Bird lists from the same observer differed more after ten years than after five years on 222 Swedish routes in 1996–2025

We tested whether recorded bird community differences increased from five to ten years when the observer was the same. Swedish fixed-route data version 1.16 supplied 1,254 matched sets on 222 routes in 1996–2025. Each set used years t, t+1, t+5 and t+10, with one observer ID and survey dates within 14 days of year. We averaged within routes, then gave routes equal weight. Mean Jaccard dissimilarity was 0.3686 at one year, 0.3916 at five years and 0.4150 at ten years. The ten-year minus five-year difference was 0.0234, with a pointwise 95% route bootstrap interval of 0.0174 to 0.0295. Relative-count Bray-Curtis gave an increase of 0.0240 [0.0189, 0.0291]. Date, effort, complete-series and dependence checks kept the main increase positive. The selected routes were mainly in southern Sweden. A fixed observer ID does not fix detection over time. The result does not separate directional change from fluctuations with long memory or identify an environmental cause. The package includes the unchanged source archive, matched sets, code, bootstrap draws, figures, provenance and reproduction instructions.

bird monitoring · temporal turnover · biodiversity · observer effects · Sweden · matched surveys