Global forestation and deforestation affect remote climate via adjusted atmosphere and ocean circulation

Large-scale tree planting is considered a crucial land-based strategy to mitigate climate change, complementing fossil fuel emission reductions. However, the impacts of both forestation (reforestation and afforestation) and deforestation extend beyond carbon cycle effects, influencing local and global climates through alterations in surface energy balance (albedo, evaporative fraction, surface roughness). These biogeophysical effects vary depending on location and can counter the temperature changes resulting from changes in land carbon storage. Globally, albedo changes tend to dominate the biogeophysical temperature response, potentially diminishing the overall mitigating effect of large-scale forestation due to warming from reduced albedo. Beyond direct thermodynamic effects, substantial changes in forest cover can alter atmospheric dynamics, a key driver of regional precipitation and temperature. Previous studies, often employing simplified ocean models, have indicated that surface temperature changes from large-scale forestation or deforestation, especially in the extratropics, can shift the intertropical convergence zone (ITCZ) and alter atmospheric heat transport. However, the interaction between ocean circulation and the atmospheric response to land-use changes is not well understood. This study addresses the knowledge gap by investigating the complex interplay between atmospheric and oceanic circulations in response to global forestation and deforestation using a fully coupled atmosphere-ocean model. This is especially pertinent given recent proposals for global-scale afforestation and ongoing large-scale tree-planting initiatives.

Existing research highlights the multifaceted impacts of land-use change on climate. Studies have shown that forestation and deforestation affect the surface energy balance, leading to localized and global temperature changes. The impact of these changes on atmospheric dynamics, specifically the shift in ITCZ and changes in cross-equatorial atmospheric heat transport, have been explored. However, many of these investigations have utilized simplified ocean models (slab ocean or prescribed sea-surface temperatures), overlooking the potential influence of ocean circulation on the climate system's response to land-use changes. The Land-Use Model Intercomparison Project (LUMIP) within CMIP6 has shown that some models exhibit substantial near-surface temperature changes over oceans in response to global deforestation. The significant role of ocean thermohaline circulation in responding to global land-use change has also been highlighted, indicating the need for coupled atmosphere-ocean models to comprehensively study these interactions. This study addresses these shortcomings by using a fully coupled climate model to investigate the impact of large-scale forestation and deforestation on both atmospheric and ocean circulations.

This study employed the Community Earth System Model (CESM) version 2.1.2, a fully coupled global climate model encompassing atmospheric, land, ocean, and sea-ice components. The model was run at approximately 1° horizontal resolution with preindustrial climate forcing held constant throughout the simulations. Three simulations were conducted: (1) a control simulation using preindustrial land cover (~30% forest cover); (2) a forestation scenario (80% forest cover), achieved by transforming grasslands, croplands, shrubs, and urban areas into forests; and (3) a deforestation scenario (0% forest cover), where forested areas were converted to grasslands. All other parameters, including CO2 mixing ratios, were kept identical across simulations. The simulations were run for 300 years, with the first 200 years in the control simulation excluded to ensure equilibrium. Analyses focused on years 50-300 to minimize the influence of initial model adjustments to the land-surface perturbations. Statistical significance of differences between scenarios and the control were assessed using a two-sided Wilcoxon rank-sum test with a Benjamini-Hochberg correction for multiple comparisons. A simplified expression was used to translate global mean top-of-atmosphere radiative forcing to changes in atmospheric CO2 concentrations for comparative analysis. Jet streams were identified using a method based on vertical average of total wind speed between 400 and 100 hPa. Deep and shallow jets were differentiated based on upper-tropospheric wind shear.

Global-scale forestation resulted in a 0.5°C global mean warming, most pronounced over northern extratropical land. This warming led to a decrease in meridional heat transport in the Northern Hemisphere, primarily due to a 10-25% reduction in poleward ocean heat transport. The Atlantic Meridional Overturing Circulation (AMOC) weakened by 22%. Warming of high-latitude land surfaces resulted in weaker and poleward-shifted weather systems, attenuating and displacing the midlatitude westerlies. Deforestation caused opposite, but larger-amplitude changes: -1.6°C global mean cooling, a stronger AMOC (49% increase), and an accelerated extratropical jet stream. The cooling was particularly strong at high latitudes due to strong snow-ice-albedo feedback. Both forestation and deforestation scenarios showed substantial changes in regional precipitation and cloud cover, often in regions remote from the altered land cover. For instance, forestation caused a decrease in annual mean precipitation in the Euro-Mediterranean region by more than 5%. Changes in near-surface winds were observed, with decreased speeds over most land areas in the forestation scenario and increased speeds in the deforestation scenario. Global mean cloud cover decreased in the forestation scenario and increased in the deforestation scenario. The initial radiative forcings were approximately 0.70 W m⁻² (equivalent to about 40 ppm CO2) for forestation and -0.85 W m⁻² (-42 ppm CO2) for deforestation. The Hadley cell showed a weakening in the Northern Hemisphere during boreal winter in the forestation scenario and strengthening in the deforestation scenario. The ITCZ shifted southward in the deforestation scenario. Changes in the extratropical atmospheric circulation were observed, with a poleward shift of the jet stream in the forestation scenario and equatorward shift in the deforestation scenario. Changes in the eddy momentum flux convergence were also observed in both scenarios, reflecting changes in the intensity and position of the jet stream.

The findings highlight the profound impact of global-scale land-use changes on atmospheric and ocean circulations, extending beyond previously understood thermodynamic effects. The significant response of the AMOC, previously overlooked due to the lack of coupled models, underscores the importance of considering ocean dynamics when assessing land-use change impacts. The reduced ocean surface heat fluxes following forestation, resulting from higher surface temperatures and reduced wind speeds, likely contributes to the AMOC slowdown by inhibiting deep water formation. The results demonstrate that even regionally limited forestation could induce remote circulation changes. This study's results necessitate a holistic assessment of biogeophysical consequences, including remote effects, in planning future large-scale forestation initiatives. The complex interactions between land-surface changes, atmospheric dynamics, and ocean circulation emphasize the interconnected nature of the Earth's climate system.

This study demonstrates the significant influence of global-scale forestation and deforestation on atmospheric and ocean circulations, resulting in substantial changes in remote climate patterns. The strong response of the AMOC highlights the importance of coupled atmosphere-ocean models in assessing land-use change impacts. Future research should explore these mechanisms under more complex model configurations, considering various scales of forestation and incorporating biogeochemical effects. Although the study does not advocate against forestation, it emphasizes the need to consider remote circulation effects when designing large-scale forestation projects to maximize benefits and minimize unintended consequences.

The study's idealized scenarios, focusing on global-scale changes in forest cover with constant CO2, might not fully capture the complexity of real-world land-use change. The sensitivity of the results to model type and land-surface parameterizations requires further investigation. The model simulations, while running for 300 years, may not have reached full equilibrium, potentially influencing the long-term impacts of land-cover changes. Future work should explore these aspects to refine our understanding of the complex feedback mechanisms.