Hi Chris,
That was me on the phone making the sweeping statement about point counts.
Perhaps I could have chosen my words more carefully. But my sense is, based
on personal experience, talking to others, and backed up by Chris and
Laura's study, as well as a study out of Clemson a couple of years ago by
Jennifer Froehly, that at least under certain circumstances, point counts do
not result in high LOSH detection rates. That may be more true in areas
that have fewer LOSH than in areas where they are more abundant. I think
that Amy's bullets below hit all the major points, and don't really have
anything more to add - sorry to clog up people's inbox, and I'm not looking
to reignite the discussion, but I just wanted to take ownership of the
statement I made. Thanks.
Sergio
-----Original Message-----
From: eloshwg-bounce@xxxxxxxxxxxxx <eloshwg-bounce@xxxxxxxxxxxxx> On Behalf
Of Christopher Hill (Redacted sender "Chill" for DMARC)
Sent: Thursday, February 13, 2020 11:59 AM
To: eloshwg@xxxxxxxxxxxxx
Subject: [eloshwg] Re: point counts
Thanks, Amy,
Most of these seem like limitations to point counts in general, and this
summary helps me a lot, so thanks. I think maybe the disconnect I was
having was with the word “unreliable.” You seem to be saying that point
counts are an imperfect method for proving that shrikes are or are not at a
particular site at a particular time. Coming from a place where shrikes are
rare and endangered, that is a big reason for you to do surveys.
What I am more interested in is answering the question “how many shrikes are
in this landscape?” in an area where shrikes are more common. If surveys do
not detect every shrike I am OK with that, and would still consider a survey
reliable if it reflected the local abundances of shrikes.
I am now seeing Chris LItuma’s email and I will give him a call.
Chris
On Feb 13, 2020, at 11:48 AM, Amy Chabot <achabot@xxxxxxxxxxxxxx> wrote:
Hi all,
I'm hoping (hint, hint) that Chris Lituma, or perhaps more so his student
Laura Graham, will be able to chime in here as "detectability" was part of
her research, and she did base methodology on point counts, but then added
in time after for looking for shrike.
From my own experience, there are a few factors:
1. LOSH don't tend to vocalize and many point counts (e.g. BBS), in my
experience are more focused on auditory detection 2. Length of time for
the point count, wrt shrike detection, is important. Again, in my limited
experience, many point count methods are too short for shrike. I usually
wait 20 minutes, scanning with binoculars, before I move on to another
site/point during surveys.
3. Point counts are better when you have a nice open habitat, but not too
wide open. i.e. sometimes it's more effective for shrike to walk through
the habitat, stopping to scan every now and then for a length of time,
than standing in one spot.
4. Given 3, it matters if the point count is limited or unlimited
distance 5. The reliability of detecting a shrike varies considerably
based on stage of reproduction. It can also vary based on weather (yes, I
know "duh" - but e.g. in high wind, shrikes often hide or don't hunt from
as high up).
6. There are simply times when shrike don't want to be found - thus, my
"go to" (and the ON survey protocol lays all this out in much more
concrete and likely understandable way), is to visit a site at least 3x,
at least a day or so apart, and preferably throughout the season, before I
mark shrike as "absent" or a site as unoccpuied. Seems that if they are
going to be found, you'll see them, if you wait 15 or 20 minutes (often
less but not always). Whereas, you can also spend hours on one day and not
see them, then, go back a day later, and, there they are.
7. Road-side versus on-site are considerations (not necessarily
related to point count methods, but to survey method more generally, and
has to do with visibility, etc.) 8. Just an fyi, there has been some work
(again, Laura can add more specifics from recent work) as to whether call
playback helps.
9. Finally, the Survey123 App being developed includes a "time to first
detection" specifically to help us figure out/adjust for detectability.
Hope this helps. Happy to talk by phone!
Amy
________________________________________
From: eloshwg-bounce@xxxxxxxxxxxxx [eloshwg-bounce@xxxxxxxxxxxxx] on
behalf of Christopher Hill [dmarc-noreply@xxxxxxxxxxxxx]
Sent: Thursday, February 13, 2020 11:31 AM
To: eloshwg@xxxxxxxxxxxxx
Subject: [eloshwg] point counts
Hi all,
There was a strong statement on the phone call today (and I don’t
recognize everyone’s voice so I don’t know who I am quoting): “We’ve all
come to the conclusion that point counts are unreliable for detecting
shrikes.”
That surprised me! I would have thought that shrikes, though they don’t
vocalize reliably, would be very detectable because of open habitats and
conspicuous perching.
Exactly how reliable or unreliable point counts are is something I’m
interested in, and potentially I could add some data to the discussion,
but I guess I should catch up with where the discussion currently is,
first. I didn’t want to sidetrack the call but I would very much like to
follow up on this point, so if anyone could provide any pointers to
written stuff on this point, or is willing to field a phone call from me,
or is willing to put together a paragraph or some bullet points on this
consensus I would be very grateful.
Anybody?
Chris? ?a k y + j rם ^ +h wb 'nj[0 Z z) ? y R {
+ ) . j % p K ?? ? u(k G zǚ Z ڝا :) fz ? ?u ̨ j)m h ? ~
+- N ? ֦z ? h ? j + ] ,ʋ zf ? ! u ڶ ~ +- ? , Xf r? z
+ !jx?ʋ r *' ֭ Ǡ oiƭ ݊? j?z 칻? &ނ h i 0 zX + b ޖ !
email filtering provided by
(((GRIFFCOMM<https://nam03.safelinks.protection.outlook.com/?url=http%
3A%2F%2Fwww.griffcomm.ca%2Fmail-filtering-r%2F%3FALS&data=02%7C01%
7Cchill%40coastal.edu%7Ce4688540bad4440d23ea08d7b0a4ad21%7Cbf1f856b8ef
84e52be9387d3c3622797%7C0%7C0%7C637172093640958226&sdata=sV1uEzBPx
gtNLpmxtEI7ot8VltgqS3VO3fxgCG5ayfM%3D&reserved=0>
a k y + j rם ^ +h wb 'nj[0 Z z) y R {
+ ) . j % p K u(k G zǚ Z ڝا :) fz u ̨ j)m h ~
+- N ֦z h j + ] ,ʋ zf ! u ڶ ~ +- , Xf r z +
!jx ʋ r *' ֭ Ǡ oiƭ ݊ j z 칻 &ނ h i 0 zX + b ޖ !
