Common Black Hawk identified as Bald Eagle, and filtered out by the range model in September (SW New Mexico)

[mryakima]

Hello BirdNET team,

I run a BirdWeather PUC (station 28942) in riparian habitat near Alma,
Catron County, New Mexico (approx. 33.5, -108.9). A pair of Common Black
Hawks (Buteogallus anthracinus) nests about 2,000 ft from the recorder.
I hear them calling at close range many times, but the station almost
never records them.

The one time a call did get through, it was labelled Bald Eagle:

• Date/time: 2026-09-21 11:23 MDT
• Prediction: Bald Eagle, confidence 0.65, location probability 0.019
• Clip: attached (28942-20260921-172320.flac); the call is at 3.0–6.0 s

I confirmed by ear that this is the Common Black Hawk, not a Bald Eagle.

I think the location filter is part of the problem. BirdWeather reports
these weekly probabilities for my location:

• Common Black Hawk is above 0.5% only from April to early September.
  Later in September it drops to 0.46% and 0.43%.
• Bald Eagle stays above 0.5% all year (1.5–2% in September).

With BirdWeather's default 0.5% cutoff, Common Black Hawk is filtered out
in September while the local pair is still here, and a similar raptor
that is still allowed is left to take the label.

Two questions:

1. Does the classifier itself confuse Common Black Hawk with Bald Eagle,
   or is the Bald Eagle label a side effect of the filter?
2. Could the range model be underestimating how late Common Black Hawks
   stay in southwestern New Mexico riparian habitat?

I can provide more close-range recordings of this pair if that would help.

Thank you for BirdNET.

Scott Ray
Catron Co, NM