Rain predicted more flood flags in Scottish beetle traps in 1994-2003 but gave no clear result in three other samples

We tested whether rain predicts recorded trap flooding in a public beetle survey. We joined UK Environmental Change Network records to Met Office rain and temperature data. We used 5,879 location-date records from ten sites in England and Scotland in 1993-2015. Four separate models controlled location by month, year, temperature and trap duration. In Scotland in 1994-2003, an extra 50 mm of rain per 30 days was linked to 7.20 more percentage points of flooded traps. The four-test adjusted interval was 3.78 to 10.62 points. Three other samples had positive estimates but intervals that included zero. The early Scottish result held with shorter intervals, ten-trap records, each site removed and two-year uncertainty groups. It did not clearly repeat in 2004-2015. This is evidence about the survey process. It is not evidence of beetle loss or a causal climate effect. Country-scale monthly weather, selected records, possible changes in flag use and few year groups limit inference. The package gives the model panel, code, source hashes, checks and a figure.

carabid beetles · rainfall · observation bias · pitfall flooding · regional replication · null results