Gridded maps of global population scaled to match the 2023 Wittgenstein Center (WIC) Population projections

The gridded population data used for calculating exposed populations is based on the population projections from the original SSPs (KC and Lutz, 2017) which were subsequently gridded (Jones and O’Neill, 2016). These gridded projections were aggregated to 0.5 ° spatial resolution and then scaled to match the latest available projections for population in line with the updated SSPs, v3.0 (KC et al., 2024). The scaling is done on a country basis for all countries included in the latest SSP projections. Countries, which are not included in these projections, remain unchanged. The scaling process is done on a country-level basis using the following step: 

The total population for the original gridded data is calculated using the ISIMIP fractional country raster (Perrette, 2023), excluding border cells containing contributions from more than one country.

The population of the fractional border cells is subtracted from the total population of the SSP population projections and the required scalar to match the population from the gridded data to the SSP population projections is calculated.

This scalar is applied to all non-fractional cells.

While it is possible to calculate the scalar for each country including the proportion of the population in the fractional border cells, this would require the scalar to also be applied to that proportion of the population in the border cells to match the overall population number for each country. Applying different scalars to the population proportions for each country in the same cell would, however, change the ratios of the population in the fractional border cells and subsequently lead to skewed results when reapplying the fractional country raster to the scaled data for the aggregation to country level. For small countries, where more population lives in fractional border cells than in non-border cells, and for countries that only consist of border cells with contributions from more than one country, all cells were used in the scaling process. It should be noted that some smaller countries cannot be scaled properly and that the latest SSP population projections do not contain values for all countries. Since there has been no release of updated gridded population projections yet, the gridded population data created using this approach still provide the closest match to the latest SSP population projections currently available.

用于计算暴露人口的网格化人口数据，其基础源自KC与Lutz于2017年提出的原始共享社会经济路径（SSPs，Shared Socioeconomic Pathways）的人口预测，后续经Jones与O’Neill于2016年的研究完成网格化处理。上述网格化预测数据先被聚合至0.5°空间分辨率，随后根据KC等人2024年发布的更新版v3.0共享社会经济路径（SSPs）的最新可用人口预测数据进行缩放校准。

该缩放校准工作针对最新SSP预测中涵盖的所有国家按国别开展，未纳入该预测的国家数据则保持不变。缩放校准流程按国别执行，具体步骤如下：

基于ISIMIP分国别比例栅格数据（Perrette, 2023）计算原始网格化数据的总人口，剔除包含多国贡献的边界栅格单元。

从SSP人口预测的总人口中减去分国别边界栅格单元的人口，随后计算出使网格化数据人口与SSP人口预测匹配所需的缩放系数。

将该缩放系数应用于所有非分国别栅格单元。

尽管可以将分国别边界栅格单元的人口占比纳入各国缩放系数的计算，但此时需将该缩放系数同样应用于边界栅格单元的对应人口占比，以匹配各国的总人口数。但若对同一栅格单元内不同国别的人口占比应用不同缩放系数，则会改变分国别边界栅格单元的人口占比，后续在将分国别比例栅格重新应用于缩放后的数据以聚合至国别层面时，将导致结果出现偏差。

对于边界栅格单元内居住人口多于非边界栅格单元的小型国家，以及仅包含多国贡献边界栅格单元的国家，缩放校准流程将使用所有栅格单元。需说明的是，部分小型国家无法完成合理的缩放校准，且最新的SSP人口预测并未涵盖所有国家的相关数据。由于目前尚未发布更新版的网格化人口预测数据，采用该方法生成的网格化人口数据仍是当前可获取的、与最新SSP人口预测匹配度最高的数据集。