Forest management in southern China generates short term extensive carbon sequestration

Xiaowei Tong, Martin Brandt, Yuemin Yue, Philippe Ciais, Martin Rudbeck Jepsen, Josep Peñuelas, Jean‐Pierre Wigneron, Xiangming Xiao, Xiao‐Peng Song, Stéphanie Horion, Kjeld Rasmussen, Sassan Saatchi, Lei Fan, Kelin Wang, Bing Zhang, Zhengchao Chen, Yuhang Wang, Xiaojun Li, Rasmus FensholtNature Communications 2020 · DOI

TL;DR

Southern China's intensive forest management policies have successfully created a significant carbon sink, increasing aboveground carbon stocks by 0.11 Pg C per year between 2002 and 2017. This sequestration offsets a substantial portion (33%) of regional fossil CO2 emissions, primarily driven by newly established forests, but challenges like forest growth saturation and soil moisture depletion threaten its long-term sustainability.

What Problem Does This Paper Address?

The global challenge of mitigating climate change necessitates understanding and quantifying the role of land-use practices, particularly forest management, in carbon sequestration. While afforestation and reforestation are known carbon sinks, their scalability and potential trade-offs (e.g., with food production and water resources) are debated. Furthermore, empirical, large-scale data on the effectiveness of diverse forest management strategies, beyond simple afforestation, is scarce. Southern China, with its intensive land-use policies aimed at both economic development and climate mitigation, presents a critical case study to investigate these complex interactions. This paper aims to quantify the carbon sequestration impact of various land-use changes in southern China and assess the sustainability of this carbon sink service.

Methods

The study employs a multi-faceted approach combining satellite observations and statistical modeling. Key methods include:

  • Forest Probability Mapping: Using daily MODIS BRDF-corrected imagery from 2000-2017, the researchers generated annual forest/non-forest probability maps. This was achieved by training a Random Forest classifier with high-resolution GF-1 satellite imagery (2m resolution) and SRTM elevation data. This approach allows for the detection of changes in forest cover over time.
  • Land Use Typology: The annual forest probability maps were further analyzed using the Landsat-based Detection of Trends in Disturbance and Recovery (LandTrendr) algorithm. This algorithm partitions time-series data into segments to identify disturbances and recovery patterns, allowing for the classification of land into eight distinct types based on management intensity, forest continuity, and dynamics (e.g., dense forest, persistent forest, non-forest, recovery, afforestation, deforestation, and two types of rotation).
  • Carbon Density Estimation: A global benchmark map of woody vegetation C density (2015) was used to train a gradient boosted decision tree model. This model, utilizing MODIS reflectance bands, NDVI, EVI2, MCD43A3 albedo, and SRTM elevation, was applied to the annual forest probability maps to derive maps of annual aboveground carbon density at a 500m resolution.
  • Soil Moisture Analysis: SMOS (Soil Moisture and Ocean Salinity) satellite data, available at a 25x25 km resolution from 2010-2017, was used to assess soil moisture trends.
  • Intercomparison and Validation: The derived forest probability maps and carbon density estimates were validated against independent datasets, including Landsat-based tree cover maps and AVHRR data, to ensure accuracy and reliability. SMOS L-band vegetation optical depth (L-VOD) data was used to cross-validate carbon density dynamics.

Datasets

  • MODIS MCD43A4 (BRDF-corrected imagery): Used for generating annual forest probability maps (2000-2017).
  • GF-1 satellite imagery (2m resolution): Used for training the Random Forest classifier for forest probability mapping (2013-2017).
  • SRTM Digital Elevation Model: Used in conjunction with satellite imagery for forest probability and carbon density modeling.
  • Landsat-based tree cover maps: Used for validation of forest cover estimates (2010).
  • AVHRR data: Used for long-term forest dynamics comparison (1982-2016).
  • SMOS (Soil Moisture and Ocean Salinity) data: Used for soil moisture trend analysis (2010-2017).
  • SMOS L-VOD data: Used for independent validation of carbon density dynamics.
  • Global benchmark map of woody vegetation C density (2015): Used for training the carbon density estimation model.
  • Chinese Forestry Statistical Yearbooks: Provided timber production data for province-level analysis.
  • Provincial CO2 emission data: Used for comparing carbon sequestration to regional emissions.

Key Insights and Ideas

  • Diverse Management, Diverse Impact: The study highlights that not all forest management is equal. Newly established forests (afforestation and recovery) are the primary drivers of carbon sequestration, contributing significantly more than existing or passively managed forests.
  • The "Managed Forest" Advantage: Actively managed forests, including those with rotation cycles, contribute substantially to carbon sequestration, even with increased timber harvesting. This suggests that sustainable forest management can balance economic output with climate benefits.
  • Non-Forests Matter Too: Persistent non-forest areas, often agricultural lands, also contribute to the overall carbon sink, indicating that optimizing land use across different categories is crucial.
  • The Soil Moisture Trade-off: A critical finding is the significant decline in soil moisture in certain areas, particularly those with intensive tree plantations. This raises concerns about water availability and the long-term ecological sustainability of these practices.
  • Short-Term Gains, Long-Term Questions: While the current carbon sequestration is substantial, the study points to potential saturation of forest growth and the need for careful consideration of the full life cycle of harvested timber to ensure continued climate benefits.

Key Results

  • Overall Carbon Sink: Southern China's land use changes increased standing aboveground carbon stocks by 0.11 ± 0.05 Pg C y⁻¹ between 2002 and 2017.
  • Contribution Breakdown:
    • Newly established forests (afforestation/recovery): 32%
    • Existing forests: 24%
    • Harvested forest areas (regrowth): 16%
    • Non-forest areas: 28%
  • Emission Offset: The carbon sequestration is equivalent to 33% of regional fossil CO2 emissions during the last 6 years (2012-2017).
  • Timber Harvest: Timber harvest tripled during the study period.
  • Soil Moisture Decline: Significant soil moisture decline was observed in 8% of the area.

Limitations and Future Work

  • Spatial Resolution: The study acknowledges that the 500x500m resolution of MODIS data can lead to mixed signals within a grid cell, potentially over- or underestimating the true extent of certain land-use changes. SMOS data's 25x25 km resolution further limits detailed analysis of specific land-use types.
  • Carbon Dynamics Complexity: The carbon sink calculations are primarily based on aboveground biomass. The study notes that changes in litter, coarse woody debris, and soil carbon are not fully accounted for, and the life-cycle analysis of harvested wood products is crucial for a complete picture.
  • Forest Growth Saturation: The paper suggests that forest growth may eventually saturate, limiting future increases in carbon sequestration from existing plantations.
  • Soil-Water Depletion: The observed soil moisture decline poses a significant challenge to the long-term sustainability of the carbon sink service, especially in drought-prone regions.
  • Future Work: The authors emphasize the need for continued monitoring of both carbon stocks and soil moisture dynamics. They also call for more refined analyses that incorporate rotation cycles, plantation types, and market scenarios to better evaluate the potential of managed forests for climate change mitigation. Understanding the full life-cycle of harvested timber is also highlighted as a critical area for future research.

Summary updated

Abstract

Land use policies have turned southern China into one of the most intensively managed forest regions in the world, with actions maximizing forest cover on soils with marginal agricultural potential while concurrently increasing livelihoods and mitigating climate change. Based on satellite observations, here we show that diverse land use changes in southern China have increased standing aboveground carbon stocks by 0.11 ± 0.05 Pg C y<sup>-1</sup> during 2002-2017. Most of this regional carbon sink was contributed by newly established forests (32%), while forests already existing contributed 24%. Forest growth in harvested forest areas contributed 16% and non-forest areas contributed 28% to the carbon sink, while timber harvest was tripled. Soil moisture declined significantly in 8% of the area. We demonstrate that land management in southern China has been removing an amount of carbon equivalent to 33% of regional fossil CO<sub>2</sub> emissions during the last 6 years, but forest growth saturation, land competition for food production and soil-water depletion challenge the longevity of this carbon sink service.
BibTeX
@inproceedings{tong2020forest,
  author = {Xiaowei Tong and Martin Brandt and Yuemin Yue and Philippe Ciais and Martin Rudbeck Jepsen and Josep Peñuelas and Jean‐Pierre Wigneron and Xiangming Xiao and Xiao‐Peng Song and Stéphanie Horion and Kjeld Rasmussen and Sassan Saatchi and Lei Fan and Kelin Wang and Bing Zhang and Zhengchao Chen and Yuhang Wang and Xiaojun Li and Rasmus Fensholt},
  title = {{Forest management in southern China generates short term extensive carbon sequestration}},
  booktitle = {Nature Communications},
  year = {2020},
  doi = {10.1038/s41467-019-13798-8},
}
Figure