Extended Data Fig. 5: Changes in the probability density of species thermal suitability and the proportion of thermally mismatched species in different ecosystems.
From: Contrasting thermophilization among forests, grasslands and alpine summits
a, Changes in the probability density of thermal suitability under local climate (Tclim) inferred from plot-level floristic temperature (Tplant) distributions. Boxplots showing changes in probability density of Tplant between baseline survey and resurvey (Δ = resurvey – baseline survey), with the centre line indicating the median, the box representing the 25th and 75th percentiles, and the whiskers showing the minimum and maximum values. These results show that the probability density at Tclim significantly and marginally decreased from baseline to resurvey in forest and grassland ecosystems, respectively, suggesting increasing climatic mismatch. b, Changes in the proportion of thermally mismatched species (Δ = resurvey – baseline survey). Mismatch is defined as background macroclimate falling outside of the 2.5–97.5% range of a species’ thermal distribution. These results suggest that the proportion of mismatched species significantly increased in forests, but not in grasslands or on alpine summits. In both panels a and b, each point represents a plot (n = 4372, 1209, and 457 for forests, grasslands, and alpine summits, respectively), and the black circles with error bars indicate means with 95% confidence intervals estimated using linear mixed-effects models with “site” as a random-effects factor, and the p values indicate whether change differs significantly from zero (linear mixed-effects models; two-sided).
