Nine burned Cedar Creek plots closed 48% of the richness gap and 23% of the biomass composition gap to controls after fertilizer stopped (2015–2022)

We tested plant community recovery after 32 years of fertilizer addition at Cedar Creek, Minnesota. We used open E002 data for 27 plots at the three highest past nitrogen rates. In each plot, fertilizer stopped on one half in 2014 and continued on the other. We compared 153 paired samples in 2015–2022 with unamended controls from the same field, fire group, and year. In nine plots with annual spring fire, stopped halves closed 48.1% of the richness gap to controls and 22.7% of the gap in relative biomass composition. The respective 95% plot bootstrap intervals were 41.0–55.1% and 15.1–30.7%, conditional on the references. All nine burned plots moved closer to control biomass composition. The mean distance decrease was 0.150 on a 0–1 scale, with a 95% interval of 0.100–0.202 when reference plots were also resampled. Species presence gave 46.2% composition recovery. Thus species gains coexisted with a larger remaining difference in biomass shares. The two unburned groups had smaller distance decreases, and the order of richness and composition fractions differed in field C. The main result was stable to distance choices, omission of individual controls or years, and data cleaning checks. The results apply to the measured strips, years, and reference states. Causal interpretation in field B requires exchangeable split sides, which we could not confirm from a separate design statement. All added nutrients stopped together. We provide raw inputs, source hashes, code, figures, and an independent calculation.

grassland · community recovery · nutrient cessation · Cedar Creek · plant richness · biomass composition · fire · open data · reproducibility