Equal tree counts reversed the change in species-list differences across 50 one-hectare Barro Colorado cells from 1985 to 2015

We tested spatial change across 50 fixed one-hectare cells in the Barro Colorado Island forest plot, Panama. We used living trees with measured main stem diameter of at least 10 mm in 1985 and 2015. Full species-list Sorensen dissimilarity rose from 0.18472 to 0.18761. In samples of 200 trees per cell, it fell from 0.45776 to 0.45061. Thus, equal counts reversed the direction of the species-list comparison. Under the same count control, exact expected abundance dissimilarity rose from 0.44546 to 0.45338. The two measures therefore gave opposite directions. Small trees, with diameter 10 to less than 100 mm, showed the same pattern. These signs held after each of 15 spatial strip deletions and across sample counts of 100 to 2000. Large-tree results were weaker. We rebuilt all eight census count tables from the original data and checked every species-by-cell entry. We preserve unchanged endpoint source tables, compact inputs, code, figures, provenance and numerical checks. This result applies to one finite forest plot. It does not identify a cause or estimate change across forests.

tropical forest · spatial biodiversity · species lists · species abundance · rarefaction · Barro Colorado Island · reproducibility