# Decadal extreme drought reduces alpine subsoil carbon stocks

**Authors:** Ronglei Zhou, Jinsong Wang, Quancheng Wang, Ning Liu, Chenglong Ye, Jingjing Shi, Mengjie Liu, Zhangwei Gao, Houkun Chu, Zhenrui Zhang, Bin Niu, Song Wang, Ruiyang Zhang, Dashuan Tian, Shuli Niu

PMC · DOI: 10.1073/pnas.2517468123 · Proceedings of the National Academy of Sciences of the United States of America · 2026-02-20

## TL;DR

Extreme drought over ten years reduces carbon storage in alpine grassland subsoil, challenging assumptions about soil resilience to climate change.

## Contribution

First quantitative assessment of subsoil carbon loss under decade-long extreme drought in alpine grasslands.

## Key findings

- Extreme drought caused 27-37% subsoil carbon loss in 20-40 cm layers, driven by reduced mineral-associated organic carbon.
- Mild to moderate drought had no significant effect on soil carbon at any depth.
- Subsoil carbon losses linked to disrupted soil-microbe-mineral interactions and reduced microbial activity.

## Abstract

Grasslands store more than 30% of global soil organic C, which is highly susceptible to drought. However, no study has quantitatively assessed how soil-profile C stocks respond to long-term, especially extreme drought. In a 10-year field experiment, we show that decade-long extreme drought decreases subsoil C stocks by 27 to 37% relative to controls, driven by reductions in mineral-associated organic C. These persistent C losses in subsoil are mainly governed by abiotic constraints on microbial functioning and the limited supply of microbial precursors. Our findings challenge the prevailing assumption that subsoil C reservoirs are resistant to drought and underscore the imperative to incorporate refined subsoil processes into Earth System Models to accurately capture soil C–climate feedback under intensifying drought.

Grassland soil organic carbon (SOC) stocks are increasingly vulnerable to intensified drought. Yet, how long-term, especially extreme, drought affects these C reserves across soil profiles remains unresolved. Here, we present the response of SOC stocks at various depths (0 to 60 cm) using an experiment involving a 10-year drought gradient (P, 1/2 P, 1/4 P, and 1/12 P, where P is ambient precipitation) in an alpine grassland. We found that extreme drought (1/12 P) caused significant SOC losses in the subsoil, whereas mild to moderate droughts had no significant effects on SOC at any depth. Under extreme drought, SOC stocks declined by 27% in the 20 to 30 cm layer and by 37% in the 30 to 40 cm layer, while no significant changes were observed in the topsoil (0 to 20 cm) and the deeper layer (40 to 60 cm). Subsoil C losses were primarily driven by reductions in mineral-associated organic C (MAOC) rather than particulate organic C. Multiyear extreme drought disrupted soil–microbe–mineral interactions, as indicated by reductions in soil nitrogen availability, microbial biomass, carbon use efficiency, and mineral-binding agents. These changes collectively undermined the formation and stabilization of MAOC. Our findings suggest that prolonged extreme drought can weaken subsoil C storage and stability, highlighting the need to incorporate subsoil processes into Earth System Models to better predict soil C–climate feedback under future drought scenarios.

## Full-text entities

- **Diseases:** ESMs (MESH:D020721), Drought (MESH:C536747)
- **Chemicals:** GalN (MESH:D005688), sodium hexametaphosphate (MESH:C009285), GluN (MESH:D005944), sulfuric acid (MESH:C033158), lignin (MESH:D008031), CO2 (MESH:D002245), chloroform (MESH:D002725), aluminum (MESH:D000535), Amino sugars (MESH:D000606), MAOC (-), MurA (MESH:D009112), CuO (MESH:C030973), ammonium oxalate (MESH:D019815), dithionite (MESH:D004227), HCl (MESH:D006851), potassium dichromate (MESH:D011192), water (MESH:D014867), phenol (MESH:D019800), Fe (MESH:D007501), Feo (MESH:C034236), Oxalate (MESH:D010070), N (MESH:D009584), PNAS (MESH:D020135), C (MESH:D002244), NaCl (MESH:D012965), P (MESH:D010758)

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC12933107/full.md

## Figures

5 figures with captions in the complete paper: https://tomesphere.com/paper/PMC12933107/full.md

## References

69 references — full list in the complete paper: https://tomesphere.com/paper/PMC12933107/full.md

---
Source: https://tomesphere.com/paper/PMC12933107