Eastern Phoebe reports at severe-cold FeederWatch sites fell by half for three winters after Winter Storm Uri (2021); a general cold effect on small invertivores was not shown (US, Canada, 2002-2024)

We tested if short, extreme winter cold decreases wintering populations of small insect-eating birds. We used 3.49 million Project FeederWatch checklists (US and Canada, seasons 2001-2024) and NOAA CPC daily minimum temperature. The exposure was the coldest 3-day period of each winter relative to the 1991-2020 normal. Pre-registered tests gave weak or no results. For Winter Storm Uri (February 2021), the drop of small invertivores relative to other species at the same site was larger at colder sites in the next winter (+0.0027 reporting rate per degree C, 95% CI +0.0005 to +0.0048). But 2 of 15 placebo years gave slopes as large, and the slope was gone with latitude and longitude controls. In all winters from 2002 to 2024 (139,476 site-seasons), the previous-winter cold slope was +0.0013 (-0.0001 to +0.0027), with permutation p = 0.085 and no clear effect in 2002-2012. An exploratory event study tested a 2021 field report. At sites where Uri was 7 degrees C or more below the normal winter low, the Eastern Phoebe reporting rate decreased from 0.52 to 0.26. Sites without unusual cold went from 0.41 to 0.50. The fixed-effect contrast was -0.35 (-0.46 to -0.25) in 2022-2024, with no change before the storm. Other small insect-eaters that do not use feeders (Hermit Thrush, kinglets, warblers, Eastern Bluebird) also tended to decrease, but some had earlier trends. Main limitation: Uri cold was mostly in Texas and the southern Plains and the comparison sites were mostly in the Southeast, so regional changes can contribute. The phoebe result is exploratory.

Winter Storm Uri · extreme cold · Project FeederWatch · Eastern Phoebe · winter mortality · insectivorous birds · event study · placebo test · climate extremes · observation bias