A High-Resolution Gridded Dataset of Vertically Resolved Fine Root Biomass Across China

Abstract. Fine root plays a critical role in ecosystem biogeochemical cycles, serving as the primary organs for water and nutrient uptake and exhibiting rapid turnover. However, a spatially explicit dataset of fine root biomass (FRB) across soil profiles and vegetation types remains lacking. To address this knowledge gap, we compiled 2,311 in-situ measurements of woody FRB and 3,580 measurements of herbaceous FRB from peer-reviewed studies across China. Using these observations, we developed random forest models to estimate surface-layer (0–20 cm) FRB and the vertical attenuation coefficient (k) of FRB for woody and herbaceous ecosystems, respectively. By integrating the predicted surface-layer FRB and vertical distribution functions, we finally generated a gridded dataset of FRB across 0–2 m soil depth (at 20 cm intervals) at 1 km spatial resolution for all woody and herbaceous ecosystems in China for 2005–2025. Model evaluation showed that our models explain approximately 86 % and 81 % of the observed variations in surface FRB across woody and herbaceous ecosystems, respectively, and reasonablely capture the interannual dynamics of observed FRB. Model predictions reveal substantial spatial and vertical variations in both woody and herbaceous FRB across China. Woody FRB density was generally higher in southern China than in central and northern China, whereas herbaceous FRB density was highest on the Qinghai–Tibet Plateau. Woody plants exhibit deeper root distributions in arid regions, whereas herbaceous species show consistently surface-concentrated rooting patterns across climate zones. The environmental controls of FRB also differed between vegetation types, with woody FRB being more strongly influenced by root traits and nutrient availability, whereas herbaceous FRB was more closely associated with climatic conditions. This dataset provides an observation-constrained, spatially explicit benchmark for quantifying belowground carbon stocks and dynamics, and it is archived and publicly available at https://doi.org/10.6084/m9.figshare.32353257 (Zheng et al., 2026).