全球变化知识资源中心(globalchange): Linearity between temperature peak and bioenergy CO2 emission rates

Many future energy and emission scenarios envisage an increase of bioenergy in the global primary energy mix^{1, 2, 3, 4}. In most climate impact assessment models and policies, bioenergy systems are assumed to be carbon neutral, thus ignoring the time lag between CO₂ emissions from biomass combustion and CO₂ uptake by vegetation⁵. Here, we show that the temperature peak caused by CO₂ emissions from bioenergy is proportional to the maximum rate at which emissions occur and is almost insensitive to cumulative emissions. Whereas the carbon–climate response (CCR; ref. 6) to fossil fuel emissions is approximately constant, the CCR to bioenergy emissions depends on time, biomass turnover times, and emission scenarios. The linearity between temperature peak and bioenergy CO₂ emission rates resembles the characteristic of the temperature response to short-lived climate forcers. As for the latter^{7, 8, 9}, the timing of CO₂ emissions from bioenergy matters. Under the international agreement to limit global warming to 2 °C by 2100³, early emissions from bioenergy thus have smaller contributions on the targeted temperature than emissions postponed later into the future, especially when bioenergy is sourced from biomass with medium (50–60 years) or long turnover times (100 years).

Bioenergy is part of many future low CO₂ emission scenarios and it is the most important renewable energy option in studies designed to align with future RCP projections, reaching approximately 250 EJ yr⁻¹ in RCP2.6 (ref. 1), 145 EJ yr⁻¹ in RCP4.5 (ref. 2) and 180 EJ yr⁻¹ in RCP8.5 (ref. 4) by the end of the twenty-first century. Integrated assessment models and policy directives have mainly focused on the quantification and mitigation of risks associated with deforestation and land-use changes¹⁰, and only recently has the default ‘carbon neutrality’ assumption applied to CO₂ emissions from bioenergy come under scrutiny by governmental authorities¹¹.

In bioenergy systems, the CO₂ exchanges with the atmosphere are usually characterized by fast emissions from biomass combustion and slow CO₂ uptake by vegetation re-growth. As succinctly mentioned in the 5th IPCC Assessment Report¹², this yields a non-zero climate forcing even if the net CO₂ fluxes sum up to zero over time. The climate impact from this temporal asymmetry can be quantified at different points of the carbon–climate cause–effect chain¹², from a simple sum of CO₂ fluxes informing about an initial carbon debt⁵ to radiative forcing and subsequent temperature change¹³. Whereas the temperature response to a CO₂ pulse from fossil fuels is sustained for many centuries at an approximately constant or slightly decreasing value^{6, 14, 15, 16}, recent studies showed that the temperature change from bioenergy CO₂ emissions is characterized by an initial warming followed by a smaller long-term cooling and asymptotically tend to zero^{12, 13, 17}. However, an analysis that disentangles the role of CO₂ emissions from bioenergy within the policy-relevant framework^{3, 7, 8, 18} linking temperature peak (ΔT_peak) and emissions is still missing. Many studies found that the temperature peak of long-lived greenhouse gases (GHG) is roughly proportional to cumulative emissions^{6, 14, 19}, whereas the ΔT_peak from short-lived species is constrained by their maximum emission rate^{7, 8, 9, 12, 20}. The reason is that the atmospheric perturbation from long-lived GHGs such as CO₂ is lasting so long that the induced temperature rise will stabilize only if emissions are reduced to zero¹⁹, whereas the temperature change from short-lived species decreases after a maximum once emission rates have peaked⁹. Within a two-basket approach in which GHGs are differentiated into long- and short-lived⁸, a specific global warming target could therefore be achieved by setting a dual objective to limit cumulative emissions of long-lived GHGs and maximum emission rates of short-lived species⁸.

Here, we show that there is a linear relationship linking the global temperature peak from bioenergy to maximum CO₂ emission rates, as it is observed for short-lived climate forcers. Using the global carbon-cycle climate model OSCAR v2.1 (ref. 21), whose technical description is available in the Supplementary Information, we investigate the climate system response to CO₂ emissions from bioenergy sourced from biomass resources with short (6 years), medium (55 years) and long (103 years) turnover times. The latter case study can be taken as the upper bound for the regeneration period of commercial forest plantations. Summarized in Table 1, the bioenergy experiments are based on post-harvest chronosequences of CO₂ net ecosystem exchanges (NEE) that dictate the rates at which the biomass energy resources can be replenished. We treat biomass as a renewable source, with the system being carbon neutral along the biomass turnover time. Simulations are performed under a constant background climate following the protocol¹⁵ recently used by the IPCC (ref. 12) for the computation of emission metrics and temperature responses (see Methods for specific details). The direct carbon and climate responses to CO₂ pulses for the cases analysed in this study are reproduced in the Supplementary Information, where the possible effects of a changing climate are also explored (Supplementary Figs 3–5).

Table 1: Characteristics of the biomass and post-harvest net ecosystem exchanges (NEE) used in the bioenergy experiments.

Full table

These results are obtained after integration of a global carbon-cycle climate model and empirical observations of biosphere–atmosphere exchanges of CO₂ following harvest disturbance. The responses in atmospheric CO₂ and global mean surface temperature are computed using OSCAR v2.1 (ref. 21), a compact coupled carbon-cycle and climate model that simulates the redistribution of anthropogenic carbon among the main carbon reservoirs (atmosphere, terrestrial biosphere and oceans). A technical description of OSCAR v2.1 is available in the Supplementary Information, together with a description of the origin of the uncertainty ranges shown in the results. The performance of OSCAR v2.1 with respect to other models is benchmarked in Supplementary Fig. 1. The carbon and climate responses of this paper are calculated under constant background climate conditions according to the standard protocol for emission metrics¹⁵—that is, the model is initially forced with historical concentrations up to the reference year (namely, 2010), thereafter the concentration and other anthropogenic forcings are stabilized at the 2010 level (for example, atmospheric CO₂ concentration is kept constant at the value of 389 ppm), and then a CO₂ emission pulse of 100 GtC is added to the atmosphere five years after the reference year (namely, in 2015). The size of this pulse is compatible with the responses to infinitely small pulses, as shown elsewhere for both the carbon-cycle¹⁵ and the climate (temperature) system⁶. The results presented in this study can thus be downscaled to characterize impacts from emissions of smaller size. The atmospheric lifetime of the CO₂ perturbation in the bioenergy cases is computed by fitting the ensemble mean curves in Supplementary Fig. 3a with a first-order decay model. CCR is computed from emission pathways growing at a rate between 0 and 6% as the ratio of the instantaneous global average surface temperature change (in °C) to cumulative carbon emissions (in TtonC). The possible influence of varying background CO₂ atmospheric concentration and climate on the temperature response and CCR are investigated by reproducing the same experiments in 2100 after letting CO₂ concentration change during the twenty-first century according to the four RCP scenarios (Supplementary Figs 4 and 5).

We created two independent groups of 500 idealized emission trajectories (with a ten-year peak phase occurring between 2030 and 2160, followed by a post-peak phase with a decline to zero within a maximum of 100 years) to study the sensitivity of ΔT_peak to ΣE or E_max. In the experiment aiming at testing the sensitivity of ΔT_peak to ΣE (Fig. 3a), emission trajectories are constrained to E_max = 10 GtonC yr⁻¹ and result in ΣE ranging between 0.3 and 2 TtonC. The dependency on E_max (Fig. 3b) is studied over emission trajectories of 5 < E_max < 30 GtC yr⁻¹ and resulting in the same amount of cumulative emissions (ΣE = 1 TtonC). The normalized temperature ranges in Fig. 3c are computed as (ΔT_peak^max − ΔT_peak^min)/ΔT_peak^max using the corresponding temperature peak dynamics. Another set of unconstrained emission trajectories in which E_max and ΣE span between 0 and 20 GtonC yr⁻¹ and between 0 and 4 TtonC, respectively, are generated to distinguish short- and long-lived species as in ref. 8 (Supplementary Figs 6 and 7). Units for the emission trajectories of the non-CO₂ GHGs: ΣE are in 1,000 Mtons and E_max are in Mtons per year for CH₄ and N₂O; ΣE are in Mtons and E_max are in ktons per year for the other GHGs.

In the bioenergy simulations, emissions from combustion are associated with the ecosystem carbon responses represented by the NEE chronosequences (Supplementary Fig. 2 shows the data used in this study). NEE values include CO₂ sequestration by net primary productivity (NPP) and ecosystem respiration (that is, oxidation of carbon from soil and dead biomass). In the chronosequences representative of a medium and long turnover time, the ecosystem is a net carbon source for approximately the first couple of decades following disturbance, owing to the dominant respiration flux (mainly from the decomposition of harvest residues) over NPP, which then becomes dominant and ecosystems turn to be net carbon sinks. These ecosystem responses are prescribed to OSCAR v2.1 in the form of a series of small pulses whose cumulative value over the turnover time is equal to the emission pulse. Many woody bioenergy cases can be expected to fall between the short and long turnover time; biomass species with turnover times shorter than six years, such as perennial grasses and short rotation coppice, cause a perturbation smaller than that shown here for the short turnover case, and can be approximately considered climate neutral. The simulations in this paper are based on the assumption that bioenergy is a renewable resource—that is, it fully regenerates along the turnover time—although on a case-specific basis some net carbon losses (for example, in the first succession) or gains (for example, in fertilized plantations) can be expected.


Email:
Passwd
验证:
	换一张
Have you forgotten your password? Stay signed in Log In

全球变化知识资源中心

用户服务

站点统计

相关链接

联系我们