the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Towards the definition of a solar forcing dataset for CMIP7
Thierry Dudok de Wit
Ilaria Ermolli
Margit Haberreiter
Doug Kinnison
Daniel Marsh
Hilde Nesse
Annika Seppälä
Miriam Sinnhuber
Ilya Usoskin
Abstract. The solar forcing prepared for the 6th round of the Coupled Model Intercomparison Project (CMIP6) has been used extensively in climate model experiments and has been tested in various intercomparison studies. Recently, an International Space Science Institute (ISSI) Working Group has been established to revisit the solar forcing recommendations, based on the lessons learned from CMIP6, and to assess new data-sets that have become available, in order to define a roadmap for building a revised and extended historical solar forcing data-set for the upcoming 7th round of CMIP. This paper identifies needs for improvements and outlines a strategy to address them in the planned new solar forcing dataset. Proposed major changes include the adoption of the new TSIS-1 solar reference spectrum for solar spectral irradiance and an improved description of top of the atmosphere energetic electron fluxes, as well as their reconstruction back to 1850 by means of geomagnetic proxy data. In addition, there is an urgent need to consider the proposed updates in the ozone forcing data-set in order to ensure a self-consistent solar forcing in coupled models without interactive chemistry. Regarding future solar forcing, we propose consideration of stochastic ensemble forcing scenarios, ideally in concert with other natural forcings, in order to allow for realistic projections of natural forcing uncertainties.
- Preprint
(681 KB) - Metadata XML
- BibTeX
- EndNote
Bernd Funke et al.
Status: final response (author comments only)
-
RC1: 'Comment on gmd-2023-100', Anonymous Referee #1, 01 Aug 2023
In their manuscript, the authors make suggestions for the construction of a solar forcing dataset for use in the upcoming phase 7 of the Coupled Model Intercomparison Project (CMIP). The final sentence of the Introduction nicely summarizes what I see as the main point of this manuscript: “An important aspect of this work is the need for community feedback, as this will eventually help us translate these suggestions into recommendations for CMIP7.” I find the publication of such suggestions very timely and useful, and I would very much like to see this manuscript be published quickly because, as the authors mention, the solar forcing dataset is a prerequisite for the construction of an ozone dataset to be used by models without interactive ozone chemistry, and it would be very useful to coordinate with the PMIP community concerning the construction of solar forcing data for the deeper past. For all modeling groups that want to participate in CMIP7 the preparation is getting more difficult the later the necessary input datasets will be available. I have only minor suggestions which I’d like the authors to consider before I can recommend publication of the manuscript. Two of them are more general, the others are listed below ordered by their occurrence in the text.
First, it would be very useful to clarify the role and mandate (if existing) of the group of authors for the construction of the envisaged CMIP7 solar forcing dataset. Concerning ozone, in Chapter 4 it is said that “This effort will be coordinated by the CMIP7 Climate Forcing Task Team.” What is this team, and which role does it play concerning solar forcing? It would also be very useful to mention where the team of authors will become active themselves in the construction and where they rely on input from other groups. For example, from my reading of Chapter 2 I got the impression that a reference quiet-Sun spectrum is available, but for the irradiance variability the authors may depend on input from the SATIRE and/or NRLSSE groups. Is there any commitment from these groups or at least an established forum for the potentially necessary coordination? It would also be useful to more clearly address a coordination with the PMIP group. Lines 106 ff. sound to me like the groups of authors will wait for a recommendation of the PMIP group. I think it would be very useful to search for a consensus proactively. Concerning particle forcing it sounds to me that the group of authors may have all the tools and data available which are necessary to produce the forcing data themselves. Is this correct or not? Please spell this out.
Second, I think that for climate modellers, and if comments from them are intended, it would be very useful to more comprehensively discuss the impacts of the choices which have to be made for constructing the solar forcing dataset. In the introduction it is said that “new data-sets, if adopted, would introduce changes in the radiative forcing of climate, either directly or via their influence on atmospheric composition”. Why not summarize and, wherever possible, quantify the major effects somewhere in the manuscript? The definition of future solar forcing is left, to my reading, relatively open in the final section of the manuscript. As also some future forcing is needed, I’d like to see a more specific suggestion also for this period. It may be useful to consider lessons learned from earlier CMIP phases. Sedlacek et al. (Earth and Space Sciences, 2023; probably published after the submission of this manuscript) come to the relatively strong conclusion that it may not be necessary to provide different future scenarios: “Our results indicate that low amplitude solar forcings such as the EXT CMIP6 or similar are not worthwhile considering during the next CMIP type of activities.“ There may be other opinions, but I think this result should be discussed. On the other hand I could imagine that the relatively large energy shifts between the visible and near-infrared parts of the solar spectrum in the new TSIS compared to earlier data, as reported by the authors, may have a non-negligible impact. Are estimates of the impact already possible?
L25: “However, the analysis of climate model simulations that did use the M17 data-sets also revealed some issues. Small changes in the shape of the solar reference spectrum, for example, impacted climate simulations and required careful tuning of the models.” In the spirit of what I wrote above: Here it would be very useful to be more specific and provide references if possible. What were the issues? Which small changes caused which impacts?
L52: “SATIRE uses the WHI spectrum, where available …” For which wavelengths?
L85: “NRLSSI2 is more data driven as it relies on solar proxies only.” I may not understand which data are referred to, here, and what “more data driven” actually means.
L106: ”Other aspects …” That’s only one aspect, right?
L113: “… in both models. Unfortunately, different versions coexist …” Which models? I guess SATIRE and NRLSSI but as they are not mentioned in the lines above it might be good to say that again. And please be more specific concerning the different versions.
L115: Why should the process “be flexible enough to allow for yearly updates”? I do see the appeal of a flexible tool, but in view of CMIP yearly updates wouldn’t be necessary.
L130: “leading to a significant underestimation of the atmospheric response in the middle and upper mesosphere (Smith-Johnsen et al., 2018; Sinnhuber et al., 2022; Szelag et al., 2022)” and L144 “lead to a stronger atmospheric response (Sinnhuber et al., 2022; Pettit et al., 2021)” I only checked the Sinnhuber et al. paper, but their conclusions don’t seem to fit well to these statements, they, e.g., write: “In the high-latitude upper mesosphere and lower thermosphere above 80 km altitude, multi-model mean results of NO using different MEE ionization rate data-sets are very similar”, and “In the high-latitude mesosphere below 85 […] it is not possible to provide a robust estimate as to which of the ionization rate data-sets perform best”, and “All three observational data-sets agree on a significant NO enhancement during and after the geomagnetic storm at and below 70 km altitude.” So I suggest to summarize the potential impacts of an improved dataset on the atmospheric in a more nuanced way.
L225: “up to 15% ozone reduction on average, solar cycle variations about the same magnitude”
Citation: https://doi.org/10.5194/gmd-2023-100-RC1 -
CC1: 'Comment on gmd-2023-100', Anoruo chukwuma, 05 Aug 2023
The content of the paper is very interesting and to improve CMIP. However, I suggest a more clear deficiency of CMIP6 and what will be added to the proposed CMIP7 in solar forcing modeling. The highlights could be done in <steps-by-steps>.
Citation: https://doi.org/10.5194/gmd-2023-100-CC1 -
CC2: 'Comment on gmd-2023-100', Gavin A. Schmidt, 07 Aug 2023
This is a timely discussion, and kudos to the authors for initiating it. I have three main suggestions for the team.
1) There are many files and inputs which are being created, but they are not all equally used. Perhaps the authors could assess the literature to see what was adpoted broadly and which datasets were not as well utilised. This may inform the prioritization of ongoing and future efforts.
2) Structural uncertainty in solar forcing, past and future, is important to characterise. While not as large as the uncertainty in aerosols, this uncertainty can play a role in the attribution of past climate changes. In PMIP3, we specifically set out alternative forcings (for the TSI and SSI) that groups could use (or not) to characterise this and this was broadly done by many groups in the last1ky experiments (Schmidt et al, 2011; 2012). I would strongly suggest doing something similar for the historical experiments. Create self-consistent separate input files based on the SATIRE and NRLSSI efforts, but do not average them to produce a 'best' guess that has not been validated - and as you note in the text, had weird inconsistencies. Ideally, and this may take some time, we would want to be creating an ensemble of reconstructions from a sampling of uncertain parameters within the reconstruction process, taking into account uncertain raw data and processing choices. Then median, maximal, and minmal reconstructions could be derived in a relatively coherent way. This might not be viable for CMIP7, but this should be the medium term aim. This would also allow for greater consistency with the PMIP5(?) forcings.
3) With respect to compositional feedbacks. First, these are not limited to ozone; there are also stratospheric water vapor effects via photolytic reactions with H2O and ozone-mediated changes in the oxidation of CH4 - which might even be more important than some of the SEP effects? Second, some thought might be directed towards creating a blended ozone product that uses the observed changes in ozone that are coherent with the solar cycle together with a model based interpolation. This would have the advantage of being calibrated to the observations and not be totally beholden to the assumptions in any specific GCM (or the inputs to it, such as the uncertain background spectra). Even better would be a calibrated parameterization that provides a delta(O3) field as a funciton of the TSI/SSI change that could be valid across the PMIP, DECK and ScenarioMIP experiments. Indeed, a linearized ozone parameterization that changes as a function of TSI, temperature, etc. might provide most of what would be seen in a fully functional climate-chemistry model at a fraction of the computational cost and which would be useful far beyond the solar forcing issue.
Thanks.
Citation: https://doi.org/10.5194/gmd-2023-100-CC2 -
RC2: 'Comment on gmd-2023-100', Anonymous Referee #2, 27 Aug 2023
The manuscript presents the strategy for improvement of the applied for CMIP6 simulations
solar forcing data set. The authors suggested major changes inspired by the new TSIS-1 solar
reference spectrum and new findings in energetic electron fluxes and spectrum description. The
authors also propose switching from deterministic to stochastic ensemble forcing scenarios. In
general, the manuscript is useful, but not mature enough and requires major improvements.
Major issues
1. The authors did not consider “the latest scientific advances made in the understanding of
climate response” declared in the introduction
2. Critical analysis of the CMIP6 solar irradiation data set is missing. Some analysis in lines
26-27 is confusing. How the climate simulations were impacted, why tuning was
necessary. It would be very helpful to mention how the improvement in CMIP6
mentioned in lines 22-25 were used by CMIP6 modeling teams.
3. Any description of the requirements from CMIP7 is missing. To understand the process,
it is necessary to have good understanding of the final users. For example, is it really
necessary to provide extensive energetic particle forcing?
4. The climate community is not represented in the co-author list. It is mentioned in the
manuscript, that the strategy for the solar forcing definition should be developed in close
collaboration with climate community. But the publication of this manuscript looks like
the authors would like to have substantial help from the people representing climate
community without involving them to the process.
5. The plan and timeline for the development of solar forcing (including the ozone fields)
for the simulations of future climate is not clearly presented. What is CMIP7 Climate
Forcing Task Team?
Minor Issues
Lines 41-44: These arguments are obvious.
Line 43: less than 1% → around 0.1%
Lines 65-68: Some illustration would be helpful. How re-normalization is performed?
Line 98: Pragmatic solution is not necessary the best.
Line 105: Which new data? All listed in the para or some of them?
Line 108: There were alternatives to the SATIRE in PMIP4.
Lines 109-115: This para is confusing. Does the future forcing will be repletion of some
particular time in the past? Which time then? Yearly update for CMIP7 is strange. Do the
authors really expect yearly rerun of all simulations?
Lines 115-122: I understand some political background of the proposed arithmetic mean, but this
approach is not scientifically solid. Potentially, the authors can establish the accuracy of both
data sets relative to the observed SSI/TSI and simply apply better (even slightly better) model.
Lines 136-137: Which minor updates are suggested?
Lines 139-149: Nesse Tyssøy et al., (2022) found large differences between different data sets?
Why Asikainen (2019) was chosen? Maybe it is better to use arithmetic mean as suggested for
SSI?
Line 153: Now it is suggested to apply van de Kamp et al. (2016). How about strong
underestimation and Asikainen (2019) mentioned in the previous section?
Lines 156-165: The proposed plan does not look feasible taking into account December 2023
deadline.
Lines 191-194: Switch from dipole to IGRF could generate some jumpsCitation: https://doi.org/10.5194/gmd-2023-100-RC2 -
CC3: 'Comment on gmd-2023-100', Valentina Zharkova, 28 Aug 2023
Dear authors,
Your solar forcing term is too simplified and not correct. You ver-averaging the data taking one measurement f TSI per year, In statistics averaging wrks only for the data with normal distribution. Wile variation off YSI per month and per year are far from normal.
I suggest that you should consider the ISI variations along the Earth orbit, which is shown too change significantly in this millennium as I shown in the book chapter https://www.intechopen.com/chapters/75534 and in other papers shown in my web page https://soolargsm.com, e.g.https://arxiv.org/pdf/2008.00439.pdf, https://www.scirp.org/journal/paperinformation.aspx?paperid=124007.
This will allow your codes t reflect the extra heating the Earth atmosphere gets every month during March-July every year in the millennium 1600-2600.
Kind regards
Prof. Valentina Zharkova
Citation: https://doi.org/10.5194/gmd-2023-100-CC3 -
RC3: 'Comment on gmd-2023-100', Tom Woods, 29 Aug 2023
This is an excellent paper summarizing the plans for updating the inputs of solar irradiance, energetic electrons, and ozone forcing data for CMIP7. This manuscript does not provide results of those updates though, but instead is a very useful progress report on this research activity.
I only have one minor (editorial) comment.
Line 108: Change “one used for for CMIP6” to “one used for CMIP6”.
Citation: https://doi.org/10.5194/gmd-2023-100-RC3 -
CC4: 'Comment on gmd-2023-100', Gareth S. Jones, 08 Sep 2023
It is very interesting to see a paper describing how the process to define solar forcing for a future phase of CMIP will be done.
It seems that the main areas of interest to those running climate simulations with solar forcings are covered.I have some views that the authors might like to consider.
* General
After reading the manuscript it is not clear to me how decisions are going to be made about what to do for CMIP7 and by who. Will it just involve the authors of this study, or a panel of some kind? Is the wider community going to be surveyed or is this paper the only route to feedback to the decision makers, via this discussion area?
* Reference spectrum (Section 2.1)It would be really helpful to include a plot showing the proposed reference spectra relative to the alternatives and past ones.
This would help to understand what possible climatic impact this may have, as referred to in Lines 67-69, and as supporting evidence for the "problem" when averaging the two solar irradiance models as referred to in Lines 93-94.
* Solar datasets for simulations of the pastIt would also be really helpful to include a plot of the historical timeseries of the different TSI datasets and the SSI changes being described in Section 2.2, including alternative datasets not currently mentioned. Figure 1b in Yeo GRL 2020 [5] is a good example of what I had in mind.
Lines 97-105 & 116-121 The authors propose to just use two models of historical solar irradiance. But what about others that are available, some of which have been mentioned by the IPCC [1]. (e.g., [2,3,4])? There may be good reasons to not use or assess them - for instance the evidence for relatively large increase in TSI since the Maunder Minimum might be lacking [5] - but it would be helpful to briefly explain why the authors are excluding some datasets.
Lines 109-110 "As for CMIP6, we are planning to provide an ensemble of forcing scenarios with daily values up to 2300". The datasets provided by Matthes et al. 2017 [6] for input4MIPs were just two scenarios. A reference ('REF') and a deep minimum future ('EXT'). Not sure that could be called 'an ensemble'.
Lines 116-121 Later on there is the discussion about future forcing uncertainty, but nothing is mentioned about what could be done for assessing historical forcing uncertainty.
It may be impractical for modeling centres to run many multiple historical simulations to sample forcing uncertainties, but that is where simple models may be useful [7]. I strongly suggest that the authors also consider providing some ways of sampling the solar historical forcing uncertainty, similar to what they propose for the future scenarios. This would be in addition to the single recommended historical TSI/SSI for use by coupled models.
* Solar datasets for simulations of the futureL239-L241 There are non insignificant differences between what solar irradiance models and observations give for recent TSI [9], with much larger uncertainties over longer timescales [5]. The reduction in TSI over the 21st century in the projection [6] should probably be mentioned. Is it too early to assess it? My opinion is that it was too speculative to propose for use with climate models.
I think rather than saying "realistic" for the Matthes et al. 2017 [6] dataset projection, a better word is 'plausible'.L251-252 I suggest expanding on what is meant by "stochastic ensemble forcing scenarios". Does this suggest a large ensemble of TSI timeseries, say, each corresponding to different plausible evolutions of solar cycle lengths and amplitudes and longer time scale magnitude?
Expecting modeling centres to use multiple future solar irradiance datasets in their coupled models may be optimistic to say the least. The effort to create ancillary files to run on each model, for each "ensemble forcing" should not be underestimated. Institutions submitted ssp245 (as an example) simulations for 46 models. Most of them (~30) submitted initial condition ensembles of three or less, while only 13 submitted greater than 5 ensemble members. Thus it is likely only a few models will be able to sample some of the proposed future solar forcing uncertainty.
A danger could be that some of those solar forcing ensembles are over sampled by those models with few ensemble members, reducing their usefulness.
It is likely that only a few modeling centres will have the resources to submit a reasonable number of simulations that sample the solar forcing ensemble.
Matthes 2017 [6] provided two future TSI/SSI scenarios ('REF' and 'EXT'). I have come across only one study that has used the 'EXT' scenario, and that study concluded that "low amplitude solar forcings such as the EXT CMIP6 or similar are not worthwhile considering during the next CMIP type of activities." [8] (are the authors aware of this reference?).While providing forcing ensembles could be useful for use in simple models [7], please consider also retaining a single recommended solar irradiance dataset for most coupled model submissions to use.
References
[1] - Gulev T et al. 2021, Changing State of the Climate System. In Climate Change 2021: The Physical Science Basis. Contribution of Working Group
I to ]the Sixth Assessment Report of the Intergovernmental Panel on Climate Change[2] - Yeo KL et al. 2017, EMPIRE: A robust empirical reconstruction of solar
irradiance variability, Journal of Geophysical Research[3] - Egorova T et al., 2018, Revised historical solar irradiance forcing,
Astronomy & Astrophysics[4] - Penza V et al., 2022, Total Solar Irradiance during the Last Five
Centuries, The Astrophysical Journal[5] - Yeo KL et al., 2020, The Dimmest State of the Sun, Geophysical Research
Letters.[6] - Matthes K et al., 2017, Solar forcing for CMIP6 (v3.2), Geoscientific
Model Development[7] - Good P et al., 2013, Abrupt CO 2 experiments as tools for predicting and
understanding CMIP5 representative concentration pathway projections, Climate
Dynamics[8] - Sedlacek J et al., 2023, Future Climate Under CMIP6 Solar Activity
Scenarios, Earth and Space Science.[9] - Coddington O et al., 2019, Solar Irradiance Variability: Comparisons of Models
and Measurements, Earth and Space ScienceCitation: https://doi.org/10.5194/gmd-2023-100-CC4
Bernd Funke et al.
Bernd Funke et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 978 | 242 | 23 | 1,243 | 5 | 5 |
- HTML: 978
- PDF: 242
- XML: 23
- Total: 1,243
- BibTeX: 5
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1