Alan/David,
Time sampling did not occur to me. But now I think I get it.
The only time I've participated in counts involving tens of thousands of
birds was during the 2020 count at Lake Abert and those flocks were pretty
much fixed in one location.
Thank you for taking the time to explain the process.
Jack Williamson
West Linn, Oregon
On Mon, Nov 28, 2016 at 9:35 PM, David Irons <llsdirons@xxxxxxx> wrote:
Jack,
I am sure that you and hopefully others reading your post realize that
Phil and those who attempt to quantify some of these large flights are not
making tens of thousands if individual identifications.
With large mono-directional flights the easiest way to come up with a
reasonable number is to make occasional sample counts ranging from 1-10
minutes in duration. Then you can use these counts to assess how many birds
per minute, per 10 minutes, or per hour are passing. I tend to take a
conservative approach when I do this by extrapolating using the lower
counts and to some degree ignoring the peak pulses.
With birds seen in smaller numbers or occasional groups, one can enter
those as they are seen (if you are eBirding using a smart phone) or jot
down tallies on a notepad.
In my experience, the average birder tends to under estimate the number
of birds in a large flock, sometimes grossly. I tend to fall back on the
order of magnitude rule. So long as my estimates are within an order of
magnitude they aren't likely to be major outliers. If I see a flock of 100
birds I'm certainly not going to estimate that only ten birds are in the
flock or estimate that 1000 are in the flock. Further, with a number that
size my estimate isn't likely to be off by more than double or less than
half the actual count.
So lets say I estimate a flock of 100 birds to have 200 birds in it and
report that via eBird. Then three other eBirders come along and see the
same flock and report counts of 80, 95 and 110 birds. These four counts add
up to 485, divided by the number of observer visits (four) averages out to
121.25 birds per visit. If we are talking about a species that is fairly
common, is this imperfect average statistically significant? In the minds
of some maybe, but not in my opinion.
I've been on CBCs where party members have fussed over whether there were
15 or 20 juncos in the flock we just pished up. Given how many uncounted
juncos there are in the typical western Oregon CBC circle, does a
discrepancy of five individuals matter? Or, if we over-count isn't it
equally likely that other parties are out there undercounting the groups of
juncos that they see?
Producing meaningful counts of relatively abundant birds doesn't have to
be an exact science and it generally isn't. It is a good idea to double
count large flocks. If I see a flock that I grossly estimate to be in the
thousands, I may start out counting by 100's and then double check myself
by counting by 50s. Over time you will find that your initial gross
estimates become increasingly accurate.
When we first got together seven years ago Shawneen thought some of my
counts were too high and I would tease her about her conservative counts.
After years of birding together and cross-checking one another our counts
and estimates now don't differ by much.
Dave Irons
Sent from my iPhone
On Nov 28, 2016, at 7:13 PM, Jack Williamson <jack.williamson.jr@xxxxxxxxx> wrote:
become even more amazing to me since Matthew Hunter's 11 November Post
Let me start this discussion by saying - I am not worthy
Phil's numbers are often staggering in their immense size, but they've
about Estimating Bird Numbers.
Sibley's practice quiz.
In the spirit of full disclosure, I admit to getting a B- (at best) on
more worthy :-)
Had I aced that quiz, I would be inclined to declare myself to myself as
does what Phil does.
Hopefully, I've set the stage for a productive discussion of how one
and 30 minute period.
Phil tallied (41) species and 35,644 (+) individual birds over a 2 hour
birds per second or [238] birds per minute for each of the 150 minutes of
That's a lot of birds!
When you break it down - those numbers translate to identifying [4]
the observation period.
Jack Williamson
West Linn, Oregon
