Beyond carbon: Does afforestation/reforestation mitigate or trigger future climate extremes?

The large-scale deployment of carbon dioxide removal (CDR) methods will be required to achieve the climate targets set in the Paris Agreement, with afforestation/reforestation (AR) currently being the most widely implemented one. While AR is recognized for its carbon sequestration potential, its biogeophysical effects remain insufficiently understood. This is particularly the case for climate extremes, since their quantification and robust statistical inference against internal variability require large simulation ensembles, which are typically lacking in modelling studies. Climate extremes such as heatwaves, floods and droughts have far-reaching consequences for ecosystems and human society, and understanding whether large-scale AR could unintentionally intensify climate extremes, or may support mitigation efforts by dampening them, is thus critical for assessing its viability as a mitigation measure. To investigate that, it is crucial to disentangle how the biogeophysically-induced effects on climate extremes are affected by the associated AR-induced global cooling and whether such effects differ across various emissions pathways.

Here, using an unprecedented multi-member ensemble of emission- and concentration-driven simulations with a fully coupled Earth System Model (MPI-ESM), we investigate the influence of large-scale AR on precipitation and heat extremes across different scenarios (SSP1-2.6, SSP5-3.4os, SSP3-7.0, SSP5-8.5). The AR scenario used includes an ambitious deployment in the range of country pledges, reaching 935 Mha by 2100 globally. Our setup featuring 120 simulations in total enables a robust quantification of changes in climatic extremes due to AR across spatial and temporal scales and their uncertainty boundaries. We assess the responses of eight climate extreme indicators for 2091-2100.

Our results reveal spatially heterogeneous but overall dampening effects of AR on end-of-century heat extreme indicators. On average, annual maximum temperature decreases by 0.24 °C [0.17 °C, 0.19 °C, 0.17 °C] under SSP5-3.4os [SSP1-2.6, SSP3-7.0, SSP5-8.5]. The number of extreme heat days decreases by 15.1 % [11.3 %, 7.4 %, 4.6 %], annual maximum wet-bulb temperature by 0.10 °C [0.07 °C, 0.08 °C, 0.07 °C] and warm spell duration by 20.3 % [13.0 %, 11.7 %, 7.6 %], respectively. Any biogeophysical warming visible in concentration-driven simulations seems to be largely offset by AR-induced cooling in emission-driven ones, although regional exceptions exist. While the emission scenarios influence the magnitude of differences, they do not alter the overall signal. The dampening of heat extremes is especially evident in major population exposure hotspots, such as Central Africa and Eastern Asia, suggesting that AR can provide co-benefits for mitigating heat-related risks.

AR effects on precipitation extremes are less consistent and exhibit strong regional and scenario-specific variability, with most changes being within the boundaries of internal variability. AR-induced intensification in some regions (e.g. in the tropics) balances out reduction elsewhere, showing a mixed global signal. Within re/afforested regions, precipitation extremes tend to be less intense in high-emission scenarios (SSP3-7.0 and SSP5-8.5). 

Overall, our study offers a robust, policy-relevant assessment of the impacts of large-scale AR on future climate extremes. Our results suggest that large-scale AR application not only contributes to climate change mitigation but also offers adaptation benefits, particularly by reducing heat extremes, offsetting any biogeophysically-induced warming.