Extended Data Fig. 8: Structural equation models showing the effects of tree diversity, alternative climatic factors and soil conditions on decadal changes in soil C and N stocks.
From: Tree diversity increases decadal forest soil carbon and nitrogen accrual
a,b, Path diagrams of factors influencing changes in soil C and N stocks in the organic horizon (n = 361). b,d Path diagrams of factors influencing changes in soil C and N stocks in the mineral horizon (n = 245). Numbers adjacent to arrows are standardized path coefficients, analogous to relative regression weights. Solid and dashed arrows represent positive and negative relationships, respectively. Different colours represent different types of explanatory variable (see Fig. 1). Only significant pathways are shown (P < 0.05). The goodness-of-fit statistics for panels a–d are: GFI = 0.988, SRMR = 0.032, P = 0.249; GFI = 0.991, SRMR = 0.029, P = 0.477; GFI = 0.987, SRMR = 0.033, P = 0.367; and GFI = 0.986, SRMR = 0.038, P = 0.391, respectively, indicating close model-data fit. ΔSoil COrganic and ΔSoil NOrganic represent decadal changes in soil C and N stocks of the organic soil horizons, respectively. ΔSoil CMineral and ΔSoil NMineral represent decadal changes in soil C and N stocks of the mineral soil horizons, respectively. GDD, mean annual growing degree-days; AGP, mean annual precipitation at the growing season; FDis, functional diversity; ΔThickness, decadal changes in soil organic horizon thickness; CWMPC2, community-weighted mean of trait values; Horizon thickness, initial organic horizon thickness. Higher CWMPC2 values indicate traits associated with lower tree maximum height (see Extended Data Fig. 1).
