After large mapped fires, open-land bird counts increased by about 10% and total counts did not change on 103 US Breeding Bird Survey routes (mostly western), 1986-2025

Question: do bird counts on Breeding Bird Survey (BBS) routes change after a large fire burns part of the route, and does the change depend on habitat guild? Approach: we joined 31,514 MTBS fire perimeters to 3,152 US BBS route paths. A route was burned when fire covered at least 20% of a 500 m route buffer in one bird-year after 5 years with no mapped fire. 103 burned routes (84 in the West, 96% wildfire area) were compared with 1,772 unburned routes in the same Bird Conservation Region. We used a stacked before-after-control model with observer-on-route and region-year fixed effects. The plan was written before we fit outcome models. Results: in post-fire years 1-5, open-land species (AVONET habitat density 3, not waterbirds) increased by 0.102 log units (95% CI 0.037 to 0.166; placebo p = 0.01). The increase repeated in fire cohorts of 1986-2004 (+0.131) and 2005-2022 (+0.093). It became larger with burned fraction (+0.070 at 10%, +0.186 at 50%). Total counts (-0.001, CI -0.052 to 0.049) and species richness did not change. The within-survey open-minus-closed contrast was +0.145 (CI 0.010 to 0.281; no pre-trend). Woodpeckers (+0.069) and closed-habitat species (-0.044) had CIs that included zero. Closed-habitat loss was clear only in early cohorts and at 50% burned. Lazuli Bunting (+0.34) and Mourning Dove (+0.27) increased most. Golden-crowned Kinglet and Warbling Vireo (-0.20) decreased most. Limitations: few eastern routes (19) and few prescribed-fire routes (2); a pre-trend for open-land species (p = 0.027); a waterbird negative control that also increased (+0.095, CI included zero); route lines from about 1998; no burn severity; and BBS counts mix abundance and detection.

wildfire · MTBS · Breeding Bird Survey · habitat guild · open habitat · closed habitat · woodpeckers · disturbance · weather extremes · event study · difference in differences · placebo test · observation bias · AVONET · biodiversity · North America