the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Nov 2024
flat10MIP: An emissions-driven experiment to diagnose the climate response to positive, zero, and negative CO2 emissions
Abstract. The proportionality between global mean temperature and cumulative emissions of CO2 predicted in Earth System Models (ESMs) is the foundation of carbon budgeting frameworks. Deviations from this behavior could impact estimates of required net zero timings and negative emissions requirements to meet the Paris Agreement climate targets. However, existing ESM diagnostic experiments do not allow for direct estimation of these deviations as a function of defined emissions pathways. Here we perform a set of climate model diagnostic experiments for the assessment of Transient Climate Response to cumulative CO2 Emissions (TCRE), Zero Emissions Commitment (ZEC), and climate reversibility metrics in an emissions-driven framework. The emissions-driven experiments provide consistent independent variables simplifying simulation, analysis and interpretation with emissions rates more comparable to recent levels than existing protocols using model-specific compatible emissions from the CMIP DECK 1pctCO2 experiment, where emissions are strongly weighted towards the end of the experiment at significantly greater than present day values. A base experiment, ‘esm-flat10’, has constant emissions of CO2 of 10 GtC per year (near-present day values), and initial results show that TCRE estimated in this experiment is about 0.1 K less than that obtained using 1pctCO2. A subset of ESMs exhibit land carbon sinks which saturate during this experiment. A branch experiment, esm-flat10-zec, measures ZEC, which we find is reduced by 25–50 % compared with 1pctCO2 branch experiments. A final experiment, esm-flat10-cdr, assesses climate reversibility under negative emissions, where we find that peak warming may occur before or after net zero and residual warming after removal of all greenhouse gases is well described by ZEC in most models and that current Simple Climate Model (SCM) distributions may be over-estimate temperature reversibility compared with ESMs. We propose a set of climate diagnostic indicators to quantify various aspects of climate reversibility. These experiments were suggested as potential candidates in CMIP7 and have since been adopted as “fast track” simulations.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Geoscientific Model Development.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.- Preprint
(3655 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
AC1: 'Missing Figure 1', Benjamin Sanderson, 15 Nov 2024
We regret that the first figure in the preprint is erroneously a duplicate of Figure 3. A high resolution version of the correct figure is attached to this comment.
Figure 1: Experiment design. a) and b) show annual and cumulative carbon emissions as a function of time for the four experiments. Panel c) shows global mean surface temperature derived from cumulative emissions, assuming a perfectly linear TCRE relationship, with expected temperature evolution assuming cumulative emissions proportionality using the IPCC AR6 WGI best TCRE estimate (solid line, 1.65°C per 1000 PgC) and likely range (shaded area, 1.0-2.3°C per 1000 PgC) (Intergovernmental Panel on Climate Change (IPCC), 2023b).
-
RC1: 'Comment on egusphere-2024-3356', Kirsten Zickfeld, 11 Feb 2025
The study tests the utility of a set of idealized experiments – the suite of emissions driven “flat10” experiments described in Sanderson et al., 2024 – using an ensemble of ESM, model of intermediate complexity and simple climate model simulations. A set of metrics (TCRE, ZEC and metrics for reversibility) is evaluated and compared to metrics evaluated using a prior set of simulations based on the concentration-driven 1pctCO2 experiment. The study finds that TCRE is slightly lower and the ZEC substantially lower than when quantified with the 1pctCO2 experiment.
This is an insightful study which demonstrates that the flat10 experimental protocol is suitable for quantifying policy-relevant metrics in a consistent way, and that these metrics exhibit important differences relative to those quantified from simulations based on the 1pctCO2 experiment. The manuscript is well structured and well written for the most part, but lacks clarity and precision in some instances, which are identified in the specific comments below. The manuscript would also benefit from a stronger processed-based interpretation of the differences relative to metrics quantified from the 1pctCO2 experiment. There are several typos, missing definitions and omissions in the figure labelling (only partly reported below) and the authors should re-read the manuscript carefully before resubmission.
Specific comments
Numbers refer to line numbers.
43: “…CMIP DECK 1 pctCO2 experiment, where emissions are strongly weighted towards the end of the experiment”. This wording is used throughout the manuscript but its meaning is unclear to me. Why are emissions “weighted”? Also, metrics are typically quantified around year 70 of the 1pctCO2 experiment, not at the end.
48-51: Long sentence. Suggest to break it into two.
58: The proportional relationship was also demonstrated with observational data – see Gillett et al., 2013.
68-69: Several corrections must be applied to calculate emissions budgets using TCRE, not just for temperature impacts of non-CO2 emissions. These include human-induced warming, Earth system feedbacks among others (e.g. IPCC AR6 WGI, Ch. 5, Fig. 5.31).
80: TCR corresponds to the term DT(70)/I_atmos(70) (not DT(70)) in this equation.
84: “deviations from cumulative emissions proportionality” manifest in different ways, not just as ZEC. They include deviations from linearity at higher levels of cumulative CO2 emissions (>1,500 GtC; IPCC AR6 WGI, Ch. 5) that are unrelated to ZEC or (ir)reversibility. Discuss these sources of deviation and clarify that ZEC is one of them or reword the sentence.
87: A clearer definition of RAZE and its relationship to ZEC would be helpful.
111-112: “…inhibit a clear diagnosis of whether and how the general response to negative emissions differs from the climate response to positive emissions”: There are two issues with the CDRMIP 1pctCO2-cdr reversibility experiment: i) the asymmetry in compatible CO2 emissions and removals, which is addressed in the flat10-cdr experiment, ii) the lagged response of the coupled climate-carbon cycle system to positive emissions, which confounds the response to negative CO2 emissions. This issue is exacerbated in the 1pctCO2-cdr experiment due to the abrupt transition from large positive to large negative emissions, but is also present in the flat 10-cdr experiment. This should be acknowledged. Two ways have been proposed to isolate the response to negative emissions: subtract the committed response to positive emissions from the overall response (Zickfeld et al., 2016; Koven et al., 2023; Chimuka et al., 2023), or apply negative emissions from an equilibrium state (Zickfeld et al., 2021). Neither approach is entirely accurate.
124: “isolate the impact of global scale removals”: How will this be done? Accurately isolating the response to removals is not possible with the proposed experimental design as argued above.
198-200, “TR1000 would be a measure…”: Unclear what TR1000 measures exactly based on this description. The warming centered around year 100 could itself deviate from TCRE nonlinearity, therefore subtracting “the expectation from TCRE” may not be the same as subtracting “a 21 year average around year 100”.
215: “This ensemble…”: There appears a sentence missing that serves as transition from the experimental to the models that will be used to run the experiments.
216; “each model is selected for its specific configuration”. Unclear – do you mean “categorized”, e.g. into ESM, EMIC, SCM ?
276, “majority stabilize rapidly”: Land carbon in most models appears to continue to decline slightly after year 100.
279-281: Repeats experimental design description.
282, “… indicative on non-proportionality of temperature with cumulative emissions”. This finding could be described in more detail as the temperature response at year 300 differs substantially from the temperature response in concentration driven simulations, with near zero warming or net cooling relative to year 0 in many models.
290: This should probably read “timing of the PEAK land sink relative to net zero”.
310-311: “Similarly…”: This sentence does not seem to belong here as the paragraph describes the carbon cycle response.
314: "Transient CLIMATE response to positive emissions”.
344-345: You could report ZEC 90 for FLATMIP simulations to allow a proper comparison with ZECMIP simulations.
350: T_1000PgC and T_100yr not defined.
365-366: Smaller ZEC in experiments with lower emission rates up to the point of net zero was noted in MacDougall et al (2020) and attributed to the warming and carbon cycle response being closer to equilibrium.
372-373: Sentence unclear.
480-481: Statement unclear.
Fig. 2: The temperature trajectory does not correspond to the actual response in the flat10-cdr experiment. In emission driven simulations warming after all CO2 is removed from the atmosphere should be closer to zero.
Fig. 3: DGMST axis labels in panels b, c: Use either T_1000PgC or T100yr consistently (assuming they are the same here).
Fig. 6: Circles could be shown in panel d) for comparison.
Fig. 10: Horizontal axis labels not positioned properly.
Fig. 12: Units missing in axis labels.
References:
Chimuka et al., 2023, Quantifying land carbon cycle feedbacks under negative CO2 emissions. Biogeosciences. 20: 2283–2299, https://doi.org/10.5194/bg-20-2283-2023.
Gillett et al., 2013, Constraining the Ratio of Global Warming to Cumulative CO2 Emissions Using CMIP5 Simulations, J. Clim., https://doi.org/10.1175/JCLI-D-12-00476.1
Zickfeld et al., 2021, Asymmetry of the climate-carbon cycle response to positive and negative CO2 emissions, Nat. Clim. Chang., 11, 613–617, https://doi.org/10.1038/s41558-021-01061-2.
Citation: https://doi.org/10.5194/egusphere-2024-3356-RC1 -
RC2: 'Comment on egusphere-2024-3356', Stuart Jenkins, 26 Feb 2025
This study analyses the flat10MIP experiments as a method to measure key coupled climate-carbon-cycle properties in Earth System Models, with the experimental protocol outlined in Sanderson et al. (2024). The authors frame the novelty of this work around the experimental design (CMIP models are forced with constant CO2 emissions at 10GtCO2/yr for a period of 100 years), meaning that models are run in emissions-driven configurations, forced at approximately present-day emissions rates and all react to the same quantity of cumulative CO2 emissions. This is a nice advancement – it is a more relevant perturbation for characterising key model response characteristics, such as the TCRE, ZEC and the extent of reversibility for CO2-induced warming.
This study estimates TCRE and ZEC from flat10 and flat10cdr experiments, finding TCRE is slightly lower on average (0.1K) than when measured with DECK 1%/yr CO2 concentration increase experiments. ZEC is shown to be 20-25% lower in this experiment configuration.
The manuscript is generally well written, although there are several places where greater discussion would be beneficial, particularly linking to processes which could be responsible for the emergent behaviours described in this work, and a handful of typos etc to be caught. Specific comments are left below. A general tightening of the written text and figure captions would be valuable before final submission.
Specific comments
ZEC needs to be qualified over a timescale throughout the paper – ZEC is time dependent and thus should not be talked about as being e.g. 20-25% smaller, without qualifying the timescale over which this was measured.
Line 50/51: “may be overestimate…” change to “may overestimate…”
Line 58: Observational constraints are also possible, although granted these do all require a model of some kind to determine characteristic warming response shapes, but not necessarily only ESMs.
Line 86/89: RAZE could instead be defined as measure of the residual warming trend over a multidecade interval following net zero, normalised by the warming level. That way it is consistent with the definition of ZEC, where both then refer to their measuring of post-net-zero warming behaviours.
Line 93-95: Can we clarify that the throughout the text ‘ZEC’ refers to ZEC over the first century following net zero? Or discuss evidence of multi-century response characteristics and how these may differ from the multidecade response.
Line 121/123: Again, clarity on the time horizon of this ZEC assessment is needed.
Line 126/127: “Further studies are needed to”? or “Regional and component responses require further study beyond the scope of the globally aggregated analysis presented here”?
Line 132/134: suggest edit for readability? “Here, Flat10MIP simulates 3 of the 4 experiments proposed in (Sanderson et al., 2023) using CMIP6 generation models, as a pilot study in preparation for CMIP7.”
Line 149/152: Slight lack of clarity on what is requested, 150 or 300 yr simulations? I think you mean in this flat10MIP experiment set you requested 150 years, and in CMIP7 iteration models request could rise to 300 years, for reasons you lay out. Is that correct interpretation? Could it be clarified that over these experiments the expectation is to run +10GtC/yr emissions throughout?
Line 157: Nice. Would be good to have this definition of ZEC earlier in paper, as per previous comments.
Line 166: “Cumulative net zero emissions” is a slightly unknown term. Could you say it without jargon? “…such that cumulative emissions return to zero, i.e. emissions and removals sum to zero.”? Maybe not, but net zero is such a well understood term it may be worth reserving it for only describing annual emissions balance?
Line 174/180: This is good, glad to see this experiment also proposed. It would be good to briefly justify the expectation of a difference in response to a sudden cessation in emissions here. I.e. why do you need a gradual-NZ experiment as well as a sudden cessation one? Do you expect your ZEC measures in this study to have been impacted by the sudden cessation approach. Is such an experiment a useful measure of a true earth system property, or an emergent measure of a model’s response to this particular scenario?
Line 216: potentially worth adding a qualifying sentence on the impact of measuring earth system properties from individual experiment runs with variability. Clearly ESMs will be impacted by this in a very different way to EMICs and SCMs.
Line 224: closing bracket is missing.
Figure 2: I imagine this figure was thrown together quite quickly. Could it be tidied up a little? E.g. bring all text within figure bounds.
Line 242/244: Long sentence which I took a few read throughs to grasp. Potentially look to clarify.
Line 279: Some repetition here on the above experimental description. Suggest remove and refer to above.
Line 283/285: Could also reference Jenkins et al. (2022) RAZE study. Negative RAZE corresponds to warming peaking before net zero. “as seen in similar experiments [Koven et al.] and ZECMIP experiments [Jenkins et al.]”.
Line 303/311: What would be more informative is an attempt to link this explanation to the underlying mechanisms and possible processes which could cause this diversity of responses.
Figure 7: Colouring is not very readable. White text on light colours particularly.
Line 348/349: Is there an explanation for this that could be offered?
Line 350/353: FaIR and MAGICC and CICERO-SCM are all tuneable to have various responses. What does the distribution of model responses actually tell us? How have the three parameter distributions been selected which you run over each model? Surely this is a key reason for an observed differences in model response distribution?
Line 364/373: Is it inevitable that the ZEC will be lower under this experiment design, compared to ZECMIP, since ZEC has demonstrated proportionality to the “average cumulative emissions over the period” [Jenkins et al 2022], which is different under each experiment design? This could be addressed as a point in the final paragraph of this section [lines 384-390].
Line 412/417: This explanation of TCRE + ZEC is not a particularly precise explanation, given the ZEC is both time and scenario dependent. RAZE approach is an alternative (and more scenario and time dependent) justification of this emergent behaviour (at least from the perspective of a global impulse response model), and robustly identifies the emergent ZEC (multi-decadal) response as impacting before and during the net zero transition. In a symmetric experiment design like the one proposed in flat10cdr, one would then expect half of the ZEC to be realised at the time of net zero, in so far as we trust a FaIR-like IR model to be representative of the coupled climate-carbon-cycle response in ESMs.
Citation: https://doi.org/10.5194/egusphere-2024-3356-RC2
Data sets
Global mean output from flat10MIP simulations B. M. Sanderson https://doi.org/10.5281/zenodo.14012042
Model code and software
Model code to reproduce study B. M. Sanderson, N. Steinert, and C. D. Koven https://doi.org/10.5281/zenodo.14012042
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 609 | 97 | 10 | 716 | 6 | 6 |
- HTML: 609
- PDF: 97
- XML: 10
- Total: 716
- BibTeX: 6
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1