Date: 10/4/19 5:07 am
From: <clearwater...>
Subject: [obol] Oregon eBird Tropical Kingbird statistics
Bob and all,

No confusion about what you posted. My posting was about establishing reasonable probabilities, not categorically excluding remote possibilities.

The sample of 12 vocal kingbirds that you brought up is enough to establish, with very high confidence, that there is not an *equal* (50-50) probability of Couch's vs. Tropical Kingbird occurring here.

A regional sample of 50-some calling kingbirds is enough to establish with high confidence that the ratio of Tropical to Couch's Kingbirds is at least 10 to 1, here in the Pacific Northwest.

That doesn't absolutely rule out that one or a few Couch's Kingbirds might have gone unrecognized as such, by keeping mum. But that possibility doesn't have much impact on the potential use of Tropical Kingbird reports for reasonable scientific purposes (e.g. establishing the typical timing for northern movements, or their habitat association while visiting the region).

Finally, the 290 reports you found without mention of vocalizations don't necessarily indicate that 96% of such reports could be in error. In many/most cases, there is more than one eBird report for a given bird, because of the tendency for birders to flock to see stakeout birds.

For example, there are 17 separate eBird reports of a single Tropical Kingbird near the boatworks on Wireless Rd near Astoria from 16 Nov through 2 Dec 2015. Fifteen of those reports don't mention vocalizations, or else note that the bird was silent. But vocalizations consistent with Tropical Kingbird were noted on 29 Nov 2015. Likely all 17 reports were of the same bird, which was positively identified by vocalizations at least once.

If you work through the full series of reports from a given location in Oregon, Washington or British Columbia (I did this for quick a sample of about 30 cases along the Washington coast yesterday), you'll see this is a frequent pattern. There are often multiple "me-too" reports of birds that were either silent or where the observers didn't bother to note vocalizations, but somewhere along the line someone has heard vocalizations consistent with Tropical Kingbird. Treating each report in such a series as an "independent" observation, as you've done with the Oregon sample, will give misleading results.

Joel

 
Join us on Facebook!