Random catch thinning reproduced 91% of the mean first-pass taxon shortfall in selected English three-pass fish surveys, 2000–2025

A first electrofishing pass catches fewer fish and can record fewer taxa than three passes. We tested how much of the taxon shortfall was expected from the lower fish count. We used 5,235 fully counted three-pass Environment Agency surveys at 1,773 English sites in May–October 2000–2025. For each survey, we calculated exact expected richness in a random sample without replacement from its pooled catch, with the first-pass fish count. We gave sites equal weight. Mean richness was 4.896 taxa in pass 1, 4.946 in the random-sample expectation, and 5.458 in three passes. Thus random thinning reproduced 91.0% of the mean first-pass shortfall. A catchment resample gave a nominal 95% range of 86.0–97.8%. A later check that also resampled common three-year blocks gave 81.9–107.5%. The observed-minus-expected difference was −0.050 taxa. Its range widened from −0.083 to −0.012 with catchments alone to −0.111 to +0.038 with shared years. The remaining mean difference is therefore uncertain. Period, method, region, and lamprey checks showed variation. This is a conditional catch comparison, not an estimate of complete fish communities or capture probabilities. Fully counted surveys are a restricted subset. An independent parser reproduced all included counts, and a separate source table agreed at 103,185 run-count cells. A clean repeat matched all 16 numerical output files. The package preserves the source data, provenance, code, results, figures, and repeat steps.

fish · electrofishing · rarefaction · sampling effort · taxon richness · England · biodiversity monitoring