- Journal List
- Data Brief
- v.40; 2022 Feb
- PMC8762352
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsem*nt of, or agreement with, the contents by NLM or the National Institutes of Health.
Learn more: PMC Disclaimer | PMC Copyright Notice
Data Brief. 2022 Feb; 40: 107804.
Published online 2022 Jan 6. doi:10.1016/j.dib.2022.107804
PMCID: PMC8762352
PMID: 35071702
Niklas Boke Oléna and Veiko Lehstenb,c,d,⁎
Author information Article notes Copyright and License information PMC Disclaimer
Abstract
We present a novel, global 30 arc seconds (∼1 km at the equator) population projection dataset covering each year from 2010 to 2100 that is consistent with both country level population and gridded urban fractions from the Coupled Model Intercomparison Project 6 (CMIP6). While IPCC population projections until 2100 are available at country level for Socio-Economic Pathways (SSPs), land cover (including the urban fraction) is only available for Representative Concentration Pathways (RCPs). To perform simulations of e.g., future supply and demand for agricultural products, fine scale projections of population density are needed for combinations of SSPs and RCPs. Therefore, we generated a 30 arc seconds dataset consistent with both SSPs and RCPs within the framework of the IPCC. This data set is useful in applications where spatially explicit projections of aspects of global change are investigated at a fine spatial scale. For example, if a link function between night-time lights and population density is found based on current satellite images and recent population density data, a projection of night-time light lights can be generated by using this link function with our projected population density. Such a projection can for example be used to evaluate the potential for future light pollution.
Keywords: Population, Scenarios, IPCC, Global, Modelling
Specifications Table
| Subject | Global and Planetary Change |
| Specific subject area | Population projection |
| Type of data | Gridded dataset |
| How data were acquired | Simulated |
| Data format | Annual files of population density at 1km resolution in Geotif format. |
| Parameters for data collection | The input data were selected to contain the necessary data to estimate the future population in relation to IPCC CMIP6 scenarios. |
| Description of data collection | Input data to model future population |
| Data source location | SSP population - https://tntcat.iiasa.ac.at/SspDb/ RCP urban fraction - https://luh.umd.edu/ WorldPop – Unconstrained global mosaics - https://www.worldpop.org/geodata/summary?id=24767 Roads - https://sedac.ciesin.columbia.edu/data/set/groads-global-roads-open-access-v1 Water bodies - https://sedac.ciesin.columbia.edu/data/set/grump-v1-national-admin-boundaries/data-download (created with national boundaries data) Country borders – http://thematicmapping.org/ Urban extent - https://www.researchgate.net/publication/339873537_MGUP_annual_global_2001_2018. |
| Data accessibility [1] | Repository name: DataGURU Data identification number: https://doi.org/10.18161/global_popcount.201610 Direct URL to data: https://dataguru.lu.se/app#worldpop |
| Code accessibility [2] | Repository name: GitHub Direct URL to code: https://github.com/niklasbokeolen/world_population |
Open in a separate window
Value of the Data
•
This dataset allows to generate projections (i.e. estimations of future development) of any issue related with global change which requires population density as a parameter.
•
Researchers working in global change issues but also spatial planners at state levels are the intended end-users/stakeholders of this dataset.
•
This data can be used as an input data set for studies evaluating aspects related to global change. For example, air pollution depends on the social settings in the development scenario, such as per capita income or mechanisation level, but at a fine spatial scale it also depends on population density, which has not been available before.
1. Data Description
1.1. Dataset
We generated population projections [1] for 6 combinations of SSPs and RCPs, see below. All of them can be accessed at the DataGURU server (https://dataguru.lu.se/, short link to this data set: https://dataguru.lu.se/app#worldpop) which also allows basic pre-processing such as selection of certain years, temporal averaging and spatial cropping. After selecting an SSP/RCP combination, the data can either be pre-processed by the DataGURU server or downloaded directly via ftp. The data are stored as one geotiff (.tif) file for each scenario combination and year with the datatype FLT4S and NA-value as -9999. They are in a longitude latitude projection with WGS84 as the datum. The files are named to correspond to year, SSP and RCP scenario numbers (information in brackets are replaced by year or number).
“CMPI6wp_grid_pop_count[year]_SSP[SSPnr]_ RCP[RCPnr]”.
Data ranges from 2010 until 2100 and are available for the six different SSP RCP combinations (SSP1-RCP2.6, SSP2-RCP3.4, SSP3-RCP7.0, SSP4-RCP3.4, SSP4-RCP6.0, SSP5-RCP8.5).
The naming of the datafiles will differ if advanced features of DataGURU such as cropping are used. The code used to generate the dataset is available on GitHub [2].
Fig.1 is a conceptual figure showing the steps taken to create the resulting dataset. It is a flow chart describing the method used to distribute population. Example made with artificial numbers. The smaller grid cells correspond to 30 arc second pixels and the full grid (9x) represents one 0.25-degree pixel. We also (for the sake of the example) assume that the full grid corresponds to one country. Green boxes with rounded corners indicate input data. The upper part represents the year 2010 and the lower one represents the procedure performed for the year 2011.
Fig. 1
Flow chart describing the method used to distribute population. Example made with artificial numbers. The smaller grid cells correspond to 30 arc second pixels and the full grid (9x) represents one 0.25-degree pixel, we also (for the sake of the example) assume that the full grid corresponds to one country. Green boxes with rounded corners indicate input data. The upper part represents the year 2010 and the lower one represents the procedure performed for the year 2011. For the year 2012 this process is repeated with the 2012 input data and the 2011 result data.
An example map of the population grid for year 2050 for SSP 3 and RCP 7. For visibility purposes the colour range has been restricted to be between zero and 1000, meaning that every pixel with 1000 or more people will be shown in the same way (dark red). Fig.2 has been made with the resulting dataset found at DataGURU.
Fig. 2
An example map of the population grid for year 2050 for SSP 3 and RCP 7. For visibility purposes the colour range has been restricted to be between zero and 1000, meaning that every pixel with 1000 or more people will be shown in the same way (dark red).
List of the input raw datasets used to create the population projections.
1.2. Raw data
The complete information on the raw data can be found in Table1.
Table 1
Input datasets used to grid future populations.
| Name | Spatial domain | Temporal domain | Type | Source |
|---|---|---|---|---|
| SSP population scenarios | Country | 2000-2100 | Continuous | SSP Database [7] |
| RCP Urban fraction | 0.25 degree | 2010-2100 | Raster | LUH2-v2f (luh.umd.edu) [9] |
| WorldPop | 30 arc-second | 2010 | Raster | https://www.worldpop.org/geodata/summary?id=24767ww.worldpop.org[6] |
| Roads | Global | 1980-2010 | Polylines | gROADSv1 [10] |
| Water bodies mask | Global | 2000 | Polygons | GRUMPv1:National-Administrative-Boundaries [12] |
| Country Borders | Global | 2008 | Polygons | Sandvik [13] |
| Urban Extent | Global | 2001 | Raster | MGUP [11]https://www.researchgate.net/publication/339873537_MGUP_annual_global_2001_2018 |
Open in a separate window
2. Experimental Design, Materials and Methods
Here, we present a yearly gridded population projection for 2010–2100 for the globe consistent with both the SSPs population and the RCPs urban fractions [3] based on the data for the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design. The CMIP framework is one of the fundamental elements for climate change studies into the future. It combines and coordinates the design of global climate model simulations [4]. Our work is based and updates on the African future population dataset for CMIP5 projections developed by Boke-Olén etal. [5]. The population dataset, presented in this paper, is useful for global scale studies using gridded RCP land use and climate data in relation to gridded population estimates, for example to assess future supply and demand of food.
The method is based on distributing country level Shared Socioeconomic Pathway (SSP) urban and rural population projections onto a 30 arc seconds (∼1km at the equator) grid. The method is designed to assure that the created gridded population is also conforming to the Coupled Model Intercomparison Project Phase 6 (CMIP6) Representative Concentration Pathways (RCP) urban fraction grid at 0.25°. We used global population data from the WorldPop Project [6] for year 2010 as a starting point for the population projections. The starting dataset was combined with population centre of gravity (one way to describe population centres using both spatial distribution and density [6]), distance to road, and distance to urban areas to allow each pixel to be ranked uniquely into urban or non-urban (see section below). A list of the used datasets can be found in Table1.
The future population distribution was modelled to be consistent with both the CMIP6 RCP-specific urban fraction dataset (LUH2-v2f, luh.umd.edu) [9] and the country level SSP population and urban fraction scenarios from the SSP database [8]. The RCP urban fractions (LUH2-v2f) come as six different scenario combinations which are all used within this study to create separate population projections. The scenario combinations are SSP1-RCP2.6, SSP2-RCP3.4, SSP3-RCP7.0, SSP4-RCP3.4, SSP4-RCP6.0, SSP5-RCP8.5.
The WorldPop gridded population datasets for year 2010 were used as a starting point for the population projections. The water body mask polygon dataset was converted to raster with an output resolution of 30 arc-seconds. This mask was subsequently used to remove water bodies from the population dataset. While this step was not necessary as the WorldPop dataset already contains a sufficiently fine scaled water body mask, we report here that we performed this step to allow replicating our work. While this step (where we use an older water bodies mask) might influence the projected population on single pixels (in case the two datasets disagree on the extend of water), we expect this not to have a measurable effect. Moreover, the number of people within an administrative unit is not affected. To make sure that each 30 arc-second pixel of the initial dataset was unique within each 0.25-degree pixel we added the information of distance to road, distance to urban extent and distance to centre of gravity to the initial dataset. The three added datasets were each rescaled and inversed to be between 1.0·10−5 and 1.1·10−5 were values closer to the road, urban extent, or centre of gravity had a higher value, similar to Boke-Olén etal.[5]. The created dataset of semi-unique values per 0.25-degree pixel is from now on referred to as unique population dataset.
2.1. Population distribution
To distribute the urban and rural population separately we need to separate the two classes. This is done by using the unique population datasets as the starting point together with the urban fraction (Fig.1). We rank each 30 arc-seconds pixel within each urban fraction grid cell and use this information to decide which are urban. The number of urban pixels within each 0.25-degree cell is calculated based on the urban fraction and is rounded to the nearest integer. The created urban pixels are used per country to distribute the urban and rural population values assuming the same relationship as in the unique population dataset. Using this method, we assure that the population size order between pixels remains if they are in the same country and both either urban or rural. This is illustrated in the example in Fig.1 where the two urban pixels still have the same ratio of population sizes (3:1) relative to each other after the population allotment. Once the population has been distributed for the first year, the new created dataset will be used as input for the next year thus creating a loop (Fig.1). An example of the global population dataset for SSP 3 and RCP 7 for year 2050 is displayed in Fig.2.
2.2. Uncertainty
The population projections were created using a deterministic method hence the uncertainty of the output is solely caused by uncertainty in the input data (Table1). Our method of creating a unique starting point by adding rescaled distance to road and urban extent might have introduced some bias caused by the influence on road and urban extent. However, since we rescale the values to be between 1.0·10−5 and 1.1·10−5 we argue that this effect can be neglected as this is way below the uncertainty in the input data sets. One major additional uncertainty lies in our assumption that the relationship between urban and rural population distribution which we estimate based on data from the start of the century remains valid until the end of the century. Comparable assumptions are also the basis of the input data adding additional uncertainty. These data sets (the input data as well as our data set) should always be seen as our best possible estimate. The user of these data should be aware that the projection will most likely be less accurate for later years compared to earlier years.
Ethics Statement
The work uses only publicly available data no personal or experimental data was used.
CRediT authorship contribution statement
Niklas Boke Olén: Data curation, Formal analysis, Writing – review & editing. Veiko Lehsten: Writing – review & editing, Writing – original draft.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships which have or could be perceived to have influenced the work reported in this article.
Acknowledgments
The research presented in this paper is a contribution to the strategic research area Biodiversity and Ecosystems in a Changing Climate, BECC.
References
1. Boke-Olén N., et al. University of Lund. DataGURU data dissemination server; 2021. A High Resolution Data Set for Global Future Population Developed with RCP and SSP Scenarios for Years 2010 to 2100. [CrossRef] [Google Scholar]
2. Boke-Olén, N., etal., Code for the Generation of a High Resolution Data set for Global Future Population Developed with RCP and SSP Scenarios for Years 2010 to 2100. University of Lund. https://github.com/niklasbokeolen/world_population
3. Hurtt G.C., et al. Harmonization of land-use scenarios for the period 1500–2100: 600 years of global gridded annual land-use transitions, wood harvest, and resulting secondary lands. Climatic Change. 2011;109(1-2):117–161. [Google Scholar]
4. Eyring V., et al. Overview of the coupled model intercomparison project phase 6 (CMIP6) experimental design and organization. Geosci. Model Devel. 2016;9(5):1937–1958. [Google Scholar]
5. Boke-Olén N., et al. High-resolution African population projections from radiative forcing and socio-economic models, 2000 to 2100. Sci. Data. 2017;4 [PMC free article] [PubMed] [Google Scholar]
6. WorldPop . School of Geography and Environmental Science, University of Southampton; Department of Geography and Geosciences, University of Louisville; Departement de Geographie, Universite de Namur) and Center for International Earth Science Information Network (CIESIN), Columbia University; 2018. Global High Resolution Population Denominators Project - Funded by The Bill and Melinda Gates Foundation (OPP1134076)www.worldpop.org [CrossRef] [Google Scholar]
7. Tellier Luc-Normand, Vertefeuille Claude. Understanding spatial inertia: center of gravity, population densities, the Weber problem, and gravity potential. J. Reg. Sci. 1995:155–164. 191935.1. [Google Scholar]
8. IIASA . 2015. SSP Database 2012–2015.https://tntcat.iiasa.ac.at/SspDb Available from: [Google Scholar]
9. Hurtt G., et al. Harmonization of global land-use change and management for the period 850–2100 (LUH2) for CMIP5 Geosci. Model Dev. 2020;213:5425–5464. doi:10.5194/gmd-13-5425-2020. [CrossRef] [Google Scholar]
10. Center for International Earth Science Information Network - CIESIN - Columbia University and Information Technology Outreach Services - ITOS - University of GeorgiaNASA Socioeconomic Data and Applications Center (SEDAC); Palisades, NY: 2013. Global Roads Open Access Data Set, Version 1 (gROADSv1) [Google Scholar]
11. Huang Xin, Huang Jiongyi, Wen Dawei, Li Jiayi. An updated MODIS global urban extent product (MGUP) from 2001 to 2018 based on an automated mapping approach. Int. J. Appl. Earth Observ. Geoinf. 2021;95 doi:10.1016/j.jag.2020.102255. [CrossRef] [Google Scholar]
12. 2011. Center for International Earth Science Information Network - CIESIN - Columbia University, etal., Global Rural-Urban Mapping Project, Version 1 (GRUMPv1): National Administrative Boundaries. [Google Scholar]
13. B. Sandvik, thematicmapping.org. Available from: http://thematicmapping.org/.
Articles from Data in Brief are provided here courtesy of Elsevier
