Temperature and Rainfall Patterns Constrain the Multidimensional Rewilding of Global Forests

Abstract The long‐term contribution of global forest restoration to support multiple dimensions of biodiversity and ecosystem function remains largely illusive across contrasting climates and forest types. This hampers the capacity to predict the future of forest rewilding under changing global climates. Here, 120 studies are synthesized across five continents, and it is found that forest restoration promotes multiple dimensions of biodiversity and ecosystem function such as soil fertility, plant biomass, microbial habitat, and carbon sequestration across contrasting climates and forest types. Based on global relationship between stand age and soil organic carbon stock, planting 350 million hectares of forest under the UN Bonn Challenge can sequester >30 Gt soil C in the surface 20 cm over the next century. However, these findings also indicate that predicted increases in temperature and reductions in precipitation can constrain the positive effects of forest rewilding on biodiversity and ecosystem function. Further, important tradeoffs are found in very old forests, with considerable disconnection between biodiversity and ecosystem function. Together, these findings provide evidence of the importance of the multidimensional rewilding of forests, suggesting that on‐going climatic changes may dampen the expectations of the positive effects of forest restoration on biodiversity and ecosystem function.


Supporting information Temperature and rainfall patterns constrain the multidimensional rewilding of global forests
Content: Table S1. Attributes used to derive the seven ecosystem functions and biodiversity.  Table S3. Literature included in our dataset for the meta-analysis. Table S4. Contributions of multiple ecosystem attributes to PCA 1 of the principal component analysis. Figure S1. Article selection process using Preferred Reporting Items for Systematic Reviews (PRISMA) guidelines. Figure S2. Global distribution of the selected stand forest. The blue and red circles indicate which experiment provided selected angiosperm and conifers stands, respectively. Based green map represents the distribution of global evergreen, deciduous and mixed forests. The observations of Angiosperm and conifers forests were 141 and 65, respectively. Figure S3. Probability density of the forest restoration effect on plant biodiversity, microbial biodiversity, plant biomass, microbial habitat, soil carbon, SOM decomposition and soil fertility. Results are based on 21 studies for plant biodiversity, 28 studies for microbial biodiversity, 58 studies for plant biomass, 29 studies for microbial habitat, 100 studies for soil carbon, 18 studies for SOM decomposition and 92 studies for soil fertility. Figure S4. Estimates (±95% CI) of the log response ratio for plant biodiversity, microbial biodiversity, plant biomass, microbial habitat, soil carbon, soil fertility and SOM decomposition of different climate type (A, dryland and mesic) and tree functional type (B, evergreen, deciduous and mixed forests). The vertical line was drawn at LnRR=0. Number values for each bar indicate the sample size. The error bars indicated the 95% confidence interval (CI). If the CI did not overlap with zero, a response was considered to be significant. Figure S5. Relationships between stand age with the response ratios (LnRR) of plant biodiversity (A), microbial biodiversity (B), plant biomass (C), microbial habitat (D), soil carbon (E), soil fertility (F) and SOM decomposition (G) for dryland and mesic forests. Figure S6. Relationships between stand age with the response ratios (LnRR) of plant biodiversity (A), microbial biodiversity (B), plant biomass (C), microbial habitat (D), soil carbon (E), soil fertility (F) and SOM decomposition (G) for arid, cold, and tropical/temperate forests.

Figure S7
Partial correlation analyses reveals that climate is significantly correlated with the restoration of multiple ecosystem attributes (lnRR) even when controlling for changes across different age ranges (lnRR stand age).

Figure S1
Article selection process using Preferred Reporting Items for Systematic Reviews (PRISMA) guidelines.

Figure S2
Global distribution of the selected stand forest. The blue and red circles indicate which experiment provided selected angiosperm and conifers stands, respectively. Based green map represents the distribution of global evergreen, deciduous and mixed forests. The observations of Angiosperm and conifers forests were 141 and 65, respectively.

Figure S3
Probability density of the forest restoration effect on plant biodiversity, microbial biodiversity, plant biomass, microbial habitat, soil carbon, SOM decomposition and soil fertility. Results are based on 180 observations for plant biodiversity, 251 observations for microbial biodiversity, 594 observations for plant biomass, 309 observations for microbial habitat, 549 observations for soil carbon, 268 observations for SOM decomposition and 922 observations for soil fertility.

Figure S4
Estimates (±95% CI) of the log response ratio for plant biodiversity, microbial biodiversity, plant biomass, microbial habitat, soil carbon, SOM decomposition, and soil fertility of different climate type (A, dryland and mesic) and tree functional type (B, evergreen, deciduous and mixed forest). The vertical line was drawn at LnRR=0. Number values for each bar indicate the sample size. The error bars indicated the 95% confidence interval (CI). If the CI did not overlap with zero, a response was considered to be significant.

Figure S5
Relationships between stand age with the response ratios (LnRR) of plant biodiversity (A), microbial biodiversity (B), plant biomass (C), microbial habitat (D), soil carbon (E), soil fertility (F) and SOM decomposition (G) for dryland and mesic forests.

Figure S6
Relationships between stand age with the response ratios (LnRR) of plant biodiversity (A), microbial biodiversity (B), plant biomass (C), microbial habitat (D), soil carbon (E), soil fertility (F) and SOM decomposition (G) for arid, cold, and tropical/temperate forests.

Figure S7
Partial correlation analyses reveals that climate is significantly correlated with the restoration of multiple ecosystem attributes (lnRR) even when controlling for changes across different age ranges (lnRR stand age).