|
I
reckon Geoff would be right, which is why I raised the aspect of that there may
be flexibility in the way the GBS sites are interpreted. This was certainly the
case in the early years of the GBS (Version 1 & 2 of the chart) when the
instructions included to include species seen outside your area (in a way
that was never explained). I did what I could to fix this, with V 3 of the
chart. This is all explained in the GBS Report. Also the questionnaire survey I
did in 1989 explored the aspect of how people interpreted the site, relative to
different species.
More
to the point and to be fair, with regard to these birds it simply is very hard
to tell at any one instant how many birds are within the area and so it is not
unexpected that people will try to count all in view and assume that potentially
all were within the cylinder of the 3.1 hectare site. That is what I do if there
are any of these swifts above, then I try to count all in view and perceive them
as the one flock and that most would potentially be within the 100 metre
radius.
A
question arises in Geoff's graph whether the number noted would be one or ten. I
say it should be zero, as none (let alone not most) of the birds are within the
area. And thus Geoff's illustration on a logic basis contains no limits. I would
say the group needs to be a flock of some cohesion and at least most observed
within the site, so that it is not unreasonable to think that they all were in
the site at the one time, to have any justification to record as one count on
the GBS.
See
this paragraph extracted from page 21 of the GBS Report. (Sadly it is not in the
index under "Needletail" because I used the general term "swift")!
Accurately counting and recording bird numbers is always a problem
for almost any bird survey. Survey methods were discussed in a series of papers
edited by Davies (1983) and Ralph & Scott (1981). The GBS is based on some
unrealistic assumptions: That each observer is continually aware of the number
of individuals of each species present in their 3+ hectare area. That they are
able to pick for each species when that number reaches its maximum and that they
will record it. However, for the common species, the best that most can do is
provide an estimate of the maximum number observed at any one time. It is
certainly difficult to count the number of swifts in a large swirling flock,
within a 100 metres' radius cylindrical shape of unlimited altitude. People have
been inconsistent in applying the distance limit. It is very likely that
uncommon birds, especially if they are large and conspicuous, such as raptors,
would have been recorded if observed from the site, even if actually outside the
100 metres. The same observers would probably not have done this for species
that occur within the site. Often it is hard to know for birds seen flying over
or heard at night (such as Masked Lapwings and Boobooks) whether they are within
100 metres or how many individuals were involved. This phenomenon has the
benefit of providing information as to their presence that would have been
missed if observers kept strictly to the rules. A bigger problem is when
observers fail to record numbers of the common species. In interpreting these
data, be aware that there is some variation in scaling of results.
Also I am certain that Martin is
right that "once a flock is greater than 5,
there is a distinct digital preference for numbers ending in zero." Back in 1989 I did a survey of preferred numbers, logging the number
of times a number was used from about ten randomly chosen charts and I observed
the result Martin tells. I did this manually as there was no way of doing it
from the way the data existed at that time, I must have been crazy but I
did it. It would be easy to do now, even from the entire database. And
maybe worth doing. Figure 34 of the GBS Report is relevant to this except it
expresses averages as logarithms. That is on the same question but I think more
useful because it tells a lot about social behaviour of the birds, rather that
just being an investigation of people's biases for easy numbers. But then I
think that is perfectly reasonable and typical of any kind of human
numerical survey. People may doubt an observation of 347 birds on a GBS
chart.
Philip
Call
me a cynic, Martin, but re our previous discussion I think the ‘elastic site’
effect might also be at work:
From: martin
butterfield [
Sent: Friday, 19 March 2010
9:32 PM
To: COG List
Subject: [canberrabirds] Re: Swifts in
Carwoola
Following my report earlier this
evening an observer commented that they had seen a single swift over their
yard. After a good description of swift jizz they concluded "... do they
fly solo or would this one have been a stray from a nearby
out-of-sight flock passing by? I thought they flocked!"
The attached
document shows the distribution of "flocks" of swifts reported in 28 years of
GBS. Clearly reports of 1 or 2 swifts are not uncommon. There are
enough references in HANZAB to 'single birds' to suggest that some birds may be
'loners' but possibly this simply reflects that relatively dispersed food
resources lead the more typical large flock to be dispersed when passing over a
small area such as a GBS site..
At a more detailed level - and possibly
of interest in view of recent posts about counting flocks - once a flock is
greater than 5, there is a distinct digital preference for numbers ending in
zero. Given the speed at which these birds travel it is not surprising
that observers round them to the nearest 10.
Martin
On Fri, Mar 19, 2010 at 6:40 PM, martin butterfield <>
wrote:
Some smoke is being blown up from the SE - presumably from
one of the many hazard reduction burns listed on the RFS web site. Some
woodswallows and the occasional Swift were cruising through this, as reported by
a resident on the Widgiewa Rd ridge. They were unobliging in visiting my
GBS site (about 1km NE).
|
