Input data set: National, gridded historical population (ISIMIP3b)

Annual national population data for 249 countries are based on the historic population estimates provided by the UN World Population Prospects (United Nations, 2019) until the year 2021. They are complemented by five yearly country-level future projections for SSP 1-5 (Version 2.0, December 2018; (K.C, 2020; European Commission and Joint Research Centre, 2018; based on KC and Lutz, 2017). From 2021-2030 there is a transition period where the difference between the observed data in 2021 and the SSP projections in 2030 is gradually reduced to zero. From 2020-2100, the original five-yearly SSP projections are linearly interpolated to generate yearly data.

Gridded population projections were created by downscaling the national population data described above. Therefore, we used the 30" resolution gridded population projections from the National Center for Atmospheric Research (NCAR) (Gao, 2017, 2020; Jones & O’Neill, 2016) as weighting factors and the extension of countries as defined by the ISIMIP fractional country map (Perrette 2022). Similarly, urban and rural population data from NCAR were processed at the 30" resolution and then scaled using the same factors as the total population to align with the national population data. The high resolution data was aggregated to the 0.5° x 0.5° grid. The original population data projections from NCAR are provided in 10 year intervals and were linearly interpolated to annual values.

- KC S., Lutz W. (2017) . The human core of the shared socioeconomic pathways: Population scenarios by age, sex and level of education for all countries to 2100, Global Environmental Change,
Volume 42, Pages 181-192, ISSN 0959-3780, https://doi.org/10.1016/j.gloenvcha.2014.06.004.
- K.C., S. (2020). Updated demographic SSP4 and SSP5 scenarios complementing the SSP1-3 scenarios published in 2018. IIASA Working Paper. Laxenburg, Austria: WP-20-016
- European Commission, Joint Research Centre, Demographic and Human Capital Scenarios for the 21stCentury: 2018 assessment for 201 countries, Wolfgang Lutz, Anne Goujon, Samir KC, Marcin Stonawski, Nikolaos Stilianakis (Eds.), Publications Office of the European Union, Luxembourg, 2018, ISBN 978-92-79-78024-0, doi:10.2760/41776, EUR 29113
- Gao, J. 2020. Global 1-km Downscaled Population Base Year and Projection Grids Based on the Shared Socioeconomic Pathways, Revision 01. Palisades, New York: NASA Socioeconomic Data and Applications Center (SEDAC). https://doi.org/10.7927/q7z9-9r69.
- Gao, J. 2017. Downscaling Global Spatial Population Projections from 1/8-degree to 1-km Grid Cells. NCAR Technical Note NCAR/TN-537+STR . https://doi.org/10.5065/D60Z721H.
- Jones, B., & O’Neill, B. C. (2016). Spatially explicit global population scenarios consistent with the Shared Socioeconomic Pathways. Environmental Research Letters, 11(8), 084003. https://doi.org/10.1088/1748-9326/11/8/084003
- Mahé Perrette (2022): ISIpedia country masks (v1.1). ISIMIP Repository. https://doi.org/10.48364/ISIMIP.635131.1
- United Nations, Department of Economic and Social Affairs: Population Division (2019), World Population Prospects 2019, archive [data set], https://population.un.org/wpp/Download/Archive/Standard/ (last access: 22 December 2023), 2019.

With version 1.1, we provide gridded population data, that is based on HYDE v3.3 (private communication with K. Klein Goldewijk, to be published). Just like the original dataset we provide total, rural and urban population per grid cell. The data for the historical period is identical to the corresponding ISIMIP3a data, adjusted for the different start and end years.

Care must be taken when combining the time series of gridded population data from ISIMIP3b with the historical gridded population data from ISIMIP3a. This is because inconsistencies occur at the grid-cell level during the transition from the observational period to the projections (2015-2016). These inconsistencies arise predominantly due to differences in the observational basis data sets applied in the projections provided for ISIMIP3b and the gridded historical data available for ISIMIP3a. While the gridded population projections provided by NCAR use observational data from Gridded Population of the World version 3 (GPWv3) population counts and the Global Rural-Urban Mapping Project (GRUMP v1) for urban and rural shares, ISIMIP3a gridded population time series are based on HYDE v3.3 (Klein Goldewijk et al. 2017; Kees Klein Goldewijk, personal communication, 2022). These different data sources may not perfectly agree on the historical spatial distribution of population, and any disagreements between them will also translate into mismatches between the spatial patterns of the year-2015 ISIMIP3a population data and the year-2016 ISIMIP3b population data. Additionally, the NCAR projections begin to deviate from the historical GPW estimates as early as 2005, whereas the projections under ISIMIP3b start only in 2015. This second source of inconsistencies may be less relevant as the differences should be largely mitigated by the national scaling factors.