Resources for the Future Socioeconomic Projections (RFF-SPs)

RFF-SPs Monte Carlo output data: This dataset includes the socioeconomic and emissions data generated by the Resources for the Future Socioeconomic Projections (RFF-SPs) model as discussed in Rennert et al. (forthcoming) (https://www.rff.org/publications/working-papers/the-social-cost-of-carbon-advances-in-long-term-probabilistic-projections-of-population-gdp-emissions-and-discount-rates/). The data take the form of a Monte Carlo simulation with n = 10,000 draws. File structure and column metadata are described here: --- death_rates/ --- rffsp_death_rates_run_1.feather rffsp_death_rates_run_2.feather rffsp_death_rates_run_2.feather ... rffsp_death_rates_run_999.feather rffsp_death_rates_run_1000.feather This folder contains 1,000 files in the .feather file format (https://arrow.apache.org/docs/python/feather.html), which is optimized for I/O speed and compressed to minimize storage requirements. Each file in this folder contains 3 columns: ISO3, Year, and DeathRate. For each row: - ISO3 contains the ISO numeric-3 code (https://www.iso.org/iso-3166-country-codes.html) of the country whose GDP and population are projected. - Year contains the calendar year of the predictions. - year contains the calendar year of the prediction. - DeathRate contains a death rate in average deaths per 1000 people. Baseline mortality data found in data/mortality was derived from the death_rates.csv provided to the RFF team on October 7th from Hana Sevcikova. In that source file, the column DeathRate is the annual deaths per 1000 people. PopAvg is the denominator (average between two time periods) and PopStart is population at the start of the time interval. Values are average deaths per 1000 people. There is one mortality death rate trajectory for each population trajectory, and each RFF-SP is matched to one of these 1000 trajectories. The file sampled_pop_trajectory_numbers.csv described below maps each of the 10,000 RFF SP scenarios to the baseline mortality scenario (out of 1000) matched to its population draw. --- emissions/ --- rffsp_co2_emissions.csv rffsp_ch4_emissions.csv rffsp_n2o_emissions.csv Each file in this folder contains 3 columns: sample, year, and value. For each row: - sample contains the number identifying which draw a prediction belongs to (from 1 to 10,000). - year contains the calendar year of the prediction. - value contains the projected annual global emissions of the gas specified in the filename. The units are as follows: rffsp_co2_emissions.csv is in gigatons C rffsp_ch4_emissions.csv is in megatons CH4 rffsp_n2o_emissions.csv is in megatons N2 --- pop_income/ --- rffsp_pop_income_run_1.feather rffsp_pop_income_run_2.feather rffsp_pop_income_run_3.feather ... rffsp_pop_income_run_9999.feather rffsp_pop_income_run_10000.feather This folder contains 10,000 files in the .feather file format (https://arrow.apache.org/docs/python/feather.html), which is optimized for I/O speed and compressed to minimize storage requirements. Each file corresponds to one draw of our socioeconomic data, and contains 4 columns: Country, Year, Pop, and GDP. The number in each filename corresponds to the "sample" column in the emissions data. For each row: - Country contains the ISO Alpha-3 code (https://www.iso.org/iso-3166-country-codes.html) of the country whose GDP and population are projected. - Year contains the calendar year of the predictions. - Pop contains the projected population for a given country and year, in units of thousands of people. - GDP contains the projected GDP for a given country and year, in units of millions of 2011 USD. --- The probabilistic population projections were produced by Adrian E. Raftery and Hana Ševčíková (University of Washington), using the methods described by Raftery and Ševčíková (2021). Please cite this reference in any publications using these projections. Their research was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) under NIH grant number R01 HD-070936. The probabilistic economic and emissions projections are from Rennert et al. (forthcoming), which in turn are based in part on Müller, Stock, and Watson (forthcoming). --- sample_numbers/ --- sampled_gdp_trajectory_numbers.csv sampled_pop_trajectory_numbers.csv These two files hold the sample IDs corresponding to the 10,000 draws from the dataset described above in the pop_income/ section. For reproducibility, these sample sets are available as inputs to the scripts which created them and used when the script is run in deterministic mode. More information on the weighting and specifications of such sampling is available in the code. --- ypc1990/ --- rffsp_ypc1990.csv This CSV file is a matrix of 10,000 rows, pertaining to the 10,000 draws, and 184 columns, pertaining to the 184 countries in this analysis. The columns are labeled with the ISO Alpha-3 code (https://www.iso.org/iso-3166-country-codes.html). The values are the GDP for a given country and sample, in units of millions of 2011 USD. <strong>References</strong> Müller, U.K, Stock, J.H., and Watson, M.W. (forthcoming). An Econometric Model of International Growth Dynamics for Long-Horizon Forecasting. The Review of Economics and Statistics, available online 30 October 2020. URL: https://direct.mit.edu/rest/article-abstract/doi/10.1162/rest_a_00997/97738/An-Econometric-Model-of-International-Growth Raftery, A.E. and Ševčíková, H. (2021). Probabilistic population forecasting: Short to very long-term. <em>International Journal of Forecasting</em>, available online 7 October 2021. URL: https://www.sciencedirect.com/science/article/pii/S0169207021001394 Rennert, K., Prest, B.C., Pizer, W., Newell, R.G., Anthoff, D., Kingdon, C., Rennels, L., Cooke, R., Raftery, A.E., Ševčíková, H, and Errickson, F. (forthcoming). The Social Cost of Carbon: Advances in Long-Term Probabilistic Projections of Population, GDP, Emissions, and Discount Rates. <em>Brookings Papers on Economic Activity.</em> Available online 27 October 2021. URL: https://www.rff.org/publications/working-papers/the-social-cost-of-carbon-advances-in-long-term-probabilistic-projections-of-population-gdp-emissions-and-discount-rates/RFF-SPs蒙特卡洛（Monte Carlo）输出数据集：本数据集包含由未来资源学会社会经济预测（Resources for the Future Socioeconomic Projections, RFF-SPs）模型生成的社会经济与排放数据，相关细节见Rennert等人（即将出版）的研究（https://www.rff.org/publications/working-papers/the-social-cost-of-carbon-advances-in-long-term-probabilistic-projections-of-population-gdp-emissions-and-discount-rates/）。该数据采用蒙特卡洛模拟形式，共包含n=10000次抽样。 文件结构与列元数据说明如下： --- death_rates/ --- rffsp_death_rates_run_1.feather、rffsp_death_rates_run_2.feather、rffsp_death_rates_run_2.feather……rffsp_death_rates_run_999.feather、rffsp_death_rates_run_1000.feather 该文件夹包含1000个.feather格式文件（https://arrow.apache.org/docs/python/feather.html），该格式针对I/O速度进行了优化并经过压缩以最小化存储开销。文件夹内每个文件均包含3列：ISO3、Year与DeathRate。各行字段说明如下： - ISO3：包含待预测GDP与人口的国家的ISO数字3位代码（https://www.iso.org/iso-3166-country-codes.html）。 - Year：预测对应的日历年。 - DeathRate：以每千人平均死亡数计的死亡率。 基线死亡率数据源自data/mortality目录，该数据由Hana Sevcikova于10月7日提供给RFF团队的death_rates.csv衍生而来。在该源文件中，DeathRate列代表每千人年死亡数，PopAvg为分母（两个时间段的平均值），PopStart为时间段起始时的人口数。所有数值均以每千人平均死亡数计。每个人口轨迹对应一条死亡率轨迹，且每个RFF-SP情景均与这1000条轨迹中的一条匹配。下文将介绍的sampled_pop_trajectory_numbers.csv文件，将10000个RFF-SP情景分别映射至与其人口抽样匹配的基线死亡率情景（共1000种）。 --- emissions/ --- rffsp_co2_emissions.csv、rffsp_ch4_emissions.csv、rffsp_n2o_emissions.csv 该文件夹内每个文件均包含3列：sample、year与value。各行字段说明如下： - sample：用于标识预测所属的抽样编号（取值范围为1至10000）。 - year：预测对应的日历年。 - value：对应文件名中指定气体的预测年度全球排放量。各气体的单位如下： - rffsp_co2_emissions.csv：单位为十亿吨碳（gigatons C） - rffsp_ch4_emissions.csv：单位为百万吨甲烷（megatons CH4） - rffsp_n2o_emissions.csv：单位为百万吨一氧化二氮（megatons N2O） --- pop_income/ --- rffsp_pop_income_run_1.feather、rffsp_pop_income_run_2.feather、rffsp_pop_income_run_3.feather……rffsp_pop_income_run_9999.feather、rffsp_pop_income_run_10000.feather 该文件夹包含10000个.feather格式文件（https://arrow.apache.org/docs/python/feather.html），该格式针对I/O速度进行了优化并经过压缩以最小化存储开销。每个文件对应一次社会经济数据抽样，包含4列：Country、Year、Pop与GDP。文件名中的数字与排放数据中的"sample"列相对应。各行字段说明如下： - Country：对应待预测GDP与人口的国家的ISO Alpha-3代码（https://www.iso.org/iso-3166-country-codes.html）。 - Year：预测对应的日历年。 - Pop：对应国家与年份的预测人口数，单位为千人。 - GDP：对应国家与年份的预测GDP，单位为2011年百万美元。 本数据集的概率人口预测由Adrian E. Raftery与Hana Ševčíková（华盛顿大学）基于Raftery与Ševčíková（2021）提出的方法生成。若使用此类预测开展研究，请引用该文献。该研究由尤尼斯·肯尼迪·施莱佛国家儿童健康与人类发展研究所（Eunice Kennedy Shriver National Institute of Child Health and Human Development, NICHD）通过NIH资助项目R01 HD-070936支持。 概率经济与排放预测源自Rennert等人（即将出版）的研究，该研究部分基于Müller、Stock与Watson（即将出版）的成果。 --- sample_numbers/ --- sampled_gdp_trajectory_numbers.csv、sampled_pop_trajectory_numbers.csv 这两个文件存储了pop_income/目录下所述数据集的10000次抽样对应的样本ID。为保证可复现性，这些样本集可作为生成数据集的脚本的输入，并在脚本以确定性模式运行时使用。关于此类抽样的权重与规格的更多信息可在代码中查阅。 --- ypc1990/ --- rffsp_ypc1990.csv 该CSV文件为10000行×184列的矩阵：10000行对应10000次抽样，184列对应本分析涉及的184个国家。列标签采用ISO Alpha-3代码（https://www.iso.org/iso-3166-country-codes.html）。各数值为对应国家与样本的GDP，单位为2011年百万美元。 <strong>参考文献</strong> Müller, U.K, Stock, J.H., 及 Watson, M.W.（即将出版）. 用于长期预测的国际增长动态计量模型. 《经济与统计评论》（The Review of Economics and Statistics），2020年10月30日在线发布. 链接：https://direct.mit.edu/rest/article-abstract/doi/10.1162/rest_a_00997/97738/An-Econometric-Model-of-International-Growth Raftery, A.E. 及 Ševčíková, H.（2021）. 概率人口预测：短期至超长期. 《国际预测期刊》（International Journal of Forecasting），2021年10月7日在线发布. 链接：https://www.sciencedirect.com/science/article/pii/S0169207021001394 Rennert, K., Prest, B.C., Pizer, W., Newell, R.G., Anthoff, D., Kingdon, C., Rennels, L., Cooke, R., Raftery, A.E., Ševčíková, H, 及 Errickson, F.（即将出版）. 碳社会成本：人口、GDP、排放与贴现率的长期概率预测进展. 《布鲁金斯经济活动论文集》（Brookings Papers on Economic Activity），2021年10月27日在线发布. 链接：https://www.rff.org/publications/working-papers/the-social-cost-of-carbon-advances-in-long-term-probabilistic-projections-of-population-gdp-emissions-and-discount-rates/