bob Vandenbosch