Earth system model simulations by Ge et al. (Nature Geoscience, 2023) showed that reducing anthropogenic aerosol emissions amplifies the local surface cooling produced by afforestation, with additional cooling of up to 0.4 degrees when emissions return to preindustrial levels and the strongest effects projected for China under strict air-pollution control. However, that study could not separate the two coupled pathways driving the amplification - the increase in surface solar radiation as the aerosol haze lifts versus the stimulated growth of vegetation (higher leaf area index and stronger evapotranspiration) - because its subgrid forest-minus-openland comparison of precomputed CMIP6 scenario output leaves these processes fully coupled. The problem asks for a quantitative, causal partition of the amplified local cooling into the direct-radiation and vegetation-mediated contributions under an explicitly declared decomposition convention, by any approach able to deliver such a partition for the forest-minus-openland contrast (for example designed coupled vegetation-chemistry-climate experiments, decomposition of paired CMIP6/AerChemMIP simulations, or observationally constrained estimates), with the vegetation-mediated share interpreted under currently incomplete representations of vegetation processes and with an assessment of whether regional atmospheric feedbacks such as cloud, circulation, and biogenic volatile organic compound induced aerosol changes modify the local signal. The precipitation response to emission reductions and the emission pathways themselves remain background context for projecting the amplification in a policy-relevant way. A 2026 follow-up by the same group quantified the local biophysical feedback of general vegetation responses to aerosol removal and a 2025 study quantified reforestation effects on biogenic secondary organic aerosol, but the attribution of the afforestation cooling amplification into its radiation and vegetation-growth pathways, and the test of whether regional feedbacks modify it, remain undone.

Background

Afforestation (planting forests on land that was not previously forested) changes local climate through biophysical processes in addition to removing atmospheric CO2: forests are darker than grassland or cropland (lower surface albedo, a warming effect), but they transfer more water to the atmosphere through evapotranspiration (evaporation plus plant transpiration, a cooling effect) and mix air more efficiently through their rough canopies. Which of these competing effects dominates depends strongly on the background climate, for example snow cover and water availability. Leaf area index (LAI, the one-sided leaf area per unit ground area) is the standard measure of vegetation density used in climate models, and the local temperature response to a unit change in LAI is commonly written as a sensitivity ∂T/∂LAI\partial T/\partial \mathrm{LAI}. Anthropogenic aerosols - suspended particulate pollution from fossil-fuel and biomass burning - scatter and absorb sunlight, so heavy pollution reduces the total solar radiation reaching the surface while increasing its diffuse fraction. As air-quality policy lowers aerosol emissions, more total solar radiation reaches vegetation, and the regional climate (temperature, precipitation) shifts as well; both changes can stimulate or suppress vegetation growth and therefore feed back on the local climate effect of forests. Ge et al. (Nature Geoscience, 2023) asked how this changing aerosol background interacts with the local effects of afforestation. They used the subgrid forest-minus-openland method: within each grid cell of scenario simulations from CMIP6 (the sixth phase of the Coupled Model Intercomparison Project, which coordinates standard climate-model experiments across modeling centers) and its aerosol-chemistry component AerChemMIP, they compared near-surface temperature and the surface energy balance between forest and open-land fractions, so the afforestation effect is inferred from the model's own juxtaposition of the two land covers rather than from paired simulations with and without deforestation. Their result: lower anthropogenic emissions modulate afforestation's local effect through the surface energy balance, and if current emissions were reduced to preindustrial levels, afforestation would produce additional local cooling of up to 0.4 degrees, most strongly in China under strict future air-pollution control - meaning enhanced afforestation cooling could partially counteract the warming that aerosol removal itself causes. The open issues this problem addresses are the following. First, the amplification is produced jointly by the increased total solar radiation (more energy to drive evapotranspiration) and by aerosol-decline-stimulated vegetation growth (higher LAI), and these pathways are fully coupled in the scenario archives that were analyzed, so their relative contributions cannot be separated without an explicit attribution convention or experiment design. Second, the vegetation response itself is uncertain because CMIP6-type models incompletely represent processes such as stomatal responses to reduced ozone pollution, nutrient limitation, and subgrid ecological constraints, so the vegetation-mediated part of the amplification may be underestimated; the biogeochemical side (the additional CO2 uptake associated with vegetation change) was quantified by the source itself and again in the 2026 follow-up, and is context here rather than an open deliverable. Third, the subgrid method contains no large-scale atmospheric feedback: widespread afforestation can itself alter regional cloud cover, circulation, and biogenic volatile organic compound (BVOC) emissions that form aerosol and ozone, so assessing whether such mesoscale-to-regional feedbacks modify the local signal requires going beyond the subgrid comparison, in which afforestation is not allowed to change atmospheric composition and cloudiness. Fourth, model agreement on the precipitation response to emission reductions is limited - the source's own nine-model analysis (Extended Data Fig. 5) marks where eight of the nine models agree on the sign of the change in downward shortwave radiation, 2-meter temperature and precipitation - and precipitation is a key driver of LAI and evaporative cooling. Fifth, CMIP6 scenario pathways may understate future regional emission reductions (for example in China; regionally targeted scenario work appears among the source's references), so regionally targeted scenarios are relevant for policy-relevant projections.

Problem Statement

Explain how and why lower anthropogenic aerosol pollution amplifies the local cooling effect of afforestation, and quantitatively attribute that amplification. Concretely: (1) causally partition the amplified local cooling of afforestation under reduced aerosol emissions - the local surface temperature response of the forest-minus-openland contrast together with its surface energy balance terms (net shortwave radiation and albedo, sensible heat flux, latent heat flux, and leaf area index) - into the contribution from the increased total surface solar radiation (and its changed diffuse fraction) as aerosol loading falls and the contribution from aerosol-decline-stimulated changes in vegetation state (higher leaf area index and the associated change in the temperature sensitivity to LAI, dT/dLAIdT/d\mathrm{LAI} (equivalently written ∂T/∂LAI\partial T/\partial \mathrm{LAI}), whose albedo/snow- and evapotranspiration-mediated components must be distinguished), reporting the relative magnitudes and signs of the two contributions under an explicitly declared decomposition convention that states how the non-additive interaction (residual) term is treated - for example the convention used by Ge et al. (2026), dT=(dT/dLAI)⋅dLAI+LAI⋅d(dT/dLAI)+residualdT = (dT/d\mathrm{LAI}) \cdot d\mathrm{LAI} + \mathrm{LAI} \cdot d(dT/d\mathrm{LAI}) + \mathrm{residual}, which decomposes the vegetation-mediated part and its residual; the radiation contribution must likewise be carried as a separate term of the declared convention. The numerical magnitude of the residual must be reported, and the signs and relative magnitudes of the two contributions must be shown to be robust to the choice of convention, tested by repeating the partition under at least two conventions. Any approach able to deliver a causal partition for the forest-minus-openland contrast is admissible, including designed factorial coupled vegetation-chemistry-climate experiments, decomposition of paired CMIP6/AerChemMIP simulations under a stated convention, or observationally constrained estimates. The partition must carry two source-based conditions: (2) CMIP6-type models incompletely represent vegetation processes such as stomatal responses to reduced ozone, nutrient limitation, and subgrid ecological constraints, so the vegetation-mediated contribution may be underestimated and the reported magnitude should carry an explicit uncertainty statement for that contribution; (3) the subgrid forest-minus-openland method (which compares forest and open-land fractions within the same model grid cell and therefore omits large-scale atmospheric feedback) yields a local result; the partition must therefore include an explicit assessment, going beyond the subgrid comparison and allowing regional atmospheric composition and cloudiness to respond, of whether regional cloud cover, circulation, or biogenic volatile organic compound driven aerosol and ozone formation materially modify the local signal, since the local magnitude may be conservative if they do. Relevant background for projecting the amplification under clean-air and climate-mitigation policy: it is conditional on the emission pathway considered (idealized removal to preindustrial levels; SSP3-7.0 versus its lowNTCF variant for strict future air-pollution control); model agreement on the precipitation response to emission reductions, a key driver of leaf area index and evaporative cooling, is limited; and CMIP6 scenario pathways may understate future regional emission reductions (for example in China), for which more regionally targeted scenarios exist. These projection factors remain background context and are not a required deliverable.

References

  1. Ge, J., Huang, X., Zan, B., Qiu, B., Cao, Y. & Guo, W. Local surface cooling from afforestation amplified by lower aerosol pollution. Nature Geoscience 16, 781-788 (2023). DOI: 10.1038/s41561-023-01251-x. https://doi.org/10.1038/s41561-023-01251-x
  2. Ge, J., Yue, X., Mu, M., Miao, X., Huang, X., Qiu, B. & Guo, W. Local biophysical climate feedback from vegetation responses to lower aerosol pollution. npj Climate and Atmospheric Science 9, 38 (2026). DOI: 10.1038/s41612-025-01310-7. https://doi.org/10.1038/s41612-025-01310-7
  3. Zhu, J., Penner, J. E., Liu, H., Guo, Q., Liu, Y., Deng, J., Zhao, X., Liu, C.-Q. & Fu, P. Reforestation-induced aerosol cooling effects divergently modulated by various types of biogeophysical feedback. National Science Review 12, nwaf323 (2025). DOI: 10.1093/nsr/nwaf323. https://doi.org/10.1093/nsr/nwaf323
  4. Luo, H., Quaas, J. & Han, Y. Decreased cloud cover partially offsets the cooling effects of surface albedo change due to deforestation. Nature Communications 15, 7345 (2024). DOI: 10.1038/s41467-024-51783-y. https://doi.org/10.1038/s41467-024-51783-y
  5. Li, Y., Ge, J., Wu, H., Tang, R., Cheng, Y., Liu, X., Wang, Y., Zhao, W., Gao, C., Duan, S.-B., Shi, Q., Leng, P., Zhao, E., Yan, G., Song, X., Zhou, C. & Li, Z.-L. Amplified local cooling effect of forestation in warming Europe. Nature Communications 16, 8412 (2025). DOI: 10.1038/s41467-025-63556-2. https://doi.org/10.1038/s41467-025-63556-2
  6. Li, Y., Li, Z.-L., Wu, H., Liu, X., Lian, X., Si, M., Li, J., Zhou, C., Tang, R., Duan, S., Zhao, W., Leng, P., Song, X., Shi, Q., Zhao, E. & Gao, C. Observed different impacts of potential tree restoration on local surface and air temperature. Nature Communications 16, 2335 (2025). DOI: 10.1038/s41467-025-57606-y. https://doi.org/10.1038/s41467-025-57606-y
  7. Liu, S., Zhou, Y., Zhao, L., Mills, G., Li, X., Zhang, H., Seto, K. C. & Gong, P. Divergent surface urban heat island responses to background aerosol mitigation. Nature Cities 3, 654-664 (2026). DOI: 10.1038/s44284-026-00463-7. https://doi.org/10.1038/s44284-026-00463-7
  8. Li, L., Wang, Z., Qiu, B., Huang, X., Guo, W., Miao, X., Zhao, S., Chen, J. & Ding, A. Terrestrial Carbon Sink and Clean Air Co-Benefits From China's Carbon Neutrality Policy. Earth's Future 12, e2024EF004631 (2024). DOI: 10.1029/2024EF004631. https://doi.org/10.1029/2024EF004631
  9. Chen, L. Cooler forests in clean air. Nature Geoscience 16, 758-759 (2023). DOI: 10.1038/s41561-023-01253-9. https://doi.org/10.1038/s41561-023-01253-9