Three-Dimensional Modeling of Soil Organic Carbon Stocks in Forest Ecosystems of Northeastern China Under Future Climate Warming Scenarios

Shuai Wang, Shouyuan Bian, Zicheng Wang, Zijiao Yang, Chen Li, Xingyu Zhang, Di Shi, Hongbin LiuForests 2025 · DOI

TL;DR

This study developed a novel 3D modeling approach for soil organic carbon (SOC) stocks in Northeast China's forests, accounting for depth-specific variations under climate warming. The research highlights that rising temperatures will likely lead to significant SOC losses, particularly in topsoil, with topography and vegetation being key drivers of these changes.

What Problem Does This Paper Address?

Soil organic carbon (SOC) is a critical component of the global carbon cycle, acting as a massive terrestrial carbon sink. Forests, especially in regions like Northeast China, play a significant role in this. However, most research on SOC stocks has focused on the topsoil (0-30 cm), leaving a substantial knowledge gap regarding how carbon is stored and how it will respond in deeper soil layers under future climate change. Understanding these depth-dependent dynamics is crucial for accurate climate modeling and developing effective carbon management strategies. This paper aims to fill this gap by providing a detailed, three-dimensional understanding of SOC stocks and their projected changes in Northeast China's forest ecosystems under various warming scenarios.

Methods

The study employed a multi-faceted methodology to achieve its goals:

  • Random Forest (RF) Modeling: This machine learning algorithm was used to predict SOC stocks. RF is known for its ability to handle complex, high-dimensional data and maintain good predictive accuracy. The researchers optimized key hyperparameters like ntree (number of trees) and nodesize (minimum observations in terminal nodes) for better performance.
  • Equal-Area Spline Profile Function: This technique was used to model the vertical distribution of soil properties (like SOC) within soil profiles. It allows for a continuous representation of soil characteristics across different depths, enhancing the accuracy of depth-specific predictions.
  • Space-for-Time Substitution: This approach was used to project future SOC stock changes under climate warming. It leverages existing relationships between SOC and environmental variables, assuming that current spatial variations can represent future temporal changes under different climate conditions. This method allows for simulating future scenarios without direct long-term observational data for those specific future conditions.

Datasets

The study utilized the following datasets:

  • Soil Profile Data: 63 complete soil profiles (up to 1m depth or bedrock) were collected from forest soils in Northeast China during 2017-2018. These provided the ground truth for SOC content and other soil properties.
  • Environmental Predictors: A suite of 12 environmental variables were used, encompassing:
  • Climatic Variables: Mean Annual Temperature (MAT) and Mean Annual Precipitation (MAP) at 1 km resolution, downscaled to 90m. Future warming scenarios (+1.5°C, +2°C, +4°C) were generated.
  • Topographic Variables: Derived from a 90m Digital Elevation Model (DEM), including elevation (ELE), slope gradient (SG), slope aspect (SA), profile curvature (PC), catchment area (CA), and topographic wetness index (TWI).
  • Soil Property Variables: Soil texture data (sand, silt, clay percentages) at 1km resolution, derived from national soil survey data.
  • Biological Variables: Normalized Difference Vegetation Index (NDVI) at 250m resolution, resampled to 90m.

Key Insights and Ideas

  • Depth Matters: The study strongly emphasizes that SOC distribution is not uniform with depth. Different environmental factors control SOC at different soil layers, highlighting the inadequacy of topsoil-only studies.
  • Topography Dominates: Surprisingly, topographic variables, particularly elevation, were found to be the most significant drivers of SOC spatial variability across all depths, contributing about 50% of the total influence. This suggests that landscape features play a crucial role in how carbon is stored and distributed.
  • Vegetation's Deep Influence: NDVI, a proxy for vegetation health and productivity, showed a surprisingly strong influence on SOC even in deeper soil layers. This might be due to the significant input of organic matter from vegetation in these forest ecosystems and the slow decomposition rates in cold climates.
  • Warming's Double-Edged Sword: While warming generally leads to SOC loss due to increased decomposition, the study hints at a complex response in permafrost regions. Moderate warming (1.5°C) might lead to less loss than higher warming levels, potentially due to partial permafrost thaw activating carbon but not fully accelerating decomposition, while higher warming might lead to more significant losses but also potentially some stabilization due to anaerobic conditions in thawed permafrost.

Key Results

  • Model Performance: The RF model demonstrated robust predictive capability across all soil depth intervals, with good accuracy metrics (high R² and LCCC, low MAE and RMSE). Accuracy generally decreased with increasing depth.
  • Spatial Patterns: SOC stocks were concentrated in central regions and lower in northern areas. Maximum accumulation was observed in the 5-15 cm layer, with decreasing amounts in deeper layers.
  • Climate Warming Impact: Under all simulated warming scenarios (1.5°C, 2°C, and 4°C), SOC stocks showed a consistent declining trend across all depths, with the most pronounced reductions under the 4°C scenario. The top 30 cm of soil experienced significant losses, accounting for approximately half of the total SOC loss in the top meter.
  • Dominant Drivers: MAT and NDVI were identified as key environmental drivers controlling SOC variability. Topography was the primary driver overall, followed by climate and biological factors. Soil properties showed less relative importance, possibly due to their covariation with topographic factors.

Limitations and Future Work

  • Limited Sampling Density: The study acknowledges that the number of sampling sites (n=63) might not fully capture the complex SOC spatial variability across Northeast China, especially in permafrost regions. This could lead to underestimation of variability.
  • Data Integration Challenges: Integrating environmental datasets from various sources with different resolutions and formats may have introduced some loss of fine-scale information.
  • Unmodeled Factors: The model did not fully incorporate factors like microbial activity, detailed soil texture variations, land-use history, or groundwater influences, which can affect SOC stabilization.
  • Simplified Climate Scenarios: The study only considered temperature increases and did not account for changes in precipitation or soil water availability, which are crucial for SOC decomposition.
  • Future Research Directions: The authors suggest future work should include:
  • More extensive soil sampling.
  • Enhanced environmental datasets, incorporating more detailed soil properties and microbial data.
  • More sophisticated climate modeling approaches that include precipitation and other factors.
  • Incorporating vegetation feedbacks and forest succession dynamics.
  • Long-term observational data to validate the space-for-time substitution approach.
  • Refining model parameterization to better represent site-specific soil processes.

Summary updated

Abstract

Understanding the detailed spatiotemporal variations in soil organic carbon (SOC) stocks is essential for assessing soil carbon sequestration potential. However, most existing studies predominantly focus on topsoil SOC stocks, leaving significant knowledge gaps regarding critical zones, depth-dependent variations, and key influencing factors associated with deeper SOC stock dynamics. This study adopted a comprehensive methodology that integrates random forest modeling, equal-area soil profile analysis, and space-for-time substitution to predict depth-specific SOC stock dynamics under climate warming in Northeast China’s forest ecosystems. By combining these techniques, the approach effectively addresses existing research limitations and provides robust projections of soil carbon changes across various depth intervals. The analysis utilized 63 comprehensive soil profiles and 12 environmental predictors encompassing climatic, topographic, biological, and soil property variables. The model’s predictive accuracy was assessed using 10-fold cross-validation with four evaluation metrics: MAE, RMSE, R2, and LCCC, ensuring comprehensive performance evaluation. Validation results demonstrated the model’s robust predictive capability across all soil layers, achieving high accuracy with minimized MAE and RMSE values while maintaining elevated R2 and LCCC scores. Three-dimensional spatial projections revealed distinct SOC distribution patterns, with higher stocks concentrated in central regions and lower stocks prevalent in northern areas. Under simulated warming conditions (1.5 °C, 2 °C, and 4 °C increases), both topsoil (0–30 cm) and deep-layer (100 cm) SOC stocks exhibited consistent declining trends, with the most pronounced reductions observed under the 4 °C warming scenario. Additionally, the study identified mean annual temperature (MAT) and normalized difference vegetation index (NDVI) as dominant environmental drivers controlling three-dimensional SOC spatial variability. These findings underscore the importance of depth-resolved SOC stock assessments and suggest that precise three-dimensional mapping of SOC distribution under various climate change projections can inform more effective land management strategies, ultimately enhancing regional soil carbon storage capacity in forest ecosystems.
BibTeX
@inproceedings{wang2025three,
  author = {Shuai Wang and Shouyuan Bian and Zicheng Wang and Zijiao Yang and Chen Li and Xingyu Zhang and Di Shi and Hongbin Liu},
  title = {{Three-Dimensional Modeling of Soil Organic Carbon Stocks in Forest Ecosystems of Northeastern China Under Future Climate Warming Scenarios}},
  booktitle = {Forests},
  year = {2025},
  doi = {10.3390/f16081209},
}
Figure