Relative decline in density of Northern Hemisphere tree species in warm and arid regions of their climate niches

Abstract

Although climate change is expected to drive tree species toward colder and wetter regions of their distribution, broadscale empirical evidence is lacking. One possibility is that past and present human activities in forests obscure or alter the effects of climate. Here, using data from more than two million monitored trees from 73 widely distributed species, we quantify changes in tree species density within their climatic niches across Northern Hemisphere forests. We observe a reduction in mean density across species, coupled with a tendency toward increasing tree size. However, the direction and magnitude of changes in density exhibit considerable variability between species, influenced by stand development that results from previous stand-level disturbances. Remarkably, when accounting for stand development, our findings show a significant change in density toward cold and wet climatic conditions for 43% of the species, compared to only 14% of species significantly changing their density toward warm and arid conditions in both early- and late-development stands. The observed changes in climate-driven density showed no clear association with species traits related to drought tolerance, recruitment and dispersal capacity, or resource use, nor with the temperature or aridity affiliation of the species, leaving the underlying mechanism uncertain. Forest conservation policies and associated management strategies might want to consider anticipated long-term species range shifts alongside the integration of contemporary within-distribution density changes.

Keywords: climate change; climatic sensitivity; forest dynamics; species density; stand development.

Fig. 1.

Decline in mean density of the 73 species analyzed across their climatic niches, with considerable variability in the direction of changes in density between species. We partitioned the climatic niche occupied by each species into nine climatic regions defined by the terciles in mean winter temperature and aridity across the species’ observed range. For each of the nine regions, we calculated mean changes in density (i.e., annual change in the number of stems per hectare for all trees larger than 12.7 cm in diameter at breast height between consecutive forest inventory censuses) considering all inventory plots sharing those climatic conditions across the period 1985 to 2019. Panels in the top row show results for two illustrative species with contrasting climatic niches (A) Quercus robur and (B) Pinus echinata. A full set of results for all species is found in SI Appendix, Figs. S2 and S3. (C) Histograms showing the mean annual change in the number of stems per hectare for all 73 species and for each of the nine climatic regions defined by temperature and aridity terciles. Numbers show the number of species with significant increases (blue) and decreases (red) in density (SI Appendix, Fig. S4), along with the overall mean change $\pm$ SE across species. Dashed lines indicate no change and negative values indicate a decrease in overall mean density. Note that legends in (A and B) are on different scales, and higher values of the aridity index imply less aridity.

Fig. 2.

Relationship between changes in density and changes in mean tree basal area between censuses. Mean tree basal area was calculated as the average basal area of living trees of each species in each plot, and changes in mean tree basal area as the annual change in mean tree basal area between censuses. Mean changes in density were calculated as the annual change in the number of stems per hectare between censuses. Both changes were quantified for all trees larger than 12.7 cm in diameter at breast height between consecutive forest inventory censuses across the period 1985 to 2019 and averaged for each species. Gray points represent mean changes in density and basal area for all analyzed species across all the plots where each species was present, while red points highlight example species covering climatic gradients in Europe and the United States. The yellow points indicate the overall mean of all species. Early, mid, and late panels include only sites that have low, mid, or high stand basal areas compared to the highest stand basal area values found in the climate space in which the stand was located (Materials and Methods).

Fig. 3.

Climate-driven changes in species density adjusting for stand development. (A) Predicted changes in species density (i.e., annual change in the number of stems per hectare for all trees larger than 12.7 cm in diameter at breast height between consecutive forest inventory censuses across the period 1985 to 2019) by species-level models when setting mean winter temperature and aridity index in cold and wet (blue), median climate (black), and warm and arid (red) conditions within each species’ climatic niche, and setting stand development in actual, early, and late stand development values (Materials and Methods). Points indicate mean changes in species density and intervals 50% equitailed CI, with positive values indicating increases and negative values decreases in density over time. The selection of species for representation was made by covering climatic gradients of Europe and the United States. A full set of results for all species is found in SI Appendix, Figs. S5 and S6. (B) Histograms showing the mean climatic sensitivity of each species in actual, early, and late stand development conditions. Species’ climatic sensitivity was calculated as the mean difference in the annual change in the number of stems per hectare expected when setting climatic conditions as cold and wet or warm and arid, according to each species’ climatic niche. Positive values indicate that in cold and wet climatic conditions, species would gain more individuals, or lose fewer individuals, than in warm and arid conditions. Negative values indicate that in cold and wet conditions, species would gain fewer individuals, or lose more individuals, than in warm and arid conditions. Numbers on each histogram panel show the number of species with significant increases (blue) and decreases (red) in density (SI Appendix, Table S2), along with the overall mean climatic sensitivity $\pm$ SE across species. Dashed lines indicate no difference between cold and wet and warm and arid climatic conditions within species.