|Application the Cloud Feedback Model Intercomparison Project (CFMIP) to become a CMIP6-Endorsed MIP
Mark Webb, Chris Bretherton, Sandrine Bony, Jen Kay, Steve Klein, Pier Siebesma,
Bjorn Stevens, George Tselioudis, Masahiro Watanabe,
Peter Good, Timothy Andrews, Roger Marchand, Robin Chadwick and Hervé Douville
Updated 3rd June 2015
The primary goal of CFMIP is to inform improved assessments of climate change cloud feedbacks. However, the CFMIP approach is increasingly also being used to understand other aspects of climate response, such as regional-scale precipitation and non-linear changes.
CFMIP started in 2003 and its first phase (CFMIP-1) organised an intercomparison based on perpetual July SST forced Cess style +2K experiments and 2xCO2 equilibrium mixed-layer model experiments containing ISCCP simulator in parallel with CMIP3 (McAvaney and Le Treut, 2003). Results from CFMIP-1 had a substantial impact on the evaluation of clouds in models and in the identification of low level cloud feedbacks as the primary cause of inter-model spread in cloud feedback, and featured prominently in the fourth and fifth IPCC assessments.
The subsequent objective of CFMIP-2 was to inform improved assessments of climate change cloud feedbacks by providing better tools to support evaluation of clouds simulated by climate models and to understand cloud-climate feedback processes. CFMIP-2 organized further experiments as part of CMIP5, introducing seasonally varying SST perturbation experiments for the first time, as well as fixed SST CO2 forcing experiments to examine cloud adjustments, and idealized ‘aquaplanet’ experiments to establish the contributions of land and zonally asymmetric circulations to cloud feedback uncertainties (Bony et al., 2011). CFMIP-2 also introduced satellite simulators to CMIP via the CFMIP Observation Simulator Package (COSP), not only the ISCCP simulator, but additional simulators to facilitate the quantitative evaluation clouds using a new generation of active RADARs and LIDARs in space. Additionally CFMIP-2 introduced into CMIP5 process diagnostics such as temperature and humidity budget tendency terms and high frequency ‘cfSites’ outputs at 120 locations around the globe. CFMIP also organized a joint project with the GEWEX Global Atmospheric System Study (GASS) called CGILS (the CFMIP-GASS Intercomparison of LES and SCMs) to develop cloud feedback intercomparison cases to assess the physical credibility of cloud feedbacks in climate models by comparing Single Column Models (SCM) versions of GCMs with high resolution Large Eddy Simulations (LES) models. Additionally CFMIP-2 developed the CFMIP-OBS data portal and the CFMIP diagnostic codes repository (see http://www.cfmip.net for more details).
Early studies arising from CFMIP-2 include numerous model evaluation studies using COSP, studies attributing cloud feedbacks and cloud adjustments to different cloud types, and the finding that idealized ‘aquaplanet’ experiments without land or Walker circulations are able to capture the essential differences between models’ global cloud feedbacks and cloud adjustments. Process outputs from CFMIP have also been used to develop and test physical mechanisms proposed to explain and constrain inter-model spread in cloud feedbacks in the CMIP5 models. CGILS has demonstrated a consensus in the responses of LES models to climate forcings and identified a number of shortcomings in the physical representations of cloud feedbacks in climate models. Additionally the CFMIP experiments have, due to their idealized nature, proven useful in a number of studies not directly related to clouds, but instead analyzing the responses of regional precipitation and circulation patterns to CO2 forcing and climate change. Studies using CFMIP-2 outputs from CMIP5 remain ongoing and many further results are expected to feed into future assessments of the representation of clouds and cloud feedbacks in climate models. For a list of publications arising from CFMIP-2, please refer to the CFMIP publications page at http://www.cfmip.net.
Given the previous record of CFMIP activities and the case outlined below we would like to request that CFMIP be endorsed as a CMIP6 project to continue support for community activities in this important area of research. We provide information on our plans for CFMIP-3 structured according to the provided criteria below.
Name of MIP: The Cloud Feedback Model Intercomparison Project (CFMIP)
Co-chairs: Mark Webb firstname.lastname@example.org, Chris Bretherton email@example.com
Members of the Scientific Steering Committee: Mark Webb (Met Office), Chris Bretherton (U. Washington), Sandrine Bony (IPSL), Jen Kay (CIRES), Steve Klein (PCMDI), Pier Siebesma (KNMI), Bjorn Stevens (MPG), George Tselioudis (NASA GISS), Masahiro Watanabe (U. Tokyo)
Link to website: http://www.cfmip.net
Goal of the MIP and a brief overview: The primary goal of CFMIP is to inform improved assessments of climate change cloud feedbacks. However, the CFMIP approach is increasingly being used to understand other aspects of climate response, such as circulation, regional-scale precipitation and non-linear changes. This involves bringing climate modelling, observational and process modelling communities closer together and providing better tools and community support for evaluation of clouds and cloud feedbacks simulated by climate models and for understanding of the mechanisms underlying them. This is to be achieved by:
Ongoing organized coordinated model inter-comparison activities which include experimental design as well as specification of model output diagnostics to support quantitative evaluation of modelled clouds with observations (e.g. COSP) and in-situ measurements (e.g. cfSites) as well as process-based investigation of cloud maintenance and feedback mechanisms (e.g. cfSites, budget tendency terms, etc.)
Ongoing development and improvement of COSP and CFMIP-OBS infrastructure.
Ongoing collaboration with the cloud process modelling community (via GASS collaboration) on CGILS and via new efforts to develop a hierarchy of experiments linking GCMs with cloud resolving models (CRMs) and Large Eddy Simulation (LES) models run on large domains (e.g. via the IMPULSE project consortium).
Organising annual meetings to provide a focus for community activities relevant to CFMIP and also to the broader community working to understand changes in clouds, circulation and precipitation which impact regional projections of climate change. (These two communities are increasingly becoming connected because the experiments designed for CFMIP are also useful in addressing a broader range of questions not directly related to clouds.)
Andrews, T., (2014), Using an AGCM to diagnose historical effective radiative forcing and mechanisms of recent decadal climate change. J. Climate, 27, 1193–1209, doi:10.1175/JCLI-D-13-00336.1.
Bony, S., Webb, M., Bretherton, C. S., Klein, S. A., Siebesma, P., Tselioudis, G., & Zhang, M. (2011). CFMIP: Towards a better evaluation and understanding of clouds and cloud feedbacks in CMIP5 models. Clivar Exchanges, 56(2), 20-22.
Good, P., Andrews, T., Bouttes, N., Chadwick, R., Gregory, J. M., Lowe, J. A. (2014). The nonlinMIP intercomparison project: physical basis, experimental design and analysis principles. In preparation; (attached)
McAvaney BJ, Le Treut H (2003) The cloud feedback intercomparison project: (CFMIP). In: CLIVAR Exchanges—supplementary contributions. 26: March 2003.
Skinner, C.B., M. Ashfaq, and N.S. Diffenbaugh (2012). Influence of twenty-first-century atmospheric and sea surface temperature forcing on West African climate. J. Climate, 25, 527-542.
Stevens B., Bony S., Frierson, D.M, Jakob, C., Kageyama, M., Pincus, R, Shepherd, T., Sherwood, S., Siebesma, A. P., Sobel, A., Watanabe, M., Webb, M.J. (2014). Clouds, Circulation and Climate Sensitivity: A Grand Science Challenge. World Climate Research Programme Report No. 8/2014
We argue below the CFMIP and its proposed experiments meet the requirements laid out by the CMIP panel, as outlined below.
1. CFMIP and its experiments directly address the key science questions of CMIP6. The question that CFMIP most directly addresses is `How does the Earth system respond to forcing?’ The CFMIP emphasis on understanding cloud feedbacks makes CFMIP highly relevant to this question. The next most relevant question is `What are the origins and consequences of systematic model biases?’ CFMIP has a strong model evaluation component via the use of satellite simulators, process diagnosis and comparisons with LES, and a proven track record in investigating the link between errors in cloud processes and cloud feedbacks. CFMIP is also relevant to the question `How can we assess future climate changes given climate variability, climate predictability, and uncertainties in scenarios?’ CFMIP will continue to supplement fully coupled CMIP experiments with idealised experiments that focus on basic understanding of the dominant uncertainties associated with cloud feedbacks. This will continue to support work which relates variability on observable timescales (e. g. seasonal to decadal) to longer term climate change responses (e.g. via 'emergent constraints'). For example the amipPiForcing experiment proposed below will support studies relating cloud variability and feedbacks on observable timescales to long term cloud feedbacks (Andrews, 2014).
Note also that the WCRP Grand Challenge on Clouds, Circulation and Climate is led by two CFMIP committee members (Bony and Stevens), and has three additional CFMIP committee members on its steering committee (Webb, Siebesma, Watanabe), including one of the CFMIP co-chairs. This puts CFMIP in an excellent position to directly address the questions arising from the WCRP Grand Challenge.
2. CFMIP builds on and connects to the shared CMIP DECK and CMIP6 historical experiments. The AMIP experiment is the control simulation for the CFMIP amip4K, amip4xCO2 and amipFuture experiments which were proposed by CFMIP for CMIP5 and which we would like to see continued in CMIP6 as Tier I experiments. The proposed Tier II experiments also connect to the AMIP DECK experiment; the AMIP preindustrial forcing experiment and amip minus 4K experiments also use the DECK AMIP experiment as a control. The abrupt +/- 4% solar constant experiments build on and contrast with the DECK abrupt4xCO2 experiment, as do the abrupt4xCO2 and abrupt0.5CO2 experiments. Additionally the atmosphere-only timeslice experiments build on the abrupt4xCO2 experiment, decomposing the regional response of each model's abrupt4xCO2 run into separate responses to each aspect of forcing and warming. Additionally CFMIP will propose additional process diagnostics and simulator outputs for the CMIP6 historical experiment, which will allow process based comparisons with the AMIP experiments to assess the impact of coupled SST errors on the simulation of clouds and regional precipitation patterns in the CMIP6 models.
3. CFMIP will continue to follow the CMIP modeling infrastructure standards and conventions, in terms of experimental design, data format and documentation. CFMIP-2 experiments were organized as part of CMIP5 and the CFMIP co-chairs have demonstrated the ability to follow all of the relevant standards in experimental protocols, in specification of diagnostic output requests, data formats and documentation. We commit to continuing in this spirit for CFMIP experiments which are coordinated through CMIP6.
4. All experiments are tiered, well-defined, and useful in a multi-model context and don’t
overlap with other CMIP6 experiments.
These are outlined below, and detailed specifications are provided in the accompanying spreadsheet. They are tiered into Tiers I and II. Additionally we give guidance on other experiments currently under development which we may propose as additional Tier II experiments in the future. Alternatively these additional experiments may be coordinated outside of CMIP.
These experiments are we believe useful in the multi-model context because the common purpose that they share is a focus on understanding the inter-model uncertainty/spread in cloud adjustments and cloud feedbacks as well as that in regional precipitation and circulation change and non-linear change. Investigation of inter-model requires multi-model analysis and hence all of these experiments are useful (and in fact require) a multi-model context. The usefulness of the Tier I experiments to a number of climate researchers has already been demonstrated by the large number of publications produced using CFMIP-2 experiments.
We have checked for overlaps with other CMIP6 experiments and are confident that links with other MIPS (e.g. nonLinMIP, GeoMIP, SolarMIP, RFMIP and PMIP) are based on complementary but non-overlapping experiments.
Summary of proposed experiments
Tier I Science questions, activities and experiments
1.1 Continuation of CFMIP-2 experiments - Lead coordinator: Mark Webb (Met Office)
Science Question: What are the physical mechanisms underlying the range of cloud feedbacks and cloud adjustments predicted by climate models, and which models have the most credible cloud feedbacks?
The CMIP5/CFMIP-2 experiments and diagnostic outputs have enabled considerable progress on these questions but participation by a larger fraction of modelling groups is required in CMIP6 for a more comprehensive assessment of the uncertainties across the full multi-model ensemble. Our proposal is essentially to retain the CFMIP-2/CMIP5 experiments in Tier I for CMIP6. The experiments to be retained are amip4K, amip4xCO2, amipFuture, aquaControl, aqua4xCO2 and aqua4K. These build on the amip DECK experiment. As the output requirements for the DECK are not yet finalised, it is possible that the DECK AMIP experiment will not contain all of the output diagnostics required for CFMIP. For this reason we also request an additional CFMIP AMIP experiment including the full set of CFMIP diagnostics, both for model evaluation and for interpretation of feedbacks and adjustments in conjunction with other Tier I CFMIP experiments. If all of the proposed CFMIP diagnostics are included in the DECK experiment, this additional CFMIP AMIP experiment will not be required.
Tier II Science questions, activities and experiments
2.1 Abrupt +/-4% Solar Forced AOGCM experiments - Lead coodinators: Chris Bretherton (UW), Roger Marchand (UW), Bjorn Stevens (MPI)
Science Question: How do responses in the climate system due to changes in solar forcing differ from changes due to CO2, and is the response sensitive to the sign of the solar forcing?
Rapid adjustments in clouds and precipitation are now recognized as significant components of models’ responses to CO2 forcing. While they can easily be separated from conventional feedbacks in SST forced experiments, such a separation in coupled models is complicated by various issues, including the response of the ocean on decadal timescales. A number of studies have examined cloud feedbacks in coupled models subject to a solar forcing, which is generally associated with much smaller cloud and precipitation adjustment, due to a smaller atmospheric absorption for a given top of atmosphere forcing. Solar forcing also has a weaker impact on the stratosphere than CO2, potentially resulting in different upper tropospheric meridional temperature gradients and storm track responses.
A +4% solar experiment would be equivalent to the abrupt4xCO2 experiment but would increase the solar constant abruptly by 4 percent, resulting in a radiative forcing of a similar magnitude to that due to
CO2 quadrupling. This would provide a useful complement to the DECK abrupt4xCO2 experiment, and would support our understanding of regional responses of the coupled system with and without CO2 adjustments. A complementary -4% abrupt solar forcing experiment would allow the examination of feedback asymmetry under climate cooling, and would also help with the interpretation of model responses to geo-engineering scenarios and volcanic forcing, and relate to past climates.
2.2 Abrupt2xCO2 and abrupt0.5xCO2 Experiments (nonLinMIP) - Lead Coordinator Peter Good (Met Office Hadley Centre)
Science Question: To what extent is regional-scale climate change per CO2 doubling state-dependent (nonlinear), and why? How does the balance of mechanisms differ for high-forcing compared to low-forcing scenarios or paleoclimate simulations?
To address this question we propose two new experiments for Tier II, abrupt2xCO2 and abrupt0.5xCO2, to explore global and regional-scale nonlinear responses, highlighting different behavior under business-as-usual scenarios, mitigation scenarios and paleoclimate simulations. Additional experiments may be proposed for Tier II in the future, or coordinated via CFMIP outside of CMIP6. These include 100-year extensions to abrupt4xCO2 and abrupt2xCO2; a 1% ramp-down from the end of the 1pctCO2 experiment; an abrupt step-down to 1xCO2 from year 100 of the abrupt4xCO2. These would be used to explore longer-timescale responses, quantify nonlinear mechanisms more precisely and understand the reversibility of climate change.
2.3 amipMinus4K Experiment: Lead Coordinator: Mark Webb (Met Office)
Science Question: Are cloud feedbacks symmetric when subject to climate cooling rather than warming, and if not, why not?
An amipMinus4K experiment would take a similar form to the amip4K experiment, except that the sea surface temperatures would be uniformly reduced by 4K. This will be used to investigate asymmetric responses of clouds to a cooling climate in an idealized experiment, providing a link to PMIP. This experiment also complements the abrupt0.5xCO2 and the -4% solar experiments in that one can identify asymmetries in the warming/cooling response with and without interactions with the ocean. This experiment has been proposed for CFMIP following discussions with PMIP representatives (Pacale Braconnot, Masa Kageyama, and Masakazu Yoshimori).
2.4 Feedbacks in AMIP experiments: Lead Coordinator: Tim Andrews (Met Office)
Science question: Are climate feedbacks during the 20th century different to those acting on long term climate change and climate sensitivity?
Experiment and rationale: The previous CFMIP design was unable to diagnose time-dependent feedbacks that potentially undermine the simple linear forcing-feedback paradigm and which may be relevant to the gap between observed and modeled estimates of climate sensitivity. To address this we propose an additional experiment called ‘amipPiForcing’ (amip pre-industrial forcing), which is exactly the same as the standard amip run (i.e. SSTs and sea-ice) but run for the period 1870-present with constant pre-industrial forcings (i.e. all anthropogenic and natural forcing boundary conditions identical to the piControl run). Since the forcing constituents do not change in this experiment it readily allows a simple diagnosis of the simulated atmospheric feedbacks to observed SST changes, which can then be compared to feedbacks representative of long term change and climate sensitivity (e.g. from abrupt4xCO2 or amip4K). This has an advantage over the alternative approach of first estimating the forcing and adjustments (e.g. from RFMIP) and removing them from the amip experiment since the approach here only requires a single experiment (rather than pairs) which reduces the noise. The experiment has the additional benefit, by differencing with the standard amip run, of providing detailed information on the transient effective radiative forcing and adjustments in models relative to pre-industrial for the standard AMIP period. The inclusion of CFMIP process diagnostics not available in the RFMIP experiments will also enable a deeper understanding of the factors underlying forcing and feedback differences in the present and future climate.
2.5 Timeslice experiments for understanding regional climate responses to CO2 forcing. Co-ordinators: Rob Chadwick (Met Office) and Hervé Douville (CNRM)
How do regional climate responses (of e.g. precipitation) in a coupled model arise from the combination of responses to different aspects of CO2 forcing and warming (uniform SST warming, pattern SST warming, direct CO2 effect, plant physiological effect)?
Which aspects of forcing/warming are most important for causing inter-model uncertainty in regional climate projections?
Can inter-model differences in regional projections be related to underlying structural or resolution differences between models through improved process understanding, and could this help us to constrain the range of regional projections?
What impact do coupled model SST biases have on regional climate projections?
We propose a set of 6 20-year atmosphere-only timeslice experiments to decompose the regional responses of each model's abrupt4xCO2 run into separate responses to each aspect of forcing and warming (uniform SST warming, pattern SST change, increased CO2, plant physiological effect). As well as allowing regional responses in each individual model to be better understood, this set of experiments should prove especially useful for understanding the causes of model uncertainty in regional climate change.
The experiments are: 1) sstPi – the same as amip but with monthly-varying SSTs and sea-ice from years 101-120 of each model’s own control run rather than observed fields; 2) sstPi4K – the same as sstPi but with SSTs uniformly increased by 4K; 3) sstPi4xCO2 – the same as sstPi but CO2 as seen by the radiation scheme is quadrupled; 4) sstPi4xCO2Veg – the same as sstPi4xCO2 but with the plant physiological response also able to respond to the increased CO2; 5) sstPiFuture – the same as sstPi but a seasonally varying monthly mean climatology of the SST pattern anomaly taken from years 91-140 of each model's own abrupt4xCO2 minus piControl is scaled to have a global mean increase of 4K and applied; 6) sstPiTot – the same as sstPiFuture but also with 4xCO2 including the plant effect. sstPiTot is used to establish whether a timeslice experiment can adequately recreate the coupled abrupt4xCO2 response in each model, and then forms the basis for a decomposition using the other experiments.
We also propose an additional amip based experiment, amipTot: the same as sstPiTot but with the SST pattern anomaly climatology from sstPiFuture added instead to the observed background SSTs and sea-ice (as for other amip experiments). Comparison of amipTot and sstPiTot should help to illuminate the impact of SST biases on regional climate responses in each model, and how this contributes to inter-model uncertainty.
2.6 Atmosphere-only experiments for understanding the role of cloud-radiative effects in the large-scale atmospheric circulation in current and perturbed climates. Co-ordinators: Sandrine Bony (IPSL) and Bjorn Stevens (MPI).
How do cloud-radiative effects impact the structure, the strength and the variability of the general atmospheric circulation in the present-day climate?
How much do cloud-radiative feedbacks contribute to the spread of circulation and precipitation responses in climate change?
Can we identify robust aspects of the climate response to global warming that do not depend on cloud-radiative feedbacks?
It is increasingly recognized that clouds, and cloud-radiative effects in particular, play a critical role in the general circulation of the atmosphere (ITCZ, MJO, storm tracks, hurricanes) and its response to global warming. A better assessment of this role would greatly help interpret model biases (how much do biases in cloud-radiative properties contribute to biases in the structure of the ITCZ, in the position and strength of the storm tracks, in the lack of intra-seasonal variability, etc) and to inter-model differences in simulations of the current climate and in climate change projections (especially changes in regional precipitation and extreme events). More generally, a better understanding of how clouds couple to circulation is expected to improve our ability to answer two of the four science questions raised by the WCRP Grand Challenge on Clouds, Circulation and Climate Sensitivity: what controls the position, the strength, and the variability of the storm tracks and of the tropical rainbelts?
These questions provided the scientific motivation for the Clouds On/Off Klima Intercomparison Experiment (COOKIE) project proposed by the european consortium EUCLIPSE and CFMIP in 2012. The COOKIE experiments, which have been run by 4 to 8 climate models (depending on the experiment), consisted in switching off the cloud-radiative effects (clouds seen by the radiation code -and the radiation code only- were artificially made transparent) in an atmospheric model forced by prescribed SSTs. By doing so, the atmospheric circulation could feel the lack of cloud-radiative heating within the atmosphere, but the land surface could also feel the lack of cloud shading, which led to changes in land-sea contrasts. The change in circulation between On and Off experiments was resulting from both effects, obscuring a bit the mechanisms through which the atmospheric cloud-radiative effects interact with the circulation for given surface boundary conditions. As the LW cloud-radiative effects are felt mostly within the troposphere (and represent most of the LW+SW cloud-radiative heating) while the SW effects are felt mostly at the surface, we could better isolate the role of tropospheric cloud-radiative effects on the circulation by running atmosphere-only experiments in which clouds are made transparent to radiation only in the LW.
We propose in Tier II a set of simple experiments similar to the amip, amip4K, aquaControl and aqua4K experiments of CMIP5/CFMIP2 (and Tier 1 of CMIP6) but in which cloud-radiative effects are switched off in the LW part of the radiation code. These experiments will be referred to as offlwamip, offlwamip4K, offlwaquaControl and offlwaqua4K. The analysis of idealized (aqua-planet) experiments will allow us to assess the robustness of the impacts found in more realistic (AMIP) configurations. It will also facilitate the interpretation of the results using simple dynamical models or theories, in collaboration with large-scale dynamicists (e.g. DynVar). The comparison of the inter-model spread of simulations between AMIP and offlwAMIP experiments for present-day and globally warmer climates will help identify which aspects of the spread depend on the representation of cloud-radiative effects, and which aspects do not, thus better highlighting other sources of spread.
Additional CFMIP experiments under consideration for the future
We also propose to use these CMIP6 experiments as the foundation for further experiments planned in the context of the Grand Challenge on Clouds, Circulation and Climate Sensitivity. These will include for example sensitivity experiments to assess the impacts of different physical processes on cloud feedbacks and regional circulation/precipitation responses, and others designed to test specifically proposed cloud feedback mechanisms. Additional experiments further idealizing the aquaplanet framework to a non-rotating rotationally symmetric case are also under development. These will be proposed as additional Tier II experiments at a future time, or coordinated by CFMIP outside of CMIP6.
5. Unless a Tier 1 experiment differs only slightly from another well-established experiment, it must already have been performed by more than one modeling group. All of the proposed Tier I experiments were previously included in CMIP5 and so are well established and already performed by multiple groups.
6. A sufficient number of modelling centers (~8) are committed to performing all of CFMIP‘s
Tier 1 experiments and providing all the requested diagnostics needed to answer at least
one of its science questions. Fourteen modeling groups have so far agreed to participate in CFMIP as part of CMIP6, implying that they are prepared to perform the Tier I experiments. These are ACCESS (Australia), BCC (China), CanESM (Canada), CESM (USA), CNRM (France), FGOALS (China), GFDL (USA), IPSL-ESM (France), MIROC6-GCM (Japan) NICAM (Japan), MPI-ESM (Germany), MRI (Japan) and UKESM (United Kingdom).
7. The MIP presents an analysis plan describing how it will use all proposed experiments, any
relevant observations, and specially requested model output to evaluate the models and address its science questions. Our analysis plan is outlined below.
We commit to contributing to the creation of the CMIP6 data request and to analyzing the data, as we did for CMIP5. This will include making proposals for an updated COSP request in CMIP6 (see the proposal from the COSP PMC), and also additional improvements to the CFMIP diagnostic specifications relating to temperature and humidity budget increments, 3D radiative fluxes, inclusion of aerosol diagnostics across CFMIP experiments, and the introduction of additional locations in the cfSites specification.
We also commit to identifying observations needed for model evaluation and improved process understanding, and to contributing directly to making such datasets available as part of obs4MIPs. For example the CFMIP community has up to now played a central role in providing versions of CloudSat and CALIPSO datasets designed for direct comparison with CMIP5 data through the CFMIP-OBS website (see http://climserv.ipsl.polytechnique.fr/cfmip-obs/) and part of this work has recently involved publishing this data via the ESG and linking into obs4MIPS (see for example references to CFMIP-OBS on the obs4MIPS website at https://www.earthsystemcog.org/projects/obs4mips/aboutus). This work will continue.
CFMIP analysis activities are ongoing and the CFMIP community is ready to analyse CMIP6 data at any time. We would like modelling groups to perform the proposed CFMIP/CMIP6 experiments at the same time or shortly after their DECK experiments. Subsequent CFMIP experiments which are not included in CMIP6 will build on the proposed DECK and CMIP6/CFMIP experiments and some will start as soon as CMIP6 DECK experiments start to become available. We envisage a succession of CFMIP related intercomparisons addressing different questions arising from the GC spanning the duration of CMIP6.
We commit to scientifically analyze, evaluate and exploit the proposed experiments, and have identified leads within CFMIP for different aspects of this activity. An overview of the proposed evaluation/analysis of the CMIP DECK and CMIP6 experiments follows:
CFMIP will continue to exploit the CMIP DECK and CMIP6 experiments to understand and evaluate cloud processes and cloud feedbacks in climate models. The wide range of analysis activities described above in the context of CFMIP-2 will be continued in CFMIP-3 using the CMIP DECK and CMIP6 experiments, allowing the techniques developed in CFMIP-2 to applied to an expanding number of models, including the new generation of models currently under development. These activities will include evaluation of clouds using additional simulators (see proposal regarding COSP below), investigation of cloud processes and cloud feedback/adjustment mechanisms using process outputs (cfSites, budget tendency terms, etc). The inclusion of COSP and budget tendency terms in additional DECK experiments (e.g. abupt4xCO2 and some scenario experiments, also see proposal for COSP below) will enable the CFMIP approach to be applied to a wider range of experimental configurations. (Lead coordinator Mark Webb).
Analysis of the +/-4% solar model runs would include an evaluation of both rapid adjustments and longer-term responses on global and regional top-of-atmosphere radiative fluxes, cloud types (using ISCCP and other COSP simulators) and precipitation characteristics, as well as comparison of these responses with responses in DECK abrupt4xCO2 experiments. GeoMIP and SolarMIP have expressed a strong interest in these CFMIP experiments and joint analysis of these CFMIP experiments with GeoMIP and SolarMIP experiments is anticipated, specifically with the goal of determining to what degree results from abrupt solar forcing ONLY experiments and abrupt CO2 ONLY experiments can be used to predict what happens when both forcing are applied simultaneously, as done in the GeoMIP experiments (Lead coordinator Chris Bretherton).
Analysis of nonlinear climate processes will primarily involve comparing the abrupt4xCO2, abrupt2xCO2 and abrupt0.5xCO2 experiments over the same timescale (Good et al., 2014). (Lead coordinator Peter Good).
Analysis of amipPiForcing has already been done in detail for a single model in Andrews (2014). We propose to use this has a starting point for a multi-model analysis. (Lead coordinator Timothy Andrews).
An overview analysis of regional responses and model uncertainty in the timeslice and amipTot experiments will be carried out by the co-ordinators, in collaboration with members of contributing modeling groups. We anticipate that further detailed analysis on the processes at work in different regions will be carried out by a variety of research groups with interest and expertise in a particular region: for example a set of similar experiments has previously been used to examine the climate response of the West African monsoon in CCSM3 (Skinner et al. 2012). The timeslice and amipTot experiments have already been successfully run with HadGEM2 (Met Office), and are currently in the planning stage for CNRM. (Lead coordinator Robin Chadwick).
When analyzed together with the amip4K experiment, the amipMinus4K experiment allows one to exploit the CFMIP process diagnostics to understand for asymmetries in the climate response to warming and cooling which have been noted in PMIP experiments. These might arise from cloud phase responses in middle- and high-latitude clouds or from the adiabatic cloud liquid water path response feedback which is important over land regions and which would be expected to be weaker with cooling because of the non-linearity in the Clausius-Clapeyron relation. (Lead coordinator Mark Webb).
8. The MIP has completed the MIP template questionnaire. We have done this.
9. The MIP contributes a paper on its experimental design to the CMIP6 Special Issue. We agree to do this.
10. The MIP considers reporting on the results by co-authoring a paper with the modelling groups. We agree to do this. Separate papers will be prepared for each of the experiment groups proposed.
Answers to other questions in the MIP template questionnaire
All model output archived by CMIP6-Endorsed MIPs is expected to be made available under the same terms as CMIP output. Most modeling groups currently release their CMIP data for unrestricted use. If you object to open access to the output from your experiments, please explain the rationale. We have no objection to this.
List of output and process diagnostics for the CMIP DECK/CMIP6 data request. Please see the accompanying spreadsheet and outline below.
Any proposed contributions and recommendations for model diagnostics and performance metrics, observations/reanalysis data products, tools, code or scripts. We have provided a database of performance metrics and codes at the CFMIP Diagnostics Code Repository and a set of observational data for comparison with CFMIP outputs at the CFMIP-OBS site. Both are accessible via the CFMIP website http://www.cfmip.net. We welcome additional contributions to both of these databases.
Any proposed changes from CMIP5 in NetCDF metadata (controlled vocabularies), file names, and data archive (ESGF) search terms. None expected.
Explanation of any proposed changes (relative to CMIP5) that will be required in CF, CMOR, and/or ESGF. None expected.
CFMIP Recommended Outputs For CMIP6 DECK experiments and CFMIP experiments.
CFMIP recommends a set of diagnostic outputs for the CMIP6 DECK and CFMIP experiments which are based on those from CFMIP-2, with some modifications. These are detailed in the accompanying spreadsheet CMIP6DataRequestCompilationCFMIP_20150331.xls, and are summarized below. The recommendations are in two parts. The first part describes updates to the CFMIP process diagnostics compared to those which were requested in CMIP5, in terms of additional variables and the experiments in which they are requested. This set was drawn up by the CFMIP committee and ratified by the modeling groups following a presentation at the 2014 CFMIP meeting. The second part describes recommendations for COSP outputs in the DECK and CMIP6 Historical experiments which were drawn up by the COSP Project Management Committee (PMC). Please refer to the request scoping worksheet in the accompanying spreadsheet for a summary of which outputs are requested in which DECK experiments CFMIP experiments.
For participation in CFMIP it is required that modeling groups commit to performing all of the Tier I experiments, and sufficient diagnostic outputs to answer at least one scientific question. Since a number of the science questions of CFMIP (e.g. those pertaining to precipitation responses) require no diagnostic outputs beyond the standard ‘Amon’ outputs from CMIP5, a modeling group may qualify for participation in CFMIP even if they run the Tier I experiments without CFMIP simulators or process outputs. Such a submission would be useful, in the main for the precipitation analysis aspects of CFMIP. However we strongly recommend that participating groups additionally submit as many of the COSP and process outputs as they are practically able to, to support investigations of the full range of scientific questions of CFMIP in CMIP6.
Proposed updates to CFMIP process outputs for the CMIP DECK, CMIP6 Historical and CMIP6 CFMIP experiments.
The diagnostic request for CMIP5/CFMIP2 is summarised and motivated in the CFMIP-2 proposal document [Bony et al., 2009], and documented in detail in the CMIP5 Standard Output documentation at http://cmip-pcmdi.llnl.gov/cmip5/output_req.html in excel spreadsheet format (Worksheet ‘CFMIP output’ indicates which tables appear in which experiments and for which periods, which other worksheets such as cfMon, cfDay etc indicate the variables in each table). Our view is that the CFMIP-2 diagnostics set is fundamentally sound and forms a suitable basis for the process diagnostics in the DECK, CMIP6 Historical and CMIP6 CFMIP experiments. Thus, we present this proposal as changes with respect to the CMIP5/CFMIP-2 protocol in the accompanying spreadsheet, which includes a request scoping worksheet indicating which outputs are requested in which experiments, including the CMIP6 DECK + CMIP6 Historical experiments and the CFMIP experiments proposed within CMIP6. In the sections below we present and motivate the specific requested changes.
In this section we cite a number of peer reviewed publications. Please refer to http://www.cfmip.net -> CFMIP Publications for full references.
cfSites Outputs: The CFMIP cfSites outputs were requested in CMIP5 for 120 locations in the amip, amip4K, amipFuture and amip4xCO2 experiments, and for 73 locations along the Greenwich meridian in the aquaplanet experiments. These outputs have so far been used to evaluate the models with in-situ measurements (e.g. Nuijens et al. (submitted), Guichard et al. (in prep), Neggers et al. (submitted) and to examine cloud feedbacks on short timescales such as over the diurnal cycle (Webb et al. 2015). For CMIP6 we have dispensed with the cfSites outputs in the aquaplanet experiments, and in amipFuture, retaining them in amip, amip4K and amip4xCO2 only. At the request of the US CLIVAR ETOS WG we have added St. Helena to the list in light of upcoming field work/additional radiosondes from these islands, increasing the total number of locations to 121. (Ascension island was also requested, but this was already present).
Temperature and humidity tendency terms: CFMIP-2 requested cloud, temperature and humidity tendency terms. In CMIP6 we have omitted the cloud condensate tendency terms because no publications have arisen from those saved in CMIP5. The temperature and humidity tendency terms from CMIP5 have been widely used however. Temperature and humidity tendency terms have been demonstrably useful for understanding the roles of different parts of the model physics in cloud feedbacks, adjustments, and present-day variability (Williams et al 2013, Webb and Lock 2013, Kamae and Watanabe 2012, Demoto et al 2013, Sherwood et al 2014, Ogura et al 2014, Brient et al. (submitted), Xavier et al. (submitted)). They have also been used to understand regional warming patterns such as polar amplification in coupled models (e.g. Yoshimori et al 2013,2014). For CMIP6 we have improved the definitions of the temperature and humidity tendency terms, and added some additional terms such as clear-sky radiative heating rates to more precisely quantify the contributions of different processes to the temperature and humidity budget changes underlying cloud feedbacks and adjustments. A shortcoming of the CMIP5 protocol was that we were unable to interpret the physical feedback mechanisms in coupled model experiments due to lack of process diagnostics. For this reason we are additionally requesting these budget terms in the DECK abrupt4xCO2 experiment and the pre-industrial control.
Additional daily diagnostics: So-called ‘clustering’ approaches are now commonly used for assessing the contributions of different cloud regimes (e.g. stratocumulus, trade cumulus, frontal clouds, etc) to present day biases in cloud simulations and to inter-model differences in cloud feedbacks (e.g. Williams and Webb 2009, Tsushima et al., 2013, Tsushima et al., submitted). We have added some additional daily 2D fields to the standard package of CFMIP daily outputs to allow further investigation of feedbacks between clouds and aerosols associated with the changing hydrological cycle (aerosol loadings and cloud top effective radii/number concentrations) and a clearer diagnosis of the roles of convective and stratiform clouds (convective vs stratiform ice and condensed water paths and cloud top effective radii/number concentrations).
Proposal of request of COSP diagnostics for CMIP/DECK, CMIP6 Historical and CMIP6 CFMIP experiments
COSP Project Management Committee