FaIR calibration data

This dataset contains the full data, input scripts and produced output data for the AR6-consistent calibration of FaIRv2.1.1. The zipfile contains everything, allowing you perform your own analysis. The GitHub version contains enough for "bare bones" reproducibility, including downloading of external datasets and generation of intermediate files. The CSV file is the output most people will probably care about, containing the constrained, calibrated parameter set. <strong>FaIR v2.1.1</strong> Obtainable from https://pypi.org/project/fair/ From the command line: <pre><code>pip install fair==2.1.1</code></pre> <strong>Calibration v1.2</strong> Along with fixing a few minor things, the biggest two changes are (1) the constraints are updated to much more recent history (2022 where available) and (2) future warming from SSPs has been dropped as a constraint, typically making projected uncertainties a little wider than IPCC (though medians are still quite close). In the calibrated_constrained_parameters.csv file, the column names have been prefixed by the component of FaIR that they relate to (done in order to reduce namespace ambiguities). clim=climate, cc=carbon cycle, ari=aerosol-radiation interactions, aci=aerosol-cloud interactions, o3=ozone, fscale=forcing scaling. 1.5 million prior ensemble Climate response calibrated on 49 abrupt-4xCO2 experiments from CMIP6 and sampled using correlated kernel density estimates Methane lifetime calibrated on 4 AerChemMIP experiments for 1850 and 2014 (Thornhill et al. 2021a, 2021b). Unlike other variables which are sampled around some prior uncertainty, only the best estimate historical calibration is used. The base (1750) lifetime has been fixed and consistently used across projections. Carbon cycle uses the parameters from Leach et al. 2021 calibrated for FaIR 2.0.0 using 11 C4MIP models. Aerosol cloud interactions depend on SO2, BC and OC, using new calibrations from 13 RFMIP and AerChemMIP models, with the APRP code fixed by Mark Zelinka (Zelinka et al. 2023). Prior of -2 to 0 W/m2. Aerosol radiation interactions use prior values from AR6 Ch6, with a factor of two uncertainty for each species and a prior in the range of -0.6 to 0.0. Ozone uses the same methodology as AR6 (Smith et al. 2021b). NOx emissions are updated from RCMIP to fix a unit conversion issue, where biomass burning emissions from GFED are in units of NO and everything else (from CEDS) is in NO2. FaIR uses NO2 units consistently. In RCMIP v5.1.0 (Nicholls and Lewis 2021), the conversion was not done. When using this calibration for future scenarios, it is important that this is carefully considered. Effective radaitive forcing uncertainty follows the distributions in AR6, <strong>with asymmetric distributions switched to skew-normal.</strong> <strong>Volcanic forcing time series updated to 2022 (from IGCC).</strong> <strong>AR6_updated (v1.2.0)</strong> 1001-member posterior (deliberately chosen). Emissions and concentrations from RCMIP (i.e. based on CMIP6), but with NOx updated as above. ssp245 chosen for post-2014. Temperature from <strong>IGCC</strong> (Forster et al. 2023) (1850-<strong>2022</strong>, mean of 4 datasets). Warming <strong>2003-2022</strong> relative to 1850-1900 range from IGCC. CO2 concentrations constrained to IGCC estimate for <strong>2022</strong>. Ocean heat content from IGCC (1971-<strong>2020</strong>), linear. two step constraining procedure used: first RMSE of less than <strong>0.17K</strong> (up from 0.16), then 8-variable distribution fitting. Aerosol ERF, ERFari and ERFaci as in AR6 WG1 <strong>No future warming constraints.</strong> <strong>Performance relative to AR6 assessed ranges</strong> Assessment Model Relative diff Metric lower central upper lower central upper lower central upper ECS 2 3 5 2.00 2.96 5.01 0% -1% 0% TCRE (likely) 1 1.65 2.3 1.14 1.54 2.06 14% -7% -11% TCR 1.2 1.8 2.4 1.31 1.80 2.44 9% 0% 2% GSAT 2003-2022 rel. 1850-1900 0.87 1.03 1.13 0.87 1.03 1.13 0% 0% 0% EEU 2020 rel. 1971 356.8 465.3 573.8 358.37 465.58 573.48 0% 0% 0% ERFaer 2005-2014 rel. 1750 -2 -1.3 -0.6 -2.03 -1.32 -0.59 2% 2% -2% ERFari 2005-2014 rel. 1750 -0.6 -0.3 0 -0.60 -0.30 0.00 0% -1% ERFaci 2005-2014 rel. 1750 -1.7 -1 -0.3 -1.71 -1.01 -0.34 1% 1% 13% WMGHG ERF 2019 rel. 1750 3.03 3.32 3.61 3.03 3.29 3.58 0% -1% -1% Methane ERF 2019 rel. 1750 0.43 0.54 0.65 0.44 0.55 0.66 2% 2% 2% AF 2xCO2 (likely) 0.47 0.53 0.59 0.50 0.51 0.52 6% -4% -12% AF 4xCO2 (likely) 0.5 0.6 0.7 0.54 0.59 0.62 8% -2% -12% CO2 concentration 2022 416.2 417 417.8 416.10 416.97 417.79 0% 0% 0% SSP119 2021-2040 rel. 1995-2014 0.38 0.61 0.85 0.33 0.58 0.90 -12% -5% 6% SSP119 2041-2060 rel. 1995-2014 0.4 0.71 1.07 0.32 0.69 1.26 -19% -3% 18% SSP119 2081-2100 rel. 1995-2014 0.24 0.56 0.96 0.16 0.59 1.38 -35% 5% 44% SSP126 2021-2040 rel. 1995-2014 0.41 0.63 0.89 0.36 0.60 0.91 -12% -5% 3% SSP126 2041-2060 rel. 1995-2014 0.54 0.88 1.32 0.47 0.86 1.44 -12% -3% 9% SSP126 2081-2100 rel. 1995-2014 0.51 0.9 1.48 0.40 0.91 1.81 -21% 1% 22% SSP245 2021-2040 rel. 1995-2014 0.44 0.66 0.9 0.40 0.61 0.87 -8% -7% -4% SSP245 2041-2060 rel. 1995-2014 0.78 1.12 1.57 0.71 1.07 1.58 -9% -5% 1% SSP245 2081-2100 rel. 1995-2014 1.24 1.81 2.59 1.11 1.76 2.81 -10% -3% 8% SSP370 2021-2040 rel. 1995-2014 0.45 0.67 0.92 0.42 0.61 0.84 -7% -9% -9% SSP370 2041-2060 rel. 1995-2014 0.92 1.28 1.75 0.86 1.19 1.62 -7% -7% -7% SSP370 2081-2100 rel. 1995-2014 2 2.76 3.75 1.85 2.59 3.60 -8% -6% -4% SSP585 2021-2040 rel. 1995-2014 0.51 0.76 1.04 0.47 0.70 0.99 -7% -7% -5% SSP585 2041-2060 rel. 1995-2014 1.08 1.54 2.08 1.01 1.46 2.07 -6% -5% -1% SSP585 2081-2100 rel. 1995-2014 2.44 3.5 4.82 2.38 3.40 4.91 -3% -3% 2% Note 1: TCRE and AF are very poor constraints, taken only from model evidence, so we should not put too much weight on them. Note 2: No future warming constraint is used: these are for sense-checking only. Note 3: present-day CO2 concentration, ERFari and ERFaci are not constraints used by AR6. <strong>References</strong> Forster et al. 2023: https://doi.org/10.5194/essd-15-2295-2023 Leach et al. 2021: https://doi.org/10.5194/gmd-14-3007-2021 Nicholls and Lewis 2021: https://doi.org/10.5281/zenodo.4589756 Smith et al. 2021a: https://doi.org/10.1029/2020JD033622 Smith et al. 2021b: https://www.ipcc.ch/report/ar6/wg1/downloads/report/IPCC_AR6_WGI_FGD_Chapter07_SM.pdf Thornhill et al. 2021a: https://doi.org/10.5194/acp-21-853-2021 Thornhill et al. 2021b: https://doi.org/10.5194/acp-21-1105-2021 Zelinka et al. 2023: https://doi.org/10.5194/acp-23-8879-2023本数据集包含适配**AR6（政府间气候变化专门委员会第六次评估报告）**的FaIR v2.1.1校准所需的全部数据、输入脚本与生成的输出数据。压缩包包含所有内容，支持用户开展自定义分析。GitHub版本具备“极简可复现”能力，涵盖外部数据集下载与中间文件生成流程。CSV文件为多数用户关注的核心输出产物，包含经约束的校准参数集。 **FaIR v2.1.1** 可从https://pypi.org/project/fair/ 获取，命令行安装命令为： <pre><code>pip install fair==2.1.1</code></pre> **校准版本v1.2** 本次更新修复了若干细微问题，两大核心变更为：（1）约束条件更新至最新可用的2022年数据集；（2）不再将社会经济路径（SSPs）的未来升温作为约束条件，该调整通常会使投影不确定性略宽于IPCC报告（尽管中位数结果仍较为接近）。在`calibrated_constrained_parameters.csv`文件中，列名已添加其对应的FaIR组件前缀，以减少命名空间歧义：clim=气候系统，cc=碳循环，ari=气溶胶-辐射相互作用，aci=气溶胶-云相互作用，o3=臭氧，fscale=强迫缩放因子。 ### 校准细节 本数据集包含150万成员的先验集合： 1. 气候响应基于CMIP6的49次突发4倍CO₂试验进行校准，并通过相关核密度估计方法进行采样； 2. 甲烷寿命基于1850年与2014年的4次AerChemMIP试验进行校准（Thornhill等，2021a、2021b）。与其他围绕先验不确定性采样的变量不同，此处仅采用最优估计的历史校准结果。基准年（1750年）甲烷寿命已固定，并在所有投影场景中一致使用； 3. 碳循环采用Leach等（2021）为FaIR 2.0.0校准的参数集，基于11个C4MIP模型； 4. 气溶胶-云相互作用依赖SO₂、BC与OC物种，采用13个RFMIP与AerChemMIP模型的新校准结果，且Mark Zelinka修复了APRP代码（Zelinka等，2023），先验范围为-2至0 W/m²； 5. 气溶胶-辐射相互作用采用AR6第六章的先验值，每个物种的不确定性系数为2，先验范围为-0.6至0.0 W/m²； 6. 臭氧采用与AR6一致的方法（Smith等，2021b）； 7. NOₓ排放数据已从RCMIP更新，以修复单位转换问题：GFED的生物质燃烧排放单位为NO，而其他来源（CEDS）的排放单位为NO₂，FaIR统一采用NO₂单位。在RCMIP v5.1.0（Nicholls与Lewis，2021）中未完成该转换，因此使用该校准针对未来情景时需谨慎处理该单位差异。 有效辐射强迫不确定性遵循AR6中的分布形式，**已将非对称分布转换为偏态正态分布**。**火山强迫时间序列已更新至2022年，数据来源为IGCC**。 **AR6_updated（v1.2.0）** 本数据集包含1001个成员的后验集合（该规模为刻意选择）。排放与浓度数据来自RCMIP（即基于CMIP6），但NOₓ排放已按前述方式更新。2014年后的情景采用SSP245路径。温度数据来自**IGCC**（Forster等，2023），覆盖1850年至**2022**年，为4个数据集的平均值。相对于1850-1900年基准期的**2003-2022年**升温数据同样来自IGCC。CO₂浓度已约束至IGCC估计的**2022年**数值。海洋热含量数据来自IGCC，覆盖1971年至**2020**年，采用线性拟合处理。 本次校准采用两步约束流程：首先筛选均方根误差（RMSE）小于**0.17K**的样本（较此前的0.16K有所调整），随后进行8变量分布拟合。气溶胶有效辐射强迫、ERFari与ERFaci的定义采用AR6第一工作组报告中的标准。**未使用未来升温作为约束条件**。 ### 相对于AR6评估范围的性能对比 | 评估指标 | 模型基准范围 | | | 本数据集结果 | | | 相对偏差 | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | 下限 | 中位数 | 上限 | 下限 | 中位数 | 上限 | 下限 | 中位数 | 上限 | | ECS | 2 | 3 |5 |2.00 |2.96 |5.01 |0% |-1% |0% | | TCRE（极可能） |1 |1.65 |2.3 |1.14 |1.54 |2.06 |14% |-7% |-11% | | TCR |1.2 |1.8 |2.4 |1.31 |1.80 |2.44 |9% |0% |2% | | 2003-2022年相对于1850-1900年的全球表面气温 |0.87 |1.03 |1.13 |0.87 |1.03 |1.13 |0% |0% |0% | | 2020年相对于1971年的海洋热含量 |356.8 |465.3 |573.8 |358.37 |465.58 |573.48 |0% |0% |0% | | 2005-2014年相对于1750年的气溶胶总辐射强迫 |-2 |-1.3 |-0.6 |-2.03 |-1.32 |-0.59 |2% |2% |-2% | | 2005-2014年相对于1750年的气溶胶-辐射相互作用辐射强迫 |-0.6 |-0.3 |0 |-0.60 |-0.30 |0.00 |0% |-1% |0% | | 2005-2014年相对于1750年的气溶胶-云相互作用辐射强迫 |-1.7 |-1 |-0.3 |-1.71 |-1.01 |-0.34 |1% |1% |13% | | 2019年相对于1750年的长寿命温室气体总辐射强迫 |3.03 |3.32 |3.61 |3.03 |3.29 |3.58 |0% |-1% |-1% | | 2019年相对于1750年的甲烷辐射强迫 |0.43 |0.54 |0.65 |0.44 |0.55 |0.66 |2% |2% |2% | | 2倍CO₂下的辐射强迫（极可能） |0.47 |0.53 |0.59 |0.50 |0.51 |0.52 |6% |-4% |-12% | | 4倍CO₂下的辐射强迫（极可能） |0.5 |0.6 |0.7 |0.54 |0.59 |0.62 |8% |-2% |-12% | | 2022年CO₂浓度 |416.2 |417 |417.8 |416.10 |416.97 |417.79 |0% |0% |0% | | SSP119情景2021-2040年相对于1995-2014年的升温 |0.38 |0.61 |0.85 |0.33 |0.58 |0.90 |-12% |-5% |6% | | SSP119情景2041-2060年相对于1995-2014年的升温 |0.4 |0.71 |1.07 |0.32 |0.69 |1.26 |-19% |-3% |18% | | SSP119情景2081-2100年相对于1995-2014年的升温 |0.24 |0.56 |0.96 |0.16 |0.59 |1.38 |-35% |5% |44% | | SSP126情景2021-2040年相对于1995-2014年的升温 |0.41 |0.63 |0.89 |0.36 |0.60 |0.91 |-12% |-5% |3% | | SSP126情景2041-2060年相对于1995-2014年的升温 |0.54 |0.88 |1.32 |0.47 |0.86 |1.44 |-12% |-3% |9% | | SSP126情景2081-2100年相对于1995-2014年的升温 |0.51 |0.9 |1.48 |0.40 |0.91 |1.81 |-21% |1% |22% | | SSP245情景2021-2040年相对于1995-2014年的升温 |0.44 |0.66 |0.9 |0.40 |0.61 |0.87 |-8% |-7% |-4% | | SSP245情景2041-2060年相对于1995-2014年的升温 |0.78 |1.12 |1.57 |0.71 |1.07 |1.58 |-9% |-5% |1% | | SSP245情景2081-2100年相对于1995-2014年的升温 |1.24 |1.81 |2.59 |1.11 |1.76 |2.81 |-10% |-3% |8% | | SSP370情景2021-2040年相对于1995-2014年的升温 |0.45 |0.67 |0.92 |0.42 |0.61 |0.84 |-7% |-9% |-9% | | SSP370情景2041-2060年相对于1995-2014年的升温 |0.92 |1.28 |1.75 |0.86 |1.19 |1.62 |-7% |-7% |-7% | | SSP370情景2081-2100年相对于1995-2014年的升温 |2 |2.76 |3.75 |1.85 |2.59 |3.60 |-8% |-6% |-4% | | SSP585情景2021-2040年相对于1995-2014年的升温 |0.51 |0.76 |1.04 |0.47 |0.70 |0.99 |-7% |-7% |-5% | | SSP585情景2041-2060年相对于1995-2014年的升温 |1.08 |1.54 |2.08 |1.01 |1.46 |2.07 |-6% |-5% |-1% | | SSP585情景2081-2100年相对于1995-2014年的升温 |2.44 |3.5 |4.82 |2.38 |3.40 |4.91 |-3% |-3% |2% | #### 注释 注释1：TCRE与AF的约束性较弱，仅基于模型模拟证据，因此不应过度依赖其结果。 注释2：本次校准未使用未来升温约束条件，上述结果仅用于合理性检验。 注释3：当前CO₂浓度、ERFari与ERFaci并非AR6原报告所采用的约束条件。 **参考文献** Forster等，2023：https://doi.org/10.5194/essd-15-2295-2023 Leach等，2021：https://doi.org/10.5194/gmd-14-3007-2021 Nicholls与Lewis，2021：https://doi.org/10.5281/zenodo.4589756 Smith等，2021a：https://doi.org/10.1029/2020JD033622 Smith等，2021b：https://www.ipcc.ch/report/ar6/wg1/downloads/report/IPCC_AR6_WGI_FGD_Chapter07_SM.pdf Thornhill等，2021a：https://doi.org/10.5194/acp-21-853-2021 Thornhill等，2021b：https://doi.org/10.5194/acp-21-1105-2021 Zelinka等，2023：https://doi.org/10.5194/acp-23-8879-2023