TL;DR
Human land use significantly drives the year-to-year fluctuations (interannual variability or IAV) of the global carbon cycle, accounting for 30-45% of this variability. This challenges previous assumptions that primarily attributed IAV to natural climate variations, highlighting the need to consider land management in climate change mitigation strategies.
What Problem Does This Paper Address?
The global carbon cycle is a complex system where the land absorbs a significant amount of atmospheric CO2. Understanding the year-to-year variations (IAV) in this land carbon sink is crucial for predicting how the land will respond to future climate change and for refining climate models. Previous research largely attributed this IAV to natural climate fluctuations, particularly in tropical regions and semi-arid vegetation. However, the increasing influence of human land use and management on ecosystems was not fully accounted for in these assessments. This paper aims to quantify the role of human land use in driving the IAV of the net land carbon balance (S<sub>net</sub>), which represents the overall carbon exchange between land and atmosphere.
Methods
The researchers employed an improved version of the ORCHIDEE Dynamic Global Vegetation Model (DGVM). This enhanced model is capable of distinguishing between managed and intact land at a sub-grid scale, a key advancement over previous models. They compared their results with a traditional bookkeeping model (HN2017) that relies on land-use reconstructions and fixed carbon densities. The ORCHIDEE model was forced with historical land-use change data and climate data. The study also involved a sensitivity run to assess the impact of shifting cultivation. Model validation was performed against various observational datasets, including deforestation area, forest biomass growth, and global biomass distribution.
Datasets
- Global Carbon Budget 2017 data: Used for atmospheric CO2 growth rate and ocean carbon sink estimates.
- LUH2 dataset: Provided historical land-use change data for agricultural area changes and shifting cultivation.
- CRUNCEP climate data: Used to force the ORCHIDEE model.
- FAOSTAT data: Used for harvested fuel wood volumes.
- Forest inventory data: Used for model validation of forest carbon sinks.
- Satellite-based tree cover loss data: Used for validation of deforestation.
- Satellite-based biomass datasets: Used for validation of forest biomass.
The paper does not introduce new benchmark datasets but rather utilizes and integrates existing ones for model development and validation.
Key Insights and Ideas
- Land use is a major driver of carbon cycle variability: The most significant finding is that human land use and management are not just passive influences but actively contribute substantially to the year-to-year fluctuations in the global carbon sink. This shifts the paradigm from solely climate-driven variability to a more integrated view.
- Managed land's dynamic role: Managed lands, such as agricultural areas and forests under management, exhibit a much higher interannual variability than previously understood. This is due to factors like agricultural practices and forest harvesting, which respond to both climate and socio-economic drivers.
- Rethinking intact land's contribution: When the role of land use is properly accounted for, the contribution of intact, undisturbed ecosystems to the IAV of the carbon sink is significantly reduced compared to prior assessments. This implies that our understanding of how natural ecosystems buffer climate change might need refinement.
- Model improvements are critical: The development of DGVMs that can resolve sub-grid scale differences between managed and intact lands is crucial for accurately attributing carbon fluxes and understanding their variability.
Key Results
- Managed land accounted for 30-45% of the interannual variability (IAV) of the net land carbon balance (S<sub>net</sub>) between 1959 and 2015.
- In contrast, previous assessments using bookkeeping models attributed only 5% of the IAV to land use.
- The contribution of intact land to the IAV of S<sub>net</sub> was reduced by more than half when land use was properly considered.
- The temperature sensitivity of the net land sink (Y<sub>LAND</sub>) was estimated at -3.4 Pg C yr<sup>-1</sup> K<sup>-1</sup>, consistent with previous estimates. However, the decomposition of this sensitivity showed that managed land plays a larger role than previously thought.
Limitations and Future Work
- Underestimation of E<sub>fire</sub> and ELUC IAV: The study acknowledges that the IAV of emissions from deforestation (E<sub>fire</sub>) and overall land-use change emissions (ELUC) might be underestimated due to smoothed input data for deforestation areas and static coefficients for biomass partitioning.
- Neglect of irrigation: The model did not simulate cropland irrigation, which can buffer the IAV of carbon fluxes in agricultural lands by relieving water stress during droughts. This omission might lead to an overestimation of ELUC's IAV.
- Uncertainty in secondary forest proportion: The exact contribution of managed land to Y<sub>LAND</sub> is subject to the assumed proportion of secondary forests, which was set to be consistent with the HN2017 model.
- Future work suggestions: The authors emphasize the need for further research on managed lands, including their carbon cycle dynamics and climate sensitivity. They also call for continued development of DGVMs to better separate managed and intact land components and to evaluate the contributions of anthropogenic versus natural factors to the land carbon balance. Future climate policy should focus on enhancing the resilience of forest sinks and minimizing legacy emissions from land use.
Summary updated