TL;DR
The Congo Basin, a critical but understudied carbon sink, is experiencing a significant expansion of its swamp forests, reversing previous drying trends. This expansion, driven by increased rainfall, has led to reduced CO2-equivalent emissions despite a rise in methane, strengthening the region's role as a net carbon sink and highlighting the importance of accurate wetland monitoring for global climate models.
What Problem Does This Paper Address?
Accurately estimating global greenhouse gas (GHG) emissions is crucial for understanding and mitigating climate change. However, a major source of uncertainty lies in the emissions from wetlands, particularly tropical ones. Wetlands are significant sources of methane (CH4), a potent GHG, and their desiccation can lead to the release of substantial amounts of carbon dioxide (CO2). The Congo Basin, home to vast peatlands and swamp forests, is a key area for carbon storage but has historically lacked sufficient data for precise GHG emission estimates. Previous studies suggested a long-term drying trend in the region, which could have implications for carbon release. This paper aims to address this knowledge gap by quantifying changes in swamp forest extent, productivity, and GHG emissions in the Congo Basin between 2007 and 2024, and to assess the impact of these changes on the region's role as a carbon sink.
Methods
The researchers employed a multi-faceted approach combining satellite Earth observation, hydrological modeling, and existing emissions datasets.
- Swamp Forest Extent Mapping: High-resolution L-band Synthetic Aperture Radar (SAR) imagery from the ALOS PALSAR and PALSAR-2 satellites (25m pixel size) was used to map swamp forest extent annually from 2007 to 2024. The double-bounce scattering mechanism in SAR data, which is enhanced by water inundation, allowed for the differentiation of flooded swamp forests from non-flooded terra firme forests. A random forest (RF) classification model was trained using ground-truth GPS coordinates to identify permanently flooded swamp forests.
- Forest Productivity Assessment: Monthly global vegetation indices from the MODIS product (1km resolution) were used to quantify the Enhanced Vegetation Index (EVI), a proxy for photosynthetic activity and forest productivity. EVI data for the April-May-June period (chosen for its lower cloud cover) from 2007 to 2024 were analyzed.
- GHG Emissions Estimation: To estimate flood depths and water table levels (WTLs), the annual swamp forest maps were combined with the Height Above Nearest Drainage (HAND) model. This allowed for the prediction of WTLs based on flood extent and terrain characteristics. GHG emissions (CO2, CH4, N2O, and CO2-equivalent) were then estimated by applying emission factors from a global tropical wetland GHG emissions database (Zou et al., 2022) to the mapped swamp forest extents and estimated WTLs.
Datasets
- Satellite Imagery:
- ALOS PALSAR and PALSAR-2 (L-band SAR imagery, 25m pixel size) for swamp forest extent mapping.
- MODIS monthly global vegetation indices 1km product (MOD13A3 V061) for forest productivity (EVI) assessment.
- Ancillary Data:
- Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) for rainfall data.
- Satellite altimeter-derived annual river level height anomalies for 52 virtual river stations across the Congo Basin.
- Global Inundation Extent from Multi-Satellites (GIEMS-2) dataset for comparison with low spatial resolution global wetland datasets.
- Height Above Nearest Drainage (HAND) model for hydrological modeling.
- Emissions Data:
- A synthesis of 504 global GHG net flux experimental studies from Zou et al. (2022), specifically focusing on tropical wetland emission factors.
- Ground Truth Data:
- Ground-truth GPS coordinates for permanently flooded peat-forming swamp forests and terra firme forests (Crezee et al., 2022).
The study highlights that existing low spatial resolution global wetland datasets (like GIEMS-2) significantly underestimate the extent of tropical swamp forests, suggesting a need for higher-resolution, ecosystem-specific data.
Key Insights and Ideas
- Reversal of Drying Trend: The most striking finding is the observed expansion of swamp forests, directly contradicting previous reports of a long-term drying trend in the Congo Basin. This suggests a significant shift in regional hydrology.
- Hydrological Sensitivity: The Congo Basin's swamp forests are highly sensitive to regional climatic fluctuations, with increased rainfall driving the observed expansion. This underscores the importance of accurate climate data for understanding these ecosystems.
- Underestimation in Global Models: The study demonstrates that current low-resolution global wetland datasets used in Earth System Models (ESMs) significantly underestimate the extent of tropical swamp forests, leading to potential underestimations of CH4 emissions.
- Net Carbon Sink Strengthening: Despite increased CH4 emissions due to wetland expansion, the overall CO2-equivalent emissions have decreased, indicating that the Congo Basin is becoming an even stronger net carbon sink. This is primarily due to reduced CO2 emissions from stabilized soil carbon stores.
- Productivity Boost: The wetting trend has also led to increased forest productivity (EVI) in both swamp and terra firme forests, suggesting enhanced carbon sequestration potential.
Key Results
- Swamp Forest Expansion: Swamp forest extent increased from approximately 195,345 km² in 2007 to 222,467 km² in 2024, representing an average annual increase of 1528 km² (95% CI; 265–2791 km²).
- Increased Rainfall and River Levels: The swamp forest expansion coincided with a significant increase in basin-wide annual average rainfall and a rise in average river levels across the basin between 2007 and 2023.
- Forest Productivity Increase: Mean EVI for all Congo Basin forests showed a significant increasing trend (Slope=0.001, p-value <0.001) over the study period. Both swamp forests and terra firme forests increased in productivity.
- GHG Emission Changes:
- CH4 emissions increased by approximately 80,000 tonnes per year (95% CI; 30,000–120,000).
- CO2 emissions decreased by 4.5 million tonnes per year (95% CI; -6.7 to -2.2).
- N2O emissions declined by 570 tonnes per year (95% CI; -950 to -190).
- Overall CO2-equivalent emissions (100-year horizon) decreased by 2 million tonnes per year (95% CI; -2.94 to -1.12) between 2007 and 2024.
Limitations and Future Work
- Water Table Level Estimation: While the study used a sophisticated method to estimate water table levels, these were derived from flood extent and terrain data, and direct in-situ measurements would provide further validation. The estimation of below-ground WTLs was based on the minimum HAND value from non-flooded pixels within the maximum flood extent, which might lead to underestimation of emissions in areas where water tables recede beyond this.
- Methane-Climate Feedbacks: The paper acknowledges that methane's global warming potential is significantly higher over a 20-year horizon than a 100-year horizon, which would amplify its contribution to CO2-equivalent emissions and potentially alter the conclusion about the net reduction in warming potential.
- Ecological Feedbacks: The study notes that prolonged inundation can lead to ecological trade-offs, such as shifts in community composition towards lower-biomass taxa, which are not fully captured by remote-sensing metrics and could influence carbon stocks.
- Future Work: The authors emphasize the need for in situ experimental studies to better understand the influence of increasing forest productivity on swamp forest GHG fluxes and to account for ecological feedbacks. Upscaling these high-resolution, ecosystem-specific methods to other remote tropical regions is crucial for improving global GHG modeling and carbon budget assessments. Further research into the persistence of the wetting trend versus the long-term drying trend is also warranted.
Summary updated