Drought lowered bird counts in the dry central US plains again in 2006-2025, about 7-9% at SPI -3 (10.7% published); ten rain-tracking species gave half and the montane response did not repeat

Albright et al. (2010) used North American Breeding Bird Survey (BBS) counts from 15 central US states for 1989-2005. They found that bird abundance in the dry Great Plains domain decreased with drought (32-week June SPI slope +0.0387 per SD), and that montane birds increased. We repeated the design with the BBS 2026 release and route-level SPI from NOAA nClimGrid (1,393 routes, 32,748 route-years, 1967-2025). We then tested 2006-2025 out of sample and added domain-by-year effects, survey-condition controls, lags, a next-year placebo, a 100-400 km ring test and regional year totals. The dry-domain slope replicated (1989-2005: +0.033) and held in 2006-2025 (+0.031, year-clustered 95% CI +0.012 to +0.049; +0.024 with domain-year effects; -7% to -9% at SPI -3). A March-June temperature control reduced it to +0.017. Ten species selected on 1989-2005 data gave about half of the 2006-2025 slope (+0.155 per SD), among them Lark Bunting, Dickcissel, Cassin's Sparrow and wetland birds. These species came back the next year. The other species decreased less (+0.012 per SD), the decrease continued into the next year, and it was absent in 1967-1988. Nearby wetter routes did not gain birds, and regional totals fell. The published montane increase and humid-domain migrant slope did not repeat after 2005. The forecast skill for route changes was small (+1.4%). Limitations: BBS counts mix abundance and detection, the guild definitions are approximate, and there are few year clusters.

drought · standardized precipitation index · Breeding Bird Survey · Great Plains · out-of-sample test · replication · nomadic grassland birds · Neotropical migrants · weather extremes · observation bias