Item

Abstract

Soil organic carbon (SOC) is the largest carbon
pool in terrestrial ecosystems and may play a key role in biospheric
feedbacks with elevated atmospheric carbon dioxide
(CO2) in a warmer future world.We examined the simulation
results of seven terrestrial biome models when forced with
climate projections from four representative-concentrationpathways
(RCPs)-based atmospheric concentration scenarios.
The goal was to specify calculated uncertainty in global
SOC stock projections from global and regional perspectives
and give insight to the improvement of SOC-relevant processes
in biome models. SOC stocks among the biome models
varied from 1090 to 2650 PgC even in historical periods
(ca. 2000). In a higher forcing scenario (i.e., RCP8.5), inconsistent
estimates of impact on the total SOC (2099–2000)
were obtained from different biome model simulations, ranging
from a net sink of 347 Pg C to a net source of 122 Pg C.
In all models, the increasing atmospheric CO2 concentration
in the RCP8.5 scenario considerably contributed to carbon
accumulation in SOC. However, magnitudes varied from 93
to 264 PgC by the end of the 21st century across biome models.
Using the time-series data of total global SOC simulated
by each biome model, we analyzed the sensitivity of
the global SOC stock to global mean temperature and global
precipitation anomalies (1T and 1P respectively) in each
biome model using a state-space model. This analysis suggests
that 1T explained global SOC stock changes in most
models with a resolution of 1–2 C, and the magnitude of
global SOC decomposition from a 2 C rise ranged from almost
0 to 3.53 Pg Cyr−1 among the biome models. However,
1P had a negligible impact on change in the global SOC
changes. Spatial heterogeneity was evident and inconsistent
among the biome models, especially in boreal to arctic regions.
Our study reveals considerable climate uncertainty in
SOC decomposition responses to climate and CO2 change
among biome models. Further research is required to improve our ability to estimate biospheric feedbacks through both SOC-relevant and vegetation-relevant processes.