Mean plant temperature scores rose on 1,699 matched Swedish forest plots from 1999–2004 to 2019–2024

We tested whether recorded plant lists shifted toward species with higher temperature scores on fixed Swedish forest plots. We used open National Forest Inventory records and EIVE 1.0 plant scores. Each of 1,699 plots had 100 square metres of productive forest at visits exactly 20 years apart. Visit days differed by at most 30 days. Each visit had at least three species with exact name matches. Mean temperature score rose from 3.2606 in 1999–2004 to 3.2789 in 2019–2024. The change was +0.01835 units on a 0–10 scale. Its 95% interval from 4,999 resamples of 61 spatial blocks was +0.01263 to +0.02377. These are indicator units, not degrees Celsius. The change left after removal of linear species-score links with light, nutrients and moisture was +0.01666 (+0.01154 to +0.02163). An exact split gave +0.03771 from additions to recorded lists and −0.01936 from losses. Thus, the rise did not require loss of cold-associated plants. Six added synonym matches gave +0.01855. Date, list-size and spatial-block checks kept the same direction. All six survey-year groups had positive means. A species-label shuffle gave a two-sided probability of 0.0441 under its exchangeability assumption. Nutrient and moisture scores also rose; the light change was uncertain. Earlier work already found trait changes in Swedish forest plants. This contribution gives a later matched-plot estimate and a repeatable score and gain/loss analysis. It does not assign a climate cause or estimate a change for all Swedish forest area. Detection change, an incomplete species list, fixed indicator scores and unmeasured forest change limit the interpretation. The package contains source extracts, EIVE data, provenance, code, all results, figures and independent checks.

plant communities · forest vegetation · Sweden · temperature indicator · EIVE · paired plots · species occurrence · spatial block bootstrap · reproducible research