From Andy Didiuk
-----Original Message-----
From: Didiuk, Andrew (EC) <andrew.didiuk@xxxxxxxxx>
Sent: Thursday, February 13, 2020 5:51 PM
To: sergio.harding@xxxxxxxxxxxxxxxxx
Subject: FW: point counts
Hi Sergio,
Perhaps you could send this to the group as well.
Good afternoon Hazel,
I do not seem to be able to respond to the group list. Could you forward
this to the group, or to Chris? Thanks
Andy
Good afternoon Chris,
An interesting initial discussion regarding detection probability of shrikes
during point counts. Always a dilemma for all of our survey efforts for all
species.
I am working on my final data analyses for detection of prairie loggerhead
shrikes during road surveys in Alberta and Saskatchewan. These road surveys
are embedded in farmland habitat with remnant patches of grassland, riparian
corridors, active farmyards, abandoned farmyards, and shelterbelts for
shrubby habitat for shrikes. I have attached a copy of the survey protocol.
To assess detectability, we selected 12 routes in farmland, with 50-52 stops
each, and used the protocol of the our farmyard road surveys in the
attachment (view, broadcast call, view) for 200m x 200m plots at the stops.
We implemented our survey protocol along each route for 4 consecutive days,
with an average of 4-6 minutes at each stop. On the fifth day, we only
traveled to those stops where no shrikes had been previously detected and
increased the survey time from 6 to 10 minutes at each stop, including
moving the vehicle to different vantage points. The intent of this effort
was to look at the cumulative detections along each route, assuming the
results following the 5th day with longer more intensive searching resulted
in 100% detection of shrikes. We know this is not true, at some stops there
are likely shrike pairs with only a small portion of their territories
within the 200m x 200m plot and more surveys would be needed to detect them,
as well as stochastic probability of no detection. Therefore, the percent
detections are biased high to some degree. However, the data do indicate an
approximate probability of detection for a single pass along the survey
route with its stops.
We repeated this procedure for all 12 routes, with 4 consecutive days of
survey followed by the 5th day of more intensive scrutiny, for four time
periods (cycles), to address potential differences during the breeding cycle
due to behaviour of the pair:
cycle 1 ~nesting and egg laying ~May 11-31
both male and female
fussing about
cycle 2 ~incubation ~June 1 - 20
female on nest with nest recesses,
male feeding female
cycle 3 ~hatch and early nest site ~June 21 - July 10
both male
and female busy feeding nestlings, some early movement of young away from
nest cup
cycle 4 ~gradual movement of your from nest site ~July 11 to
July 31youngmore mobile, still being fed by male and female
When I complete these analyses and report I can provide if you wish. The
data suggests so far that a single pass during our surveys resulted
generally in a 25-35% detection of shrikes that might be present within the
survey plot (i.e. a portion of a shrike territory within the plot). This
is similar to some findings by Dave Prescott in Alberta, and I will try to
dig up that report and forward to you. The final analyses will provide a
more detailed assessment in terms of variation within and among survey
cycles.
In the prairie provinces, the shrike is extremely local in its distribution,
with pairs widely distributed. Depending on the region, I have to look at
100-200 habitat patches to find a shrike (see the attached protocol to get
an idea of the survey effort). Having said that, I believe there is some
potential for deriving a detection probability for our single pass survey
protocol at a given stop, but we will see how the final analyses unfold.
I also did a bit or work looking at variability of observers upon shrike
detection, with three observers positioning themselves to minimize influence
of behaviour of the other observers (see pic), each scanning the plot, then
entering the truck and recording their data as per the protocol. I will
have this worked up and provide if wish.
Andy
Andrew B. Didiuk
Wildlife Biologist
Canadian Wildlife Service
Environment and Climate Change Canada / Government of Canada
andrew.didiuk@xxxxxxxxx 306-975-4005
Biologiste de la Faune
Service canadien de la faune
Environnement et Changement climatique Canada / Gouvernement du Canada
Government of Canada | Gouvernement du Canada andrew.didiuk@xxxxxxxxx
306-975-4005
-----Original Message-----
From: eloshwg-bounce@xxxxxxxxxxxxx <eloshwg-bounce@xxxxxxxxxxxxx> On Behalf
Of Christopher Hill
Sent: February 13, 2020 10:31 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 !
Attachment:
Example of Route Detection.jpg
Description: JPEG image
Attachment:
Praire Loggerhead Shrike Road Survey Protocol Mar 15-18.pdf
Description: Adobe PDF document
