Observer changes had no clear link to pooled route call-list turnover on 506 NAAMP routes in 2001–2015

We tested whether annual frog and toad call lists changed more when the observer changed. We used the USGS NAAMP release and matched the same eight to ten sites over three survey visits in each of two adjacent years. The main sample has 1,590 year pairs on 506 random routes in 21 US states in 2001–2015, including 135 observer changes. The response was Sorensen distance between annual taxon lists pooled across a route. A linear model had route and state-year effects and controls for survey date and air temperature changes. The adjusted observer-change estimate was +0.0196 on a zero-to-one scale. Its nominal 95% interval was −0.0198 to +0.0590, with clusters by route and year. All planned route-level checks had intervals that included zero. A planned secondary measure, mean distance between lists at matched individual sites, gave +0.0533, with interval +0.0205 to +0.0861. Thus the main route-level association was uncertain, while the site-level response had a positive association under the model. These records do not identify true colonization, extinction, or observer error. Non-random observer changes, local habitat changes, unmeasured survey conditions, and the selected complete-visit sample limit interpretation. The package contains unchanged source data, source hashes, a local plan, matched-site tables, all results, figures, code, and an independent raw-data and model check.

amphibians · NAAMP · observer changes · species turnover · call surveys · sampling effort · reproducibility