Mystery Bird Meets Imperial Woodpecker – Trail Cam Photos Revisited (Part 2)

If you haven’t done so already, please read Trail Cam Photos Revisited for a more comprehensive discussion of the image that’s central to this post (including an explanation of our conclusion that the mystery bird is larger than a Pileated Woodpecker). At the time of writing, I didn’t envision doing a follow-up, but the nagging sense that the Rhein Imperial Woodpecker film might be even more relevant than I thought initially led me to go through the film again and pull an additional frame that showed the body profile, with neck extended, more fully and accurately. While I added an update to the original post that included a brightened composite for comparison, there’s a bit more to say.

We’re aware that many (perhaps most) in the scientific and birding communities will accept nothing less than a clear, high quality photograph (or series of photographs) or video. We’re also aware that many people will dismiss any post-processing whatsoever, even when intermediate steps are shown and the processing is relatively limited (in this case only involving the removal of motion blur). I must add, as should be evident from the images in this post, my skills with photographic post-processing tools are very limited. (Patricia Johnson, my wife, had to talk me through the rotation of the images in Photoshop.) Nevertheless, we think comparing the de-blurred mystery bird with frames from the Rhein film showing an Imperial Woodpecker in flight with similarly positioned wings makes a compelling case that our mystery bird is indeed an Ivory-billed Woodpecker.

Our choice of location for the game cam deployment was not random. As with this image (from our old search area), which was obtained a week after a sighting in the same location, the camera was deployed in an area where we’d had recent possible contacts – multiple double knocks (scroll to the end of the trip report) heard within a few hundred yards before and during the deployment, about a week before the image was captured. We also recorded an apparent double knock on the day we retrieved the cards. Thus, in both instances there was a close temporal association between a putative encounter and obtaining (at worst) a strongly suggestive trail cam capture. But I digress . . .

To return to the Imperial Woodpecker, these two screen captures are the most salient.

Screen cap of Imperial Woodpecker in flight, shortly after take-off, at a different angle, but with similar wing position.

Screen capture of Imperial Woodpecker in flight, shortly after take-off, at a different angle, but with similar wing position.

Screen Shot 2015-09-05 at 6.15.38 AM

Screen capture of Imperial Woodpecker in flight.

I was unable to find frames in the film that replicate the angle from which the mystery bird was shot. In the first of these two frames, the bird is flying downward and is angled slightly away from the camera, obscuring the bill and foreshortening the neck and tail. The second is a ventral view from behind, and the bird is angled downward. Our mystery bird is ascending and is seen in profile. Nonetheless, the similarities in both the extent of white on the wing and physical structure are striking. This becomes even more apparent when the Imperial frames are rotated and sized to match the mystery bird. Be sure to click on the images to see the full sized versions.

Rotated, resized image of Imperial Woodpecker in flight. The bird is angled downward and slightly away from the camera obscuring the bill and foreshortening the profile.

Rotated, resized image of Imperial Woodpecker in flight. In the original, the bird is angled downward and slightly away from the camera obscuring the bill and foreshortening the profile.

Imperial Woodpecker in flight. Rotated and re-sized for comparison with mystery bird.

Imperial Woodpecker in flight. Rotated and re-sized for comparison with mystery bird. Although this is a more ventral and posterior view, the similarities in structure are dramatic.

I concluded the previous post by observing, “We realize that this is far from conclusive but can think of few alternative interpretations, all of which are problematic.” Based on the comparison with known stills of the Imperial Woodpecker, I am now firmly convinced that the mystery bird is an Ivory-billed Woodpecker and do not think there’s a reasonable alternative explanation. Frank’s comment was, “People have been executed on far flimsier evidence.”


Trail Cam Photos Revisited

Frank recently found a series of images from our trail cam deployment discussed here. These photographs, taken with a second camera, are of very poor quality, but they show what is clearly a Pileated Woodpecker on the target snag. Frank’s discovery led us to re-examine some of the images discussed in Frank’s post and elsewhere on the site because they gave us a reference object to assess the size of the birds in two other low quality trail cam photos.

IMG_2735-1 IMG_2736 IMG_2737 IMG_2738 IMG_2739

Based on this reference object and informed by outside evaluations, we’re confident that the “mystery bird” discussed in this post is in fact a Red-headed Woodpecker. (We still concur with the author’s analysis of the other image discussed in the post.) The bird is clearly behind the tree but not very far from it, and it is simply too small to be anything else. Despite my initial interest in this photo, I had been leaning toward Red-headed based on feedback from a number of people and on the length of the tail.


I have always thought that this was the most intriguing image in the series, although some reviewers have disagreed.


I’ve also always believed that the bird was behind the snag and in front of the somewhat more distant small branches, which would mean it’s large. Frank, who has by now reviewed perhaps 1 million trail cam images, has always agreed with this interpretation.

Frank’s discovery of the Pileated sequence led me to re-examine this photograph and dig a little deeper. One reviewer suggested that motion blur made it impossible to make any judgments about size or distance from the camera. In response, I did a bit of research and found Focus Magic, a forensic program designed to reduce or eliminate motion blur.

The results of running the image (bird only) through Focus Magic are interesting. (No other processing was done.)

Image processed with Focus Magic. Setting for a 35 degree angle and a blur distance setting of 13 out of 20

Image processed with Focus Magic. Setting for a 35 degree angle and a blur distance setting of 13 out of 20. The object was selected using the lasso tool in Photoshop, making it possible to de-blur the moving bird without affecting the rest of the image.

I shared this processed image with Louis Shackleton, a professional photographer friend who has a background in ornithology. I also sent Louis one of the Pileated images and these two other trail cam photos that were taken within an hour and ten minutes of the first, on December 7, 2014.



The first of these shows what we believe to be either a Red-headed or an Ivory-billed Woodpecker in flight and behind the snag. The other shows an intriguingly shaped but badly blurred bird in flight, passing between the camera and the snag. (The motion blur in this image is so severe that I was unable to make any meaningful improvements using Focus Magic.) Louis had this to say about the photographs he reviewed:

“ . . .I concur that the first image is the most interesting. Comparing that and the image of the PIWO on the tree, it seems to be larger, [and] as you say, I think it’s beyond the snag. The second image, it’s also beyond the snag, but there’s no way to gauge how far. The third image, it’s definitely in front of the snag, but also no way to tell how far in front.”

Frank and I interpret the photograph as follows. It shows a long-necked, long-tailed, slender-bodied bird that is somewhat larger than a Pileated Woodpecker flying upwards at about a 35 degree angle. There is white on the trailing edge of the wing, although it’s unclear whether this white is on the underside of the left wing, the upper side of the right wing, or both. We do not believe this white to be an artifact, since it appears faintly in the unaltered image; it becomes more fully resolved when the blur is eliminated; it is still present even at a blur distance of 20, the highest Focus Magic setting, when image clarity breaks down significantly.

Focus Magic processed image with blur distance set at 5

Focus Magic processed image with blur distance set at 5

Focus Magic processed image with blur distance set to 20

Focus Magic processed image with blur distance set to 20

A couple of considerably more ambiguous features are also intriguing. The bird appears to have a fairly distinct and sizeable bill, and in the Focus Magic iteration in which the white is most clearly defined (blur distance 13), there’s a hint of red on the head, although this could easily be an artifact. While the William Rhein film of an Imperial Woodpecker in flight was shot at a different angle, we think the profile and structure of the IMWO in that footage strongly resemble our mystery bird.

Screen capture of Imperial Woodpecker in flight

Screen capture of Imperial Woodpecker in flight from the film by William Rhein

Screen cap of Imperial Woodpecker in flight, shortly after take-off, at a different angle, but with similar wing position.

Screen capture of Imperial Woodpecker in flight, shortly after take-off, at a different angle, but with similar wing position.

Edited to add: To facilitate comparing our mystery bird with the Imperial Woodpecker in the frame shown above, I’ve created a composite image using the 13-35 de-blurred image, which I’ve also brightened. In addition to flying downward, the Imperial is angled slightly away from the camera, foreshortening the neck and obscuring the bill.Composite

It has been observed that there’s nothing to prove our mystery bird is a woodpecker, and that’s a fair point; however, the size, shape, and apparent white on the back of the wing are all consistent with Ivory-billed Woodpecker. We realize that this is far from conclusive but can think of few alternative interpretations, all of which are problematic.

Brief Update – Frustration

We recently purchased a new Plotwatcher Pro trail cam and are very hopeful that it will be a major improvement over the old Reconyx. Frank Wiley attempted to visit the search area today to place a trail cam on the large, downed sweet gum limb shown and discussed here. The road leading to the location is impassable; the entire area appears to be flooded, and placing the camera would probably be impossible even if the site could be reached.

I was planning to make a trip at the end of this month, but I need to have minor toe surgery. Even if that weren’t in the offing, it seems unlikely that the trip would be worthwhile. We just have to keep fingers crossed that Frank can get a camera on the limb before woodpeckers start to feed on it. There won’t be any news for the next few months, and I won’t be able to get back into the field until late fall or early winter. I may add a couple of pages in the interim.

Trip Report, Part Two: March 31-April 5, 2015 – Sighting on 4/3/15

As mentioned in the Part One of this trip report, I was alone in the search area from March 31st to April 2nd. Frank Wiley joined me from April 3rd-5th. He had a very robust, brief sighting on the morning of the 3rd, and we had a possible auditory encounter on the 4th. Conditions changed dramatically over the course of the week, and by the 4th, leaf-out had progressed to the point where examining tall trees for cavities and scaling had become very difficult. Weather conditions were generally good, although winds were high for much of the time. It rained heavily on the night of 3rd, and all we did on the 5th was set up a trail cam.

March 31, 2015

I hiked into what we now think is a hot zone and worked my way south, finding approximately a dozen downed sweet gums and the standing sapling that had been heavily stripped of bark. I did several playbacks and got reactions from pileateds and red-bellieds – calling and drumming but not attractions. I called it a day when I found a secondary feather that I suspected might come from an ivorybill. It seemed large for a red-headed and was white, except for the base, which was black on both sides. I bagged the feather and hiked back out. Upon examination, it became clear that it was indeed a RHWO feather; the differences are subtle. The innermost secondary of the Agey and Heinzmann feather, which has been confirmed as an innermost ivorybill secondary, is 3.5” long (thanks to Fredrik Bryntesson for providing the information on this feather, which comes from a letter from Heinzmann to Wetmore) and is all white. I have seen an ivorybill secondary that is 4.25″, though I’m not sure of the number. It had black on one side. The difference is a very fine one, especially since the innermost RHWO secondaries also have black on one side and white on the other. This was false alarm, but better safe than sorry. The U.S. Fish and Wildlife Service’s Feather Atlas is a great resource; I only wish that it included the ivorybill. Photo of RHWO feather taken with pocket camera in the field.

April 1, 2015

To my relief, nothing much of consequence happened on April Fools Day. I returned to the same area and did several playbacks over the course of the morning, generating increased drumming and calling from pileateds, red-bellieds, and red-headeds. The playbacks also seemed to cause Barred Owls and American Crows to vocalize, and after one series at 9:00 am, two crows came in silently to investigate and then called a little. I also found the “chopped” sweet gum sapling that day; it was within a hundred yards of the scaled sapling, in an area where there was an abundance of work on sweet gums.

April 2, 2015

I visited three other locations. I began with the area where I found a concentration of bark scaling in 2012 and where I recorded calls in 2013 and where the tree on the homepage is located. The hydrology in this area has changed considerably for several reasons, and I saw nothing of interest. I did encounter a couple of large moccasins. BigMoc Playbacks provoked some reactions from pileateds and red-bellieds. Strangely enough, the pileated did not respond to PIWO playbacks.

I then went to another patch where there have been contacts over the years, and where we found a heavily scaled downed sweet gum last spring. There has been no further work on that tree, and the decay has progressed to the point where the bark is loose and Bess beetles, skinks, and slugs are residing underneath. This tree has now been down for between between one and two years, and the twigs and small branches are starting to fall away. The bark too is getting looser.

Impressionistically, this is quite consistent with Tanner’s characterization of the decay process. We suspect that this lack of return visits is typical of ivorybill feeding behavior, since we’ve monitored multiple feeding trees and have seen this pattern repeatedly. In one case, the small scaled oak in the old search area, there was one round of scaling found in January and a second in May or June, but in all the other monitored trees have been scaled once and that’s all. In my experience pileateds typically return to feeding trees on a fairly regular basis, and in the case of this particular tree, there’s obviously an abundant food supply under the bark.

April 3, 2015

Frank and I hiked into the ‘hot zone’. We did some playbacks at approximately 8 am, although I did not record anything about them in my notes. We proceeded in a more or less southerly direction with Frank in the lead by about 10 yards; I was walking slowly and looking up and to my right. As we approached a body of water, Frank stopped and blurted something unintelligible. I caught up with him, and he said he had gotten a very good look at a male Ivory-billed Woodpecker that had flushed, presumably from a fallen log lying in the water or possibly from water’s edge. The distance was no more than 20 yards. I handed Frank my field book so that he could draw what he saw and record his observations. I’ve included the sketch and transcribed his description, with a few redactions related to the specific location. These were done immediately after the sighting and without reference to a field guide. Not included in the description is his estimate that the sighting lasted 2-3 seconds, as the bird, which was in the open for perhaps 10 yards (we’ll measure this distance at the next opportunity), flew upward into an opening in the woods across the water, and in his excitement, he mistakenly gave the date as April 4. Screen Shot 2015-04-08 at 4.58.02 PM

  1. Big traffic cone shaped WHITE bill (3”ish?)
  2. Solid black head/face – light colored eye (whiteish)
  3. Bright red crimson crest puffed up not what you’d expect.
  4. Stripe on face beginning behind.
  5. Stripes on back form chevron over rump.
  6. Wings long/thin shallow rapid flaps.
  7. Rear 1/3 to 1/2 of wings white, all the way out to primaries.
  8. Long tapered tail.

Later that day in an email exchange, we added the following comments.


“At the first sign of movement, I assumed a Wood Duck had flushed, looked in that direction, and immediately saw the crimson red of the crest. I then thought, “PIWO” but noticed the big white traffic cone bill, and an almost entirely black face. There was a white stripe that started below/behind the light (I got the impression of white – not yellow) colored eye. The crest was not “groomed” as is usually seen in most of the artwork – rather it was puffed up as if the bird were agitated.”


“I just have a few things to add to this. I was about 30 feet behind Frank and was looking in the opposite direction, so I didn’t see it. Frank kind of blurted something, and when I got to him, he was clearly and deeply shocked and absolutely sure of what he saw. I told him to sketch it and then write a description. I didn’t hear any splashing or wood duck sounds, and I’m sure I would have if it had been a duck. He mentioned that he didn’t think the bird was terribly frightened and had probably flushed due to the sound of something approaching, since he was likely out of view.  We were approaching the area where I found a concentration of bark scaling, including both the hatchet sapling and the other scaled one I found on Tuesday. (A better photo and a detail of a worked exit tunnel on that one are attached) and were perhaps 100 yards away. I had not mentioned this to Frank until after the sighting. It’s also not at all far from the cluster of old cavities we found last spring. I guessed that the bird was drinking from the creek, and we are contemplating putting a camera on the log, in the event that it’s a place that might be a preferred spot.” The only other item of note on the third was our finding an interesting looking old cavity in a fallen sweet gum. Though partially healed over, it appears to have been quite large when fresh, at least 6″x4″ and the area beneath it had been stripped of bark. DownedCav

April 4, 2015

We returned to the same general area on the morning of the 4th, covering a lot of territory and finding more scaling on downed sweet gums. At approximately 11 am, we heard a loud knock or knocks. Frank heard a softer first knock and a much louder second one. I only heard the second, but it was loud, sharp, and woody. It’s worth noting that though louder second knocks are unusual, they are by no means unheard of . . . “the second blow is louder in 21 out of 119 recorded examples of double-rap displays by seven Campephilus species we studied . . .” The knock or double knock had come from the northeast.

We decided to do some playback, after which we heard two more possible double knocks in close succession, now from the west. I was not overly impressed by these, feeling they somehow lacked the energy I associate with really good knocks. Frank agreed about the first but thought the second was as good as any he’s heard. Slight differences in perception like these are not at all uncommon in our experience.

April 5, 2015

On the morning of the 5th, we went out to place a trail camera where Frank had his sighting. We’re hopeful that this new higher resolution camera will provide unambiguous results. It’s worth noting that we flushed a Pileated Woodpecker from virtually the same spot and from a similar distance as we approached the water. There was no mistaking it, and within a roughy similar time frame, and with a virtually identical flight path, it was possible to note an equivalent number of field marks, including the facial pattern; we agree that this pileated appeared to be a female. Weather and lighting conditions on both days were similar with overcast skies, perhaps slightly darker on the morning of the 5th.

Encounter Rate

      A poster on the Ivory-billed Woodpeckerl Researchers Forum raised a couple of questions about sightings and their quality, and about our encounter rate. I thought I’d take this opportunity to respond here.
      A number of people have had possible sightings in this area. Quality is a subjective term, but it’s safe to say the quality has varied. In most cases, the sightings have been by single observers. A few have involved perched birds. Distinctive features – black crests, malar and dorsal stripes and bills – have been noted in several instances; however, most have been fleeting glimpses of birds in flight. One recent flyover may have involved a pair – field marks were seen on one of the two birds. Just as we see little point in posting additional audio unless conclusive documentation is obtained, we don’t think that posting details of sightings serves much purpose, unless it’s a robust, two-person observation or is accompanied by a clear photograph.
      A little background on the encounter rate. I made my first search trip to Arkansas back in April 2007. Since then I have spent countless days in the field, probably upwards of 100 by now, in Arkansas, South Carolina, Florida, and Louisiana. I’ve learned a lot over the years; I’ve made some mistakes; I’ve written off some possible encounters as wishful thinking and have tried to maintain a healthy self-skepticism. I’ve also been somewhat obsessed with finding a category of feeding sign that is beyond the physical capacity of a Pileated Woodpecker to produce and is therefore diagnostic. As discussed in some of the recent posts, I believe I have succeeded in this, but of course my belief remains just that and is entirely unproven.
      While we don’t count finding suggestive foraging sign in estimating the ‘encounter rate’, which is actually a possible or putative encounter rate, the quantity of scaling that I consider to be Grade A or that is congruent with Tanner’s description far surpasses what I have seen anywhere else. Our old search area was considerably more compact, and in 2009 and 2010, before the adjacent parcel was logged, it was fairly easy to find suggestive feeding sign within that compact area. Since that time, I’ve refined my criteria for Grade A scaling and have made them somewhat more stringent. Nevertheless, this year in the new search area, I have found sign that meets my criteria or fits the Tanner model on virtually every field day. During trips last year, fresh scaling seemed to have been concentrated in one or two locations, and this year it seems to be more scattered. This is purely impressionistic, but whatever the case, there is an abundance of feeding sign that is unmatched by anything I’ve seen elsewhere. And I have never seen scaling that resembles what’s shown in my May, July, and October posts anywhere else, save in the former Project Coyote search area.
      Regarding the encounter rate itself: we include all putative encounters – DKs, whether ambient or in response to ADKs, putative kents, and possible sightings – even weak possibles. We estimate that we have been having one putative encounter for every 20-24 person hours in the field. The rate was considerably higher on my recent trip but probably nowhere near high enough to change the estimate. I should also add that, purely impressionistically, this is the only location where I have heard frequent apparent responses to ADKs. We had a possible attraction in the old search area and did get apparent response DKs from time to time, but nowhere near as often as in the new search area.
      The challenge lies in figuring out where the birds are roosting and nesting, if indeed they are present. We suspect there may be a nesting pair, as auditory encounters in the southern sector have frequently involved two sources of sound. To date, we have no indication that more than one bird is present in the northern sector. We have had very little success finding cavities in either sector, but it has also been difficult to find Pileated Woodpecker roosts in the deeper parts of the swamp, even in obvious, defended home ranges.
      I will be making brief trips in November and December and probably a somewhat longer one in January.