Hedging our bet on forest permanence

doi:10.21203/rs.3.rs-2922314/v1

Download PDF

Physical Sciences - Article

Hedging our bet on forest permanence

https://doi.org/10.21203/rs.3.rs-2922314/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Forests currently absorb a quarter of anthropogenic carbon emissions. Moreover, establishing additional forests is presently the most cost-effective and scalable option for carbon removal services. Thus, mitigation pledges and scenarios aiming for the Paris Agreement climate goals became heavily dependent on maintaining the existing forest carbon sink and expanding it further. However, a mounting body of evidence questions the permanence of the carbon stored in forests under the pressure of both climate change and direct human intervention. Here we show that rapid decarbonization can secure mitigation goals despite potential carbon loss of forest disturbances. Yet, the projected CO₂ price to stay below 1.5°C of warming is 71$/tCO₂ higher (+22%) compared to a world without increasing disturbance trends. Likewise, it entails a 20% (-1.2GtCO₂/yr) reduction in allowed yearly emissions and an expansion of 69 million hectares (+17%) of dedicated mitigation land. Responding late to the disturbance, rather than taking preemptive action, doubles the projected cost and effort until mid-century in multiple key sectors. Our results underline the urgent need to safeguard forests for the economic feasibility of the Paris Agreement climate goals. Deferring fossil fuel phase-out leaves us ill-prepared if trends of increasing forest disturbance rates continue.

Earth and environmental sciences/Climate sciences/Climate change/Climate-change mitigation

Scientific community and society/Forestry

Earth and environmental sciences/Environmental social sciences/Climate-change mitigation

The terrestrial carbon sink is critical to mitigate climate change by removing more than 13 billion tons of CO₂ from the atmosphere yearly, equivalent to a third of all anthropogenic emissions^26,27. Any conceived climate stabilization target relies on the continuity of this vast carbon removal service. Forests accumulate most of the carbon sequestered this way, constituting two-thirds or 8.8 billion tons of CO₂ per year¹. This carbon sink in forests acts unassisted in the sense that human actions only indirectly modify it. E.g., by unintentional tree regrowth on abandoned land or by the fertilizing effect of the added CO₂ emitted to the atmosphere^28,29. However, additional carbon dioxide removal (CDR) is needed on top of these natural, unassisted sinks to keep warming well below 2°C^30,31 despite hard-to-mitigate or even perpetual emissions sources like concrete manufacturing, aviation, and food production^5–9. Establishing new forests is the foremost method to provide low-cost and large-scale CDR with present-day technology^2–4.

However, forests are vulnerable to changes in environmental conditions and land-use, rendering the permanence of the carbon stored in them uncertain^32–34. Our exposure to potential forest carbon release is threefold. We rely on 1) the permanence of existing forest carbon stocks, 2) the continued removal of a major fraction of anthropogenic emissions from the atmosphere, and 3) on vastly increasing the global forest carbon stock by large-scale afforestation/reforestation (A/R) to provide us with the much-needed CDR services. Thus, our plan to stay below critical warming thresholds becomes a bet on the world’s forests to stay healthy, undisturbed, and productive. Despite that, a growing body of evidence puts the safety of this bet into question.

Recent research reveals that the models on which today’s mitigation pathways are based might be too optimistic about the future forest carbon storage in three key aspects. 1) Models neglect disturbances such as windfall, pests, and disease³⁵ that make up the bulk of current forest damage and show the most pronounced increasing trends in many regions^22,25,36. Already modelled disturbances, such as fire, fail to approximate historic trends³⁷. 2) Models might overestimate the positive influence of warmer temperatures and CO₂ fertilization³⁸. Findings of Dow et al. in 2022 revealed that warmer spring temperatures do not lead to more carbon being accumulated in temperate forests as projected by models¹⁵. Further, models project the Amazon rainforest to act as a major carbon sink continually³⁹. However, more recent studies provide direct evidence of phosphorous limitation potentially cutting biomass carbon growth in half^16,17. 3) Deforestation and degradation are likely underestimated in scenario-building models that currently assume perfect efficiency in replacing forest land with agriculture. A study in 2022 by Pendrill et al. found that only 45% to 65% of deforested land for agriculture was used within years after clearance due to mismanagement, land speculation, and uncontrolled fire clearings²⁰.

Future disturbance levels are still highly uncertain, especially given the wide variety of biotic and abiotic sources modified by climate change accompanying degradation and clearing. Nevertheless, historic disturbance rates and their progression can provide insight, albeit limited to current climatic conditions. Historic disturbance rates in Europe and Canada are estimated between 0.26 %/yr and 0.57 %/yr^36,45–47. However, only the highest (0.57%/yr) assessment included direct human interventions⁴⁶. Rates in the Amazon could already be much higher as untracked degradation alone was found to surpass even deforestation by now affecting an average of 0.46% of forest area per year over the 22-year study period⁴⁸. Finding a marked increase in historic disturbance rates over the last decades^{24,25,36,45,49–52}, a future increase under accelerating climate change and resource demand-induced pressure is deemed likely^{21,23,24,45,53–56}. Increases in disturbance rates were found globally except for the Congo basin, where no significant increase was observed²⁴. Moreover, recently observed spikes in single-agent disturbances could indicate non-linear behaviour in response to direct and indirect human-induced stress^{23,25,36,51,53}. Bark beetle disturbance rates doubled in Europe over the last 20 years³⁶. Further, bark beetle damage in the Czech Republic surged from a long-term range of 0.2%/yr - 1.4%/yr to up to 5.4%/yr in most recent years²⁵. Similarly, disturbance related to declining forest health overtook all other types of disturbances in the US since the late 1990s, increasing dramatically from 0.07%/yr in 1985 to 2.82%/yr in 2001⁵¹.

We assess mitigation policy reaching climate targets despite looming impermanent forest carbon storage. In a set of 42 scenarios, we explore modified mitigation pathways along three dimensions (i) the degree of annual forest disturbance (n = 4 & control), (ii) different long-term climate targets (n = 2), and (iii) varying policy regulations (n = 5) within the integrated assessment model REMIND-MAgPIE^43,44(Figure 1b). In the following, we contrast the foresighted policy taking immediate action with the myopic response, both aiming at the 1.5°C-consistent budget under the annual disturbance rate of four per thousand trees per year (Figure 1a/c). However, results from all 42 scenarios are made available in the extended data figures.

Reaching the 1.5°C target despite an annual disturbance rate of four in a thousand trees is achieved by measures that further lower emissions over all CO₂-emitting sectors (Figures 2 & 3), expanding renewable energy supply (Figure 3), ramping up CDR deployment (Figures 2, 4 & 5), and boosting the price on carbon (Figures 2 & 6). The adjustment to the forest carbon loss (FCL) can be less abrupt and sweeping if the response is foresighted and preemptive (SSP2-1.5°C-FCL-Foresight). As the time window for mitigation narrows, the regulation needs to become more severe. Thus, a myopic policy response to the same forest disturbance rate (SSP2-1.5°C-FCL-Myopic) comes at a much greater cost in terms of CO₂ prices required to achieve the carbon budget (Figure 2).

Implications for the energy sector

In the SSP2-1.5°C scenario without FCL, renewable energy supply is expanded by 122 EJ/yr between 2030 and 2050 to limit cumulated emissions to 500 GtCO₂ (Figure 3a). Consequently, emissions from the energy, industry, transport, and buildings sector are reduced by 69% (Figure 3c). To achieve a further 40% reduction in emission by the end of the century, an even higher renewable energy share is required (Figure 3a/c).

Both the foresighted and myopic FCL scenarios project a higher renewables and bioenergy share in the first half of the century (Figure 3b). Preemptive action minimizes the adjustments needed later in the century. In the SSP2-1.5°C-FCL-Foresight scenario, fossil-fuelled energy sources are further reduced by -5.8% (-17 EJ/yr) by 2030 compared to the scenario without FCL, and yearly emissions decrease by an extra -1.5 GtCO₂/yr, mainly from the energy supply and industry sectors (Figure 3d).

In the myopic policy response scenario (SSP2-1.5°C-FCL-Myopic), the additional renewables capacity in 2050 is tripled, increasing from +2.2EJ/yr to +7.3EJ/yr, and fossil-fuelled energy sources are phased out 1.6 times faster (-37.5EJ/yr) than in the foresighted adjustments (-23.3EJ/yr) (Figure 3b). As a result, the emission rate in the myopic scenario is 734MtCO₂/yr lower than in the foresighted scenario. Two-thirds (492MtCO₂/yr) of these further emission reductions come from the industry and transport sectors. Additionally, emission reductions in the buildings sector double between the foresighted and myopic scenario (Figure 3d).

Negative emissions

Energy-side mitigation measures are accompanied by negative emissions to reach the 1.5°C target. Technical CDR capacity (BECCS and Direct Air Capture with Carbon Capture and storage (DACCS)) in the SSP2-1.5°C scenario without FCL doubles every 2.6 years from 2022⁵⁷ to 2030, adding on average 40 MtCO₂/yr of additional capacity per year (Figure 4a).

The SSP2-1.5°C-FCL-Foresight scenario has a 40% higher yearly growth rate in the first half of the century, with +132 GtCO₂ more cumulative negative emissions by 2100 (Extended Data Fig. 4).

Delayed implementation of negative emission infrastructure in the SSP2-1.5°C-FCL-Myopic scenario leads to a steeper growth rate in the second half of the century, almost doubling the rate of the foresighted scenario from +100MtCO₂/yr to +198MtCO₂/yr (Figure 4a). The myopic scenario also relies on more DACCS reaching a capacity of 8.6MtCO₂/yr compared to 3.7MtCO₂/yr in FCL-Foresight and 1.0MtCO₂/yr in the SSP2-1.5°C scenario without FCL. As a result, the cumulative negative emissions in 2100 are +242 GtCO₂ higher in SSP2-1.5°C-FCL-Myopic compared to the SSP2-1.5°C scenario, resulting in almost twice the additional cumulative carbon removal than the foresighted policy in response to the same forest disturbance rate (Extended Data Fig. 4).

Land-use demand

Using land-based methods to remove carbon dioxide from the atmosphere, such as BECCS and A/R, requires vast areas of land. In the SSP2-1.5°C scenario, 224Mha of land is reserved for mitigation until 2030 (Figure 4b). However, this expansion is accompanied by a simultaneous reduction of 247Mha of crop- and pastureland despite a growing population creating the need for more efficient agricultural practices. By the time cumulated emissions peak in 2050, land-use for CDR efforts and reduction in crop- and pastureland will almost double. Under the foresighted FCL response, an additional 69Mha of land is allocated for CDR by 2050. However, the myopic scenario uses +149Mha more land for mitigation, effectively doubling the additional land-use for CDR compared to the foresighted response (Figures 2 & 4b).

Economic cost and drivers

Carbon pricing is central to incentivizing the growth of renewables and CDR. However, such climate mitigation policies may also slow economic growth measured by GDP. The SSP2-1.5°C scenario requires a carbon price of 115 $/tCO₂ in 2030, rising to 327 $/tCO₂ in 2050 (Figure 6). The peak GDP loss due to the climate policy is projected to be -2.1% by mid-century compared to a world without strengthened climate policies (Extended Data Fig. 7).

Taking immediate foresighted action on FCL results in a 22% increase in carbon price in 2030 and 2050 and a peak GDP loss of -2.4%. Myopic policy adjustments lead to a 48% increase in the carbon price, with a peak GDP loss of -2.7%. Thus, the additional carbon price and GDP cost incurred by the FCL response more than doubles between the foresighted and myopic scenarios facing the same forest disturbance rate (Figures 2 & 6, Extended Data Fig. 7).

Our results suggest that immediate and ambitious mitigation action is the best way to prepare for so-far underexplored consequences of human activity like an increase in forest disturbances. Achieving mitigation goals despite FCL required cuts in projected emissions from all sources, as well as large-scale CDR deployment. Implementing a higher carbon price in the modelled scenarios was crucial to achieving the adjustments to the reference mitigation pathway. Delayed and myopic decarbonization efforts exacerbated the problem, resulting in a more than doubled CO₂ price increase (+71$/tCO₂ to +156$/tCO₂), relative GDP loss (-0.25% to -0.63%), and added mitigation area (+69Mha to +149Mha).

Despite assuming disturbances, A/R can still aid mitigation efforts if their regrowth outpaces emissions from disturbed forests. However, without a conservative (large) buffer fraction, establishing additional forests increases the exposure to future CO₂ emissions from disturbances. In addition, a large buffer likely lowers the appeal of A/R establishment as it would increase the price per absorbed CO₂. Further, the buffer could exacerbate issues linked to poor A/R decisions hiking prices by competing for agricultural land and biodiversity loss of plantations. However, positive feedback could also become more extensive as native forests can bolster species richness. To effectively use forests in mitigation efforts, reliable estimates of future disturbance rates are essential.

The A/R establishment in scenarios presented here is partially exempt from the response policies' delayed and myopic behaviour. As the revenue for A/R is formed over a 50-year planning horizon (methods), the A/R decision partially reaches beyond the period of fixed policies of the delayed and myopic scenarios. However, future A/R revenue is discounted by the interest rate. Thus, short-term revenue affected by the delayed and myopic policies is weighted more in the A/R decision.

Future forest disturbance rates, their spatial distribution, and evolution are still highly uncertain. Some sources, like fire, are better quantified than others, such as windfall and pests³⁵. Disturbances from fire are already included in the model chain used here. However, the implemented fire dynamics only lower the average carbon storage and, thus, only indirectly induce emissions by decreasing the potential storage.

Our study aimed to explore the space of societal responses to various disturbance rates resulting from different, unaccounted-for sources of disturbances. However, we did not assess spatial heterogeneities, so we cannot draw any regional conclusions from our results. Additionally, we did not examine differences in the timing of disturbances. In all scenarios, CO₂ emissions from forests were introduced before the cumulative emission peak was reached, allowing for mitigation efforts to adjust and achieve their targets in almost all cases. Only scenarios that aimed to limit warming to 1.5°C with an annual disturbance rate of 1.6% and myopic actions fail to meet their goal. Introducing forest disturbances closer to the peak would likely result in more scenarios being inconsistent with their initial target. The scenarios explored here anticipate sustained negative emissions after the peak is reached. Thus, any CO₂ released from forests late in the century would likely induce less response as the negative emissions form a buffer to the peak of CO₂ in the atmosphere.

As signs of decreasing forest resilience become more and more evident, it is crucial to include major forest disturbance processes in models that determine the global carbon budget. If these processes are not taken into account, unanticipated emissions from increasing rates of forest disturbances could consume a considerable portion of the remaining emission allowance, potentially jeopardizing mitigation efforts.

Data availability

All scenario results are available at https://polybox.ethz.ch/index.php/s/RRsGqrcxlZfdIGJ for the purposes of the review process. Further, the data supporting the analysis of this paper will be archived at Zenodo and receive a Digital Object Identifier before publication.

Code availability

The source code for REMIND 3.0 is available at https://github.com/remindmodel/remind, together with the model documentation at https://rse.pik-potsdam.de/doc/remind/3.0.0/. The source code of MAgPIE 4.6 is available at https://github.com/magpiemodel, with the model documentation at https://rse.pik-potsdam.de/doc/magpie/4.6.0/.

Acknowledgements

This research received funding from the German Federal Ministry of Education and Research (BMBF) and the German Aerospace Center (DLR) via the LAMACLIMA project as part of AXIS, an ERANET initiated by JPI Climate (http://www.jpi-climate.eu/AXIS/Activities/LAMACLIMA, grant no. 01LS1905A),with co-funding from the European Union (grant no. 776608). Further funding was provided by the RESCUE project (grant no. 101056939) within the European Union’s Horizon 2020 research and innovation program.

Author contributions

M.G.W, F.H, and A.P conceived the study. M.G.W, F.H., L.M, N.B, and A.P. designed the scenarios. M.G.W produced the REMIND-MAgPIE model runs with assistance from F.H and L.M. MAgPIE code was extended by M.G.W and F.H. M.G.W drafted the manuscript and created the figures with significant input from F.H., L.M, N.B, G.L, J.P.D, J.H, C.M, and A.P. All authors provided comments on the manuscript.

Corresponding author

Correspondence and requests should be addressed to Michael Gregory Windisch ([email protected]).

Competing interests

The authors declare no competing interests.

Pan, Y. et al. A large and persistent carbon sink in the world’s forests. Science 333, 988–93 (2011).
Fuss, S. et al. Betting on negative emissions. Nature Climate Change 2014 4:10 4, 850–853 (2014).
Fuss, S. et al. Negative emissions - Part 2: Costs, potentials and side effects. Environmental Research Letters 13, 063002 https://doi.org/10.1088/1748-9326/aabf9f (2018).
Bui, M. et al. Carbon capture and storage (CCS): the way forward. Energy Environ Sci 11, 1062–1176 (2018).
Popp, A. et al. Land-use futures in the shared socio-economic pathways. Global Environmental Change 42, 331–345 (2017).
Bodirsky, B. L. et al. Global food demand scenarios for the 21st century. PLoS One 10, 0139201 (2015).
Bock, L. & Burkhardt, U. Contrail cirrus radiative forcing for future air traffic. Atmos Chem Phys 19, 8163–8174 (2019).
Paltsev, S., Morris, J., Kheshgi, H. & Herzog, H. Hard-to-Abate Sectors: The role of industrial carbon capture and storage (CCS) in emission mitigation. Appl Energy 300, 117322 (2021).
Luderer, G. et al. Residual fossil CO2 emissions in 1.5–2 °C pathways. Nature Climate Change 8, 626–633 (2018).
Roe, S. et al. Contribution of the land sector to a 1.5 °C world. Nature Climate Change 9, 817–828 (2019).
Rogelj, J. et al. Scenarios towards limiting global mean temperature increase below 1.5 °c. Nature Climate Change 8, 325–332 (2018).
Forzieri, G., Dakos, V., McDowell, N. G., Ramdane, A. & Cescatti, A. Emerging signals of declining forest resilience under climate change. Nature 608, 534–539 (2022).
Reich, P. B. et al. Even modest climate change may lead to major transitions in boreal forests. Nature 608, 540–545 (2022).
Dial, R. J., Maher, C. T., Hewitt, R. E. & Sullivan, P. F. Sufficient conditions for rapid range expansion of a boreal conifer. Nature 608, 546–551 (2022).
Dow, C. et al. Warm springs alter timing but not total growth of temperate deciduous trees. Nature 608, 552–557 (2022).
Fleischer, K. et al. Amazon forest response to CO2 fertilization dependent on plant phosphorus acquisition. Nature Geoscience 12, 736–741 (2019).
Cunha, H. F. V. et al. Direct evidence for phosphorus limitation on Amazon forest productivity. Nature 608, 558–562 (2022).
Boulton, C. A., Lenton, T. M. & Boers, N. Pronounced loss of Amazon rainforest resilience since the early 2000s. Nature Climate Change 12, 271–278 (2022).
Curtis, P. G., Slay, C. M., Harris, N. L., Tyukavina, A. & Hansen, M. C. Classifying drivers of global forest loss. Science 361, 1108–1111 (2018).
Pendrill, F. et al. Disentangling the numbers behind agriculture-driven tropical deforestation. Science 377, eabm9267 (2022).
Uribe, M.d.R. et al. Net loss of biomass predicted for tropical biomes in a changing climate. Nature Climate Change 13, 274-281 (2023).
Hartmann, H. et al. Climate Change Risks to Global Forest Health: Emergence of Unexpected Events of Elevated Tree Mortality Worldwide. Annual Review of Plant Biology 73, 673–702 (2022).
Hammond, W. M. et al. Global field observations of tree die-off reveal hotter-drought fingerprint for Earth’s forests. Nature Communications 13, 1–11 (2022).
McDowell, N. G. et al. Mechanisms of woody-plant mortality under rising drought, CO2 and vapour pressure deficit. Nature Reviews Earth & Environment 3, 294–308 (2022).
Hlásny, T. et al. Devastating outbreak of bark beetles in the Czech Republic: Drivers, impacts, and management implications. For. Ecol. Manage. 490, 119075 (2021).
Keenan, T. F. & Williams, C. A. The Terrestrial Carbon Sink. Annual Review of Environment and Resources 43, 219–243 (2018).
Friedlingstein, P. et al. Global Carbon Budget 2021. Earth Syst Sci Data 14, 1917–2005 (2022).
Zhu, Z. et al. Greening of the Earth and its drivers. Nature Climate Change 6, 791–795 (2016).
Pugh, T. A. M. et al. Role of forest regrowth in global carbon sink dynamics. Proc. Natl. Acad. Sci. USA 116, 4382–4387 (2019).
Roe, S. et al. Contribution of the land sector to a 1.5 °C world. Nature Climate Change 9, 817–828 (2019).
Rogelj, J. et al. Scenarios towards limiting global mean temperature increase below 1.5 °c. Nature Climate Change 8, 325–332 (2018).
Anderegg, W. R. L. et al. Climate-driven risks to the climate mitigation potential of forests. Science 368, eaaz7005 (2020).
Albert, J. S. et al. Human impacts outpace natural processes in the Amazon. Science 379, eabo5003 (2023).
Lapola, D. M. et al. The drivers and impacts of Amazon forest degradation. Science 379, eabp8622 (2023).
Canadell, J. G. et al. Global Carbon and Other Biogeochemical Cycles and Feedbacks. Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, pp. 673–816 (IPCC, 2021).
Patacca, M. et al. Significant increase in natural disturbance impacts on European forests since 1950. Glob. Chang. Biol. 29, 1359–1376 (2023).
Jones, M. W. et al. Global and regional trends and drivers of fire under climate change. Reviews of Geophysics 60, e2020RG000726 (2022).
Wang, S. et al. Recent global decline of CO2 fertilization effects on vegetation photosynthesis. Science 370, 1295–1300 (2020).
Huntingford, C. et al. Simulated resilience of tropical rainforests to CO2-induced climate change. Nature Geoscience 6, 268–273 (2013).
Forzieri, G., Dakos, V., McDowell, N. G., Ramdane, A. & Cescatti, A. Emerging signals of declining forest resilience under climate change. Nature 608, 534–539 (2022).
Keenan, T. F. et al. Net carbon uptake has increased through warming-induced changes in temperate forest phenology. Nature Climate Change 4, 598–604 (2014).
Richardson, A. D. et al. Terrestrial biosphere models need better representation of vegetation phenology: results from the North American Carbon Program Site Synthesis. Glob. Chang. Biol. 18, 566–584 (2012).
Dietrich, J. P. et al. MAgPIE 4 – a modular open-source framework for modeling global land systems. Geosci. Model Dev. 12, 1299–1317 (2019).
Baumstark, L. et al. REMIND2.1: transformation and innovation dynamics of the energy-economic system within climate and sustainability limits. Geosci. Model Dev. 14, 6571–6603 (2021).
Seidl, R., Schelhaas, M. J., Rammer, W. & Verkerk, P. J. Increasing forest disturbances in Europe and their impact on carbon storage. Nature Climate Change 4, 806–810 (2014).
Senf, C. & Seidl, R. Mapping the forest disturbance regimes of Europe. Nature Sustainability 4, 63–70 (2020).
White, J. C., Wulder, M. A., Hermosilla, T., Coops, N. C. & Hobart, G. W. A nationwide annual characterization of 25 years of forest disturbance and recovery for Canada using Landsat time series. Remote Sens. Environ. 194, 303–321 (2017).
Matricardi, E. A. T. et al. Long-term forest degradation surpasses deforestation in the Brazilian Amazon. Science 369, 1378–1382 (2020).
Senf, C. & Seidl, R. Storm and fire disturbances in Europe: Distribution and trends. Glob. Chang. Biol. 4, 3605-3619 (2021).
Feng, Y. et al. Doubling of annual forest carbon loss over the tropics during the early twenty-first century. Nature Sustainability 5, 444–451 (2022).
Cohen, W. B. et al. Forest disturbance across the conterminous United States from 1985–2012: The emerging dominance of forest decline. For. Ecol. Manage. 360, 242–252 (2016).
Bauman, D. et al. Tropical tree mortality has increased with rising atmospheric water stress. Nature 608, 528–533 (2022).
Hartmann, H. et al. Climate Change Risks to Global Forest Health: Emergence of Unexpected Events of Elevated Tree Mortality Worldwide. Annual Review of Plant Biology 73, 673–702 (2022).
McDowell, N. G. et al. Multi-scale predictions of massive conifer mortality due to chronic temperature rise. Nature Climate Change 6, 295–300 (2015).
McDowell, N. G. et al. Pervasive shifts in forest dynamics in a changing world. Science 368, eaaz9463 (2020).
Anderegg, W. R. L. et al. Climate-driven risks to the climate mitigation potential of forests. Science 368, eaaz7005 (2020).
International Energy Agency. Carbon Capture, Utilisation and Storage – Analysis - IEA. https://www.iea.org/reports/carbon-capture-utilisation-and-storage-2 (accessed 03 May 2023).

Model description

REMIND-MAgPIE combines the energy-economy model REMIND (regional model of investments and development) with the land-use allocation model MAgPIE (model of agricultural production and its impact on the environment). The dynamic global vegetation model LPJmL (Lund-Potsdam-Jena managed land) provides biophysical constraints of the land-use system, which is driven by bias-adjusted climate projections from ISIMIP, based on the CMIP6 earth system model MRI-ESM2 for ssp119 and ssp126 as provided by the ISIMIP project⁵⁸.

The models reach equilibrium through iterative runs of each individual model. After each run, REMIND reports GHG prices and bioenergy demand to MAgPIE. Subsequently, MAgPIE finds an optimal land-use pattern fulfilling the bioenergy demand and adds an incentive/disincentive to A/R / deforestation through the GHG price. MAgPIE, in turn, reports back emissions of the land sector and the price of delivering the required bioenergy, which is then accounted for in the next REMIND iteration.

MAgPIE is a partial equilibrium model minimizing the global cost of agricultural production. The demand for these products like food, feed, and fibre depends on the socioeconomic development following SSP scenarios. Biophysical constraints for yield, water availability, and carbon density are provided by the vegetation model LPJmL. The GHG price drives A/R. A 50-year planning horizon is used to establish the value of the mitigation forest. The removed carbon and the future carbon price are known in this decision. However, the revenue created by the future GHG price is discounted by the expected interest rate. Forests accumulate carbon along sigmoidal growth curves reaching their local storage potential provided by LPJmL. Fire dynamics, as implemented in LPJmL, lower this carbon storage potential.

REMIND is a general equilibrium model combining a Ramsey-type growth model of the economy with an engineering-based energy system model. The two models link the final energy demand generated by the economic activity of major sectors like transport, industry, and buildings. In turn, the economic core is informed about the cost incurred by its energy use. The energy sector's transformation is constrained by inertia and path dependencies created by existing infrastructure and learning curves/adjustment costs incurred by adopting new technologies. Emissions of major GHGs are linked to the primary energy sources. The simple climate model MAGICC explores the resulting change in radiative forcing and global mean temperature.

Peak cumulated emission budgets of 500GtCO₂ and 1150GtCO₂ aim at peak warmings of 1.5°C and 2°C. The temperature is allowed to be at or slightly above its peak but is required to return below the warming threshold by 2100. More than 50% of 1.5°C and 67% of 2°C scenarios reach the threshold by the end of the century⁵⁹. The scenarios establishing the two emission budgets are independent of the scenarios produced in this study. The difference between carbon released and regrown induced by the forest disturbance is accounted for under net land-use emissions taking up part of the emission budget.

Disturbances

MAgPIE does not simulate trees individually. Instead, forests consist of different age classes that develop and accumulate carbon over time. Previously, the carbon storage of any forest in MAgPIE only grew over time as the age-class distribution aged. In this study, a share of each age class is set back to the first class, releasing the carbon difference between them.

Four disturbance rates are explored. Each higher level doubles the disturbance factor from two to four to eight to 16 per thousand per year. Each disturbance rate is introduced in 2030 and built up linearly until the full level is reached in 2050.

The disturbed share of the forest is allowed to regrow immediately. Thus, no forest area is lost to the disturbance. Primary forests are the exception to this as the land type is exclusive to pristine forests in the model. As such, disturbed primary forest is reclassified and regrown as secondary forest. Between perpetual disturbance and regrowth, forests will move towards a new equilibrium carbon storage.

REMIND-MAgPIE runs in timesteps of five years to 2060 and in ten-year timesteps onward to 2150. The yearly disturbance value is multiplied by the timestep length to form the disturbance’s extent between two timesteps. The modified age-class distribution is allowed to age and accumulate carbon immediately. Thus, the released carbon from the loss of storage is partly offset by the regrowth produced by the first age class.

Because the disturbance is summarized over the whole timestep, all yearly disturbances within that time are allowed to regrow the full length of the timestep. Hence, the resulting carbon release is slightly smaller than an implementation where disturbances could be introduced yearly.

Scenario setup

Five scenarios investigate each disturbance rate, differing in their response to the impermanent carbon storage in forests. The scenarios are shaped to explore the space of possible reactions to forest impermanence ranging from immediate, proactive actions to myopic scenarios in which action is only taken years after the forest disturbance is evident.

The most optimistic scenario assumes perfect foresight of the disturbance. Therefore, actions are taken immediately, preparing for forest impermanence even before they take effect in 2030.

Two less optimistic scenarios do not prepare in advance and only act after the forest disturbance becomes evident in 2030. The first/second scenario acts in the first/second timestep (2035/2040) after carbon is first released from forests delaying action by five/ten years. However, after the initial delay, the response is again drawn from a perfect knowledge of future disturbance rates.

The two least optimistic scenarios initially follow the five/ten-year delayed scenarios. However, instead of the perfect foresight of increasing disturbance rates after the delay, they myopically only respond to the disturbance rate experience in that timestep. Thus, the rising disturbance rate between 2030 and 2050 is continuously underestimated. E.g., the myopic scenario acting with a five-year delay only responds in 2035 and as if the disturbance rate remains at its 2035 level of only 40% of its total value. Subsequently, with each timestep (five years), the action is adjusted to prepare as if the disturbance rate will stay at the newly reached point in time indefinitely.

Two cumulative emission peaks are investigated. The more stringent 500 GtCO₂ peak corresponds to the 1.5°C target, and the 1150 GtCO₂ peak aims at the 2°C target detailed in the Paris Agreement.

All response scenarios are explored under middle-of-the-road socioeconomic conditions of the SSP2 pathway. They are compared to a reference SSP2 scenario without forest disturbances to investigate the necessary mitigation adjustments to deal with the disturbance and the cost this incurs. In addition, the response scenarios are also contrasted against each other to probe the cost and effort associated with neglecting disturbances.

Method references

58. Warszawski, L. et al. The inter-sectoral impact model intercomparison project (ISI-MIP): Project framework. Proc. Natl. Acad. Sci. USA 111, 3228–3232 (2014).

59. Riahi, K. et al. Mitigation Pathways Compatible with Long-term Goals. Climate Change 2022: Mitigation of Climate Change. Contribution of Working Group III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC, 2022).

There is NO Competing Interest.

ExtendedDataFig1.png
Extended Data Fig.1 Global energy supply by source of all scenarios. a) Annual primary energy supply in the respective control scenario without forest carbon loss (FCL). b) Difference in annual primary energy supply between the control (a) and the FCL scenarios. The energy supply is categorized by sources: Biomass, nuclear, renewable energy, and fossil fuel. The left column (sub-figures A, C, E, G) shows results of the 1.5°C mitigation scenarios. The right column (sub-figures B, D, F, H) depict results of the 2°C mitigation scenarios. Rows from top to bottom display the four disturbance rates of 2/1000 (A, B), 4/1000 (C, D), 8/1000 (E, F), and 16/1000 (G, H) trees per year. The 1.5°C mitigation scenario under 16/1000 trees per year disturbance (G) did not find a feasible solution for the “FCL-Myopic-10yr” scenario. Thus, no results are shown for the SSP2-1.5°C-FCL-Myopic-10yr scenario.
ExtendedDataFig2.png
Extended Data Fig.2 Global emissions by sector of all scenarios. a) Annual emissions in the respective control scenario without forest carbon loss (FCL). b) Difference in annual emissions between the control (a) and the FCL scenarios. The emissions are categorized by sector: Energy supply, buildings, industry, and transport. The left column (sub-figures A, C, E, G) shows results of the 1.5°C mitigation scenarios. The right column (sub-figures B, D, F, H) depict results of the 2°C mitigation scenarios. Rows from top to bottom display the four disturbance rates of 2/1000 (A, B), 4/1000 (C, D), 8/1000 (E, F), and 16/1000 (G, H) trees per year. The 1.5°C mitigation scenario under 16/1000 trees per year disturbance (G) did not find a feasible solution for the “FCL-Myopic-10yr” scenario. Thus, no results are shown for the SSP2-1.5°C-FCL-Myopic-10yr scenario.
ExtendedDataFig3.png
Extended Data Fig.3 Global negative emission capacity of all scenarios. Annual negative emission rates by source: Direct Air Carbon Capture with Storage (DACCS), Bioenergy with Carbon Capture and Storage (BECCS), and Afforestation/Reforestation (AR). Yearly growth rates in technical carbon dioxide removal (CDR) capabilities (BECCS and DACCS) are shown in pink. Numbers represent the average growth rate needed between given points in time. In the case of 2030, the technical CDR growth rates are indicated for the IEA’s reported values in 2022 (45 MtCO₂/yr installed CCUS capacity⁵⁷) and the model’s output of 2030. The left column (sub-figures A, C, E, G) shows results of the 1.5°C mitigation scenarios. The right column (sub-figures B, D, F, H) depict results of the 2°C mitigation scenarios. Rows from top to bottom display the four disturbance rates of 2/1000 (A, B), 4/1000 (C, D), 8/1000 (E, F), and 16/1000 (G, H) trees per year. The 1.5°C mitigation scenario under 16/1000 trees per year disturbance (G) did not find a feasible solution for the “FCL-Myopic-10yr” scenario. Thus, no results are shown for the SSP2-1.5°C-FCL-Myopic-10yr scenario.
ExtendedDataFig4.png
Extended Data Fig.4 Global cumulative negative emissions of all scenarios. Cumulative negative emission by source: Direct Air Carbon Capture with Storage (DACCS), Bioenergy with Carbon Capture and Storage (BECCS), and Afforestation/Reforestation (AR). The left column (sub-figures A, C, E, G) shows results of the 1.5°C mitigation scenarios. The right column (sub-figures B, D, F, H) depict results of the 2°C mitigation scenarios. Rows from top to bottom display the four disturbance rates of 2/1000 (A, B), 4/1000 (C, D), 8/1000 (E, F), and 16/1000 (G, H) trees per year. The 1.5°C mitigation scenario under 16/1000 trees per year disturbance (G) did not find a feasible solution for the “FCL-Myopic-10yr” scenario. Thus, no results are shown for the SSP2-1.5°C-FCL-Myopic-10yr scenario.
ExtendedDataFig5.png
Extended Data Fig.5 Global land-use change of all scenarios. Difference in primary land-use area by type in comparison to their respective extent in 2020. A/R actions are divided into efforts planned under Nationally Determined Contributions (Aff NDC) and forest establishment driven solely by carbon dioxide price (Aff CO₂-Price). The combined area of total land carbon dioxide removal (CDR) differences, including both A/R and bioenergy, in 2100 is highlighted in black. The left column (sub-figures A, C, E, G) shows results of the 1.5°C mitigation scenarios. The right column (sub-figures B, D, F, H) depict results of the 2°C mitigation scenarios. Rows from top to bottom display the four disturbance rates of 2/1000 (A, B), 4/1000 (C, D), 8/1000 (E, F), and 16/1000 (G, H) trees per year. The 1.5°C mitigation scenario under 16/1000 trees per year disturbance (G) did not find a feasible solution for the “FCL-Myopic-10yr” scenario. Thus, no results are shown for the SSP2-1.5°C-FCL-Myopic-10yr scenario.
ExtendedDataFig6.png
Extended Data Fig.6 Global CO₂-price development of all scenarios. CO₂ price of all five disturbance response scenarios (dark blue, dark/light green and red) next to their respective control scenario (light blue) without forest carbon loss (FCL). The left column (sub-figures A, C, E, G) shows results of the 1.5°C mitigation scenarios. The right column (sub-figures B, D, F, H) depict results of the 2°C mitigation scenarios. Rows from top to bottom display the four disturbance rates of 2/1000 (A, B), 4/1000 (C, D), 8/1000 (E, F), and 16/1000 (G, H) trees per year. The 1.5°C mitigation scenario under 16/1000 trees per year disturbance (G) did not find a feasible solution for the “FCL-Myopic-10yr” scenario. Thus, no results are shown for the SSP2-1.5°C-FCL-Myopic-10yr scenario.
ExtendedDataFig7.png
Extended Data Fig.7 Global GDP loss development of all scenarios. Percent GDP loss inferred by climate action above the baseline of currently implemented national policies (NPi). Shown are all five disturbance response policy scenarios (dark blue, dark/light green and red) next to their respective control scenario (light blue) without forest carbon loss (FCL). The left column (sub-figures A, C, E, G) shows results of the 1.5°C mitigation scenarios. The right column (sub-figures B, D, F, H) depict results of the 2°C mitigation scenarios. Rows from top to bottom display the four disturbance rates of 2/1000 (A, B), 4/1000 (C, D), 8/1000 (E, F), and 16/1000 (G, H) trees per year. The 1.5°C mitigation scenario under 16/1000 trees per year disturbance (G) did not find a feasible solution for the “FCL-Myopic-10yr” scenario. Thus, no results are shown for the SSP2-1.5°C-FCL-Myopic-10yr scenario.

Download PDF

Version 1

posted

You are reading this latest preprint version

Hedging our bet on forest permanence

Status:

Version 1

Abstract

Figures

Main

Discussion and conclusion

Declarations

References

Methods

Additional Declarations

Supplementary Files

Status:

Version 1