Winter Sage Sparrow identification dilemma: Sagebrush or Bell's Sparrow?

Presumed Bell's Sparrow (canescens). Borrego Springs, California. December 26, 2014. Greg Gillson.
Note the sharply-defined black malar and unstreaked back.

In 2013 the American Ornithologists' Union split the Sage Sparrow into Sagebrush Sparrow (Artemisiospiza nevadensis) of the Great Basin and NW states, and Bell's Sparrow (A. belli) found in California and western Nevada.

The Bell's Sparrow includes the dark nominate form (A. b. belli), a resident of coastal California, and the migratory "Mojave" form (A. b. canescens) of eastern California and western Nevada--a "tweener" population, genetically more similar to Bell's than Sagebrush. But ID can be tricky, especially in winter when both forms (nevadensis and canescens) mix together on the wintering grounds in the American SW deserts (from central Arizona west to SE California, and NW Mexico).

This past week I had the opportunity to observe over 50 birds in 3 locations around Borrego Springs in eastern San Diego County. Well, I saw that many birds, but certainly not very well--they prefer low scattered creosote and other desert bushes in the sand. I tried photographing as many as I could during my (unsuccessful) search for LeConte's and Crissal Thrashers. The desert at dawn was just below freezing, but warmed quickly to about 60 degrees by late morning.

When submitting my reports to eBird of Bell's and Sagebrush Sparrows from this field trip, I received this reply: 

Despite recent papers describing A. b. canescens (Bell's Sparrow) and nevadensis (Sagebrush Sparrow) in museum trays, many Southern California experts feel strongly that such winter birds cannot be diagnosed to species in the field without a strongly supportive series of photos.  Phil Unitt wrote in the San Diego Bird Atlas: "Patten & Unitt found too much overlap between nevadensis and canescens for the latter to be distinguished.” 

While there is much overlap, some expert feel that with excellent photos of the front, back and sides that it *might* be possible to assign a species to limited birds.  Therefore, each observation of wintering Bell's (A. b. canescens) or Sagebrush (A. nevadensis) must to be carefully supported with a good series of photos per individual before record confirmation is considered in eBird.

This is certainly commendable restraint to the data integrity of eBird. However, I've never been one to shy away from bird identification when it was difficult. In fact, just the opposite.

One of the "recent papers" in the eBird reply is probably Peter Pyle's preliminary ID discussion based on museum specimens. (pdf here)

My summary of his findings...

"It appears a combination of bolder back streaking and reduced malar can be used to separate most or all nevadensis from most or all canescens....I did not find any intermediates in fresh plumage [September-December] that could not be placed to species."

On the other hand, Patten and Unitt's paper (pdf here), not only argued that nevadensis and canescens weren't good subspecies (before the split), they argued that the definition of subspecies was too broad. They argued against mean differences as the determining factor for subspecies recognition and for diagnosability exclusively.

Patten and Unitt looked primarily at mantle shade, wing cord, and back streaking. However, their definition of back streaking was in an all-or-nothing manner, not degree. Thus, they said belli has a streaked back, when other authors use the unstreaked (or faintly streaked) back as an identification mark for belli to separate it from the significantly streaked back of nevadensis. They used wing cord for size determination and found more overlap than previous authors between nevadensis and canescens, and showed a cline for this measurement.

So, with eBirds warnings and Pyle's encouraging words, I dove right in. The top photo I believe is Bell's (canescens) and the photo immediately following I believe to be Sagebrush (nevadensis).

Presumed Sagebrush Sparrow. Borrego Springs, California. December 26, 2014. Greg Gillson.
Note the striped back and paler more diffuse malar.

For comparison, below are two photos of breeding Sagebrush Sparrows (A. nevadensis) from SE Oregon. The first shows a stronger malar (lateral throat) stripe than typical. And you can't see back stripes from this angle. There are dark interior shaft streaks on the lesser coverts and scapulars--these are wing feathers, though, not back feathers.

Confirmed Sagebrush Sparrow. Malheur NWR, Oregon. May 26, 2007. Greg Gillson.
This individual has a stronger darker malar than most.

This second photo of Sagebrush Sparrow (below) is typical, showing a weak malar and strongly striped back.

Confirmed Sagebrush Sparrow. Malheur NWR, Oregon. May 27, 2009. Greg Gillson.
Note the densely streaked back and weak malar, typical of this species.

So now, for eBird documentation, I will show a series of photos of the best photographed presumed Sagebrush and Bell's (canescens) Sparrows.

I photographed 9 different birds. Of these I believe 3 show characters of Bell's (canescens) and 3 show characteristics of Sagebrush. Three others don't show the right angle or are out-of-focus enough that I can't tell, though 2 of them I lean towards Bell's. All non-photographed birds, as well as ambiguous ones go down on my eBird list as Sagebrush/Bell's Sparrow (Sage Sparrow).

I have selected the 2 best series of photos of both presumed Bell's and presumed Sagebrush sparrows.

Bird A, presumed Bell's (canescens): Photos 9784, 9786, 9787, 9788
Old Springs Road Open Space Preserve, Borrego Springs, San Diego County, California
December 26, 2014.
[Click photos for full-sized view.]

Bird B, presumed Sagebrush: Photos 9790, 9791, 9792, 9794
Old Springs Road Open Space Preserve, Borrego Springs, San Diego County, California
December 26, 2014.
[Click photos for full-sized view.]

Update: During the winter of 2017-2018 I was sent a quiz of sorts. It was a photo of 10-12 winter "sage sparrow" museum specimens of both Bell's (canescens) and Sagebrush Sparrows. There was wide variation between darkness of head and malar thickness. I identified at least half the birds with confidence--the others I would not have attempted to identify in the field. How did I do? Was I fooling myself? No. All the birds I felt confident about I identified correctly! Of the others? Back streaking (not scapular streaking) was the clinching ID mark. If one gets a clear view of the center of the upper back, identification can be made with certainty.