Removing an earlier fitted pond response lowered mean correlation between duck survey changes in 22 prairie areas in 2000–2019

We tested whether an earlier pond model could account for shared changes between duck taxa in later survey years. We used ten taxa in 22 fixed prairie areas from the Waterfowl Breeding Population and Habitat Survey. We fitted pond slopes to annual changes in square-root duck estimates in 1975–1999. We kept those slopes fixed and removed the predicted pond component in 2000–2019. Mean correlation between taxa fell from 0.302 to 0.213. The change was −0.089. A nominal 95% block range was −0.119 to −0.051, based on 4391 of 4999 resamples with a defined result. Sparse records made the other results undefined. A check without one sparse area gave a change of −0.090 and used all 4999 resamples. Mean correlation remained positive. A same-period check gave a smaller decrease with uncorrected aerial densities than with corrected estimates. The result supports transfer of a statistical pond relation across periods. It does not identify a causal pond effect. Ducks and ponds share a survey, and joint observation error is not resolved. The package preserves source data, provenance, code, results, figures and checks from a clean repeat.

waterfowl · wetlands · species synchrony · survey error · temporal validation · open data