|
Thanks Geoffrey,
There is a whole set of literature, mainly in
Europe, about the decline of House Sparrows as shown in urban bird surveys. This
has been discussed before. Although the reasons may of course vary from those
here.
In preparing the information for the GBS Report I
did of course look at all the trends for all the species in the various
statistics: A, F, R, W & G (as explained therein). I decided to concentrate
on A for several reasons:
Abundance is the most useful measure, the one that
previous writers of the ABR had used most often and is what really sets COG's
GBS off, against all those many, less useful surveys that only use some measure
of recording rate or distribution rate in their stats. Also the values of A, F
& R are all related, in that, broadly increases or decreases in each, will
go together. This is mostly statistical consequence. It was complicated enough
to include the graphs of A for long term and monthly for 120 species. It would
have been too much to include graphs of either or both F & R as well. I
didn't want it to be too statistically complex or cost too much for me to
produce (as I was financing the GBS Report from the largess of the unemployment
benefit.) Actually the stats for F & R don't vary much for most species. I
have commented in the text for a small number of species when there was
something useful to say about these aspects. Examples that come to mind are the
Satin Bowerbird and Common Myna. As for F, it doesn't take much for a species to
be observed just once at a site in a year. Also some years we only had 43 sites,
so one random yes or no makes for a >2% difference. Also F tended to be very
influenced by just a few sites, eg stats for birds that were only recorded from
sites near water or on the fringes of Canberra were unreasonably variable,
depending on the list of sites for the year.
Another point to note is that you wrote "Given that that is over a full year". Well yes, potentially but
not in practice. Figure 7 (page 26) and surrounding text demonstrates that
although the modal number of weeks per chart is 52, the average number of
weeks per chart is only 40, with a broad spread. I would fully expect that the
same trend has continued after Year 21 (but the data are not assembled on
that).
Philip
|
