Flood flags gave no clear taxon richness difference in four UK beetle trap samples in 1993-2015

We tested whether flood-marked pitfall traps had lower recorded taxon richness after control for collection and fixed trap effects. Four England and Scotland samples in 1993-2003 and 2004-2015 used 6,299 trap events in 643 mixed-status collections. Flood coefficients were -0.103, -0.060, -0.219 and -0.048 distinct positive taxon codes per trap. All four year-cluster intervals included zero after a parameter-count correction and adjustment for four tests. Removing explicit no-beetle codes left 109 to 351 events per sample and changed one slope sign. Collection-only models gave two negative intervals that did not persist with fixed trap control. These results do not establish a common richness loss or validate flooded zeros. Few year groups, selected mixed collections, local habitat and reporting differences limit inference. The test concerns weather-related observation bias. It does not estimate a climate change or air pollution effect. The package gives source hashes, code, aggregate model evidence, a figure and two numerical checks. External raw files and record-level data remain outside the bundle.

carabid beetles · observation bias · flooding · taxon richness · fixed effects · null result · regional replication