Recorded species lists became more similar on 3,564 Swedish forest plots between 1999–2004 and 2019–2024

We tested whether recorded species lists became more similar between the same Swedish forest locations over 20 years. We used open National Forest Inventory records for 3,564 matched plots. Each endpoint survey covered 100 square metres on productive forest land. We compared 193 monitored species at 6.35 million fixed plot pairs. Mean Simpson dissimilarity fell from 0.4129 in 1999–2004 to 0.3927 in 2019–2024. The change was −0.0202, or a 4.9% decrease. Its 95% interval from 100 km spatial blocks was −0.0260 to −0.0146. Mean recorded richness rose from 15.64 to 16.83 species. The decrease in dissimilarity also appeared with 200 km blocks, the same survey month, vascular species alone, and a wider sample with 19-year or 20-year gaps. All 21 within-county comparisons had the same direction. An independent species-set calculation confirmed the main means for every plot pair. The result concerns recorded lists in a fixed full-area sample. It assumes that repeated published coordinate pairs identify the same plots. It does not correct detection error, establish a national area-weighted trend, or identify a cause. The package preserves compact source data, native identifiers, provenance, code, selection tables, results, figures, and repeat instructions.

biodiversity · forest vegetation · Sweden · species composition · spatial similarity · repeat surveys · reproducibility