In many cases, modelled future pathways include an overshoot period where temperatures first exceed the climate target and are later reduced by the use of CDR at scales exceeding residual emissions, so that net-negative values are achieved. Due to delayed and indecisive climate action, overshoot pathways are becoming increasingly likely and have recently gained significant research interest, but uncertainties remain with regards to Earth system responses to temperature overshoot.
Even though a wide range of novel land- and ocean-based CDR methods are being considered, their technological readiness levels currently remain low, and further research is needed. Therefore, virtually all of CDR being applied to date consists of conventional land-based CDR methods, and in particular Afforestation/Reforestation (AR) and sustainable forest management. AR includes the conversion of land into forest through tree planting efforts. The new forest sequesters more carbon compared to the prior non-forested ecosystem, and therefore more CO2 is taken up from the atmosphere. AR can include the reforestation of historically deforested land, or the conversion of land that has not historically been a forest, but rather a different type of natural ecosystem (afforestation).
Given the multitude of co-benefits and the carbon sequestration potential of tree planting, many countries around the world have decided to include AR as a core tool in their portfolio of net-zero and long-term climate pledges. The countries' pledges amount to a total of 490 Mha by 2060 and are expected to increase as more and more countries submit their pledges. However, converting land to a forest can be a challenging task, since it can trigger biodiversity loss, and it can compete with agricultural land use, which might threaten food security.
A team around Yiannis Moustakis investigated the mitigation potential of ambitious AR deployment in the range of country pledges under a temperature overshoot scenario. Their results - entitled “Temperature overshoot responses to ambitious forestation in an Earth System Model”- were recently published in Nature Communications.
For their study, they developed a spatio-temporally explicit AR scenario in the range of country pledges, taking into account techno-economic considerations, future land availability, food demand, and biodiversity protection and utilizing >1,200 modelled future socioeconomic scenarios produced by so-called Integrated Assessment Models. In this scenario, AR reaches 595 Mha by 2060 and 935 Mha by 2100, which corresponds to an area slightly smaller than 3 times the area of India. In this scenario, heavily managed grazing land over deforested areas is prioritized for tree planting, while natural grass- and shrub-lands are not converted. This will protect biodiversity as much as possible while making the areas needed for AR available.
Video1: Visualization of the simulated afforestation/reforestation to mitigate climate change (animation produced by DKRZ). The globe one the left shows the forestation pattern compared to 2015 described by the constructed scenario, which is constrained by techno-economic considerations, restoration potential and biodiversity maps, reaching 595 Mha by 2060, and 935 Mha by 2100. The resulting total global forest area change is shown in the lower left.
The team ran this scenario from 2015 to 2100 with the state-of-the-art Max Planck Institute’s Earth System Model (MPI-ESM), which features a fully interactive carbon, energy, and water cycle between the land, the ocean, and the atmosphere globally, and can thus represent all the necessary feedbacks and climatic effects that emerge following the application of large-scale AR. Apart from the developed AR scenario, the MPI-ESM simulations are driven by fossil fuel emissions following the Shared Socioeconomic Pathway (SSP) 5-3.4. This is an emission scenario that features high emissions until the mid-century, followed by a rapid decarbonization of the economy, reaching net-negative during the late century. Apart from the AR scenario, also a reference (REF) scenario was run, featuring no land-use or land-cover change since 2015, which serves as the counterfactual scenario.
Given the inherent internal variability of the Earth system, and the subtle dynamics of temperature overshoot trajectories, the scientists have undertaken the heavy computational effort of running 10 ensemble members for each of the REF and AR scenarios. All simulations have been run on DKRZ high-performance computing system Levante. In total 1,700 model-years, which translates to ~4,600 node-hours, were simulated and produced about 10 Terabytes of output data. Such an amount of resources deployment is not typical for a modelling study, but is necessary here, in order to allow for the footprint of AR to emerge, thus making statistical inference more robust.
Overall, the scientists show that ambitious AR in the range of country pledges can reduce peak global temperature by 0.08°C, and end-of-century temperature by 0.2°C, while the duration of the overshoot period is reduced by 13 years, compared to the REF scenario. The footprint of AR on global temperatures becomes already evident in 2052, with AR having reached 495 Mha at that point. By the end of the century, the achieved temperature mitigation is the result of a reduction of atmospheric carbon by 283 Gt CO2 compared to the REF scenario. This is however lower than the carbon sequestration over land, which reaches 382 Gt CO2. This difference occurs due to the fact that, as atmospheric CO2 levels get lower compared to the REF scenario, the diffusion of CO2 into the ocean also gets lower, and therefore under the AR scenario the ocean takes up less carbon overall, thus slightly offsetting the land uptake. This has crucial implications for monitoring, reporting, and verifying CDR, since it implies that the actual atmospheric CO2 reduction is different than the amount of sequestration that one measures on the field.
At the same time, it is assumed that the planting of trees in some regions could change surface energy and heat fluxes in such a way that local warming can occur. This is mainly the result of changes in the reflectivity (albedo) and the roughness of the land surface following AR. Such a warming could likely increase adaptation needs locally, and thus should be treated with caution. However, the study shows that by the end of the century, the strong reduction in atmospheric CO2 (compared to REF) facilitates a widespread cooling across the globe, which can at least compensate for any potential local warming that would have occurred. The team further showed that tree planting leads to an overall wetter hydroclimate over forested regions, with increased precipitation, evapotranspiration, relative humidity, and cloudiness levels.
Despite the capacity of large-scale AR in sequestering carbon and reducing temperatures, the authors make a cautionary note on its potential socioeconomic side-effects, and call for careful and holistic planning of future AR. Conversion of agricultural land into forest can deprive people of their livelihoods, and even physically displace populations. It might as well possibly disturb local food networks, thus threatening food security. At the same time, most modelled future pathways disproportionally allocate AR across the Global South. This does not only reasonably raise questions about fair burden-sharing, but also suggests that large-scale AR would heavily rely on countries whose governance and institutional capacities might not be strong enough to facilitate a transparent, inclusive and successful large-scale CDR application. Weak rule of law and insecure land tenure in such countries could also threaten the permanence of the new forest, and therefore of the sequestered carbon. Given the above, the authors highlight that AR is not a silver bullet, and rather make an urgent call for rapid emissions reduction.
References:
Moustakis, Y., Nützel, T., Wey, HW. et al. Temperature overshoot responses to ambitious forestation in an Earth System Model. Nat Commun 15, 8235 (2024). https://doi.org/10.1038/s41467-024-52508-x
Author:
Yiannis Moustakis, Department of Geography at Ludwig Maximilians University, Munich