Scoring Scopoli's

A recent thread on birdforum got me thinking about Scopoli's Shearwater again.

The original poster provided two images for ID, with one suspected of being a Scopoli's (top here) and obviously so, and the other a Cory's in similar conditions.

After years of watching both Scopoli's and Cory's in seawatching conditions (rather than from pelagics) I was of the opinion that Scopoli's often has an appearance of having a narrower, paler more broken leading edge to the hand between the carpal joint and the base of P10.

So I decided to run an image survey of both Scopoli's and Cory's to try and assess this. (Fair use applies to all images used here).

Scopoli's examples

Cory's examples

Quantification of these things can be difficult, but I attempted to apply a scoring system, between 0 and 3 to the primary lesser and median coverts, and the marginal coverts in terms of "thickness" (i.e how far dark feathering extended towards the median of greater primary coverts, and "density" i.e. how black and widespread these feather tracts were on a whole, leading to the impression of a dark hand region, or where these tracts broken, interspersed with white or fringed white (less dense).

Birds like these Scopoli's scored low in both thickness and density as examples.

Where as Scopoli's like this scored higher.

I will add all sorts of caveats here right at the outset. This is a quick and dirty analysis. Images are of course affected by all manner of parameters, including lighting and post processing etc.

I have endeavoured to ensure that all my samples were correctly identified however, and that there's no duplicate individuals (though I won't swear in court on that).

I post merely as a pointer. A search image feature that may help seawatchers ( I hope). I will leave it to seabird researchers to grill skins, or to go and try obtain sets of images of birds, ideally where both species are congregating in numbers under the same viewing conditions for the sake of robustness.

My sample set for Scopoli's numbered 128 individuals.

My sample set for Cory's numbered 153 individuals.

The average combined score for both parameters for Scopoli's was 2.0

The average combined score for both parameters from Cory's was 3.2

So some apparent difference between the two species is obviously there.

How did that break down?

Well as you might think, Scopoli's was heavily weighted to the lighter side of the spectrum. 10.9% scored zero values in thickness, 21.1% scored zero values on density.

69.5% scored a One on thickness and 63.3% scored a One on density. (Note, these values did not always align. I.e. a value of One for density could have a value of 2 for thickness etc.)

What this does suggest however is that Scopoli's does show a strong tendency to have a paler forewing on the underside of the hand, with perhaps 4 out of 5 individuals on the pale side of the problem.

How about Cory's then?

Well... Cory's is a much more variable beast, with a wider, more even spread of values. However it was noted that individuals with very narrow thickness were in short supply, with the weighting of birds scoring mostly One or Two for each parameter.

This of course means some serious overlap.

That said, there really is a je ne sais quoi when it comes to Scopoli's, a paleness and visual effect that really jumps out more than the number imply, and birds with virtually no black leading edge "should" be Scopoli's.

That said again, I would expect there to be exceptions out there, waiting to be snapped.

Indeed, I found outliers to some of the existing features. E.g. Scopoli's with a spot at the bases of both P10 and P9 exist.

How does this play out with Irish birds?

The first Irish Scopoli's off Cork, fulfilling all the ID criteria outlined by Flood et al, so checking the leading edge of the wing is somewhat unnecessary, but with a thickness scoring of about One and density markings of One also this does line up well.

Images by Paul Connaughton from Rare Bird Alert article.

The Bridges of Ross "claim" is more difficult to assess due to the poor quality of the images.

I had discussed this previously on the Band group when the news was put out.

(All images harvested from publicly available video on ebird and Vimeo)

Looking at this again in even greater detail, it's clear that you can't really even say where P10 is or P9 is and what is white and what isn't.

Indeed, the original images seem somewhat cherry picked, with the level of inter frame artefact and variability being extremely high between other frames in the video.

With none of the features really being assessable, I don't think this claim can pass muster, (however with the state of the IRBC nothing would surprise me any more).

The leading edge thickness change and warps between frames, but is consistently present at least. However, not to a degree where it can be scored reliably.